1.0
Views: 677Likes: 5https://www.scribd.com/doc/117960604/AjoyGhatakOptics4E2008McGrawHill536sOCR
03/21/2015
text
original
Emeritus Professor Department of Physics Indian Institute of Technology Delhi
Tata McGrawHill Publishing Company Limited
NEW DELHI
McGrawHill Offices
New Delhi New York St Louis San Francisco Auckland Bogota Caracas Kuala Lumpur Lisbon London Madrid Mexico City Milan Montreal San Juan Santiago Singapore Sydney Tokyo Toronto
Tata McGrawHill
Published by the Tata McGrawHill Publishing Company Limited, 7 West Patel Nagar, New Delhi 110 008. Copyright © 2009, by Tata McGrawHill Publishing Company Limited. First reprint 2008
DLXLCRYXRBXL:1.
No part of this publication may be reproduced or distributed in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise or stored in a database or retrieval system without the prior written permission of the publishers. The program listings (if any) may be entered, stored and executed in a computer system, but they may not be reproduced for publication. This edition can be exported from India only by the publishers, Tata McGrawHill Publishing Company Limited. ISBN13: 9780070262157 ISBN10: 0070262152 Managing Director: Ajay Shukla General Manager: PublishingSEM & Tech Ed: lrbha Mahajan Sponsoring Editor: Shalini Jha Senior Copy Editor: Dipika Dey Production Executive: Suneeta S Bohra General Manager: MarketingHigher Education & School: Michael J. Cruz Product Manager: SEM & Tech: Ed: Biju Ganesan ControllerProduction: Rajender P Ghansela Asst. General ManagerProduction: B L Dogra Information contained in this work has been obtained by Tata McGrawHill, from sources believed to be reliable. However, neither Tata McGrawHill nor its authors guarantee the accuracy or completeness of any information published herein, and neither Tata McGrawHill nor its authors shall be responsible for any errors, omissions, or damages arising out of use of this information. This work is published with the understanding that Tata McGrawHill and its authors are supplying information but are not attempting to render engineering or other professional services. If such services are required, the •assistance of an appropriate professional should be sought. Typeset at Tej=Composers, WZ391, Madipur, New Delhi 110063, and printed at Avon Printers, Plote No. 16, Main Loni Road, Jawahar Nagar, Industrial Area, Shahdara, Delhi 110 094 Cover printed at: SDR Printers
The first laser was fabricated in 1960, and since then there has been a renaissance in the field of optics. From optical amplifiers to laser physics, fiber optics to optical communications, optical data processing to holography, optical sensors to DVD technology, ultrashort pulse generation to super continuum generation, optics now finds important applications in almost all branches of science and engineering. In addition to numerous practical applications of optics, it is said that it was the quest to understand the `nature of light' that had brought about the two revolutions in science: the development of quantum mechanics started with an attempt to understand the `light quanta' and the starting point of the special theory of relativity was Maxwell's equations which synthesized the laws of electricity and magnetism with that of light. Because of all this, an undergraduate course in optics has become a `must' not only for students of physics but of engineering as well. Although it is impossible to cover all areas in one single book, this book attempts to give a comprehensive account of a large number of important topics in this exciting field and should meet the requirements of a course on optics meant for undergraduate students of science and engineering. The first edition of this book appeared in 1977 and it is indeed a matter of great satisfaction that during the past 31 years, the book has been received well by the academic community in India; the first edition also got translated into Chinese and Persian and the third edition got reprinted in Singapore. I have myself used this book several times in teaching the undergraduate course on Optics to engineering students as well as to students of Engineering Physics at ITT, Delhi and have felt the necessity of rewriting certain portions of the book. Thus, while preparing the present edition of the book, extensive revisions have been made  a few chapters have been added and a number of new topics have been introduced. New to the Edition A summary of changes in the content of the book are given below. • The book now starts with a large number of coloured photographs portraying interesting topics in Optics and a new chapter (Chapter 1 on History of Optics) that tracks the evolution of this subject. • Significant revisions made in Chapter 2 on What is Light and new figures integrated in Chapter 3 on Fermat's Principle and its Application make the presentation more clear. • Chapter 9 on Dirac Delta Function and Fourier Transform is a new chapter that covers representations of the Dirac delta function, Fourier integral theorem; and two and three dimensional Fourier transforms; these are very important for understanding the field of Fourier Optics. • In Chapter 15 on Interference by Division of Amplitude, figures have been replaced and the section on Fiber Bragg Gratings has been rewritten. Chapter 18 on Fraunhofer Diffraction: I has been rewritten with inclusion of new figures and Chapter 19 on Fraunhofer Diffraction: II and Fourier Optics is new to this edition. • Chapter 22 on Polarization and Double Refraction and Chapter 25 on The Particle Nature of Radiation have been revised with addition of new figures. • 10 additional figures for Chapter 26 on Lasers: An Introduction form the new highlights of this chapter. • Chapters 27, 28 and 29 are three new chapters, written on fiber optics and waveguide theory  an area which has, during the last 35 years, revolutionized communications. These chapters focus on Basic Concepts and Ray Optics Considerations, Basic Waveguide Theory and Concept of Modes and Single Mode Fibers. I may mention that my own research interests are in the general area of fiber optics and I have found that there are many beautiful experiments in fiber optics (not very difficult to set up) which not only allow us to understand difficult concepts but
vi
Preface to the Fourth Edition
they also find very important applications. For example, the working of a Fiber Bragg Grating is a striking application of the interference phenomenon and finds important applications in sensors and other optical devices. Similarly, the prism film coupling experiment allows us to understand the concept of quantization and the Faraday rotation in optical fibers finds important application in the industry. Further, an almost monochromatic laser pulse propagating through a special optical fiber can lead to the `awesome' supercontinuum generation. There are many such examples and in the revised edition, I have included some of them. Organization of the Book The book attempts to give a balanced account of traditional optics as well as some of the recent developments in this field. The plan of the book is as follows: • Very often, a good photograph clarifies an important concept and also makes the student interested in the subject. It is with this intention that we have started the book with a few colored photographs that describe important concepts in optics. • 'The first chapter of the book gives a brief history of the development of optics. I have always felt that one must have a perspective of the evolution of the subject that he (or she) wants to learn. Optics is such a vast area that it is extremely difficult to give a historical perspective of all the areas; as such, I may have missed names of persons who have made important contributions. Fortunately there is now a wealth of information available through the Internet; I have included a large number of references to various books and websites. I have highlighted the evolution of fiber optics and related areas, and the names of many individuals who have made important contributions to the growth of optics. • The second chapter gives a brief historical evolution of different models describing the nature of light. It starts with the corpuscular model of light and then discusses the evolution of the wave model and the electromagnetic character of light waves. Then we have a discussion of the early twentieth century experiments, which could only be explained by assuming particle nature of light, and ends with a discussion as to how we reconcile to `waveparticle duality'. • Chapters 3 to 6 are on geometrical optics. Chapter 3 starts with Fermat's principle and discusses ray tracing through gradedindex media explaining in detail the phenomena of mirage and looming, ray propagation through graded index optical waveguides and also reflection from the ionosphere. Chapter 4 is on ray tracing in lens systems and Chapter 5 is on the matrix method in paraxial optics, which is used in the industry. Chapter 6 gives a brief account of aberrations. • Chapters 7 to 12 discuss the origin of refractive index and the basic physics of wave propagation including Huygens ' principle. Many interesting experiments (like the redness of the setting sun, water waves, etc.) are discussed. The concept of group velocity and the dispersion of an optical pulse as it propagates through a dispersive medium have been discussed in detail. Self Phase Modulation (usually abbreviated as SPM), which is one of the phenomena leading to the supercontinuum generation, has also been explained. • Chapters 13 to 16 cover the very important and fascinating area of interference and many beautiful experiments associated with it  the underlying principle is the superposition principle, which is discussed in Chapter 13. Chapter 14 discusses interference by division of wave front including the famous Young 's double hole interference experiment. In Chapter 15, interference by division of amplitude is discussed which allows us to understand colors of thin films and applications like nonreflecting films, etc. The basic principle of the working of the Fiber Bragg Gratings (usually abbreviated as FBG) is discussed along with some of its important applications in the industry. In the same chapter, Michelson Interferometer is also discussed which is perhaps one of the most ingenious and sensational optical instrument for which Michelson received the Nobel Prize in Physics in 1907. Chapter 16 discusses the Fabry  Perot interferometer that is based on multiple beam interference and is characterized by a high resolving power and hence finds applications in highresolution spectroscopy. • Chapter 17 discusses the basic concept of temporal and spatial coherence. The ingenious experiment of Michelson, which used the concept of spatial coherence to determine the angular diameter of stars, has been discussed in detail. Topics like optical beats and Fourier transform spectroscopy have also been discussed. • Chapters 18, 19 and 20 cover the very important area of diffraction and discuss the principle behind topics like diffraction divergence of laser beams, resolving power of telescopes, laser focusing, Xray diffraction, Fourier optics and spatial frequency filtering, etc. • Chapter 21 is on holography giving the underlying principle and many applications. Dennis Gabor received the 1971 Nobel Prize in Physics for discovering the principle of holography.
Preface to the Fourth Edition
vii
• The next three chapters are on the electromagnetic character of light waves. Chapter 22 discusses the polarization phenomenon and propagation of electromagnetic waves in anisotropic media including first principle derivations of wave and ray velocities. Phenomena like optical activity and Faraday rotation (and its applications to measuring large currents) have been explained from first principles. In Chapter 23, starting with Maxwell's equations, the wave equation has been derived which had led Maxwell to predict the existence of electromagnetic waves and also to propound that light is an electromagnetic wave. Reflection and refraction of electromagnetic waves by a dielectric interface have been discussed in Chapter 24. Results derived in this chapter directly explain phenomena like Brewster's law, total internal reflection, evanescent waves, FabryPerot transmission resonances, etc. • Chapter 25 is on the particle nature of radiation  for which Einstein received the 1921 Nobel Prize. The chapter also discusses the Compton Effect (for which Compton received the 1927 Nobel Prize in Physics), which established that the photon has a momentum equal to h/1.. • Chapter 26 is on Lasers  a subject of tremendous technological importance. The basic physics of optical amplifiers and of lasers along with their special characteristics are also discussed. • The last three chapters are on waveguide theory and Fiber Optics  an area that has revolutionized communications and find important applications in sensor technology. Chapter 27 discusses the light guidance property of the optical fiber (using ray optics) with applications in fiber optic communication systems. The chapter also gives a very brief account of fiberoptic sensors. Chapter 28 discusses basic waveguide theory with Maxwell's equations as the starting point. The last chapter discusses the propagation characteristics of single mode optical fibers, which are now extensively used in optical communication systems. In summary, the book discusses some of the important topics that have made tremendous impact in the growth of science and technology.
Salient Features of the Book
• A large number of figures correspond to actual numerical calculations which were generated using software like GNUPLOT and Mathematica. There are also some diagrams which give a three dimensional perspective of the phenomenon. • Every chapter starts with important milestones in the area. This would give a historical perspective of the topic. • All important formulae have been derived from first principles so that the book can also be used for self study. • Numerous worked out examples are scattered throughout the book, this should help clarify difficult concepts. • Each chapter ends with , a summary of important results derived in the chapter. Problems for practice reinforce understanding of concepts. Also, references and suggested readings give the reader leads to avail more information. • Appendices include Gamma Functions and integrals involving Gaussian Functions, Diffraction of a Gaussian Beam and TE and TM modes in Planar Waveguides.
Explore the Web
A website accompanies this book to provide additional resources for learning and teaching. Supplements available in electronic form include articles on Optical Instruments for the students and Solution Manual and chapter wise PowerPoint Slides for the instructors. I dedicate this book to my students: my continuous interactions with them have led to a deeper understanding of optics. I end with the quotation (which I found in a book by G.L. Squires): I have learnt much from my teachers, but more from my pupils. To all my pupils, I owe a very special debt. I will be very grateful to receive suggestions for further improvement of the book. My email addresses are ajoykghatak@yahoo.com _and ajoykghatak@gmail.com . Acknowledgements At IIT, Delhi, I was very fortunate to have the opportunity to interact with outstanding colleagues and students, which made teaching any course, a pleasure and a challenge. We had the opportunity and freedom to modify and develop any course and present it in a form, which would make the subject more interesting. That is how the present book has evolved.
it
viii
Preface to the Fourth Edition
In this evolution, many persons have assisted me and provided valuable comments. I would first like to mention the name of my very close friend and colleague, Professor Ishwar Goyal who had used earlier editions of this book many times while ' teaching Optics at IIT, Delhi and had made numerous suggestions and constructive criticisms. I am sure he would have been very happy to see the present edition of the book but unfortunately, he is no more with us  I greatly miss my interactions with him. I am also very grateful to Professor M.S. Sodha for his constant encouragement and support. I am indebted to Professor K. Thyagarajan for continuous collaboration and also for letting me use some of his unpublished notes. I would like to express my sincere thanks to Professor Arun Kumar, Professor Lalit Malhotra, Professor Bishnu Pal, Professor Anurag Sharma and Professor K. Thyagarajan (from IIT, Delhi); Dr Kamal Dasgupta and Dr Mrinmay Pal (from CGCRI, Kolkata); Dr Rajeev Jindal and Mr Giriraj Nyati (from Moser Baer in Noida); Professor Vengu Lakshminarayanan (from University of Waterloo, Canada) and Professor Enakshi Sharma (now at University of Delhi, South Campus) for their help in writing some portions of the book. My profound thanks to Dr Gouranga Bose and Dr Parthasarathi Palai (now at Tejas Networks in Bangalore), Professor Chandra Sakher, Professor R.S. Sirohi and Dr Ravi Varshney (from IIT, Delhi); Professor Govind Swarup (from GMRT, Pune); Dr Somnath Bandyopadhyay, Dr Shyamal Bhadra, Dr Kamal Dasgupta, Dr Tarun Gangopadhyay, Ms Atasi Pal and Dr Mrinmay Pal (from CGCRI, Kolkata); Dr Suresh Nair (from NeST, Cochin); Mr Avinash Pasricha (from the US Information Service at New Delhi); Professor Ajoy Kar and Dr Henry Bookey (from Heriot Watt University, Edinburgh); Dr R.W. Terhune, Professor R.A. Phillips and Dr A.G. Chynoweth (from USA) and Dr R.E. Bailey (from Australia) for providing me important photographs that I have used in this book. I thank Mr V.V. Bhat for providing me literature on the contributions made by scientists and technologists in ancient India. I would also like to express my gratitude to my other colleagues  Professor B.D. Gupta, Professor Ajit Kumar, Professor M.R. Shenoy and Professor Kehar Singh for collaboration in research and stimulating discussions. I also thank all the authors and their publishers for allowing me to use many diagrams from their published work. I wish to express my sincere thanks to Dr V. Balasubrahmanyam (from Osmania University in Hyderabad), Dr Ramakrishna Damle (from Bangalore University), Dr R.K. Kar (from Vishwa Bharti in Shanti Niketan), Dr Sanjay Kumar (from St. Stephen's College in Delhi), Dr N. Ramamurthy (from Annamalai University in Tamil Nadu), Dr Amarjyoti Choudury (from Tezpur University in Assam) and Dr K. Porsezian (from Pondicherry University in Puducherry) for reviewing the earlier edition of the book and providing very valuable suggestions. I am grateful to Professor G.I. Opat of University of Melbourne for his invitation to attend the 1989 conference on Teaching of Optics which gave me many ideas as to how to make difficult concepts in optics easy to understand. I extend my appreciation to Dr Sunil Khijwania, Mr Monish Das and Mr Debasish Roy for their help in the preparation of the manuscript and drawing of some difficult diagrams. A part of the present revision was carried out with support from Department of Science and Technology, Government of India; I gratefully acknowledge this support. Finally, I owe a lot to my family  particularly to my wife Gopa  for allowing me to spend long hours in preparing this difficult manuscript and for her support all along.
New Delhi June 2008
Ajoy Ghatak
Preface
1. History of Optics
References 1.15
1.1
2. What is Light?
2.1 Introduction 2.1 2.2 The Corpuscular Model 2.1 2.3 The Wave Model 2.3 2.4 The Particle Nature of Radiation 2.5 2.5 Wave Nature of Matter 2.6 2.6 The Uncertainty Principle 2.7 2.7 The Single Slit Diffraction Experiment 2.8 2.8 The Probabilistic Interpretation of Matter Waves 2.9 2.9 An Understanding of Interference Experiments 2.10 2.10 The Polarization of a Photon 2.12 2.11 The Timeenergy Uncertainty Relation 2.14 Summary 2.14 Problems 2.15 Solutions 2.15 References and Suggested Readings 2.16
2.1
Part 1
Geometrical Optics
3. Fermat's Principle and Its Applications
3.1 Introduction 3.3 3.2 Laws of Reflection and Refraction from Fermat's Principle 3.3 Ray Paths in an Inhomogeneous Medium 3.8 3.4 The Ray Equation and its Solutions 3.12 3.4
3.5 Refraction of Rays at the Interface between an Isotropic Medium and an Anisotropic Medium Summary 3.21 Problems 3.21 References and Suggested Readings 3.24
3.18
4. Refraction and Reflection by Spherical Surfaces
4.1 Introduction 4.1 4.2 Refraction at a Single Spherical Surface 4.2 4.3 Reflection by a Single Spherical Surface 4.3 4.4 The Thin Lens 4.4 4.5 The Principal FOCI and Focal Lengths of a Lens 4.6 The Newton Formula 4.7 4.7 Lateral Magnification 4.7 4.8 Aplanatic Points of a Sphere 4.8
4.1
4.5
x
Contents
4.9 The Cartesian Oval 4.10 4.10 Geometrical Proof for the Existence of Aplanatic Points 4.11 The Sine Condition 4.11
Summary 4.13 Problems 4.13 References and Suggested Readings 4.14
4.10
5. The Matrix Method in Paraxial Optics
5.1 Introduction 5.1 5.2 The Matrix Method 5.2 5.3 Unit Planes 5.7 5.4 Nodal Planes 5.8 5.5 A System of Two Thin Lenses
5.1'
5.9 Summary 5.11 Problems 5.11 References and Suggested Readings 5.12
6. Aberrations
6.1 Introduction 6.1 6.2 Chromatic Aberration 6.1 6.3 Monochromatic Aberrations 6.4
Summary 6.12 Problems 6.12 References and Suggested Readings 6.13
6.1
Part 2
7.1 Introduction 7.3 7.2 Simple Harmonic Motion 7.3 7.3 Damped Simple Harmonic Motion 7.4 Forced Vibrations 7.9 7.5 Origin of Refractive Index 7.11 7.6 Rayleigh Scattering 7.15
Vibrations and Waves
7.3
7. Simple Harmonic Motion, Forced Vibrations and Origin of Refractive Index
7.7
Summary 7.16 Problems 7.16 References and Suggested Readings 7.18
8. Fourier Series and Applications
8.1 Introduction 8.1 8.2 Transverse Vibrations of a Plucked String 8.3 8.3 Application of Fourier Series in Forced Vibrations 8.4 The Fourier Integral 8.6
Summary 8.7 Problems 8.8 References and Suggested Readings 8.8 8.5
8.1
9. The Dirac Delta Function and Fourier Transforms
9.1 Introduction 9.1 9.2 Representations of the Dirac Delta Function 9.1 9.3 Integral Representation of the Delta Function 9.2 9.4 Delta Function as a Distribution 9.2 9.5 Fourier Integral Theorem 9.3 9.6 The Two and Three Dimensional Fourier Transform
Summary 9.6 Problems 9.6
9.1
9.5
Contents
xi 10.1
10. Group Velocity and Pulse Dispersion 10.1 Introduction 10.1 10.2 Group Velocity 10.1 10.3 Group Velocity of a Wave Packet 10.5 10.4 Self Phase Modulation 10.11
Summary 10.13 Problems 10.14 References and Suggested Readings 10.15
11. Wave Propagation and the Wave Equation 11.1 Introduction 11.1 11.2 Sinusoidal Waves: Concept of Frequency and Wavelength 11.3 11.3 Types of Waves 11.4 11.4 Energy Transport in Wave Motion 11.4 11.5 The Onedimensional Wave Equation 11.5 11.6 Transverse Vibrations of a Stretched String 11.6 11.7 Longitudinal Sound Waves in a Solid 11.7 11.8 Longitudinal Waves in a Gas 11.8 11.9 The General Solution of the Onedimensional Wave Equation 11.9
Summary 11:13 Problems 11.13 References and Suggested Readings 11.14
12. Huygens' Principle and Its Applications 12.1 Introduction 12.1 12.2 Huygens' Theory 12.1 12.3 Rectilinear Propagation 12.2 12.4 Application of Huygens' Principle to Study Refraction and Reflection 12.5 Huygens' Principle in Inhomogeneous Media 12,9
Summary 12.9 Problems 12.10 References and Suggested Readings 12.10_
12.1
12.3
__
Part 3
Interference
13. Superposition of Waves 13.1 Introduction 13.3 13.2 Stationary Waves on a String 13.3 13.3 Stationary Waves on a String Whose Ends are Fixed 13.5 13.4 Stationary Light Waves: Ives and Wiener's Experiments 13.6 13.5 Superposition of Two Sinusoidal Waves 13.6 13.6 The Graphical Method for Studying Superposition of Sinusoidal Waves 13.7 The Complex Representation 13.9
Summary 13.9 Problems 13.9 References and Suggested Readings 13.10
13.3
13.7
14. Two Beam Interference by Division of Wavefront 14.1 Introduction 14.1 14.2 Interference Pattern Produced on the Surface of Water 14.2 14.3 Coherence 14.5 14.4 Interference of Light Waves 14.6 14.5 The Interference Pattern 14.7 14.6 The Intensity Distribution 14.8 14.7 Fresnel's Twomirror Arrangement 14.13 14.8 Fresnel Biprism 14.14 14.9 Interference with White Light 14.15
14.1
xii
Contents
14.10 Displacement of Fringes 14.15 14.11 The Lloyd's Mirror Arrangement 14.16 14.12 Phase Change on Reflection 14.16
Summary 14.17 Problems 14.17 References and Suggested Readings 14.18
15. Interference by Division of Amplitude
15.1 Introduction 15.1 15.2 Interference by a Plane Parallel Film when Illuminated by a Plane Wave 15.2 15.3 The Cosine Law 15.3 15.4 Nonreflecting Films 15.4 15.5 High Reflectivity by Thin Film Deposition 15.7 15.6 Reflection by a Periodic Structure 15.8 15.7 Interference by a Plane Parallel Film when Illuminated by a Point Source 15.12 15.8 Interference by a Film with Two Nonparallel Reflecting Surfaces 15.14 15.9 Colours of Thin Films 15.17 15.10 Newton's Rings 15.18 15.11 The Michelson Interferometer 15.22
Summary 15.25 Problems 15.25 References and Suggested Readings 15.26
15.1
16. Multiple Beam Interferometry
16.1 Introduction 16.1 16.2 Multiple Reflections from a Plane Parallel Film 16.3 The Fabryperot Etalon 16.3 16.4 The Fabryperot Interferometer 16.5 16.5 Resolving Power 16.6 16.6 The LummerGehrcke Plate 16.9 16.7 Interference Filters 16.10
Summary 16.11 Problems 16.11 References and Suggested Readings 16.11 16.1
16.1
17. Coherence
17.1 Introduction 17.1 17.2 The Linewidth 17.3 17.3 The Spatial Coherence 17.4 17.4 Michelson Stellar Interferometer 17.6 17.5 Optical Beats 17.7 17.6 Coherence Time and Linewidth via Fourier Analysis 17.9 17.7 Complex Degree of Coherence and Fringe Visibility in Young's Doublehole Experiment 17.8 Fourier Transform Spectroscopy 17.12
Summary 17.17 Problems 17.17 References and Suggested Readings 17.18
17.1
17.10
Part 4 18. Fraunhofer Diffraction: I
18.1 Introduction 18.3 18.2 Singleslit Diffraction Pattern 18.4 18.3 Diffraction by a Circular Aperture 18.8 18.4 Directionality of Laser Beams 18.10 18.5 Limit of Resolution 18.15
Diffraction
18.3
1 Introduction 20.1 20.3 19.10 The Fourier Transforming Property of a Thin Lens 19.17 .10 Xray Diffraction 18.18 Problems 20.8 Array of Identical Apertures 19.3 22.27 18.26 18.6 Fraunhofer Diffraction by a Rectangular Aperture 19.5 Gaussian Beam Propagation 20.31 18.9 Spatial Frequency Filtering 19. Polarization and DoubleRefraction 22.20 20.4 The Fraunhofer Approximation 19.2 20.5 The Phenomenon of Double Refraction 22.10 22.6 20.1 _ 22.1 20.36 References and Suggested Readings 18.11 The Selffocusing Phenomenon 18.1 19.19 References and Suggested Readings 20.6 22.2 Production of Polarized Light 22.1 21.12 References and Suggested Readings 19.12 Optical Media Technologyan Essay 18.7 Fraunhofer Diffraction by a Circular Aperture 19.7 Nslit Fraurithofer Diffraction Pattern 18.4 19.38 19.8 20.3 Requirements 21.9 21.3 19.15 Summary 20.6 21.2 Fresnel Halfperiod Zones 20.12 Problems 19.20 18.2 Theory 21.23 18.3 Malus' Law 22.36 Problems 18.Contents 18.9 22.1 21.9 Oblique Incidence 18.7 19.4 Fresnel DiffractionA More Rigorous Approach 20.10 20.7 Diffraction of a Plane Wave by a Long Narrow Slit and Transition to The Fraunhofer Region 20.6 Interference of Polarized Light: Quarter Wave Plates and Half Wave Plates 22.1 Part 5 Electromagnetic Character of Light 22.6 Diffraction by aStraightEdge20.3 19.2 The Fresnel Diffraction Integral 19.8 The Diffraction Grating 18.1 19.4 Superposition of Two Disturbances 22.6 Twoslit Fr unhofer Diffraction Pattern 18.17 18.12 19. Holography 21.1 Introduction 19.5 19.3 21.9 References and Suggested Readings 21. Fresnel Diffraction 20.13 22.33 Summary 18.6 Summary 21.6 19.10 Summary 19.3 Uniform Amplitude and Phase Distribution 19. Fraunhofer Diffraction: II and Fourier Optics 19.8 Problems 21.4 Some Applications 21.3 The Zoneplate 20.1 Introduction 21.4 20.5 Fraunhofer Diffraction by a Long Narrow Slit 19.1 Introduction 22.
2 Reflection at an Interface of Two Dielectrics 24.5 Angular Momentum of a Photon Summary 25.4 25.5 Energy Density and Intensity of an Electromagnetic Wave 23.9 Change in the SoP (State of Polarization) of a Light Beam Propagating Through an Elliptic Core Single Mode Optical Fiber 22.17 References and Suggested Readings 24.20 22.39 Contents s 23.18 24.5 23. Reflection and Refraction of Electromagnetic Waves 24.4 The Poynting Vector 23.13 Photons 25.7 The Wave Equation in a Conducting Medium 23.15 Faraday Rotation 22.34 Summary 22.4 23.1 Introduction 25.1 24.1 Introduction 26.15 23.12 Summary 23.9 Physical Significance of Maxwell's Equations 23.11 Rochon Prism 22.36 Problems 22.10 25.3 26.2 The Fiber Laser 26.14 24.15 Summary 24.1 24.1 Maxwell's Equations 23.13 References and Suggested Readings Part 7 26.26 22.4 Reflectivity of a Dielectric Film 24.21 22.3 Reflection by a Conducting Medium 24.4 The Photon Mass 25.3 The Threedimensional Wave Equation in a Dielectric 23.3 25.2 Plane Waves in a Dielectric 23.10 25.10 23.12 Plane Wave Propagation in Anisotropic Media 22.6 25.24 22.4 25.32 22.3 The Compton Effect 25.16 Theory of Optical Activity 22.11 Lasers & Fiber Optics 26.9 26.10 Wollaston Prism 22. The Particle Nature of Radiation 25.8 Optical Activity 22.2 The Photoelectric Effect 25.13 Ray Velocity and Ray Refractive Index 22.7 Analysis of Polarized Light 22.14 Problems 23.8 The Continuity Conditions 23.2 23.6 Radiation Pressure 23.33 22.xiv 22.1 24.11 23. Electromagnetic Waves 23.3 The Ruby Laser 26.25 22.14 Jones Calculus 22.9 23.1 Introduction 24.12 Problems 25.8 23. Lasers: An Introduction 26.16 Problems 24.37 References and Suggested Readings 22.3 .1 Part 6 25.22 22.30 22.1 23.14 References and Suggested Readings 23.
8 Typical Parameters for a Ruby Laser 26.3 Guided Modes of a Step Index Fiber 29.9 The Attenuation Limit 27.2 28.9 Monochromaticity of the Laser Beam 26. 1 I.9 27.1 Appendix A: Gamma Functions and Integrals Involving Gaussian Functions Appendix B: Evaluation of the Integral Appendix C: Diffraction of a Gaussian Beam Appendix D: TE and TM Modes in Planar Waveguides Name Index Subject Index A.5 Optical Resonators 26.8 27.7 28. 1 B.16 References and Suggested Readings 27.2 Te Modes of a Symmetric Step Index Planar Waveguide 28.6 Dispersion Compensating Fibers 29.9 Problems 29.4 The Optical Fiber 27.1 Introduction 27.10 References and Suggested Readings 28.8 Attenuation in Optical Fibers 27.30 Problems 26.14 27.18 27.25 26.7 The Lineshape Function 26.24 26. 1 D.12 29.27 Summary 26.6 29.5 Pulse Dispersion in Single Mode Fibers 29.1 Introduction 28.6 27.31 References and Suggested Readings 26.7 29.32 27.15 Problems 27. 1 I.2 27.14 26.10 Raman Amplification and Raman Laser 26.11 28.12 Fiber Optic Sensors 27.4 Single Mode Fiber 29.26 xv 26.2 Some Historical Remarks 27.4 27.7 The Numerical Aperture 27.26 26..5 Why Glass Fibers? 27.1 29.Problems 28. Fiber Optics III: Single Mode Fibers 29.3 27.2 27.1 29.11 Dispersion and Maximum Bit Rates 27.10 Pulse Dispersion in Multimode Optical Fibers 27. 1 C.3 Total Internal Reflection 27.3 29.7 27.4 Te Modes of a Parabolic Index Planar Waveguide 28.13 26..8 . 1 . Fiber Optics II: Basic Waveguide Theory and Concept of Modes 28.11 28. Fiber Optics I: Basic Concepts and Ray Optics Considerations 27.7 27.3 Physical Understanding of Modes 28.8 28.2 Basic Equations 29.6 Waveguide Theory and Quantum Mechanics 28.1 28.6 Einstein Coefficients and Optical Amplification 26.6 The Coherent Bundle 27.Contents 26.12 References and Suggested Readings 29.5 28.11 27.1 Introduction 29.4 The HeNe Laser 26.5 Tm Modes of a Symmetric Step Index Planar Waveguide 28.1 29.
.
.
.
.
.
.
.
.
.
.
.
physics and mechanics at the University of Alexandria. mathematician. also known as Euclid of Alexandria. Vasco Ronchi in his famous book On The Nature of Light has written: Today we tend to remember only Newton and Huygens and consider them as the two great men who laid the foundations of physical optics. Portrait from Ref.org . 3. astronomer and statesman. Euclid's Optics was translated as the Kitab Uqlidis Fi Ikhtilaf AlManazir. fortunately. the eye. 1 and 2).347 BC) was a Greek philosopher. It seems that Euclid 's work on optics came to the West mainly through medieval Arabic texts. _He believedthat visioninvolvesraysgoingfrom the eyes tothe object seen and he studied the relationship between the apparent sizes of objects and the angles that they subtend at Justus van Ghent's 15thcentury depiction of Euclid. was a Greek mathematician who was born between the years of 320 and 324 BC. reflection and the use of mirrors.. at the end of the chapter. Archytas (428 BC .wikipedia. Euclid. . Most figures in this chapter have been adapted from www. 10 . see the book review of The Arabic Version of Euclid's Optics in Ref. You will find many references from the Internet. 46. Genesis Optics is the study of light that has fascinated mankind ever since he could see. 7. In reality. It is said that he had propounded the idea that vision arises as the effect of an invisible `fire' emitted from the eyes so that on encountering objects it may reveal their shapes and colors (see Ref. apart from the ones given. He wrote Catoptrica which described the propagation of light... Hero (or Heron) of Alexandria (c. No likeness or description of Euclid's physical appearance made during his lifetime survives. It is with this perspective that I thought it would be appropriate to give a very brief history of the development of optics.70 AD) 'lived in Alexandria. the discussion on the nature of light was fully developed even before these two men were born.there is a wealth of information that is now available through the Internet. For those who want to know moreof thehistory. The cover of the Latin translation of Euclid's OPTICS (1458) is from Ref. Let there be light: and there was light. In his Optica (about 300 BC) he noted that light travels in straight lines and described the law of reflection.And God said. This is not really true and perhaps this tendency is due to the distance in time which as it increases tends to strengthen the contrast and to reduce the background. Roman Egypt and was a teacher of mathematics.
Witelo wrote an exhaustive tenvolume work on optics entitled Perspectiva which was Witelo of Silesia (1230. 11. This treatise discussed concave and convex mirrors in both cylindrical and spherical geometries. Alhazen is considered as 'father of optics' because of the tremendous influence of his Book of Optics (Arabic: Kitab alManazir. Part V of Roger Bacon's "Opus Majus" is practically an annotation to Ibn al Haitham's Optics. anticipated Fermat's law of least time. In the 5a' century. caws ^ a fti p urn v^ RI14eo 1*e4G o 6ta*tao. In this he was anticipating Fermat's Principle of Least Time by many centuries.. . Erazmus Ciolek Witelo (born ca. 8). in Latin. Alhazen had also studied the reverse image formed by a tiny hole and indicated the rectilinear propagation of light. see References 1215. 168 AD) known in English as Ptolemy. Turingorum et Polonorum filius meaning `a son of Poland and Thuringia'. which study led Alhazen to the belief that light consists Optics of rays which originate in the object seen. some discussion on Alhazen's work can be found in Ref.000dinar note (Ref. in passing through a medium. a view contrary to that of Euclid and Ptolemy. He made experimental contributions of the highest order in optics. 10).2 Claudius Ptolemaeus (ca. 1230 and died around 1275) was a theologian. Elliot wrote the following about the book: Alhazen was one of the ablest students of optics of all times and published a sevenvolume treatise on optics which had great celebrity throughout the medieval period and strongly influenced Western thought. 90 ca.g. He enunciated that a ray of light. physicist. 15.550) is the first of the great mathematicianastronomers of the classical age of Indian mathematics and Indian astronomy.r th u ti u l lew :ii6i#^hd:oatb^ " . and not in the eye.1276) largely based on the work of Ibn alHaytham and served as the standard text on the subject until the seventeenth century. The adjacent painting ig the impression of Ptolemy by a medieval artist (Ref. He also measured the angle of refraction in water for different angles of incidence and made a table of it. e. In it. and considered refraction and the magnifying power of lenses. uta.1. According to the ancient Greeks. . notably that of Roger Bacon and Kepler. the eye was assumed to be a source of light. refraction and colour. was a Hellenistic mathematician and astronomer who lived in Roman Egypt. It contained a remarkably lucid description of the optical system of the eye. Latin: De Aspectibus or Perspectiva). Witelo called himself. Ibn alHaytham (9651039) (often called as Alhazen) was born in Basra. 10). this was also assumed by the early Indian philosophers. 9). takes the path which is easier and `quicker'. Robert S. The above statue is on the grounds of IUCAA. Cover of lrtellonis Thuringopoloni opticae libri decem (Ten Books of Optics by ThuringoPole Witelo) Ref. Aryabhatta (AD 476'. To quote Nobel Prize winning physicist Abdus Salam: Ibn'alHaitham was one of the greatest physicists of all time. Ref.. Pune in India (Ref. Aryabhatta reiterated that it was light arr$ing from an external source at the retina that illuminated the world around us.uuta . Ptolemy's Optics is a work which survives only in a poor Arabic translation and in Latin translation of the Arabic. There are many books written on the work of Alhazen (see. . natural philosopher and mathematician. 34131 0 Ct)..ciKI 4Pfii. including reflection. . he wrote about properties of light.t. The adjacent photograph is on an Iraqi 10. Iraq (Mesopotamia).
he used a telescope as a compound microscope." Reviewing the book. there is lot of controversy on that. and the mathematics of refraction. and persuasive power of Kepler's arguments.8 January 1642) is often referred to as the `father of modem physics'. 1619) was a Dutch eyeglass maker. Johannes Kepler (December 27. Galileo Galilei (15 February 1564 . the camera obscura was published in the book De radio astronomico et geometrico liber. The result is a work of extraordinary breadth whose significance transcends most categories into which it might be placed.trologer. 1452 May 2. quim ut et* cohtingitihoceltfiinzaiofup riatpan dehvfi! ut rails apparebitinfurkr ficerevtraiioeiigitoptica. Lippershey saw two children. .History of Optics 1. Safi 114. and a key figure in the 17th century astronomical revolution. In 1545. astronomer and as. playing_with _lenses _in_hisshop. Galileo was among the first to use a refracting telescope as an instrumentto observe stars and planets. 20) says "Optics was a product of Kepler's most creative period._19). who used it to study perspective. Gemma Frisius described his observation of the solar eclipse in 1544 (see Refs 17 and 18). In 1609. The announcement (see Ref. 16). . (a) Self portrait of Leonardo da Vinci (April 15. hepublished_ the_ book entitled Ad Vitellionem paralipomena. David Lindberg writes: In this book Donahue has performed service of enormous value to Kepler scholars and historians of early optics. 1630) was a German mathematician. imKnimust defiant Sit nosenai Anno Loumit pauld p us Viz. Some people believe that it wds Leonardo da Vinci who had first observed diffraction.‘ 24fl i (ha a f44. Many historians believe that in 1608. 1571 = November 15. obfenuuimus`. In 1610. The photograph is that of a GDR stamp featur_ingJohannes Kepler_(Ref. rigor. however. the physiology of vision. Many historians believe that Lippershey also invented the compound microscope. This inspired Lippershey to the creation of the first telescope. the first detailed description of the pinhole camera (camera obscura) was given in the manuscript Codex atlanticus (c. This appears to be the first clearly documented use of the compound microscope. see Refs 21 and 22. Some historians credit Galileo Galilei for the invention of the first telescope. 1485) by Leonardo da Vinci. In1604. and discovered that images were clearer when seen through two lenses. 1519) around 1514 and (b) his famous Mona Lisa painted somewhere between 1503 and 1505 (Ref. It began as an attempt to give astronomical optics a solid foundation. quibus Astronomiae pars optica traditur. the German physicist and mathematician. An English translation Hans Lippershey (1570 Sept. but also reveals the clarity. but soon' transcended this narrow goal to become a complete reconstruction of the theory of light. In the above diagram. Mum in tabula perradio:Solis. Although Alhazenhad studied the reverse image formed by a tiny hole. and he made improved microscopes in 1623 _ _ and after. Donahue) has recently been published as Johannes Kepler Optics. His lucid translation of the difficult Latin of Kepler's great optical treatise not only affords ready access to Kepler's optical achievement.3 (by William H.
and presented in detail in his book Opticks (1704). Clerselier. 32 and is not authenticated.. 24.4 Optics the corpuscular model. writer. FRS (July 18. he discovered the law of refraction that is referred as the Snell's law.treatise which attracted Isaac Newton to the study of optics. 1650) was a highly influential French philosopher. It was this. for more details and additional references. that nature always acts by the shortest and simplest path. 23). 1663). 1703). The adjacent portrait is taken from Ref. not a physical oneit is not and can not be the cause of any effect in nature.1. In 1669. the explanation came later. A replica of the earliest surviving telescope attributed to Galileo Galilei. Fermat showed that the law of refraction can be deduced by assuming that the path of a refracted ray of light was that which takes the least time! Fermat's Principle met with objections. Rene Descartes (March 31. Around 1660. Francesco Maria Grimaldi (April 2. Descartes. In May 1662.4 Nov 1698) was a Danish scientist. scientist and. In 1621.March 3. 1665) was a French mathematician and never went to college (Ref. 1601January 12. gave the fundamental laws of propagation of light: the laws of reflection and refraction. 1662). In his 1664 book Micrographia. Robert Hooke observed diffraction in 1672.December 28. Hooke had also observed the colours from thin sheets of mica which was much later explained through interference of light. 29. Robert Hooke was the first to describe `Newton 's rings'. He interpreted the phenomenon by stating that light had to consist of a very fine fluid of some sort in a state of constant vibration. He formulated a geometrical basis for a wave theory of light in his Physicomathesis de lumine (1666). For more details see. 2629). on display at the Griffith Observatory (Ref. an expert in optics. wrote: The principle you take as a basis for your proof to wit. In Ref. In a letter to Cureau de la Chambre (dated January 1. 1596 . Robert Hooke. regarding lumen as a swarm of spherical corpuscles(see Ref. e. Refs 30 and 31. in his book entitled Dioptrique (1638). Willebrord Snel van Royen (15801626) was a Dutch astronomer and mathematician. 1635 . which means `breaking up'. Newton deals with the diffraction problems of Grimaldi in Part III of his Opticks (1704). 25). Pierre de Fermat (August 17. He laid the groundwork for the later invention of `diffraction grating' . The rings are named after Newton because Newton explained it (incorrectly) in a communication to the Royal Society in December 1675. Rasmus Bartholin (Latinized Erasmus Bartholinus. mathematician. 13 Aug 1625 . is only a moral principle. he discovered double refraction of a light ray by calcite and wrote a 60page memoir about the results.g. Grimaldi discovered the diffraction of light and gave it the name `diffraction'. He also put forward .February 11. it has been shown that: Descartes' insightful derivation of Snell's law is seen to be largely equivalent to the mechanicalparticle or corpuscular derivation often attributed to Newton (who was seven years old at Descartes' death). see Ref. 1618 .
Rendered into English by SILVANUS P. 1773 May 10. 36 Ole Christensen Romer (25 September 1644 .Thomas Young used his wave theory to University of Chicago Press The entire book can be read at the website given in Ref. The preface by Huygens ends with: TREATISE ON LIGHT in which are explained the causes of that which occurs OPTICKS or:. withstood the test of time: and even now the exquisite skill with which he applied his conception of the propagation of waves of light to unravel the intricacies of the phenomena of the double refraction of crystals. a Christiaan Huygens (April 14.measurements. . on dispersion.fupttficieeadem deoraim target.Emit obJecli iw u Eansad EE altcra U in11g. THOMPSON The front covers of the first edition (1704) and of the fourth edition (1730) of OPTICKSor A Treatise of the Reflections.a Dutch mathematician. 1629 .a1MLbtTUM IX. In 1672.19 September 1710) was a Danish astronomer who in 1676 made the first quantitative . however. In 1801. 36.ors.ofthespeedoflight. will excite the admiration of the student of Optics. Rcfac7iaus. s or RLtLEXIONS. y Prirnu oligaa den6or dam circomvalvirur ineumlwm. t.5 Sir Isaac Newton (4 January 1643 . & t. pr Adeanbm apparcr imu a duabusbe ineon• bifoc rma I:XPF. 40. and I shall owe much in return to him who shall be able to supplement that which is here lacking to me in knowledge. noar pars FC1ir nbkmioLpcr Ye Rcaiore. propagating in. 1829) was an English scientist. Thomas Young (June 13.dnnnev.di. rprxfcnariin riperli• obiaOll • SP rn Ill`. limefpecieram. Hooke had earlier observed the colours from thin sheets of mica . he produced his famous book on optics: Traite de la Lumiere.). & In REFRACTION and particularly in the strange REFRACTION OF ICELAND CRYSTAL by CHRISTIAAN HUYGENS Finally.a straight line (see Ref. or the separation of light into a spectrum of its component colours. Thomson) in his Prefacewrites: The Theatise on Light ofHuygens has. Inflections and Colours of Light by Sir Isaac Newton. 34). 37.Newton explained this by saying that the needle exerts a force that `pulled' the light from the straight line path. The 4th edition of the book is available as a Dover reprint (Ref. there remains much more to be investigated touching the nature of Light which I do not pretend to have disclosed. Emsma.lc Early( &Tear. Huygens gave the theory of double refraction which was discovered by Bartholinus_ in 1669. the English translation of the book is now available as a Dover reprint (Ref. Ninmm & arirm aculnruri.y5TUD1 Figures. Curvilinear by Sir ISAAC N. The book ' s first page is shown below.VI'ON. U to lima of A. 38). houa. 35) and the entire book can be read at the website given in Ref. In 1678. In REFLEXION. pan dilmiur. astronomer and physicist.tra 15 injollt4 REFRACT 10 detcgimr.fe: wn=row lvx.J COLOURS TREATISE {LIG H T. In 1690. he proposed the wave theory of light. L I G H. . REFRACTIONS INFLEXIONS .Hibtory of Optics Pages 1 and 14 of Erasmus Barthollnus Opus (Ref. EXPERIMENTA CRYSTALLI ISLANDICI DISDIACLAS'I•ICI Q. 33) Minsk' BARTaor.Newton explained this by `fits of easy transmission and reflection' of the light rays. res.elper Mimi corpus pPeeiuodmn vidcunu. Refractions. he made a systematic study of light and published it in the form of a book in 1704. first reported in 1672.bm e.July 8. In this book. Macm.evanr suemiuta OF & CE.. Two TREATISES {S PI CI LS coil 4PTLCKS: O R•A TREATISE or T IIE Reflcl7iars. Ieu. The translator (S.r . Fonnn. animadvcaenmsio tetm Mier. cli prseedente. Am. In addition to his numerous contributions to science and mathematics. in a communication to the Academie des Science in Paris. Inflellionr and Colours. and in particular demonstrated how waves might interfere to form a wavefront. AI AG. this experiment is considered one of 10 most beautiful experiments in physics (Refs 41and42)..31 March 1727) is considered as one of the greatest figures in the history of science (for more details. ronrp. 1695). 39) and also in the website given in Ref. ' for more details see Ref. alievandonpparebir um. Grimaldi had earlier observed light entering the shadow of a needle . ' The Hague. Newton describes his experiments. see Ref. was .. spa A.n0N. Young demonstrated the wave nature of light through a simple twohole interference experiment.E. and of the refraction of the atmosphere.inf 1. BAara GLint mothm.. A :. P. The 8 January 1690.
the plane of polarization of linearly polarized light (propagating through a material medium) gets rotated by the application of an external magnetic field aligned in the direction of propagation. physicist. 47. He also invented the diffraction grating and demonstrated the accurate measurement of the wavelength. In the photograph above. and relating colour to wavelength. . In 1859. was a French physicist. he calculated the approximate wavelengths of the seven colours recognized by Newton. Fraunhofer invented the spectroscope and discovered 574 dark lines appearing in the solar spectrum. 1853) was a French mathematician. 1786 October 2. for more details see Ref. 49 and 50). This relation subsequently was one of the four equations of Maxwell and is referred to as Faraday's law. Poisson deduced from Fresnel's theory the necessity of a bright spot at the center of the shadow of a circular opaque obstacle. 1765 .. 1826) was a German optician. Optics Joseph Nicephore Niepce (March 7.g. at Paris. In 1823. 1867) contributed significantly to the fields of electromagnetism and electrochemistry. as substitutes for mirrors. Fraunhofer published his theory of diffraction. Poisson had hoped to disprove the wave theory. By the year 1821. Faraday discovered the phenomenon that is now called the Faraday rotation. In this experiment. Kirchhoff and Bunsen explained these lines as atomic absorption lines. astronomer. For more details see. for more details see Ref. 51). Faradayis holding a glass bar of the type he used in 1845 to demonstrate the Faraday rotation.. Francois Jean Dominique Arago (February 26. In 1814. Faraday wrote in his notebook: I have at last succeeded in . 1787 . The experiment established that magnetic force and light were related. 4345. 48.1. With this result. AugustinJean Fresnel (May 10. and politician. however Arago experimentally verified the prediction. magnetising a ray of light. Faraday had established that a changing magnetic field produces an electric field. he became the 25th Prime Minister of France. for more details see Ref.July 5. Fresnel contributed significantly to the establishment of the wave theory for light. 1827). e. Michael Faraday (September 22. Arago observed the rotation of the plane of polarization in quartz. for which he was the first to construct a special type of lens. Joseph von Fraunhofer (March 6. In 1817.June 7. Birth of Photography: Joseph Niepce's 1822 photograph `La Table Servie' (see Ref. In 1845. Ref. (Ref. 1833) was a French inventor and a pioneer in photography. 1788 . 46. he wrote a memoir on diffraction for which in the ensuing year he received the prize of the Academie des Sciences. he was nominated a commissioner of lighthouses. he was able to show via mathematical methods that polarization could be explained only if light was entirely transverse.. Although this spot is usually referred to as the Poisson spot but many people call it as Arago's spot. now called a Fresnel lens.6 explain the colours of thin films (such as soap bubbles). In 1818. In 1819.July 14. In 1811.August 25. 1791 . he proposed that light waves were transverse and thus explained polarization. these lines are referred to as Fraunhofer lines. In 1818.
1775 . This is the reason why sky appears blue. Around 1865..colour has a wavelength which is about 1. (11 December 1781 . Thus the blue colour is scattered 10 times more than the red colour (because the red . for more __ details see Refs 52 and 53.History of Optics 1. 5759). however. the colour looks blue and the light coming out straight appears reddish. the shorter blue wavelengths are scattered more strongly than the red. see Refs 54 and 55. 1812) was a French engineer. thus from the side. James Clerk Maxwell (13 June 1831 . For more details about Maxwell. Max Planck had said. This synthesis represents one of the great scientific achievements of the 19` hcentury.5 November 1879) was an outstanding Scottish mathematician and theoretical physicist. Some of the original papers of Maxwell can be seen in the website in Ref. Albert Einstein had said. the sky does not appear violet because there is very little violet in the sunlight! Some of the scientific papers of Lord Rayleigh can be seen at the website given in Ref. the reflected ray makes an angle of 90 with the refracted ray. John William Struttusually referred to as Lord Rayleigh (12 November 184230 June 1919) In 1869.. In 1931 (during the birth centenary celebration of Maxwell). this made him predict that light must be an electromagnetic wave. these equations are known as Maxwell's equations and appeared in his book A Treatise Electricity and Magnetism published in 1873. he promptly called this as Brewster's law! Malus is best known for the law named after him which states that the intensity of light transmitted through two polarizers is proportional to the square of the cosine of the angle between the polarization axes of the polarizers. In 1810. David Brewster repeated the experiments of Malus for many materials and realized that when a ray is polarized by reflection. John Tyndall had discovered that when light passes through a transparent liquid with small particles in suspension (like a small amount of milk put in water). 60.10 February 1868) was a Scottish scientist. physicist. Malus had published his theory of double refraction of light in crystals. he was unable to obtain the relationship between the polarizing angle and refractive index: In 1811. and mathematician. Maxwell also predicted the existence of electromagnetic waves (which were later observed by Hertz) and showed that the speed of propagation of the electromagnetic waves is approximately equal to the (then) measured value of the speed of light. the most profound and the most fruitful that physics has experienced since the time of Newton". he wrote: This velocity is so nearly that of light that it seems we have strong reason to conclude that light itself (includ John Tyndall (August 2. Maxwell showed that the laws of electricity of magnetism can be described by four partial differential equations.this as Tyndall scattering but it is more often referred to as Rayleigh scattering because Rayleigh studied this phenomenon in great detail and showed (in 1871) that scattering is inversely proportional to the fourth power of the wavelength (see Ref. Malus had published his discovery of the polarization of light by reflection. Lord Rayleigh received the 1904 Nobel Prize in Physics. David Brewster.. In 1809. "(Maxwell's theory). remains for all time one of the greatest triumphs of human intellectual endeavor". Although many people call .7 ing radiant heat and other radiations) is an electromagnetic disturbance in the form of waves propagated through the electromagnetic field according to electromagnetic laws.. "(The work of Maxwell was) . 1893) was an Irish natural philosopher. .February 24. 1820December 4.75 times the wavelength of blue). FRS. Although violet has even smaller wavelength. In 1864. EtienneLouis Malus (July 23. 56.
P UN[DA Heinrich Hertz and James Clerk Maxwell on a Mexican stamp (Ref. Hertz was also the first scientist to observe photoelectric effect. To quote from Ref. Michelson built the famous interferometer which was later called the Michelson interferometer. Born and raised in Edinburgh. These waves. is named (Ref. he found that the maximum spark length. In 1897. Heinrich Rudolf Hertz (February 22. John Tyndall had demonstrated light guidance in water jets. 65). J J Thomson discovered electrons and in 1899. though Joseph Henry was not aware of . But they are there. In 1902. he showed that electrons are emitted when light falls Optics on a metal surface. the book by David Park (Ref. 1857 . 62). He was awarded the 1907 Nobel Prize in Physics (the first American to receive the Nobel prize in science) for his optical precision instruments and the spectroscopic and .January 1. and then to the United States in 1871. the SI unit of frequency. 1894) was a German physicist after whom the Hertz. Hertz generated electric waves using an electric circuit. Hertz had later said. Hertz reported the observations but did not pursue further and also did not make any attempt to explain them. We should mention here that in 1842 (when Maxwell was only 11 years old) the American physicist Joseph Henry had magnetized needles at a distance of over 30 feet (with 2 floors. "I do not think that the wireless waves I have discovered will have any practical application ". In 1887.8 In 1854. 67). The above is a 1876 photo of Alexander Graham Bell (3 March 1847 . Hertz proved that these waves were transmitted through air by detecting them with another similar circuit some distance away. Hertz was a very modest person..2 August 1922) speaking into prototype model of the telephone (Ref. Albert Abraham Michelson (19 Dec 18529 May1931) was born in Strelno. . 1880. 64) and the original collection of Henry's papers referenced in Park's book. "Nothing. asked one of his students at the University of Bonn. and when sparks crossed this gap violent oscillations of high frequency were set up in the rod. for more details see e. the circuit contained a metal rod that had a small gap at its midpoint.. "This is just an experiment that proves Maestro Maxwell was right. what next?". duplicating but not acknowledging Babinet (see Ref.it.g. we just have these mysterious electromagnetic waves that we cannot see with the naked eye.1. Philip Lenard observed that (i) the kinetic energy of the emitted electrons was independent of the intensity of the incident light and (ii) that the energy of the emitted electron increased when the frequency of the incident light was increased. 61 for more details).: was reduced when the apparatus was put in a black box (this is due to the fact that the box absorbed the ultraviolet radiation which helped the electrons in jumping across the gap). he had produced and detected electromagnetic waves. He also showed that like light waves they were reflected and refracted and. while receiving the electromagnetic waves in a coil with a spark gap. Prussia and moved to the US at the age of 2 (Ref. he emigrated to Canada in 1870. The photophone was invented jointly by Alexander Graham Bell and his assistant Charles Sumner Tainter on February 19. each 14 inches thick) from a single spark." "So. Bell believed the photophone was his most important invention. both in the form of light and radio waves. 63: In 1888. 66). after the discovery he had said. conclusively confirmed Maxwell's prediction on the existence of electromagnetic waves. that they traveled at the same speed as light but had a much longer wavelength. I guess". originally called Hertzian waves but now known as radio waves. most important. in a corner of his physics classroom at the Karlsruhe Polytechnic in Berlin. Thus.
this prediction led to the development of the laser. 1867July 9. possessed of a degree of accuracy never hitherto attained.9 tum theory of radiation. In 1917.. By its means we are enabled to ensure that the prototype of the metre has remained unaltered in length. during the centenary celebration of Einstein's Year of Miracles. In 1897.. 64) has written: He (Michelson) was 34 when he established that ether cannot be found. Einstein. E = Inc'. In 1887. 69). David Park (Ref. he and Edward Morley carried out the famous MichelsonMorley experiment which proved that ether did not exist. Professor Mukunda while concluding a talk on Einstein's Life and Legacy said. 70.. and to restore it with absolute infallibility. Einstein with Rabindranath Tagore. "How then to conclude this talk? I can think of no better way than to say of Einstein what he said about Gandhi in 1939: Generations to come will scarce believe that such a one as [Gandhi] ever in flesh and blood walked upon this earth. 69). the President of the Royal Swedish Academy of Sciences said.November 28. Albert Einstein (14 Mar 1879 18 Apr 1955) was an outstanding theoretical physicist. but by non participation in anything you believe is evil. Maurice Paul Auguste Charles Fabry (June 11. 1945) and JeanBaptiste Alfred Perot (November 3. while rederiving Planck's law. he made delicate optical measurements for 44 more years and to the end of his days did not believe there could be a wave without some material substance to do the waving.Your interferometer has rendered it possible to obtain a nonmaterial standard of length. during their widelypublicized July 14. Einstein in 1905 put forward that light consists of quanta of energy. 68." Some of Einstein's early papers can be found in the website in Ref. In 2005. 1863. supposing it were to get lost ". 1925) were French physicists. He received the 1921 Nobel prize in Physics for his services to Theoretical Physics. 1.History of Optics metrological investigations carried out with their aid. was able to predict the process of stimulated emission and almost 40 years later. 1930 conversation (Ref. Max Planck presents Einstein with the MaxPlanck medal in Berlin on June 28. We should strive to do things in his spirit: not to use violence in fighting for our cause. 1929 for extraordinary achievements in theoretical physics (Ref." Einstein was also a great humanist and a great admirer of Mahatma Gandhi about whom he had said: I believe that Gandhi's views were the most enlightened of all the political men in our time. In the presentation speech. . Fabry and Perot published their important article on what we now call the FabryPerot interferometer. Albert Einstein was named Time magazine's `Person of the. In the adjacent photograph. For more details about them see Ref. where the lead article said "He was the embodiment of pure intellect. ". in a paper entitled On the quan In 1999. Century'... Einstein is best known for his theory of relativity and specifically massenergy equivalence. this eventually led to the development of quantum theory. and especially for his explanation of the photoelectric effect.
27 June 1975). At a time when no single known fact supported this theory. De Broglie received the 1929 Nobel Prize in Physics for his discovery of the wave nature of electrons. the photograph is from Ref. He made this discovery in 1912.20 Oct 1984) and Werner Karl Heisenberg (December 5. This discovery. 75).. his assistant Walter Thompson is on the left. In 1924. not just light. father and son were jointly awarded the Nobel Prize in Physics For their services in the analysis of crystal structure by means of Xrays. 77. In the presentation speech it was mentioned Louis de Broglie had the boldness to maintain that . 1901  . this led the Nobel prize winning physicist PAM Dirac to make the famous statement that "each photon then interferes only with itself '. matter is. 13 Jan 1936 (from Ref. de Broglie (pronounced in French as de Broy) formulated the de Broglie hypothesis.Taylor has often been described as "one of the great physical scientists of the 20th century"..10 Optics Arthur Holly Compton (10 Sep 1892 . Compton found that the energy of an Xray or gamma ray photon decreases due to scattering by free electrons. 74). the adjacent photograph is from Ref.g. He discussed his ideas with his father (William Henry Bragg) who developed the Xray spectrometer in Leeds. claiming that all matter. a wave motion. In 1915. he related wavelength to the momentum. Louis de Broglie asserted that a stream of electrons which passed through a very small hole in an opaque screen must exhibit the same phenomena as a light ray under the same conditions. has a wavelike nature. 76.19 Mar 1987) was a French physicist. 71. On the right is Geoffrey Ingram Taylor (7 March 1886 .. Refs 72 and 73.15 Mar 1962) on the cover of Time magazine. by its nature. Louis de Broglie (15 Aug 1892 . a fact that upset the son! (Ref. De Broglie's formula was confirmed three years later for electrons with the observation of electron diffraction in two independent experiments. William Henry Bragg (2 July 186210 March 1942) William Lawrence Bragg (31 March 18901 July 1971) William Lawrence Bragg (the son) discovered the most famous Bragg's law which makes it possible to calculate the positions of the atoms within a crystal from the way in which an Xray beam is diffracted by the crystal lattice. Taylor demonstrated interference fringes using an extremely feeble light source.. during his first year as a research student in Cambridge. The collaboration between father and son led many people to believe that the father had initiated the research. known as the Compton effect. The research papers of Compton can be found in the website given in Ref. demonstrates the corpuscular nature of light. see e. Compton received the 1927 Nobel Prize in Physics for his discovery of the effect named after him. 1 Paul Adrien Maurice Dirac (8 Aug 1902 . For more details.1. In 1922. In 1909.
The very idea of God is a product of human imagination. Sometimes thesetwo photons would have to annihilate one another and other times they would have to produce four photons. He once said "God used mathematics in creating the world". . the Raman effect was discovered ... Landsberg and Mandel'shtam (in Russia) were also working on light scattering and according to Mandel'shtam.. scientists from Russia kept calling Raman scattering__ as Mandel'shtamRaman scattering. Dirac said: "I cannot understand why we idle discussing religion. Chandrasekhara Venkata Raman (7 Nov 188821 Nov 1970) Kariamanikkam Srinivasa Krishnan (4 Dec 189813 June 1961) On 28th February 1928. I do not recognize any religious myth. today we routinely talk about Raman amplification in optical fibers. we should have half the total number going into each component. 1976) were both celebrated theoretical physicists.S. Heisenberg and Dirac took part in it. Werner Heisenberg recollects a friendly conversation among young participants at the 1927 Solvay Conference about Einstein and Planck's views _ on religion. Subsequently. What they did not clearly realize. The uncertainty principle (which can be derived directly from the axioms of quantum mechanics) can be used to explain the diffraction of a photon (or an electron). the paper had Raman as the author and therefore thephenomenon came to be known as Raman effect although many scientists (particularly in India) kept on referring it as the RamanKrishnan effect. there were several papers written by Raman and Krishnan. The new theory.. This would contradict the conservation of energy. in his famous book Principles of Quantum Mechanics. Heisenberg was one of the founders of quantum mechanics and is also wellknown for discovering one of the central principles of modern physics. Interference between two different photons never occurs. 80. he acknowledged that the observations were made by K. which connects the wave function with probabilities for one photon gets over the difficulty by making each photon go partly into each of the two components. According to Ref. be able to interfere with one in the other. we should require a photon in one component to. Among other things. Each photon then interferes only with itself. . they observed the `Ramanlines' on 21 S` February 1928. Dirac can be considered as the creator of the complete theoretical formulation of quantum mechanics. but here he used `God' as a metaphor for nature. On the assumption that the beam is connected with the probable number of photons in it. he communicated a paper entitled A Change of Wavelength in Light Scattering to Nature..we cannot help but admit that any religion is a pack of false statements. however. wrote: Some time before the discovery of quantum mechanics people realized that the connection between light waves and photons must be of a statistical character. In 1928. deprived of any real foundation. the Heisenberg uncertainty principle which he developed in an essay published in 1927. Raman observed `Raman effect' in several organic vapours like pentane . we refer the reader to Ref.V.. But the results were presented in April 1928 and it was only on 6`h May 1928 that Landsberg and.which they called `the new scattered radiation' .and as scientists honesty is our precise duty . "Dirac to whom in my opinion we owe the most logically perfect presentation of quantum mechanics". 79). But by then it was too late! Much later. If we are honest . Krishnan and C. Krishnan and himself. K.70 years later it has become an important mechanism for signal amplification in optical communication systems. Dirac is widely regarded as one of the greatest physicists of all time. Raman got the 1930 Nobel prize in Physics for "his work on the scattering of light and for the discovery of the effect named after him". . 78.Wolfgang Pauli. Suppose we have a beam of light consisting of a large number of photons split up into two components of equal intensity. was that the wave function gives information about the probability of one photon being in a particular place and not the probable number ofphotons in that place. the paper was published on 21st April 1928. '. Dirac did not believe in God. At about the same time. Raman made a newspaper announcement on 29th February and on 8th March 1928.. Mandel'shtam communicated their results to the journal Naturwissenschaften. at least because they contradict one another. Pauli jokingly said: "Well. Albert Einstien had said. our friend Dirac has got a religion and the first commandment of this religion is: God does not exist and Paul Dirac is his prophet". Although in the paper. S.. Dirac. For a nice historical account of Raman effect. (Ref. If the two components are now made to interfere.History of Optics February 1. The importance of the distinction can be made clear in the following way.
Schawlow got the Nobel prize much later. Townes . According to Townes. they coined the word `maser' for this device. adapted from Ref. he conceived the idea of amplification through population inversion in 1951 (see Ref. The adjacent photograph is that on a 1988 Hungarian stamp honoring Gabor (from Ref.9 Feb 1979. 2005). The most important concept in the development of laser is that of stimulated emission which was introduced by Einstein in 1917. 87). an American physicist. USA in 1963 and Yuri Denisyuk in the Soviet Union.5 May 2007) and the components of his first ruby laser (Ref. absorption would always dominate over emission. and (e) the first page of Gould's laser notebook written in November 1957. London). which is an acronym for microwave amplification by stimulated emission of radiation.January 8.g.1921 . (d) Gordon Gould (July 17. the term `laser ' was first introduced to the public in Gould 's 1959 conference paper `The LASER. 81). In this article. Theodore Harold Maiman (11 July 1927 . In A Century of Nature: TwentyOne Discoveries that Changed Science and the World. In 1947. he shared the 1981 Nobel Prize in Physics with Nicolaas Bloembergen and Kai Siegbahn for their contributions to the development of laser spectroscopy.1. while working in the area of electron optics at British ThomsonHouston Co. 1920 . 1922July 1. many people believe that Gordon Gould (while he was a graduate student at Columbia University) is the inventor of the laser. Budapest . In fact. Townes wrote an article entitled The First Laser (Ref. the books by Taylor (Ref. On the first page of Gould's laser notebook (written in November 1957) he coined the acronym LASER and described the essential elements for constructing the laser. He was MAGYAR POSTA awarded the 1971 . Townes. Light Amplification by Stimulated Emission of Radiation ' . Around the same time. In 1958.April 28. the field of holography advanced only after development of the laser in 1960.. Schawlow and Townes published a paper entitled `Infrared and Optical Masers' in Physical Review showing how stimulated emission would work with much shorter wavelengths and describing the basic principles of the optical maser (later to be renamed a laser). A portion of Bertolotti's book can be read at the website given in Ref. a Russian physicist born in Australia. 1916 . 83) and in early 1954. in the UK. which has led to the construction of oscillators and amplifiers based on the maserlaser principle. both American physicists. Gordon and Zeiger (at the Physics Department of Columbia University) published a paper on the amplification and generation of electromagnetic waves by stimulated emission. 88). 1999).H.12 Dennis Gabor (5 Jun 1900. 85). Townes.September 16. initiating this new scientific field. because under normal conditions.1915) and (his sister's husband) Arthur Leonard Schawlow (May 5. However. Optics (a) Charles Hard Townes (b. However. Gould fought for thirty years United States Patent and Trademark Office for recognition as the inventor of the laser. 84) and by Bertolotti (Ref. (c) Aleksandr Mikhailovich Prokhorov (July 11. a Russian physicist and educator. July 28. The first holograms that recorded 3D objects were made by Emmett Leith and Juris Upatnieks in Michigan. 2001). C. Dennis Gabor invented holography. during the 40th anniversary celebration of the `invention of the laser' at the 1998 CLEO conference in San Francisco. It took over 35 years to realize amplification through stimulated emission primarily because stimulated emission was long regarded as a purely theoretical concept which never could be observed. 86. e.Nobel Prize in Physics for his invention and development of the holographic method. Basov and Prokhorov shared the 1964 Nobel Prize in Physics for their fundamental work in the field of quantum electronics. see. 82. Basov and Prochorov at the Lebedev Institute in Moscow independently published papers about the maser. 2002). Half of the prize was awarded to Townes and the other half jointly to Basov and Prokhorov. (b) Nikolay Gennadiyevich Basov (December 14.
Ref. 93. in June 1960. in a landmark theoretical paper (published in Proceedings of TEE). Maiman then turned to Nature. who was an editor of Physical Review Letters at the time. (Ref. but the editors turned it down. usually even more selective than Physical Review Letters. July 2.g. e. China) and George Hockham of ..13 wrote "Theodore Maiman made the first laser operate on 16 May 1960 at the Hughes Research Laboratory in California. Snitzer also invented both neodymiumand erbium doped laser glass. an article on the excitation of ruby with light. a continuous laser light (at 1. He promptly submitted a short report of the work to the journal Physical Review Letters. and also in surgery. The above 1964 photograph of Dr Patel with the carbon dioxide laser is from Ref.15 microns) from a gas laser.1938) developed the carbon dioxide laser in 1963. with an examination of the relaxation times between quantum states. has said that he turned down this historic paper because Maiman had just published. November 4. In 1966. 1960. 91 and 92 C. In 1961. Eager to get his work quickly into publication. and Donald Herr/off produced. On December 12. Kumar N. see. Patel (b. for the first time. Some have thought this was because Physical Review had announced that it was receiving too many papers on masersthe longerwavelength predecessors of the laserand had announced that any further papers would be turned down. within one year of the development of the first laser. 89 and 90). William Bennett. it is now widely used in industry for cutting and welding.History of Optics 1. Pasternack's reaction perhaps reflects the limited understanding at the time of the nature of lasers and their _ significance. All Javan. Charles Kuen Kao (b. But Simon Pasternack. by shining a highpower flash lamp on a ruby rod with silvercoated surfaces. where the paper was better received and published on 6 August". and that the new work seemed to be simply more of the same. Elias Snitzer and his coworkers developed the first fiber optic laser. 1933 in Shanghai.
.compared with existing coaxial cable and radio systems. They said that if the impurities could be removed. then (to quote from Ref. UK.. Japan and Germany started working on purifying glass and the first breakthrough was reported in 1970. 94) "the new form of communication medium. Alferov shared the 2000 Nobel Prize in Physics.1. After this paper. 95.14 Standard Telecommunications Laboratories in UK pointed out that the loss in glass fibers was primarily caused by impurities and therefore it was not a fundamental property of the fiber itself. In about 10 more years. Optics Diode Laser Anatomy Semiconductor Laser Diode 61=4 izuzzh. has a larger information capacity and possible advantages in basic material cost".or may be even less.this breakthrough was the starting point of the fiber optic revolution. UK. Window Connector Pins Bonding Metallic Wires Strip Monitor Photodiode In 1970. and Peter Schultz successfully prepared the first batch of optical fiber with sufficiently low loss so as to make fiberoptic communication a reality . by Izuo Hayashi and Morton Panish at the Bell Labs design the (Ref. and in June 1970. the photograph is from Ref. France. Murray Ramsay of Standard Telecommunication Laboratories (now Nortel) demonstrates fiber optic video communications to HRH Queen Elizabeth II in May 1971 at the Institute of Electrical Engineers in London. Robert Maurer. If this could he achieved. optical fibers became so transparent that more than 95% of the signal power would pass after propagating through 1 km of the optical fiber. 97.. Corning Glass Works scientists Donald Keck. scientists in USA. the loss can be brought down to about few dB/km . Semiconductor lasers that operate continuously at room temperature were first fabricated in May 1970 by Zhorev Alferov and his group in Leningrad. the above photograph is from Ref. as research continued. This was a major turning point towards the development of the fiber optic communication system. 96).
P. http://en. 1. Canada.uiuc.org/wiki/Fran%C3%A7ois_Arago 47.org/wiki/Archytas_of_Tarentum 2.15 18. Joyce and Alice Joyce.htm 23. V Ronchi.org/expo/vatican.wikipedia. REFERENCES 1.fairfield.edu/cgilocal/DHI/ot2wwwdhi?specfile=/texts/english/dhi/ dhi.wikipedia. Simpson. http://www. http://en.com/WSRNet/D1/hist.exhibit/exhibit/dmathematics/Greek math. Eric Renner.virginia. 36.wikipedia. http://www.aol. http://en. http://en. Heinemann.edu/jmac/sj/scientists/grimaldi . 5. http://en. The EDFA brought about a revolution in fiber optic communication systems.wikipedia. http://www.org/wiki/AIHazen 1L.org/wiki/Olaf_Roemer 38. London (1972). The Nature of Light (Translated by V Barocas).100.com/cgibin/SoftCart.org/wiki/Aryabhatta 10. and J.cz/enlpinholecameras/whatis.org/files/14725/14725h/14725h.org/wiki/Kepler 20. Dover publications (1962).html 8. Newton and Snell's Law. The Dictionary of the History of Ideas at the University of Virginia Library (http://etext.wikipedia.wikipedia. Rediscovering a Historic Technique.umd.virginia.B.wikipedia. http://en. W. http://www. The erbium doped fiber amplifier (usually abbreviated as EDFA) was invented in 1987 by a group including David Payne.org/wiki/Descartes 27.org/wiki/Fermat 26. http://en. http://www.aoogle. from the University of Southampton and a group from AT&T Bell Laboratories.org/phys/tyoung/tyoung. http://wwwgroups. Focal Press (1995). Dover Publications (1952). http://en.org/wiki/David_brewster 54. http://members. http://www. Desurvire.org/wiki/Willebrord_Snellius 25.org/wiki/Robert_Hooke 33.com/watbooks/2915. Descartes.htm 44.org/wiki/Ptolemy 9.stand. Mears and L.aol.edu/optics/timeline/people/ witelo. Becker.html 4. J. http://www. Soc. http://en. http://en.wikipedia.ibiblio.html 98).org/wiki/Joseph_von_Fraunhofer 48.wikipedia.com/books?id=GnAFAAAAQAAJ& pg=PA3 81 &dq=Newton%27 s+book+OPTICKS#PPA219.wikipedia. 3.edu/ece549/lecture23.wikipedia.org/wiki/Galileo 24.History of Optics In 1978.html 13.htm 37. http://www.wikipedia.com/WSRNet/D1/hist.ru/Physics/English/top_ref. http://en.org/wiki/EtienneLouis_Malus 53.wikipedia.wikipedia.ece.pinhole.org/wiki/Michael_Faraday 52.ibnalhaytham. R. http://en.net/custom. 40.wikipedia. http://en.com/young.lib. http://members. Vol 66. 13 August 1944). http://en.nad.wikipedia. Payne (b.org/wiki/Fresnel 49.ece.wikipedia.wikipedia. http://www.phy. http://en. http://en.edu/cgilocal/DHI/ot2wwwdhi?specfile=/texts/english/dhi/dhi. http://en.htm 32. http://en. Reekie.greenlion.org/wiki/Leonardo_da_Vinci 17.exe/ optics. pl.em?pid=673913 12.polarization.org/wiki/Witelo 16.ecademics.o2w&act=text&offset=13050720&query=optics &tag=OPTICS+AND+VISION).wikipedia. http://en. http://en.org/wiki/Niepce 50. http://en. http://courses. http://books. Treatise on Light. New York.htm 42. Holt Reinhart and Winston. the photosensitivity of germaniumdopedcore optical fibers was discovered by Kenneth Hill while working at the Communications Research Centre in Ottawa.dcs.edu/taylor/optics3.wikipedia.. http://www. 43.htm 46. Ml 41. He also demonstrated the first infiber Bragg grating (see Ref.org/wiki/Isaac_Newton 39. E. http://www. http://en.wikipedia.faculty.html .com/history/bart. 30.org/kheirand/publications/reviews . http://en.hr/dpaar/fizicari/xmaxwell.org/wiki/Euclid 6. 1976.org/wiki/Christiaan_Huygens 35.wikipedia.html?E+scstore 21.wikipedia.pdf 51.org/wiki/Francesco_Maria_Grimaldi 31. http://en. http://en.wikipedia. C Huygens. http://en.wikipedia. 1959.html. http://en.uk/history/PictDisplay/ Heron. "Great Experiments in Physics" p96101.polybiblio.html 34. Opticks.fsu.magnet.htm 15.org/wiki/Thomas_Young_%28scientist %29 45. 19. http://physics. David N. Pinhole Photography. Opt. Am.html 14.htm 7.cavendishscience. http://www.gutenberg. 28.o2w&act=text& offset=13050720&query=optics&tag=OPTICS+AND+ VISION 29. http://micro. http://etext.wikipedia.ac.wikipedia.org/wiki/Hans_Lippershey 22.manhattanrarebooksscience. Isaac Newton. http://www.org/wiki/James_Clerk_Maxwell 55.
org/wiki/John_Tyndall 58.ieee. http://www.pdf 92.wikipedia.org/wiki/Louis_de_Broglie Optics 78. New York (2001). http://en. http://www. (1966). The Fire within the Eye: A Historical Essay on the Nature and Meaning of Light.press.org/articles/news/8/2/9/1/queen 98. 86.edu/brenner/taylor/physic_today/ taylor. and the ThirtyYear Patent War. http://en. 74.harvard.wikipedia. Hecht. http://www. 95.belllabs. http://www.org/wiki/Paul_Adrien_Maurice_Dirac 79.M l 87. http://www.physics.com/history/laser/contrib. Simon & Schuster Publishers.org/wiki/Laser 83. http://en. Nick Taylor.wikipedia. http://www. Princeton University Press (1997).edu/Mist/Chicago/284158_ townes.beyonddiscovery.gov/nif/library/aboutlasers/how. http://en.html 89. The History of the Laser.org/wiki/Albert_Abraham_Michelson 68. 69. Batchelor.html 88.llnl. http://www.com/books?id=gWMSAAAAIAAJ& pg=RA3PA426&1pg=RA3PA426&dq=scientific+papers +of+rayleigh&source 61.org/wiki/Einstein 70. 1151.belllabs. http://en. lEE.I. Journey into Light: Life and Science of C. http://th. http://en. 85. http://en.google. Institute of Physics Publishing.wikipedia.info/hertz. J.org/wiki/Alexander Graham_Bell 67.org/wiki/Dennis_Gabor 82.org/wiki/Arthur_Compton 76.seas. http://fibers.wikipedia.com/Maxwell/maxwell_oplf.htm 72. Kao and G.wikipedia.cap.html 97. "Who were Fabry and Perot?". Proc.html 94.113 (7). http://www.com/history/laser/contrib.wikipedia.wikipedia. Am. http://en. Venkataraman. html 93.com/books?id=WcDPrrfloQC&pg =PA56&ots=FNeFkfdlW_&dq=Papers+of+Albert+Einstein &sig=OADkr l YehpFPm9xCR21pK1YNs z0#PPA 110. F.org/?id=3717 84. Hockham.com/history/laser/contrib. A. David Park. G. 65. Oxford (1999) 62.shtml 59. No.de/jr/gif/stamps/ sm_maxwell.com/laser. 80. Dielectric fibre surface waveguides for optical frequencies. http://www.com/books?id=JObDnEtzMJUC&pg =PA241&lpg=PA241&dq javan+bennett+and+herriot&source =web&ots=tvP2kA2aNb&sig=K7KDtOYXj gHjTPaYwpOx XquumT4#PPP9.unifrankfurt. K.htm 90. Phys. the Nobel Laureate. http://books.uklgeneral/history/john_tyndall_ biography. http://books.belllabs. G. Vol. http://phisicist. Mario Bertolotti.pdf ].org/wiki/Heinrich_Hertz 63.html 64.1. W Heisenberg.U.jpg 66. http://books.uchicago.page asp?I=448 96. City of Light.org/wiki/Geoffrey_Ingram_Taylor 73. Laser: The Inventor.. http://en.google.tyndall. Physics and Beyond: Encounters and Conversations.pdf 57. Vol.wikipedia. http://books.16 56. http://en.ac. September 1998 [available at the website http://www.physik. The Life and Legacy of G. 66.html .farhangsara.org/organizations/foundation/donors .greatachievements. http://vacuumphysics.org/content/view. http://www.wikipedia. Cambridge University Press (1994).rutgers.. J. http://en. http://en. http://en. http://www. pp798802. J.com/ooks?id=98sCh99YIJsC&dq = Papers+of+Arthur+H+Compton 77. Penguin Books (1994).com/C125716F004E0776/0/D94B 4F 1 F28466E2AC 125717100513E 1 A/$File/Innovation_ 7_35. http://www. New York: Harper & Row (1971). http://www.wikipedia.wikipedia. Mulligan. Philadelphia (2005). Taylor.ca/awards/press/1998hill.wildpedia. 81. C.zeiss. 9.google.edu/ugrad/387/ Mulligan98. Raman.html 91.google.org/wiki/William_Lawrence_Bragg 75.M 1 71.org/wiki/John_William_Strutt 60.
Now we have given up. and then in many respects it behaved like a wave. protons. protons. the electron. Isaac Newton. We say: `it is like neither'. we are assuming that the motion is confined to the xy plane. The trajectory of the . So it really behaves like neither.INTRODUCTION Ever since man could see. 2. alphaparticles.1. There is one lucky break. Much later. I will reflect on what light is Albert Einstein. a luminous body emits a stream of particles in all directions. it was shown that phenomena such as the photoelectric effectand the Compton effect could only be explained if we assume a particle model of light. a collision between the two particles rarely occurs. Albert Einstein. The particles are assumed to be very tiny so that when two light beams overlap. it was found that light did indeed sometimes behave like a particle. we usually tend to visualize them as tiny particles. for example. Near the end of the chapter. Indeed electrons. Using the corpuscular model. Historically. as we know. are neither particles nor waves. Now. Thus.electrons behave just like light. In the early days. . photons. alphaparticles. we will qualitatively discuss how the wave and the particle aspects of radiation can be explained on the basis of the uncertainty principle and the probabilistic interpretation of matter waves. was thought to behave like a particle. but then it was discovered that it behaves like a wave. and so on) is the same for all.1): Newton thought that light was made up of particles. However. the answers to the questions like `What is an electron or what is light' are very difficult. etc. The modern quantum theory describes them in a very abstract way which cannot be connected with everyday experience. however (in the beginning of the twentieth century). In this chapter. In' order to understand refraction we consider the incidence of a particle at a plane surface (y = 0) as shown in Fig. the phenomena of interference and diffraction were demonstrated which could only be explained by assuming a wave model of light.1 . The reflection law follows considering the elastic reflection of a particle by a plane surface. are known to a tremendous degree of accuracy . a light beam was thought to consist of particles. one can explain the laws of reflection and refraction in the following manner. he has been interested to know as to what light is. Later. neutrons. The quantum behaviour of atomic objects (electrons. protons. 1951 2. Later. 2. According to it. To quote Feynman (from_Ref. however . Thus.2 THE CORPUSCULAR MODEL The corpuscular model is perhaps the simplest model of light. the values of the mass and charge of electrons. CA 1917 All the fifty years of conscious brooding have brought me no close to the answer to the question. we will make a brief historical survey of the important experiments which led to models regarding the nature of light. photons. etc. wrote `Are not the ray of light very small bodies emitted from shining substance?'. but he is deluding himself.approximately one part in a billion! Their velocities can also be changed by the application of electric and magnetic fields. or whatever you want to call them. in his book Opticks (Ref. they also exhibit diffraction and other effects ' which can only be explained if we assume them to be waves. 2). they are all `particlewaves'. neutrons.For the rest of my life. `What are light quanta?' Of course today every rascal thinks he knows the answer.
2. this derivation is equivalent to the corpuscular derivation which is usually attributed to Newton. 2. discovered the law of refraction which is known as Snell's law. Heath & CO. 2.2 Stroboscopic pictures of a ball moving with a certain speed on a horizontal surface moving down to a lower horizontal surface through a slope [Adapted from PSSC. if the ball initially moves on the lower surface approaching the slope.. however. Two stroboscopic pictures of the motion of the ball are shown in Fig. Using this table.. the component perpendicular to the edge increases in value resulting in an increased speed of the ball. Conversely.2. The above equation directly gives Snell 's law sine1 sin 62 _ P2 P1 _ V2 V1 (2) In order to understand the explanation of Snell's law of refraction using the corpuscular model.e. 2..2 Y Optics x Fig.1. The slope can be approximately assumed to represent the interface between the two media. Rene Descartes derived Snell's law.1 Refraction of a corpuscle. 2]. In 1637. The component of the momentum parallel to the edge of the slope does not change. Mass. this is consistent with the reversibility of rays undergoing refraction. Newton (16421727) was only about eight years old when Descartes (15961650) died and therefore Descartes did not get the corpuscular model from Newton! The first edition of Newton's Optiks (in which Newton had discussed the corpuscular model) was published in 1704 [Ref. as we shall see later. particle is determined by the conservation of the xcomponent of the momentum (= p sin 0) where 9 is the angle that the direction of propagation makes with the yaxis. an . The conservation condition leads to the following equation: pi sin e1 = P2 sin 02 (1) where the angles 01 and 02 are defined in Fig. The wave theory does make the correct prediction of the ratio of velocities of waves in the two media. Fig. It is probably because of the popularity of Newton's Optiks that the corpuscular theory is usually attributed to Newton. We may mention here that in 140 AD.2. in 1621.C. used with permission]. Boston. it predicts that if the ray moves towards the normal (i. we consider a simple experiment in which a ball moving with a certain speed on a horizontal surface moves down to a lowerhorizontal surface through a slope. is not consistent with experimental observations. Although the simple corpuscular model of light explains Snell's law of refraction satisfactorily. if the refraction occurs at a denser medium) its speed would become higher which. Physics. the speed decreases as it goes up the slope. D. 1965. Claudius Ptolemy measured the angle of refraction in water for different angles of incidence in air and made a table of it. Willebrord Snell.
2. and (b) that light could propagate through vacuum.. Using this principle. there came up a large number of experimental observations (like interference.light is a fluid that exhibits wavelike motion. the mechanic and. [Photograph courtesy: United States Information Service. Fermat derived Snell's law (see Chapter 3) and showed that if the velocity of light in the second medium is less. some light does enter the geometrical shadow which is due to the phenomenon of diffraction. corpuscles of different sizes give rise to the sensation of different colors at the retina of the eye. contrary to what predicted by the `corpuscular theory'. the sky appears perfectly dark (see Fig.3). We should mention here that if we are below the shade of a building then under the shade we can always read a book . Francesco Grimaldi.3). . He explained the prismatic spectrum by assuming that particles of different sizes refract at different angles. 4]. then the shadows would have been extremely dark which is indeed the case on the surface of the moon (see Fig.to quote from the internet . because it alone can afford us the enjoyment of a look at the personal activity of this unique man. Historically. were observed by Hooke around the middle of the seventeenth century. Ref.3 A photograph of the man on the moon. 3]. even on the surface of the moon. not the least.In one person he [Newton] combined the experimenter. the ray would bend towards the normal. Perhaps the two most important experimental facts which led to the early belief in the corpuscular model of light were: (a) the rectilinear propagation of light which results in the formation of sharp shadows. Hooke also observed this phenomenon. As we shall see in later chapters.. As will be discussed in later chapters. This phenomenon is essentially due to the wave character of light and cannot be explained on the basis of the simple corpuscular model. dark and we would never be able to read a book in our own shadow! And also. This phenomenon of scattering is also responsible for the blue color of the sky and the red color of the setting sun. Once again. is nevertheless Fig. an Italian physicist. Einstein wrote: . `Newton's rings ' . was probably the first person to observe the phenomenon of diffraction of white light as it passed through small apertures.. using the corpuscular model. which are a beautiful manifestation of the wave character of light. Later. Since the moon does not have an atmosphere.3 THE WAVE MODEL Although the corpuscular model explains the propagation of light through free space and can be made to predict the correct forms of the laws of reflection and refraction. New Delhi]. Grimaldi concluded that . according to Newton. However.If the earth did not have an atmosphere. the explanation of Newton's rings on the basis of the wave model is discussed in Chapter 15.the light that enters the shadow is not due to diffraction but due to scattering of light by air molecules. shadows are not perfectly dark.this new edition of his [Newton's] work on Optics to be welcomed with warmest thanks. on the surface of the moon. Newton's explanation of the rings can be found at many places [see for example. Descartes' theory remained undisputed until about 1662 when Fermat enunciated the principle of least time. diffraction effects are usually difficult to observe because the wavelength associated with light waves is extremely small.What is Light? 2. the rings are named after Newton because he had given an explanation of their formation. Finally. We may mention here that commenting on Isaac Newton's Optiks.. a satisfactory explanation of the diffraction phenomenon can only be given if one assumes a wave model of . 2. the shadows would be extremely 2. diffraction and polarization) which could not be explained on the basis of the corpuscular model of light.3 English translation of Descartes' original paper appears in a paper by Joyce and Joyce [Ref.. a small amount of light does enter the geometrical shadow because of diffraction. certain and alone: his joy in creation and his minute precision are evident in every word and in every figure. which was later found to be quite unsatisfactory. the theorist. The domain of optics in which light is assumed to travel in straight lines is known as geometrical optics which can easily be explained on the basis of the corpuscular model of light. as careful experiments later showed. Around 1665. 2. He stands before us strong. the artist in exposition. Notice the dark sky.
and this results in the propagation of the electromagnetic wave even in free space. Associated with a light wave are changing electric and magnetic fields. In 1816. 2. the diffraction effects are small and therefore light approximately travels in straight lines. Poisson. Indeed. Heinrich Hertz carried out experiments which could produce and detect electromagnetic waves of frequencies smaller than those of light.14858 x 108 mis. we will not go into the details of the various theories. This drawback was also one of the reasons for the immediate nonacceptance of the wave model. Faraday carried out experiments which showed that a varying magnetic field induces an electromotive force. In 1888.see Sec.2. The frequency of the emitted electromagnetic waves depended on the values of the inductance and capacitance of the circuit. Soon after. A derivation similar to the original derivation is given in Ref. by assuming that light waves were transverse in character. Around 1830. 5]. light waves are electromagnetic waves. The sole fact that the two values were very close to each other led Maxwell to propound (around 1865) his famous electromagnetic theory of light according to which. Optics The nineteenth century also saw the development of electricity and magnetism. Cauchy and many other physicists contributed to the development of the ether theory which also necessitated the development of the theory of elasticity. 1 The derivation of the Fresnel laws on the basis of electromagnetic theory will be discussed in Chapter 24. 22. Navier.107 x 108 m/s. In 1816. He found that this value was very close to the measured value of the speed of light which according to the measurement of Fizeau in 1849 was 3. In 1808. he showed that the velocity of the electromagnetic waves can be calculated from experiments in which a certain quantity of electric charge is measured by two different methods. Indeed. at the time of Huygens. Because of the smallness of the wavelength. in Fig. Young showed that the wavelength of light waves was about 6 x 10 5 cm. However. From these equations.4).1. There were considerable difficulties in the explanation of the models and since we now know that ether does not exist. . Young gave a satisfactory explanation of the formation of Newton's rings.4. Huygens could explain the laws of reflection and refraction (see Chapter 12) and he could also interpret the phenomenon of double refraction (see Chapter 22) discovered in 1669 by the Danish physicist Erasmus Bartholinus. 2 This follows directly from the dipole radiation pattern . From the wave equation so derived. the changing magnetic field produces a time and space varying electric field and the changing electric field produces a time and space varying magnetic field. the elastic ether theory was developed. The electromagnetic waves could be detected by means of a detector and it was found that a signal was not received when the detector was placed parallel to the source2 (see Fig. the branch of optics in which one completely neglects the finiteness of the wavelength is called geometrical optics and a ray is defined as the path of energy propagation in the limit of A > O. This model was first put forward by Huygens in 1678 [Ref. In addition. In 1820.4. In 1802. similar experiments were also carried out by Henry around the same time and the law is also referred to as the FaradayHenry law. Maxwell found that the speed of the electromagnetic waves in air should be about 3. Using the wave model. so compelling was Newton's authority that it is said that people around Newton had more faith in his corpuscular theory than Newton himself and no one believed in Huygens' wave theory until 1802 when Thomas Young performed the famous interference experiment which could only be explained on the basis of a wave model of light (see Chapter 14). Malus observed the polarization of light but he did not try to interpret this phenomenon. Soon afterwards. 6. which was explained by Young. the wave theory seemed to be well established and since it. he derived a wave equation and predicted the existence of electromagnetic waves (see Chapter 23). Fresnel along with Arago performed the famous experiment on the superposition of linearly polarized light waves.4 light. light was thought to travel in straight lines and Huygens tried to invoke unrealistic assumptions in order to explain the rectilinear propagation of light using his wave theory. the dipole is oscillating along the zaxis and the electric field on the yaxis is along the zaxis. He summed up all the laws of electricity and magnetism in the form of equations [which are now referred to as Maxwell's equations (see Chapter 23)]. In the second quarter of the nineteenth century. These waves were produced by discharging electrically charged plates through a spark gap. Maxwell generalized Ampere 's law by stating that a changing electric field can also set up a magnetic field. 22. Ampere found that two parallel wires carrying currents attract each other. These measurements were carried out in 1856 by Kohlrausch and Weber. in 1832 Fresnel derived the expressions for the reflection and transmission coefficients ) by using models for ether vibrations. and from their data. was thought that a wave required a medium for its propagation. Oersted discovered that currents caused magnetic effects. Fresnel gave a satisfactory explanation of the diffraction phenomenon by means of a wave theory and calculated the diffraction patterns produced by various types of apertures and edges.
What is Light?
2.5 a metal sheet he could demonstrate the laws of reflection. Hertz's experimental results provided a dramatic confirmation of Maxwell's electromagnetic theory. In addition, there were so many other experimental results, which were quantitatively explained by using Maxwell's theory that towards the end of the nineteenth century, physicists thought that one had finally understood what light really is.
2.4 THE PARTICLE NATURE OF RADIATION
In 1887, while receiving the electromagnetic waves in a coil with a spark gap, Hertz found that the maximum spark length was reduced when the apparatus was put in a black box; this is due to what is now known as the photoelectric effect and the box absorbed the ultraviolet radiation which helped the electrons in jumping across the gap. Hertz reported the observations but did not pursue further and also did not make any attempt to explain them. In 1897, J J Thomson discovered electrons and in 1899, he showed that electrons are emitted when light falls on a metal surface. In 1902, Philip Lenard observed that (i) the kinetic energy of the emitted electrons was independent of the intensity of the incident light and (ii) that the energy of the emitted electron increased when the frequency of the incident light was increased. As will be discussed in Sec. 25.2, this phenomenon cannot be explained by a theory based on the wave model of light. In 1905, Einstein interpreted the photoelectric effect by putting forward his famous photon theory according to which light_ consisted of quanta of energy E=hv (3)
(a)
(b) Fig. 2.4 Schematic of Hertz's experiment for generation and detection of electromagnetic waves. Sparks across the gap P produce electromagnetic waves whose frequency depends on the inductance and capacitance of the circuit. The electromagnetic waves can be detected by means of a detector D which is nothing but a short wire bent in the form of a circle with a small gap. A signal is detected if the gap in the detector is parallel to the line joining the knobs of the spark gap P as shown in (a); if the gap is at right angles as shown in (b), no signal is received. Hertz also produced standing electromagnetic waves by gettingthem reflected by a metal sheet (see Figs 13.3 and 13.4). He could calculate the wavelength of the waves and knowing the frequency, he showed that the speed of the electromagnetic waves was the same as the that of light. Using a collimated electromagnetic wave, and getting it reflected by
where v is the frequency and h (= 6.626 x i034 Js) is the Planck's constant; and that the emission of a photoelectron was the result of the interaction of a single quantum (i.e., of the photon) with an electron. In his 1905 paper (Ref. 7 and reprinted in Ref. 8), Einstein wrote: Monocromatic radiation behaves as if it consists of mutually independent energy quanta of magnitude hv. In the same paper he also wrote: According to Maxwell's theory, energy is considered to be a continuous spatial function for all purely electromagnetic phenomena, hence also for light.... The wave theory of light, which operates with continuous spatial functions, has proved itself superbly...
2.6 In 1909, Einstein wrote [quoted from Ref. 9] It is undeniable that there is an extensive group of data concerning radiation which shows that light has certain fundamental properties that can be understood much more readily from the standpoint of the Newton emission (particle) theory than from the standpoint of the wave theory. It is my opinion, therefore, that the next phase of the development of theoretical physics will bring us a theory of light that can be interpreted as a kind of fusion of the wave and emission theories. We may note the prediction of Einstein. To quote from Ref. 10: Owing to Einstein's paper of 1905, it was primarily the photoelectric effect to which physicists referred as an irrefutable demonstration of the existence of photons and which thus played an important part in the conceptual development of quantum mechanics. It was only in 1926 that Gilbert Lewis, an American chemist, coined the word `photon' to describe Einstein 's `localized energy quanta'. Einstein received the 1921 Nobel Prize in Physics for his services to Theoretical Physics, and especially for his explanation of the photoelectric effect. Einstein's photon theory predicted that if the frequency v of the incident radiation was greater than the critical frequency vc, then the kinetic energy of the emitted electron would be h(v which was later verified. by Millikan for visible light,. by de Broglie for Xrays and by Thibaud and Ellis for yrays. Einstein also showed that the photons, in addition to having an energy equal to hv, should have a momentum given by
Optics of (especially) solid bodies. All my attempts however, to adapt the theoretical foundation of physics
to this (new type of) knowledge failed completely. It was as if the ground had been pulled out from under one, with no firm foundation to be seen anywhere, upon which one could have built. That this insecure and contradictory foundation was sufficient to enable a man of Bohr's unique instinct and tact to discover the major laws of the spectral lines and of the electron shells of the atoms together with their significance for chemistry appeared to me like a miracle  and appears to me as a miracle even today. This is the highest form of musicality in the sphere of thought. [Quoted from the autobiographical notes by Einstein in Albert Einstein: Philosopher, Scientist, edited by P. A. Schilpp, Tudor Publishing Co., New York, 1951]. In making this transition from Planck' s quantised oscillators to quanta of radiation, Einstein had made a very important conceptual transition, namely, he introduced the idea of corpuscular behavior of radiation. Although Newton had described light as a stream of particles, this view had been completely superseded by the wave picture of light, a picture that culminated in the electromagnetic theory of Maxwell. The revival of the particle picture now posed a severe conceptual problem, one of reconciling wave and particle like behaviour of radiation. It also soon became apparent that matter also exhibited both types of behaviour. For example, an electron with an accurately measured value of mass and charge could undergo diffraction in a manner similar to that of light waves. We will now give a brief account of some of the other important experimental evidence showing waveparticle duality that led to the development of the quantum theory.
p = cv =
h
(4)
which was verified experimentally in 1923 by Compton. This experiment is known as the Compton effect and will be discussed in detail in Sec. 26.3. Compton received the 1927 Nobel Prize in Physics for his discovery of the effect named after him. We may mention here that in 1900, Max Planck had put forward his famous theory of blackbody radiation, the derivation of which presupposed that energy can be absorbed and emitted by an individual resonator only in quanta of magnitude hv... According to Einstein: ,... I could nevertheless see to what kind of consequences this law (i.e., Planck's law) of temperatureradiation leads for the photoelectric effect and for other related phenomena of the transformation of radiationenergy, as well as for the specific heat
2.5 WAVE NATURE OF MATTER
Experiments by Wilson with his cloud chamber had clearly shown theparticle like behaviour of alpha and beta particles. These are emitted by radioactive elements and when they pass through supersaturated vapour, they form tracks of condensed droplets. For alpha particles, these tracks are nearly straight lines; however, for electrons, the tracks are irregularly curved. The existence of continuous tracks suggests that the emissions from the radioactive substance can be regarded as minute particles moving with high speed. Further, the fact that electrons could be deflected by electric and magnetic fields and also the fact that one could accurately determine the ratio of their charge to mass suggest very strongly that
s
What is Light?
2.7
electrons are particles. This view remained unchallenged for where p is the momentum of the electron. Shortly afterwards, a number of years; C.T.R. Wilson was awarded the 1927 in 1928, G.P. Thomson carried out electron diffraction exNobel Prize in Physics for his method of making the paths of periments by passing electrons through thin polycrystalline electrically charged particles visible by condensation of metal targets (see Sec. 18.10 for more details). The diffracvapour. tion pattern consisted of concentric rings similar to the At this stage, one could ask if matter may not show wave DebyeScherrer rings obtained in the Xray diffraction patlike behavior also just as light exhibited corpuscular and tern. By measuring the diameters of the rings and from the wavelike behaviour. In 1925, de Broglie proposed just such known structure of the crystals, Thomson calculated the a hypothesis and argued that the relation given by Eq.(4), wavelength associated with the electron beam which was in between wavelength and momentum applied for electrons as agreement with the deBroglie relation [Eq. (5)1. In 1937, well. In his 1925 paper, he wrote: Davisson and Thomson shared the Nobel prize for their experimental discovery of the diffraction of electrons by The basic idea of quantum theory is, of course, the crystals. Max Jammer [Ref. 10] has written: Thomson, the impossibility of considering an isolated fragment of father, was awarded the Nobel prize for having shown that energy without assigning a certain frequency to it... the electron is a particle, and Thomson, the son, for having de Broglie was awarded the 1929 Nobel Prize in Physics for shown that the electron is a wave. his discovery of the wave nature of electrons. In the presenIn Fig. 2.5 we have shown Debye Scherrer rings produced tation speech (on December 12, 1929), the Chairman of the by scattering of Xrays [see (a)], and by scattering of elecNobel Committee for Physics said: trons [see (b)] by an aluminum foil . The two figures clearly show the similarity in the wavelike properties of Xrays and Louis de Broglie had the boldness to maintain that of electrons. not all the properties of matter can be explained by the theory that it consists of corpuscles... At a time when no single known fact supported this theory, Louis de Broglie asserted that a stream of electrons which passed through a very small hole in an opaque screen must exhibit the same phenomena as a light ray under the same conditions.... The experimental results obtained have fully substantiated (a) Louis de Broglie's theory. Hence there are not two (b) worlds, one of light and waves, one of matter and Fig. 2.5 The diffraction pattern of aluminum foil produced corpuscles. There is only a single universe. (a) by Xrays and (b) by electrons; notice the similarity in the diffraction patterns. Photograph Later, de Broglie wrote [quoted from p. 58 of Ref. 9]: courtesy McGraw Hill Digital Access Library. I was convinced that the wave particle duality discovered by Einstein in his theory of light quanta was 2.6 THE UNCERTAINTY absolutely general and extended to all of the physiPRINCIPLE cal world, and it seemed certain to me, therefore, that the propagation of a wave is associated with the The reconciliation of the corpuscular nature with the wave motion of a particle of any sort photon, electron, character of light (and also of the electron) has been brought proton or any other. about through the modern quantum theory; and perhaps the In 1927, Davisson and Germer studied the diffraction of best known consequence of waveparticle duality is the unelectrons from single crystals of nickel and showed that the certainty principle3 of Heisenberg which can be stated as diffraction patterns could be explained if the electrons were follows: assumed to have a wavelength given by the de Broglie relaIf the xcoordinate of the position of a particle is tion known to an accuracy Ax, then the xcomponent of h the momentum cannot be determined to an accuracy (5) p 3We may mention here that the uncertainty principle can be derived from the Schrodinger equation (see, e.g., Chapter 5 of Ref. 12). And as mentioned by Richard Feynman, "Where did we get that [the Schrodinger equation] from? Nowhere. It is not possible to derive it from anything you know It carne out of the mind of Schrodinger."
2.8 better than Apx = h/L,x, where h is the Planck's constant. Alternatively, one can say that if Ax and Apx represent the accuracies with which the xcoordinate of the position and the xcomponent of the momentum can be determined, then the following inequality must be satisfied
Ox
Optics S
S
•
. by
'
t
Apx h
(6)
We do not feel the effect of this inequality in our everyday experience because of the smallness of the value of Planck's constant (= 6.6 x 10 27 erg sec). For example, for a tiny particle of mass 106 gm, if the position is determined within an accuracy of about 106 cm, then according to the uncertainty principle, its velocity cannot be determined within an accuracy better than Av = 6 x 1016. cm/sec. This value is much smaller than the accuracies with which one can determine the velocity of the particle. For a particle of a greater mass, Ov will be even smaller. Indeed, had the value of Planck's constant been much larger, the world would have been totally different. In a beautifully written book, Gamow (Ref. 13) has discussed what our world would be like if the effect of the uncertainty principle were perceivable by our senses.
S' Fig. 2.6 Diffraction of a photon (or an electron) beam by a narrow slit of width b cular model is quite incapable of explaining the phenomenon of diffraction. However, if we use the uncertainty principle in conjunction with the corpuscular model, the diffraction phenomenon can be explained in the following manner: When a photon (or an electron) passes through the slit, one can say that Ax=b which implies that we can specify the position of the photon to an accuracy b. If we now use the uncertainty principle, we would have h (7) APx b i.e., just by making the photon pass through a slit of width b, the slit imparts a momentum in the xdirection which is h/b. It may be pointed out that before the photon entered the slit, px (and hence Opx) can be made arbitrarily small by putting the source sufficiently far away. Thus we may write Apx 0. It would however be wrong to say that by making the photon pass through the slit, Apx Ax is zero; this is because of the fact that Apx 0 before the photon entered the slit. After the photon has entered the slit, it is confined within a distance b in the x direction and hence Apx = h/b. Further, since before the photon entered the slit px = 0, we will therefore have h I Px' APx ' b But px = p sin e, where 0 is the angle that the photon coming out of the, slit makes with the yaxis (see Fig. 2.6). Thus psin0 or sin 0
2.7 THE SINGLE SLIT DIFFRACTION EXPERIMENT
We will now show how the diffraction of a light beam (or an electron beam) can be explained on the basis of the corpuscular nature of radiation and the uncertainty principle. Consider a long narrow slit of width b as shown in Fig. 2.6. Now, one can always choose the distance between the source and the slit large enough so that p, can be assumed to have an arbitrarily small value. For example, for the source at a distance d, the maximum value of px of the photons approaching the slit will be b _ by b Px^P2 cd which can be made arbitrarily small by choosing a large enough value of d. Thus we may assume the light source to be sufficiently far away from the slit so that the photons approaching the slit can be assumed to have momentum only in the ydirection. Now, according to the particle model of radiation, the number of particles reaching the point P (which lies in the geometrical shadow) will be extremely small; further, if we decrease the width of the slit, the intensity should decrease, which is quite contrary to the experimental results because we know that the beam undergoes diffraction and the intensity at a point like P would normally increase if the width of the slit is made smaller. Thus, the classical corpus 
b
=p b
(8)
The above equation predicts that the possibility of a pho ton traveling at an angle 0 with the ydirection is inversely
What is Light?
2.9 tion of neutrons by a single slit and his experimental results agree with the intensity distribution as predicted by the wave theory with A given by Eq.(5).
proportional to the width of the slit; i.e., smaller the value of b, greater is the value of 0 and greater is the possibility of the photon to reach deep inside the geometrical shadow. This is indeed the diffraction phenomenon. Now, the momentum of a photon is given by h P_A Thus Eq. (8) becomes sin 0 A (9)
2.8 THE PROBABILISTIC INTERPRETATION OF MATTER WAVES
In the previous section, we have seen that if a photon lasses through a slit of width b, then the momentum imparted in the xdirection (which is along the width of the slit) is = h/b. The question arises whether we can predict the trajectory of an individual photon. The answer is no. We cannot say where an individual photon will land up on the screen; we can only predict the probabilities of arrival of the photon in a certain region of the screen. We may, for example, say that the probability for the arrival of the photon in the region lying between the points A and B (see Fig. 2.6) is 0.85. This would imply that if the experiment was carried out with a large number of photons, about 85% of them would land up in the region AB; but the fate of an individual photon can never be predicted. This is in contrast to Newtonian mechanics where the trajectories are always predetermined. Also, if we place a light detector on the screen, then it will always record one photon or none and never half a photon. This essentially implies the corpuscular nature of the radiation. However, the probability distribution is the same as predicted by the wave theory and therefore if one performs an experiment with a large number of photons (as is indeed the case in most ex periments) the intensity distribution recorded on the screen is the same as predicted by the wave theory. In order to explicitly show that diffraction is not a many photon phenomenon, Taylor in 1909 carried out a beautiful experiment which consisted of a box with a small lamp which casts the shadow of a needle on a photographic plate (see Fig. 2.8). The intensity of light was so weak that between the slit and the photographic plate, it was almost impossible to find two photons (see Example 2.1). In fact, to get a good fringe pattern, Taylor made an exposure lasting several months. The diffraction pattern obtained on the photographic plate was the same as predicted by the wave theory. The corpuscular nature of radiation and the fact that one cannot predict the trajectory of an individual photon can be seen from Fig. 2.9, which consists of series of photographs showing the quality of pictures obtainable from various number of photons (Ref. 16). The photograph clearly shows that the picture is built up by the arrival of concentrated packets of energy and the point at which a particular photon will arrive is entirely a matter of chance. The figure also shows that
which is the familiar diffraction theory result as will be discussed in Sec. 18.3. We can therefore say that the waveparticle duality is a consequence of the uncertainty principle and the uncertainty principle is a consequence of the waveparticle duality. To quote Max Born [Ref. 14]: Physicists of today have learnt that not every question about the motion of an electron or a photon can be answered, but only those questions which are compatible with the uncertainty principle. Returning to Eq. (7), we may mention that de Broglie had suggested that the equation A = h/p is not only valid for photons but is also valid for all particles like electrons, protons, neutrons, etc. Indeed, the de Broglie relation has been verified by studying the diffraction patterns produced when electrons, neutrons, etc., pass through a single crystal; the patterns can be analyzed in a manner similar to Xray diffractions (see Sec. 18.10). In Fig. 2.7, we show the experimental data of Shull (Ref. 15) who studied the Fraunhofer diffrac900 600
Slit width
6.10 sec 770 microns
1 0 20 10 0 10 20
Angular position (seconds arc) Fig. 2.7 Angular broadening of a neutron beam by small slits [After Ref. 15].
2.10
Optics
the photograph is featureless when a small number of photons are involved and as the number of photons reaching the photographic plate increases, the intensity distribution becomes the same as would be predicted by the wave theory. To quote Feynman: ...it would be impossible to predict what would happen. We can only predict the odds! This would mean, if it were true, that physics has given up on the problem of trying to predict exactly what will happen in a definite circumstance. Yes! physics has given up. We do not know how to predict what would happen in a given circumstance, and we believe now that it is impossiblethat the only thing that can be predicted is the probability of different events. It must be recognized that this is a retrenchment in our earlier idea of understanding nature. It may be a backward step, but no one has seen a way to avoid it. A somewhat similar situation arises in radioactivity. Consider a radioactive nucleus having a halflife of say 1 hour. If we start with 1000 such nuclei, theft on an average 500 of them would undergo radioactive decay in 1 hour and about 250 of them in the next 1 hour and so on. Thus, although to start with, all nuclei are identical, some nuclei would decay in the very first minute and some nuclei can survive for hours without undergoing radioactive decay. Thus, one can never predict as to which nucleus will undergo decay in a specified period; one can only predict the probability of its undergoing decay in a certain interval of time. This is indeed a manifestation of quantum mechanics. To quote Feynman again: A philosopher once said it is necessary for the very existence of science that the same conditions always produced the same results. Well they don't!
e
Slit Smoked Glass Fig. 2.8 Schematic diagram of the experimental arrangement of Taylor to study the diffraction pattern produced by a weak source. The whole apparatus was placed inside a box.
Weak Source
Needle Photographic Plate
2.9 AN UNDERSTANDING OF INTERFERENCE EXPERIMENTS
We consider the twohole interference experiment similar to that performed by Young (see Secs 14.4 and 14.5). The experimental arrangement is shown in Fig. 2.10 where a weak light source So illuminates the hole S and the light emerging from the holes St and S2 produces the interference pattern on the screen PP'. The intensity is assumed to be so weak that in the region between the planes AB and PP' there is almost never more than one photon (see Example 2.1).. Individual photons are also counted by a detector on the screen PP' and one finds that the intensity distribution has a cos 2 pattern
Fig. 2.9 Photographs showing the quality of a picture obtainable from various numbers of photons: (a), (b), (c), (d), (e) and (f) correspond to 3 x 103 photons, 1.2 x 104 photons, 9.3 x 104 photons, 7.6 x 105 photons, 3.6 x 106 photons and 2.8 x 107 photons respectively. [From Ref. 16; reprinted with permission].
What is Light?
A
2.11
So
B Fig. 2.10
P'
Young's double hole experimental arrangement for obtaining the interference pattern. So represents a point source.
similar to that shown in Fig. 14.8. The corpuscular nature of the radiation is evident from its detection in the form of a single photon and never a fraction of a photon. The appearance of the interference pattern is because of the fact that a photon interferes with itself. The quantum theory tells us that a photon partially passes through the hole SI and partially through S2. This is not the splitting of the photon into two halves but only implies that if we wish to find out through which hole the photon passed, then half the time it will be found to have passed through the hole SI and half the time through S2. The photon is in a state which is a superposition of two states, one corresponding to the wave emanating from. hole SI and the other to the one emanating from hole S2. The superposed state will give rise to an intensity distribution similar to that obtained by considering the superposition of two waves. It may be noted that if we had employed a device (like a microscope) which would have determined which hole the photon had passed through, then the interference pattern on the screen would have been washed out. This is a consequence of the fact that a measurement always disturbs the system. This is very nicely discussed in Ref. 1. Thus we may say that the photons would arrive as packets of energy but the probability distribution (on the screen) will be proportional to the intensity distribution predicted by using a wave model. In a recent paper, Tonomura and his coworkers (Ref. 17) have demonstrated the single electron build up of an interfer Fig. 2.11 Buildup of the electron interference pattern. Number of electrons in (a), (b), (c), (d) and (e) are ence pattern. Their results are shown in Fig. 2.11. It may be 10, 100, 3000, 20000 and 70000 respectively. seen that when there are very few electrons they arrive ran[Adapted from Ref. 17]. domly; however, when a  large number of electrons are involved, one obtains an intensity distribution similar to the duced by a Michelson interferometer is discussed in Chapter 15. one predicted by wave theory. According to Dirac [Ref.18]: We next consider the interference experiment involving Some time before the discovery of quantum mechathe Michelson interferometer in which a light beam is parnics people realized that the connection between tially reflected by a beam splitter and the resulting beams are light waves and photons must be of a statistical made to interfere (see Fig. 2.12); the interference pattern pro
2.12
M1 ,s\\\ \ \\ \ \ \ \\ \ \ \\ \\'
Optics into which one original beam has been split. In the accurate mathematical theory each translational state is associated with one of the wave functions of ordinary wave optics, which may describe either a single beam or two or more beams into which one original beam has been split. These translational states can be superposed in a manner similar to the one employed while considering the interference of two beams. Thus, each photon goes partly into each of the two components and interferes only with itself. If we try to determine the fate of a single photon by measuring the energy in one of the components then Dirac argues: The result of such a determination must be either a whole photon or nothing at all. Thus the photon must change suddenly from being partly in one beam and partly in the other to be entirely in one of the beams. This sudden change is due to the disturbance in the translational state , of the photon which the observation necessarily makes. It is impossible to predict in which of the two beams the photon will be found. Only the probability of either result can be' calculated .... Our description of the photon allows us to infer that, after such an energy measurement, it would not be possible to bring about any interference effects between the two components. So long as the photon is partly in one beam and partly in the other, interference can occur when the two beams are superposed, but this probability disappears when the photon is forced entirely into one of the beams by an observation.
ti

Fig. 2.12 Schematic of the setup of the Michelson interferometer. G represents a beam splitter, Ml and M2 represent plane mirrors. character. What they did not clearly realize, however, was that the wave function gives information about the probability of one photon being in a particular place and not the probable number of photons in that place. The importance of the distinction can be made clear in the following way. Suppose we have a beam of light consisting of a large number of photons split up into two components of equal intensity. On the assumption that the beam is connected with the probable number of photons in it, we should have half the total number going into each component. If the two components are now made to interfere, we should require a photon in one component to be able to interfere with one in the other. Sometimes these two photons would have to annihilate one another and other times they would have to produce four photons. This would contradict the conservation of energy. The new theory, which connects the wave function with probabilities for one photon gets over the difficulty by making each photon go partly into each of the two components. Each photon then interferes only with itself Interference between two different photons never occurs. In the Michelson interferometer experiment, Dirac argues: ... we describe the photon as going partly into each
of the two components into which the incident beam
2.10 THE POLARIZATION OF A PHOTON
Let us consider the incidence of a plane electromagnetic wave on a polaroid whose pass axis is along the ydirection (see Fig. 2.13); obviously, the electric vector of the transmitted wave would be along the ydirection (see Chapter 22). Thus, if the electricvector associated with the incident wave oscillates along the xaxis, the wave will be absorbed by the polaroid. On the other hand, if the electric vector oscillates along the yaxis, it will just pass through the polaroid. Further, if the electricvector makes an angle 9 with the pass axis, then the intensity of the transmitted beam will be Io cos 2 0, where to represents the intensity of the incident beam (this is known as Malus' law which will be discussed in Sec. 22.3). In the photon theory also one can associate a certain state of polarization with every photon. One can argue that if the electricvector associated with the photon is along the y (or
is split. The photon is then, as we may say, in a translational state given by the superposition of the two translational states associated with the two components.... For a photon to be in a definite translational state it need not be associated with one single beam of light, but may be associated with two or more beams of light, which are the components
What is Light? y
2.13 similarly, the number of photons passing through either St or S2 will approximately be 2.5 x 1010 x 2 x rc x (0.02)2 = 1000 per second 27n x (100)2 where we have assumed that after passing through S, the photons are evenly distributed in the hemisphere; this is strictly not correct because the diffraction pattern is actually an Airy pattern (see Chapter 18)  nevertheless, the above calculations are qualitatively correct. The distance between the planes AB and PP' is 100 cm which will be traversed by a photon in time 3 x 109 s. Thus, approximately every thousandth of a second a photon enters the region and the space is traversed much.before the second photon enters. Therefore, in the region between AB and PP' one will (almost) never find two photons. This is somewhat similar to the case when, on an average, 100 persons pass through a room in one year and the time that each person takes to cross the room is 1 s, thus it will be highly improbable to have two persons simultaneously in the room.
x
Fig. 2.13
The incidence on a polaroid of a linearly polarized light beam whose electricvector makes an angle 6 with the yaxis; the pass axis of the polaroid is along the yaxis.
the x) axis, then the photon will pass through or get absorbed by the polaroid. The question now arises as to what will happen to a single photon if the electricvector makes an angle 8 with the pass axis. The answer is that the probability for the photon to pass through the polaroid is cos 2 0 and if the experiment is conducted with N photons (and if N is very large) then about N cos 2 0 photons will pass through; one cannot predict the fate of an individual photon.
Example 2.2 In this example we will use the uncertainty principle to determine the size of the hydrogen atom4. Although this example is not directly related to optics, it demonstrates the farreaching consequences of the uncertainty principle. We consider the hydrogen atom which consists of a proton and an electron. Since the proton is very much heavier than the electron, we consider only the motion of the electron. Let the electron be confined to a region of linear dimension a. Thus according to the uncertainty principle:
p =gyp h/a (10)
Example.'2.1 Let a source (with A. = 5 x 10_5 cm) of power 1 W
be used in the experimental arrangement shown in Fig. 2.10. where /t = hl27c ; the reason for using h rather than h will be men(a) Calculate the number of photons emitted by the source per tioned later. The kinetic energy of the electron will be given by second. h2 (11) K.E. = P = , (b) Assume the radii of the holes S, S1 and S2 to be 0.02 cm and 2m 2ma SoS = SS1 = SS2 = 100 cm and the distance between the planes Now, there exists an electrostatic attraction between the two AB and PP' to be also 100 cm. Show that in the region beparticles; the corresponding potential energy being given by tween the planes AB and PP' one can almost never find two 2 photons. (12) P.E. = 4 eoa
;Solution:
(a) The energy of each photon will be 6.6 x 1034 (J . s) x 3 x 10 8 (m/s) 5 x 107 (m) 4x1019 J Thus the number of photons emitted per second will be 1W = 2.5 x 10 18 4 x 1019 J (b) The number of photons passing through the hole S will approximately be
by
=_
he
where q ( 1.6 x l019 C) represents the magnitude of the charge of the electron and e0 (= 8.854 x 1012 CN2 m 2) represents the permittivity of free space. Thus the total energy is given by
E=K.E.+R.E. q2 h2
(13) 47c eoa The system would settle to a state of lowest energy; thus we must set dE / da equal to zero: 2 dE 0==+ q2 l?
2ma2 da ma3 4nE 0 a 2
2.5x1018xx(0.02)2= 2.5 x 10 per second 4xn. x(100)2
___
0
implying
4 The analysis is adapted from Ref. 1.
2.14
Optics
a =an=
h2 q2 m 4'eo r
(14)
If we substitute the values of h(=. 1.055 x 10 34 Js), rn (= 9.11 x 31 kg), e0 and q we would obtain 10 (15) Thus we get the remarkable result that the size of the hydrogen atom is a direct consequence of the uncertainty principle. To quote Feynman:
which can be interpreted as follows: if At represents the uncertainty in the time at which a timedependent process takes place, then the uncertainty DE in the energy of this process will be h/At. Assuming At = 109 s,
DE =
h
At
x 106 eV
a=a 0 =0.53x 1010 m=0.53.E
Summary
♦ The corpuscular model of light is due to Descartes rather than to Newton. The law of refraction was discovered experimentally in 1621 by Snell. In 1637, using corpuscular model, Descartes derived Snell's law of refraction. ♦ The wave model of light was first propounded by Huygens in 1678. Using the wave model, Huygens could explain the laws of reflection and refraction and he could also interpret the phenomenon of double refraction. o Around the middle of the nineteenth century, Maxwell generalized Ampere's law by stating that a changing electric field can also produce a magnetic field. He summed up all the laws of electricity and magnetism in the form of equations which are now referred to as Maxwell's equations. From these equations, he derived a wave equation and predicted the existence of electromagnetic waves and showed that the speed of the electromagnetic waves in air should be about 3.107 x 108 m/s which was very close to the measured value of the speed of light. The sole fact that the two values were every close to each other led Maxwell to propound his famous electromagnetic theory of light according to which, light waves are
we now understand why we do not fall through the floor.... In order to squash the atoms close together, the electrons would be confined to smaller space and by the uncertainty principle, their momenta would have to be higher on the average, and that means high energy; the resistance to atomic compression is a quantum mechanical effect...
So
We next substitute the value of a from Eq. (13) in Eq. (12) to obtain
q2 \ _ h 2 m q2 L q2 _ 4trso \h2 4tteo E 2nz(h 2 4trEo 2 N2 m q 2h 2 Ottc o, Substituting the values of h, m, q and so, we get
(16)
E=2.17x10 19 J 13.6 eV (17) which is nothing but the ground state energy of the hydrogen atom. Thus, that the ionization potential of hydrogen atom is = 13.6 eV follows from the uncertainty principle: We may point out that the uncertainty principle can be used to give only an order of magnitude of the size of the hydrogen atom or its ionization potential; we had intentionally chosen the constants in such a way that the ground state energy comes out to be correct. It is for this reason that we had chosen h instead of h in Eqs (10) and (11).
electromagnetic waves.
♦ In 1905, Einstein interpreted the photoelectric effect by putting forward his famous photon theory according to which the energy in a light beam of frequency v was concentrated in corpuscles of energy hv, where h represents Planck's constant. ♦ The consequence of waveparticle duality is the uncertainty principle of Heisenberg according to which if the xcoordi
2.11 THE TIMEENERGY UNCERTAINTY RELATION
When an atom makes a transition from an energy state E2 to an energy state El (E2 > Et) a photon of frequency v = (E2 El)/h is emitted. This emission is essentially a pulse of a duration = 10 9 s; this duration is usually denoted by T. This leads to a frequency width Av and as will be shown in Chapter 17, TAv1 (18) Multiplying both sides by h, we get the timeenergy uncertainty principle:
nate of the position of a particle is known to an accuracy Ax, then the xcomponent of the momentum cannot be determined to an accuracy better than Apr = h/Ax, where h is the Planck's constant.
♦ The classical corpuscular model is quite incapable of explaining the diffraction of light by a single slit. However, if we use the uncertainty principle in conjunction with the corpuscular model, the;diffraction phenomenon can be explained. o In the Young's double hole interference pattern, the corpuscular nature of the radiation is evident from its detection in the form of single photons and never'a fraction of a photon. The appearance of the interference pattern is because of the fact that a photon interferes with itself. The quantum theory tells us that a photon partially passes through the two holes. This is not the splitting of the photon into two halves but only implies that the photon is in a state which is a superposition
At LE h
(19)
3 A photon of wavelength 6000. must be negative and greater in magnitude than the kinetic energy.2 In continuation of the previous problem. is passed through a slit of width 0. [Ans: = 4 x l0° photons] Solutions Problems 2. 2. one corresponding to the wave emanating from the first hole and the other to the one emanating from the second hole. we get E (1. estimate the kinetic energy of the proton inside the nucleus and the strength of the nuclear interaction.14 . Assume A. From this result. or h pI0 Therefore.1 An electron of energy 200 eV is passed through a _4 circular hole of radius i0 cm. AB = 6 x 10 radians] 2.6 µm.6xl0 Although electrons do emerge from nuclei in /idecay. since the rest mass of the electron is very much smaller than that of the proton.6 The lifetime of the 2P state of the hydrogen atom is about 1. Therefore . What would be the kinetic energy for an electron. A Laser beam P 2.14. is the mass of the proton.7 A 1 W laser beam (of diameter 2 cm) falls normally on two circular holes each of diameter 0. (V).05 x 1027 ergsec) 2 2 x 1. the rare occasions when /3decay occurs may be attributed to the transformation of a neutron into a proton and an electron (and the neutrino) so that the electron is in fact created at the instant the decay occurs. what would be the corresponding uncertainty for a 0.(V) 20 MeV which indeed gives the correct order of the potential energy.1 g lead ball thrown with a velocity 103 cm/sec through a hole 1 cm in radius? [Ans: Ae 5 x 1030 radians] 2.What is Light? of two states. 2. [Ans: = 6 x 10 8 s1 ] 2. On substitution.67 x 10 24 g x (1013 cm)2 = 3 x 105 ergs = 20 MeV Since the proton is bound inside the nucleus. The uncertainty in momentum for the electron is again V/r0 .15 Calculate the average number of photons that will be found between the planes AB and PP ' . however.np 2rnp rot where in. 5 x 10 24 g cm/s. they seldom have energies exceeding a few million electron volts. B P' [430 cm H Fig. Thus one does not expect the electron to be a basic constituent of the nucleus.05 cm as shown in Fig. [Ans: = 1. The superposed state will give rise to an intensity distribution similar to that obtained by considering the superposition of two waves. What is the uncertainty introduced in the momentum and also in the angle of emergence? 6 [Ans: Ap .5 The proton is confined within a sphere of radius r° = 1013 cm.200 MeV 1013 x 1.5 x 10 20 photons/s] 2.2 mm (a) Calculate the uncertainty introduced in the angle of emergence.5 Calculate the uncertainty in the momentum of a proton which is confined to a nucleus of radius equal to 1013 cm.4 A 50 W bulb radiates light of wavelength 0./b) where b is the width of the slit.05 x 1027) MeV . Calculate this angle and compare with the angle obtained in part (a). Calculate the number of photons emitted per second. the velocity of the electron is very close to c and we have to use the extreme relativistic formula for the energy.6 x 109 s. the kinetic energy of the proton will be given by 2 h2 E=P 2. the average of the potential energy. 2. if it had to be confined within a similar nucleus? 2. ch E=cp= re (3 x 10 10 ) (1. [Ans: Ae = 3 x 103] 2. Use the time energy uncertainty relation to calculate the frequency width Av. Thus the uncertainty in the momentum must be at least of the order of V/r0 . = 6 x 105 cm and the distance between the planes AB and PQ to be 30 cm. (b) The first minimum in the single slit diffraction pattern occurs at siri 1 (A.
New Delhi (2005). Treatise on Light. Tompkins in Wonderland. Tonomura. Newton and Snell's law. Isaac Newton. Descartes.G. 10.16 REFERENCES AND SUGGESTED READINGS 1. London (1962). III. Optiks. Macmillan India. R. Dover Publications (1962). V. Proceedings of the Symposium held at CNEN Casaccia Center in September 1968. (1957). reprinted by Kluwer Academic Publishers.. Cropper. Rome (1970). 66. Einstein. A. (1976). Joyce and A. John Wiley & Sons. Einstein's Miraculous Year: Five Papers that changed the face of Physics. Kawasaki and H.B. Shull. Ezawa.. 12. Stachel (Ed.2. 6. Dover Publications (1952). Leighton and M. Mass.. T. Soc. 57 (2). C. A. America. R. Vol. T. Feynman.M. W. Joyce. CNEN. 17. Prtinceton University Press. The Feynman 2.B. Reading. New York (1970). R. Academic Press. Annalen der Physik. Princeton. Lectures on Physics. J. A. On a heuristic point of view concerning the production and transformation of light. Atomic Physics. 4. (1965).B. Opt.H. (1965). Reading. 14. Feynman. Quantum Physics. New Delhi. Addison Wesley Publishing Co. R. The Quantum Physicists and an Introduction to their Physics. Born. Jammer.P. Leighton and M. A Text Book on Light. New York (1965). Dirac. 1998. Neutron Diffraction: A General Tool in Physics in Current Problems in Neutron Scattering. 3. Demonstration of singleelectron build up of an interference pattern. A. J. 8. Dordrecht (2004).). Phys. Resnick. II. McGrawHill. Optics 7. Eisberg and R.P. A. Cambridge University Press. Sands. The Conceptual Development of Quantum Mechanics. C. 117 (1989). M. 16. Barton. Sands. J. Ghatak and S Lokanathan. 17. Matsuda. Gamow. New York (1974). Mass. Vol. J. 1905. 132. Addison Wesley Publishing Co.A. 9. Rose. 15. Quantum Mechanics: Theory and Applications. Cambridge (1940). Longmans Green & Co. Oxford University Press. Oxford University Press. Endo. G. Blackie & Son. R. W. London (1939). M. Vol. W. Mr. . 18. The Feynman Lectures on Physics. 11. reprinted by Shrishti Publishers. Quantum Effects in Human Vision. Advances in Biological and Medical Physics Vol. P. 1. Am. 5. Huygens. 13. Oxford (1958). The Principles of Quantum Mechanics.
Z Geometrical PART This part (consisting of four chapters) is entirely based of geometrical optics and includes Ray tracing through gradedindex media explaining in` detail the phenomena of mirage and looming and also reflection from the ionosphere. which is extensively used in the industry. . A detailed description of the matrix method in paraxial pptics. Ray tracing through a system of lenses leading to various concepts used in the design of optical instruments. • A study of aberrations of optical systems.
.
we want more than just a formula. and even for extremely small diameters of the aperture. First we have an observation. 140 AD 3. this is called a ray. The field of optics under such an approximation (i. contrary to what predicted by the 'corpuscular theory'. consider a circular aperture in front of a point source P as shown in Fig. 1621 Although the above mentioned numerical table was made in 140 AD. Thus. 1657 Pierre de Fermat enunciated his principle of `least time' and derived Snell 's law of refraction and showed that if the velocity of light in the second medium is less. a Dutch mathematician.1 cm). then at first the size of the patch starts decreasing. then we have a law which summarizes all the numbers.1 mm) then the pattern obtained on SS' ceases to have welldefined boundaries. a ray defines the path of propagation of the energy in the limit of the wavelength going to zero. But the real glory of science is that we can find a way of thinking such that the law is evident. then on the screen SS '. we will obtain a welldefined shadow on the screen SS ' . the diffraction effects will be absent. I Important Milestones Greek physicist Claudius Ptolemy measured the angle of refraction in water for different angles of incidence in air and made a table of it. . and it is called the principle of least time. then we have numbers that we measure. In Chapters 18 and 20 we will discuss the phenomenon of dif fraction in great detail and will show that the diffraction effects become smaller with the decrease in wavelength and indeed in the limit of A > 0. his derivation assumed corpuscular model of light. Since light has a wavelength of the order of 105 cm. in many applications. or Fermat's principle.1 INTRODUCTION The study of the propagation of light in the realm of geometrical optics employs the concept of rays. The field of geometrical optics can be studied by using Fermat's principle which determines the path of the rays. it was only in 1621 that Willebrord Snell. discovered the law of refraction which is now known as Snell's law. one can. According to this principle the ray will correspond to that path for which the time taken is an extremum in comparison . Snell's law. 3.. in the zero wavelength limit one can obtain an infinitesimally thin pencil of light. neglect the finiteness of the wavelength. To understand what a ray is.). 1637 Descartes derived the..etc. Vol. the ray would bend towards the normal.e. mirrors.FERMAT'S . The first way of thinking that made the law about the behavior of light evident was discovered by Fermat in about 1650. one can see a patch of light with welldefined boundaries. but when the size of the aperture becomes very small (< 0. Richard Feynman in Feynman Lectures on Physics.ITS APPLICATIONS Now in the further development of science. When the diameter of the aperture is quite large (. which is small compared to the dimensions of normal optical instruments like lenses. the neglect of the finiteness of the wavelength) is called geometrical optics. This phenomenon is known as diffraction and is a direct consequence of the finiteness of the wavelength (which is denoted by A. When we start decreasing the size of the aperture.PRINCIPLE AND: . and therefore.1.
it is either a minimum or a maximum or stationary*. then 2 will be B C nr B in ds (1) A Fig. (3) and in this formulation. if A and B are two points in a homogenous medium. Thus according to Fermat' s principle. the ray paths may be maxima. 2). i.e. z) represent the position dependent refractive index.. Thus. Let n(x. is an extremum**. Let i' be the time taken along the nearby path AC'B (shown as the dashed curveFig. in a medium whose refractive index is constant at each point). The correct form is: The actual ray path between two points is the one for which the optical path length is stationary with respect to variations of the path.1 The light emitted by the point source P is allowed to pass through a circular hole and if the diameter of the hole is very large compared to the wavelength of light then the light patch on the screen SS' has well defined bouridaries. 3. 3. Since c is a constant. From the above principle one can immediately see that In a homogenous medium (i. then the ray path will be along the straight line ACB because any nearby path like ADB or AEB will correspond to a longer time. if 'r represents the total time taken by the ray to traverse the path AB along the curve C (see Fig.2).2 If the path ACB represents the actual ray path then thetime taken in traversing the path ACB will be an extremum in comparison to any nearby path AC'B. the symbol A B below the integral represents the fact that the integration is from the point A to B through the curve C.e.3. ** A nice discussion on the Extern= principle has been given in Chapter 26 of Ref. and if ACB indeed represents the path of a ray.2) then 2 The actual path between two points taken by a beam of light is the one which is traversed in the least time. This is expressed 'by Eq. Fermat's principle can be derived from Maxwell's equations (see Ref. 3. . Here. 3.3. y. greater than or equal to 't' for all nearby paths like AC'B. = L. We may mention here that according to the original statement of Fermat: ds _ n ds c c/n will represent the time taken to traverse the geometric path ds in a medium of refractive index n. 1. the above integral represents the optical path from A to B along C. one can alternatively define a ray as the path for which f n ds A CAB (2) B S' Fig. 3. the ray would follow the path for which s Ends =0 A>B (3) to nearby paths.. 3.e.2 LAWS OF REFLECTION AND REFRACTION FROM FERMAT'S PRINCIPLE We will now obtain the laws of reflection and refraction from Fermat's principle. The above statement is incomplete and slightly incorrect. i. the light ray would follow that path for which the time taken is an extremum.4 Optics A S P either less than.. and of course. = I A where dsi represents the ith arc length and the corresponding refractive index. minima or stationary. the rays will be straight lines because a straight line will correspond to a minimum value of the optical path connecting two points in the medium. Consider a plane mirror MN as shown in * The entire field of classical optics (both geometrical and physical) can be understood from Maxwell's equations. out of the many paths connecting the two points. c represents the speed of light in free space. Thus referring to Fig. Then where the lefthand side represents the change in the value of the integral due to an infinitesimal variation of the ray path. 3.r nl ds.
PS and PB are in the same plane and ZAPS = LSPB. 3. we have to determine the path from A to B (via the mirror) which has the minimum optical path length. BN = h2 and MR = x. Fig. the refracted ray and the normal to the interface must all lie in the same plane. 3.x. the angle of incidence i (= LAPS) and the angle of reflection r (= LSPB) Further. Since the path would lie completely in a homogenous medium. . Simple geometric considerations show that ZAPS = LSPB Lop =n 1 AR + n2 RB = n.4 The shortest path connecting the two points A and B via the mirror is along the path APB where the point P is such that AP. intersect the interface at R and proceed to B along RB. Clearly. Thus we have to minimize the length A 'PB.e. A'. Thus the points A. nix n2 (Lx) (5) p Q A' Fig. Let a ray starling from the point A. To determine that point R for which the optical path length from A to B is a minimum. as can be seen from Fig. there may be more than one ray path connecting two points.5 A and B are two points in media of refractive indices nl and n2. L ' Fig. 3. for minimum optical path length. Clearly.Fermat's Principle and Its Applications 3. the incident ray. where MN = L is a fixed quantity. Jx2 + hl + n2 j(L . let PQ be a surface separating two media of refractive indices n1 and n2 as shown in Fig.x)2 + h22 (4) To minimize this. in the presence of the mirror there will be two ray paths which will connect the points A and B. The straight line path AB is also a ray. we drop perpendiculars AM and BN from A and B respectively on the interface PQ. 3. PS being the normal to the plane of the mirror. is A Fig. These form the laws of reflection. we drop a perpendicular from A on the mirror and let A' be a point on the perpendicular such that AR = RA ' . we will have a ray. and thus. The optical path length from A to B.5 must be equal and the incident ray. 3. the two paths will be AB and APB. To obtain the laws of refraction. Thus we have to find that path APB for which AP + PB is a minimum. thus AP = PA' and AQ = A'Q where AQB is another path adjacent to APB. 3. in general.. light rays in a homogenous medium are straight lines. 11 x2+hi A i. Fermat's principle tells us that whenever the optical path length is an extremum. RN = L . we need to minimize only the path length. we must have dLop = dx =0 1/x2+h2 If(Lx)2+h2 Thus for minimum optical path length.3 Since the shortest distance between two points is along a straight line. The ray path connecting A and B will be such that ni sin 81 = n2 sin 92. P must be on the straight line A 'B. the reflected ray and the x normal to the surface at the point of incidence on the mirror sin 01 = I must be in the same plane. P and B will be in the same plane and if we draw a normal PS at P then this normal will also lie in the same plane. by definition. all nearby paths like AEB or ADB will take longer times. Let AM = hi.5.5 It should be pointed out that. To obtain the laws of reflection.4. for A'PB to be a minimum. To find the position of P on the mir ror. Then since A and B are fixed.
Example 3. for PQ' + Q ' L' to be a minimum. etc. It is for this reason that antennas (for collecting electromagnetic waves) or solar collectors are often paraboloidal in shape. However. Fig. 3. parallel to the axis of the parabola. like in radio astronomy (see Figs 3. parallel to the axis. 3.8). From the point Q' we drop a perpendicular Q'L'. 3. all rays parallel to the axis will pass through S and conversely. The laws of reflection and refraction form the basic laws for tracing light rays through simple optical systems. a paraboloid is obtained by rotating a parabola about its axis. This is the.6 (L . Let the ray path be PQ'S. 3. Show. incident at the point Q (see Fig.6). Show that all rays emanating from the point Si will pass through the point S2 after undergoing a reflection. this procedure will be quite cumbersome and as we will show. (5) becomes nl sin el = nl sin e2 (6) which is the Snell ' s law of refraction.9). Thus PQ'+Q'S=P_Q'+Q'L' Example 3. GMRT. and thus the actual ray which connects the points P and S will be PQ + QS where PQ is parallel to the axis. C e Fig.6). The GMRT consists of 30 dishes of 45m diameter with 14 antennas in the Central Array. Then. the point Q should lie on the straight line PQL. 3. 2 in the prelim pages. Therefore. reason why a paraboloidal reflector is used to focus parallel rays from a distant source.8 Fully steerable 45m paraboloidal dishes of the Giant Metrewave Radio Telescope (GMRT) in Pune. It can be shown from geometrical considerations that the reflected ray QS will always pass through the focus S.6). A colour photograph of the above figure appears as Fig. by using Fermat's principle. . like a system of lenses and mirrors. 3. one has to draw a normal at the point Q and then draw the reflected ray. 3. A colour photograph of the above figure appears as Fig.6 All rays parallel to the axis of a paraboloidal reflector pass through the focus after reflection (the line ACB is the directrix). In order to find the reflected ray. 1 in the prelim pages. In order to use Fermat's principle we try to find out the ray connecting the focus S and an arbitrary point P (see Fig. Photograph courtesy: Professor Govind Swarup. According to Fermat's principle the ray path will correspond to a minimum value of PQ' + Q 'S. Photograph courtesy: McGrawHill Digital Access Library.x) Optics and sin 02 = Thus Eq. Pune. Solution: Consider a ray PQ. that all the rays will pass through the focus of the paraboloid. the use of Fermat's principle leads us to the desired results immediately. From the definition of the parabola it follows that Q'L' = Q'S. A L L' Fig. on the directrix AB. Let L be the foot of the perpendicular drawn from the point P on AB.7 and 3.3. all rays emanating from the point S will become parallel to the axis after suffering a reflection.7 A paraboloidal satellite dish. India.2 Consider an elliptical reflector whose foci are the points SI and S2 (see Fig.1 Consider a set of rays. incident on a paraboloidal reflector (see Fig.
in a ruby laser (see Chapter 26) one may have a scheme in which the laser rod and the flash lamp coincide with the focal lines of a cylindrical reflector of elliptical crosssection. The point C represents the center of the spherical surface SPM. the sign convention is discussed later on in this problem. 3. i. [Note: We reserve the symbol R to represent the radius of curvature of a spherical surface which will be positive (or negative)_ depending upon whether the center of curvature lies on the right (or left) of the point P. The quantity r represents the magnitude of the radius of curvature which. the quantities x and y are the magnitudes of the distances. Fig. 3.10. unless the quantity inside the square brackets is zero we must have 9 = 0 implying that the only ray connecting the points 0 and Q will be the straight line path OPQ which also follows from Snell's law because the ray OP hits the spherical surface normally and should proceed undeviated.7 M Fig. of course. 0 is assumed to be smallwhich is the paraxial approximation. 1 Solution: Consider an arbitrary point P on the ellipse (see Fig. by considering the triangle SCQ we would have SQY _l r2 1_1 e2 2 r y Thus the optical path length OSQ is given by Lop =n1 OS + n2 SQ l Ir 1 L For the optical path to be an extremum we must have dep = (n x + n2 y)+ 2 r2 x+ n_n2 Y n1e2 Y _0=r2 Example 3.3 r l+ n 2_ n2Y nl 1 e Lx y (8)  Solution: From the triangle SOC we have OS = [(x + r)2 + r2 . C and Q are in a straight line.2(x + r) rcos 9]/ Thus. for Fig. 3. .. if the point I corresponds to P1= yo (see Fig. Because of the above mentioned property of elliptical reflectors. On the other hand. if the value of y was such that the quantity inside the square brackets was zero.10). 3. Consider two points 0 and Q such that the points 0. Now. y.10) then all paths like OSI are allowed ray paths implying that all (paraxial) rays emanating from 0 will pass through I and I will therefore represent the paraxial image point._ n2. no other ray (emanating from either of the foci) will pass through an arbitrary point Q which lies on the axis. assuming the angle 0 to be small.x. they are often used in laser systems.10).nt (9) r Yo x then dLop/d9 would vanish for all values of 9. i. It is well known that SIP + S2P is a constant and therefore. Similarly. Use Fermat's prin ciple to find the ray connecting the two points 0 and Q. (ii) The above considerations will remain valid even for an ellipsoid of revolution obtained by rotating the ellipse about its major axis. 3.10 SPM is a spherical refracting surface separating two media of refractive indices nl and n2. determine the paraxial image of the point 0. if y was equal to yo such that n2 nt .9 All rays emanating from one of the foci of an ellipsoidal reflector will pass through the other focus.e. happens to be R.] = [x 2 +2rx+2r 2 _2(xr+r 2 )[ 2 x (^rl+ rxr x2 LL e21 x+ ! r 2IY + 02 where we have assumed 9 (measured in radians) to be small so that we may use the expression 2 cose1 2 and also make a binomial expansion. it is neither a maximum nor a minimum but has a constant value for all points lying on the mirror. 3. we may note the following two points: (i) Excepting the rays along the axis. For example. Obviously.. As a corollary. Notice that here we have an example where the time taken by the ray is stationary.Fermat's Principle and Its Applications 3. Also. all rays emanating from the point SI will pass through S. Consider a spherical refracting surface SPM separating two media of refractive indices n1 and n2 (see Fig. C represents the center of the spherical surface. Calculate the optical path length OSQ in terms of the distances . Similarly.e. r and the angle 0 (see Fig. all rays like OSI (which start from 0 and pass through I) take the same amount of time in reaching the point I.9). 3. such a configuration leads to an efficient transfer of energy from the lamp to/ the ruby rod.
3. we may state that the product n(x) cos 0(x) n(x) sin fi(x) (14) ` + n2 SP should bean extremum n2 is an invariant of the ray path. in the limiting case of a continuous variation of refractive index..12(a)] nl OP+n2 P1 =n1 OS+n2 SI Thus. if y > yo (i. At each interface. d2Lop/d92 will vanish implying that the extremum corresponds to stationarity. In order to determine whether the ray path OPQ corresponds to minimum time or maximum time or stationary. the air near the ground has a higher temperature than the air which is much above the surface. 3. Thus u = x. Thus the refracted ray must appear to come from the point I. we must determine the sign of d2Lop/dd2 which is given by 2 2 ddb r r2 n Cx + Y n1 Fig. (13) and is an extremum.. (9) is a particular form of the equation determining the paraxial image point n2 . the light ray satisfies Snell's law and one obtains [see Fig. We will restrict ourselves to the special case when the refractive index changes continuously along one direction only. Thus..n1 v u  Optics n2 . Let P be an arbitrary point in the second medium and we wish to find the ray path connecting the points 0 and P.3 RAY PATHS IN AN INHOMOGENEOUS MEDIUM In an inhomogeneous medium. the quantity ni OS . For example. Thus. v = +y and r = +R. we have n1 OP+n2 PQ =n1 OS + n2 (SIPI+PQ) = nt OS + n2 (SI + IQ) >n1 OS+n2 SQ implying that the ray path OPQ corresponds to a maximum.12(a). the refractive index varies in a continuous manner and. in the paraxial approximation.11). 3. in general. n1 OS Thus. the ray paths are curved. For OSP to be an allowed ray path Lop . we have the stationarity condition. the point I is such that the first quantity is already an extremum thus.4 We again consider refraction at a spherical surface. On the other hand. all rays emanating from the point 0 will take the same amount of time in reaching the point I.3. one can argue that if I is the paraxial image point of P then n1 OS  n2 SI (12) an extremum. the point Q is on the right of the paraxial image point I) d2Lop /d92 is negative and the ray path OPQ corresponds to maximum time in comparison with nearby paths and conversely. Alternatively. We will use Snell's law (or Fermat' s prin ciple) to determine the ray paths in an inhomogeneous medium. 3. Y Yo Obviously. We may therefore say that for a virtual image we must make the quantity n1 n2 r  =r2 n2 [1  11. when Q lies on the right of the point I.n2 SI is independent of 0 (11) sin O = n2 sin 02 = n3 sin 03 = . The inhomogeneous medium can be thought of as a limiting case of a medium consisting of a continuous set of thin slices of media of different refractive indicessee Fig. This leads to the phenomenon known as mirage. Similarly. if y = yo. when Q lies on the left of the point I then the ray path OPQ corresponds to a minimum and when Q coincides with I.11 The refracted ray is assumed to diverge away from the principal axis. the piecewise straight lines or Lop = (n1 OS  SI) + n2 (IS + SP) should be an extremum where we have added and subtracted n2 SI. Thus. however.e. 3. 1 Example 3. + SI should be an extremum and therefore it should be a straight line. Now.n1 R (10) with the sign convention that all distances measured to the right of the point P are positive and those to its left negative. the refracted ray is assumed to diverge away from the principal'axis (see Fig. we will denote this invariant by 16 The value of this invariant may be determined from the fact that if the ray initially makes an angle e1 (with the zaxis) at a point where the refractive index is then the value ' of /3 is n1 cos e1.8 We should mention that Eq. on a hot day. Let us consider paraxial rays and let I be a point (on the axis) such that n1 OS . we assume this direction to be along the xaxis. the quantity SP . for paraxial rays.n2 SI is independent of the point S. Since the density of air decreases with increase in temperature. the refractive index increases continuously as we go above the ground.
7'e To ne ne so that 9e = (16) 2(1 . the ray path bends in such a way that the product n(x) cos 0(x) remains constant [see Fig. In Fig.13 Ray paths in a medium characterized by a linear variation of refractive index [see Eq.13.z (m) 1000 1200 shown in Fig. This is therefore called the shadow region.5 m above the ground. 0. 3. an eye in this region can neither see the object nor its image. .5 m and the curves correspond to + 0.325°. At the eye position E (x = x e ).234 x 105 m l.2°.4). thus. On a typical hot day the temperature near road surface To 323 °K (= 50 °C) and.. 0.00026 giving 9e 5. The object point is at a height of 1. and if at that point the ray makes an angle 9e with the horizontal then 73 = no = ne cos 9e (17) Usually 9e << 1 so that no =cos 0 e =126 Be At constant air pressure 112(1  n1 ne (18) (no . the refractive index variation can be approximately assumed to be of the form n(x) no + kx 0 < x < few metres or ne . We consider a ray which becomes horizontal at x = 0.12(b)]. the eye at E will see the mirage and not see the object directly at P. As mentioned earlier. 0 200 400 600 800 . 0°.13 we have shown rays emanating (at different angles) from a point P which is 1. 3. 0. at 30 °C.ne (1 To (20) where no is the refractive index of air at x = 0 (i. The exact ray paths (see Example 3.2°. just above the ground) and k is a constant. Te ti 303 °K (= 30 °C).3486° and 0. 3.no _ ne1 1. (19) we get * For more details. (b) For a medium with continuously varying refractive index.8) are shown in Fig. 3. n(x) cos 9(x) = n1 cos 01 = 73 Fig. about 1. no i = 1 __z e0 3. the only ray path connecting points P and E will be along the curve PME and that a ray emanating horizontally from the point P will propagate in the upward direction as PC as shown in figure.Fermat's Principle and Its Applications 3.(16)] with k = 1. if the refractive index is ne. Equation (15) can be used to derive the ray equation (see Sec.12(a) form a continuous curve which is determined from the equation (15) implying that as the refractive index changes. thus each ray has a specified value of the invariant /3 (= n1 cos 01). the ray bends in such a way that the product n.1 ) To = (ne 1)Te (19) From Eq.3.12 (a) In a layered structure. Thus. there is also a region R1 where only the object is directly visible and the virtual image is not seen. ne = 1.5° The shading shows that the refractive index increases with x. see Refs 48. 3. Now. the ray path bends in such a way that the product n(x) cos 9(x) remains constant.5 m above the ground. in such a condition. Indeed. cos 9i remains constant. The figure shows that when the object point P and the observation point E are close to the ground. We also find that there is a region R2 where none of the rays (emanating from the point P) reaches. on a hot day the refractive index continuously decreases as we go near the ground.9 Fig.67 x 103 radians = 0. 3.28°.e. Furthermore.1 The Phenomenon of Mirage* We are now in a position to qualitatively discuss the formation of a mirage. 3.
the upper edge will travel with a smaller speed in comparison with the lower edge. (15) because at such a point. 3.14. 1r/16. 0 = 0 and one may expect the ray to proceed horizontally beyond the turning point as shown by a dotted line in Fig. In Fig.3. 3. 3.06455 (= nl) and different rays correspond to different values of 01. . it approaches (n(2 + exact ray paths are obtained by solving the ray equation (see Example 3. x represents the height above the ground. the angle that the ray makes with the zaxis at the point P.14 and 3.13.13).x> 0 (21) The actual values of the refractive index for parameters given by the above equation are not very realisticnevertheless.43 m and 2. tc/ 30.14 and 3.n (x) Fig.15. they correspond to the following values of various parameters: and no = 1.10) and are shown in Figs 3.8 m and the curves correspond to 01 (the initial launch angle) = 0. ic / 10 and ic/8.14 Ray paths in a medium characterised by Eqs (21) and (22). 3. The shading shows that the refractive index increases with x. lc/ 30.000233.15 show the ray paths emanating from points that are 0.45836 a = 2.. The refractive index at x = 0 is ) The no and for large values of x. 0. Jr/ 60.00 0 6 z(m) . 't / 60. lc/ 15 and jc / 10. Figures 3. . 3.05 1. The object point is at a height of 2.15 Ray paths in a medium characterized by Eqs (21) and (22).10 We should mention here that the bending up of the ray after it becomes parallel to the zaxis cannot be directly inferred from Eq. a straight line path like BB' does not correspond to an extremum value of the optical path.v/ 11.303 m1 (22) e"x).8 m above the ground respectively. it allows us to understand qualitatively the ray paths in a graded index medium.43m) and the curves correspond to e1 (the initial launch angle) = + r/10. The shading shows that the refractive index increases with x. it immediately follows that the ray path should be symmetrical about the turning point and hence bend up. From Fig. Fig.10 1. the bending of the ray can be understood by considering a small portion of a wave front such as W (see Fig. and this will cause the wave front to tilt (see W') making the ray to bend. from considerations of symmetry and from the reversibility of ray paths. Furthermore. n2 and a are constants and once again.14 we again see that when the object point P and the 1. n2 = 0. However. the point at which 0 = 0 is known as the turning point. We next consider a refractive index variation which saturates to a constant value as x f 00: n2 (x) =no+n2 (1  Optics where no. The object point is at a height of 1 / a (=0. Physically. the point P corresponds to a value of the refractive index equal to 1. 3.
poznan. The actual formation of mirage is shown in Figs.2 m. an eye in this region can neither see the object nor its image. Once again. /3= 1. Thus for a ray launched with 91 = tr/8. A suitable refractive index variation for such a case can be written as: n2(x) = no + n2 e' (23) The equation describing the ray path is discussed in Problem 3.14. S.15. above cold sea water. If the eye is at E.3 The Graded Index Atmosphere One of the interesting phenomena associated with imaging in a graded index medium is the noncircular shape of the setting . in such a condition. 3.16 A typical mirage as seen on a hot road on a warm day. 3.17. A few seconds later (notice the motion of the bird to the left of the Sun!). thus.18. the only ray path connecting points P and E will be along the curve PME and that a ray emanating horizontally from the point P will propagate in the upward direction. photograph adapted from http://fizyka. if points P and E are much above the ground (see Fig.13.5 As an example.14) then n1 cos 91 = 1. Thus.jpg.073 m n2 Fig. A colour photograph of the above figure appears as Fig.17 This is actually not a reflection in the ocean. Naval Observatory and are on the Naval Observatory's website. the eye will see the object almost directly (because of rays like PCE) and will also receive rays appearing to emanate from points like P'. 3.20). For an object point P at a height of 0. 3. 3.11 observation point E are close to the ground. at x = 0.01627 Thus if the ray becomes horizontal at x = x2 then n (x2) = and x = a 1n 1.19 and 3.1 2 73 = 1.3.n2(x2)nol 2. but the miraged (inverted) image of the Sun's lower edge. Ref: http://mintaka.6 In Fig. 3.edu/GF/explain/simulations/infmir/Kaplan_photos. 3. if 91 represents the angle that the ray makes with the zaxis at the point P (see Fig.8 m where n(x) = 1. We assume the values of no. for an object shown in Fig. n(x) = 1.03827 Example 3. 3.html . 3 in the prelim pages.(22).2 The Phenomenon of Looming The formation of mirage discussed above occurs due to increase in the refractive index of air above the hot surface. 3. the object cannot be observed directly. then it will receive rays appearing to emanate from P'.3. Such a phenomenon in which the object appears to be above its actual position is known as looming. 3. the eye at E will see the mirage and not see the object directly at P.2 in.03827 x cos 0 implying 91 = 13° Further. shown as PC in Fig. It may be readily seen that different rays do not appear to come from the same point and hence the reflected image seen will have considerable aberrations. A colour photograph of the above figure appears as Fig.5 m. for the ray which becomes horizontal at x = 0. there is a shadow region R2 where none of the rays (emanating from the point P) reaches there.14.03827 Thus.1. 4 in the prelim pages. 3. n2 and a to be given by Eq.1 cos 91 =1. let us calculate the angle at which the ray should be launched so that it becomes horizontal at x = 0.15).16 and 3. The photographs were taken by Dr.sdsu.put. Example 3.Fermat's Principle and Its Applications 3.phys. the ray paths are shown in Fig.pl/pieransk/ Physics%20Around%2OUs/Air%20mirror. since no other rays emanating from P reach A. Now. George Kaplan of the U. However. the air near the water surface is colder than the air above it and hence there is an opposite temperature gradient. the object point corresponds to x = 2.01627 L = 0. On the other hand. Fig. it is commonly observed in viewing ships over cold seawaters (see Figs 3. Moreover. the reflection fuses with the erect image.2 m the value of the invariant is given by /3 = 1.
fi/netcomm/news /showarticle. It is for this reason that the setting sun appears flattened and also leads to the fact that the days are usually about 5 minutes longer than they 3.5 m.com/ see/weather/elements/miragel. (22). This condition is common over snow. n2 and a are given by Eq. Figure adapted from http: / /www. Obviously. A colour photograph of the above figure appears as Fig.18 Ray paths corresponding to the refractive index distribution given by Eq. The refractive index of the air gradually decreases as we move outwards. We will restrict ourselves to the special case when the refractive index changes continuously along only one direction. if we were on the surface of the moon. ice and cold water surfaces. the solution of which will give the precise ray paths in an inhomogeneous medium. would have been in the absence of the atmosphere.3. For it to be seen. the values of no.21 and Fig.finland. (23) for an object at a height of 0.12 Optics z(m) Fig. which we assume to be along the xaxis. Thus the sun (which is actually at S) appears to be in the direction of S'. 3.islandnet. thus tricking our eyes into thinking an object is located higher or is taller in appearance than it actually is. the air close to the surface must be much colder than the air above it. 8 in the prelim pages). Fig. 3. 3. This can easily be understood in the following manner.asp? intNWSAID=25722.22. the rising or the setting sun would not only look white but also circular in shape! or the rising sun (see Fig: 3.20 A house in the archipelago with a superior mirage. we will derive the ray equation.4 THE RAY EQUATION AND ITS SOLUTIONS In this section. If we approximate the continuous refractive index gradient by a finite number of layers (each layer having a specific refractive index) then the ray will bend in a way similar to that shown in Fig.19 The superior mirage occurs under reverse atmospheric conditions from the inferior mirage. light rays are bent downward toward the surface. 6 in the prelim pages. A colour photograph of the above figure appears as Fig. Figure adapted from http://virtual . When very cold air lies below warm air. Fig. 3. 5 in the prelim pages.htm. This medium can be thought of as the limiting case of a medium comprising of a continuous set of thin slices of media of dif .
if we refer to Fig. 3. ferent refractive indices. In such a case. then (ds) 2 = (dx) 2 + (dz) 2 or where X x+ no and x= k (31) ( ds  )^=Cd ds + (25) Thus the ray path is given by x(z) Now.12(a) forms a continuous curve as in Fig. the ray equation [Eq. as it ought to be in a homogeneous medium. the RHS of Eq. (28)] takes the form (24) or d2X = xz X(z) dz2 (30) d2x dz2 1 dn2 = [no + kx] z 2/32 dx /1 Furthermore.Fermat's Principle and Its Applications 3. 3.13 Thus Eq. Example 3. 3.12(b). (28) is zero and one obtains d2x = 0 dz2 Integrating the above equation twice with respect to z. we find that dz = no + CI e" + C2 e ' (32) = cos 0 = n(x) (26) where the constants Cl and C2 are to be determined from initial conditions. If ds represents the infinitesimal arc length along the curve.21 The noncircular shape of the setting sun (see also Fig. As discussed earlier. (28). let us con• sider a homogeneous medium for which n(x) is a constant. (27) in a slightly different form by differentiating it with respect to z: 2 dxd 2 x dz dz2 __ 1 dn2 dx 1:3 2 dx dz 1 or Fig. for a continuous variation of refractive index. Eq. we obtain x = Az + B which is the equation of a straight line.7 As a simple application of Eq. 3. for a continuously varying refractive index. 3. the product n(x) cos 0(x) is an invariant of the ray path which we denote by /3 : n(x) cos 0(x) = /3 Example 3. light from S appears to come from S'.12(b). thus . the ray is launched at x = x1 making an angle 91 with the zaxis. (27) can be integrated to give the ray path x(z). the piecewise straight lines shown in Fig. however. (25) becomes dzJ J Z = n2(x) 1 73 2 (27) For a given n(x) variation. d2x dz2 = dn2 dx 2/3 2 (28) Both Eqs (27) and (28) represent rigorously correct ray equations when the refractive index depends only on the xcoordinate.22 Because of refraction. We assume that at z = 0.8 We next consider the ray paths in a medium characterized by the following refractive index variation n(x) = no + kx (29) For the above profile. it is often more convenient to put Eq. 8 in the prelim pages). Fig.
In2 (x)a 2 /3 Substituting for n2 (x). cos 01 1 and all rays have the same periodic length. Indeed. 8. the ray will be incident at the corecladding interface at an angle and the ray will be refracted away.00026 and k 1. (27) to determine the ray paths. . 3.4. the ray path will become horizon f = ± 1 dz f (37) tal at the corecladding interface.23.14 x(z = 0) = xi and dx = tan 91 dz z=0 Optics We may mention here that we could have also used Eq.410 (< n2 )the last ray undergoes refraction at the corecladding interface. the rays will be guided in the core if n2 < l3 <n1 .485 (= n2) and 1. nl = 1. in the paraxial approximation.5.23 (with 01= 4° and 8. we have plotted typical paraxial ray paths for rays launched Ray paths in such media are of tremendous importance as they readily lead to very important results for parabolic index fibers which are extensively used in fiberoptic communication systems (see Sec.8796 mm respectively.23. the ray paths shown correspond to = 0 and 01 = 4°.Xi) sin [r(z .3.25. 3. 1 1 1 nl  )2 3 (39) and r=Y (40) the corresponding values of /3 are approximately 1. Thus nl 20 Y= a (44) In a typical parabolic index fiber. 1. (43) I x I < a cladding .234 x 10 5 ml /3 = n1 cos = 2 [xi + k (n +n sin 0 )] = 2 + k sin 0 0 l 1 [xl (n0 . (28) to obtain the ray path.13°) the values of zp would be 0. Now. It may be readily seen that the periodical length zp of the sinusoidal path is given by zp = Writing x = x0 sin 0 and carrying out the straightforward integration we get x = ^. Obviously. 1 )^ (34) The region Ixl< a is known as the core of the waveguide and the region Ixl> a is usually referred to as the cladding. 27.13° and 20°.1 Ray Paths in Parabolic Index Media We consider a parabolic index medium characterized by the following refractive index distribution: 12 2(x) and y 1. In Fig. 0 = 0.nl (33) n2 (x)=n1[l201] . (32) with x1 = 1.01 .13 shows the ray paths as given by Eq. Ixl<acore = n2 = nl (1. For /3 < n2.z0)] (41) We can always choose the origin such that z0 = 0 so that the general ray path would be given by x=±x0 27r r tic a cos e 1 2D (47) sinrz (42) Thus for the two rays shown in Fig. n1 = 1. 3.um giving n2 1.485 (45) 3. x0 =. 3. a = 20 . in an optical waveguide the refractive index distribution is usually written in the form*: 2 Elementary manipulations would give us C1 and C2 where n1 = no + kx1 represents the refractive index at x = x1 and we have used the fact that 91 (35) Figure 3. we get = n2. we may write n2 < /3 < nl Guided rays Refracting rays (46) f where 1 x0x 2 `x r fdz (38) z0 /3 < n2 In Fig. Equation (27) can be written as .20). Thus. When /3 We will use Eq.0607 x 104 m1 = ni  y2 x2 (36) Typical ray paths for different values of 01 are shown in Fig.7).5 m.496 (> n2).8864 mm and 0.
4. Physically.wikipedia. Such small size lenses find many applications. 3.12 (b)]: .24 2 2a A gradientindex lens with a parabolic variation of refractive index.2 for exact calculation). 3. 3.25). figure adapted from http: / / en.25. Such a calculation is of considerable importance in fiber optic communication systems (see Sec. A colour photograph of the above figure appears as Fig. t Fig. for instance.25 Paraxial ray paths in a parabolic index medium. all rays must take the same time to go from P to Q.23 Typical ray paths in a parabolic index medium for parameters given by Eq. the ray path (inside the core) is given by x = x0 sin Fz (49) where x0 and F have been defined through Eqs (39) and (40). 3. (26)].13° and 20°. typically zp = few cm and the diameter of the lens would be few millimeters. 3. For example a GRIN lens can be used to couple the output of a laser diode to an optical fiber. As shown in Sec. along the zaxis. (iv) The rays periodically focus and defocus as shown in Fig. Four interesting features may be noted: (i) In the paraxial approximation (/3 ni) all rays launched horizontally come to a focus at a particular point. a GRIN lens of length zp/2 can be used to transfer collimated light from one end of the lens to the other. 3. (36). Different rays (shown in the figure) correspond to different values of /3 . all rays emanating from P will focus at Q and if we refer to our discussion in Example 3. 27. 27. the rays emerging from point P) get trapped in the medium and hence the medium acts like a `guide'.4.1.org/wiki /Refractive_index . (iii) Ray paths would be allowed only in the region where /3 is less than or equal to n(x) [see Eq.4.e. the length of such a GRIN lens would be zp/4 (see Fig. (27). Further. characterized by parabolic variation of refractive index in the transverse direction. It is for this reason that parabolic index waveguides are extensively used in fiberoptic communication systems (see Sec. 20 0 I° zp Z (mm) Fig. the ray would become parallel to the zaxis) when n(x) equals /3 . Thus the medium acts as a converging lens of focal length given by: (4g) D (ii) Rays launched at different angles with the axis (see. although the ray PLQ traverses a larger path in comparison to PMQ. are now commercially available and find many applications (see Fig. it does so in a medium of `lower' average refractive indexthus the greater path length is compensated for by a greater `average speed' and hence all rays take the same time to propagate through a certain distance of the waveguide (see Sec.7). 3.7). Indeed such media are referred to as optical waveguides and their study forms a subject of great contemporary interest.2 Transit Time Calculations in a Parabolic Index Waveguide In this section we will calculate the time taken by a ray to traverse a certain length through a parabolic index waveguide as described by Eq. Similarly. Let da represent the time taken by a ray to traverse the arc length ds [see Fig.15 We may mention here that GradientIndex (GRIN) lenses.3. 3. 3..24). dx/dz would be zero (i.Fermnat's Principle and Its Applications 3. x z(mm) Fig. 8. Notice the periodic focussing and defocussing of the beam.(45) for 01 = 4°. this immediately follows from Eq. 3. 7 in the prelim pages. In the paraxial approximation. The lens focuses light in a way similar to a conventional lens.
zo) 4 z<0 where we have used Eq. (50) as da = For the fiber parameters given by Eq.y2 xo sine r z] dz c where in the last step we have used Eq. c c n2 (x) dz Example 3.gx (58) (51) [nl .16 ds c/n(x) Optics where in the last step we have assumed (50) 0 da = n2 n2 nl n2 l 2n1 = 0. z = 0) as shown in Fig.26 0 2 ^z ( n1) Z(N = n2)  TO' = nl) . 3. (49). 3. Thus if 2(z) represents the time taken by the ray to traverse a distance z along the waveguide then or nl j T _ J c/3 o Thus. n2(x) decreases linearly with x and Eq.26. (26)] we may write Eq. (45). ($ A2 (55) or AT = 2c (nl n2 _n2)2 Parabolic ray paths (corresponding to 01 20°. 45° and 60°) in a medium characterized by refractive index variation given by Eq.z2 (z . in the region x > 0. (52) would make a negligible contribution to r(z) and we may write 'C(Z) = (62) 2 P+ 2 nl . the pulse will get temporally broadened. if a pulse of light is incident on one end of the waveguide. 30°. Thus K2 =0 /3 = n1 cos e1 (60) (61) Further.3.21 sin 21 (52) 22 x Consider a ray incident on the origin (x = 0. When /3 = sponds to the ray along the zaxis) (which cone (53) 2'(z) = c/n i which is what we should have expected as the ray will always travel with speed c/n 1 . The ray paths in the region x < 0 are straight lines. it would in general excite all rays and since different rays take different amounts of time. (39).25 ns/km n2 (56) where c is the speed of light in free space. 2^2 dz2 The general solution of which is given by x(z) dz  y2 xo c/3 J j^ 1. For large values of z. the second term on the RHS of Eq. this broadening will be given by 42= 2 Fig. Z (54) Now. we get AT (57) We will use this result in Chapter 27. Thus. for a parabolic index wayeguide. Since dz n(x) ds = f3 [see Eq.9 We next consider the ray paths in a medium characterized by the following refractive index variation l[ni y 2 x2 ]dz n2 (x) = ni x<0 x >0 = dT =nl .(z . .(58).zo) 2 4 z _ .cos (2rz) dz 2 0 = 4 ^2 and z2 +K1 z+K2 (59) c/3 [nl Y Z 2 2 YZ xo ] z + 2c/3. or 2(z) = 1 l + 2]z + (14162 ) sin 2rz 2c/3 4 cy n1 dx =K1 =tan 91 dz z=0 Thus the ray path will be given by (tan O1 )z x(z) =. (28) takes the form g d2x =.
854x10 12 x(2Trx4. 0 = m/2 and nT = 0 implying mep o. then of = c (66) Thus as the electron density starts increasing from 0 (beyond the height of 60 km) the refractive index starts decreasing and the ray paths would be similar to that described in Example 3. Thus. by measuring the delay At. the echo will be received after about 670 µs. This is how the short wave radio broadcasts (X.28 we have plotted the frequency dependence of the equivalent height of reflection (as obtained from the delay time of echo) from the E and F regions of the ionosphere.6x 10 19 ) 2 = 2.60 x 1019 C represents the charge of theelec tron m = 9. The shading shows the variation of electron density. the ray path is a parabola.9.27) /3 = cos 01 = nT (64) where h represents the height at which it undergoes reflection.6x10 6 ) 2 (1.= 20 m) sent at a particular angle from a particular city (say London) would reach another city (say New Delhi) after undergoing 9.854 x 1012 x l0 12 C 2/Nm 2 represents the dielectric permittivity of vacuum (65) In a typical experiment.5 to 20 MHz) is sent vertically upwards and if the echo is received after a delay of At seconds. an electromagnetic pulse (of frequency between 0. Alternatively.4x10 100 km 3.3 Reflections from the Ionosphere The ultraviolet rays in the solar radiation results in the ionization of the constituent gases in the atmosphere resulting in the formation of what is known as the ionosphere. Ne (100 km) mEo (2nv) 2 2 Thus if an electromagnetic signal is sent from the point A (at an angle 01) is received at the point B.6 x 106 Hz. 3. for normal incidence.6 x 10 11 electrons/m3 . Typical ray paths are shown in Fig. 3.11x10 31 x 8. 3. and Eo = 8.5.27 Reflection from the E region of the ionosphere.26.11 x 1031 kg represents the mass of the electron. (76) of Chapter 7): n2 (x) Al N 8 Fig. Thus if electromagnetic pulse is reflected from the E layer of ionosphere (which is at a height of about 100 km). The point T represents the turning point. one can determine the height (at which the pulse gets reflected) from the following relation h ZLt '(67) In Fig. the refractive index is given by (see Eq. the calculations corresponds to n1 =1. The ionization is almost negligible below a height of about 60 km. reflection from the ionosphere. From the figure we find that at v = 4.1m 1 and different rays corresponds to 01 = 9' 6' 4 and 3 180 km J 11 m 3 (Ne)max.4. Further. g=0. 3.17 F region where 2 zo = ii sin 201 g Thus in the region 0 < z < zo. echoes suddenly disappear from the 100 km height.) 2 Ne(xT) = q 2 =1 N e( x) q2 mEo a)2 (63) where Ne (x) represents the number of electrons/unit volume in m3 x represents the height above the ground in meters represents the angular frequency of the electromagnetic wave q 1. one can determine the refractive index (and hence the electron density) of the ionospheric layer where the beam has undergone the reflection.Fermat's Principle and Its Applications 3. We should mention that if nT represents the refractive index at the turning point (where the ray becomes horizontal) then (see Fig. Because of the presence of the free electrons (in the ionosphere).
For more details of the studies on the ionosphere.p2 ) 1/2 n2 fi(x) = Ke ax12 3. 9].10 In this example we will obtain the solution of the ray equation for the refractive index variation given by n2(x) = no + 14 = 1 cosh1 (Keax1/2 ) 7 (75) (1 . Since x = x1 at z = 0 (the initial point) C Further. Further./3 2 2 n0 + nP n12 Example 3. Ray paths in biaxial media are discussed in Ref. 11. (69) correspond to a ray going up and a ray going down respectively.n0 +nP . 3. 10).e a`) /3dx (68) Substituting in Eq. = a 1n[ K a K n2 cosh y(C ± z)] (73) where 7= (74) 2/3 which gives us the ray path. Carrying out the elementary integration. C = 0. 2 1/2 2 2 K e ax h /2 . F region of the ionosphere.sign in Eq.3.n2 [K2 eax _ 1' 1/2 Thus for a ray launched horizontally at x = x1. . [Adapted from Ref. however. Mitra [Ref. is quite complicated. K. some of the The + and .14 and 3.15. 9] x(z) If we further increase the frequency. z = 0 and n1= n(x 1).28 Frequency dependence of the equivalent height of reflection from the E and F regions of the ionosphere. typical examples are glass. (27) we would obtain ±dz = x 1/2 (76) {(np + n2 /32) n2 e eaxl2 dx . the proof.5 REFRACTION OF RAYS AT THE INTERFACE BETWEEN AN ISOTROPIC MEDIUM AND AN ANISOTROPIC MEDIUM In this section we will use Fermat's principle to determine the direction of the refracted ray for a ray incident at the interface of an isotropic and an anisotropic medium*. we get * A proof for the applicability of Fermat's principle in anisotropic media has been given by Newcomb (Ref. in an anisotropic medium. the echoes appear from the. water and air. or Where and d'1 ±dz= K an2 (02 _ 1) 1/2 2/3 (69) (70) (71) K = 1 (no + n2 . On the other hand. Typically ray paths (for different values of 91) are shown in Figs 3. the reader is referred to one of the most outstanding texts on the subject by Professor S. We may point out that in an isotropic medium the properties remain the same in all directions.18 Optics 300 200 100 0 1 2 3 4 5 6 7 8 9 10 Frequency of the exploring waves in MHz Fig. (72) /3 = n1 cos ei is the angle that the ray initially makes with the zaxis at where e1 x = x1.
Fermat's Principle and Its Applications properties (such as speed of light) may be different in different directions. 3.29.3. We will now use Fermat's principle to study the refraction of a ray when it is incident from an isotropic medium into an anisotropic mediumboth media are assumed to be homogeneous. the refractive index variation for the [ extraordinary ray is given by [see Eq. Thus the ordinary ray obeys Snell's laws but the extraordinary ray does not. then no = 1. it (in general) splits .5. (121) of Chapter 22] n2(0) = no cost 0 + ne sin2 0 (77) x + x2 319 cos 0 = (z +112 x2)112 we have Lop = nl [h? + (L x)]112 + [n o h Z + ne x2]112 (79) For the actual ray path. we may mention here that when a light ray is incident on a crystal like calcite. In a uniaxial medium.29 . with the normal.nl sin 2 i using which we can calculate the angle of refraction for a given angle of incidence (when the optic axis is normal to the surface). however.1 Optic Axis Normal to the Surface We first consider the particularly simple case of the optic axis being normal to the surface. we assume the first medium to be air so that n1 = 1. 3.We next consider a more general case of the optic axis makmal to the surface. when 0 = 0). we will consider anisotropic media in greater detail. (80) simplifies to nl sin a = n2 sin r which is nothing but Snell's law.5.e. 3. its speed is c/no and when it propagates perpendicular to the optic axis (0= n/2) its speed is c/ne.o dx implying nl (Lx) _ [hl +(Lx) 2 ]1 12 or l sin l = ne tan r [no + ne tan 2 /2 (80) .sin 2 i If we assume the second medium to be calcite.1° It may be seen that if no = ne = n2 (say) then Eq. Then tan r = no sin i ne jne . we readily get r = 31. .48641 (when nl = 1) (82) (81) a Fig. Since Simple manipulations give us tan r = no nl sin i ne \Ine .nex oh2 +nex 2]1l2 2 2 and sin 0 = )u2 n where no and no are constants of the crystal and 0 represents the angle that the ray makes with the optic axis. In Chapter 22.65836. we must have dL0 . ing an angle . Obviously. rJ where we have used the fact that the angle of refraction r = 0 and tan r = . As a simple example. the refracted ray and the optic axis to lie in the same plane. (83) ne = 1. The velocity of the ordinary ray is the same in all directions.2 Optic Axis in the Plane of Incidence* The direction of the refracted extraordinary ray when the optic axis (of the uniaxial crystal) is nor. Referring to Fig. the optical path length from A and B is given by 2 ] 112 + x2]112 (78) Lop = n 1[ h i + (L .. and Thus for i = 45°.into two rays known as ordinary and extraordinary rays.x) n(0)[hi + where nl is the refractive index of medium I and we have assumed the incident ray. the optic axis is * May be skipped in the first reading. when the extraordinary ray propagates parallel to the optic axis (i.
20 assumed to lie in the plane of incidence as shown in Fig. the refracted ray does not lie in the plane of incidence. it can be shown that if the optic axis lies in the plane of incidence then the refracted ray also lies in the plane of incidence.3. we must have dLop 0 dx implying nl (L. i = 0.x) [h? +(Lx) 2 ]'2 2 no (h2 cosO+ x sin 0) sin ¢ + 11.x) 2 ] 1/2+ h2 sin 0) 2] 'i2 (86) [no (h2 cos 0 + x sin 0)2 + ne (x cos . the above equation can be solved to give the values of 0 and hence the angle of refraction r (= 9 + ^).30) is given by x2 ] '/2 (84) Lap = nl (g.. (80). Eq. (87) simplifies to Fig. the optical path length from A to B (see Fig.h2 sin 0)2 or nl sin i = no cos9 sinO+ne sine cost [no cos" 9 + n^ sin2 611/2 (87) For given values of the angles i and 0. 3.h2 sin 0)2] 112 (85) and Lop = n l [hi + (L .. In the present calculation. Eq.h2 sin 0) cos cp ]'i2 [no (h2 cosO + x sin0) 2 + n^ (x cosO . i.e. we consider normal incidence. (ii) When 0 = 0. in an anisotropic medium. nl sin i = n2 sin (9 + 0) = n2 sin r which is nothing but Snell's law.e. However.0) sin ^i + ne sin (r . (87) gives us ne2 Thus n(9) = cos0 ^h2 +x2 2 sink or no cos 0 sin 0+ n! sin 9 cos 0= 0 no cos (r . The above equation is identical to Eq. 3. (87) becomes nl sin i = no Since 0 = r .30 The direction of the refracted extraordinary ray when the optic axis (of the uniaxial crystal) lies in the plane of incidence making an angle with the normal to the interface. Some interesting particular cases may be noted. i. sin2 0 + ne + ne (x cos 0 .+ n. the anisotropic medium becomes isotropic and Eq. Thus.0) cos 0 = 0 I h22 +x2 1 [no (h2 cos 0 + x sin 0)2 or cos r [no cos 0 sin 0 . sin2 (r = no (cos r cos + sin r sin 0) 0)2 + ne sin 9 91 1/2 cos' e + ne sin ne sin r r]112 [no cos' r + no' sin2 (88) ne (sin r cos x 0  cos r sin 2 0)2 sin cos O + = no h2 2 h22 + x2 ljh2 + x2 x I 1h2 +x2 + where we have used the fact that r = 0. We may mention here that in general. + (L . (iii) Finally.0. we are assuming this and finding the direction of the refracted ray for a given angle of incidence.30.x) 2 ] u2 + n(9)[h2 + Optics For the actual ray path. (x cos ¢ . the optic axis is normal to the surface.ne sin 0 cos 0] + sin r [n. we have n2(0) = no cos2 (r . (i) When no = ne = n2. 3. Now.
22. 3. The exact ray paths are determined by solving either of the equations: dx where u and v are the object and image distance and R is the radius of curvature with the sign convention that all distances to the right of P are positive and to its left negative. y2and r and the angle 9 are defined in Fig. I where x = 0. V11 2 (x)/3 2 =± dz or d2x _ 1 dn2 (x) 2/3 2 dx dz2 Fig.Fermat's Principle and Its Applications 3. ♦ For an inhomogeneous medium characterized by the refractive index variation n(x). 3. the above analysis is valid for an arbitrary orientation of the optic axis.16(b)]. ♦ In a parabolic index medium n2 (x) = n y2 x2. Consider an arbitrary point Q on the axis of the system and using a method similar to that used in Example 3. Summary 3. 0 is assumed to be small. ♦ Laws of reflections and Snell's law of refraction (n1 sin 01= n2 sin 02. the refracted extraordinary ray undergoes finite deviation. ♦ Fermat's principle can be used to study refraction of rays at the interface of an isotropic medium and an anisotropic medium. Determine the paraxial image point and show that the result is consistent with the mirror equation 1+1 = 2 (91) u v R • ♦ The slightly modified version of Fermat's principle is: the actual ray path between two points is the one for which the optical path length is stationary with respect to variations of the path.32. Rays launched at different angles take approximately the same time in propagating through a large length of the medium. when the crystal is rotated about the normal. where 01 and 02 represent the angles of incidence and refraction) can be derived from Fermat's principle. Problems Optic axis Fig.31)./32 and we have assumed z = 0 /3 Y Equation (89) shows that in general r ^ 0 (see Fig. looming and also reflections from the ionosphere. here 0(x) is the angle that the ray makes with the zaxis.21 where the invariant /3 is determined from the initial launching condition of the ray. Furthermore. the ray paths are sinusoidal: x(z) = ± x0 sin Fz where I' = 1 . in general.3.z+ 1 . the refracted (extraordinary) ray lies in the plane containing the normal and the optic axis. for normal incidence. 3. ♦ Ray paths obtained by solving the ray equation can be used to study mirage. 2 .1 In this and the following two problems we will use Fermat's principle to derive laws governing paraxial image formation by spherical mirrors. this constant is denoted by /3 which is known as the ray invariant. for normal incidence. the refracted ray also rotates on the surface of a cone [see Fig. We may mention here that. the ray proceeds undeviated when the optic axis is parallel or normal to the surface.32 Paraxial image formation by a concave mirror. Consider an object point 0 in front of a concave mirror whose center of curvature is at the point C.31 For normal incidence. 3. x0 = 1 . However. . (89) we note that when the optic axis is normal to the surface (0 = 0) or when the optic axis is parallel to the surface but lying in the plane of incidence (0 = 7r/2). the ray paths [x(z)] are such that the product n(x) cos 0(x) remains constant. r = 0 and the ray goes undeviated. Returning to Eq.292 y (90)__ where the distances x. show that the optical path length Lop (= OS + SQ) is approximately given by Lap = x+Y+ r2 .
We now have to consider the optical path length Lop = n1 OS .n2 SQ. M Fig.6 C is the center of the reflecting sphere of radius R (see Fig.1 +16 2 (92) Fig.3 Proceeding as in the previous problem. Consider an object point 0 forming a virtual image at the point I.7 SPM is a spherical refracting surface separating two media of refractive indices n1 and n2. y and r and the angle 6 are defined in Fig. 3. Show that Lop= OSSQ=xy+Zr 2 1 Optics JJ Y [ where the distances x.SQ. (10). We assume that all rays emanating from 0 appear to emanate from I so as to form a perfect image. 3. 3. the image distance v and the radius of curvature 1? are negative since the image point and the center of curvature lie on the left of the point P. ellipsoid pf revolution will focus to one of the focal points of the ellipse provided the eccentricity = nr/n2 . In this case also one obtains a virtual image.34). v and R are all negative quantities because they are on the left of the refracting surface. provided the eccentricity of the ellipse equals nt /n 2 . One should now assume the optical path length Lop to be OS .r .36). 3.33). 3. Show that the paraxial image is formed at y = yo which is given by . PI and P2 are two points on a diameter equidis SQ 2 y x 1 2 [n2 nl n2 n l xn2 y. 3. Show by using Fermat's principle that all rays parallel to the major axis of the ellipse will focus to one of the focal points of the ellipse (see Fig. we must have + OP2 n1 OSn2 S1=n 1 OPn2 PI . (a) Obtain the optical path length Pro as a function of 0. 3.22 3. and (b) find the values of 6 for which P I OP 2 is a ray path from reflection at the sphere. show that it is given by Lop =n1 OSn 2 Fig.2 Fermat's principle can also be used to determine the paraxial image points when the object forms a virtual image.35). 3. QB+n 2 BC and show that the point B (whose coordinates are x and y) lies on the periphery of an ellipse). that u. (91) because whereas the object distance u is positive. use Fermat's principle to determine the mirror equation for an object point at a distance less than R/2 from a concave mirror of radius of curvature R. the minus sign occurs because the rays at S point away from Q (see Example 3.34 Paraxial image formation by a concave refracting surface SPM. 3. Let Q represent an arbitrary point on the axis. C represents the center of curvature.3. 3. Thus according to Fermat's principle. 3. 3. 3. 16 2 r (94) Also show that the above expression leads to the paraxial image point which is consistent with Eq. Consider an object point 0 in front of the convex mirror SPM (see Fig.35).33 Paraxial image formation by a convex mirror.4 We next consider a point object 0 in front of a concave refracting surface SPM separating two media of refracting indices n1 and n2 (see Fig. we may note tant from the center. (Hint: Start with the condition that n2 AC' = n1 1 1 2 r x yo (93) which is consistent with Eq..5 If we rotate an ellipse about its major axis we obtain what is known as an ellipsoid of revolution.4). 3.33. (see Fig.35 All rays parallel to the major axis of the.
485.12 In an inhomogeneous medium the refractive index is given by 2(x) [Hint: We consider a point C which is at a distance d1 from the point 0 and d2 from the point I. 3.48. where S is an arbitrary point on the refracting surface. S Fig. 4. Calculate the angle that the ray will make with the zaxis at the launching point. (a) Assume rays launched on the axis at z = 0 (i. 3. calculate the focal length for each ray and qualitatively plot the ray paths. a ray is launched at x = 0.495. 3.475 and 1.0.43 m making an angle 7r/60 with the zaxis (see Fig.23 = n1 z1  n2 (z2  z1) (97) where the origin is assumed to be at the point 0 and the coordinates of P and I are assumed to be (0. Assuming the right hand side to be zero.12 and Sec. In each case find the height at which the ray becomes horizontal. show that the equation of the refracting surface (separating two media of refractive indices n1 and n2) is given by =+ =1 L for x > 0 for x < 0 . a ray is launched at x = 2.d1. 3.8 m such that it becomes horizontal at x = 0. ±30 gm.. show that the refracting surface is spherical.n2 (x2 + y2 + =n1 ± di) . (see also Fig. a = 50 pm. Calculate the value of A.50.480.e. ±20 mm.15). (b) Assume rays incident normally on the plane z = 0 at x = 0.20) l2 Ixl < a =ni(12A)=nZlxl>a Fig. 3. 1.14). Find the corresponding values of /3 .z) 2] 112 3.0. x = 0 when z = 0) with = 1.490.470 In each case calculate the angle that the ray initially makes with the zaxis (01) and plot the ray paths.37. with the radius given by r Thus show that nd1 2l Fig.36 A spherical reflector. y. 4. 3. Assume n1 = 1.9 For the refractive index variation given by Eqs (21) and (22). ±40 pm. Thus A2) 112 n1 (x2 + y2 + z 2) U2 . z) represent the coordinates of the point S.8 Referring to Fig. if I represents a perfect image of the point 0. 3.11 Consider a parabolic index medium characterized the following refractive index variation: n2 (x) (x a = n? [l. z2) respectively.Fermat's Principle and Its Applications n1 [x2 + y2 + z2] 1/2 + n2 [x2 + y2 + (z2 . Assume the origin to be at 0 and let (x. 3.] 3. The surface corresponding to Eq. 1. nl OP nt + n2 (95) = n 2 d2 = n 1 n 2 r (96) where d1 and d2 are defined in Fig. [Ans: 01 = 19°] 3.37 All rays emanating from 0 and getting refracted by the spherical surface SPM appear to come from I.2m (see Fig.38.10). The above equation would give the equation of a sphere whose center is at a distance of n2r/n 1 (= d1) from 0. (97) is known as a Cartesian oval. z1) and (0. n2 = 1. ±10 pm. 1.n2 (r ± d2) ° where A = d2 .38 The Cartesian oval.10 For the refractive index variation given by Eqs (21) and (22). All rays emanating from 0 and getting refracted by SPM pass through I. 3. 1. 3. Calculate the value of x at which it will become horizontal.
Pergamon Press.K. Kaw. 18. 46. Contemporary Optics. 1920.39 4 6 8 z (mm) 10 12 P Notice that the periodic length 27r zp= g Ray paths in a graded index medium characterized by Eq. New York. 338. vol. Plenum Press. W. (100). Leighton and M. UK. `A note on mirage formation'.S. Mach.K. American Journal of Physics. (98) to determine the ray paths. New York. 3.G2 Optics is independent of the launching angle (see Fig. Mass. The Asiatic Society. 1974. Physics of the Air. Bush and R. R. Ghatak.39) and all rays rigorously take the same amount of time in propagating through a distance zp in the zdirection. 90. `Generalized Fermat's principles'. E. 7.J. vol. AddisonWesley Publishing Co.H. 10. Thyagarajan. New Delhi].B.S. Aggarwal and P. (27) and integrate to obtain x(z) = (100) 1 g In2 1 sinh_ [1? 2  sin gz (101) 2 Fig. 1976. 6. American Journal of Physics. K. 11. Second Edition. 1983. REFERENCES AND SUGGESTED READINGS 1. A. M. vol. Reading. Oxford. [Reprinted by Macmillan India. R. (27) can be written in the form + aKln2 dz = 2/1 dG .. 503. Born and E. 12. 3. 1976. Fraser and W. Sodha. 102. R. 9.. Scientific American. 297. Principles of Optics. show that Eq. (23). 42. Ghatak. Khular. 0) where its orientation with respect to x axis is 45°. WA. Khular. Thyagarajan and A. `Mirages'. 4. Humphreys. I. (1978).K. 3. Calcutta. A. `A Note explaining the mirage'. 1967. Kluwer Academic Publishers. 1965. 2002. Thyagarajan and A. 45. 5.n0 n2 2 and G(x) = K1 eaxi2 (99) 2 Integrate Eq. 1952. Lagrangian Optics. Mitra. V Lakshminarayanan. [Hint: While carrying out the integration.K. vol. Ghatak and K Thyagarajan. `Ray tracing in uniaxial and biaxial media'.T. 234. American Journal of Physics. 8. 3. K.14 Consider a graded index medium characterised by the following refractive index distribution n2 (x) = ni sech2 gx Substitute in Eq.24 Write down the equation of a ray (in the xz plane) passing through the point (0.3. McGrawHill Book Co. vol. Vol. E. A. 3. K. Sands.P. 1977. `Image formation by an optically stratified medium: Optics of mirage and looming'. Ghatak and K. 2. Optik. make the substitution: (98) c= l in? 73 2 a sinh gx] where K1 = 2 /3 . 51. January. A. Robinson.B. Wolf. K. M. The Upper Atmosphere. Newcomb. British Journal of Applied Physics. 0. The Feynman Lectures on Physics. Feynman. vol. 774.13 For the refractive index profile given by Eq. S.. . 1975.
Although there is no additional physics involved (other than the Snell's laws) in the tracing of rays. Greek philosopher and mathematician (6th century BC). 23 Alhazen had used spherical and parabolic mirrors and was aware of spherical aberration. the refracted ray and the normal (to the surface) lie in the same plane. and it is possible that Chinese and Greek knowledge were both derived from a common source in Mesopotamia.1 INTRODUCTION In this chapter we will study the formation of an image by simple optical systems.1).. another Greek philosopher (300 BC). His work_ was translated into Latin and became accessible to later European scholars. . 4. The idea that light is emitted by a source and reflected by an object and then enters the eye to produce the sensation of sight was known to Epicurus.* In order to trace a ray through such an optical system. From the Internet 4. The Greek mathematician Euclid (300 BC). * The optical system may also consist of mirrors. India or Egypt. such numerical computations are usually done on a highspeed computer. acting like feelers.. 4. Pythagoras. Vol. who accepted the Pythagorean idea. the design of even a simple optical system involves tracing many rays and therefore considerable numerical computations.3). suggested that light consists of rays that. and (b) if 01 and 02 represent the angles of incidence and refraction respectively. We will assume the optical system to be made up of a number of refracting surfaces like a combination of lenses. Nowadays. It may be of interest to note that optical designers were among the first to make use of electronic computers when they were introduced in the early fifties. the more mysterious sense of sight is explained in terms of the intuitively accepted sense of touch. References to burning mirrors go back almost to the start of history. travel in straight lines from the eye to the object and that the sensation of sight is obtained when these rays touch the object. In this way. The Pythagorean hypothesis was eventually abandoned and the concept of rays traveling from the object to the eye was finally accepted about AD 1000 under the influence of an Arabian mathematician and physicist namedAlhazen. The New Encyclopedia Britannica. in which case the reflection of rays should also be taken into account (see Sec. knew that the angle of reflected light rays from a mirror equals the angle of incident light rays from the object to the mirror. then sin 01 _ n2 sin 02 n1 (1) where nl and n2 are the refractive indices of the two media (see Fig. It is only necessary to reverse the direction of these rays to obtain the basic scheme of modern geometrical optics. He also investigated the magnification produced by lenses and atmospheric refraction.The use of plane and curved mirrors and of convex and concave lenses were discovered independently in China and in Greece. it is necessary only to apply Snell's laws at each refracting surface which are as follows: (a) the incident ray.
2 Optics We next make use of the paraxial approximation. 4. 4. then the slope angle is negative. so that we may write sin 01= tan 01 01 etc. viz.1 (a) Paraxial image formation by a spherical refracting surface separating media of refractive indices n1 and n2. 02. where the angles are obviously measured in radians.1(a) and we have assumed that the foot of the perpendicular (D) is very close to the point P so that OD = OP = x. We will consider a point object 0 emitting rays in all directions.1(a) we have = r +x (2) 02 =/3.1(a) if 01 and 02 are the angles that the rays OS and SI make with the axis. Let C represent the center of curvature of the spherical surface. 2. We refer to Fig. 4. 4. a1. 0 represents the object point and I the paraxial image point. in the absence of a refracting surface.3). we have sin 01= $1= /3 + a1 = tan /3 + tan a1 and h h (3) r y where the distances h. Thus in Fig. Thus. 4. in terms of the angles defined in Fig. 4. (b) corresponds to positive u. 3. then 61 = a1 and 02 = .. Conversely. y and r are defined in Fig. if we consider only those rays which make small angles with the line joining the points 0 and C then all rays do converge to a single point I [see Fig.2 REFRACTION AT A SINGLE SPHERICAL SURFACE We will first consider refraction at a spherical surface SPM separating two media of refractive indices n1 and n2 (see Fig.(3) to obtain (in the paraxial approximation): sin 02 4.4.n1 y x r (4) 4.1 The Sign Convention Before we proceed further we should state the sign convention which we will be using throughout in the book.1 (a).1(a) and consider the point P as the origin of the coordinate system.1(b). For u to be positive. we must have a situation like the one shown in Fig. The angle that the ray makes with the axis is positive if the axis has to be rotated in the anticlockwise direction (through the acute angle) to coincide with the ray. . x.2.a2 =tan/3tan a2= 0 1 =/3+a1 and 02 =/3a2 u<O. 4. 4.a2.1(a)]. all angles O1. We may mention that not all rays emanating from O converge to a single point. the rays converge to a point to the right of P. etc. a2 and /3 are small. The sign convention is as follows: 1. aZ and /3 represent the magnitudes of the angles. (If the final result does not depend on the angles. Thus in Fig. All distances to the right of the point P are positive and distances to the left of the point P are negative. however.1(a)). ID = IP = y. The rays are always incident from the left on the refracting (or reflecting) surface. We now use Eqs (1) . We will use Snell's laws of refraction to determine the image of the point O. a1. the object distance u is a negative quantity and the image distance v and the radius of curvature R are positive quantities.R>O nl sin 01 = n2 sin 02 n 1 l 12 + h r x or n2 =n2 ^h 1 r h yJ nl = n2 . 4. This is known as the paraxial approximation and according to Fermat 's principle all paraxial rays take the same amount of time to travel from 0 to I (see Example 3. then it is more (b) Fig.v>O. Now. if the axis has to be rotated in the clockwise direction (through the acute angle) to coincide with the ray.
v = y and R = r. obviously. u = + 150 cm.Refraction and Reflection by Spherical Surfaces convenient to use the magnitude of the angles as has indeed been done in Sec. We proceed in a manner exactly similar to that in Sec. in general. On the other hand. 4.1(b) both 01 and 02 are negative quantities.3 to obtain 0' 1< 40 I 25 30  333 01 180 Fig. u = x. the point C represents the center of curvature. corresponding to Fig. Thus 1. Obviously u=40cm.1 Consider a medium of refractive index 1. 4.0andn2 =1. It should be noted that. u is negative and v positive.5 1 0.2 The radii of curvature of the two surfaces are 15 and 25 cm with their centers at CI and C2 respectively. 4.5 = + 0_5 v 150 25 giving cm 3 and a real image is formed on the right of P2 at a distance of 33 3 cm. and all distances measured in the downward direction are negative. not all rays from an offaxis point will intersect the axis and after refraction at the second surface will.3 REFLECTION BY A SINGLE SPHERICAL SURFACE We next consider the imaging of a point object 0 by a spherical mirror SPM (see Fig. thus any ray from the point 0 (like OS) will be in the plane containing the axis and the normal at the point S and consequently. the axis of the system is defined by the line joining the object point 0 and the center of curvature C.1(b) u and v are both positive.2). (4) becomes n2 _ nl n2 . Example 4. We should point out that while considering refraction by a single surface (as in Fig. 4. 4. 4. 4. All distances measured upward from the axis (along a perpendicular to the axis) are positive. 4. 4. all rays emanating from a point on the axis are meridional rays. the image is formed at 0' at a distance of 150 cm from P2.3 In the absence of the second surface. In Fig. There is an object at a distance of 40 cm (from PI) on the line joining CI and C2. nl = 1.5 and n2 = 1. 4.n 1 =1. R = 25 cm. then for the ray diagram shown in Fig.1).3) in the paraxial approximation. The angle that a ray makes with the normal to the surface is positive if the normal has to be rotated in the anticlockwise direction (through the acute angle) to coincide with the ray. 10 of the previous chapter). these rays are known as skew rays.2 Paraxial image formation by a medium of refractive index 1.0 1.Thus 1.5 bounded by two spherical surfaces S 1 P 1M1 and S 2P2M2 .1(a). Thus.R=+15cm. whereas for Fig. and conversely.5.1 (a).1(a). Equation (5) is known as the Gaussian formula for a single spherical surface. as can be readily seen. the refracted ray will always intersect the axis. not remain confined to a single plane. 4. 5.5 15 v+ 40 v = +180 cm 4.nl (5) v u R which gives the image point due to refraction at a spherical surface (see also Eq. in Fig. All distances are measured in centimeters.2 and we refer to Fig. Rays which remain confined to a plane (containing the axis) are known as meridional rays. while considering refraction by the second surface. Determine the position of the paraxial image. 4.2. v =+33 4. 4.2 The Gaussian Formula for a Single Spherical Surface If we now use the sign convention discussed above. 0' now acts as a virtual object and since it is to the right of S2P2M2 we have. .0.5 bounded by two spherical surfaces S I P I M 1 and S2P2M2 as shown in Fig. 4. In the latter case. Solution: We first consider refraction by S I P1M 1 . 0I and 02 are positive quantities.2 or as in a lens) then the line joining the two centers of curvature is defined as the axis. 4. Thus Eq. if there is second refracting surface (like in Fig.
Fig. It is interesting to note that if we set n2 = .4 THE THIN LENS A medium bounded by two spherical refracting surfaces is referred to as a spherical lens. ^1=$at= Yx h and Y r where the distances x.1). We would like to determine the final image position of the object point 0 which is at a distance of 80 cm from the point P1. 4.x. (1) becomes the law of reflection if we have n2 = .6. 4. . Example 4. This follows from the fact that Snell's law of refraction Eq. giving 2 11 v=+40cm 20 'v 8 Thus the final image is formed on the right of S2P2M2 at a distance of 40 cm. Different types of lenses are shown in Fig. we will consider the paraxial image formation by a thin lens. If the thickness of such a lens (shown as tin Fig. we get Eq.4 Paraxial image formation by an optical system comprising a concave mirror S 1P1M 1 and a convex mirror S 2P2M2. thus we obtain the mirror equation u v R + 4. The line joining the centers of curvature of the spherical refracting surfaces is referred to as the axis of the lens.y and R = .3 Paraxial image formation by a spherical reflecting surface SPM. 4. (7).4). it may have cylindrical surfaces).n1 in Eq. In this section.2 Consider an optical system consisting of a concave mirror S1P1M1 and convex mirror S2P2M2 of radii of curvatures 60 and 20 cm respectively (see Fig.3.4. a real image will be formed at 71 which now acts as a virtual object for S2P2M2. the two mirrors being separated by a distance of 40 cm. most lenses employed in optical systems have spherical refracting surfaces. y. then is = .g. we have (considering imaging by S2P2M2). We first consider the imaging by S 1P1M1 . (7) which is the same as was derived by using Fermat 's principle (see Problem 3. (5). The corresponding considerations for a thick lens will be discussed in Problem 4. However. a lens may have nonspherical refracting surfaces (e. v = . u = . We illustrate the use of Eq.r. we have v=48 cm 60 80 + v In the absence of the mirror S2P2M2. Since 01 = 02 (the law of reflection). 4. we get 1+1=2 x y r =a2P (6) If we again use the sign convention that all the distances to the right of P are positive and those to its left negative. h and r are defined in Fig. Therefore. we will simply use the term `lens' to imply a spherical lens. Since I1 is to the left of P2. since u = . In general. which happens to be the point P1.60 cm (because both 0 and C are on the left of P 1).20 cm.5) is very small compared to object and image distances and to the radii of curvature of the refracting surfaces then the lens is referred to as a thin spherical lens.4 Optics 14 80 rj M Fig.8 cm and R = . (7) through an example. 4.6.n1. 4.80 cm and R = .
5 Fig.5.7(b)). Similarly. However. The lens is placed in a medium of refractive index nl and the refractive index of the material of the lens is n2. 4. the image of the point 0 would have been formed at Q whose position (given by v') is determined from the following equation (see Eq. the point Fl is the first principal focus. if (1/R 1) . if the double convex lens is placed in a medium whose refractive index is greater than that of the material of the lens. 4. is given by f =(n1) 2) /Rl R ) (12) nl . The line joining the two centers of curvature is known as the axis of the lens [u = x.nl Rl (8) where u is the object distance which is negative for the object point 0 shown in the figure. In Eqs (8) and (9) the distances are measured from the centre of the lens P. for the lens shown in Fig. known as the focal length of the lens. 4.(1/R2)] is a positive quantity then the focal length is positive and the lens acts as a converging lens [see Fig.7(d)].5 THE PRINCIPAL FOCI AND . 4.7(c)]. then the point Q lies to the right of the surface and if v' is negative then Qlies to the left of the surface.n2 R2 (9) . the first principal focus is defined as the point (on the axis) such that a ray passing through that point will.7(a)].5. similarly for the double concave lens [see Fig. For a diverging lens. we get v u (n 1) (R 1 (a) R1 > 0 R) z (10) (b) R1 < 0 R2 >0 (c) R1 < 0 (d) R1>0 where _ n2 nnl Equation (10) is known as the thin lens formula and is usually written in the form 1 1 v u f R2<0 R2 <0 R2>0 Fig. (5)) n2 v' u where f. 4. then the focal length becomes negative and the lens acts as a diverging lens [see Fig. this is justified because the lens has been assumed to be thin.FOCAL LENGTHS OF A LENS For a converging lens. 4.Refraction and Reflection by Spherical Surfaces 4. 4. 4. Obviously if v' is positive.5 Image formation by a thin lens. RI is positive and R2 is negative.n2 . the ray which (in the n2 v' nl . 4. if the second refracting surface had not been there.8(a). the image formed by the first surface is considered as the object (which may be real or virtual) for the second surface. The point Q now acts as the (virtual) object for the second refracting surface and the final image is formed at I whose position is determined from the equation nl v For a lens placed in air (which is usually the case). after refraction through the lens. v' = y' v = y]. Thus. emerge parallel to the axissee ray 1 in Fig. n > 1 and if [(1/Rl) .(1/R2) is a negative quantity then the lens acts as a diverging lens (see Fig. 4. Adding Eqs (8) and (9). Let RI and R2 be the radii of curvature of the left and right surfaces of the lens. In order to determine the position of the image we will consider successive refractions at the two surfaces. We consider a point object 0 on the axis of a (thin) lens as shown in Fig.6 Signs of R1 and R2 for different lens types.
(c) and (d) correspond to the situation when the refractive index of the material of the lens is smaller than that of the surrounding medium and therefore a biconvex lens acts as a diverging lens and a biconcave lens as a converging lens. 4. 4. and v are positive quantities. the values of fl and f2 can be readily obtained by considering the thin lens formula [see Eq. (10)] and one gets Fig. 4. For a thin lens placed in a medium such that the refractive indices on both sides of the lens are the same (n3 = nl in Fig. f2  A = [(n 1) I R1 .8) and we have 1 nl 1 n2 nl + n3 n2 Rl RZ (15) X1 K. 4. u and v are negative quantities and x2 and fl are positive quantities. here xi. after refraction by the lens. The point F1 is the first principal focus and its distance from the lens (denoted by f1) is known as the first focal length of the lens.fi and u are negative quantities and x2. 4.f2 is positive for a converging lens and negative for a diverging lens.+ 9] = f (13) However. (b) Paraxial imaging by a diverging lens. 4. f2. when v = oo.8(a) and (b)]. xi.8(b).4. For a converging lens the point at which the ray will intersect the axis [shown as F2 in Fig. Similarly.7 (a) and (b) correspond to the situation when the refractive index of the material of the lens is greater than that of the surroundings and therefore a biconvex lens acts as a converging lens and a biconcave lens acts as a diverging lens. f2.X2 (b) Fig.6 Optics (c) n2 < ni (d) absence of the lens) would have passed through the first principal focus emerges. As can be seen from Fig. as a ray parallel to the axissee ray 1 in Fig. u = fi (ray 1 in Fig. The distance of the second principal focus from the lens is known as the second focal length and is denoted by f2. Obviously.8.2): n3hi v u n 2 n l+ n3 n 2 R1 R2 (14) Now. 4. for a diverging lens. if n3 ^ nl then the thin lens formula assumes the following form (see Problem 4.8(b)]. . the point at which the ray would have intersected the axis (if produced backwards) is the second principal focus [see the point F2 in Fig.8).8(a)] is known as the second principal focus of the lens.8 (a) Paraxial imaging by a converging lens. We next consider a ray which travels parallel to the axis [see ray 2 in Figs 4. fl is negative for a converging lens and positive for a diverging lens. 4.
+ xl fl xl x2 f2 (21) 4.8(a) and (b)] 4. (13) we have xlxz = f 2 y' y  x2 x1 (20) Fig. and vice versa. Considering similar triangles in Fig.9 Imaging of an object of height y by a spherical refracting surface. the magnification produced by a single refracting surface is given by m = Y' Y and considering triangles AOC and ICB. v is also equal to zero [see Eqs (10) and (14)]. if m is positive. Considering either Fig. When the thin lens has the same medium on the two sides.f.8(b)]. 4. then the image will lie on the right of the second principal focus. Equations (17) and (18) give flf2 = x lx2 (19) where we have made use of Eqs (17) and (18). It may be noted that for a diverging lens [see Fig.1).nl y nt u nt R and un1 R 1 R+1_ A xi (17) (22) v n2 _ n2nl u n2 R which is known as the Newtonian lens formula.n2 1 . .8(b) we readily get y' v u Y _ f2+x2 .f2 where the vertical distances are positive if measured above the line and negative if measured below the line (see Sec.= Rl n3 R2 f2 (16) 4. referring to Fig. the image is erect [as in Fig. when u = 0. 4.00. after refraction.8(a) and (b)] (ii) A ray parallel to the axis will. (5) gives us n2 v = n2 . _ vR u+R Y Now. Eq. * This follows from the fact that.nl + n3 . 4. after refraction. 4. 4. 4.8(b) and conversely if m is negative. emerge parallel to the axis [see ray 1 in Figs 4.7 LATERAL MAGNIFICATION The lateral magnification m is the ratio of the height of the image to that of the object. Once we know fl andf2 (and therefore the positions of the first and second principal foci) the (paraxial) image can be graphically constructed from the following rules: (i) A ray passing through the first principal focus will.8(a)]. we get Y. 4. either pass through or appear to come from (depending on the sign off. 4.f2 which are identical to Eqs (17) and (18). then using Eq.) the second principal focus [see ray 2 in Figs 4. Eqs (17) and (18) would be . for a thin lens. when u = . The magnification can also be calculated as the product of the individual magnifications produced by each of the refracting surfaces. 4. we have Y' Y and y' _ x2 (18) Y .8(a) and (b)] (iii) A ray passing through the center of the lens P will pass through undeviated* [see ray 3 in Figs 4. v =f2 (ray 2 in Fig. the image is inverted as in Fig.6 THE NEWTON FORMULA Let xi be the distance of the object from the first principal focus Fl (x1 will be positive if the object point is on the right of Fl and conversely) and let x2 be the distance of the image from the second principal focus F2 as shown in Figs 4. Thus if the object lies on the left of the first principal focus.8) and we have n2 .8(a) or 4. Obviously.2.8(a) and (b).8(a).Refraction and Reflection by Spherical Surfaces Similarly.9.7 showing that xi and x2 must be of opposite signs.
4. erect and smaller in size by a factor of 2. The two lenses are separated by 8 cm.5 cm Further.5 cm on the left of the concave lens. we get (23) n2u Thus. _+ 1 40 20 40 Thus.10.. v = +40 cm and m1= 1. in general. all rays emanating from a given object point were found to intersect at one point which is the image point. For an object of height 1 cm (at a distance of 40 cm from the convex lens). (11) we get 1 + 1. Example 4. i. i. If we had considered rays which make large angles with the axis. This image acts as a virtual object for the concave lens with u = +32 cm and f = 10 cm. 4. However.14. (The same problem will be solved again in Chapter 5 by using the matrix method. using Eq. The image is virtual. there exist two points for which all rays emanating from one point intersect each other at the other point.e.8 APLANATIC POINTS OF A SPHERE In Sec. . 4. the image is of the same size but inverted.mlm2 u consistent with Eq.5 'I Fig. (21). Thus v (24) m __. 4.. (22).11) and a perfect image is not formed. (after refraction) they do not pass through the same point on the axis (see Fig. we had considered rays which made small angles with the axis. 4.40cm.f=+20 cm Therefore. if ml and m2 represent the magnifications produced by the two refracting surfaces in Fig. for a given spherical surface.2 n2 v m2 = n v 1 where v' represents the distance of the image formed by the first refracting surface.) Solution: { Let us first calculate the position and size of the image formed by the first lens: 22 The final image is formed at a distance of 14. This point is at a distance equal to n2 IRI/n 1 from the center of the spherical surface and a virtual image is formed at a distance of n1 IRI/n2 +20 10 u.3 Consider a system of two thin lenses as shown in Fig.14. The convex lens has a focal length of +20 cm and the concave lens has a focal length of 10 cm. 4. m2 =Thus m =m1m2=+ 320/22 32 _ 1 2.10 Paraxial imaging by an optical system consisting of a converging lens of focal length 20 cm and a diverging lens of focal length 10cm separated by 8 cm. All distances in the figure are in centimeters. In this approximation. 4. The image is said to be afflicted with aberrations.. calculate the position and size of the image. it was found that the images of point objects are perfect. while discussing image formation by a single refracting surface we had made use of the paraxial approximation.2. Thus v u f 1=1+1 40 I4.e. then nl v' m=n2 u and y Optics m=' =my 1 v giving 1 32 1 __ 22 10 320 v = .8.8 Substituting for v/R and u/R in Eq. then we would have observed that.2.
The object 0 is immersed in the oil and the distance OC is made equal to n 3 1R 1 I/n2 so that the point 0 is the aplanatic point with respect to the hemispherical surface.9 Now.1 The Oil Immersion Objective Fig. The hemispherical lens L1 is placed in contact with a drop of oil whose refractive index is the same as that of the lens. (a) and (b) correspond to n2 < nl and n2 > nl respectively. .7) or by using geometrical methods (see Sec.10).12(a) and (b)].8. will not pass through the point I this leads to aberrations in the image. 4.12 (b) (a) O and I represent the aplanatic points of a spherical surface. 4. we may assume 0 to be embedded in a medium of refractive index n2. Now L2 is an aplanatic lens with respect to the object point at I1 and therefore a perfect image of I1 is formed at I.e. Therefore. The lateral magnifications caused by the refracting surface R1 and lens L2 are ml = n2(I1 P1) n3 ( 0P1) (27) + n11 [1 n2 (25) and _ n 4( IP 3) n5 (11 1'9 and IP2 = IR21 (28) [1 + j?2 n1 (26) Thus the oil immersion objective reduces considerably the angular divergence of the rays and results in an increase in lateral magnification without introducing spherical aberra n2 < nl n2 > n1 Fig. 4.Refraction and Reflection by Spherical Surfaces 4. Hence all rays emanating from 0 hit the first surface normally and move on undeviated. This can be easily proved by using Fermat's principle (see Problem 3. The two points are said to be the aplanatic points of the sphere and are utilized in the construction of aplanatic lenses (see Fig.14. all rays emanating from O appear to come from I. in general.. from the center [see Figs 4. However.11 The point I represents the paraxial image point of the object point 0 formed by a spherical refracting surface SPM. 4.13). for all practical purposes. the radius of curvature of the first surface (= R1) is such that the point 0 coincides with its center of curvature. Thus OP2 = IR 2 1 The principle of aplanatism has a very important application in microscope objectives where one is interested in having as wide a pencil of light as possible without causing any aberrations. 4. if we consider nonparaxial rays like OS1 (which make large angles with the axis) then the refracted ray. which is why a perfect (virtual) image is formed at I1.13) which are used in wide aperture oil immersion microscope objectives. 4. We refer to the optical system shown in Fig. The points 0 and I are the aplanatic points of the spherical surface of radius R2 (see Fig. 4. i. A perfect (virtual) image of 0 is formed at I.
oil immersion objectives have a certain degree of chromatic aberration. tion. 4. for two points to form perfect images of Fig. The refracting surface is known as a Cartesian oval. We consider a spherical 4. the refracting surface should not be spherical. Figure 4. Fig. mention that a perfect image is formed only of one point and therefore nearby points have some aberrations. The points 0 and I1 are the aplanatic points corresponding to the hemispherical surface of radius R1.In this section we will show the existence of aplanatic points soid of revolution (see Problem 3. the image is other at the other point I.5) and under certain using geometrical considerations.9 THE CARTESIAN OVAL M In general. Moreover. The points 0 and I are the aplanatic points of the spherical surface S 2P2M2thus a perfect (virtual) image is formed at I.14 The oil immersion objective.10 Optics Fig. We should. the surface becomes an ellip.15 The refracting surface (known as the Cartesian oval) is such that all rays emanating from the each other. 4. however. point 0 intersect at I.12(a) and (b)]. the lens L2 acts as an aplanatic lens for the (virtual) object at I1. 4.15 is the locus of the point S such that n1 OS + n2SI = constant (29) The refracting surface is obtained by revolving the curve shown in the figure about the zaxis (see also Problem 2. however.10 GEOMETRICAL PROOF FOR THE EXISTENCE OF APLANATIC POINTS . The object point 0 is at the center of curvature of the first surface S 1P1M1.8).4.13 The aplanatic lens.15 shows the two points 0 and I such that all rays emanating from 0 (and allowed by the system) intersect each circumstances the surface is spherical. z 4. 4. When the object point is at infinity. Aplanatic lens L2 Fig. Thus the curve SPM shown in then virtual [see Figs 4.
0 and I are the aplanatic points. With C as center. 4. i.u and using Eq. r CS OCrlµ (33) Thus the two triangles SOC and SIC are similar. We next consider a slightly offaxis point 0' (directly above 0) and. we get sin a sin /3 _ 4. 4.16). The inner and outer spheres are of radii r/µ and .17. This implies that the optical system has no spherical aberration corresponding to O.Refraction and Reflection by Spherical Surfaces refracting surface SPM of radius r separating two media of refractive indices nl and n2 (see Fig. Let IOCP represent any common diameter of the three spheres intersecting the outer and inner spheres at I and 0 respectively. then all rays emanating from the point 0 will appear to come from I. Now. (34).11 THE SINE CONDITION We consider a general optical system as shown in Fig. considering the triangle SOC we have sin a _ r/u _ 1 r sin O1 .ur respectively. . and therefore a= O2 and . the linear magnification is given by m `2 for all values of 01. 4.e. The system is said to be free not only from spherical aberration but also from coma. 4. we draw an arbitrary line hitting the refracting surface at the point S. The point C represents the centre of the spherical surface SPM. If we can show that sin a sin /3 1 1 n2 µ sin /3 sin a nl (36) proving that 0 and I are aplanatic points.11 (35) where t > 1. We join I and S and extend the line further as SQ. CS and IC = µr =µ r IC CS (32) 4.(34) Fig. Furthermore. We assume that the point 0 (on the axis of the system) is perfectly imaged at I.16 Geometrical construction for the derivation of aplanatic points. and 0 and I will be the aplanatic points for the spherical refracting surface SPM. SPM is the refracting the surface of radius r. all rays emanating from 0 intersect each other at I. we draw two spheres of radii µr and r/u as shown in Fig. We also have sin Ol sin O2 nl n2 (37) µ (31) It is obvious that the points 0' and I' will also be aplanatic and therefore the image formed by a small planar object at 0 will be sharp even for the offaxis points. We will assume n2 < nl and define nl (30) Now. From the point 0. /3 = L ISC = 01 .16.
indeed the condition for sharp imaging of 0 =and 01 is quite different (see Problem 4. It is of interest to note that according to Eq. 4.11). Further OPL[O'B 1 B2I' ] = OPL[O'D 1D2I'] where the point G is the foot of the perpendicular drawn from the point 0 on ray 4. 3. (39). (45) from Eq. Also. we get OPL[O'D ID2 1'] = OPL[OC1 C2 1] (45) Now.17 The optical system images perfectly the points 0 and 0' at I and I respectively.17. See Ref. the rays 3 and 4 meet at infinity and intersect at F'. The ray O'B 1 is parallel to the ray 0A1 and the ray O 'D1 is parallel to 0G 1 . 4. Now. when the condition given by Eq. the rays O'B I and 0AI meet at infinity and therefore OPL[O'B 1B2 F] = OPL[OA IA2F] (42) (39) We next consider the triangle FII' Fl' = [FI2 + 1y2121112 where 61 and 02 are defined in Fig. sharp imaging of a nearby point on the axis (like 01) is not obtained. .11. since I is the image of the point 0.4. (39) perfect imaging of (nearby) offaxis points requires a condition to be satisfied by rays from an onaxis point. (39) is satisfied.1 Proof of the Sine Condition** We refer to Fig. ** For a rigorous proof of the sine condition. We will assume that the axial point 0 is perfectly imaged at I and will use Fermat's principle to determine the condition for perfect imaging of the nearby offaxis point 0'. Thus the linear magnification will be constant if the ratio sin 61 /sin 02 is constant for all points on the refracting surface and the image will be free from the aberration known as coma. we get OPL[O 'B IB21'] = OPL[OA1A21 ] OPL[OC I C2I] (44) 4. 4. we get m= Y2 = Yi 1! n2 n1 sin 91 n2 sin e2 consistent with Eq.12 Optics Fig.OPL[GO '] = OPL[F ' l] (41) (47) * It may be noted that if we use Eqs (37) and (38). If we add Eqs (42) and (43). We subtract Eq. we have OPL[OA 1A21] = OPL[OC1 C21] Since the left hand side of the above equation is OPL [O 'DID21']. for 0' to be sharply imaged at I we must have* sin 01 Y2 = linear magnification yi n2 sin 02 = n1 Now. = Fl 1 + 1 1'212 2 FI2 Thus Fl' = Fl (43) where we are assuming that 1y 2 1 is small enough so that terms proportional to 1y 2 1 2 can be neglected.17. according to the sinecondition. (46) to obtain OPL[F 'I'] . so that OPL[GD1D2F'] = OPL [OCI CF ' ] (46) (40) where OPL stands for the optical path length.
Assume n1= n3 = 1 and n2 = 1.F 'I) But GO' = yl sin 01 and F'I' . assume nl = n3 = 1.nl R v u The sign convention is as follows : 1. A concave lens of focal length 20 cm is placed beyond the convex lens at a distance of 25 cm.Refraction and Reflection by Spherical Surfaces or n2(F 'I ') . [Ans: x1 = 250 cm. derive Eq.4 (a) In Fig. let R1 and R2 be the radii of curvature of the left and right surfaces of the lens. Problems 4. 4. 4. the refracting surface is a Cartesian oval. 4.13 ♦ For two points to form perfect images of each other. ♦ For a thin lens of refractive index n (placed in air).5. Using Eq. All distances to the right of the refracting surface are positive and distances to the left of the refracting surface are negative. Also calculate x1 and x2 and verify Newton's formula [Eq. The rays are always incident from the left on the refracting surface.5. the quantity f is known asthe focal lengthof thelens. The two points are said to be the aplanatic points of the sphere and are utilized in the construction of aplanatic lenses.18.60 cm respectively.3 Referring again to Fig. (51) yl showing that the linear magnification is constant if the ratio sin 61/sin 02 is constant for all points on the refracting surface.1 (a) Consider a thin biconvex lens (as shown in Fig. Summary ♦ Consider refraction at a spherical surface separating two media of refractive indices nl and n2.F'I = HI' = (Y2) sin (02) (50) (49) (48) 4. The lens is placed in air (i. (5) and considering successive refractions at the two surfaces. [Ans: Real image at a distance of 60 cm from the concave lens. 2. (20)]. nl = n3 = 1). (14).5 Consider an object of height 1 cm placed at a distance of 24 cm from a convex lens of focal length 15 cm (see Fig. The sine condition is of extensive use in the design of optical systems. then the image distance is given by nl n2 n3 Rt lRi R2 which is usually referred to as the 'thinlens formula'. Draw the ray diagram. This point is at a distance equal to n21 RI/n 1 from the center of the spherical surface and a virtual image is formed at a distance of n 1 IR1/n 2 from the center.nl = n2 .5 and n2 = 1. For an object at a distance of 100 cm from the lens. we get nl sin N1 = n2 sin 02 y2 = linear magnification .6. For u = 50 cm determine the position of the (paraxial) image. vu =(n1) V Fig. 4. (b) In (a).19). the paraxial image is formed at a distance v where n2 . 4.0 but n3 = 1. For a point object at a distance Iui on the left. 4.18) made of a material whose refractive index is 1. there are two points for which all rays emanating from one point intersect each other at the other point. Substituting the above two equations in Eq. let nl and n3 be the refractive indices of the media on the left and on the right of the lens respectively. determine the position and linear magnification of the (paraxial) image.] . Determine the position of the image and draw an approximate ray diagram for u = 100 cm.18 R2 4.n 1 (GO ') = n2(F 'I) or n l (GO ') = n2 (F 'I' . Also determine the first and second principal foci and verify Newton's formula.e. What is the qualitative difference between the systems in (a) and (b)? where H is the foot of the perpendicular from the point I on ray 4. ♦ For a given spherical surface.3. x2 = 576 cm] 4. Draw the ray diagram and determine the position and size of the final image. (48). 4. assume the convex lens to be replaced by a (thin) biconcave lens with 112 1 1 = 100 cm and IR 21 = 60 cm. IR 21= 60 cm with nl = 1. Repeat the calculations and draw the ray diagram. The radii of curvature of the first and second surfaces (R1 and R2) are +100 and .2 Consider a thin lens (made of a material of refractive index n2) having different media on the two sides.18 assume a biconvex lens with IRII = 100 cm. [Ans: x1 = 25 cm and x2 = +225 cm] (b) Repeat the calculations of the above problem when the object is at a distance of 50 cm.
5.5. Plenum Press. Contemporary Optics. Fig. Draw the ray diagram for the axial point of the object. New York. [Reprinted by Macmillan India. Wolf. Thyagarajan. it is made of.14 +15 20 Optics I<. R.7 cm from the. R. New Delhi. Show that the paraxial focal point is at a distance of 6. 1965. [Ans: At a distance of about 4. Show that parallel rays will focus at a point 40 cm from P2 (see Fig. For an object of height 1 cm at distance of 90 cm from the first surface. Pergamon Press. 4.7 In Problem 4.] 4. 4. [Ans: Real image at a distance of about 6.10 Consider a lens of thickness 1 cm. Ghatak and K. 4. determine the position and size of the image. Determine the point at which parallel rays will focus. Mass. Vol. AddisonWesley Publishing Co. All distances are measured in centimeters..K.21). 24.B. 1975. [Ans: Real image at a distance of 60 cm from the second surface.] REFERENCES AND SUGGESTED READINGS 1. Reading. Penfield. Vol.6 (see Fig.] 4. American Journal of Physics. Principles of Optics.19 An optical system consisting of a thin convex and a thin concave lens.. Leighton and M.55 cm from the second surface. 1956.24 ^I ( 25 Fig. 3. M. . The thickness of the lens is 5 cm and the refractive index of the material.H. `Consequences of parameter invariance in geometrical optics'.4. is 1.6 Consider a thick biconvex lens whose magnitude of the radii of curvature of the first and second surfaces are 45 and 30 cm respectively. Born and E.2 cm from the first surface. Feynman. I. point P2. A. 4. The Feynman Lectures on Physics.6 assume that the second surface is silvered so that it acts like a concave mirror. 1978. 2.9 Consider a hemisphere of radius 20 cm and refractive index 1.5. made of a material of refractive index 1. The radii of curvature of the first and second surfaces are +4 cm and 4 cm respectively. For an object of height 1 cm at a distance of 90 cm from the first surface determine the position end size of the image and draw the ray diagram.21 4. 4. (Remember the sign convention)] 4.P. 19. placed in air. 4. Oxford. Sands.20). R.8 Consider a sphere of radius 20 cm of refractive index 1.
in this chapter. Let . Before we describe the matrix formulation of geometricoptics it is necessary to mention the rule of matrix mi ltiplication and the use of matrices for solving linear equations.Chapter THEMATRIX METHOD In dealing with a system of lenses we simply chase the ray through the succession of lenses. develop the matrix method which can be applied with ease under such situations. In order to obtain the position of the final image due to such a system.1 Refraction of a ray by a surface SQS' which separates two media of refractive indices nl and 11 2 . . If the refracting surface is spherical then the normal NQN' will pass through the centre of curvature C. * The author thanks Professor K. g^ B= h ni n2 \. thus the matrix A= (a b c d e f (2) has 2 rows and 3 columns and has 2 x 3 = 6 elements. NQN' denotes the normal at the point Q. are made up of a large number of refracting surfaces (like in a combination of lenses) and any ray can be traced through the system by using the above conditions. then sin 61 _ sin 62 n2 n1 (1) Optical systems. and (b) if 61 and 62 represent the angles of incidence and refraction respectively. /h ^lJ r(ag + bh + ci) \(dg + eh + fi)) (4) (da e f) .l) represent_a (2 x 1) matrix.1). That is all there is to it.11 tems. one has to calculate stepbystep the position of the image due to each surface and this image will act as an object for the next surface. We shall.=S= Fig.1 INTRODUCTION* Let us consider a ray PQ incident on a refracting surface SQS' separating two media of refractive indices nl and n2 (see Fig. Then the product AB = b c (3) . A (m x n) matrix has m rows and n columns and has (m x n) elements. 5. The direction of the refracted ray is completely determined from the following conditions: (a) the incident ray. Let NQN' denote the normal to the surface. This method indeed lends itself to direct . A (m x n) matrix can be multiplied only to a (n x p) matrix to obtain am x p matrix. will be a (2 x 1) matrix. Richard Feynman in Feynman Lectures on Physics 5. Such a stepbystep analysis becomes complicated as the number of elements of an optical system increases. in general. use in computers for tracing rays through complicated optical sys. and the product BA has no meaning. Thyagarajan for his help in writing this chapter. the refracted ray and the normal lie in the same plane. 5.
when it strikes an interface of two media. The rays undergo translation when they propagate through a homogeneous medium as in the region PQ (see Fig. in Fig. it undergoes only two operations: (a) translation and (b) refraction. 5.e''=eand f' = f. we will specify the quantity. all the elements must be equal. (9) d)\g h )\Z2) X=BZ. However. or where X and Z represent (2 x 1) matrices: x2Z X= . = n cos ty (= n sin a) (10) (1. (5) 5. if we have and then (8) (Y2) h) \z2 Yi = ezi. Z2 and B represents a (2 x 2) square matrix f\ h. which represents the product of the refractive index with the sine of the angle that the ray makes with the zaxis this quantity is known as the optical direction cosine. 5. for example.e.2.+ fz2 Y2 = gzi + hz2 (7) Consequently. it undergoes refraction. 5. b' = b. The quantities (x i . 5. However. nonparaxial rays lead to what are known as aberrations which will be discussed in Chapter 6. The axis of symmetry is chosen as the zaxis. We will be considering only paraxial rays in this chapter.5. of Further. In the paraxial approximation we may confine ourselves to rays which pass through the axis of the system. The set of two equations xi = ayl + bye x2 = cyt + dye can be written in the following form: (x2) _ which can be verified by direct substitution. 2 ((Y1 +dy 2))  C d) C y2 (6) the last step follows from the rule c matrix multiplication. c' = c.2).2. Such a ray can be specified by its distance from the axis of the system and the angle made by the ray with the axis. We will now study the effect of translation and of refraction on the coordinates of the ray. . Now.2 In a homogeneous medium the ray travels in a straight line. al) represent the coordinates of the ray. (a) Effect of Translation (12) Consider a ray travelling in a homogeneous medium of refractive index nl which is initially at a distance xi from the P Fig. d'=d. i. the point P on the ray is at a distance xi from the axis and makes an angle al with the axis. instead of specifying thengle made by the ray with the zaxis. when a ray propagates through an optical system.2 THE MATRIX METHOD We will consider a cylindrically symmetric optical system similar to the one shown in Fig.2 If we define a (2 x 3) matrix A' Optics Equations (9) and (12) tell us that and xi = (ae + bg) zl + (af + bh) z2 x2 = (ce + dg) zl + (df + dh) z2 (13) (a' b' c' d' e' f' A' = A then if and only if a' = a. We will now use the matrix method to trace paraxial rays through a cylindri cally symmetric optical system. these rays remain confined to a single plane.
(15) reduces to x2 = xi + a1D (16) If (17) =n l ai and (18) 12 = n 2a2 then. (20). Let (x 2 .3 zaxis and makes an angle al with the axis (see point P in Fig.2). before and after refraction. al is very small and hence we can make use of the approximation tan al = al.. Also. then (15) Since we are interested only in paraxial rays.3). refraction through a spherical surface can be characterized by a 2 x 2 matrix: 1 0 P = n2 sin 02 (23) 1 (31) . since 01 is small. we may write x (26) Now.e. Thus. 5. Eq. one can make use of the approximation sin 9 = 9. 5. (23) reduces to n101 = n2 02 (24) (19) From Fig.3 it follows that 91 = + al and 92 01 + a2 (25) where al. a2 and 01 are respectively the angles that the incident ray. Notice that det T = (b) Effect of Refraction We will now determine the matrix which would represent the effect of refraction through a spherical surface of radius of curvature R. is the same (i.The Matrix Method in Paraxial Optics 5. Since the medium is homogeneous. if a ray is initially specified by a (2 x 1) matrix with elements Al and xi. Consider the ray AP intersecting a spherical surface (separating two media of refractive indices nl and n2 respectively)' at the point P and getting refracted along PB (see Fig. Also. Thus Eq. for the refracted ray. where al is measured in radians. the ray travels in a straight line and. The matrix T is known as the translation matrix. we get and A'2 =a 1 x2 = xl + D a. x2 = xi) we obtain. the refracted ray and the normal to the surface make with the zaxis.nl R is known as the power of the refracting surface. if PP' and MM' are perpendiculars on the axis and if P'M' = D. 5. Since we are dealing with paraxial rays.3 The refraction of a ray at a spherical surface.. nl x2 = xi + D tan al Fig. since the height of the ray at P. If 01 and 02 are the angles made by the incident and the refracted ray with the normal to the surface at P (i.e. is completely given by the 2 x 2 matrix 01 T= I 1 D/n1 1 (21) x (27) and the final ray is given by Eq. (26). using Eqs (15) and (17). therefore. then the effect of translation through a distance D in a homogeneous medium of refractive index ni. (42) _ (1 P) (A l j x2 01 xi (30) Thus. 5. then according to Snell's law nl sin 01 where we have used Eq. Thus =Px (28) (29) =1 (22) where P n2 .2). from Eqs (24) and (25). 5. we get or n1 (0 1 + al) = n 2(01 + a2) n2a2 n l al n2 R nl which may be combined into the following matrix equation: (20) Thus. a2) represent the coordinates of the ray at the point M (see Fig. with the line joining P to the centre of curvature C)'. a2 = al (14) Further.
the system matrix is. the coefficient of 21 should vanish and therefore u.P)( 1 1 1 (x2) .4 It may be noted here that det = 1 P 0 1 =1 (32) Optics In general. x2) represent the coordinates of the ray at 0. If a ray is specified by (:) when it enters an optical 2 X2 Fig.2.4 Imaging by a spherical refracting surface separating two media of refractive indices n1 and n2. x") v.lu1 i 2 + 1 L P (38) 1X1 5. x1). In general. in general.4).(_db where the matrix S = r^ C) (xi ) a) c is called the system matrix and is determined solely by the optical system. the units will not be given. x ). (A.2. the same problem was discussed in the previous chapter using the stanand dard geometrical method.1 .ad= 1 (35) (36) v/n2 1 0 (x2 ) Simple manipulations give Pu 1+nil (1 u/nl P nl (A2) x2 \n2 (xl ) (1 n2 )i (37) + nu) from which we obtain x2 . A' (just before refraction). Thus system. by a Spherical Refracting Surface Imaging For a ray emanating from an axial object point (i. The quantities a and P have the dimension of inverse length and the quantity d has the dimension of length. a product of refraction and translation matrices. can be characterized by the refraction and translation matrices. on the image plane . nl or n2 As a simple illustration of the use of the matrix method we consider imaging by a spherical surface separating two media of refractive indices n1 and n2 (see Fig. 1+ Pu n2 n1 n1  v u P n2 . (A2. however. write (33) 1 (34) (0 P) (x) ( 1 0) (A"' v/n2 or (1 0) (1 .1 (1 We should mention here that the quantities b and c are dimensionless. + nu) .".1). in general. . Thus in the above equation. we obtain det S = 1 be . The negative signs in some of the elements of S have been chosen for convenience. We will be using the analytical geometry sign convention so that the coordinates on the left of the point P are negative and coordinates on the right of P are positive (see Sec. 4. Hence. Let (A 1 . then one can. ' ' (a. 5.5. it will be implied that a and P are in cm1 and d is in cm. A" (just after refraction) and at I respectively.n1 R (39) which is the same as derived in the previous chapter. for x1 = 0) the image plane is determined by the condition x2 = 0.e. an optical system made up of a series of lenses. using the property that the determinant of the product of matrices is the product of the determinant of the matrices. Since the only two operations a ray undergoes in traversing through an optical system are refraction and translation. Also. 5. and is specified by when it leaves the system.
Let us consider a ray O'P starting from the point 0' which lies in the object plane. Let (A1 ..5 last surface.5 (the point I is the paraxial image of the point 0 and the image plane is defined to be the plane which contains the point I and is normal to the axis). if D2 is found to be positive the image is real and is formed on the right of the refracting surface. (23) of the previous chapter. if D2 is found to be negative. according to our sign convention. Thus.e. the point I' is assumed to lie on the image planesee Fig. . (A.2 Imaging by a Coaxial Optical System We will next derive the position of the image plane for an object plane.1 + (c .aD2 ) xi For a ray emanating from the axial object point (i. x ').aD2 (x1 ) (42) Thus x2 = (bD 2 + aD 1D2 . 5.The Matrix Method in Paraxial Optics i 5. on the other hand.5). P. (39) gives Thus m= my . x 1). which is at distanceD 1 from the first refracting a (see Fig.d). for points on the left of a refracting surface. Let the image be surface of the optical system formed at a distance D2 from the last refracting surface.cD1 .2. Further. the second matrix correspond to the system matrix of the optical system. 5. 1 \ 1+ Pu n1 0 P VP (1 1) n2 j (40) giving x2= 1.d = 0 (43) Fig.cD1 . x") and (A2. (A".d c . 5. Q and I' respectively. we obtain x2 1j d c)D1 A2 _' 1 0\ b a l l 0 2 1) 1 b+aD1 x2 . Then 1 0\ / A.(bD2 + aD1D2 . 5.P n2 ( vPl x1 Thus the magnification is given by m = x2 = 1 _ vP x1 n2 which on using Eq. the image will be virtual and will be formed on the left of the last refracting surface. thus D1 is an intrinsically negative quantity. Let QI' be the ray emerging from the (41) (x1 J \D2 where the first and the third matrices on the RHS correspond to translations by distances D2 and (D1) respectively (in a medium of refractive index unity). for the image plane we must have bD2 + aD 1D2 . for x1 = 0) the image plane is determined by the condition x2 = 0.'.5 The object point 0 is at a distance (Dl) from the first refracting surface. Carrying out the matrix multiplications. The paraxial image is assumed to be formed at a distance D2 from the last refracting surface. x2) represent the coordinates of the ray at 0'.cD1 . the distances will be negative and for points on the right of the refracting surface the distances will be positive. Now.D1 1 xi1 x'1 (A' b a) x""/ d c l x' J 1 (A'2 l)" n2u consistent with Eq.
(1_ 2 .6 A paraxial ray passing through a thick lens of thickness t. 5. then for a general optical system we may write ( Al x2) _ 1 PI P2 S (0 1 (52) = C l/0 M) ^ x i) (47) Thus for a thin lens.1 i Obtain the system matrix for a thick lens and derive the thin lens and thick lens formulae. if x1 and x2 correspond to object and image planes. we have (b+0 (x2) caD2 ) Optics v Dl xi (44) For x2 ^ 0.5. (53) and d=0 Substituting the above values of a. Thus the system matrix for a thin lens is given by (57) 0/ P1 = R1 and n1 n1 1n P2 = R2 = .6).R2 (50) * Notice that since we are dealing with paraxial rays.(0 1 ) (t/n 1 . Equation (55) is the wellknown thin lens formula.1) ( 1 . we obtain a=P1 +P2. x1 and x2 are the distances of the points P and Q from the axis (see Fig.I li n n i For a thin lens.P1. c and din Eq. The ray is assumed to strike the first surface of the lens at P and emerge from the point Q. M = x2 l xl would be given by M X2 2 =caD2 xl Fig. the distance between P and Q is approximately t. (The signs of R1 and R2 for different kinds of lenses are shown in Fig.aD2 represent the powers of the two refracting surfaces.7). (45) Further. Thus our system matrix is given by 01) (1 1 b S= C d c) . in propagating from P to Q. Thus D2 1 f (55) P+P2 =1)Irf [(n R1 R2J] (56) (x2) where CO 1 ) tin \0 .b=1. the magnification of the system.0 and the system matrix takes the following form: l\ 0 . corresponding to the image plane. Thus. t .6). (43). Let RI and R2 be the radii of curvatures of the two surfaces. b. 5.P2( 1 .aD2 M (46) Hence.D1 = 0. Let us consider a lens of thickness t and made of a material of relative refractive index n (see Fig. Example 5. . undergoes two refractions [one at the first surface (whose radius of curvature is R1) and the other at_the second surface (whose radius of curature is R2)] and a translation through a distance* t in a medium of refractive index n.! Pl /Jl) _ (51) t 1. since b+aD1 a c . let the coordinates of the ray at P andQ be Solution: or D2  A (P1 + P2) (54) = (n . The ray. we obtain b+aD1= =1 1 c .1 ) \ x1 / (49) represents the focal length of the lens. 5. 5. we obtain x2 = (c .1) R2 Rl 1 1 D1 xl J and r I X2 (48) or where where A1 and A2 are the optical direction cosines of the ray at P and Q.c=1 D2 + (P1 + P2) D 1D2 .6 which would give us the relationship between the distances and D2.aD2 )x 1 Consequently.
we get the required relation between D1 and D2. c1 a (63) 5. 5.t PI ). i. then U2 is the corresponding image plane. A ray emanating at any height from the first unit plane will cross the second unit plane at the same height. if dui and di2 represent the distances of the unit planes from the refracting surfaces (see Fig. 5. one each in the object and the image space.P2t I n} n =1d n n j (58) 1b (60) a c1 (61) do.7 b + adui = Or dui 1 =1 c . c and din Eq. For a thick lens.e. Thus if u is the distance of the object plane from the first unit plane and v is the distance of the corresponding image plane from the second unit plane (see Fig. It will be convenient to measure distances from the unit planes. * Obviously. .The Matrix Method in Paraxial Optics 5. any paraxial ray emanating from the unit plane in the object space will emerge at the same height from the unit plane in the image space. we get c(1b) d+cu+ c1 a v + a b + au + (1 .ad uz (59) (a) R1 > O (b) Ri < O R2> O (c) R1<O R2<O (d) Ri>0 R2>O R2<O Fig.8 Ul and U2 are the two unit planes. (51) a=Pi +P2 (1 . b=1.8)* we obtain from Eq. from Eq. Thus. if we consider Ui as an object plane.8). we would obtain 1b (62) Di =u+dui=u+ and D2 = v + du2 = v+ aNow. _ a Hence the unit planes are determined completely by the elements of the system matrix S. however. 5.b) DZ _ If we substitute the above values for a. we have from Eq.. for thick lenses it is more convenient to define the unit and the nodal planes which we shall do in the following sections.3 UNIT PLANES The unit planes are two planes. (43) we have d + cDi (64) b + aDi Substituting for Di and D2 from Eqs (62) and (63). (46): u or v= adbe+c(au+l)(c1) (1+au) a(l+au) a (1 + au) (65) <dui =1 b a a D2 U U2 Object plane First unit plane Second unit Image _ plane plane Fig. (43). 5. b.7 Signs of RI and R2 for different lens types. between which the magnification M is unity.
this condition requires the equality of a'1 and . since we are considering an axial object point.4 NODAL PLANES Nodal points are two points on the axis which have a relative angular magnification of unity. can also lie inside the optical system as shown in (b).8 where we have used the condition that det S = be . we get from Eq. Thus 1 =(n1) f Comparing with Eq. The positions of the unit planes are shown in Fig. 5. Thus Pl t dui = n 1 ^. To determine the position of the nodal points. (58). N1 . The planes which pass through these points and are normal to the axis are known as nodal planes. (44) X12 On simplification. Since we have assumed the media on either side of the system to have the same refractive index.t n1 n R ] (70) = (b + ad. From the definition of nodal points.9. 5.9 Unit planes of a thick biconvex lens. [using Eqs (58). In order to calculate the focal length we note from Eq.10).10). This has arisen because of the equality of the indices of refraction on either side of the optical system.10 Fig. (60) and (61)]: 1 = P2t n [Pi +P2 lLPi)i n and dug = P1 [P1+P2I1 1' P1 n (69) (68) /J For a thick double convex lens with 1R 1 1 = IR21 n1 P2 = R where R = IR1 1 = IR21.tnt PI) f where we have used Eq. a ray striking the first point at an angle a emerges from the second point at the same angle (see Fig. we obtain 1 1 =a (67) v u Thus 1/a represents the focal length of the system if the distances are measured from the two unit planes.a)A 1 = a 1 (75) (76) Thus b + adn1 = 1 or b (71) d 1 = 1 a (77) and t 1 t (72) d„2=n [2t n =2n nR ] where we have assumed t << R which is indeed the case for most thick lenses. 5. we require that a ray incident at an angle al on the point N1 emerge from the optical system at the same angle al from the other point N2. For example.e. Similarly we can get dn2 = c1 a (78) / 1 1l + (n1) 2 t R1 R2 J nR1R2 (74) (b) Fig. for a thick lens one obtains. 5. we consider two axial points N1 and N2 at distances dn1 and dug from the two refracting surfaces respectively (see Fig. and N2 denote the two nodal points of an optical system. (60) we find that dn1 = dui.12 . Also. i. 5.ad = 1 (66) Optics 5. (67) that l (73) 1 =a=Pi +P2(1. x1 = 0. t 2n 2. The nodal points.5.
2 cm d = .11).5x0.25.5 Transmission through glass ( 1 40/ 1. (51) 1 v) (Al (x2) .7 cm b = 0.9167 Thus a=f = To determine the focal distance v. Consider a thick equiconvex lens (made of a material of refractive index 1. This gives us 0. Solution: .125) \ 1 1 _ 0. 1 + (0. 5. c and d which are also called the Gaussian constants of the system).9.6 1)/20) 1 A SYSTEM OF TWO . the nodal planes coincide with the unit planes.25 v)^.0375 25 + 0.6667 b = 0.7cm The system matrix elements are 0. 1b = 20 cm a c1 Example 5.THIN LENSES We finally use the matrix formulation for the analysis of a combination of two thin lenses of focal lengths f1 and f2 separated by a distance t.6)/20) C^ 5. (1 0) (0. Obtain the system matrix and determine the focal length and the positions of unit planes.25 .3 a = 20 cm Thus both the unit planes pass through the centre of the sphere.25 0. we get the positions of the unit planes 1_b d = 0. In general.0375 v Thus at the image plane. one can obtain all the properties of the system. 5.6 O) C^ Refraction at the first surface (1.0.25 or v=6. when the media on either side of an optical system have the same refractive index (which is indeed the case for most optical systems).240 0.Thematrices from_the first refracting surface to the image plane are given by Second surface to image (1 0) Refraction at second surface 1 (11.0375 cm c = 0.25 01 15x1 1 1.9167 0.9 Thus.5 v 0.0.6 (see Fig. if we know a.9167 = c. consider a ray incident parallel to the axis for which 2'1 = 0. The focal plane would be that plane for which x2 is also zero.1250. The magnitudes of the radii of curvature of the two surfaces is 4 cm. Solution: R 1 =+4 cm R2 F Example 5.6667 0.The Matrix Method in Paraxial Optics 5.5 This gives us x2 = (25 + 0.35 cm a = f= 1 0. = c1 =0. = f = 26. The thickness of the lens is 1 cm and the lens is placed in air. Find the positions of the paraxial focal point and the unit planes. Consider a sphere of radius 20 cm of refractive index 1.5) of the type shown in Fig. the ray coordinates are =4 cm t=1cm Both surfaces have equal power P1 =P2= nR l = 05 = 0.125(1. Using Eqs (60) and (61). b.25 0.e.0375 _ 0.(25 00.25 0 0. 5. The system matrix for the combination .24 f= 4.9.03751 v1 25 0. d = 25 cm 1 The unit planes are given by du1= and dr2 = Thus the unit planes are as shown in Fig.125 cm 1 Thus the system matrix is from Eq.0.25 ( 0.6.0375 v = 0. if we know the elements of the system matrix S (i.11 Imaging by a sphere of radius 20 cm and refractive index 1.2 40 cm^ Fig.125 1.25 v 0.25 . 5.0375 v) x1 (0.35cm do. The nodal planes coincide with the unit planes because the lens is immersed in air.25.
5 cm 2.10.12 and Problem 4.2 v = 0 v = 14.t . 3 d=25 1 . Determine the system matrix elements and the positions of the unit planes. For an object (of height 1 cm) placed at a distance of 27. 20 t=25 cm f Ill t f2 (1 o) [1 _ f. 10 f c=?.5 \/ ' k 14101 I and the matrix for translation through a distance t (in air is) is (80) 25  1440/3N Fig.5 cm u = 27.5.5 cm. i f2 =20 cm b= 45.20 cm) separated by 25 cm (see Fig.(.fi + .12 Solution: Thus the system matrix S is given by: f1 =+15 cm 1.12.f2 c=1d=t Since f = + 10 cm. 4. (80)]. 5. we readily get a= 0 1 = 1.f2 fi 1.5) = 15 cm Thus the distance of the object from the first unit plane is given by Thus t a=t+t.fi. Solution: Let v be the distance of the image plane from the concave lens. the element a in the system matrix represents the inverse of the focal length of the system. b=1. The magnification is given by M= v f As already noted. t ) Thus.b tf a c1 f2 tf (84) a fi It is easy to see that if we have a system of four thin lenses. (82)..fi.5). Thus the matrix. Thus the image is at a distance of 30 .2 . we simply have to multiply seven matrices [four of them being of the type given by Eq. 14153 14 27. (1_ t (81) and dui = ca0 lab = .O1v) The image plane would correspond to or 32 + 2. which when operated on the object column matrix gives the image column matrix. (v 01) (+32 2 0.5 . (67)] v =30cm (82) which represents the distance of the image plane from the second unit plane.5 cm from the convex lens.5 Consider a system of two thin lenses as shown in Fig. Concave Concave lens lens to image 0) r1 +'1 0 Convex lens to cocave lens (8 1) Convex lens 1i20J Object to convex lens 1 0) ^40 1 (v Example 5.6 1 0. (57)] 1 0 Optics of +15 20 1 0 1` f2 (79) 1.(50/3) = 40/3 cm from the concave lens. For a 1 cm tall object at a distance of 40 cm from the convex lens. Thus. using Eq. the focal length of the combination is 1 1+1_t (83) =a f fi f2 fif2 the unit planes are given by [see Eqs (60) u = 2 The positions and (61)]: of Example 5.f2 fi f2 .2v+32 0. determine the size and position of the image.6+0. due= l =.(12.01 = (2. calculate the position and size of the image.4 Consider a lens combination consisting of a convex lens (of focal length + 15 cm) and a concave lens (of focal length . 5. (79) and three of them of the type given by Eq.f2 . is given by dui = dug = 1 .10 of the two lenses can be obtained by noting that the matrix the two lenses are [see Eq. we get [using Eq.
we have dt1 = 1b = 80 cm a ca 11 and d. The planes which pass through these points and are normal to the axis are known as nodal planes. one each in the object and the image space.6 + 0. d=8 If we now use Eqs (60 and (61). The magnification is va M where the refraction matrix is given by (1 P 0 1 with p = n2 . unit planes are two planes. Using the unit planes draw the ray diagram. [Ans: (a) a = 1/15 b = 1/3 c =l d = 20. The thickness of the lens is 5 cm and the refractive index of the material of the lens is 1.. it is at a distance of 14.01 32 =+1 2.10. Use Eq.The Matrix Method in Paraxial Optics 5.e.second lens.] 5.2 Summary ♦ In the paraxial approximation we may confine ourselves to rays which pass through the axis of the system..^(xl1 X (9/5 1/1001 3/5 JI 8 Thus a b= 100 f = 100 cm c= 5.2 Consider a thick biconvex lens whose magnitudes of the radii of curvature of the first and second surfaces are 45 cm and 30 cm respectively. (b) The final image is virtual and is 15 cmaway (on the. between which the magnification M is unity. (a) Determine the system matrix elements and the positions of the unit planes. (b) Assume a parallel beam of light incident from the left.e.6 In the above example determine the system matrix and hence the positions of the unit planes. Solution: (x2 / T where the translation matrix T is given by 1 0 T= Dlnt 1) The system matrix is given by (1 1/i 0)( O) (Ol 1 20) S= 8 1 ♦ The effect of refraction through a spherical refracting surface (separating media of refractive indices ni and n2) is given by ( A l . 5.01 v=0.5 cm to the left of the second unit plane or at 14. = n sin a which represents the product of the refractive index with the sine of the angle that the ray makes with zaxis. use Eq.2 2. Such a ray can be specified by its distance from the axis of the system x.5. i. a ray striking the first point at an angle a emerges from the second point at the same angle.left) from the . is given by i. these rays remain confined to a single plane. Determine the elements of the system matrix and positions = u =+ 1 2. 4.nl R ♦ By successive application of the above matrices one can study paraxial imaging by a coaxial optical system.60.2 N ' A. If we compare this with Eq.5 cm to the left of the concave lens as shown in Fig. ♦ In an optical system.e. i. ♦ If a ray is initially specified by a (2 x 1) matrix with elements Al and x1. (67) to determine the position of the image.. (45). we obtain M = 0.2 = = 40 cm Thus the first unit plane is at a distance of 80 cm to the right of the convex lens and the second unit plane is at 40 cm to the right of the concave lens. Finally. 1 _ Problems 5.I ) Example 5. the first convex lens is in the middle of the two unit planes. (67) to obtain 1 22 1 __ 1 + u 100 120 1200 v = 600cm 11 Thus the image is at 54. any paraxial ray emanating from the unit plane in the object space will emerge at the same height from the unit plane in the image space. ♦ Nodal points are two points on the axis which have a relative angular magnification of unity. The object distance from the first unit plane is therefore given by u = (80 + 40) = 120 cm We now use Eq. and the quantity R. then the effect of translation through a distance .11 D in a homogenous medium of refractive index n1.5 cm to the left of the concave lens.1 Consider a system of two thin convex lenses of focal lengths 10 and 30 cm separated by a distance of 20 cm in air. (67) and the positions of the unit planes to determine the image point.
2 = 40/91 cm] 5.12.5.13 5. Determine the positions of the principal planes. 1966. Geometrical Optics: The Matrix Theory. Simmons and M.0455.. Interpret the negative a1 ^t sign in the expression for magnification. the image is virtual nl J l and at a distance of (1+ nt r from the surface.. 32. [Ans: a = 0. show that the focal length is 40 cm. 5.3333. d = 3.9444.14. The refractive index of the material of the lens is 1. 5.9630.= a2) is given by fl/f2.12 of the unit planes and use Eq. `An application of matrix optics. Vol. Using matrix method show that for an object at a distance of C 1 + n2 r from the surface. 6. 7.02716. Burch. and S. States. If H1 and H2 denote the positions of the first and second principal points. then show that AH1 = 13. and J.N. 4.' American Journal of Physics. Geometric Optics: An Introduction.J.3636. Further. d„1 = 2. . AddisonWesley Publishing Co. W.5. Show that the angular magnification of the lens combinations (which is just.5 Consider a combination of two thin lenses of focal lengths f1 and f2 separated by a distance (fi + f2). A Nussbaum. Eakin. Brouwer. c 0.13. AddisonWesley Publishing Co. Gerrard.1. 4. K. [Ans: d„ 1 = 20/91 cm. John Wiley & Sons. 5. Mass. the radii of curvature of the first and second surfaces are 10 cm and + 20 cm respectively and the thickness of the REFERENCES AND SUGGESTED READINGS 1: J. 758.M. 1971. 4. Marcel Dekker. Waves and Photons: A Modern Introduction to Light. J. 3. `Matrix representation of Gaussian optics. New York. 1964.P. Reading. New York.0 cm. Blaker. Benjamin. 90. 34. 5. n2 J Fig. Final image at a distance of 60 cm from the second surface.6 Consider a spherical refracting surface as shown in Fig.4 Consider a thick lens of the form shown in Fig. A. Optics Fig. Halbach.3 Consider a hemisphere of radius 20 cm and refractive index 1. D. 1970. d. Vol. 2 = . Mass. Introduction to Matrix Methods in Optics.3 cm and that H2 lies on the second surface as shown in Fig. 2. 1964.14 lens is 1.' American Journal of Physics. 5.. New York. Davis. b = 0. Guttmann.] 5.5. Matrix Methods in Optical Instrumental Design. (67) to determine the image point of an object at a distance of 90 cm from the first surface.M.W. 1975. 1968. Reading. 5.
1 INTRODUCTION In (chapter 4. Richard Feynman in Feynman Lectures on Physics. this is known as chromatic aberration. In practice. i. So the subject is really ultimately quite simple. I 6. The five coefficients represent the spherical aberration. neither of the above assumptions is true. including analysis of aberrations. astigmatism.Geometrical optics is either very simple or else it is very complicated If one has an actual. and involves no new principles. It should be mentioned that if a polychromatic source (like white light) is used for image formation (which is indeed the case for many optical instruments) then. 6.e. nonparaxial rays also take part in image formation. Thus. they are also called monochromatic aberrations. For a polychromatic source. all rays emanating from a single object point converged to a single image point and the magnification of the system was a constant of the optical system. while studying the formation of images by refracting surfaces and thin lenses. coma. the images of objects were perfect. One can set up the problem and make the calculation one ray after another very easily. People have said that this is too tedious.2 CHROMATIC ABERRATION Let us consider a parallel beam of white light incident on a thin convex lens as shown in Fig. Vol.1. independent of the particular ray under consideration. the image will appear to be coloured. It can be shown that the primary aberrations of any rotationally symmetric system can be specified by five coefficients. This will be followed by a discussion of monochromatic aberrations. We had found that in the realm of paraxial optics. . chromatic aberration is due to the dependence of the refractive index of the material of the lens on wavelength of the radiation under consideration. detailed problem in lens design. with computing machines. it is the right way to do it. one in fact has to deal with rays making large angles with the axis. the point at which the blue light would focus is nearer the lens than the point at which the red light would focus. 6. This departure leads to what are known as aberrations. the actual images depart from the ideal images. Since in real optical systems. the images will be coloured. Since blue light gets refracted more than red light. the wavelength dependence of the refractive index results in the coloured image. curvature of field and distortion. different wavelength components (after refraction) proceed along different directions and form images at different points. this leads to coloured images. In this chapter. These are called the Seidal aberrations. Since images formation is accompanied by refraction at refractive index discontinuities. in general. it may be mentioned that this aberration is independent of the five Seidel aberrations to be discussed in later sections. but today. then he has to simply trace the rays through the various surfaces using the law of refraction and find out where they come out and see if they form a satisfactory image. Since chromatic aberration is the easiest to understand. we will consider the five kinds of aberrations separately and discuss the effect on the image when each one of them is present separately. we would discuss this first. Physically. Since these aberrations are present even for light of a single wavelength. we had made the assumption that the object point does not lie far away from the axis of the optical system and that the rays taking part in image formation are essentially those which make small angles with the axis of the system. The domain of optics dealing with rays lying close to the optical axis and making small angles with it is called paraxial optics.
Similarly. then where co' = nn.. (9) would imply f = . then each wavelength refracts by different amounts. this leads to chromatic aberration (see also Fig. Ri R2 J (4) where R1 and R2 represent the radii of curvatures of the first and second surface for the first lens and. as before. ny (6) =(n1) I R1 R2J (1) (7) If a change of n by Sn (the change of n is due to the change f and f represent the focal lengths of the first and second lens in the wavelength of the light) results in a change off by Sf corresponding to a mean colour which is around the yellow then we obtain by differentiating Eq.nb1 1 +nb1 1 . If fb and fb' represent the focal lengths for the first and the second lens corresponding to the blue colour.2 1 Fb Optics 1 1 + 1^ = (nb . the focal length of the combination corresponding to the red colour would be given by Sn 1 Sn n 1 f Rl R2 nr 1 1 +nr1 1 1 f2 n1 f n'1 f' Fr Sn (2) i. 6. lr (10) are known as the dispersive powers.f . one of the lenses may be made of crown glass and the other of flint glass. represent the refractive indices for the blue and red . nb'. (9). A lens combination which satisfies Eq.1 An achromatic doublet of focal length 20 cm is to be made by placing a convex lens of borosilicate crown glass in . Let nb.nr 1 1 + nr1 1 colours respectively. the expression for chromatic aberration can easily be derived. ny' and nr' represent the corresponding refractive indices for the second lens. This aberration is independent of five Seidel aberrations. for an achromatic doublet the two lenses must be of different materials. 9 in the Prelim pages).1 1 n1 f n'1 f' (5) 1 . For the case of a thin lens. co nnlr and (9) 6. yellow and red colours respectively. f1 R2 f n= f (n' 1) IR R li lzJ nb + nr nb + nr 2 . The focal length of a thin lens is given by 1 Fb where nb 1 1 + nb.1 ( R2 fb fb + (nb' 1) l11 .2. We will find the condition for this lens combination to have the same focal length for the blue and red colours. For example..2). Since w and co' are both positive. then n1 f n' 1 f' n1 f n' 1 f' fr . we may write Fig. Sf = f n1 For the focal length of the combination to be equal for which represents the chromatic aberration of a thin lens. If nb blue and red colours. It may be mentioned that if the two lenses are made of the same material.e. Similarly. Example 6.1) RI .fb =f (nb nlr) would represent the chromatic aberration.1.n = 2 . and if Fb represents the focal length of the combination of the two lenses (placed in contact). we must have and n. f and f must be of opposite signs for the validity of Eq. (9) is known as an achromatic doublet (see Fig. Thus.. such a combination will have an infinite focal length. Thus.6. ny and nr represent the refractive indices for the material of the first lens corresponding to the blue.1 When white light consisting of a continuous range of wavelengths is incident on a lens. 6. (3) or f + f. the primed quantities refer to the second lens.1 The Achromatic Doublet We will first consider an optical system of two thin lenses made of different materials placed in contact with each other. (1) region.(n1) (Rl 1 . then co = 0)' and Eq.
6.56942 w co' t (13) f' _ (ro+w') ff' Now. Assuming nr = 1. Solution: n nb + nr = to t Fig.52264. f = f or and f = 20 x 0. (12).2.1. here the unprimed and the primed quantities refer to the borosilicate crown glass and dense flint glass respectively. for the combination to have the same focal f + f' 20 length for blue and red colours we should have or t(co +co') co co' = [1 . then co = co' and the above equation simplifies to f+ f' t(15) 2 implying that the chromatic aberration is very small if the distance between the two lenses is equal to the mean of the focal lengths. contact with a diverging lens of dense flint glass.3).62901. . for the lens combination to be of focal length 20 cm we = f + must have where.15.61216 = 0.62901+1.01546 1.R2 1 2 n Thus. we obtain Af .1) Rl .62058 2 with a similar expression for 1/f'.62058 1 ++f2 = f2 f. (11). =1.61216 and n 'b =1.56942 or wf'+co'f (14) 6.046 + 0. 6.56942] ff' + ' 20 f. This is indeed the case for the Huygens' eyepiece.2 An achromatic doublet (see Fig. differentiating Eq.l)f' f (n ' 1)f' f' (n1)f or = .61 cm f'= f = . would be If both the lenses are made of the same material. we obtain 1.3 Crown Flint A Fig.0. calculate the focal length of each lens. (9).0. then by differentiating Eq.629011.715 = _ An An' _ t An' An t f f' (n1)f + (n' .51863 1 and 1 _ 1 RI _ R2 An (n . The focal length of the combination F. nb =1.Aberrations f f' 6.61216 = 1.3 The separated doublet.On . we obtain 0.)f Thus. 10 in the Prelim pages). Consequently. w and w' represent the dispersive powers.2 f f.52264 1. n.51462.43058 = 8.2 Removal of Chromatic Aberration of a Separated Doublet Let us consider two thin lenses of focal lengths f and f' and separated by a distance t (see Fig.51462 = 0. as before.f 2 .51863 2 = 1.02715 w= AF Af Of' t Of' t A 1.1.2 f' F2 Substituting in Eq. 1=1+1 t F f f' ff' The focal length of the first lens would be given by (11) (12) f = (n . 6. +nr 2 1.51462 = " 1.1 cm 0. If Af and An represent the changes in the focal length and in the refractive index due to a change Aa in the wavelength.52264 +1.
on a plane AB the circular patch has the least diameter. 6. the image will suffer only from spherical aberration. The distance between the two foci is a measure of spherical aberration in the lens.4(a)].5 The spherical aberration of a convex lens (photograph courtesy Dr. 6. will be absent. for a convex lens. the image of a point object will not be a point. however. hit the axis at different points. K. To see how the rays hitting the refracting surface at different heights could focus to different points on the axis.aberration of a diverging lens.4(a)].5). then different rays emerging from the object converge to different points. This is called the circle of least confusion (see Fig. then in general.3.4(b)]. where the image would have the minimum diameter. rays which are incident at different heights on the lens.3 MONOCHROMATIC ABERRATIONS 6. For example. The plane z 0 separates two media of refractive indices nl and n2 (see Fig. rays which are incident farther from the axis appear to be emerging from a point which is nearer to the lens [see Fig. The combined effect of defocusing and spherical aberration leads to the formation of a circle of least confusion. in the figure we have assumed . If we restrict ourselves to the paraxial region. Similarly. Similarly. 6.6). the distance between the paraxial image point and the point at which the marginal ray strikes the paraxial image plane is called the lateral spherical aberration [see Fig. 6. etc. All other offaxis aberrations like coma. (b) The spherical. The distance along the axis between the paraxial image point and the image corresponding to marginal rays (i. the marginal rays (which are incident near the Lateral spherical aberration Fp spherical aberration (a) periphery of the lens) focus at a point closer than the focal point of paraxial rays [see Fig. 6.K. Let the plane of the refracting surface be chosen as the plane z = 0.1 Spherical Aberration Let a beam of light parallel to the axis be incident on a thin lens (see Fig. Fig. The distance between the paraxial focal point and the marginal focal point is known as the longitudinal spherical aberration and the radius of the image at the paraxial focal plane is known as the lateral spherical aberration. Gupta). The image on any plane (normal to the zaxis) is a circular patch of light. let us consider the simple case of a plane refracting surface as shown in Fig..4).6. then we can see that all rays cross the zaxis at the same point which is at a distance fp from the lens. 6. 6. 6.4 Optics 6. fn represents the paraxial focal length of the lens. 6. If one does not restrict to the paraxial region.. for a concave lens. The point at which the paraxial rays strike the axis (Fe) is called the paraxial focus and the point at which the rays near the periphery strike is called the marginal focus (FM). consequently. Thus if 0 represents an axial object. 6.4(a). Fp (b) Fig.e. as can be seen from Fig. It may be mentioned that for an object lying on the axis of a cylindrically symmetric system (like a system of coaxial lenses). The light rays after passing the lens bend towards the axis and cross the axis at some point. astigmatism. rays striking the edge of the lens) is termed longitudinal spherical aberration. Let P be the object point. The zaxis is chosen to be along the normal (PO) from the point P to the surface.6.4 (a) For a converging lens the focal point for marginal rays lies closer to the lens than the focal point for paraxial rays.
6). at which the ray strikes the refracting surface.7 The aplanatic points of a spherical refracting surface. To the next order of approximation. Thus. The calculation of the spherical aberration even for a single spherical refracting surface is quite cumbersome (see.sin e /3 2 1+1 2 R zo x or = z1 nh 1. both zo and z1 would be negative quantities and the distances OP and OQ would be zo and z1 respectively (see Fig. from Fig. . (21) is an exact expression in terms of zo. Let the zcoordinates of the points P and Q be zo and z1 respectively.0. 5). 6.nZ2 2 0 h2 1 + Z2 0 (21) Fig. i.1) (24) Equation (24) gives the longitudinal spherical aberration. From the above example. z1.6 we have (z 1 ) (17) h cot / 3 = sin / 3 " 1. for paraxial rays. we get (22) z1 = .nlzol 1+ z 1. The refracted ray appears to emerge from the point Q. nlzol 1+ n2 > nl. e. Since.Aberrations 6.g.5 The value of z1 given in Eq. spherical refracting surfaces and thin lenses must also suffer from spherical aberration. is a complicated function of the height h. Consider a ray PM incident on the refracting surface (from the object) at a height h as shown in Fig. We assume the origin to be at the point O.2 z 2n zo 2zo . Obviously. it can be seen that even a single plane refracting surface suffers from spherical aberration.) n z zo n1 where a and /3 are the angles that the incident and refracted rays make with the zaxis and n2 n=n1 Now. 6.1 sin e a n2 sin a h (18) where we have used Eq. z1 =  2 n Izol I 1+ h2 L zo r 1'2 h2 1. (16). we get h2 he z= .e. sin a = we obtain z1 = C (19) n2+ n1 + \ h2 R nl zo (25) V h2 + zo h2 ne z (h + zo ) 1/2  nh 2 2 1/z +zo) h (20) or. In the limit of h . 6..nlzol Q which is the expression for the image distance in the paraxial region. 6.2n Izol (nz . The negative sign implies that the nonparaxial rays appear to emanate from a point which is farther away from the paraxial image point. It can at once be seen that the image distance. Ref.n2 . we just give the final results: 0z=n2 (n2 n1) 1+ n1R 2n.6. From Snell's law we know that sin a=nsin/3 (16) r L 2 2zo n 2 2 (n2 1) (23) Thus the aberration is given by z Az . assuming Ih/zol 1. We have to determine z1 in terms of zo.
For a plane surface R = oo. it can be rigorously shown that all rays emanating from the point A appear to diverge from the point B (see also Sec. 4. For a given focal length of the lens. The corresponding image point B is at a distance . (25) van ishes and the spherical aberration is zero. Figure 6. Example 6. Show that for a point A [see Fig. 6. nl and n2 represent the refractive indices of the media on the left and right of the spherical surface (see Fig.7). The coefficient A is such that when it is multiplied by the cube of the height of the ray at the lens. Solution For zo = nl + n2 R. one can show that the coefficient of spherical aberration of a thin lens made of a material of refractive index n Rl 3 (R2 nf 1)(R2  f) 2] (30) For a converging lens.7. with the surfaces having radii of curvatures R1 and R2. Indeed. (25) reduces to Eq. called the shape factor. [Example 6.1) RI R2 J (28) and are utilized in microscope objectives. 6.6 where R represents the radius of curvature of the surface.2 ! Consider a spherical refracting surface of radius R. for a set of rays incident parallel to the axis.3). then show that the coordinates of the point P (x. by the following relation: _ R2 + R1 R2 .9 shows the variation of spherical aberration with q for n = 1. Thus the lateral spherical aberration for rays hitting the lens at a height h would be Slat = Ah3 = f(n1)h3 2n2  RZ  P)2{ A2 . of a given focal length).e.PJ A2. For a thin lens of given power (i.P(n + l) } + Rl (29) The longitudinal spherical aberration which corresponds to the difference between the marginal focal length and the paraxial focal length would be given by Slong = Ah 2f .5. 6. Siong will always be negative implying that the marginal rays focus closer to the lens.f(n1) 2nA2. Thus. If we assume n 1 (QP) + n 2(PF) = n2(BF). It can be seen that for values of q lying near q = + 0. Eq.6. P = 0.7(b)] such that zo = n R 1 Optics and placed in air. = f= I (n . one can control the spherical aberration by changing the value of q. (24) with n = n2/n 1 . by .nl The points A and B are known as the aplanatic points n2 zo. one of the factors in Eq. This procedure is called bending of the lens. 6. where R1 and R2 are the radii of curvatures of the two surfaces. one can define a quantity q.8). Notice that both R and zo are negative quantities.P(n + 1)1 + x ll2 l nl+n2 (26) where P the spherical aberration is zero. y) will satisfy the equation of the ellipse.3 Consider arefracting surface obtained by revolving an ellipse about its major axis.025 cm I) and h = 1 cm.(n1)f 2h2 2 n2 x where a and b are the semimajor and semiminor axes respectively (see Fig. the (magnitude of the) spherical aberration is minimum (but not zero).e. Show that all the rays parallel to the major axis will focus at one of the foci if the eccentricity of the ellipse is equal to n l ha2.3.f = 40 cm (i.] In a similar manner. would be given by A=. one obtains the lateral spherical aberration. [Hint: The eccentricity of the ellipse is given by OF 8==111r 2 a2 represents the power of the lens.8 For Example 6.Rl (31) Fig.
(b) Fig.11). For calculating the coma we have assumed tan 0 = 1. i. From triangle PA Q.u) =h (I u) = f (34) .10 (a) Refraction at the two refracting surfaces of a choosing proper values of the radii. fi h C C D ^] f1x L2 Fig.12 where a ray PA gets refracted along A Q after suffering a deviation through an angle O.. Thus the spherical aberration is dependent on how the deviation is divided between the surfaces. S1 = CS2 . 6.11 Condition for minimum spherical aberration for a combination of two thin lenses.Using the criterion of equal deviation discussed above. The physical reason for the minimum of ISIo„gl to occur at q = 0.30 Fig.9 Variation of spherical aberration and coma with the shape factor of a thin lens with n = 1.7 is as follows: It has already been mentioned before that (for a converging lens) the marginal rays undergo a large deviation which results in the spherical aberration [see Fig. (b) For a prism. we will determine the separation between two thin lenses which would lead to minimum spherical aberration. 6. the minimum deviation position corresponds to Sl = S2. for a pianoconvex lens with the plane side facing the incident light RI = ° and q = 1.4 (a)]. As in the case of the prism [see Fig. As such we should expect the spherical aberration to be minimum when the angle of deviation S [see Fig. the spherical aberration can be minimized. 6. we get (33) 01 = 02 To obtain an expression for the deviation suffered by a ray when it encounters a lens.Aberrations 0.. then for minimum spherical aberration. Indeed for q = 0.5. If 01 and 02 represent the deviations of the ray at the two lenses. we refer to Fig.e. 6. we can see that h 0= 01+02=h+ v (. 6. 6. 6.10 6.10 (a)] is minimum. 6. the diagram is exaggerated to show clearly the angles. Let LI and L2 be two lenses of focal lengths f1 andf2 respectively separated by a distance x (see Fig.10 (b)]. i. (s = S1 + 52) (32) thin lens. On the other hand. It may be mentioned that the value q = +1 implies R2 = oo and hence it corresponds to a pianoconvex lens with the convex side facing the incident light. the deviations suffered at each of the surface are equal and one obtains minimum spherical aberration. this would occur when the deviations suffered 'at each of the refracting surfaces are exactly equal. rays make an angle of 45° with the axis.7.e. .f = 40 cm and h = 1 cm.7 0.
(15)] and minimum spherical aberration [see Eq.00022 mm respectively (see Sec. the focal length of the field lens is 3f paraxial image point (see Fig. where we have used the paraxial relation 1_1 1 v u f (35) The quantity u is an intrinsically negative quantity. We can immediately see that the conditions for achromatism [see Eq. for a.22)i. 6. (38)] are simultaneously satisfied. the image of the cross wires (which are placed in the plane PQ) will show aberrations.x .3.6. (33) becomes h1 __ h2 (36) fi f2 From similar triangles ACID and BC2D (see Fig. 6. = 50001. 18.6.14 A perfectly spherical wave (converging on the plane PP') will produce an Airy pattern in the image plane. the radii of the first and second dark rings in the Airy pattern will be about 0. Often one uses a `stop' to restrict to the paraxial region. . Indeed.14 is highly magnified. D = 5 cm. however.f1f2 Thus the spherical aberration of a combination of two thin lenses is a minimum when their separation is equal to the difference in their focal lengths.3).13 The Huygens eyepiece. if a perfectly spherical wave is emanating from a lens. Thus Eq. Since the eyepiece as a whole is corrected and the individual lenses are not. 6. we obtain h1 from the (37) (38) x = . It should be mentioned that even when the system is free from all aberrations the image of a point object will still not be a point because of diffraction effects (see Sec.2 Coma Fig.3). in the Huygens eyepiece (see Fig. where f represents the focal length of the eye lens. (36) and (37). 6.14) where D is the diameter of the exit pupil.8 Optics Fig.11). 18. f D we can write h2 . For example. For example.8 (c)]. The distance between the two lenses is 2f. f = 10 cm. 6. at high apertures aberrations degrade the image and at low apertures diffraction degrades the image.00012 mm and 0. 6. if the diameter of the `stop' is made very small then the diffraction effects would dominate. the ray theory predicts a point image whereas the diffraction theory (which takes into account the finiteness of the wavelength) predicts that the image formed in the image plane will be an Airy pattern [see Fig.fi . For offaxis P f P' P' Fig. As mentioned earlier for a point object lying on the axis the image will suffer only from spherical aberration. 6. 6. Indeed. and the first dark ring will occur at a distance of 1. The Airy pattern shown in Fig. A discussion of the procedure for reducing the aberrations in various optical instruments requires a very detailed analysis involving the tracing of the rays. 18.13).fi If we use Eqs. which is beyond the scope of this book. a camera gives best image when fID 5. The spatial extent of the Airy pattern will become larger with decrease in the value of D.12 Calculation of the angle of deviation.
g. 6. 6.15(a) we have shown only those rays which lie in the meridional plane. Fig.e. Ref. curvature of field and distortion. The first offaxis aberration is coma. assuming that all other aberrations are absent. it appears that the magnification is different for different parts of the lens. The Ronald Press Co. It may be mentioned that if we consider the image formation by different zones of a lens. the image will also suffer from coma. the image will suffer from spherical aberration and coma only. The image of a point object thus has a comet like appearance and hence the name coma (see Fig. As the radius of the zone [shown as h in Fig. . Vol. used with permission.16 Image of a point source showing coma. Physics Demonstration Experiments.F. 1970.* In Fig. 6.9 focus at a point different from that of the marginal rays. The radius of the circle and the distance at which the centre lies from the ideal image point measures the coma. one can show that the coma in the image is given by (see. 6. then the spherical aberration arises due to the fact that different zones have different powers and coma arises due to the fact that different zones have different magnifications.. Thus the composite image will have a form as shown in Fig. Meiners. for points lying very close to the axis. In this section we will briefly discuss the effect of coma.15(a). Thus.) * It must be mentioned that a proper understanding of the aberrations can only be had by a careful and thorough mathematical analysis.e. 6. 6. This.15 (c). II. i. Rays which intersect the lens at diametrically opposite points focus to a single point on the paraxial image plane. (b) a three dimensional perspective is shown. To see the shape of the image one has to consider the complete set of rays. e. New York. In (a) we have shown only those rays which lie in the meridionalplane. The rays which proceed near the axis of the lens (c)  Fig.15 The image formation in the presence of coma. i. In (c) we have shown the composite image.15 (b)] increases. however. For a parallel bundle of rays incident on a lens and inclined at an angle 0 with the zaxis (see Fig. 6. 6. These different pairs of rays focus to different points in the image plane such that these foci lie on a circle. astigmatism. that plane containing the optical axis and the object point. 6..Aberrations 6. the centre of the circle also shifts away from the ideal image. (After H. 1): points.15(b) we have shown aThreedimensional perspective in which we have considered a set of rays which hit the lens at the same distance form the centre. In Fig. interested readers may look up Refs 1 and 3. The effect of coma is schematically shown in Fig. is beyond the scope of this book.16).17)..
6.18 shows'the image formation when the optical system suffers from astigmatism only.3 Astigmatism and Curvature of Field When an optical system is free from spherical aberration and coma then the system will image sharply those object points lying on or near the axis. The plane containing the axis and the object point is called the meridional plane and the plane perpendicular to the meridional plane (containing the axis) is called the sagittal plane. 4. Since at the point T. The rays in the meridional plane converge at a different point as compared to those in the sagittal plane. coma = 3(2 1) fh2 tan20 x (n1)(2n+1) _ n2 n1 _ n (39) nRl R2 n2 R1 R2 In Fig.3. It can immediately be seen that for a lens with q = + 0.0 and as such pianoconvex lenses are extensively used in eyepieces.11 we had derived the Abbe sine condition which when satisfied. this is called the sagittal focal line. But for points far away from the axis. It can also be seen that both spherical aberration and coma are close to a minimum for a pianoconvex lens (with the convex side facing the incident light) for which q = 1. 6.8. coma is zero. one in fact has a focal line which is normal to the meridional plane. For example.6. rays P4 and PB focus at the point T and rays PC and PD focus at a point S which is different from T. the rays in the sagittal plane have not still focused.9 we have plotted the variation of coma with the shape factor q. it converges to a single Fig. Consider an object point P far away from the axis. 6:18 Image formation in the presence of astigmatism. the rays in the meridional plane have defocussed. one obtains a focal line lying in the tangential plane. The distance between S and T is a measure of astigmatism. the optical system is free from spherical aberration and coma. . This focal line T is called the tangential focus. Figure 6. Fig. the image of a point will not be a point and then the optical system is said to be afflicted with astigmatism. Similarly since at S.17 Parallel rays (inclined at an angle 6 with the axis) incident on a thin lens. We may mention here that in Sec. To see the origin of astigmatism one observes that for a point on the axis (when the lens is free from other aberrations) the wavefront emerging from the lens is spherical and thus as the wavefront progresses.10 Optics 6.
the spokes will be in focus and the rim will not be in focus as shown in Fig. Thus the tangential foci and the sagittal foci of points at different distances from the axis lie on two surfaces as shown in Fig. each point will be imaged as a point but if the system suffers from nonuniform magnification. Fig. C' and D' respectively.22(b) corresponds to a negative value of E and is known as barrel distortion. the points having coordinates (0. As an example of image formation in the presence of astigmatism.19. Somewhere between the two focal lines.21 In the presence of a pinhole on the axis. This is because of the fact that corresponding to any point in the object plane.e. 6.19 Yd =Myo+ E(x o+ y o) y o Tangential and sagittal foci.4 Distortion The last of the Seidel aberrations is ' called distortion and is caused by nonuniform magnification of the system. all other aberrations will be absent. Similarly. The optical system will be said to be free from astigmatism when the two surfaces coincide. it does not focus to a point but to two lines.20(b).21).22. the image suffers only from distortion.Aberrations 6. point. M represents the magnification of the system and E represents the coefficient of distortion. 6. then the emerging wavefront is not spherical and thus as the wavefront converges. Yd) represent the coordinates of the object and the image point respectively. 6. 6. consequently. . When we discussed spherical aberration we had mentioned that for a point object on the axis of the optical system. 6. for such a configuration.20 (a) Spoked object coaxial with the axis of the lens. Figure 6. 0). the images will suffer only form spherical aberration. C and D which are imaged as A'. 6. The distance between the tangential and the sagittal foci increases as the object point moves away from the axis. The distortion of the image can be easily understood if we consider the imaging of a square grid as shown in Fig. the image is circular in shape and is called the circle of least confusion. Mathematical analysis shows that 3 : Xd A V = Mxo + E(xo + yo) xo (40) (41) and S Paraxial image plane Fig. since on the Ssurface the image of a point is a line in the meridional plane. the complete rim of the wheel where (xo. yo) and (Xd.3. Assuming unit magnification (i. 6. Fig. which are normal to each other and called the tangential and the sagittal focal lines.11 will be in focus while the spokes will be out of focus as shown in Fig.20(c). B. then the image will suffer only from distortion. if we have a pinhole on the axis at any plane of the optical system (see Fig. 6. 6. only one of the rays emanating from this point will pass through the pinhole.20(a). Similarly. (h. consider a spoked wheel coaxial with the lens axis as shown in Fig. But when the object point is nonaxial. This defect of the image is termed curvature of the field. Since on the Tsurface the image of a point source is a line perpendicular to the meridional plane. Obviously. (b) and (c) show images on the Tsurface and Ssurface respectively. on the Tsurface. This can be illustrated if we consider the imaging of four equally spaced points A. M = 1). But even when they coincide it can be shown that the resultant image surface will be curved. 6. the image will be distorted. B'.
. but the image is distorted because of nonuniform magnification. h + Eh 3 ). If the reader actually plots these points. By simple application of Snell's law obtain an expression for the spherical aberration of the slab. nr. 3h + 27Eh 3 ). height at which the ray strikes the slab. (0. (3h. 2h).12 Optics E<0 (a) (b) E<0 Fig. [Ans: Spherical aberration = . (h. Similarly for E > 0. 0). obtain the spherical aberration and coma for the lens for various values of the shape factor q and plot the variation in a manner similar to that shown in Fig. (h + Eh3. (b) represents the image when 0)..6. 0). 6.4 Does the image formed by a plane mirror suffer from any aberration? 6. 6. h). 0). h).49776) and a crown glass lens (nb 1.3 Obtain an expression for the chromatic aberration in the image formed by a plane glass slab. (0. and (c) when E > 0. Calculate the radii of curvatures of the different surfaces and the focal lengths of each of the two lenses. Similarly. ny and nr represent the refractive indices for the material of the first lens corresponding to the blue.nr n1 and of = nb nr n' 1 are known as the dispersive powers. (0. and u is the distance of the object point from the front surface of the slab. 2h). respectively. different wavelength components (after refraction) proceed along different directions and form images at different points. the focal length of the combination will be the same for blue and red colours if co' f +f' where Q) = nb . 6. this leads to chromatic aberrations. 6. . (2h. (2h + 8 Eh3. yellow and red colours respectively.9. ♦ For a lens. placed in air.22 (a) shows the object. (h. h).5 and whose curved surface has a radius of curvature of 10 cm. 3h). he would obtain Fig. The distance along the axis between the paraxial image point and the image corresponding to marginal rays rays striking the edge of the lens) is termed longitudinal spherical aberration. If we consider two thin lenses made of different materials placed in contact with each other. 6.2 Why can't you obtain an expression for the spherical aberration of a plane glass slab from Eq. R2 to co? 6.6 Consider a lens made up of a material of refractive index 1. where h is the 2n3'2h ..22(c). the marginal rays (which are incident near the periphery of the lens) focus at a point which is different than the focal point of paraxial rays. Since co and co' are both positive. 2h + 8 Eh). (27) by tending R1. (0.50529. f and f' must be of opposite signs. 0).] 6. What are other kinds of aberrations that the image will suffer from? z 2 Summary ♦ For a polychromatic source.are imaged at (0. Assuming h = 0. = 1.5 Calculate the longitudinal spherical aberration of a thin planconvex lens made of a material of refractive index 1.1 Consider a plane glass slab of thickness d made of a material of refractive index n. ny and n. .22(b).66270. n = nb+nr =n . represent the refractive indices for the second lens. (2h. 2h + 8Eh 3 ).(n . nr = 1. (0.7 An achromatic cemented doublet of focal length 25 cm is to be made from a combination of an equiconvex flint glass lens (nb = 1.5 cm and 0 = 45°. Further. 6. 0). h + Eh 3 ). 0).5 with a focal length 25 cm. 6... (2h + 8Eh 3 . ♦ The spherical aberration of a combination of two thin lenses is a minimum when their separation is equal to the difference in their focal lengths. (h + Eh3. then for E < 0. (0.64357). Notice that each point is imaged at a point. Problems 6. n'= nb+nr =ri y 2 y 2 where nb. h + Eh 3 ). for rays incident at a height of 1 cm. nb'. (3h + 27Eh3 . he would obtain a figure like the one shown in Fig. Compare the values of the aberration when the convex side and the plane side face the incident light. (h + Eh 3 .
H. 1962. Part III. A. Plenum Press. Pergamon Press.H Hopkins. 7. 5. 1960. `A Degree Physics'. 1950. London. T. Welford. London. Oxford.Aberrations 6. Smith. Amsterdam. M. and E. New Delhi.. 1978. SpringerVerlag. M. T. North Holland Publishing Co.13 REFERENCES AND SUGGESTED READINGS 1. Berlin. Francon and J. 1962. Born. 2. W. Principles of Optics. Wolf. New York. Wave Theory of Aberrations. 3. Oxford University Press. Contemporary Optics. 1974.C. . Welford. 6. Cagnet. Thyagarajan. Optics. Aberrations of the Symmetrical Optical System. Thierr. Geometrical Optics. New York. Ghatak and K. Edward Arnold Publishers.J. Atlas of Optical Phenomena. W. [Reprinted by Macmillan India. 1975. C. Academic Press.] 4. M.
.
. Chapter 12 discusses Huygens' principle which is used to derive the laws of reflection and Snell's law of refraction. shock waves. Chapters 8 and 9 discuss Fourier series and Fourier transforms which are extensively used in studying the distortion of optical pulses as it propagates through dispersive media (Chapter 10). redness of the settin i sun. The derivation and solutions of the wave equation represents the basic physics of wave propagation which have been discussed in Chapter 11. water waves. etc.2 Vibrat ions and ave s PART This part (consisting of five chapters) discusses many interesting experiments like the physics behind ionospheric reflection. Chapter 7 starts with simple ha monic motion (which is the most fundamental vibration associated with wave motion) and is followed by a derivation of the refractive index variation with frequency. pulse dispersion.
.
6).2 SIMPLE HARMONIC MOTION A periodic motion is a motion which repeats itself after regular intervals of time and the simplest kind of periodic motion is a simple harmonic motion in which the displacement varies sinusoidally with time. 7. 7. produces a changed field which is equivalent to a phase shift of the original wave.4 we will study such vibrations which will allow us to understand the origin of refractive index (see Sec. 7. producing oscillating dipole moments. 7. 7.1 INTRODUCTION The most fundamental vibration associated with wave motion is the simple harmonic motion. Richard Feynman in Feynman Lectures on Physics. 7. makes them behave like little oscillators. In order to understand simple harmonic motion we consider a point P rotating on the circumference of a circle of radius a with an angular velocity co Fig. FORCED VIBRATIONS AND ORIGIN OF REFRACTIVE INDEX The correct picture of an atom. (see Fig. Because this phase shift is proportional to the thickness of the material.Chapter Seven SIMPLE HARMONIC MOTION. which is responsible for the red colour of the setting (or rising) sun and blue colour of the sky. This new field. the effect is equivalent to having a different phase velocity in the material. I 7. . At an arbitrary time t the point will be at the position P where ZPOP0 = cot.1). which is given by the theory of wave mechanics.. 7. The foot of the perpendicular on any one of the diameters executes simple harmonic motion. If a periodic force acts on a vibrating system. together with their mass m. Po is the position of the point at t = 0. Vol.2 we will discuss simple harmonic motion and in Sec.3 we will discuss the effects (on the vibratory motion) due to damping. with a resonant frequency coo The electric field of the light wave polarizes the molecules of the gas. in Sec. We choose the center of the circle as our origin and we assume that at t = 0 the point P lies on the xaxis (i.1 The point P is rotating in the anticlockwise direction on the circumference of a circle of radius a. at the point Po).e. So we shall suppose that the electrons have a linear restoring force which. says that. in Sec. interfering with the old field. the system undergoes what are known as forced vibrations. the electrons behave as though they were held by springs.5) and even Rayleigh scattering (see Sec. with uniform angular velocity co. Y 7. so far as problems involving light are concerned. The acceleration of the oscillating charges radiates new waves of the field.
discussion that the value of 0 is quite arbitrary and depends. 7. (1) because when P coincides with PI. Let the angles LP'OX and LQ'OX be dicular on any one of its diameters will execute simple 6 and 0 respectively (see Fig. (1)] is known as a simple harmonic We next consider two points P and Q rotating on the circle motion. After crossing the point Po. . the distance OA = a cos cot (1) . 7.3). When P coincides with P2. the point A moves to and fro about the origin on the diameter. Since the angular veloc(6a) xp = a cos (cot + O) ity is co. cot = 7r/2 and hence a cos cot = a cos 7r/2 = 0. then the foot of the perpendicular is at O. The choice of the time t = 0 is arbitrary and we could have chosen time t = 0 to be the instant when P was at P' (see Fig. Clearly at an arbitrary harmonic motion. This can also be seen from Eq.point P moves from P2 to P3. T.0= 0 (or an even multiple of ti) the __ 1 __ co v 27r T or co = 2 irv (3) It should be pointed out that we could as well have studied the motion of the point B. then OA = OP2 = a. The initial phases of A and B are O and respectively. 7. which is the foot of the perpendicular from the point P on the yaxis.0 co represents the phase difference between the two simple harThe inverse of the time period is known as the frequency: monic motions and if 0.= 0.7. and finally acquires the value a when P coincides with Po.3 The points A and B execute simple harmo'riic motions with the same frequency c . 7. (1) because cot would then be greater than n/2. The quantity a is called the amplitude of time t. After P crosses P 3. 7. When the . the foot of the perpendicular would lie on the other side of the origin and thus OA would be negative as is also evident from Eq. will be the time origin would be required to complete one revolution. (7) (cot + B) .2 At t. A motion in which the displacement varies sinusoidally on the instant from which we start measuring time.1) OB = y = a sin cot (4) We had conveniently chosen t = 0 as the time when P was on the xaxis.4 Let A be the foot of the perpendicular from the point P on the xaxis.tive positions at t = 0. Clearly. and the period of the motion. The distance OB is given by (see Fig. If the angle LP' OX = 6 then the projection on the xaxis at any time t would be given by (5) OA=x=acos(cot+6) The quantity (an. (cot + 0) 27r/co. Thus. Optics and as the point P rotates on the circumference of the circle. OA starts decreasing and it finally goes to zero when P coin. The quantity T2 (2) . when a point rotates on the circumference of a with the same angular velocity and P' and Q' be their respeccircle with a uniform angular velocity. the distance of the foot of the perpendiculars from the the motion.(cot + 0) = 0.Fig. When the point P is at PI. the initial cides with P2. the foot of the perpen.+ 0) is known as the phase of the motion and ° represents the initial phase. the point P is at P' and therefore. As the point still moves further.OA starts increasing again phase is O.2). It is obvious from the above Fig. the motion repeats itself. Thus. the time taken for one complete revolution will be (6b) xQ = a cos. with time [as in Eq.
4(b)]. (Here the point x = 0 is the fixed point and is usually referred to as the equilibrium position. the phase difference (0 . Thus the displacement of a particle. Harmonic Motion In this section we will discuss three simple examples of simple harmonic motion. We could have equally well started from Eq.1 Examples of Simple.Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index motions are said to be in phase. However.0 = it (or an odd multiple of 7c) the motions are said to be out of phase. the equation of motion would be m d2x = dt2 or 2 dx + k x=0 m dt2 = kx or d 2x +w2x=0 dt2 or 7. Under such an approximation we may assume that this force is always directed towards the point B and the magnitude of this force will be* mg sin 9 = mg x (16) Thus the equation of motion will be F=m= mg x d2x l dt2 (17 (18) * We will be assuming that 6 is small so that sin 9 = 0. The motion of the bob is along the arc of a circle but if the length of the pendulum is large and the angle 0 is small.4) then the forces acting on the bob are the gravitational force mg acting vertically downwards and the tension T.) If we multiply Eq. the motion can be assumed to be approximately in a straight line [see Fig. Equation (10) can be used to define the simple harmonic motion as the motion of a particle in a straight line in which the acceleration is proportional to the displacement from a fixed point (on the straight line) and always directed towards the fixed point. the quantities 9 and 0 would change by the same additive constant. then we obtain the following expression for the force acting on the particle: f_ F=mf=mw2 x or F = kx (11) where k (= m co2) is known as the force constant. in the direction B'A. This can easily be seen by noting that since the force is acting in the xdirection. the velocity and the acceleration of the particle would be given by the following equations: v and d2 = awe sin (wt + 0) f dt2 or. In the equilibrium position (AB) the tension is equal and opposite to the gravitational force. consequently. from the equilibrium position (see Fig. The above approximation is valid for 9 < 0.0) is independent of the choice of the instant t = 0. 7. (a) The simple pendulum The simplest example of simple harmonic motion is the motion of the bob of a simple pendulum in the gravitational field. If the bob of the pendulum is displaced slightly. 7.07 radians (= 4°). where B is in radians. which executes simple harmonic motion. (10) by the mass of the particle. (11) and obtained simple harmonic motion.5 d2x (12) + w2x = 0 dt2 where cot = k/m. If we choose a different origin of time.2. can be written as: x = a sin (wt + 9) (8) or x = a cos (cot + 0) which describes a simple harmonic motion. (12) can be written in the form x = A sin wt + B cos wt which can be rewritten in either of the following forms: x = a sin (wt + 0) (14) (15) (13) = d = acv cos (cot + 0) (9) 7. and if 9 . d2x (10) dt2 Equation (10) shows that the acceleration of the particle is proportional to the displacement and the negative sign indicates that the acceleration is always directed towards the origin. in the displaced position the tension T is not in the direction of the gravitational force and if we resolve the gravitational force along the direction of the string and perpendicular to it. we see that the component mg cos 0 balances the tension in the string and the component mg sin 0 is the restoring force. Therefore. The general solution of Eq. .
5(b). The restoring force is FS which is equal to mg sin O. positions of the pendulums.5 (a) and (b).4(a) it can immediately be seen that Potential energy.e. then the equation of motion would have been 2 mgl (l) 2 2 m ( ) x2 g (24) or x2 =a sin cot = a cos (cot +2 V = 2 m cot a2 cos2(cot + 0) (25) . (21) (a) (b) Fig. show the motion of two identical pendulums which are vibrating with the same amplitude but having a phase difference of n/2. the bob of one of the pendulums be at its extreme right position. Since.7. 7. the kinetic energy is zero and when the particle passes through equilibrium position.6 Optics (a) Fig. the kinetic energy gets transformed into potential energyFrom Fig. moving towards the right [Fig. the kinetic energy is maximum. 7. the velocity of the particle would be given by .5). 7. The small circles denote the position of the bobs at t = O. 7.5(b)]. thus the motion of the bob is simple harmonic with its time period given by the following equation: dx = cal) sin (cot + 0) dt Thus the kinetic energy of the mass would be T= 12 (22) 2'n (d) T= 2^ =2tr g (19) It should be pointed out that the expression for the time period. V = mgh = mgl (1 . 7.cos 0) = mgl 2 sin /2 [0 measured in radians] %2  2m g l(^)  x2 = a sm cot = a cos (cot 2) (20) 1 Thus the two bobs execute simple harmonic motion with a phase difference of 7r/2 and in fact the first pendulum is ahead in phase by 7r/2. Let. is fairly accurate (i. the displacement of the bob of the pendulum can be written as x = a cos (cot +4)). If we measure the displacement from the equilibrium. (12). 7. then the displacements would be given by xi = a cos cot 1 (23) ma2co2 sine (cot + 0) 2 Comparing Eqs (21) and (23).4 (a) The forces on the bob of the pendulum when it is displaced from its equilibrium position. the motion is approximately simple harmonic) as long as 0 < 4°. At the extreme positions. We next consider the motion of two identical simple pendulums vibrating with the same amplitude a (see Fig. (b) If the angle 0 is small. the motion of the bob can be approximately assumed to be in a straight line. in general. we see that when the particle is at its extreme positions. It may be mentioned that in Fig. at t 0. and then the second pendulum would have been ahead of phase by 7r/2. where cot = Equation (18) is of the same form as Eq. if the bob was moving towards the left..
held by two stretched springs on a smooth table as shown in Fig.6(a)] and corresponding to the equilibrium position of the mass. all particles execute simple harmonic motions with same frequency and same initial phase but having different amplitudes. 7.7).6. then each point on the first half of the string vibrates out of phase with each point on the other half. 7. Equation (32) tells us that the motion will continue forever. We can also see from Eq. if the string is vibrating in its first harmonic.^ 1 the solution of which is given by 99999 1 I x(t) = A cos (coot + 0) (32) (c) Fig.6 Two springs of natural length to [see (a)] are stretched to a length 1 [see (b)] to hold the mass. the damping forces are primarily due to the viscosity of the surrounding medium.2. 7. If the mass is displaced slightly from the equilibrium position. Consequently. 7. Fig. We may mention ment and directed towards the equilibrium position and conthat the expression for potential energy could have been di. Once again we get a force which is proportional to the displace DAMPED SIMPLE HARMONIC MOTION 0 to vcxxxxxi "i to vcxxxxxi In Sec. The displacement can be written in the v = kx dx = kx2 form 0 where we have assumed the zero of the potential energy to be cos cot (30) y = a sin I L at x = 0. Thus the total energy E would be given by xJ f 2 E=T+V= 2 mco2 a2 (28) which.Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index 7. as expected. then each point on the string executes simple harmonic motion with different amplitudes but having (27) the same initial phase. the mass will execute simple harmonic motion. (26) that the energy associated with the simple harmonic motion is proportional to the square of the amplitude and the square of the frequency. 7. where A represents amplitude and coo the angular frequency of motion. the motion of the mass on the frictionless table rectly written down by noting the fact that if the potential will be simple harmonic in nature. is independent of time. If the mass is displaced by a small distance x from its equilibrium position [see (c)]. the damping forces will be much .sequently. then the resultant force acting on the mass will be F = k [(lx)1o] k[(l+x)1o] =2kx The amplitude is therefore zero at x = 0 and at x = L and is maximum at x = L/2. On the other hand. (b) Vibrations of a mass held by two stretched springs Another simple example is the motion of a mass m. the equation of motion will be of the form d 2 dt 2 +coox(t)=0 (31) (a) 999999 (b) k. the bob of a pendulum comes to rest after a certain period of time. For a vibrating pendulum.x . 7.7 where we have used the fact that cot = g/l. we know that in actual practice the amplitude of any vibrating system (like that of a tuning fork) keeps on decreasing and eventually the system stops vibrat ing. Similarly. we had shown that for a particle executing SHM.7 When a string clamped at both the ends is made to vibrate in its fundamental mode.3 where k represents the force constant of the spring. energies at x and at x + dx are V and V + dV. However. This phenomenon is due to the presence of damping forces which come into play when the particle is in motion. then (26) (c) Vibrations of a stretched string When a stretched dV = F dx = + kx dx string (as in a sonometer) is made to vibrate in its fundamenThus tal mode (see Fig. The two springs are of natural length 10 [Fig. the lengths of the stretched springs are 1. (29) 7.
The figure corresponds to Jr =1sand K=0. 2 coo . we must consider three cases./K2 . 7. the force constant is now denoted by k° to avoid confusion with the wave vector k. the quantity (coo .K4(t)1 e Kt ate eKt d2 2K d + K2 fi(t) dt2  d .jK2 .5s1. Notice that the amplitude decreases exponentially with time . Case I (w > K2) If the damping is small. the equation of motion will be given by 2 and =I 2 k°x (33) Optics where A and 9 are constants which are determined from the amplitude and phase of the motion at t = 0. Consequently.K 2 Case II (K2 > coo) If the damping is too large.wo) dt2 4(t) = 0 (40) the solution of which is given by 4(t) = A exp C IK2 . This is also consistent with the fact that there are no damping forces acting on the particle when it is at rest. depending on the strength of the damping force. and the time period of vibration (= 2tc/ . 7. A typical example is the motion of a simple pendulum in a highly viscous liquid (like glycerine) where the pendu . 7.coo t] Thus. (31).8 The exponential decrease of amplitude in a damped sim le harmonic motion.7. In general.cvo )t] (41) (42) (38) and we can have two kinds of motion.K2) Equation (37) is similar to Eq.col). and the solution of Eq. Equation (33) can be rewritten in the form d2x + 2 K dt2 where 2K + coo x(t) = 0 (34) dt = m and coo = m (35) Fig.K2 t + 01 (39) B exp [_.(K2 . and Eq. the exact dependence of the damping force on the velocity of the particle is quite complicated.8).K2) can be positive. In either case there are no oscillations and the motion is said to be overdamped or dead beat. Equation (39) represents a damped simple harmonic motion (see Fig.coo t] + On substitution in Eq. however. reaches a maximum and then decreases to zero (see Fig. d and d2x r . K2 is greater than coo. goo is greater than K2. in which the displacement first increases. (37) should be written in the form In order to solve Eq. or the other. In this model. one in which the displacement decreases uniformly to zero. as a first approximation we may assume it to be proportional to the velocity of the particle. dt d where the constant I determines the strength of the damping force. (34) we introduce a new variable (t) which is defined by the following equation: x(t) = fi(t) aKt (36) Thus.9). ) t] + B exp [(_K .8 larger in liquids than in gases. however. (37) would be of the form 4(t) = A cos [/a . x(t) = A exp [(K + JK2 .K2) is greater than in the absence of damping.K2 t + 61 or x(t) = AeKt cos [Jcoo . (34) we get d24 2 (37) 4(t) = 0 dt + (o . I . negative or zero.
Thus. 0 is uniquely determined by noting that sin 0 should be positive. weak damping).0) dt2 =acosin(wt0) (48) Substituting the above forms for x(t).) =JACK` cos iJwo .0) cos 0 . we get a= Further G Uz (51) (52) (44) (53) [(wo . (47) is independent of x. a(co w2)cos(cot ¢)2Ka co sin(cot 0) = G cos (cot .3. (33)]: m d2x = F cos wt F. Eq. w and a are positive.w2. i. (34) is said to be homogeneous. 7. Eq.61 + a cos (cot . w2)=Gcos 0 2 K a co = G sin If we square and add.. (37) becomes =0 dt2 the solution of which is given by =At+B Thus x(t) = (At + B) e Kt (45) The motion is again nonoscillatory and is said to correspond to critical damping. 7.9 d 2x (47) + 2K + coo x(t) = G cos cot dt2 where G = F/m and other symbols have been defined in Sec. Eq.0) sin (50) For Eq. (33). If the frequency of the external force is co then the equation of motion would be [cf. such an equation is said to be an inhomogeneous equation.3) on the motion of a vibrating system.e. dt and 2 t(s) Fig. For the particular solution of Eq.. (42)]. 8. (47). In carrying out the calculations we have assumed K = 2 s1 and /K2 c0 2 =1s1.w2 )2 + 4K 2 co 2Kw (54) 2 (coo . . (50) to be valid for all values of time we must have (49) + 0].ko x t (46) dt2 where the first term on the RHS represents the external force. K2 = wo. Case III (K2 = coo) d2x _ aw2 cos (wt .4 FORCED VIBRATIONS We consider the effect of a periodic sinusoidal force (see also Sec.0) + 0] where we have written G cos cot as G cos [(cot Thus. An equation of the type given by Eq. lum can hardly complete a fraction of the vibration before coming to rest. we obtain aw2 cos (cot 0)2Kacosin (cot 0)+awocos (cot0) = G cos [(cot .. The solid and the dashed curves correspond to B = 0 and B = A/2 respectively [see Eq.e. the other terms are the same as in Eq. the general solution of Eq. (34). To the solution given by Eq. we must add the solution of the homogeneous equation. (48).9 The variation of displacement with time in an overdamped motion. dx/dt and d2x/dt2 in Eq. (43) a(col. (47) we try d x(t) = a cos (cot . (47) will be of the form rr x(. assuming wo to be greater than K2 (i.) Since K. When.G sin (cot . .Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index 7.0) Thus. must be either in the first or in the second quadrant.K2 t .¢) [ (55) tan 0 _ 7. Equation (46) is rewritten in the form* * Notice that the RHS of Eq.
1 Resonance The amplitude of the forced vibration.4. 7. K=0 Optics 0. In the figure. z 4. In order to discuss the phase of the forced vibrations.e. the maximum becomes very sharp and the amplitude falls off rapidly as we go away from the resonance. the phase of forced vibrations is tr/2 ahead of the phase of the driving force. we refer to Eq. i. As the * There is no resonance condition . 7.12. For co = coo. (57). Notice that as the damping decreases.. with increase in damping. G [(cop 0) 2 ) 2 +4 (56) K2 col 1/2 0.10 vz (57) = w0 1 Thus the amplitude is maximum'' when w is given by Eq. the phase also increases and approaches tc (see Fig. 7.co2)2 + 4K2 co2 is a minimum. frequency of the driving force is increased beyond coo.. The second term represents the steady state solution which corresponds to the forced vibrations imposed by the external force.11 The dependence of the phase of the forced vibration on the frequency of the driving force. The maximum value of a is given by G Amax = The variation of amplitude with the frequency of the external driving force for various values of K.10. AC is a metal rod with a movable bob 2 (2K2 ) 2 +4K2 coo 1_ 2K o?o G 2144 K 1/2 J 1/2 G 2K[co 2 +K2 ] ''2 (58) Thus. 7. The variation of the amplitude with co is shown in Fig. The calculations correspond to coo = 5 s1 and the values of K are in sec1. the resonance occurs at a frequency very close to the natural frequency of the system. . When damping is extremely small. This is known as amplitude resonance.10 The first term on the RHS represents the transient solution (corresponding to the natural vibrations of the system) which eventually die out. (54) from where we find that for small damping the phase angle is small unless it is near resonance. the maximum occurs at lower values of co and the resonance becomes less sharper. Notice that the frequency of the forced vibrations is the same as that of the external force. All the salient features of forced vibrations can be easily demonstrated by means of an arrangement shown in Fig.6 0. i.co2)2 + 4K2co2] = 0 4 5 > m(s 1) 6 7 or 2(co(i a)(2w)+8K2co = 0 or 2 K2 00 Fig.11). Notice that with increase in damping.e. the resonance occurs at a smaller value of w. 7. when K2 m(sec 1) ^ Fig.7. when dco [(coo .4 7.2 depends on the frequency of the driving force and is a maximum when (coo . tan 0 = o and 0 is tr/2.
ko the force constant and coo (= . If the bob B is further moved upward.= 0).e. i. In the presence of an electric field. the equation of motion for the electron.e. In particular.e. the centre of the negative charge (due to the electrons) is assumed to be at the centre of the nucleus.. 6 for N2. represents the unit vector in the xdirection and k = 2t/A„ A. if we have a positive charge +q at the origin and a negative charge q at a distance x. etc. With B at the bottom. In this model.5 ORIGIN OF REFRACTIVE INDEX In this section we will study the origin of refractive index..12.r12. w 2 ) (65) is known as the electric susceptibility of the material. in the presence of an external electric field E. when I.qE° cos(kz m wt) (62) where we have replaced the vectors by the corresponding scalar quantities because the displacement and the electric field are in the same direction. Eq. would be m or d2x dt2 d2 dtX In the simplest model of the atom. and the amplitude of the forced vibrations decreases. We know that an atom consists of a'heavy positively charged nucleus surrounded by electrons. In the simplest model of the atom. when the rod AC is set in motion. B and LM is a simple pendulum with a bob at M.w2) (67) * Notice that in the absence of damping (i. This number is 2 for H2. 0 = 0. The metal rod and the simple pendulum are suspended from a string PQ as shown in Fig.. the centre of the negative charge gets displaced from the nucleus which results in a finite value of the dipole moment of the atom. (54). i. the simple pendulum is at the extreme left position.12 An arrangement for demonstration of forced vibrations.. At resonance. (62) is similar to Eq. ** The number of `dispersionelectrons' in a molecule of an ideal gas is the valence number of the molecules. thus. then the dipole moment would be qx. the pendulum LM also vibrates. 7.qE m (60) £ EO =1+ Nq ins() (w0 . a restoring force (proportional to the displacement) will act on the electrons which will tend to return the electrons to their rest positions. the electrons are assumed to be bound elastically to their rest positions. dipole moment per unit volume) would be given by 2 P = Nq x = Nq E m(wo 0 2 ) (64) =xE where Nq2 x= m(w (. see Eq. m and. when the metal rod is at its lowest position and moving towards right. thus. the time period decreases and the frequency of the rod becomes closer to the natural frequency of the simple pendulum and eventually the resonance condition is satisfied. As the bob B is moved upwards. .6 x 1019 C). q represent the mass and charge of the electron (q = + 1. (46) and therefore. the frequency increases. We assume E = Eo cos (kz . if there are N dispersionelectrons** per unit volume then the polarization (i.w2) wt) (63) 7.Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index 7.e. the amplitude of vibration of the simple pendulum is maximum and the phase difference between the vibrations is nearly . the field is in the xdirection having an amplitude E0 and propagating in the + z direction. representing the wavelength. Except for the damping term. The dielectric permittivity is therefore given by (see Chapter 23) E=E+ + k0 x = qE (59) or (66) 2 2 +wo x= . when these electrons are displaced by an electric field.wt) (61) C Fig. 7. jko /m) represents the frequency of the oscillator. the solution corresponding to the forced vibrations will be given by* x=q E° cos (kz m(coo . Thus d 2 x dt2 + coox = .11 where x represents the position of the electron.
1x10 4x10'm2 31 q mso wz o +2 wo Nq2 471. the (75) 3x10 10 m2 47(2 c2 eom 2 1 is positive in the entire visible region.6 eV from which one obtains coo = 2 x 101 6 s l. if we beam Xrays on the matter. the refractive index increases with frequency.2 0 which is the wellknown `Cauchy relation'. N q2 471.e. coo = 0 (there is no elastic restoring force).7.c 2 x9x1016 x8. this does not imply that one can send signals faster than the speed of light in free space (see Chapter 10).721 x 104 + 2. 2.85x1012 x9. 2 c2 Thus 2 2. and as an example if we consider hydrogen. n2 also increases. 9].wo [ and n2=1+ N 2 thus. For hydrogen.721 x 104 (72) mso wo * This also follows from the fact that according to classical electrodynamics. . Hence. This frequency corresponds to the far ultraviolet.6x10 19 )2 4x. To quote Feynman: For free electrons. where N is now to represent the density of free electrons (number per unit volume) in the stratosphere. (67)] z n2 =1.11x10. or radiowaves (or any electric (71) where the wavelength is measured in metres. Equation (75) shows that the refractive index is less than unity. then ht coo = 13. s/s0 is the dielectric constant. (75). the above numbers correspond to 0° C and 76 cm of Hg [see Ref. Thus.. as co increases.11 x 10 1s 2. Nq and 47(2 c2 Nq2 mso coo =2.0 is approximately given by the following relation: n2 = 1 + 2.62 Now at NTP.12 Optics If we divide the second equation by the first. Assuming that the characteristic Nq2 frequency coo lies in the far ultraviolet [see Eq. It is of interest to mention that for a gas of free electrons (as we have in the upper atmosphere) there is no restoring force and we must set coo = 0. quantity 1. Thus the expression for the refractive index becomes [see Eq.2 c2 eo m 5x1025 x(1. then [ .18 m2 (73) 2 = 2. the experimental variation of n2 with X1.11 x 1018 471.721 x 104 2 1 co 12 (68) or n2 = 1 + (74) 3x10 15 s1 vo= 2^ Imo ^o coo which is indeed in the ultraviolet region. however. we would get Now. (74)]*.2 0 which qualitatively agrees with Eq. which is equal to the square of the refractive index (see Chapter 23).2 c2 Ng 2 1 (69) mso 4 ''1 + mso 2 coo ) coo + where A0 = 27rc/co is the free space wavelength. coo J Further. i. But let us look again at the equation. this is known as normal dispersion. N=2x 6x10 23 m3 =5x1025 m3 22400 x lo' 1 + 2 wo where the factor 2 arises from the fact that a hydrogen molecule consists of two electrons. 22400 cc of H2 contains 6 x 1023 molecules.:. an oscillating dipole vibrating with requency coo will radiate electromagnetic waves with frequency coo. If we further assume co/coo << 1. coo from Eqs (72) and (73) to obtain this is known as dispersion. One can eliminate showing that the refractive index depends on the frequency. Equation (69) can be written in the form n2 =A (70) + A.Nq (76) mso co2 where N represents the density of free electrons. Setting coo = 0 in our dispersion equation yields the correct formula for the index of refraction for radiowaves in the stratosphere.
Wood discovered that alkali metals are transparent to ultraviolet light. The theoretical and experimental values of AP for Li. the refractive index tends to 00. Indeed in 1933. for sodium if we assume that the refractive index is primarily due to the free electrons and that there is one free electron per atom then 6 x 1023 x0. Eq.854 x 1012 C/Nm 2 we would get A. In spite of the fact that it is said that you cannot send signals any faster than the speed of light. the term (co. then the wave number k. (80). (68). (46)] 2 (79) and x +I'dr +kox=gEocos(kzcot) dtt In order to derive an expression for the refractive index. . This is due to the fact that we have neglected the presence of damping forces in our treatment. K and Rb are discussed in Problem 7. which equals nco/c. Thus. q = 1. it is more convenient to rewrite the above equation in the form dx d2x z q Eo f(kz . . Thus n2= s w t 2iKco] l = 2098 A Eo = 1 + Eo N q2 (84) Thus for A. (_ 2^cc cop 2iKw + coo)A = qE° m q E0 m[wo co t 2iKco] or* A = (82) Thus we get N q2 P= m[co co t E 2iKco] (83) The electric susceptibility would therefore be given by N q2 mpg. 7. (76) gives the correct dependence of the refractive index of the stratosphere for radiowaves. the refractive index of Na becomes real and the metal would become transparent. which implies absorption of the propagating electromagnetic wave. (62) would modify to [see Eq. If =1+ mEO [wp w 2 2iKco] Notice that the refractive index is complex.(85) _ where T1 and K are real numbers. Indeed. 7. 2535x10 22 cm 3 N= 22. < 2098 A. As mentioned above. the corresponding experimental value is 2100 A. if we substitute a solution of the type x(t) = A e'(kz .7. in Sec.9712 g/cm 3.cut) (78) in Eq. Eq. Equation (76) is usually written in the form n 2 =1where cop = P we do take into account the damping forces. however.13 waves) on free electrons. we note that as co * coo. (80). we would obtain (cot  (77) co (81) is known as the plasma frequency. For example. Substituting the values of m = 9.Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index 7. 3. it is nevertheless true that the index of refraction of materials at a particular frequency can be either greater or less than 1. if we write  n=7j+ix  .wt) + e (80) + co° x 2K dt = in dt2 where the solution of Eq. the steady state solution will correspond to frequency co. and we obtain the result that n is less that one. Returning to Eq.9712 0.99 and its density is 0. (81) and take the real part we would get the same expression for x(t) as we had obtained in Sec. the refractive index is real.4.109 x 1031 kg. (82) in Eq.3 we had used Eq. if we substitute the expression for A from Eq. The solution of the homogeneous equation will give the transient behaviour which will die out as t + 00 (see Sec. That means that the effective speed of the waves in the substance is faster than c! Can that be correct? It is correct. the refractive index is purely imaginary which gives rise to attenuation and for co > cop .602 x 10 19 C and eo = 8.4. Notice that for co < cop.99 where we have assumed that the atomic weight of Na is 22. (76) to study reflection of electromagnetic waves by the ionosphere. would be given by k = (rl ± i c (86) * Notice that A is complex.4).w2) become negative. (79) will be the real part of the solution of Eq.
. (92) describes correctly the variation of refractive index for most gases. the function (rig .zz I  0 zj K (87) which shows an exponential attenuation of the amplitude. 7. (84) would get modified to the following expression:* n2=1+ Ng2 me0 fJ [wj .wt^ J = E0 exp [i CO (t .. This is indeed borne out by the data shown in Fig. . for example. the molecules are very close to one another and the dipoles interact between themselves. Figure 7. Ref.14 shows the dependence of the refractive index of sodium vapour around 2i = 5800 A.13 Qualitative variation of (T1 2  K2) and 2rpc with 0. In order to obtain expressions for r) and K. Eq. Notice that when n is very close to unity (i. It should be pointed out that.. =1+ or rj2 and Ng2 2riK = ms° [(o 2Kco 0) IC 2. Since Dl and D2 lines occur at 5890 A and 5896 A. an atom can exn21 _ Nq2 ' ff ecute oscillations corresponding to different resonant 2 3ms 0 w2 w2 n2 + where we have neglected the presence of damping forces which is justified except when one is very close to the resonance.w2 .vl v 2v2 (93) where we have introduced the following dimensionless parameters: 2_2 2K N g22 and /3 = S2 = a Zw° w° wo mEO w° The qualitative variations of T12 . in general.0 z Fig. It may be worthwhile mentioning that in a liquid. Nq2 (w (i . Ref.x2 and 277K with S2 are shown in Fig.14 If we consider a plane electromagnetic wave propagating in the +z direction. 1. for example.K2) attains its maximum and minimum values respectively. we substitute the expression for n from Eq. Eq. 6).w 2 +2i Kw) me ° (wo w 2 2iKw) (wow 2 +2iKw) Ng 2 (wo w 2) + mE 0[( w o w 2 ) 2 +4K2 W 2 ] = (88) frequencies and we have to take into account the various contributions. consequently E = E0 e`(kz Out) Optics 2 ?pc = E0 exp[i{(ri+iK) z .[S2 2 + /32 (S2 + 1)l a/3 ji+52 [S22 + /32 (f + 1)l (91) where KJ represents the damping constant corresponding to the resonant frequency cop Indeed. (85) in Eq. for a dilute fluid).wt)].(92).0 0 E 0102_41 a) 0 2. The variation of the refractive index can be accurately fitted with the formula n2 d 1+ A 2+ B2 2 v . This should not be unexpected because damping causes a loss of energy. ** See.e.13. col . It can be easily shown that at S2 = /3 and at S2 = + f3.14. we would get** (94) * Quantum mechanics also gives a similar result (see. one should expect resonant oscillations around these frequencies.7. 7. (84) to obtain (rf+iK)2 .2i K co] (92) 2 ) 2 +4K2 0) 2 ] an (89) The above equations can be rewritten in the form x2 (90) and 2r7K= = 1 .. If we take this interaction into account. (94) reduces to Eq. then its z and t dependence would be of the form exp [i(kz . . represent the resonant frequencies and represents the fractional number of electrons per unit volume whose resonant frequency is cop Eq. If w°. 7.
Eq. Indeed. an oscillating dipole given by P = Po e (coo . H2O) one has to carry out a different analysis. (94) gives a fairly accurate description. then y becomes proportional to coo or y 1 « A4 (103). To keep the analysis simple.g. However. We assume the electromagnetic wave to be propagating along the xdirection. The measurements are of Roschdestwensky.g. for the hydrogen atom h coo = few electron volts. As discussed in Sec. 02.(o z ) radiates energy at a rate (see Sec. the figure has been adapted from Ref.14 The measured variation of refractive index of sodium with frequency around the Dl and D2 lines.1)2 1275£ 0 c3 mz (coo . 1. (100) where _ N cn4 4 (101) y 67r£oc 4 m2 (coo. H2. 37c N k= cc (106) . Now. if the colour of the setting (or rising) sun is deep red. etc.m2 ) The integration of Eq.6 RAYLEIGH SCATTERING We end this chapter by giving a brief account of Rayleigh scattering. 7.1) 2 .1 = m£0 where coo represents the natural frequency of the atom. Now. 7. liquids whose molecules posses permanent dipole moments (e. Thus if we assume co << coo. for example. 0  wz) E (95) which represents the famous 1/1. venient form 4 127r £0 c3 y 375N 4 (c' ) . where we have neglected the presence of damping. then the total energy radiated away (per unit volume) would be NP.4.. the blue component of the light coming from the setting sun is predominantly scattered out resulting in the red colour of the setting sun. the incident electric field E produces a dipole moment given by [see Eqs (64) and (65)] qz P m(c. For most atoms coo lies in the ultraviolet region. z nz . (78) of Chapter 23] I = 2 £0 cEo z (99) Thus the change in the intensity of the electromagnetic wave as it propagates through a distance dx is given by dl = NP dx or Fig.) Eq. For air.co z ) z q4 Eo (98) _ 2k4 (n .Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index 7.1) P or P = = 04 Po using which. we are neglecting the .wz ) Nq (104) [See Eq.5. (101) can be written in the following con(9'7) . (68)].15 D2 Thus if N represents the number of atoms per unit volume. 7. since the refractive index is very close to unity. For liquids. (100) is simple: ='o eYX (102) implying that y represents the attenuation coefficient. one can infer that the pollution level is high. effect of damping although it can be taken into account without much difficulty.4 Rayleigh scattering law and is responsible for the blue colour of the sky (because it is the blue component which is predominantly scattered). we may write Nq2 n1 = imt (96) (105) 2m£0 (coo . Similarly. Throughout our analysis we will assume that each scattering center behaves independentlyan assumption which will be valid for a gas where the average interatomic spacing is greater than the wavelength. The intensity of the wave is given by [see Eq. for a gas. whose molecules do not have a permanent dipole moment (e. 23.
7 x 10 19 molecules/cm3 we obtain L = = 27 km. Assume a = 0.0 cm. l R .] 7. 3 cm and 4 cm. wavelength and the frequency of the wave.0. study the time variation of displacement in each loop and show that alternate loops vibrate in phase (With different points in a loop having different amplitudes) and adjacent loops vibrate out of phase. The quantity L represents the distance in which the intensity decreases by a factor of e.15 For Problem 7. It is a matter of common observation that the blue of the sky is highly variable. It executes simple harmonic motion with period 0. t) = y(x = 1. A and v represent the amplitude. The electric field of the lightwave polarizes the molecules of the gas.z27cvt C27r J . We may conclude this chapter by mentioning that in the 1929 edition of Encylopaedia Britannica. 13.. 0. Interpret the plots physically. we have what is known as forced vibrations. (see Sec. the lightwave gets attenuated. producing oscillating dipole moments from which one can make a first principle calculation of the refractive index to obtain n2(co)=1+ Nq` map (alp . Plot the time dependence of the displacement at x = 0. This takes place in a plane passing through the sun.and y. • When an external sinusoidal force is applied to a vibrating system. A mass is dropped at the point A along the tunnel.1 sec.5 cm. A = 4 cm.. q the magnitude of the charge of the electron. t) = 2a sin ♦ The most fundamental vibration associated with wave motion is the simple harmonic motion. t) because the two points are 2 apart.15. 2iKcv) where m is the mass of the electron. .25 cm] 7.1 = 2.. ♦ When a lightwave interacts with an atom. In steady state. and attains a maximum about 90° therefrom.[OZ  [Ans: The time period will be T = 27c.. The colour usually deepens toward the zenith and also with the elevation of the observer. What will the time period be? y(x.78 x 104 in the entire region of the visible spectrum.when the mass was attached? [Ans: Ox = 0. 2L13. we may assume the electrons to behave like oscillators with resonant frequency coo.16 For air atNTP. 5900 A (yellow) and 6500 A (red) respectively. even on days that are free from clouds. N is the number of electrons per unit volume and K is the damping constant.t)=acos A .1 cm. the frequency of the forced vibrations is the same as that of the external force.. Show that it will execute simple harmonic motion.. the quantity n . Lord Rayliegh wrote in an article on SKY: SKY: The apparent covering of the atmosphere. Closely associated with the colour is the polarization of light from the sky.directions such that x(z. Summary Problems Optics 7. Because of the fact that an oscillating dipole radiates energy.0.7. 1. 7.3. 7.] Vg Fig. t) = a cos ^ x27cvt C where a. the foot of the perpendicular on any one of its diameters will execute simple harmonic motion. v sec 1. Consider the case when A= 2L/5. this leads to the fanhous 4 Rayleigh scattering .5 cm. ♦ When a point rotates on the circumference of a circle with a uniform angular velocity. the overarching heaven.2 The displacement associated with a standing wave on a sonometer is given by the following equation: x cos 2 7rvt C2^ If the length of the string is L then the allowed values of A are 2L. 7. etc. [Ans: y(x = 3.1 The displacement in a string is given by the following equa tion: 2tt 1 y(x. By how much distance had the spring stretched . 128 km and 188 km 7 for A = 4000 A (violet).5 A stretched string is given simultaneous displacement in the x. 1. With N = 2.2). 2L12.law which is responsible for the red colour of the rising sun and blue colour of the sky. 2 cm.3 A tunnel is dug through the earth as shown in Fig. 7.4 A 1 g mass is suspended from a vertical spring.
Simple Harmonic Motion Forced Vibrations and Origin of Refractive Index 7. For co = 108 CI calculate the complex dielectric constant and compare its value with the one obtained for infrared frequencies. Assuming that the refractive index is primarily due to the free electrons and that there is one free electron per atom. 7. for Xray frequencies all the electrons may be assumed to be free (see Problems 7.12). 2884 A and 3214 A.12x10 15 . Discuss the validity of the above argument.09+ 6.12 In an ionic crystal (like NaCl. 3150 A and 3400 A respectively]. On the other hand. alkali metals are transparent to ultraviolet light. only one of the electrons of a copper atom can be considered to be free. It may be noted that for small frequencies. the corresponding experimental values are 1551 A. Cl.6 In Problem 7.8) is essentially real with frequency dependence of the form x=1where co = Nq mE0 2 w w2 2 X 1/2 is known as the plasma frequency.27c v t what will be the resultant displacement? 7. show that P = Nqr= N q2 2 E m(w2 +icov) 6(w)= where M represents the reduced mass of the two ions and p represents the valency of the ion (p = 1 for Na+. Assuming that at such frequencies all the electrons can be assumed to be free. mass number is 63.10 Show that for high frequencies (w» v) the dielectric constant (as derived in Problem 7.26x109 ^2 8.2 2 2 A1 Ai A2 12 A2 where = col A1 27rc which represents the polarization.532 g/cm3. [Ans: 9 x 10 16 sec I] 7. calculate wi. t) =acos ^ z27cvt C2^ / Show that the string will vibrate along a direction making an angle 7r/4 with the x and y axes.5. p = 2 for Ca'.z27rvt and y(z.13 The refractive index variation for CaF2 (in the visible region of the spectrum) can be written in the form* 15 + 5.854 x 1012 C/Nm2.7 As mentioned in Sec. K and Rb. all the electrons can be assumed to be free. Show that the above equation can be written in the form* n =n°°+ A. [Ans: 1550 A. The drift velocity of the electron satisfies the following equation m J for Li. 7. the values of various physical constants are: m = 9.10.2 c2 e0 M  which represents the dielectric constant variation for a freeelectron gas.870 and 1.94.^2 1.5. show that v = 4'x 10 13 s l.11 For sodium.10x109 n2=6. one has to take into account infrared resonance oscillations of the ions and = Eq. F2).9 Assuming that each atom of copper contributes one free electron and that the low frequency conductivity o.10 and 85. 7. if (27r x(z. You may assume that the atomic weights of Li. Also. t) = a sin (27c z . and density is 9 g/cm 3 . 0.534. (a) In a metal. K and Rb are 6.109 x 1031 kg. measurements are of Paschen. 39.is about * Quoted from Ref.17 and y(z. Using this value of v. t) =acos . at 2 = 1 A. the electrons can be assumed to be essentially free.). 9.48 respectively and that the corresponding densities are 0. show that the conductivity is almost real for w< 1011 s l. dt + mvv = F = q E0 e The above dielectric constant variation is indeed valid for Xray wavelengths in many metals. etc.11 and 7. 7. q = 1. CaF 2.602 x 1019 C and Eo 8. 7.88x10 . 7. under this assumption show that wp = 3 x 1016 s1 and n2 = 1 and the metal will be completely transparent. 7. for copper for which the atomic number is 29. calculate 2 = 2 7r c wp 6 x 107 mhos/metre. (68) modifies to n2=1+ Nq2 pN 2 q 2 2+Me 0 (02 co) meo(co w) where v represents the collision frequency. Using the above equation show that x(w)=1mE0(w2 +iwv) N '12 _ 27r c w2 4 1 Nq2 AZ__ 47x2 c2 e0 m pNg2 47. Calculate the steady state current density (J = Nqv) and show that the conductivity is given by N q2 1 m viw (b) If r represents the displacement of the electron.
R.J. 1965. 1973.7. Crawford. A2. On the other hand. McGrawHill Publishing Co. J. Waves and Oscillations: Berkeley Physics Course. New York. N =10 12 electrons/m3 and that the electron density increases to 2 x 10 12 electrons/ m3 at x = 300 km.. Vol. R. and A2 show that m/M = 2. 1973. B. Academic Press. 7. (b) From the values of Al. John Wiley & Sons. R.S.J. Halliday. John Wiley & Sons. .6. the electron density decreases. plot the corresponding refractive index variation. New York. Part I.14 (a) The refractive index of a plasma (neglecting collisions) is approximately given by (see Sec. Leighton and M..M. the maximum value of No is 10 10 10 12 electrons/m3. n 2 (x) =. for low frequencies. [Ans: For 2 x 105 m<x<4x 105 m. (c) Show that using the constants AI. 1968. In the ionosphere. This fact is used in long distance radio communications (see Fig. 1968. the refractive index is imaginary (like in a conductorsee Sec. I.18 where A is in metres (a) Plot the variation of n2 with A in the visible region. Pain. H. Braddick. 1952.20). Theory of Electric Polarization.2 5.] REFERENCES AND SUGGESTED READINGS 1.P. Feynman. Al and A2 we obtain n. 2. Resnick and D. Oxford. Optics. R. Bottcher. 5.. London.J. 8.6) 0)2 Optics Calculate the plasma frequency. Vol..F. The Quantum Theory of Light.73 which agrees reasonably well with the experimental value given above. Loudon. 1964.3) and the beam gets reflected. McGrawHill Book Co. New York. (b) Assume that for x = 200 km.. Sands. w where co is measured in CI and x in m. thus high frequency waves (like the one used in TV) are not reflected by the ionosphere..1 . Radiation and Optics. ArnoldHeineman India. AddisonWesley Publishing Co. 7.P. Assuming a parabolic variation of N. Vibrations and Waves. Vibrations. H. 9. London. Mass. 3. The Feynman Lectures on Physics. French. Clarendon Press. 4. 1965.07 x 105 and compare this with the exact value. For x < 300 km. Elsevier Publishing Co. Sommerfeld. Physics. 6. New Delhi. C. Notice that at high frequencies n2 1. 1966. III. A. 3. Reading. 1963. Stone. A. 7. The Physics of Vibrations and Waves.J. Amsterdam.414 N1/2 s1 is known as the plasma frequency. 10.4x2015 [15x 10"(x3x 105)2 n2 = 1 co Nq mE0 2 )1/2 where co 56. 24. McGrawHill Book Co. F. New York. Waves and Diffraction.
..A. for ni ^ n to+T f cosnwt cosmwt dt to to +T = 1 f [cos (n . can be expanded in the form: f(t) = 2 ao+cos( n =1 The coefficients an and b.. f(t + nT) =f(t). any periodic vibration can be expressed as a sum of the sine and cosine functions whose frequencies increase in the ratio of natural numbers.FOURIER SERIES AND APPLICATIONS Reimann (in one of his publications in 1867) asserts that when Fourier. Now. Actually. his statement so surprised Lagrange that he denied possibility in the most definite terms. (b) piecewise continuous (i. in the interval to < t < to + T) must be (a) single valued. to +T to +T (2) = 2 ao + i an cos (n co t) + l bn sin (not) n=1 n=1 where co _ 27r T (3) f cos ncot cos mcotdt to f cost n wt dt to to +T = Z f [1+cos 2nwt]dt to =2 represents the fundamental frequency. can easily be determined by using the following properties of the trigonometric functions: to+T f cos nwt cos moot dt = to to+T {0 T/2 if if min m=n min m=n (4) f sin n cot sin man dt to to +T if ^0 T/2 if (5) ± 2. we devote this chapter to the study of Fourier series and Fourier integrals. It should also be noted that he (Fourier) was the first to allow that the arbitrary function might be given by different analytical expressions in different parts of the interval. As such.e.. a periodic function with period T. ..1 INTRODUCTION Fourier series and Fourier integrals are extensively used in the theory of vibrations and waves. Similarly. the function f(t) _must satisfy certain conditions. (1) f sin ncot cos mwtdt =0 to (6) 2 Tn tJ+sb„sin(2n n=1 T t) The above equations can easily be derived. for rn = n. in his first paper to the Paris Academy in 1807. n = 0. The results obtained will be used in subsequent chapters. For example.. stated that a completely arbitrary function could be expressed in such a series. H.e. These conditions are known as Dirichlet's conditions and are almost always satisfied in all problems that one encounters in physics. ± 1. The conditions are that the function f (t) in one period (i. according to Fourier's theorem.m) wt + cos (n + m) wt] dt 2 to . for the expansion to_ be possible. it can have at most a finite number of finite discontinuities) and (c) can have only a finite number of maxima and minima.e.. Thus. Carslaw (1930) 8. i.
if we multiply Eq. (9) t _. It should be pointed out that the value of to is quite arbitrary.. since f(t) is an odd function of t. we first multiply Eq. 1..T/2 +T/2 T f . f(t)=t where we have used Eqs (4) and (6) for m = 0. = 0 and f(t + 2nT) = f(t) b„ = 7.2.. .. f(t) = f(t)) or odd (i. out the integration we obtain = mr'T. In this example. 3.. (2) by dt and integrate from to to to + T: to+T f f (t) dt = to 00 = f(t) sin ncot dt. 8.e. n = 0... Now.. 8. S2 and S3 represent the partial sums corresponding to the saw tooth function.. 2. 3..T/2 f(t)cosncotdt.8. We may combine the above equation with Eq. 1.m) Cot 2 (n . In some problems it is convenient to choose to = T/2 . (7) to write an = = to +T T f to f (t) cos ncot dt . we will expand the above function in a Fourier series. 2 ao n=1 to+T f dt + ^ ctn to to+T to to+T n=1 to+T to f cos ncot dt + Eb„ f sin ncot dt = T ao (7) Such a choice is particularly convenient when the function is even (i. In some problems. Such a function is referred to as a saw tooth function. (8) Similarly.1 The saw tooth function.tn) co + Optics 1 then +T/2 an= T f . 2. In the former case b„ = 0 whereas in the latter case an = 0. f(t) = f (t)). (4) and (6). n = 1.1). to +T T f to  bn f (t) sin nco t dt ... . Carrying • E n=1 bn f cos (man) sin (ncot) dt to t cos nwt + 1 sin nwt nw nco {nw } cos nit = (1)n+1 1 1 = n)r nn T = 2 am where we have used Eqs. therefore. Thus ao = 2 f f(t) dt to for T<t<+x Next. S1..2 1 sin (n . 2 Fig.. n=0. 2. Example 8.. n = 0. . a.f f(t) sin (ncot) dt 2. (n +m) (1 co sin (n + m) wt l to =0 and bn In order to determine the coefficients an and b. (2) by cos (moot) dt and integrate from to to to + T we would obtain to+T f f (t) cos (man) dt to • E an n=1 = to+T 2 ao 'cos (man) dt + to = b„ 2 a2 f tsin(ncot)dt co to +T f cos (man) cos (n cot) dt + to to+T 0 Notice that the periodicity is 2t and.e.1.. it is convenient to choose to = 0.1 Consider a periodic function of the form (10) 1p (see Fig.
. Let us consider a stretched string. 11.. 8.3.6 that the displacement y(x. the corresponding shape of the string is shown as dashed line in Fig.(1)1 sin nwt d (14) (L . B . 8. Once again the function is an odd function.cos n7r] = 2A ntr [1 (1) n ] Thus 2A f(t) =  n=1.x) for a<x<L La where L represents the length of the string.3 The plucked string. d A L Fig. S1. 8.a„ = 0 and b„ = T 2 SA sin (nwt) dt = 4T nw [cos nwt]ol2 0 = T/2 n7r 2A [1.v2 at2 = 4 sin wt. Example 8. We will show in Sec. the equation of the string (in its displaced position) would be given by the following equation: for 0<x<a S3 = i sin wt L 2 sin 2wt+ 3 sin 3wt It can be seen from the figure that as n increases...x in study .. T being the tension in the string and p.3 Thus f(t) = ^ (_ lnn+1 n=1.. we will Fourier expand the function defined by the following equations: f(t) =A for  2 <t<0 =+A for 0 < t < t and f(t + 7) = f(t) +2 (13) The function is plotted in Fig. we would like to determine the shape of the string at any subsequent time. the mass per y 8. One of the ends (A) is chosen as the origin. Sz and S3 represent the corresponding partial sums. by S1 = !z sin wt.] 1 a2y (15) ax2 . consequently. A point of the string is moved upwards by a distance d.1 we have also plotted the partial sums which are given sin wt  2 sin 2cot Fig. S2 = z (12) In Fig. The dashed lines show the displaced position at t = 0. t) satisfies the following wave equation: a2y = 4 [sin wt+ The partial sums S1 3 sin 3wt+ 5 sin 5wt+. 8. the sum S„ approaches the function f(t).2. 8. waves. 1 n [1. 8. fixed at the two ends A and B. (13). sin5on] are also plotted in Fig. In the equilibrium position of the string. if the string is released from this position at t = 0. AB represents the equilibrium position. If the displacement occurs at a distance a from the origin.2 A plot of the periodic step function defined by Eq..2....2 TRANSVERSE VIBRATIONS OF A PLUCKED STRING An interesting application of the Fourier series lies ing the transverse vibrations of a plucked string. 8.Fourier Series and Applications 8. Now. sin ncOt ^ = 2a r 1 1 sin cot .2 sin 2wt + 3 sin 3wt . it is assumed to lie along the xaxis (see Fig.. 7r S2 = 4 (sinwt+ 7r r 2 sin 3wt\l where v (= Tlp) represents the speed of the transverse S3 = 4A [sinOt+sin3wt+ .3.2.2.2 In this example.3).
We would like to solve Eq. (22) _ w2 v2 Thus.8. Thus the solution of Eq.. Since y (x. n= 1. t = 0) =a d x for for 0<x<a a<x<L (18) = La (Lx) we must have sin kL = 0 Assuming a time dependence of the form cos cot (or sin wt): y(X. (15) satisfying the boundary condition given by Eq. these are given by X (x) n1c kn= L giving n=1. (17) where we have absorbed A in C and D.. n = 1. . t) = (A sin kx + B cos kx) (C cos cot + D sin cot) (20) Equation (24) gives the frequencies of the various modes of the string..2. (24) where k= 10v The solution of Eq. Thus d2T +w 2 T(t) = 0 dt2 and d2 X + k2X(x) = 0 dx2 where w=kv . thus we assume y(x.2. Substituting in Eq. we get 1 d2X 1 1 d2T _ k2 X(x) dx2 __ v2 T(t) dt2 2 Since the term 1 d2X is a function of x alone and the term X X dx2 1 2T is a function of t alone. = nLv . 2. t) Ix = o = 0 for all values of t Thus B = 0 and we obtain y(x. (25) * Rigorously we should proceed by using the method of separation of variables. The mode corresponding to the lowest frequency (n = 1) is known as the fundamental mode.. only discrete values of k (and hence of co) are permissible. (16) would be given by y(x. (15). d2X dx2 + k 2X(x) = 0 (19) co. (23) or.. 3..4 unit length.. t) (21) = E sin knx [Cn cos wnt + Dn sin wnt] n=1. t) = sin kx [C cos wt + D sin wt] (16) for all values of x. each term must be equal to a constant v2 1 T d dt2 which we have put equal to k2.2. t) = X(x) cos wt we obtain d2X dx2 or kL = nit.. t) Ix = L = 0 (for all values of t) (ii) y(x. (15) subject to the following boundary conditions: (a) y = 0 at x = 0 and x = L for all values oft (b) At t = 0 =0 (i) at Optics Now y (x.3. t) = X(x) T(t) where X(x) is a function of x alone and T(t) is a function of t alone.. (19) is simple*: X(x) = A sin kx + B cos kx Thus y(x...
. L the relation 0 = if min m= n (30) f sin nlrx sin m7rx L L 0 L/2 if [cf. 7.. n = 1.e. are absent (i.Fourier Series and Applications 8. t) = for all n 8d 7t ^(1) m+i m (2m1)7rx 1 x L (2m1)2 sin cos (2m 1) (7r v t) L (32) Cn sin knx cos coat n=1. (28) by sin L xJ dx and integrate from 0 to L to obtain: Cm L u:. 6. 2.x) sin a al L xl dx F(t) = F„ sinncot n 2ai 7r (_I) n+l sinncot n=1.2.n sin rntc 1 L aJ x COn Dn COS Co nt] t=0 = Ico n Dn sin knx n Since ay at we must have Dn = 0 Thus y(x.. if F(t) is periodic then we can apply Fourier's theorem to obtain a solution of Eq. (33). we multiply both sides of Eq. (33) is a simple harmonic motion with the frequency of the external force.3 APPLICATION OF FOURIER SERIES IN FORCED VIBRATIONS Let us consider the forced vibrations of a damped oscillator. Substituting the expression for y(x.5 On substituting in Eq. we obtain Cn = L a f (L xsin o 2dL2 x dx + ) L La f (L .. the even harmonics are absent) and Eq. (31) simplifies to C L L y(x. terms corresponding to n = 2. (27).2. 7.1 and is of the form: (35) (28) The above equation is essentially a Fourier series and in order to determine Cn. 0) sin 0 CL 1 x l dx (29) where we havL. 0) from Eq. (33) is difficult to obtain... For example.3) and F represents the external force. .. t) 2dL2 = 2 Differentiating partially with respect to t.2. If F(t) is not a sine or cosine function. (18).t it fy L (x. t) = or y(x....e. a = L12).4 that if F > 0 and F(t) = Fo cos (pt + 9) (34) then the steady state solution of Eq..3. . 0) = n (26) =0 t=0 for all values of x cos (31) (n v tJ xJ Equation (31) can be used to determine the shape of the string at an arbitrary time t.. we get ay at = Isin knx [con Cn sin cons + t=0 n a(La)TL E 1a sin n=1. we finally obtain y(x.. n (36) L a(La)7c 2 n2 sin (n .. The equation of motion would be m d22 +I d+ dt =F(t) (33) dt2 where F represents the damping constant (see Sec. If the string is plucked at the centre (i. (27) Cn sin n7z Lx 8.3. let F(t) = a t for 'r < t < 'r and F(t + 2na) = F(t). The Fourier expansion of such a function was discussed in Example 8. a general solution of Eq. It has been shown in Sec. however.. Eq. 4. (5)].
1 we had shown that expanded in the form f(t) where dt2 + T' dyt + /ia y„ = F„ sin nwt a periodic function can be = Z ao+d [an cosnwt +b..t')] dt' (48) Thus the steady state solution can be written in the form y= G. +T = T f f (t) cosnwt dt (44) ra An __ Fn m (1) 0+1 2 a'r n 7rm (38) b„ and 27r (46) T= co On substituting the above expressions for a„ and b. in Eq. sin (nwt + On) (41) n=1 toles n where the amplitude G... cos nmt and the solution of Eq.Dn sin ncot] +wo [Cn sin not + Dn cos ncot] = A. notice that when T * function is no more periodic. (33) will be of the form Y = Eyn (39) Dn n In order to determine C.. and solution in Eq. (44) and (45)]: f(t) tp+T = T f f (t) sin ncot dt to (45) The steady state solution of Eq. (43) we get [we must replace t by t' in Eqs.. 8.4 THE FOURIER INTEGRAL In Sec.6 We next consider the solution of the differential equation m or d2 y.8.. is given by 1/2 Gn = (C2 + I)! ) n D2 [(w As = T = 0) co. sin nwt + D. (37) will be of the form yn = C'. sinnco t] n=1 (43) g=m. We let T + oo so that As a 0. nciK (a . (37) to obtain we substitute the above n2 w2 [Cn sin nwt + Dn cos nwt] + nwK [Cn cos nwt .n2 w2 )2 + n2 0)2 +x1& K2 An f(t) = 2_ As f f (t') dt' + tries w2 n2 co' (coo ..n2 w2 )2 +n2 w2 K2 An where +toles 1 °s f f (t') cos [n As (t ... Thus if the integral (42) f jf(t')ldt' the Z . we get Dn = and Cn . and wo=_o an t.n2 w2 )2 + n2 w2 K2 ] 1/2 An . sin nwt Thus and (cog n2 w2 ) Cn nwKDn =An (o n2 w2 )Dn ncoKCn =0 (40) = +T/2 T f f(t') dt' + T/2 1[T n=1 +T/2 cos mot f f(t ') cos nwt ' dt' T/2 +T/2 + or T sin nwt f f(t ') sin nwt' dt' T12 (47) Solving the above equations. + K do wgyn An sin nwt = dt2 where (37) Optics 8.
Thus. . Further. we may write +00 f(t) = 27c f f f (t') cos [s(t . Equations (54) and (55) are also written in the form F(w) with +00 to +T b„ = T f f(t) sinnwt. if F(co) = then f(t) 1 27c = 2ao+ Ian cos(nwt)+lbn sin(ncot) n=1 n=1 where 27r co= T f f f(t) e ±itt dt (54) represents the fundamental frequency. (48) would go to zero.0 0 8. Summary .. n = 0.e.1. ±2.Fourier Series and Applications exists (i. 27c f f(t) etwr dt . In the next chapter we will introduce the Dirac delta function and rederive Eqs (54)(59) and work out a few examples to illustrate the physics and applications of the Fourier transform. f(t + nT) =f(t). . ♦ For a time dependent function f(t).0 27r If we add get (or f f (t ') sin [s(t . In Chapter 10. can be expanded in the form f(t) = an + lan cos( 7 1tJJJl+I bn sinl rtJ l n=1 subtract) the above two equations. ±1. Equation (53) is usually referred to as the Fourier integral theorem.3.1. F(co) is usually referred to as its frequency spectrum.t ')] dt' ds = 0 (52) ♦ A periodic function with period T. to n= 0. if it has a finite value) then the first term on the RHS of Eq. to n = 0.. i. The function F(co) is known as the Fourier transform of f(t).(56)  ♦ Transverse vibrations of a plucked string and forced vibrations can be studied by using Fourier series.2....e. its Fourier transform is defined by the equation F(co) = ff(t)etiwt dt 27t co ft) = f F(co) e iwt dco (57) .7 We can also write +00 1 F(n As) n=1 +00 G(k) __ t f(x) 27t f eikx dx (58) s (49) with f(x) (50) we have f f (t ') cos [s(t .3.2.t ')] dt' ds = +f G(k) e+`kx dk (59) Equation (50) is known as the Fourier integral.27c an = T f f(t) cosnoot. since $ F(s) ds = lim Ls. The above infinite series is known as the Fourier series and the coefficients an and bn are given by F(co) e+^rut to +T dco (55) and ‘/..t')] is an odd function of s. Since the cosine function inside the integral is an even function of s. For a time dependent function f(t). we will +00 +00 f(t) = 27c f f f (t•) + e±iw(tt') dt' dco (53) n=1 where we have replaced s by co. Further.. since sin [s(t . . +00 +. we will use Fourier transforms to study the propagatign of optical pulse in dispersive and nonlinear media.t')] dt' ds (51) where k is often referred to as spatial frequencya concept that is extensively used in Fourier Optics [see Chapter 19].
1937. 3.j7r 0 < t < T/2 f '11(x. What will be the Fourier expansion corresponding to full wave rectification? 8. Fo + Z Fo sin co t 1 cos2wt+ i cos4wt+.2. Arsac.8 Then F(t) Optics = f f( eTifUt dw w) '1`(x. New Delhi. 1950. 1966. K. Chua.C. E. Goyal and S. Introduction to the Theory of Fourier Integrals. Show that a(p) = 1 .S.8. . 5. New York.C. Dover Publications. PrenticeHall. New York. Introduction to the Theory of Fourier Integrals. Oxford University Press. Macmillan India Ltd. Ghatak.2 In quantum mechanics. Mathematical Physics. the solution of the one dimensional Schrodinger equation for a free particle is given by Fol where p is the momentum of the particle of mass m.Po) 2] Also show that f l 'I`(x.3 In continuation of Problem 8. Carslaw. C. Titchmarsh. 0) exp . Fourier Transforms and the Theory. if we assume 2 'P(x. H. J. 0) 12 dx represents the probability of finding the particle between x and x + dx and la(p)12 dp represents the probability of finding the momentum between p and p + dp and we would have the uncertainty relation Lx^p^h REFERENCES AND SUGGESTED READINGS 1.262 (Ir62)v4' Po exp [*] then show that a(P) _ 2 2 h2 exp 2 (P .1 Consider a periodic force of the form: F(t) = Fo sin cot for =0 for and F(t + 7) = F(t) where 2tr T Show that F(t) = . 3 J One obtains a periodic voltage of the above form in a half wave rectifier.. Clarendon Press. 0) 2 dx = 1 = f a (p) 2 dp 1r Indeed I'P(x. I. 4. Titchmarsh. Introduction to the Theory of Fourier Series and Integrals. t) = 1 f a(p) 27c li e ti (px_ zn^ d p 22 t) Problems 8. 0) e T px dx T/2<t<T 8. Oxford. Englewood Cliffs. 1995. (1959). 2. J. E. A..
a) would have the dimension of (time)1 .1 INTRODUCTION The Dirac delta function is defined through the equations 8(xa) =0 a+p R a (x)= 2 6 for aa<x<a+a x^a (1) (2) = 0 for Ix . Similarly.. but can be regarded only as a limit of a certain sequence offunctions. S '(x). 8(x . 621641 (1. 113.2 REPRESENTATIONS OF THE DIRAC DELTA FUNCTION There are many representations of the Dirac delta function.1 `t It is readily seen that if x has the dimension of length. 9. Now.a)dx = 1 aa f Ra (x)dx = 26 f dx = 1 (irrespective of the value of a) aa where a.. For an arbitrary function that is continuous at x = a. /3 > 0. the function RQ(x) has all the properties of the Dirac delta function. Fora . Dirac in The Physical Interpretation of Quantum Dynamics.0.(x) is plotted in Fig.. Perhaps the simplest representation is the limiting form of the rectangle function R a (x) defined through the following equation x Fig. which are even more discontinuous and less `proper' than S(x) itself. the function Ra (x) becomes more and more sharply peaked but the area under the curve remains unity.Chapter Nine }'THE DIRAC DELTA FUNCTION AND FOURIER TRANSFORMS Strictly of course. if x has the dimension of time then 8(x .a) would have the dimension of inverse length.04.M. a= 0.al > a (4) The function Ra.f(a) (3) 6 4 2 ea=0. One can also use the differential coefficients of S(x). Thus the delta function has an infinite value at x = a such that the area under the curve is unity. P. a=0. All the same one can use S(x) as though it were a proper function for practically all the purposes of quantum mechanics without getting incorrect results.A.1 and 0. Proceedings of the Royal Society of London (A). 0.04 =. namely.4. For a a 0. S "(x).1 for various values of a.1 Plots of RQ(x) for a = 2 and a = 0. we have n+/3 a+/3 Rectangle function 14 12 10 8 f f(x) 6 (x . 9.a)dx= f(a) f 8(x .4 _ 4 00 9. 9 26)_ . 9. S(x) is not a proper function of x. In the limit of a4 0.a)dx [using Eq. In each case the area under the curve is unity. +a+a f 8(x . the function Ra (x) has all the properties of the delta function and we may write 1 2 3 . (1)] aa aa .
a) dx = a> 0 f f(x) R a.f( a ) 9. f g etik(xa) dk (9) = . the function is very sharply peaked at x = a and has all the properties of Dirac delta function. N (E) dE represents the number of molecule whose energies lie between E and E+ dE. it has all the properties of the delta function and we may write f 0 N(E) dE = No It may be noted that whereas No is just a number. 7C(x . Thus lim f f(x)S(x . The total number of molecules is given by No f (8) N(E)dE = No 2 3/2 f Eve eE/kT dE o (kT) o irrespective of the value of g which is assumed to be greater than zero.2 5(x .lim g. in the infinitesimal interval a . T the absolute temperature and m the mass of each molecule.4 DELTA FUNCTION AS A OF THE DELTA FUNCTION DISTRIBUTION An extremely important representation of the Dirac delta function is through the following integral: . thus in the limit of g > 00. if we ask ourselves how many molecules have the precise speed El the .(x)dx a+a Fig.= =No f 0 x2 ex dx J g Thus for a large value of g. we have shown that +sin gx dx = 1. +g = lim a>o 2a 1 f(a) f dx a>a sin g(x a) = lim 1 S x . the function sin g(x . 20. let us consider the Maxwellian distribution . Further.a) 7c(xa) (7) In Appendix B. In each case the area under the curve is unity. N(E) dE = No Eve e kT dE E dk (6) (kT)3/2 (10) In order to prove Eq.a < x < a + a.a) from which Eq.°° 27r ( g>. 9. For g > 00. In order to understand this.a) = 1 27t f etik(xa) We should point out that the delta function is actually a distribution. Thus is very sharply peaked around x = a (see Fig.a) .0. the quantity N(E) has dimensions of (energy) 1 .2) and has a unit area under the curve irrespective of the value of g. (10).2 Plots of the function [sing (x . 9. + S(x . In Eq. The integral is I'I 2 = 2 .3 INTEGRAL REPRESENTATION 9.a) = lira R a(x) a40 Optics Now a+a f f(x) R Q(x)dx = 26 f f(x)dx (5) aa We assume the function f(x) to be continuous at x = a. Obviously.9. Thus when a .a) 7c(xa) where x = E/kT.a) for a = 2 and g = 7c (x a) 5. f(x) may be assumed to be a constant [= f(a)] and taken out of the integral. lim sin gx x+0 x . g > 0 7Cx where k represents the Boltzmann's constant. (6) readily follows. (6) we first note that +00 1 27c f f etik(xa) dk = sing(x .
9. This is a characteristic of a distribu. We may note that the Gaussian function given by Eqs (19) has a spatial width given by [see Fig.3 +^ f(x) dx must exist. 9. in addition to the distribution given is no reason why the factors e `kx and e `' cannot be interby Eq.f(x) dx' dk (14) where we have made use of the following integral (see Appendix A): f Thus if we define F(k) then f(x) = 2L f e` 2+px dx = a exp 4a J l. e. Example given by 9.3 answer would be zero. N(E) = No E 2zc then +f(k) = F(x) e . we have shown the following integral representation of the Dirac delta function However.1 As an example we consider a Gaussian function = A exp .The Dirac Delta Function and Fourier Transforms 9. (19). thus the Fourier transform of a Gaussian is a Gaussian. On the other hand.e. .0 f f(x) e ikx dx (15) F(k) e+ikx dk As can be seen from Eqs (20) the function F(k) is also Gaussian.3(a)] Ax a Its Fourier transform..From Eq.) Since f(x) we may write f(x) = 2n 1 f e±tk(xX) dk (12) f(x) (19) Its Fourier transform is given by + = f s(x 2r f f x)f(x')dx F (k) = A 2n (13) f exzi2v2 e lkx dx or F(k) = A a exp [k 2cs2 1 (20) etik(xX) .(20)]. (10). (b) The Fourier transform of the Gaussian function is also a Gaussian in the kspace [see Eq.x2 a 22 s(x.x. It may however have a finite number of finite discontinuities. if we do have NI molecules all of them having changed.(14) it is obvious that in Eqs (15) and (16) there tion.El) represents the Dirac delta function and has the dimensions of inverse energy. Re a> 0 (21) = +. in all of what follows we will use the definitions given by Eqs (15) and (16). i.g. the corresponding distribution function +„ e+`kx dx would be given by (17) F(k) = f f(x) 1 2 Ewe kT + Nt 3(E . (16) enables us to calculate the original function from the Fourier transform.3 (b)] (16) The function F(k) is known as the Fourier transform of the functionf(x) and Eq. F(k) has a width given by [see Fig. we could have defined the same energy El. Equation (14) constitutes what is known as the Fourier Integral Theorem that is valid when the following conditions are satisfied (see.5 FOURIER INTEGRAL THEOREM In the previous section. I I (a) The Gaussian function f(x) as given by Eq. (ii) The integral Ak1 a (22) f Fig.El) (11) V 7c (k71)3/2 where 5(E . References 4 and 5 of the previous chapter): (i) The function f (x) must be a single valued function of the real variable x throughout the range oo < x < 00.kx dx 2^c f (18) 9. 9.
• et J p to `°'r k (b) Fig. (21).5) f(t) = A exp . the Fourier transform of the sinc function is the rectangle function: F (sinx) = x 2 rect( k ) 2 (31) sine 4 (26) For a time dependent function we can write the Fourier transform in the following form [see also Sec. 9. 8. 9. Ixl< f(x) = recd ('E) = Example 9. As another example. = 5expl.. Thus. 9. we calculate the Fourier transform of the rectangle function 1. we consider the Fourier transform of the Gaussian function (see Fig. (b) The Fourier transform of the rectangle function. 9. Using Eq.3 i As an example..4 Thus Lxik1 which is a general characteristic of the Fourier transform pair.4 (a) The rectangle function.t t2 0 (34) .5. the Fourier transform is given by [using Eq. f f(t) ef10't dt (32) The inverse Fourier transform will then be given by +00 f(t) = 27r f F(co) a (0t dw (33) a/2 a/2 The above equations are nothing but Eqs (15) and (16) with x and k replaced by t and co respectively. (32)] 2 F(w) = $. Equation (25) can be written in the form F(k) = 2tc Ak / l rect 12 J = 11. (16) we can write +.0 4= 2 ka (27) a 2 a (24) a or recd 0. the rectangle function has a width Ax = a and its Fourier transform has a width a giving Ax Ak .4]: F[f(t)] = F(co) = 1 +. We denote the .2 Optics where (23) and sin x sinc x =_ (28) x is known as the `sine function'. (35)] is also plotted in Fig.4) +a/2 f e rkx dx F (k) = 1 27r J a/2 sin(ka/2) F(k) = (25) k V 7C Once again. EXamp a 9.w 4^ J (35) where we have used the integral given by Eq. I x I aJ = ^^ f 2.1.1 [ where x X _ a/2 i^ +r. The function F(co) [as given by Eq. The function F(w) is usually referred to as the frequency spectrum of the time dependent function f (t). J sine e` d4 (29) (30) /2 times Thus.9.sin 4 e'kxdk >2 Its Fourier transform will be given by (see Fig.
y. v. then its Fourier transform F(co) will have a spectral width Aco .y) e ii ±i (ux+vy) _09) )] = 27c f where u and v are referred to as spatial frequencies.1 derived above. full width at half maximum (usually abbreviated as FWHM) of f(t) by At. y) is defined through the equation + 00 +00 x 1 2tc dxg(xx')e`k(xx') In the second equation. z)e ±i(ux+vy+wz) dxdydz (41) ^ t ^ co Fig 9. the function f(t) attains half of its maximum 9.5 We will use Eqs (29) and (30) in Sec. v.y. v) e+`("x+vy) dudv f(x. +00 (38) The above equation may be compared with the relation 9. y) = 2tc f dx g(x) e. then [see Fig. The convolution can be used to obtain the Fourier transforms of the product of two functions: F(f(x) g (x f(x) g(x)e `h dx +.4 to two or three dimensions.1 2zc f f (x.6] Aw At .x' )dx' = g(x)* f(x) (43) Similarly. we substitute (x . w)e +i(ux+v)'+WZ) dudvdw (42) =± At.67 to = = f f(x') g(x .0 F(u.z)= (27r) 13/z fff F(u.6.6 THE TWO AND THREE DIMENSIONAL FOURIER TRANSFORM One can generalize the analysis of Sec. 19. thus at t value: 2 2A = Aexp^^2 r 2 with its inverse Fourier transform given by +0.0 +0. v) _. Similarly.x') by F(f(x)* g(x)) = 2TcF(k) G(k) 4 to obtain where F(k) and G(k) are Fourier transforms of f(x) and g(x) respectively. +. The proof is as follows: 3.5 The Fourier transform of a Gaussian temporal function is a Gaussian function in the frequency space.The Dirac Delta Function and Fourier Transforms f(t) F(co) 9.4 and Sec. if Aco denotes the FWHM of F(w). f(x. 9.6. The inverse transform would be given by +(40) F(u.At (37) F (f(x)* g (x)) f f dxe `kx dx' f (x ') g(x . we can define the 3dimensional Fourier transform +00 +00 +00 F(u.1 The Convolution Theorem The convolution of two functions f(x) and g(x) is defined by the relation f(x)* g(x) o Thus At = In 2 to 1.x') giving the uncertainty relation [see also Example 10. For example. 17. (9.ikx 2L f F(k')e i1' dk' Z7z f f .1 Ax Ak . the 2 dimensional Fourier transform of a function f(x. w) = (2703/2 f f f f(x.5)] Aco _ 4 ln2 to The convolution has this important property: The Fourier transform of the convolution of two functions is 2tt times the product of their Fourier transforms.34 to (36) Thus if a time dependent function f(t) has a temporal width At.
a + a) for Ix al < a 1 for x>a+a' (45) ♦ The Fourier transform of the Gaussian function t2 f(t) = A exp . The inverse transform would be given by fix. F(cv) will have a spectral spread Ow 1/At.a) dx = f(a) x^a G 6 (x) = . y) +00 1 2^t f F(k') G(k = ode F(k)* G(k) = 2f f F(u.f(x) +. y) e .3 Consider the symmetric function t/7(x) = A exp (Klxl) Show that ty"(x) = K2 1/7(x) .0 and for a well behaved function f(x). Problems Summary ♦ The Dirac delta function is defined through the equations 5(xa) =0 x=a f(x)5(x . v) 21c TTf(x. 9.a). if a function has a temporal spread of At.k')x where u and v are referred to as spatial frequencies.9.0.1 f dxg(x)e. a' > 0 2a2 i +. Hence show that 5(xa) = lim a'>0 ♦ For a time dependent function f(t).6 +00 Optics 1 2^c f dk'F(k') 2^ i +00 1.2 Consider the ramp function defined by the following equation 0 FQ(x) = F(t) = f flu)) e to t dt for x<aa (x . ♦ The 2 dimensional Fourier transform of a function f(x.1 Consider the Gaussian function 1 exp . (4). v)Ji(«+vy) dudv 7c +00+00 ♦ The convolution of two functions f(x) and g(x) is defined by the relation f(x)* g (x) Thus. Thus we get the following important result: If a function has a discontinuity of a at x = a then its derivative (at x = a) is aS(x . F(f(x) g (x)) 2 i = f f(x) g(x . its Fourier transform is defined by the equation +00 exp[(xa)2 1 6 2r 2a2 1 (44) F( co) Then.a) = = Atp eco' t0 4 2^ . times the convolution of The Fourier transform of the convolution of two functions is Tr times the product of their Fourier transforms.0 The above result tells us that the Fourier transform of the product of two functions is their Fourier transforms. 9.0 and 10.x')dx' = g (x)''' . 5.2AKS(x) F(u.+ 0 show that S(x .(x )2 J. which is continuous at Using Eq.a) where H(x .t(k . dx H(x . (21) show that f G6(x)dx = 1. where R a(x) is the rectangle function defined by Eq.a 2tc 9.2 is given by F(w) Show that dda = RQ(x).0. 2tc f e+«>t d f(t) which is the Gaussian representation of the delta function. Taking the limit a. then its Fourier transform .a) is the unit step func ♦ In general. Plot GQ (x) for a = 2 and a = 1. y) is defined through the equation + ` ( `" n) dxdy tion.
4 Consider the function f(t) = Aet212t2 ei got Calculate its Fourier spectrum F(co) = 2rc l r+^ J f(t)e"0tdt and evaluate approximately &Mt. 9. Evaluate f(t) using the expression for F(co).The Dirac Delta Function and Fourier Transforms 9. 9.I^11 ( b2 )J x2 =ab [a2 +b 2 ] exp ( a2 +b2 .7 In each case make an estimate of t x and Ok interpret physically.6 Show that the convolution of two Gaussian functions is another Gaussian function: [expLa jiiexp 1/2 9.5 Calculate the Fourier transform of the following functions (a) f(x) = Ae `k o x IxI < L/2 =0 xI > L/2 (b) f(x) = A exp [.
.
so you cannot use it for a timing signal.(k . which will be discussed in this chapter. 10. as the pulse propagates. This pulse propagates through a medium with what is known as the group velocity.1(b) we have shown the variation of the slowly varying envelope term.Ak) z] or 'f(z.. we produce a pulse. May be. the reader would like to go through chapter 11 first before going through this chapter. in particular.1(a) we have shown the variation of the rapidly varying cos(cot .(k .2 GROUP VELOCITY Let us consider two plane waves (having the same amplitude A) with slightly different frequencies co + Ow and co . t) at t=0 and t=At Obviously the rapidly varying first term moves with the velocity v P 10.Chapter Ten In a perfect wave. t) = 2A cos (wt .Act)) t . Richard Feynman in Feynman Lectures on Physics.* A study of this distortion of optical pulses is a subject of great importance in many areas. 10.(Ok) z] at t = 0.2(a) and (b) we have plotted ‘P(z.(k + Ok)z] t) = A cos[(U) . . make a notch in it. 10.Oco respectively. That means that you have to have more than one frequency in the wave. In Fig. In Fig.(Ok)z] (3) In Fig. From this experiment. t) = Acos[(co + Oco)t . I Important Milestone 1672 Isaac Newton reported to the Royal Society his observations on the dispersion of sunlight as it passed through a prism.k) z] (1) (2) w = k (4) * This chapter assumes a knowledge of waves which will be discussed in the next chapter. Vol. it undergoes distortion which will also be discussed. make it a little bit fatter or thinner. but upon the way that the index changes with the frequency. it has very important significance in fiberoptic communication systems which will be briefly discussed in Chapters 27 and 29. The superposition of the two waves will be given by 'ls(z. t) = A cos [(co + Aco) t .kz) cos [(Oco)t . you cannot say when it starts. Newton concluded that sunlight is composed of light of different colours which are refracted by glass to different extents. where k + Ok and k . and it can be shown that the speed at which signals travelis not dependent upon the index alone. In addition. In order to send a signal you have to change the wave somehow. represented by cos [(Oco)t .Aco)t .(k + Ak) z] + A cos [(co .1 INTRODUCTION When we switch a light source on and off.Ok are the wave numbers corresponding to the frequencies co + Ow and co .Oco propagating along the +z direction: 'F1 (z. L 10. the distance between two consecutive peaks is 2r/t k .kz) term at t = 0. the distance between two consecutive peaks is 2ir/k.
The distance between two consecutive peaks is 2nc/Lk. and vg are known as the phase velocity and the group velocity respectively. t= At) .10. 10. t= 0) (a) Y'(z.kz) term at t = 0. represented by cos [(Ow)t (&k)z]. the distance between two consecutive peaks is 2r/k. v . in a medium characterized by the refractive index variation n(co) k(co) = co n(co) c (7) The quantities v. t) at t = 0 and at t = At .2 (a) and (b) show the variation of `F'(z.1 (a) Variation of the rapidly varying cos (cot . (b) Variation of the slowly varying envelope term. indeed in the next section we will rigorously show that a temporal pulse travels with the group velocity given by 1 = = c [n(w) + w doo] c (8) dw In free space n(w) = 1 at all frequencies. `Y(z. and the slowly varying envelope [which is represented by the second term in Eq. 10. hence v g =vp =c (9) vg dk/dco (6) . The group velocity is a concept of great importance. the envelope moves with the group velocity Ow/Ak. V • V • v Y J Y V Y V Y V v (a) (b) Fig. (3)] moves with velocity Ow vg Ok (5) Thus vg Now.2 Optics 4/ cos kz . (b) z v Fig. at t = O.
3 .6 µm can be assumed to be given by the following approximate empirical formula: n(A0) = C0  a2i. vg = c/ng = 2.85 µm. on the other hand. every source of light would have a certain wavelength spread. do/dAo and ng(Ao) for pure silica as a function of the free space wavelength XD. 0 Ao ng (Ao) = 1.2 0 Returning to Eq. In or er tote this broadening. we may notice that the group velocity attains a maximum value at A0 = 1.451 D 2irc (10) and Thus do dco do d) o d^.6) are given in Table 10.&zm_ dz da.0464 x 108 m/s Example 10.0 = 0..1.o o° Fig.um would have a spectral width of about 2 nm.um < X10 < 1. a light emitting diode (usually abbreviated as LED) would have a spectral width of about 25 nm and a typical laser diode (usually abbreviated as LD) operating around 1.0444 x 108 m/s (13) and for X10 A0 The group index ng is defined as ng = v = n (2L 0) . vg = cing = 2. in general. As we will show later in this chapter (and also in Chapter 27).003 and Ao is measured in . co= + Thus at Ao = 1 pm. (8). More accurate values of n(Ao) and ng (A0) (as obtained by using the expression given in Problem 10.80 ..p + (14) where Co = 1. z (17) Aod^2 0 The quantity A'c.27 µm. result in the broadening of the pulse.3 ng (^1o) = co + aA2 0 (15) 2. [n(Ao) g c ^.5 .27 .um.o indicating that the difference between group and phase velocities is about 0.27 µm) higher wavelength components travel faster.um lower wavelength components travel faster.A0 dA g 0 c In Table 10. we note that the time taken by a pulse to traverse a length L of the dispersive medium is given by (16) Since the RHS depends on A0.1 For pure silica the refractive index variation in the wavelength domain 0.6]. this spectral width is usually denoted by 010 . Thus the pulse broadening will be given by . 10. we may mention that it is customary to express in terms of the free space wavelength Ao which is related to co through the following equation.1 we find that in pure silica. a 0.1 we have tabulated n(Ao). the above equation implies that different wavelengths will travel with different group velocities in propagating through a certain length of the dispersive medium. 10.451. Now. 0 dco do 27rc dA. = 0. In Fig.3 Variation of the group velocity vg with wavelength for pure silica. Simple algebra shows implying that (for X10 < 1. which is usually referred to as the spectral width of the source.Group Velocity and Pulse Dispersion 10.463 or. for . (12) Using Table 10. [A more accurate expression for n(A0 ) is given in Problem 10.um. n(A ) = 1.3 we have plotted (for pure silica) the wavelength variations of the group velocity vg.8%. Thus a white light source (like coming from the sun) would have a spectral width of about 3000 A. this wavelength is of great significance in optical communication systems. similarly for 20 > 1.1. Since each wavelength component (of a pulse) will travel with a slightly different group velocity it will. is usually referred as material dispersion because it is due to the material properties of the medium .
14 ' 2. * The numerical values in the Table have been calculated 'using the refractive index variation as given in Ref. In Eq.44954 1.0037 0.01131637 0.46401 1.01211873 Dm (ps/nm. for a nearly monochromatic source.95 14.02276 0.55 1.1 Values of n.0059 0. 0 around Ao 1.39 6.40 1. The quantity D.° ^0 den do x 104 ps/km. From Eq.01189888 0.0297 0.00153 0.00055 0.km) 0.0120 0.2 66.44498 ` 1.45013 1. the quantity inside the square brackets is dimensionless. more details will be given in Chapter 27.01130300 0.0) Dm Optics for pure silica* (tun_ ') d2n ng (Ao) dA2 (µm 2) ca. a pulse of temporal width i° will get broadened to 2f where 'Sf .46332 1.46744 1. ng and 20 (µm) n(2.01958 0.00365 0. after propagating through a length L of the dispersive medium.n is positive and it is said to be characterized by negative dispersion when D.45456 1.46197 1.75 0.nm (19) where A° is measured in pm and we have assumed c 3 x 108 m/s.0164 0.10.1 30.00062 0.011400.00416 0.0741 0.45 1.44439 1.20 + (fr2 m) 2 (18) In the next section we will explicitly show this for a Gaussian pulse. .64 hence the subscript m.46209 1. We assume OA° = 1nm = 10 9m and L = 1 km = 1000 m We may mention here that the spectral width of a pulse is usually due to the intrinsic spectral width of the sourcewhich for a typical LED is about 25 nm and for a Commercially available laser diode is about 12 nm.44896 1.45075 '1.01725159 0.20 1. Thus a 20 ps (Fourier Transformed) pulse will have a spectral width Ov 1 12 5x1011 Hz 20 x 10 implying Ov c We may see that d 2n da.50 1.90 0. On the other hand.95 1. A medium is said to be characterized by positive dispersion when D.35 1.52 24.01327862 0.00 1.4nm 1 2 D"` 3A.01170022 0.0221 0.n is usually referred as the material dispersion coefficient (because it is due to the material properties of the medium) and hence the subscript m on D.6 84.1 21.70 0.46189 1.7 14.01125037 0. Indeed the wavelength A0 = 1270 nm is usually referred to as the zero material dispersion wavelength and it is because of low material dispersion.00235 0.44556 ' 1.46283 1.45561 1.01552236 0.45282 1.44670 1.00462 172.9 40. Indeed.0541 0.00305 0.46489 1.46924 1.01146001 0. ° LOA.01153568 0..60 1.4 Table 10.3 106.46253 1.80 0. the second and third generation optical communication systems operated around Ao = 1300 nm.0020 0.46189 1.87 10.o  and define the dispersion coefficient as 02 m 0.10 1.27 pm.01423535 0.01257282 0.5 8.72 18.46601 1.47154 1. (17) we see that the broadening of the pulse is proportional to the length L traversed in the medium and also to spectral width of the source A A°.01170333 0.9 135.01125123 0.58 2.05 1.45139 1.44726 1.0086 0.01206070 0.46229 1. the intrinsic spectral width could be extremely small and the actual spectral width of a pulse is determined from its finite duration (such a pulse is often referred to as a Fourier transformed pulse).45208 1.23 21.46279 1.40 0.44839 1.85 0.45364 1. 0.44379 1. 2 (see Problem 10.6).44613 1.4 51.30 1.46196 1.15 1.46318 0.46241 1.0400 0.46214 1. is negative. (17).44783 1.25 1.
0042 (µm)2 d21. (28) in Eq. = + 21. one uses laser diodes with 0 = 1. . ns increases with A 0 ). t) = E0e To a+iwot (28) Ifwe substitute Eq.wo )t dt (29) expr. therefore.7 ps/km. given by Actual electric field = Re(E) = I A I cos(cot . (20) is a practical impossibility because at an arbitrary value of z. (27) we would obtain tz A(w) To E0 f e e t (w .coo ) Z To l 20 . Now at A. t) we know we can determine E(z. (27) substitute it in Eq. = . Thus for 4210 = 2 nm. (20) is said to describe a monochromatic wave which propagates with the phase velocity given by v= w =c (22) n k n E(z = 0. the actual broadening of the pulse will be DT = 2. t) 10.kz) (20) e tmt dt (27) Thus if E(z = 0. A wave packet can always be expressed as a superposition of plane waves of different frequencies: E(z t) giving D.. A may be '. t) is the Fourier transform of A(w) and using the results of the previous chapter we obtain A(co) i 2i += f E (` = 0.1 that for 0 1.Group Velocity and Pulse Dispersion 10. t) = Ae+`Q't.85 gm and AA0 = 25 nm. In practice.3 GROUP VELOCITY OF A WAVE PACKET The displacement corresponding to a onedimensional plane wave propagating in the +z direction can be written in the form E(z. Thus for & = 25 nm.4 Gaussian Pulse: As an example. t) using the following recipe: We first determine A(w) from Eq.4 (co . Thus.1 ns after traversing through 1 krn of the silica fiber.55 µm X1. t) = Ae t(wt . E(z = 0. = f A(w)ei[wtkz] dc0 (25) Obviously E(z = 0. the displacement is finite for all values of t.oQ < t < oo (24) Example 10.85 gm X10 d2n 0.55 gm and z 20 = 2 nm.0 = 0. Now at 20 = 1. Example 10. The plane wave represented by Eq. the actual broadening of the pulse will be X1. for example. (20) becomes E IAIe' twtkz+ The actual displacement is the real part of E and is.85 ps/km. t) = f A(w)e +twt dw (26) 0T„ = 43 ps/km implying that the pulse will broaden by about 43 ps after traversing through 1 km of the silica fiber. The wave described by Eq.1 ns/km implying that the pulse will broaden by 2.3 d2n = 0. in general. In the IVgeneration optical communication systems. the displacement is finite only over a certain domain of time and we have what is known as a wave packet. we consider a Gaussian pulse for which we may write t2 2 where A represents the amplitude of the wave and k(co) = co n(w) (21) c n being the refractive index of the medium.030 (µm)2 dA. one used LED's with = 0.nm We may mention here that.nm the positive sign indicating that higher wavelengths travel slower than lower wavelengths.27 µm. (Notice from Table 10.2 In the Igeneration optical communication system.2:) which corresponds to a sinusoidal variation for all values of time. E(z = 0. (25) and carry out the resulting integration.0 giving D..complex and if we write A=IAIe`° then Eq.5 Example 10.kz + 0) (23) the negative sign indicating that higher wavelengths travel faster than lower wavelengths.
2. e.06 CUo = w o+ Aw A(c))ei[mt.Ow.4(b)] so that we may write E(z.coo) dk 2 We may mention here that in order to have clarity in the figure we have chosen a very small value of To.(°82s6 10. thus at w=coo+co The right hand side is a function of (z .6 tr x 1(3.3. Returning to Eq.(z a)2 exp[i `^° (zct)] a)2 L c 2° r (36) which represents a distortionless propagation of a Gaussian pulse in a nondispersive medium*.2 Propagation in a Dispersive Medium For a wave propagating in a medium characterized by the refractive index variation n(w). (35).ct) and thus any pulse would propagate with velocity c without undergoing any distortion. (35). the corresponding spectral density function S(w) is plotted in Fig. 10. E(z. (28).4(b) will be much more sharply peaked. 10. (25). S(w) is a very sharply peaked function of co around co = coo.. 10. (25) can be written in the form o 47r exp[2 (wwo)2 Tp] (32) E(z.4(b)..18 x 10 14 Hz Thus Aco =0. 10. t) = Eo exp[.. for the Gaussian pulse given by Eq. we will have c Now. Thus. t) = f A(w)e i9e'(zct) dw In Fig. A(cv) can be complex and as such one defines the power spectral density f +"e ax Optics 10. (36) follows directly from. A(co) is a very sharply peaked function [see. we will discuss this in greater detail in the chapter on coherence (Chapter 17). (29) in Eq. (28)] for a 20 fs pulse (To = 20 x 10 15 s) corresponding to A = 1 um (wo . we would readily get Eq. usually To has a much larger value. In this tiny domain of integration. 10.4 (a) we have plotted the function . The full width at half maximum of S(w) (usually abbreviated as FWHM) is denoted by Ow.1. thus k(co) S(cv)=IA(cv)IZ For the Gaussian pulse S(w) = E 22 (31) = w c (35) and Eq. the value of dw is obtained from the following equation 1 = exp .5 we have shown the distortionless propagation of a 20 fs pulse. it is left as an exercise to the reader to show that if we substitute for A(co) from " Eq. we may make a Taylor series expansion of k(co) k(w) = k(coo) + (co .k(o))z] dw f (37) w o. In general. (36). Fig.10. S(w) attains half of its maximum value.6 where we have u s e f l e f s /4a (30) v as J ^ (see Appendix A). 14 Hz). the function A(co) is negligibly small. (38) =ko+v (cocoo)+2 g co0 0) 2 r * Whereas Eq. in Fig. we consider the following cases: dw + to =w o 2(c) cvo) or k(co) +.Ora because for co > coo + ico and for co < wo . As can be seen.g. t) k(co) _ co n(co) [ or FWHM = = 2 21n2 . dw 2 ra=mo . in most problems. A larger value of To will imply a: much smaller value of Oco (resulting in greater monochromaticity of the pulse) and obviously Fig.35 To To (33) Thus the Gaussian pulse of temporal width 20 fs has a frequency spread Ow given by (34) ow .2 2 E0e TO cos(coot) [which is the real part of Eq. .1 = I e +/3 Propagation in a NonDispersive Medium For electromagnetic waves. Eq.3. the free space is a nondispersive medium in which all frequencies propagate with the same velocity c.
10. (39).7 II I i '' 50 25 0 I 25 t (fs) 50 (a) Fig. if we retain only the first two terms on the RHS of Eq. (b) the corresponding frequency spectrum which is usually a very sharply peaked function around co = coo. then Eq. Now.a We may mention here that we have now defined vg through Eq. (37) would give us +0 E(z.) w=u. t) d2k Y= dw2 w=wo (42) f ^wo A(co) exp [i (ko z + v zcot 8 do) (43) . (41)we will show below that the envelope of the pulse moves with velocity vg which is the group velocity.15 s) Gaussian pulse corresponding to Ao = 1 pm .4 (a) A 20 fs (= 20 x 10 . where ko = k(coo) (40) (41) 1 vg and dk dc.Group Velocity and Pulse Dispersion 10.
if we take into account all the three terms in Eq. E(z. (39)].. the contribution from the region Ico . 10. (30) to carry out the integration.0 duo dw (54) d2. We define Elie pulse broadening Oz as Oz = VT2 (z) = I p''To 1 (dk/dw) t. A (w) is given by Eq. propagates it undergoes temporal broadening.. we would readily obtain do 6. 10. in any case. where we have replaced the limits from co to +00 because.)J duo For the Gaussian pulse [see Eq. (39). we would obtain E(z.11 it t (fs) Fig. thus.vgt. (52) we find that as the pulse.10. (29).8 Optics z= z0\ ^I .5 Distortionless propagation of a Gaussian pulse in a nondispersive medium. 1I II ._ J vg) Phase Term _co Envelope Term n ^ 0. z and t do not appear independently but only as z .o . t) = e i(motkoz) Phase Term s (zvgt) f di2 J A(S2)eEnvelope Term PE 2yz 2 To (50) (45) by (46) O° The corresponding intensity distribution would be given where S2= wwo We see that in the envelope term. (43) can be rewritten in the form +0. if we now substitute A (w) in the above equation and use Eq..o (47) Thus if we neglect y [and other higher order terms in Eq. Writing cot =(wwo)t+coot Eq.6 we have plotted the time variation of the intensity at different values of z. Next.cool > Ow is going to be extremely small. (28)]. the envelope of the pulse moves undistorted with the group velocity Vg = I(z. From Eq. the pulse moves undistorted with group velocity V g. t) = where E0 +ip e`(metkoz) exp (49) (44) E(z. t) +. t) = z( )/2o exp (51) where r2(z) = 'to (1 + Pz) (52) In Fig. = 21212 0 (53) ei(W O t=koz) Now dw2 = dw[c( 1 d ccao[n(Ao) ^o 2tcc2 d?n r A(52)exp iS2 t.
55 x 1.. (53) we get from Eq. where the quantity inside the square brackets is dimension.3 2 ^0 Substituting for in Eq.Group Velocity and Pulse Dispersion 10. Eq.On the other hand. (49) can be written in the form: A.o o d4 t? vs AT= z A.55 x 0.50 fs 01 . t) = coot + 2 t Thus _ Y1. notice the temporal broadening of the pulse.8). E(z.oc 2 which is identical to the result obtained in the earlier section [see Eq. at z = 2 less.50 fs ^I 14. For exp where the phase term is given by (58) .1) d2n = dA.7 and 10. t) . we assume pure silica.t (fs) Fig.004165 (µm) 2 0(z.2x2.55 pm. Further.1 Ps (1010) and 2 tan p (59) (60) Equation (59) represents the phase term and the instantaneous frequency is given by . for a 10 fs pulse. t) = exp [T(z)/r 0 ] v2 Eo x T2 (z) [i (O(z. (54) The Chirping of the Dispersed Pulse If we carry out simple manipulations.koz] Example 10.5 As an example.0.2 d2n AA. (56) and for y from Eq. (17)]. 10. at this wavelength (see Table 10.6 The time variation of the intensity at different values of z.3.743 x 1026 m is2 For a 100 ps pulse propagating through a 2 km long fiber .0 = 1.743x102' x2xi03 =1.o 2.55 x 10 6 [1. AA o I 1 Aco 2 To 1 2rtc (56) 10.004165] 2rr x 9 x 1016 2. since the spectral width of the Gaussian pulse is AT = 22 fs given by [see Eq.9 14. (33)] implying Aco = 2 zo = 4 mm we will have (55) T =[TO+(OT) 2] 1/2 =25ps f TO we may write 1 To showing that a 10 fs pulse doubles its temporal width after propagating through a very small distance (see Figs 10.
7 The temporal broadening of a 10 fs unchirped Gaussian pulse (. we assume = coo + 21c t .011 l . Now P2yz 22 = 0 2 x 2.743 x 1026 x 2 x 103 (100 x10122 ) co 2ic+ t vg (62) 0. it will get compressed until it becomes unchirped and then it will broaden again with opposite chirp.10 = 1.5.5 w I NI 1fr V 50 fs ^I II 4 50 fs I 4 i 50 fs 19504 25 0 25 9752 t (fs) Fig.10 = 1. z=0 I PI z=2z0 11 25 0 t (fs) 25 Fig. The frequency chirp is therefore given by ( Aco =co(t)  =a Example 10.10.vg l (61) .10 Optics 0. 10. Notice that since dispersion is positive.6 In continuation of Example 10. w(t) ll showing that cv(t) changes within the pulse.55 pm) propagating through silica. 10. the pulse gets down chirped.8 If a downchirped pulse is passed through a medium characterized by negative dispersion.55 pm and consider the chirping produced in a 100 ps pulse propagating in pure silica at z = 2 km.
. At t . for fused silica. the spot size w0 of the beam is about 5 um (see Examples 29. p and x will also be negative implying that the instantaneous frequency (within the pulse) decreases with time (we are of course assuming z > 0).6.8 and 29. From Example 10..e. the frequencies are slightly higher which is usually referred as `blue shifted'.8).. but when the beam propagates over an optical fiber over long distances (a few hundred to a few thousand kilometers).. For example. If the intensities are large. For a 5 mW laser beam propagating through such a fiber. it will have frequency lower than coo) and the trailing edge of the pulse (t > z/vg) will be blueshifted (i. 10.e. the resultant intensity is given by _3 I= p 5 x 10 W (65) = 108 W/m2 50 x 1012 m2 Aeff Thus the change in refractive index is given by An = n21 3.e. Vg =+50ps at the trailing edge of the pulse) Aco=. it will get compressed until it becomes unchirped and then it will broaden again with opposite chirp. the effective propagation constant is given by .47 and n2 3.2 x 1020 m2/W. Notice 8 Aco = 9 x 10 wo At t=? Aco=0 Vg 10. we can conclude the following: For positive dispersion (i. Now in a single mode fiber. it will get compressed until it becomes unchirped and then it will broaden again with opposite chirp (see Fig. the frequencies are slightly lower which is usually referred as `redshifted'.i = 50 ps Vg (i. From Eq. negative value of y). Thus the effective* crosssectional area of the beam. if the effective area of the light beam is Ae ff . Aeff = 7two = 50 µm2.1x108 Hz Thus.4 SELF PHASE MODULATION It may be mentioned that as a pulse propagates through a dispersive medium. If a downchirped pulse is passed through a medium characterized by negative dispersion.. at the trailing edge of the pulse. the accumulated nonlinear effects can be_significant. p (and therefore x) will be positive and the leading edge of the pulse (t < z/vg ) will be redshifted (i.2 x 1012 (66) Although this is very small. at the front end of the pulse) 2 2p Oco = (50 x 1012 ) T 0 (1+p 2 ) + 2x0. it will have frequency higher than w0): This implies that we will have an upchirped pulse.. 2.Group Velocity and Pulse Dispersion 10. * Values adapted from Ref.011x50x10 12 (100 x 10 12 )2 = +1. Further.e.1.e. no = 1.e. 10. New frequencies are generated when the medium is nonlinear we will briefly discuss this here. it will get further broadened and also get further down chirped. Thus if an upchirped pulse is passed through a medium characterized by positive dispersion.1 x 108 Hz Thus at the leading edge of the pulse.7 where at t = 0 we have an unchirped pulse. As the pulse propagates further.e. This is shown in Fig. (61) it can be readily seen that at negative values of z. it has frequency lower than co0).. it has frequency higher than c)o) and the trailing edge of the pulse (t > zlv g) is redshifted (i. this is known as a downchirped pulse in which the leading edge of the pulse (t < z/vg ) is blueshifted (i. the frequency spectrum remains the samei. Different frequencies superpose with different phases to distort the temporal shape of the pulse (see Problem 10. then the intensity is given by I= P Aeff (64) where P is the power associated with the light beam. The refractive index of any material is a constant only for small intensities of the propagating laser beam.11 Similarly we can discuss the case of negative dispersion (implying a positive value of y). the refractive index variation is approximately given by n = no + n21 (63) and at t(i.the beam remains confined to a very small area for long distances! We consider a laser pulse (of frequency coo) propagating through an optical fiber. That is the great advantage_of the _optical fiber . where n2 is a constant and I represents the intensity of the beam.9). no new frequencies are generated.10)..e.
10.12 w0 = c [no + n2 1]
Optics
k
where we have neglected dispersion [ i.e., p = 0 in Eqs (49) and (52)]. Thus rr 2Itw(t)
= \2
(4)
no +nz'Alf wo
1
J (67) Thus, for such a propagating beam, the phase term is approximately given by
e+i(wotkz) =
+
2g z Pol z l exp twozo l vg J
sJ
2 TO
For 2.o = 1.55 ym = 2rx3x105 1.22x 1015s 1 _6 1.55 x 10 )o Further, for Po = 15 mW, 'co = 20 fs and z = 200 km wo =
27rc 2gzPo t_ z
exp +l coot  6)0 (no
C
+
n2 P(t) Aeff
Z
=
C
+i^
where the phase 1 is defined as (I) (z,t)
=
coot 
o (no +
n2
Ac 1
eff JJ
(68)
wozo Il
vg 1.22 x 10 15 x (20 x 101S ) 2 t?
vg
We can define an instantaneous, frequency as [cf. Eq. (61)]: w(t) where
2x2.6x103 x2x105 x15x103
a
=wo  g dP(t) z
=3.2x10 13 t?
vg
n2 coo  2r n2 gcAeff 2, O A eff
+0.64 for t  ? = 20 fs (trailing edge of the pulse)
(69) vg
For Aeff = 50 µm2, Xo = 1.55 pm and n2 = 3.2 x 1020 m2/W; g 2.6x10`5 W1 m 1 . Now, for a Gaussian pulse propagating with group velocity vg [see Eq. (51)]:
2
0.64 for t 
v
20 fs (trailing end of the pulse)
g
P(z,t) =
Po
exp
2 TO
Thus the instantaneous frequency within the pulse changes with time loading to chirping of the pulse as shown in Fig.. 10.9; this is known as self phase modulation (usually abbreviated as SPM). Note that since the pulse width has not changed, but the pulse is, chirped, the frequency content of the pulse has increased. Thus SPM leads to generation of new frequencies. Indeed by passing a pulse through a fiber char
z = 200 km
50. . ,
0
50
Fig. 10.9 Due to self phase modulation, the instantaneous frequency within the pulse changes with time leading to chirping of the pulse. Calculations correspond to P0 = 15 mW, .10 = 1550 nm, 'co = 20 fs, Aeff = 5µm2 and vg = 2x10 $ m/s.
Group Velocity and Pulse Dispersion
10.13
Fig. 10.10 Laser pulses of 80fs duration having a wavelength 800 rim (and total energy of 1.6 nJ) are incident on a special optical fiber known as a holey fiber in which a silica core is surrounded by a periodic lattice of air holes; holey fibers are characterized by very small mode field diameters which leads to very high intensities. Because of the high intensities, SPM (Self Phase Modulation) and other nonlinear effects can be observed; these nonlinear effects result in the generation of new frequencies. In this experiment, the entire visible spectrum gets generated which can be observed by passing the light coming out of the optical fiber through a prism. The repetition rate of the laser pulses is 82 MHz. The special fibers were fabricated by Dr. Shyamal Bhadra and Dr. Kamal Dasgupta and their group at CGCRI, Kolkata and the supercontinuum generation was observed by Professor Ajoy Kar and Dr. Henry Bookey at Heriot Watt University, Edinburgh. A color photograph appears on the cover of the book. acterized by very small crosssectional area (so that the value of g is large) it is possible to generate the entire visible spectrum (see Fig. 10.10).
° dg o After traversing through a distance L in a dispersive medium,
a pulse will broaden by an amount
_ LA. °
d2n
Summary
o When we switch a light source on and off, we produce a pulse. This pulse propagates through a medium with what is known as the group velocity, which isgiven by 1 vs
dkldco
A0c where DA° is the spectral width of the source; the subscript m denotes that the fact we are considering material dispersion. The dispersion coefficient is given by
2 m D = At =  3 0
Ap d n x 104 ps/km.nm dg "
For a medium characterized by the refractive index variation
n(co)
n(w), c the group velocity is given by k(w) v g
where A0 is measured in p.m and we have assumed c = 3 x 108 m/s. For example, for silica, at A0 = 1.55 pm, dzn/dA = 0.00416 ,um 2 and Dr„ = +22 picoseconds per kilometer (length of the medium) per nanometer (spectral width of the
2 source). On the other hand, for silica d n
c
[nao)_ A0
d2 = 0 A0 = 1.27 pm. Indeed the wavelength A0 = 1.27 pm is usually dA0
around
where A0 is the wavelength in free space and c = (3 x 108 m/s) I is the speed of light in free space.
referred to as the zero material dispersion wavelength and it is because of low material dispersion, the second and third generation optical communication systems operated around A0 = 1.3 pm.
10.14 ♦ For a Gaussian pulse
E(z =
Optics
r 2 0, t) = E0 expl i2
o
a
1° o'
the temporal width after propagating through a distance z is given by z(z) = zo'l + p2,; thus the temporal broadening is given by
AT
= .‘Iz2 (z)zo = Iplzo
where 2 Ao 2 d2n ' 2ir2 Ao PAZ z 0 Thus at XD = 1.55 pm, for a To = 100 ps pulse (propagating in pure silica), AT = 0.55 ps/km.
p = 'r0
where Ai = 0.6563 µm and A2 = 0.4861 pm. (a) Calculate the values of A and B. (b) Using the Cauchy formula calculate the refractive index at 0.5890 pm and 0.3988 pm and compare with the corresponding experimental values: (i) (1.51124 and 1.52546) for borosilicate glass and (ii) (1.45845 and 1.47030) for vitreous quartz. [Ans: (a) For borosilicate glass A = 1.499, B = 4.22 x 1015 m2 giving n = 1.51120 at A. = 0.5890 µm, and n = 1.52557 at A = 0.3988 pm; for vitreous quartz A = 1.44817, B = 3.546 x 10 15 m2] 10.6 The refractive index variation for pure silica in the wavelength region 0.5 pm < Ao < 1.6 pm is accurately described by the following empirical formula n (Ao) =
co
+C I A0+C2 Ao+
23
(Ao 1) C4 C5 + (Al  1)2 + (AO l)3
Problems
10.1 Using the empirical formula given by Eq. (14) calculate the phase and group velocities in silica at 2i = 0.7 pm, 0.8 pm, 1.0 pm, 1.2 pm and 1.4 pm. Compare with the (more accurate) values given in Table 10.1. [Ans: n(Ao) = 1.456, 1.454, 1.451, 1.449, 1.455; ng (Ao) = 1.4708, 1.4670, 1.4630, 1.4616, 1.4615]. For pure silica we may assume the empirical formula 10.2 n(Ao) = 1.451  0.003 [ AO 2
where Co = 1.4508554, C1 = 0.0031268, C2 = 0.0000381, C3 = 0.0030270, C4= 0.0000779, C5 = 0.0000018,1= 0.035 and Ao' is measured in pm. Develop a simple program to calculate and plot n(Ao) and d 2n/d2 in the wavelength domain 0.5 < Ao < 1.6 pm and compare with the results given in Table 10.1. 10.7 (a) For a Gaussian pulse given by
_ t2
l
0
where Ao is measured in µ.m. (a) Calculate the zero dispersion wavelength. (b) Calculate the material dispersion at 800 nm in ps/km.nm. [Ans: 1.32 pm; 101 ps/km.nm] 10.3 Let n(Ao) = no + AAo where Ao is the free space wavelength. Derive expressions for phase and group velocities [Ans: vg = c/no] 10.4 Consider a LED source emitting light of wavelength 850 nm and having a spectral width of 50 nm. Using Table 10.1 calculate the broadening of a pulse propagating in pure silica. [Ans: 4.2 ns/km] 10.5 In. 1836 Cauchy gave the following approximate formula to describe the wavelength dependence of refractive index in glass in the visible region of the spectrum n(A) = A + 4
0
E=E0a zp
e imo'
the spectral width
Aco
is
approximately given by
1
zo
Assume Ao = 8000 A. Calculate Oct for zo = 1 ns and for zo = 1 ps. w0 (b) For such a Gaussian pulse, the pulse broadening is given by AT =
2z I
yl where y= d2z . Using Table 10.1, caldo)
a
culate AT and interpret the result physically. [Ans: (a) Oct = 4 x 107 and 4 x 104; w0 (b) y= 3.62 x 1026 m 1 s2; Oz = 0.072 and = 72 ps/km for zo = 1 ns and 1 ps respectively] 10.8 As a• Gaussian pulse propagates the frequency chirp is given by
2p t z vg zo(1+p2)
Now (see also Table 12.2) n(A l ) =1.50883 for borosilicate glass n(A2 ) =1.51690,, n(A l ) =1.45640 .46318} for vitreous quartz n(2. 2 ) =1
where p is defined in Eq. (50). Assume a 100 ps (=To) pulse at Ao = 1 pm. Calculate the frequency chirp '60 at o
Group Velocity and Pulse Dispersion t  zlvg = 100 ps, 50 ps, +50 ps and +100 ps. Assume z =
10.15
A(w)e ik(w)z implying that no new frequencies are generateddifferent frequencies superpose with different phases at different values of z. 10.11 The time evolution of a Gaussian pulse in a dispersive medium is given by
1 km and other values from Table 10.1. 0 +4.5 x 108 at (t  z/vg) = 100 ps, 50 ps, +50 ps and +100 ps respectively]. 10.9 Repeat the previous problem for 20 = 1.5 µm; the values of io and z remain the same. Show that the qualitative difference in the results obtained in the previous and in the present problem is the fact that at A = 1 pm we have negative dispersion and the front end is red shifted (Act) is negative) and the trailing end is blue shifted. The converse is true at A = 1.5 ,um where we have positive dispersion. 10.10 The frequency spectrum of E(0, t) is given by the function A(w). Show that the frequency spectrum of E(z, t) is simply [Ans:
w _  4.5 x 108 ,  2.25 x 108, +2.25 x 108 and
E(z, t) =
EO
ei(%tkoz)
l+ip
exp
where p = 2 Calculate explicitly the frequency spectrum
0 of E(0,t) and E(z,t) and show that the results agree with that
of Problem 10.10.
REFERENCES AND SUGGESTED READINGS
1. R. P. Feynman, R. B. Leighton and M. Sands, The Feynman Lectures on Physics, Vol. I, AddisonWesley Publishing Co., Reading, Massa., 1964. 2. U.C. Paek, G.E. Peterson and A. Carnevale, `Dispersionless single mode light guides with a index profiles' Bell System Technical Journal, Vol. 60, 583, 1981. 3. A. Ghatak and K Thyagarajan, Introduction to Fiber Optics, Cambridge University Press, Cambridge, 1998, (Reprinted in India by Foundation Books, New Delhi).
WAVE PROPAGATION AND THE WAVE EQUATION
If you are dropping pebbles into a pond and do not watch the spreading rings, your occupation should be considered as useless', said the fictional Russian philosopher, Kuzma Prutkoff. And, indeed we can learn much by observing these graceful circles spreading out from the punctured surface of calm water. Gamow and Cleveland
11.1 INTRODUCTION
In this chapter we willdiscuss the phenomenon of waves. A wave is propagation of a disturbance. For example, when we drop a small stone in a calm pool of water, a circular pattern spreads out from the point of impact. The impact of the stone creates a disturbance which propagates outwards. In this propagation, the water molecules do not move outward with the wave; instead they move in nearly circular orbits about an equilibrium position. Once the disturbance has passed a certain region, every drop of water is left at its original position. This fact can easily be verified by placing a small piece of wood on the surface of water. As the wave passes, the piece of wood makes oscillations and once the disturbance has passed, the wood comes back to its original position. Further, with time the circular ripples spread out, i.e., the disturbance (which is confined to a particular region at a given time) produces a similar disturbance at a neighbouring point at a slightly later time with the pattern of disturbance roughly remaining the same. Such propagation of disturbances (without any translation of the medium in the direction of propagation) is termed as a wave. It is also seen that the wave carries energy; in this case the energy is in the form of kinetic _energy of water molecules. We will first consider the simplest example of wave propagation, viz., the propagation of a transverse wave on a string. Consider yourself holding one end of a string, the other end being held tightly by another person so that the string does not sag. If you move the end of the string up and down a few
times then a disturbance is created which propagates towards the other end of the string. Thus, if we take a snapshot of the string at t = 0 and at a slightly later time At, then the snapshots will roughly* look like the ones shown in Figs 11.1 (a) and (b). The figures show that the disturbances have identical shapes except for the fact that one is displaced from the other by distance v 0 t where v represents the speed of the disturbance. Such a propagation of a disturbance without its change in form is a characteristic of a wave. The following points may, however, be noted: (a) A certain amount of work is done when the wave is generated and as the wave propagates through the string, it carries with it a certain amount of energy which is felt by the person holding the other end of the string. (b) The wave is transverse, i.e., the displacement of the particles of the string is at right angles to the direction of propagation of the wave. Referring back to Figs 11.1 (a) and (b), we note that the shape of the string at the instant At is similar to its shape at t = 0, except for the fact that the whole disturbance has travelled through a certain distance. If v represents the speed of the wave then this distance is simply v 0 t. Consequently, if the equation describing the rope at t = 0 is y(x) then at a later instant t, the equation of the curve would be y(x  vt) which simply implies a shift of the origin by a distance vt. Similarly, for a disturbance propagating in the x direction, if the equation describing the rope at t = 0 is y(x), then at a later instant t the equation of the curve would be y(x + vt).
* We are assuming here that as the disturbance propagates through the string, there is negligible attenuation and also no change in the shape of the disturbance.
11.2 y (x, t)
Optics
t=o
v
4
(a)
t=t At
'lp
yl'
(b)
Fig. 11.1 A transverse wave is propagating along the +xaxis on a string; (a) and (b) show the displacements at t = 0 and t = At respectively.
Example 11.1 Study the propagation of a semicircular pulse in the +x direction whose displacement at t 0 is given by the following equations: 2 112 IxlR y(x, t = 0) =[R x2] (1) IxlR =0
t=0
(a) Solution: For a wave propagating in the +x direction the dependence of y(x, t) on x and t should be through the function (x  vt). Consequently,
Y(x, t) = [R2 =0
u2 (x  vt)2]
x
t = to v to
Ix vtl R Ix  vtl>R
(2)
The shape of the pulse at t = 0 and at a later time to is shown in Fig: 11.2. Equation (2) immediately follows from the fact that y(x, t) has to be of the form y(x vt) and at t 0, y(x, t) must be given by Eq. (1). [ Example 11.2 Consider a pulse propagating in the minus xdirection with speed v. The shape. of the pulse at t = to is given by
y(x, t = to) = b2 a2 +(xxo) 2
(b) Fig. 11.2 The propagation of a semicircular pulse along the +xaxis; (a) and (b) show the shape of the pulse at t = 0 and at a later time to respectively. (Such a pulse is known as a Lorentzian pulse.) Determine the shape of the pulse at an arbitrary time t.
Solution: The shape of the pulse at t = to is shown in Fig: 11.3 (a). The maximum of the displacement occurs at x = xo. Since the
(3)
Wave Propagation and the Wave Equation
11.3 and y(x) = a cos k(x  vet) at t = 0 t (7)
The two curves are the snapshots of the string at the two instants. It can be seen from the figure that, at a particular instant, any two points separated by a distance (a) Y
v (t t0)
= 2irIk
(8)
(b)
x
Fig. 11.3
The propagation of a Lorentzian pulse along the minus xaxis; (a) and (b) show the shape of the pulse at t = to and at a later instant t respectively.
have identical displacements. This distance is known as the wavelength. Further, the displaced curve (which corresponds to the instant t = At) can be obtained by displacing the curve corresponding to t = 0 by a distance v0 t; this shows that the wave is propagating in the +x direction with speed v. It can also be seen that the maximum displacement of the particle (from its equilibrium position) is a, which is known as the amplitude of the wave.
y(x, t)
pulse is propagating in the x direction, at a later time t, the maximum will occur at xo  v(t  to). Consequently, the shape of the pulse at an arbitrary time t would be given by y(x, t) =
b2 a2 + [x  x0 + v(t
 t0 )]2
(4)
Equation (4) could have been written down directly from Eq. (3) by replacing x by x + v(t  to).
11.2 SINUSOIDAL WAVES: CONCEPT OF FREQUENCY AND WAVELENGTH
Till now we have been considering the propagation of a pulse which lasts for a finite amount of time. We will now consider a periodic wave in which the displacement y(x, t) has the form y(x, t) = a cos [k (x + vt)
Fig. 11.4 The curves represent the displacement of a string at t = 0 and at t = At respectively when a sinusoidal wave is propagating in the +xdirection. In Fig. 11.5 we have plotted the time dependence of the displacement of the points characterised by x = 0 and x = 0 x. These are given by and where
w=kv
+ 0]
(5)
y(t)=a cos cot y (t) = a cos (wt  kix)
atx=0 at x = ix
(9)
where the upper and lower signs correspond to waves propagating in the +x and x directions respectively. Such a displacement is indeed produced in a long stretched string at the end of which a continuously vibrating tuning fork is placed. The quantity 0 is known as the phase of the wave (see Chapter 7). We may, without loss of generality, assume 0 = O. Thus for a wave propagating in the +x direction, y(x, t) = a cos k(x  vt) (6)
(10)
The curves correspond to the time variation of the displacement of the two points. Corresponding to a particular point, the displacement repeats itself after a time
T = 2,r/w
(11) (12)
In Fig. 11.4 we have plotted the dependence of the displacement y on x at t = 0 and at t = At. These are given by att=0 y(x)=a coskx
which is known as the time period of the wave. The quantity v = 1/T is known as the frequency of the wave and represents the number of oscillations that a particle carries out in one sec
11.4 y(x, t) = a cos (kx  wt)
Optics
It should be pointed out that the entire discussion given above would remain valid for an arbitrary value of the phase factor O.
11.3 TYPES OF WAVES
As mentioned earlier, when a wave is propagating through a string the displacement is at right angles to the direction of propagation. Such a wave is known as a transverse wave.* Similarly, when a sound wave propagates through air the displacement of the air molecules are along the direction of propagation of the wave; such waves are known as longitudinal waves. However, there are waves which are neither longitudinal nor transverse in character; for example, when a wave propagates through the surface of water, the water molecules move approximately in circular orbits.
Fig. 11.5 The curves represent the time variation of the displacement at x = 0 at x = Lx respectively when a sinusoidal wave is propagating the +xdirection. ond. It can be seen from the two curves in Fig. '11.5, that the two points x = 0 and x = z x execute exactly similar vibrations except for a phase difference of k&. In fact any two points on the string execute simple harmonic motions with the same amplitude and same frequency but with a phase difference of kxo where xo represents the distance between the two points. Clearly if this distance is a multiple of the wavelength, i.e., xo=mA,,m=1,2,... then luxo = mA, = 2 mgr
11.4 ENERGY TRANSPORT IN ' WAVE MOTION
A wave carries energy; for example, when a transverse wave propagates through a string, the particles execute simple harmonic motions about their equilibrium positions and associated with this motion is a certain amount of energy. As the wave propagates through, the energy gets transported from one end of the string to the other. We consider the time variation of the displacement of a particle, which can be written as y = a cos (cot + 0) The instantaneous velocity of the particle would be (16)
which implies that two points separated by a distance which is a multiple of the wavelength vibrate with the same phase. Similarly, two points separated by a distance A,, i A,, .. . vibrate in opposite phase. In general, a path difference of xo corresponds to a phase difference of Using Eqs (10)(12), v or 1 T
7 2.
7 xo.
we get
w 2tc kv v 27r A. (13)
+ 0) dt =(T)acv sin (cotgiven by \ Thus, the kinetic energy would be
v
=
(17)
v = vA,
T=
Notice the similarity in the variation of the displacement with respect to x (at a given value of time) and with respect to t (at a given value of x); see Figs 11.4 and 11.5. The similarity can be expressed by writing Eq. (6) in the form y(x, t) a cos I x
dy12 1 2 m dtJ
=
2 ma2 w sin2 (cot + 0)
ma2w? [sin2 (wt + 0].
(18)
The total energy (E) will be the maximum value of T
E = ( T)max
T t)

(14)
which shows that the wavelength A. in Fig. 11.4 plays the same role as the time period Tin Fig. 11.5. Equation (14) is often written in the form
=
2 ma?co2
(19)
* Electromagnetic waves are also transverse in character. However, it should be mentioned that the electromagnetic waves have also a longitudinal component near the source which dies off rapidly at large distances (see Sec. 23.4).
Wave Propagation and the Wave Equation For a sound wave propagating through a gas, the energy per unit volume, e, would be given by 1 mna2 d e= = fpa2 w2 = 21c2pa2v2 (20)
11.5 where ao represents the amplitude of the wave at unit distance from the source.
where n represents the number of molecules per unit volume and p (= nm) the density of the gas. With such a wave, we can associate the intensity which is defined as the energy flow per unit time across a unit area perpendicular to the direction of propagation. Since the speed of propagation of the wave is v, the intensity (1) would be given by* I = 2xpva2v2 (21)
Example 11.3 A source of sound is vibrating with a frequency of 256 vibrations per second in air and propagating energy uniformly in all directions at the rate of 5 Joules per second. Calculate the intensity and the amplitude of the wave at a distance of 25 m from the source. Assume that there is no absorption [speed of sound waves in air = 330 m/sec; density of air = 1.29 kg/m 3 ].
Solution: Intensity I = 5 J/s
4,r x (25) 2
6.3x104
m2 Jsec l m2
5
Thus
a=[
= 3 x 10
7
1
1/2 1
8n3 x 1.29 x 330 x 256 x 256
25
Thus the intensity is proportional to the square of the amplitude and square of the frequency. Let us consider a wave emanating from a point source in a uniform isotropic** medium. We will assume that there is no absorption and that the source is emitting W joules per second (W represents the power of the source). Consider a sphere of radius r whose centre is at the point source. Clearly, W joules per second will cross the spherical surface whose area is 4n r2. Thus, the intensity I would be given by 1= W 4rc r2 (22)
m.
Example 11.4 Show that when a transverse wave propagates through a string, the energy transmitted per unit time is z pco2a2v where p is the mass per unit length, a the amplitude of the wave and v the speed of propagation of the wave. Solution: The energy associated per unit length of the string is
2 pco2a2; since the speed of the wave is v, the result follows.
11.5 THE ONEDIMENSIONAL WAVE EQUATION
In Sec. 11.1 we had shown that the displacement yyof a onedimensional wave is always of the form y! = f(x  vt) + g(x + vt) (24)
which is nothing but the inverse square law. Using Eqs (21) and (22) we obtain W = 2n2pva2v 2 47r r2 or a= 1/2 1 W 1 8lr pvv 2  r (23)
showing that the amplitude falls off as Y . Indeed, for a spherical wave*** emanating from a point source, the displacement is given by
f = ao sin (kr  cot)
r
where the first term on the RHS of the above equation represents a disturbance propagating in the +x direction with speed v and similarly, the second term represents a disturbance propagating in the x direction with speed v. The question now arises as to how we can predict the existence of waves and what would be the velocity of propagation of these waves? The answer to this question is as follows: If we can derive an equation of the form
* This can be easily understood from the fact that if we have N particles per unit volume, each moving with the same velocity v, then   the number of particles crossing an unit area (normal to v) per unit time would be Nv. ** Isotropic media are the ones in which physical properties (like velocity of propagation of a particular wave) are the same in all directions. In Chapter 22 we will consider anisotropic media. *** When waves emanate from a point source in an isotropic medium, all the points on the surface of a sphere (whose centre is at the point source) have the same amplitude and the same phase; in other words, the locus of points which have the same amplitude and the same phase is a sphere. Such waves are known as spherical waves. Far away from the source, over a small area, the spherical waves are essentially plane waves. 
i. (32a) (32b) where f and g are arbitrary functions of their argument. etc. (26) indeed satisfies Eq. Although the general solution of Eq.vt) + g(x + vt) (26) Optics a2 1/! tion of the form of Eq. x x+d x Similarly. then we can be sure that waves will result and yr will represent the displacement associated with the wave. 11.9 we will discuss the general solution of the wave equation. We rewrite Eq. Instead of sinusoidal variation it is often more convenient to write the solution in the form yi = A exp [i(Icx ± cot + 0)] (33) and rf =x+vt Thus a (77) a1^ = f' () a + g' ax ax ax where primes denote differentiation with respect to the argument. The force at A the upward direction is in ax + g" (^j '1) a ax (29) =f"() + g"(71) ' g'(i1)] (30) and aalti2f Tsin61 =. the speed of which would be v.e. In writing Eq. Due to the tension T. We must mention that the simplest particular solutions of the wave equation correspond to sinusoidal variation: ty = A sin [k(x ± vt) + 0] or t! = A cos [k(x ± vt) + As shown in Sec. Let us consider a small length AB of the string and calculate the net force acting on it in the ydirection. the end points A and B experience force in the direction of the arrows shown in Fig. If the string is pulled in the ydirection then forces will act on the string which will tend to bring it back to its equilibrium position. we obtain] ax =f'( ) + g'(^1) Differentiating once more. a =1 = ax [see Eq.6. (25) is of the form y^ = f(x . In its equilibrium position the string is assumed to lie on the xaxis.6 1 a2w (25) v2 at'. This follows from the fact that the general solution of Eq.vt where A. 11. a a = f' (4) a + g (ID at at at = v [f ( ) + 11. f'() Since where.2. . (25) will be derived in Sec. (26) indeed satisfies Eq. (25). we get ax2 = f "(4) Similarly.Ttan ei=  = v2 [f+ g "(77)] " Tax. a? from physical considerations. 11. is the wavelength and v the frequency of the wave. (26) in the form yr =f(g ) + g(t1) and kv = w = 2 tc v (27) (28a) (28b) where =x . as before. (33). we can predict the existence of waves. 11.* In Sec. k = 27c/A. In the next three sections we will derive the wave equation for some simple cases. A and 0 represent the amplitude and initial phase of the wave.11. (25).9. (28)]. we will show here that the solution expressed by Eq. if we ever obtain an equa v az (31) Tsin92 =TtanQ2 =Tax *In Chapter 23 we will derive the wave equation from Maxwell's equations and thereby obtain an expression for the speed of electromagnetic waves. the force at B in the upward direction is Thus (4) + On) = at =f which proves that the solution expressed by Eq. it is implied that the actual displacement is just the real part of ly which is Acos (kx±cot +0) = d f . Consequently.6 TRANSVERSE VIBRATIONS OF A STRETCHED STRING Let us consider a stretched string having a tension T. (25) from physical considerations.
In the displaced position. the displacement is not rigorously of the form given by Eq. (24). There is.2. ^. .Ax and. It should be mentioned that in an actual string.7 x+ dx Fig.Ax . 11. (34) where we have used the Taylor series expansion of I aye ax 1x+dx about the point x: ray ax /x+dx i aY a (ay) ay x x ax/x + a dx and have neglected higher order terms because dx is infinitesimal.7 Propagation of longitudinal sound waves through a cylindrical rod.6 Transverse vibrations of a stretched string. where we have assumed 6 and 02 to be small.^] P R Q S x 4(x) H^I P' 4(x + ox) IHI R' 0 T x Q' S' cg x iix+ Ax Fig. in general. Thus we may conclude that transverse waves can propagate through a stretched string and if we compare the above equation with Eq. the longitudinal strain would be where Am is the mass of the element AB. Thus the net force acting on AB in the ydirection is 11.7 LONGITUDINAL SOUND WAVES IN A SOLID In this section we will derive an expression for the velocity of longitudinal sound waves propagating in an elastic solid. (25) we obtain the following expression for the speed of the transverse waves: v = Tlp (36) The vibrations of _a clamped string will be discussed in Sec. then Am=pdx and we get a Increase in length Original length _ ax Ax Lx _ a4 ax (37) a2y a2y (35) ax2 Tip at2 which is the onedimensional wave equation. this is a consequence of the various approximations made in the derivation of the wave equation.4(x) + Ax a2Y at2 = T a2 I dx ax2 =Ax+ Ox The elongation of the element would be . Let the longitudinal displacement of a plane be denoted by 4(x). 13. where we have chosen the xaxis to be along the length of the rod (see Fig.Wave Propagation and the Wave Equation 11.4(x) + Ax = 4(x) +a Ax .7). Thus the displacements of the planes PQ and RS would be 4(x) and 4(x + Ax) respectively. Let PQ and RS be two transverse sections of the rod at distances x and x + Ox from a fixed point 0. the distance between the planes P'Q' and R'S' would be T (ay Jx+dx (ay axx 2 axe dx. therefore. Let us consider a solid cylindrical rod of crosssectional area A. an attenuation of the wave and also the shape does not remain unaltered. The equation of motion is therefore Om 4(x + Ox) . 11. If p is the mass per unit length. . 11.
6. ** See. the velocity of which would be given by [see. if one carries out a rigorous analysis of the vibrations of an extended isotropic elastic solid.8).* The above derivation is valid when the transverse dimension of the rod is small compared with the wavelength of the disturbance so that one may assume that the longitudinal displacement at all points on any transverse section (like PQ) are the same.AP(x + Ox)]A = YA aax2 2 If p represents the density. Thus F(x) = and. In general. (45)] is due to the restoring forces arising because of the elastic properties of the material. Let the pressure of the gas in the absence of any disturbance be Po. Once again. Thus the force acting on the column P'Q'S'R' would be [AP(x) . because of a longitudinal displacement. Now. if we consider the volume P'Q'S'R' then a force F is acting on the element P'Q' in the negative xdirection and a force F(x + Ax) is acting on the plane R'S' along the positive xdirection. whereas corresponding to the transverse waves discussed in Sec. modulus of rigidity and bulk modulus respectively. force acting on the element P'Q'S'R' will be F(x + Ax) . the equation of motion for the column P'Q'S'R' would be ax (AP)AAx = pAAx ate * In a similar manner one can consider transverse waves propagating through an elastic solid. therefore.2 (1 + a') v` Now. rf and K represent the Poisson ratio. 8] yr = where 17 represents the modulus of rigidity. We must mention that the transverse wave [whose velocity is given by Eq. we consider a column PQSR as shown in Fig.11. the plane PQ gets displaced by (x) and the plane RS gets displaced by a distance (x + dx).8 LONGITUDINAL WAVES IN A GAS Ax (41) In order to determine the speed of propagation of longitudinal sound waves in a gas. 11. 11. Ref. one =C1 P1v2 (43) =a  (AP) &A (46) where A represents the crosssectional area. Thus the resultant. (45) P) where a. aF ax YA ax2 (40) YA ax (39) 1 __ Y p .8 (a). Thus the equation of motion will be pAAx a2 = YAAx a2 ate ax2 or a2 _ 1 a24 ax2 v? at2 vl (42) where represents the velocity of the waves and the subscript 1 refers to the fact that we are considering longitudinal waves. . 11. 11. Ref. Consequently.F(x) = aF Ax . then the mass of the element would be p A A x. Let Po + A P(x) and Po + AP(x + Ax) denote the pressures at the planes P'Q' and R'S' respectively. for example. we have Longitudinal stress Optics can show that the velocities of the longitudinal and transverse waves will be given by** (1a) v/ = (pY (1+6)(12a)) 1/2 = 1/2 4 \u2 K+ rt 3 p 1/2 =A y = Y x Strain (38) (44) ax where F is the force acting on the element P'Q'. if we consider the column P'Q 'S 'R' then the pressure Po + AP(x) on the face P'Q' acts in the +x direction whereas the pressure Po + AP(x + dx) on the face R'S' acts in the x direction.8 Since the Young's modulus (Y) is defined as the ratio of the longitudinal stress to the longitudinal strain. for example. 5. the string moved as a whole and the restoring force was due to the externally applied tension. (see Fig.
11. if we assume y= 1. then we obtain v = Ax (b) Fig. K Now. a change in pressure gives rise to a change in volume. (53) where we have used the fact that the modulus of rigidity (rl) for a gas is zero. If we differentiate the above expression.40. Thus. the pressure fluctuations will be rapid and one may assume the process to be adiabatic. and if the frequency of the wave is large (> 20 Hz).dx = Thus.3 x 103 g/cm3.9 P or  (AP) = rP . P = 1.rPa4 A Ox Atx ax (59) (60) * This section may be skipped in the first reading. (44). where p represents the density of the gas. For air. . r a2v v2 at2 (56) a Ax ax2 we introduce two new variables 4=xvt (57) (58) A V= 4 AAx ax rJ=x+vt and write Eq. we obtain Q S x (a) fi(x) h^I P' a2 __ i a2 v2 at2 ax2 where (52) t. 11.4(x) + Ax] . we get AP VY + y0IPA V = 0 AV V The change in the length of the column PQSR is AP = [4(x + &) . we obtain Eq. the change in the volume (49) =1= Ks yP (55) and if we substitute this expression for K in Eq.(x+Ox) I*^I R' Q'  S' x P\ v2 1 (53) CyP represents the velocity of propagation of longitudinal sound waves in a gas. (47) and (51). Thus v = 330 m/sec The adiabatic compressibility of a gas is given by x 1 'aV 1 = yP VaP) s (54) Tx (AP) = p at2 (47) where the subscript s refers to the adiabatic condition (constant entropy). The bulk modulus (K) of a gas is the inverse of Ks.9 THE GENERAL SOLUTION OF THE ONEDIMENSIONAL WAVE EQUATION* In order to obtain a general solution of the equation a2.01 x 106 dynes/cm2 and p = 1. ax2 (51) Using Eqs. Thus _. Now.8 Propagation of longitudinal sound waves through air. we may write PVY = constant (48) where y= Cp/C. (56) in terms of these variables. ax or (50) av = ty a ax a4 + an a4 ax + a17 ax The original volume V of the element is A Ax.Wave Propagation and the Wave Equation ox ^ R 11. represents the ratio of the two specific heats.
(56) we obtain a2v a or a2v +2 a ti + an = a42 2 a2v an +a (63) an a J The LHS is a function of x alone and the RHS is a function of t alone. the analysis given in Sec. . it can be an arbitrary function of : = _k2 (69) * The method of separation of variables is a powerful method for solving certain kinds of partial differential equations. ly) a + a ra. ** Notice that partial derivatives have been replaced by total derivatives.vt) represents a disturbance propagating in the +x direction with speed v and the function g(x + vt) represents a disturbance propagating in the x direction.^l ail at anlan) at t) = X(x) T(t) (67) or where X(x) is a function of x alone and T(t) is a function of t alone. (60) with respect to x. In the method of separation of variables. we set this constant equal to k2.2.g.vt) + g(x + vt) (65) or a2v = a2v a2v' 2 a2v^ ak2 + aka + a71 2 ax2 Similarly where f and g are arbitrary functions of their argument. we try a solution of the wave equation aty a + v an ae = ate v +v [ [(N ap lay)) a a a a. see. Substituting in Eq. (67)]. thus 1 d2X X(x) dx2 1 1 d2T v2 T(t) dt2 Thus avi/a has to be independent of rl. 8. we obtain the following as the most general solution of the wave equation w= f() + g (el) = f(x . if the variables separate out then the method is said to work and the general solution is a linear sum of all possible solutions. e.2 a2y + a2' (62) anal aTl e ale Substituting ' the expressions for a2 vdax2 and a2 v/ate from a2W at2 ve _ 1 2 (68) 2 or** 1 dex X(x) dx2 1 d2T v 2 T(t) dt2 Eqs (61) and (62) in Eq. The function f(x . J The constant of integration can be an arbitrary function of 71 and since the integral of an arbitrary function is again an arbitrary function.. If the variables do not separate out one has to try some other method to solve the equation.5 Solve the onedimensional wave equation [Eq. each function depending only on one independent variable [see Eq. (66). This is possible only when each side is equal to a constant.11. This implies that a function of one independent variable x is equal to a function of another independent variable t for all values of x and t..10 where we have used the fact that = F'( 1 and ax = 1 optics ) or ty = (64) ax  Differentiating Eq. According to this method the solution is assumed to be a product of functions. Solution. atV = ate a + av an at a4 at an at v and Example 11. (25)] by the method of separation of variables* and show that the solution can indeed be expressed in the form given by Eqs (32) and (33). we get T(t) d2X dxe X(x) a v2 dt a2v . we get a2v a2 = a) ax(a a (a)ax a axanIan ax ax + + a1(a)ax an (61) b F ( ) d + constant of integration. On substituting this solution.^ + an^a)at] at a2v ax2 v2 at2 a2yr (66) of the form vi(x. however.
however..k2 z Z de where kx 2. One can also have lg(x. ky and k respectively. but the frequency and wavelength have to be related through Eq.kz X dx2 s 1 d2Y = .kz=0 (83b) . etc. we obtain 2 1 XYZ d T YZT d2X + XZT d2Y + XYT d2Z = dt2 v2 dy2 dx2 dz2 we have a wave propagating along the xaxis.(kxz +kyz +kz)  or or 1/!(x. all values of the frequencies are possible. the second term is a function of y alone. Notice that for a given value of the frequency. 11. For example.9). The threedimensional wave equation is of the form z v y! = The solutions of Eqs (78) and (79) could be written in terms of sine and cosine functions. t) = X(x) Y(y) Z(z) T(t) (76) Consider a vector r which is normal to k. Substituting in Eq.cot + 0) (82) 1 a2ty v2 at2 + (74) where v2 1f/ = ax + ay2 az (75) Solve the threedimensional wave equation by the method of separation of variables and interpret the solution physically. y and zcomponents are kx. each term must be set equal to a constant. the phase fronts are parallel to the yz plane. r ± cot + 0)] (81) where the vector k is defined such that its x. we can have waves propagating in different directions depending on the values of kx. In general. The direction of propagation of the disturbance is along k and the phase fronts are planes normal to k. t) _ (A cos kx + B sin kx) (C cos cot + D sin cot) Suitable choice of the constants A. thus k • r = 0.k= k . it is more convenient to write them in terms of the exponentials: yi=Aexp[i(kx x+ky y+kz z±cot+0)] = A exp [i(k . we write tif(x. Example 11.6 Till now we have confined our discussion to waves in one dimension. Thus 1 Thus tp(x. y. (74). C and D would give yi(x. the value of k2 is fixed [see Eq. etc. 8. Solution: Using the method of separation of variables. Similarly. We write 1 d2X = . if kx=k and ky =kz =0 (83a) where X(x) is a function of x alone. such waves are known as plane waves (see Fig. (80)]. t) = a exp [± i(kx ± cot + 0)] as a solution. However.2). ky and kz. (72). term on the LHS is a function of x alone. t) = a cos (kx + cot + 0) where cot = kzv2 and k2 =kx +ky +kz (80) representing waves propagating in the +x and x directions respectively. there are systems (like a string under tension and fixed at both ends) where only certain values of frequencies are possible (see Sec. ky and kz are constants.kz 2ttv (72) represents the angular frequency of the wave. The solutions of Eqs (70) and (71) can easily be written down: X(x) = (A cos kx + B sin kx) dye (78) and T(t) = (C cos cot + D sin cot) 1 d2Z = . consequently at a given time the phase of the disturbance is constant on a plane normal to k. for k= . One could have also written yi = A cos (k • r . B.cot + 0) d2T + co2T(t) = 0 ate (79) (73) v2 1 d2T T dt2 z =.Wave Propagation and the Wave Equation or d2X + k2X(x) =0 dx2 (70) or dividing throughout by 2d2Y 1 d2X [X dx2 ] + [ dy tV 1 d2Z dz2] v2 [T dt2 ] (77) and d2T + co2T(t) = 0 d t2 (71) where co=kv= Since the first. z. t) = a cos (kx .
(88) becomes 1 a2u Since rz = x2 + y2 + z2 therefore. (89) is therefore given by = f(rvt) + g(r+vt) r r r ar (90) a2>l< law l< ar a = +x = 1 al a a^ a)a ax2 .= t)) . 11.7 For a spherical wave the displacement v1 depends only on r and t where r is the magnitude of the distance from a fixed point. Thus V2tg= V2 = r ar r2 ar2 l a1 + y2 aV _ y2 2 a>l< r3 ar z Fig. [kx=k. v1 is a function of r alone and the last two terms vanish. 0 and 0 and ally dig ar + _a O + ay1 a¢ ax ar ax ae ax DO ax However. For time dependence of the form exp (± icot) one obtains yr = A r ar r r ar2 r2 ar [1 a2>l< _ 1 a>l< exp [i(kr ± cot)] (91) * In general is a function of the three spherical polar coordinates r. Obtain a general solution of the wave equation for a spherical wave. 11. t) (88) We will first show that for a spherical wave 2 r then (84) a(2al r2 ar ar _ a2y1 2a y/1 a 2a^v ar2 + r ar r2 ar (r ar) yar a = l ar ax Now* ax la( au r2 ar r ar ) 1 a2u .1 al p  the first and the second terms (on the RHS) representing an outgoing spherical wave and an incoming spherical wave respectively. + r2 ar2 az2 r3 ar Adding. . we get 3 ayr + a2 v1 _ 1 ayr r ar ar2 r ar where we have used Eq.and yaxes (see Fig.11. for a spherical wave.= 2x or ar ax 1 1 a2u r ar2 = a2 r at2 or x r __ a2u 1 a2u ar2 = v2 at2 (89) Thus atIf ax x ay1 which is of the same form as the onedimensional wave equation.9 Propagation of a plane wave along the direction k. 2r. Solution: =aw+?aw= a r2 al ar2 r ar r2 ar ar ) (87) Thus.1 aw z2 a2 _ z2 aw r at. The general solution of Eq.9). kz _ the waves are propagating in a direction which makes equal angles with x.r ar 2 (85) (86) Thus Eq. (86). Example 11. the wave equation for a spherical wave simplifies to 1 a 2 aw _ 1 a z O2v1= w v2 atz ar r2 ar If we make the substitution u(r.12 or Optics a21g = 1 all/ x2 ay/ _ x2 al(/ r ar + r2 ar2 ax2 r3 ar Similarly a2>v ay2 and a2w .
and. v = 10 cm/s] tJ(x. 0) is an even function of x. a(k) = 0. o)(= 27tv) the angular frequency of the wave.48 s l. the solution obtained by the method of separation of variables represents the general solution. The first term on the RHS of the above equation represents a l disturbance propagating in the +x direction with speed . ♦ For a spherical wave. tp(x.2 sin (0. 0) = a(k) cos kx dk J b(k) sin k(x . The quantity 0 is known as the phase of the wave. Calculate the wavelength.vt) dk (93) is of the form t/i=f(xvt)+g(x+vt) J 0 + J 0 b(k) sin kx dk (94) If tg(x. 0) is an odd function of x. Such a displacement is indeed produced in a long stretched string at the end of which a continuously vibrating tuning fork is placed. Solution: An arbitrary disturbance propagating in the +x direction can be written as a superposition of the functions cos k(x . both a(k) and b(k) will be finite).8 As mentioned earlier. therefore.Wave Propagation and the Wave Equation 11. Show that a pulse of the form W(x.2t) .5x + 3t) (iii) y(x.32 s . Now. (iii) v = 1 s l. Notice that the factor 1/r term implies that the amplitude of a spherical wave decreases inversely with r.(x. ♦ The most general solution of the wave equation a2w.2) = a(k) cos kx dk 0 (95) a (96) Using the Fourier cosine transform. (ii) v = 0.vt) dk J 0 + 0 where the functions a(k) and b(k) are to be determined from the form of th(x.0)=Aexpla 2J which is an even function of x.t) where in each case x and y are measured in centimetres and t in seconds.vt) Thus v. t v = 6 cm/s.) the wave number and ) represents the wavelength associated with the wave.13 Hence A exp (x . the displacement is given by `F = a cos [kx ± oit + 0] where a represents the amplitude of the wave. Problems 11.lx .5 sin 2tt(O. 0 22 = A exp k4 . (If t//(x.2x . and therefore. In the present case 2 where f and g are arbitrary functions of their argument. 0) is neither even nor odd. t) = 0. frequency and the velocity in each case.(ii) y(x. then we can be sure that waves will result and IV will represent the displacement associated with the wave. the intensity will fall off as 1/r 2 . 0) = A exp .vt) and sin k(x . t) = A exp .v and similarly. Thus if we can derive the wave equation from physical considerations. the second term represents a disturbance propagating in the x direction with speed v. amplitude. [Ans: (i) v = 0. t) = 0. ax2 a2y i 1 v2 at2 can be expressed as a superposition of the solutions obtained by the method of separation of variables. b(k) = and if tp(x. t) = a(k) cos k (x .signs correspond to incoming and outgoing waves respectively. a(k) = 0 1. The upper and lower signs correspond to waves propagating in the x and +x directions respectively. t = 0).1 The displacement associated with a wave is given by (i) y(x. t) 0.wt) dk (97) l J Summary ♦ For a sinusoidal wave.u (x.1 cos (0. the displacement is given by `I' = A e i(k• ±o)t) r where the + and .(x 2 t)2 J (92) f exp r I  k24 2l I cos (kx . the intensity will fall off as 1/r 2 . Thus 2 tp(x. v = 10crrt/s.vt)2 1 62 J L6A ^ Notice that the factor 1/r term implies that the amplitude of a spherical wave decreases inversely with r. k (= 27r/a. Example 11. 0) cos kx dx 2 = 2 A f exp _z cos kx dx tc 0.
5 cos (0. Blackie & Son Ltd. H.14 11. what will be the corresponding wavelength range? [Ans: 16. Oliver & Boyd Ltd. H. London. Coulson.11 Obtain the unit vector along the direction of propagation for a wave. Waldson. Waves and Oscillations. t).26 x 105 cm/s] 11.5t] where x. Physics of Waves. 5. Dover Publications. t) = a cos [2x + 3y + 4z . New York. Freeman).0.5 sin (0.004(x 10)] 11. p = 2.(k . A.. t) = a exp ] 2 11. What will be the wavelength and the frequency of the wave? [Ans: 2 z+ 3 y+ 4 z 29 29 29 11. What is the wavelength and the velocity associated with the wave? [Ans: y(x. J. 1969. I (xbv(tta))2 1 [Ans: y(x. t = 5) = 0.4 sin (0. Heath and Co. [Ans: v1= 5.9 A sonometer wire is stretched with a tension of 1 N. 1969. [Ans: v = 707 cm/s] 11. Compare this value of velocity with the speed of sound in air.lx + 7r/3) 11. t) = 0.13 Calculate the velocity of longitudinal elastic waves in aluminium (Y = 0.5 sin [0.. New York. Write the equation describing the wave and if p = 0. Pain. III.5 m > A> 0. McGrawHill Publishing Co. J.cot] Show that the wave propagates along a direction making an angle 30° with the xaxis. . Obtain the displacement (as a function of x) at t = 10 sec. 1955. G. C. At t = 5 sec the displacement associated with the wave is given by the following equation: y(x. 8. [Ans: A = 1571 cm: y(x. t) = 0.4 eX )262 + P ..70 x 10 12 dynes/cm2. the displacement of which is given by tp(x. At t = 0.2 A transverse wave (A= 15 cm..5 Consider a wave propagating in the +xdirection with speed 100 cm/sec. 1964. t = to) = a expl 6 Find y(x. Electromagnetism. t Ans: y(x. 1955.(k + 4 )262 ]^J 1 5) = 0.000 Hz. y and z are measured in centimetres and t in seconds. Waves and Diffraction. z. Physics. z. 2..1 x 105 cm/s] REFERENCES AND SUGGESTED READINGS 1. y. Theoretical Physics (translated by I. Mass.97 x 104 ergs/cm] 11.4t . F. 1967.. the point x = 0 is at its equilibrium position moving in the upward direction. Berkeley Physics Course.5 cm. calculate the energy associated with the wave per unit length of the wire.7 g/cm3). R. where x and y are measured in centimetres and t in seconds. The Physics of Vibrations and Waves.7 Repeat the above problem corresponding to y(x. [Ans: 308 m/s. (b) 1.lx) + 0.12 At t = 0. 1965. [Ans: Energy associated with the wave = 1. Vibrations. C. a wave packet is described by the following displacement: yi(x.t=0) =aexp 2 x2 6 cos kx Express the above displacement as a superposition of plane waves and interpret it physically if ko >> 1. M. London. v = 200 sec1 ) is propagating on a stretched string in the +xdirection with an amplitude of 0. C. Braddick.11.5]) 11. 9. J.8 A Gaussian pulse is propagating in the +xdirection and at t = to the displacement is given by 11. t) = 0. Heald.lx + 200tr(t . D. 11. 7.H. Vol.4 Calculate the speed of longitudinal waves at NTP in (a) argon (y= 1. 1968. 3.5 cos [0. W. McGrawHill Publishing Co. Maidenhead. y. Joos. New York. Slater and N.6 Consider a wave propagating in the xdirection whose frequency is 100 see r . The displacement at x = 10 cm is given by the following equation: y(x = 10..41).67). 4. A.. 1968.4 t) where x and y are measured in centimetres and t in seconds.1 g/cm.10 The displacement associated with a threedimensional wave is given by 1/i(x.3 Assuming that the human ear can hear in the frequency range 20 < v < 20. t) = a cos [ 2 +2 kx ky . Waves and Oscillations. McGrawHill Book Co.lx) Optics (x b)2 Y(x.0165 m] 11. Edinburgh.2 g/cm. (b) Hydrogen (y= 1. t = 0) = A(k) J A(k)e'1 dk. A. J. Calculate the wavelength and the frequency associated with the wave and obtain an expression for the time variation of the displacement at x = 0. John Wiley & Sons. Boston. = 47t k ga'Z j exP . Waves. Crawford. Seventh Edition.. Frank. Physical Science Study Committee. C. S.5 cos (0. r+c. . Van Nostrand Publishing Co. Elmore and M. Calculate the velocity of transverse waves if p = 0. 6. London.
which is just the converse of the prediction made from Newton's corpuscular theory (see Sec. the locus of points which have the same amplitude and are in the same phase are spheres. 12. when Huygens put forward his wave theory. everyone believed in Newton's corpuscular theory. The medium . It should be pointed out that Huygens did not know whether the light waves were longitudinal or transverse and also how they propagate through vacuum. if we have a point source emanating waves in a uniform isotropic medium. the rectilinear propagation of light and the fact that light could propagate through vacuum. could the nature of light waves be understood properly. a small portion of the sphere can be considered as a plane and we have what is known as a plane wave [see Fig. So empowering was Newton's authority that the scientists around Newton believed in the corpuscular theory much more than Newton himself. refraction. Huygens explained satisfactorily the phenomena of reflection. Huygens' theory predicted that the velocity of light in a medium (like water) shall be less than the velocity of light in free space. refraction and total internal reflection and also provided a simple explanation of the then recently discovered birefringence (see Chapter 22). The envelope of these wavelets gives the shape of the new wavefront. each point of a wavefront is a source of secondary disturbance and the wavelets emanating from these points spread out in all directions with the speed of the wave. In Fig.Now. . 2. if the shape of the wavefront at an earlier time is known. for example.1(b)]. each point on the circumference of the circle (whose center is at the point of impact) oscillates with the same amplitude and same phase and thus we have a circular wavefront. At a later date. no one really believed him. 12. circular ripples spread out from the point of impact. including the double refraction in calcite discovered by Bartholinus. From the Internet 12. which had satisfactorily explained the phenomena of reflection. when Maxwell propounded his famous electromagnetic theory. He considered that light is transmitted through an allpervading aether that is made up of small elastic particles.2. In this case we have spherical wavefronts as shown in Fig. As we will see later. propounded his wave theory of light (published in his Traite de Lumiere in 1690). During that period. On the other hand.2). 12. 12. Huygens explained many of the known propagation characteristics of light. At large distances from the source.1 INTRODUCTION The wave theory of light was first put forward by Christiaan Huygens in 1678. It was only in the later part of the nineteenth cen tury.1(a). a Dutch physicist. the data on the speed of light through transparent media were also available which was consistent with the results obtained by using the wave theory. according to Huygens' principle. The wave character of light was not really accepted until the interference experiments of Young and Fresnel (in the early part of the nineteenth century) which could only be explained on the basis of a wave theory. as such. On this basis. S1S2 represents the shape of the wavefront (emanating from the point 0) at a particular time which we denote as t = 0. On the basis of his wave theory. each of which can act as a secondary source of wavelets. in a communication to the Academie des Science in Paris.Christiaan Huygens. A wavefront is the locus of the points which are in the same phase.2 HUYGENS' THEORY Huygens' theory is essentially based on a geometrical construction which allows us to determine the shape of the wavefront at any time. if we drop a small stone in a calm pool of water.
it is maximum in the forward direction and zero in the backward direction*. According to the rectilinear propagation of light (which is also predicted by corpuscular theory) one should obtain a shadow in the region PQ of the screen. say t = 0. light was known to travel in straight lines and Huygens explained this by assuming that the secondary wavelets do not have any amplitude at any point not enveloped by the wavefront.2. a small portion of the spherical wavefront can be approximated to a plane wavefront thus resulting in plane waves. the obliquity factor is 1 (thereby giving rise to maximum amplitude in the forward direction) and when 8 = the obliquity factor is zero (thereby giving rise to zero amplitude in the backward direction). referring back to Fig.3). This backwave is shown as Si'SZ in Fig. 12. is assumed to be homogeneous and isotropic. Then. the secondary wavelets emanating from a typical point B will give rise to a finite amplitude at B' only and not at any other point.5 we will show how Huygens' principle can be used in inhomogeneous media. If we draw a common tangent to all these spheres. 12. Finally.12. 12. in Sec. 12. 12. The absence of the backwave is really justified through the more rigorous wave theory.2 Optics (a) (b) Fig. . As we will see in a later chapter.2. which is of the form (1 + cos 8) where 8 is the 'angle between the normal to the wavefront and the direction under consideration. we draw spheres of radius v At. In Huygens ' theory. Fig. (b) At large distances.2 Huygens' construction for the determination of the shape of the wavefront. with each point on the wavefront as center. i. which is again spherical and centered at O. S 1S2 is a spherical wavefront centered at 0 at a time. because we also obtain a backwave which is not present in practice. However. In Sec. RECTILINEAR PROPAGATION Let us consider spherical waves emanating from the point source 0 and striking the obstacle A (see Fig. this is not rigorously true and one does obtain a finite intensity in the region of the geometrical shadow.e. then we obtain the envelope which is again a sphere centered at O. The dashed curve represents the backwave. In the next section we will discuss the original argument of Huygens to explain the rectilinear propagation of light. one drawback with the above model. the presence of the backwave is avoided by 8' assuming that the amplitude of the secondary wavelets is not uniform in all directions. given the shape of the wavefront at an earlier time. Thus. SIS corresponds to the state of the wavefront at a time At. Thus the shape of the wavefront at a later time At is the sphere Si Sz There is. 12. however. where v is the speed of the wave in that medium.4 we will derive the laws of refraction and reflection by using Huygens ' principle. the medium is characterized by the same property at all points and the speed of propagation of the wave is the same in all directions. Clearly when 8 2 = 0. 12. 12. • 0 Fig. 12.3 . at the time of Huygens..3 Rectilinear propagation of light.1 (a) A point source emitting spherical waves. Let us suppose we want to determine the shape of the wavefront after a time interval of At. 0 is a point source emitting spherical waves and A is an obstacle which forms a shadow in the region PQ of the screen. * Indeed it can be shown from diffraction theory that one does obtain (under certain approximations) an obliquity factor.
Then B1 B3 = v 1 r. The above explanation of the rectilinear propagation of light is indeed unsatisfactory and is incorrect. etc.Huygens' Principle and Its Applications 12. .3 not have any amplitude at any point not enveloped by the wavefront.1 Refraction of a Plane Wave at a Plane Interface We will first derive the laws of refraction. 12. This fact is in direct contradiction to the original proposition of Huygens** according to which the secondary wavelets do S 12.1 mm or less.5. Let A 1 B 1 be a plane wavefront incident on the surface at an angle i. It can easily be seen that the incident and S' Fig.4 (a) A plane wavefront is incident on a pin hole.4. A 1 B 1 represents the position of the wavefront at an instant . ** Use of the Huygens' principle in determining the shape of the wavefront in anisotropic media will be discussed in Chapter 22. Let S1 S2 be a surface separating two media with different speeds of propagation of light v1 and v2 as shown in Fig. 12. B 1 B3 . It may be mentioned that if a plane wave is allowed to fall on a tiny hole.4(a) and (b)).4 (b) Diffraction of straight water waves when it passes through an opening (adapted from Ref. 6 * By a tiny hole we imply that the diameter of the hole should be of the order of 0. is known as the HuygensFresnel principle. (Note that the lines A 1A3. If the diameter of the pinhole is small (compared to the wavelength) the entire screen SS' will be illuminated. one does observe a finite intensity of light in the geometrical shadow. see also Fig. In the same time the light would have travelled a distance A 1A3 = v2r in the second medium. who postulated that the secondary wavelets mutually interfere. Further. 12. Fig. t = O. The Huygens' principle along with the fact that the secondary wavelets mutually interfere.* then the hole approximately acts as a point source and spherical waves emanate from it (see Figs 12. A satisfactory explanation was put forward by Fresnel. these represent rays in isotropic mediasee Chapter 4). 17 in the prelim pages. as pointed out earlier. Let z be the time taken for the wavefront to travel the distance B 1 B3 . are always normal to the wavefront. it can be explained satisfactorily on the basis of HuygensFresnel principle.4 APPLICATION OF HUYGENS' PRINCIPLE TO STUDY REFRACTION AND REFLECTION 12. however. as we will see in the chapter on diffraction.
33 1.53303 1. The envelope of these secondary wavelets is shown as A 3 C3 B 3 . when refracted into a rarer medium the wavelength and the speed of propagation will increase.890 x 105 cm) 1 Material n Material n Vacuum Air Water 1. when a wave gets refracted into a denser medium (vl > v2) the wavelength and the speed of propagation decrease but the frequency (= v/2.562816 x 105 cm 5. In Table 12.4 Optics If c represents the speed of light in free space then the ratio v (where v represents the speed of light in the particus lar medium) is called the refractive index n of the medium. Thus C 1 C2 = v1 T1 . 12. = 5. In order to determine the shape of the wavefront at the instant t = T we consider an arbitrary point CI on the wavefront. Table 12. thus. Similarly from the point AI we draw a secondary wavelet of radius v 2 T. sini _ B2 B3 /C2 B3 . From Fig.52704 1.5 Refraction of a plane wavefront A1 B I by a plane interface SIS2 separating two media with different velocities of propagation of light v1 and v2 (< v1). the angle of incidence is greater than the angle of refraction and consequently sini >1 sin r which implies vl > v2. Thus if n 1 (= s and n2= v2 are the refractive indices of vl J the two media.1 Refractive Indices of various Materials Relative to Vacuum (Adapted from Ref. A 2 C2B 2. Let the time taken for the disturbance to travel the distance C1 C2 be TI.T1) _ v2(T T1) yl Let A1 C1B1 . Clearly. Notice that r<i. Huygens' theory predicts that the speed of light in a rarer medium is greater than the speed of light in a denser medium.1 we have given the indices of refraction of several materials with respect to vacuum.46318 Note: The wavelengths specified at serial numbers 1.27 Refractive Indices of Telescope Crown Glass and Vitreous Quartz for Various Wavelengths (Adapted from Ref.45845 1. LB2 C2 B 3 = i (the angle of incidence) and LC2 B3 C3 = r (the angle of refraction). refracted rays make angles i and r with the normal. 12.B2 B3 C2 C3 sin r C2 C3 /C2 B3 _ vl ( Z .) remains the same.r1). see also Problern 10. In the rightangled triangles B 2 C2 B 3 and C3 C2 B3 .52 1. If 2'1 and 22 denote the wavelength of light in medium 1 and medium 2 respectively then.52441 1. the distance B IB2 (= B2B 3 = C1 C2) will be equal to 2. 1) (For light of wavelength 2.12. the prediction of the wave theory was indeed correct. 2 and 3 correspond roughly to the red. A 3 C3 B 3 and A4C4B4 denote the successive positions of crests. yellow and blue colours.66 which is known as the Snell's law.. This prediction is contradictory to that made by Newton's corpuscular theory (see Sec.5.861327 x 10`5 cm 1.2.46 1.889953 x 105 cm 4.54 1. i and r are the angles of incidence and refraction respectively. The shape of the wavefront at the intermediate time TI is shown as A 2 C2B 2 and clearly B 1 B2 = CI C2 v1 T1 and A IA 2 = v2 T1 . A2C2B 2 corresponds to the shape of the wavefront at an intermediate time TI. It is observed that when light travels from a rarer to a denser medium.45640 1. The table shows the accuracy with which the wavelengths and refractive indices can be measured. In Table 12.1 212 _ sini _ vl v2 sin r (3) or v1/21 = v 2/22 (4) Thus.0000 1. 7) Wavelength Telescope crown Vitreous quartz 1 _ 2 3 6.5 it is obvious that 2. then Snell's law can also be written as n1 sin i = n2 sin r (2) Fig.54 v2 (1) Quartz (crystalline) Quartz (fused) Rock salt Glass (ordinary crown) Glass (dense flint) 1. the wavelength dependence of the Table 12. .1 and the distance A IA 2 (= AZ A 3 = C2 C3 ) will be equal to 22. 2.0003 1.2) and as later experiments showed.From the point C2 we draw a secondary wavelet of radius v 2 ('r .
'r1 ). and since both the triangles are rightangled triangles. gives the critical angle.r1 )]. the light wave is incident on a denser medium. This is known as specular reflection. Fig.2 Total Internal Reflection In Fig. v2 > v 1 ). we draw a sphere of radius vet. then at a later time r the position of the wavefront would have been CB'. . 12.. 12. The three wavelengths correspond roughly to the red.4. reflected wavefront.4.= B1B2 . We draw a tangent plane on this sphere from the point B'. 12. In order to determine the shape of the reflected wavefront at the instant t= z. if the angle of incidence is such that v2a is greater than A 1B2 . Clearly. This corresponds to the case when v2 < v1. denotes the critical angle and n12 represents the re.4. 12. 12. when a plane wavefront gets reflected from a plane surface.5 the angle of incidence has been shown to be greater than the angle of refraction.vi =n 12. we do not have a well defined reflected wave. yellow and blue colours. From the point PI. what is fractive index of the second medium with respect to the first. If the mirror was not present.6 In the above we have considered the reflection of light from a smooth surface. Notice that the angle of refraction r is greater than the angle of incidence i. if the second medium is a rarer medium (i. If we consider triangles P2 P 1 B' and B 1P 1B' then the side P 1B' is common to both and since P 1 P ' = B 'B2. We consider the reflection of the plane wave and try to obtain the shape of the Fig. it can Refraction of a plane wavefront incident on a rarer medium (i. 12. Since BB1 = PP1 = v r1. we have the law of reflection.e. Al B2 v2 (5) C Fig.8) we have. total internal reflection. It is evident that one does not have a welldefined reflected beam. Indeed. i and r correspond to angles of incidence and reflection respectively. The former is the angle of reflection and the latter is the angle of incidence.surface is irregular (as shown in Fig. Notice the accuracy with which the wavelength and the refractive index can be measured.7. The critical angle will correspond to A1B2 = v2T Thus sini. we will have total ing from the irregular surface travel in many directions and internal reflection. then the refracted wavefront will be absent and we will have. LP2B'P 1 = LB 1 P1B '. v1 < v2) then the angle of refraction will be greater than the angle of incidence. we consider an arbitrary point P on the wavefront AB and let Z1 be the time taken by a disturbance to reach the point P1 from P.e. where BB' = PP' = AC = va and v is the speed of propagation of the wave.8 Diffuse reflection of a plane wavefront from a rough surface.7 Reflection of a plane wavefront AB incident on a plane mirror. If the where i.e.6. the angle of reflection is equal to the angle of incidence and the reflected wave is a plane wave.3 Reflection of a Plane Wave by a Plane Surface Let us consider a plane wave AB incident at an angle i on a plane mirror as shown in Fig. However.Huygens' Principle and Its Applications 12. 12. The value of i. Thus. i. The secondary wavelets emanatFor all angles of incidence greater than ic.5 refractive index for crown glass and vitreous quartz are given. 12. when r is equal to 7r/2. the distance B 1B ' will be equal to P 1P2 [= v(r.4 Diffuse Reflection 12. Let the position of the wavefront at t = 0 be AB. and a typical refracted wavefront would be of the form as shown in Fig. known as diffuse reflection. where B 1 B2 = vl T and A 1A2 = v2 a. A'B' is the reflected wavefront. 12. what is known as.
what are known as aberrations. in particular. the radius of the spherical wavefront from the point B. will not be a sphere. in general. The shape of the reflected wavefront is obtained by drawing a common tangent plane to all these spheres. If the time taken for the disturbance to traverse the distance QQ' be zl then. measured to the left of the point B. In order to derive a relation between u. Thus PB = u 12. Further. 12. 12.6 be shown that if the irregularity in the surface is considerably greater than the wavelength. Optics 12. (A 1 G)2 = GB x (2R . converge towards the point M and hence the point M represents the real image of the point P.10(b)]. we will have diffuse reflection. in order to determine the shape of the reflected wavefront. in general. are negative and all distances measured to the right of the point B are positive. In the absence of the spherical surface.10(b) the diameter B'OB intersects the chord A 1GC1 normally.12. the shape of the wavefront at a later time would have been A 1 B 1 C 1 where AA1 = QQ1 = BB1 = CCi = vz. Q being an arbitrary point on the wavefront. We consider an arbitrary point Q on the wavefront ABC and let it be the time taken for the disturbance to reach the point Q' (on the surface of the spherical wave).4. ABC is the incident wavefront (which is spherical and centred at P) and A 1 B'i C1 is the corresponding reflected wavefront (which is spherical and centered at P'). from the point B we have to draw a sphere of radius' vs. . Let the shape of the wavefront at the time t = 0 be ABC [see Fig. then (A i G)2 = 2R(GB) * The fact that the refracted wavefront is not. P' is the virtual image of P.4.6 Refraction of a Spherical Wave by a Spherical Surface Let us consider spherical waves (emanating from the point P) incident on the curved spherical surface SBS' . therefore. we draw a sphere. The spherical wavefront will. the shape of the wavefront at a later time would have been A 1B 1 Ci where AA1 = BB1 = CC1 = viz. we have BM=v and similarly. In order to determine the shape of the refracted wavefront at a later time z. which is equal to BB2 will be v2 z. Let ABC denote the shape of the wavefront at time t = O. in particular. v and R we use a theorem in geometry.. according to which.9 P is a point source placed'in front of a plane mirror MM'. It can immediately be seen that A 1B1C1 will have an exactly similar shape as A 1B1C1 except that A iBi Ci will have its center of curvature at the point P' where PB = BP'. We may draw similar spheres from other points on the spherical surface. 12. a sphere leads to. a small portion of any curved surface can be considered as a sphere and in this approximation we may consider A 1B2C1 to be a sphere whose center of curvature is at the point M.of radius v(zzi) whose center is at the point Q'. 12.5 Reflection of Light from a Point Source Near a Mirror Let us consider spherical waves (emanating from a point source P) incident on a plane minor MM'. BO=R where 0 represents the center of curvature of the spherical surface. In Fig. We adopt a sign convention in which all distances. In the absence of the mirror. we draw a sphere of radius v2 (r TO from the point Q'. The envelope of these spherical wavelets is shown as A 1B2C1 which.10(a)]. where u itself is a negative quantity. If GB <<R.GB) Fig. (6) where G is the foot of the perpendicular on the axis PM [see Fig. 12.9. as shown in Fig. which is shown as A 1B1'C1 in the figure. In a similar manner we can draw the secondary wavelets emanating from other points on the mirror and.* However. Let the refractive indices on the left and on the right of the spherical surface be ni and n2 respectively. thus QQ' vi z1 . since the point M lies on the right of B. Thus the reflected waves will appear to emanate from the point P' which will be the virtual image of the point P.
Thus u R which may be rewritten in the form n2 . 12. A converging spherical wavefront will propagate in a manner shown in Fig.11.ni u v R n2 v n2 .GB2) or ni (A1G)2 2R (A1G) 2 . 1j .Huygens' Principle and Its Applications 12. However. (A I G)2 = (2R . 0 is the center of curvature of SS'. much beyond the focal point the wavefronts again become spherical.n2 nl [ (AIG) 2 2R (A1G) 2 2v where we have used Eqs (7). Now BB1 = vlz and Therefore BBl BB2 vi . n2 > nl. (We are of course assuming that the second medium is a denser medium. (9) * Very close to the focal point.nl (b) ' Fig. Clearly.GB)GB = 2R(GB) we will obtain a virtual image. (A I G)2 = 2(u)GB I (8) Fig. = n2 V2 BB2 = v2 'r nl or n I BB 1 = n2BB2 or n 1 (BG + GB 1 ) = n2(BG .e. if lul > B n 2. Thus M is the real image of P.7 Since u is a negative quantity.B2M. if n2 < ni.* (7) where we have assumed GB << R. 8).n l R or IuI< Rn1 n2 . i. (8) and (9). In a similar manner.nl + n2 .nl (10) Thus. 12.. Similarly by considering the spherical surface A 1 B2C 1 (whose center is at the point M) we obtain (A1 G)2 = 2v(GB2) where v BM .11 Propagation of a converging spherical wave using Huygens' principle. 12. A 1 B2C 1 is the refracted wavefront. 12. which is approximately spherical and whose center of curvature is at M. (A1 G)2 is positive. (b) The diameter B'OB intersects the chord AIGCI normally. Beyond the focal point it will start diverging as shown in the figure.10(a)] whose radius is R. 2u . Consider the spherical surface SBS ' [see Fig. we will always have a virtual image).10 (a) Refraction of a spherical wave ABC (emanating from the point source P) by a convex spherical surface SBS' separating media of refractive indices n1 and n2 (> n 1 ). one has to use a more rigorous wave theory and the shape of the wavefront is very much different from spherical (see Ref.
14). n2 placed in a medium of refractive index n1.(n2 \ RI (14) or. This image now acts as an object to the spherical surface R2 on the left of which is the medium of refractive index n2 and on the right of which is the medium of refractive index nl. Thus. in a medium in which the velocity of propagation of the wave is v (< V). the relation derived is valid for any lens.) B=sin l (17) Using Eq.nl Rl (12) Solution. Thus no matter what the values of u and R may be. 12. then nl n2 . R1 is positive and R2 is negative and for a double concave lens.13). Thus from the point Po we draw a sphere of radius va. Let the speed of propagation of the wave in the medium be v. 12. for example. 12. Notice that we do not have to worry whether v' is positive or negative. it is automatically taken care of through the sign convention. P is the image (at a distance v from the point 0) of the point object P (at a distance u from the point 0). Example 12. 5.nl .: Let at t = 0. RI is negative and R2 is positive. The radii of curvatures of the two surfaces are R1 and R2. 12. We next consider the waves emanating Fig.8 In a similar manner we can consider the refraction of a spherical wave from a surface SBS ' shown in Fig. .12. P' is the virtual image of P. Let the radii of curvatures of the first and the second surface be RI and R2 respectively. Then n2 v' u nl . for a double convex lens. We assume a thin lens made of a material of refractive index n2 to be placed in a medium of refractive index nl (see Fig.12 Refraction of a spherical wave by a concave surface separating media of refractive indices n1 and n2 (> n1).n2 v v' R2 (13) Adding Eqs (12) and (13).6). if v is the distance of the final image point from 0. 1 where 1 . Further. (10) we can easily derive the thin lens formula.14 Generation of a shock wavefront by a vibrating particle Po moving with a speed V. Let v' be the distance of the image of the object P if the second surface were not present.1 Consider a vibrating source moving through a medium with a speed V. we obtain nl _ nl _ nl) 1 _ 1 Rz v u . Show that if V > v then a conical wavefront is set up whose halfangle is given by (. 12. Here the center of curvature will also lie on the left of the point B and both u and R will be negative quantities. all the distances are measured from the point 0).12 (n2 > n l ). the source be at the point Po moving with a speed V in the xdirection (see Fig.n2 . v will be negative and we will obtain a virtual image. Similarly it follows for other types of lenses (see Fig.13 A thin lens made of a medium of refractive index Fig.n2 n1 (1 f nl 1 1 f v u (15) 1 (16) Rl R2) Fig. We wish to find out the wavefront at a later time T. 12. The disturbance emanating from the point Po traverses a distance vrin time T. Optics (Since the lens is assumed to be thin.
12. The envelope of these wavelets gives the shape of the new wavefront.15 as A 2B2. A similar phenomenon also occurs when a charged particle (like an electron) moves in a medium with a speed greater than the speed of light in that medium. y).TO centered at P1. y) is the velocity of the wave at the point (x. we consider a medium whose refractive index decreases continuously from a given axis. 12. At should be small so that during this short interval the secondary wavelets may be assumed to be spherical. Let the source be at the position Q at the instant T. Let us try to determine the shape of the wavefront at a time At. is known as the Huygens Fresnel principle. each point of a wavefront is a source of secondary disturbance and the wavelets emanating from these points spread out in all directions with the speed of the wave. For the above construction to be valid. 12. y) = ni  y 2 (x2 + y2) (18) where the refractive index on the zaxis. 12.15. on the sphere whose origin is the point P1. where n represents the refractive index. a shock wavefront is always set up.5 HUYGENS' PRINCIPLE IN INHOMOGENEOUS MEDIA Huygens' principle can also be used to study the propagation of a wavefront in an inhomogeneous medium. which increases as x and y increase. ** See Sec. this is because of the Cerenkov radiation emitted by the fast moving electrons. which we define as the zaxis. For definiteness. Summary ♦ According to Huygens' principle. Then PoQ = VT We draw a tangent plane from the point Q. x and yaxis being the transverse axis. Thus it is evident that in` the present case the wavefront is getting focused.1 .15 The focusing of an incident plane wavefront in an inhomogeneous medium characterized by a refractive index variation given by Eq. At time position P1. 3. where v(x. Let the plane wavefront be incident along the zaxis as shown in Fig. Huygens' principle along with the fact that the secondary ♦ wavelets mutually interfere.Ti) PL = v PQ V and P1Q = V(T  T1) sing= (independent of TO Since 0 is independent of T1.15). you will find a blue glow coming out from it. A simple example is a Selfoc fiber. the speed of light will be about 2.4. y) At. (<T). If we again use the same procedure.Huygens' Principle and Its Applications from the source at a time T. It is interesting to point out that even when the source is not vibrating. centered at (x. Thus the radii of the spheres increase as we move away from the axis and if we draw a common tangent to all these spheres then the resulting wavefront is shown in Fig. PoP1 Tl 12. given that the wavefront at t = 0 is a plane wavefront A 1B1 (see Fig. It should be borne in mind that since we are considering an inhomogeneous medium.25 x 108 m/sec and the speed of the electron could be greater than this value. if its speed is greater than the speed of sound waves. (18) away from the axis. y). Since the refractive index decreases as x and y increase. we draw a sphere of radius v(T . This plane is known as the shock wavefront and propagates with a speed v. then the shape of the wavefront at time 2At (say) is shown as A 3B3. all the spheres drawn from any point on the line Po Q will have a common tangent plane.** whose refractive index variation is of the form n 2 (x. It is at once evident that the wavefront which was initially plane has now become curved. For example in water. 12. The speed of light in a medium will be equal to chi. Fig. the speed of the secondary wavelets emanating from portions of the incident wavefront will increase as we move nl is * This does not contradict the theory of relativity according to which no particle can have a speed greater than the speed of light in free space (= 3 x 108 m/sec). consequently.9 let the source be at the = VT1 In order to determine the shape of the wavefront at T. the refractive index varies continuously with position. Since P1L = v(T . If you ever see a swimming pool type reactor. * The emitted light is known as Cerenkov radiation. We will have to draw spheres of radius v(x.
PSSC.J DeWitte. Development of Concepts in Physics. White. Waves and Diffraction. construct the transmitted wavefront and show that the deviation produced by the prism is given by S=i+t . H. McGrawHill. 5. 1962. 1959.E. 8. 2. Mass. 3. Treatise on Light. D. Fundamentals of Optics. `Equivalence of Huygens' principle and Fermat's principle in ray geometry'. London. ♦ Using Huygens' principle one can derive the lens formula Optics of radius of curvature R and obtain the mirror equation 1+ v R u 12. A.. 293.C. 1969. 27. Born and E.B.12. F. 3rd Edn. Oxford University Press. America Journal of Physics. Baker and E. A.10 ♦ Laws of reflection and Snell's law of refraction can be derived using Huygens' principle.J. Physics. Dover Publications. 465. C.2 Consider a plane wave incident obliquely on the face of a prism. Mass.A. London. Braddick. 4. 1965. 6. The Mathematical Theory of Huygens' Principle. 1965.B.. 1965.A Problems 12. M. 7.J.. 1957. Vol. Oxford. Principle of Optics. Using Huygens' principle.J. Wolf. Boston. Arons. REFERENCES AND SUGGESTED READINGS 1. Reading. McGrawHill Publishing Co:. Pergamon Press. Vibration. i and t are the angles of incidence and transmittance. Jenkins and H. Huygens. B. Heath and Company.. Copson. .1 Use Huygens' principle to study the reflection of a spherical wave emanating from a point on the axis at a concave mirror where A is the angle of the prism. p. 1975. AddisonWesley Publishing Co. New York.
`:. Starting with the Young's double hole in. instrument for which Michelson received the Nobel Prize in Physics in 1907.Perot interferometer which is based on multiple beam interference and are characterized by a high resolving power and hence find applications in high resolution spectroscopy. optical beats and Fourier transform spectroscopy have also been discussed.PART 3 te rfe re n This part commences with the principle of superposition of . the various chapters discuss many i teresting experiments associated with the area of interference. Chapter 17 discusses the basic concept of temporal and spatial coherence. The ingenious experiment of Michelson (which used the concept of spatial coherence to determine the angular diameter of stars). .Waves which is the bask physics behind all interference experiments. In Chapter 15 the Michelson Interferometer is discussed in detail which is perhaps one of the most ingenious and sensational optical. Chapter 16 discusses the Fabry . terference experiment.
.
whereas the solid curve shows the resultant displacementobtained by algebraic addition of each displacement. a triangular pulse is generated which propagates to the right with a certain speed v.1).1(c) represents a snapshot at an instant when the two pulses interfere. this pulse would have propagated in the +xdirection without any change in shape.1 The propogation in opposite directions of two triangular pulses in a stretched string. The solid line gives the actual shape of the string. 1).2 STATIONARY WAVES ON A STRING Consider a string which is fixed at the point A (see Fig. Figure 13. the dashed curves represent the profile of the string if each of the impulses was moving all by itself. 13. At a little later time each pulse moves close to the other as shown in Fig. radiant heat. (a).2). happened. 13. eminently adapted to remove any doubt as to the identity of light. The phenomenon of interference contains no more physics than embodied in the above example. (As has been shown in Sec. the snapshot of the string is shown in Fig.1(d)] the two pulses exactly overlap each other and the resultant displacement is zero everywhere (where has the energy gone?). Shortly later [Fig.1(a).6 the speed of the wave is determined by the ratio of the tension in the string to its mass per unit length. At a much later time the impulses sort of cross each other [Fig.) At t = 0. From the end A. In the following sections we will consider some more examples. 13. 13. at any rate. without any interference. 13. we are. and electromagnetic wave motion. Heinrich Hertz (1888)* 13. (b). (c). we consider a long stretched string AB (see Fig. (d) and (e) correspond to different instants of time. As a simple example.1(b). In the absence of any other disturbance. of course. . 11. 13. We next assume that from the end B an identical pulse is generated which starts moving to the left with the same speed v. I believe that from now on we shall have greater confidence in making use of the advantages which this identity enables us to derive both in the study of optics and of electricity.1 INTRODUCTION In this chapter we will discuss the applications of the principle of superposition of waves according to which the resultant displacement (at a particular point) produced by a number of waves is the vector sum of the displacements produced by each one of the disturbances.The experiments described appear to me. 13. neglecting any attenuation or distortion of the pulse.1(e)] and move as if nothing had A//\ (a) A (b) A (c) A (d) Nye B B B A(e) / B Fig. 13. This is a characteristic feature of superposition of waves. A transverse sinusoidal wave is sent down * The author found this quotation in the book by Smith and King (Ref.
The displacement at any point on the string due to this wave would be given by y . at A 3A. 13. The corresponding frequency (= 2. 2 . the string along the xdirection.13. In Fig. T represents a transmitter of electromagnetic waves (the wavelength of which may be of x=o = a sin (2 7r v t) (3) where v = v/2 and we have assumed the point A to correspond to x = 0. . thus we may write yi = a sin [ (x + vt)] (2) = 2 p(2a) 2 co 2 cos2 cot = 2pa2 oi cos2 cot (10) = a sin [22r (x + v t)] Thus. P. (4) where the subscript r refers to the fact that we are considering the reflected wave.2 Reflection of a wave at x = 0. If the frequency is changed. .4.e. 13... Q and R in Fig. the displacement at the point A would have been yt where co = 2irv is the angular frequency and p the mass per unit length of the string. We can also carry out a similar experiment for electromagnetic waves.3 An arrangement for studying standing electromagnetic waves. R represents a reflector which may be a highly polished metal surface and D represents the detector which can measure the variation of the intensity of the electromagnetic waves at different points. Two consecutive maxima are separated by about 5. = a sin [ (x + vt) + ct] (1) the amplitude of the vibration is maximum. the xcoordinates of the nodes are given by T x=0 (7) the order of few cm). 2A.a sin (2 7C v t) D R Fig.: yr x=o = .2. one can observe the change in the distance between the antinodes.3. cm.6 x 109 sec1) can easily be generated in the laboratory. 13. One may approximately assume plane waves to be incident on the reflector.. x=0 Fig.8 cm. One should notice that the minima do not . because of the incident wave.4 Optics x = 0.. The displacements at these points (which are knowns as antinodes) are given by y = ± 2a cos 2 tc v t (9) At the antinodes the kinetic energy density would be given by (see Sec. Since the reflected wave propagates in the +xdirection we must have yr=+ a sin27c(i _ Vt) (5) The resultant displacement would be given by y = y= + yr = a [sin22r(+vt)+sin22r(. 5A. The result of a typical experiment is shown in Fig. 7._ vt)] = 2a sin It should be seen that for values of x such that sin T x cos 2 tc vt (6) the displacement y is zero at all times. A. (8) and are marked as points A.13. . there must be a reflected wave such that the displacement due to this reflected wave (at the point A) is equal and opposite to y. Such points are known as nodes.2) Kinetic energy/unit length where the subscript i refers to the fact that we are considering the incident wave. One can see the periodic variation of intensity.6 . 13. i. Without any loss of generality we can set 0 = 0. . the incident and reflected waves interfere and produce nodes and antinodes. Since the point A is fixed. x=4. thus A. The nodes are separated by a distance A/2 and at the midpoint between two consecutive nodes. 13.
3. When A.. Here E represents the electric field associated with the electromagnetic wave.3 STATIONARY WAVES ON A STRING WHOSE ENDS ARE FIXED 30 40 50 60 70 80 Distance from reflecting plane. if a string of length L is clamped at both ends (as in a sonometer wire) then it can only vibrate with certain well defined wavelengths. therefore.lt + fan ) L 2L = all: {sin [27r ( f + vt)] + sin [27r I f .4 A typical variation of the intensity between the reflector and the transmitter [adapted from Ref.2. Similarly when A.e. one can introduce a coefficient of reflection (r) which is defined as the ratioof the energy of the reflected beam to the energy of the incident 'beam. 13. This is because of the fact that the incident wave is really not a plane wave* and that the reflection is not really perfect. . sin (2c LI =0=sinn7r (16) = a sin [27C (1. * A plane wave is obtained by a point source at a very large distance from the point of observation (see Chapter 11).5 Standing waves on a stretched string clamped at both ends.Superposition of Waves 8 N 6 aci 4 C _ 2 0 10 20 13. 2]. (6)]: y =2a sin (22. t) = n=1 an sin x cos (27r v. n = 1) the string is said to vibrate in its fundamental mode [Fig.. cm Fig. 2.+ vt I] (11) or 2L n=1.. n The corresponding frequencies are A=fin= (12) v" (17) then the reflected wave would be given by Ereflected = atI Sin [21r I vt)] where the plane x = 0 corresponds to the plane of the reflector. x) cos(2Irvt) (14) If the other end of the string (say at x = L) is also fixed. = 2L/2 and 2L/3 the string is said to vibrate in its first and second harmonic.2. while discussing the stationary waves on a string we had assumed only one end of the string (x = 0) to be fixed. then we must have 1 2a sin L) cos (2 7r vt) = 0 (15) (Ili Eq. Thus the ratio of the amplitudes would be and if the incident wave ' is given by Eincident In Sec. 3. 13. really correspond to zero intensity and the intensities at the maxima are not constant. = 2L (i. (18) 2L' n A../ sin +a(1 f + vt )] (Y xJ cos 27r vt (a) 1=L VT' ) sin [27c (i+ (c) The first term represents the stationary component of the wave and the second term (which is small if r is close to unity) represents the progressive part of the beam. In fact. . Thus the resultant field would be given by Eresultant = Eincident + Ereflected = a sin [21r( + vt)] + a^ sin [27r ( f . In general. 13. and the resultant displacement was shown to be given by [see Eq..5(a)].. (15) is to be valid at all times. if the string is plucked and then made to vibrate then the displacement would be given by y(x.vt)] }  (Pn) (19) +a (1. 13. Fig. n Thus./) sin [27r I = 2a.vt)] = 1.5 13.
13.7 The experimental arrangement of Wiener for studying stationary light waves. one could have superposition of displacements which are in different directions. 13. The energy density is maximum at the antinodes and minimum at nodes. they may have different amplitudes and different initial phases. 13. . The distances between two successive antinodes and successive nodes are A/2.5 SUPERPOSITION OF TWO SINUSOIDAL WAVES Let us consider the superposition of two sinusoidal waves (having the same frequency*) at a particular point. however. for example.13.6 The experimental arrangement of Ives for studying stationary light waves. in general. * Indeed in Sec. For A. Wiener overcame this difficulty by placing the photographic film at a small angle and thereby increasing considerably the distance between the dark (or bright) bands (Fig.25 mm 2x103 On the other hand. the angle between the film and the mirror was about 103 radians. In the experimental arrangement of Ives. 5 x 10 5 cm). distance between two consecutive dark bands (which is equal to A/2) one can calculate the wavelength. The cut section was viewed under a microscope and bright and dark bands (separated by regular intervals)) were observed. the superposition of two linearly polarized waves to produce a circularly polarized wave (see Chapter 22).2. t = 0) and Optics 0„ are determined by the values of these are known as the initial condit=o tions. Because of the small wavelength of light. = 5 x 10 5 cm what would be the distance between two consecutive dark bands? Solution: The required distance is 2a _ 5 x 105 cm = 0. A more detailed discussion on the vibration of stretched strings has been given in Sec. 8. Example 13. 13. we had. (10)].5 x 104 mm. By measuring the Incident Light Beam Mirror Fig. 13. Let and x1 (t)=a 1 cos (cot + el) x2 (t) = a2 cos (wt + e2 ) (20) Glass Emulsion Mercury Fig.7). difficult to measure. therefore. according to the superposition principle the resultant displacement x(t) would be given by * In Chapter 17 we will consider the superposition of waves having nearly equal frequencies which leads to the phenomenon of beats. This is because of the fact that light wavelengths are extremely small (=. The beam was reflected on the mercury surface and the incident wave interfered with the reflected wave forming standing waves. in the set up of Ives the distance would be 2. while discussing stationary waves on a string. Now. at a particular value of x. However. A section of the photographic film was cut along a plane normal to the surface. A parallel beam of monochromatic light was allowed to fall normally on the glass plate. the emulsion side of a photographic plate was placed in contact with a film of mercury as shown in Fig. It should be seen that when a string is vibrating in a particular mode there is no net transfer of energy although each element of the string is associated with a certain energy density [see Eq. represent the displacements produced by each of the disturbances: we are assuming that the displacements are in the same direction*. 13. at . the distance between two consecutive dark (or bright) bands was extremely small and was.1 In a typical experimental arrangement of Wiener.4 STATIONARY LIGHT WAVES: IVES AND WIENER'S EXPERIMENTS It is difficult to carry out experiments in which one obtains stationary light waves. 13.2.6 where the constants a„ and y(x.6. two sinusoidal waves of the same frequency (but having different initial phases) superposing on each other.
a2 cos 02 (35) consistentwith Eqs (23) and (24). if 01 e2 =^c.. this is known as constructive interference.8). If we square and add Eqs (23) and (24).) and destructive interference occurs at x = 0. (i.e. A. We use the law of parallelograms to find the resultant OR of the vectors OP and OQ . Q. then the resultant amplitude will be the sum of the two amplitudes.. P. 2. there is no violation of the principle of conservation of energy.3tt.. We draw a circle of radius a1 and let the point P on the circle be such that OP makes an angle 01 with the xaxis** (see Fig. then the initial phase of the resultant will be 0. This can be easily seen by noting that OR cos 0 = OP cos 0 1 +PRcos02 = a1 cos Similarly.a2 (30) and the resultant amplitude is the difference of the two amplitudes.. (i.7 In general. as the vectors OP and OQ rotate on the circumference of the circles of radii a1 * In Chapter 14... 9 2 = 0. ... then we can see that constructive interference 3 5 occurs at x = 4 .. We next draw a circle of radius a2 and let the point Q on the circle be such that OQ makes an angle 02 with the xaxis...the energy is merely redistributed. This is known as destructive interference. (29) then a = a1 .. This method is particularly useful when we have a large number of superposing waves as it indeed happens when we consider the phenomenon of diffraction...e. at points A.... at the points P' Q' R'. (26). 4 .. however.* It may be mentioned that when constructive and destructive interferences occur.. 3212.. we will study the interference pattern produced by the superposition of spherical waves emanating from two point sources.. + an cos On xn = a.. (25) we find that if 01 .57r. if we have n displacements (21) x1 = al cos (cot + 01) x2 = a2 cos (cot + 92 ) = al cos (cot + 91) + a2 cos (cot + 02) which can be written in the form x(t) = a cos (cot + 0) (31) (22) (23) (24) and a sin 0 = a1 sin 91 + .2. (27) then a = a1 + a2 (28) Thus. sin 0„ (34) then x=xi +x2+.02 )]1'2 (25) Further tan 0 = a1 sin 91 + a2 sin 02 a1 cos 01 + a2 cos 02 (26) 13.).+x„=acos (cot +9) where a cos 9 = al cos 9l + . cos (cot + On ) where a cos 9 = al cos 01 + a2 cos 02 and a sin 9 = a1 sin 01 + a2 sin 02 (32) (33) Thus the resultant disturbance is also simple harmonic in character having the same frequency but different amplitude and different initial phase. The length of the vector OR will represent the amplitude of the resultant displacement and if 0 is the angle that OR makes with the xaxis. (20) using the graphical method... then cos 0 and sin 0 can be determined from Eqs (23) and (24) which will uniquely determine O.6 THE GRAPHICAL METHOD FOR STUDYING SUPERPOSITION OF SINUSOIDAL WAVES In this section we will discuss the graphical method for adding displacements of the same frequency..r 47c. Similarly.Superposition of Waves x(t) = x1 (t) + x2(t) 13. we would obtain a = [ai + a2 + 2a1a2 cos (91 . 13. R.. if we assume the vector OP to rotate (in the anticlockwise direction) with angular velocity w then the xcoordinate of the vector OP will be a1 cos (wt + 01) where t = 0 corresponds to the instant when the rotating vector is at the point P. Let us first try to obtain the resuliant of the two displacements given by Eq.. + a. Further.. 212. . OR sin 0 = a1 sin 01 + a2 sin 02 (36) 91 + It should be pointed out that 0 is not uniquely determined from Eq. 4 . if we assume a to be always positive. From Eq. 13. ** Clearly.. . If we refer to Fig. if the two displacements are in phase.
13.1)00] .. P3 . 13.. from the tip of this vector we must draw another vector ( PR ) of length a2making an angle 02 with the axis. P1. Thus. Therefore. therefore.. (20) then we must first draw a. the vector OR' will represent the resultant of xi.I Optics In Fig. if we wish to find the resultant of the two displacements given by Eq.8 The graphical method for determining the resultant of two simple harmonic motions along the same direction and having the same frequency..9 The graphical method for determining the resultant of N simple harmonic motions along the same direction and having the same frequency...(38) Fig. sin 00 /2 = 2LO sin and OPN = 20C NBo a sin NO0 2 80 sin (39) Fig. x2. PI P2 . and a2. It can be easily seen that if we have a third displacement x3 = a3 cos (cot + 03) (37) then from the point R we must draw a vector RR' of length a3 which makes an angle 03 with the axis. As an illustration of the above procedure we consider the resultant of N simple harmonic motions all having the same amplitude and with their phases increasing in arithmetic progression.13. The length of the vector OR will represent the resultant amplitude and the angle that it makes with the axis will represent the initial phase of the resultant displacement. The resultant is denoted by the vector OPN . PN will lie on the circumference of a circle whose centre is L and radius is LO. Thus x1 =acos cot x2 = a cos [cot + 00] + (N . xN = a cos [cot . the points 0. P2. x2 and x3. the vector OR rotates on the circumference of the circle of radius OR with the same frequency. respectively. L0 a12 7C 200 and.8 I . It is easy to prove that ALQ 1P 1 ° z LQ 2P 1 Thus LO = LP1 = LP2 . Let Q1L and Q2L be the perpendicular bisectors of OP1 and P 1 P2. 13. x3 . P2 P3 . LLP I O = LOLP1 = 00. . Further.. Thus..9 the vectors OPI . correspond to x1.vector ( OP) of length a1 making an angle 01 with the axis.
. [Ans: . Assume = 6 x 105 cm. the phase of the resultant displacement would be LPNOX = Thus a cos cot+acos(cot +0°)+. While using the complex representation.. ae iNeo/2 i't e e ieo/2 eiNOo/2 .) denotes the `real part of' the quantity inside the brackets. one must be careful in calculating the intensity of a wave which is proportional to the square of the amplitude.. (46). An interesting illustration of the usefulness of this method is to consider the resultant of the N displacements described by Eq. then the resultant amplitude will be the sum of the two amplitudes. Summary ♦ According to the principle of superposition of waves.e ie0 /2 13. It may be noted that whereas Re (xi ) + Re (x2 ) = Re (xi + x2) (Re xi )(Re x2) ^ Re (xix2) (42) 13.1) 0° (40) a sin sin NO° Oa 2 r exp I i {cot + (N 1) ° (47) L N00 2 (41) sin 8° which is consistent with Eq. calculate the frequency of the vibrations. one must calculate the amplitude first and then the intensity. whatwill. we would obtain Eqs (23) and (24). Further...3). aeiwt [1 + e ie0 + e2ieo + . The length of the string is 50 cm and xi = al cos (cot + Oi) is written as xi = aie`(uu+ei) (43) (44) where it is implied that the actual displacement is the real part of xi. (38).1/4°] 13. The length of the string is 30 cm and it vibrates in 3 loops. ♦ If the two displacements (produced by two sinusoidal waves) are in phase. if the string is made to vibrate in its fundamentalmode. 10. Problems 13.2 In Problem 13.1 Standing waves are formed on a stretched string under tension of 1 Newton.1. this is known as constructive interference..3 In the experimental arrangement of Wiener. The complex representation is also very useful in considering the spreading of a wave packet (see Sec. Thus.Superposition of Waves Further. this is known as destructive interference. + e `(Ni)eo ] = ae imt 1e Ni° ° . 13. ♦ The stationary waves on a string and the formation of standing electromagnetic waves are formed by the superposition of waves traveling in opposite direction. the resultant displacement (at a particular poin roduced by a number of waves is the vector sum of the isplacements produced by each one of the disturbances.7 THE COMPLEX REPRESENTATION Often it is more convenient to use the complex representation in which the displacement where Re (. then the resultant amplitude will be the difference of the two amplitudes. If the mass per unit length of the wire is 10 mg/cm. On the other hand..9 2 (N .+acos[Cot+(N1)0°] = A cos (cot + 0) where a sin A= and but 0= 2(N1)0° We will use this result in Chapter 18. if x2 =ate i(wr + 92) then xi + x2 = (a i ei °l + a2eie2)eirur = ae i(wr+ e) where aeie = ai e`B1 + a2ei82 (45) (46) If we equate the real and imaginary parts of Eq. (40). what should be the angle between the film and the mirror if the distance between two consecutive dark bands is 7 x 10 3 cm.be the frequency of vibration? 13.e iNOo/2 e ieo/2 ..4 Standing waves with five loops are produced on a stretched string under tension. if the two displacements are r out of phase. Thus we write x1 = aeit t+ x2 = ae i(wr + e°) Hence x=xi +x2 +.
x2] u2 and [R2 .5 The displacements associated with two waves (propagating in the same direction) having same amplitude but slightly different frequencies can be written in the form a cos2ir (Vt  Optics 13. Av = 2 sec l and a = 0. Plot the time variation of the intensity at x = 0.10 1 the frequency of vibrations is 250 sec . 13. 5 cm.9 In Example 11. Calculate the time variation of the displacement of the points which are at distances of 2 cm.5 assume v = 330 m/sec. REFERENCES AND SUGGESTED READINGS See at the end of Chapter 14.2. Plot the resultant disturbance at t = R/v.6.8 Discuss the superposition of two plane waves (of the same frequency and propagating in the same direction) as a function of the phase difference between them. see Sec.1 cm. 18 cm.6 In Problem 13.5R/v and 10R/v.(x . 35 cm and 45 cm from one end of the string. 13. Consider two semicircular pulses propagating in opposite directions.5R/v. A and A . v = 256 sec 1. 7. (Such a situation indeed arises when a plane wave gets reflected at the upper and lower surfaces of a glass slab. At t = 0.) Discuss the superposition of the displacements and show that at a particular value of x.13. . 2.10R)2] 1" (Such displacements are indeed obtained when we have two tuning forks with slightly different frequencies.1 we had discussed the propagation of a semicircular pulse on a string. 15 cm. 20 cm.) 13. the intensity will vary with time.7 Use the complex representation to study the time variation of the resultant displacement at x = 0 in Problems 13. the displacement associated with the pulses propagating in the +x and in the x directions are given by [R2 . where v denotes the speed of propagation of the wave. an a cos 2tr ((v + Ov) t  13. 5R/v. 15.5 and 13. respectively.
whenever two waves superpose. However. In this chapter.1 INTRODUCTION In the previous chapter.* although the interference does take place (see Sec. Under the first category.. 13. the first clear proof that light added to light can produce darkness. a beam is allowed to fall on two closely spaced holes and the two beams emanating from the holes interfere. In general. This phenomenon is called interference..`The wave nature of light was demonstrated convincingly for the first time in 1801 by Thomas Young by a wonderfully simple experiment. . this is known as multiple beam interferometry and will be discussed in Chapter 16. Thus.5). It will be shown that multiple beam interferometry offers some unique advantages over two beam interferometry. this time with spirit flame as light source. one cannot observe interference between the waves from two independent sources. * It is difficult to observe the interference pattern even with two laser beams unless they are phase locked. This will be discussed in the next chapter. From his discoveries in medicine and science.He let a ray of sunlight into a dark room. December 11. The methods to achieve this can be classified under two broad categories. we had considered the superposition of onedimensional waves propagating on a string and had shown that there is a variation of energy density along the length of the string due to the interference of two waves (see Fig. 14. a beam is divided at two or more reflecting surfaces and the reflected beams interfere. namely the superposition principle. Helmholtz concluded: `His was one of the most profound minds that the world has ever seen. we will consider the interference pattern produced by waves emanating from two point sources. which gave him sufficient encouragement to repeat the experiment.. 1971 Thomas Young had amazing broad interests and talents . This time he saw a number of dark lines. This method is known as division of wavefront and will be discussed in detail in this chapter. From the Internet 14. In the other method. one tries to derive interfering waves from a single wave so that the phase relationship is maintained.. Thomas Young had expected it because he believed in the wave theory of light. emphasize that the present and the following chapters are based on one underlying principle. for light waves. regularly spaced.4). Dennis Gabor in his Nobel Lecture. known as division of amplitude. pierced with two small pinholes.. placed a dark screen in front of it. however. We must. with a little salt in it. He then saw two darkish lines at both sides of a bright line. It may be mentioned that it is also possible to observe interference using multiple beams. to produce the bright yellow sodium light. this is also_ the case for microwaves. one obtains an intensity distribution which is known as the interference pattern. It may be mentioned that with sound waves the interference pattern can be observed without much difficulty because the two interfering waves maintain a constant phase relationship. in a typical arrangement. and beyond this. at some distance a white screen. due to the very process of emission.
Thus.. S1 and S2 go up simultaneously. then waves emanating from St will nterfere with the waves emanating from S2.roduce a crest at a distance p from S2.1). at an arbitrary point A (on ae perpendicular bisector) we may write the resultant disturance as Y =YI + Y2 =I= S1B = A/2 (2) At such a point the disturbance reaching from the source SI will always be out of phase with the disturbance reaching from S2.1 Waves emanating from two point sources S1 and S2 vibrating in phase. where the solid curves represent (at a )articular instant) the positions of the crests due to distur.1. The solid and the dashed curves represent the positions of the crests and troughs respectively. i. Consequently. if the distances involved are large (in comparison to the wavelength). 14. Similarly. Next. Notice that at all points on the perpendicular bisecOY the disturbances reaching from SI and from S2 will [ways be in phase. We may have. However. We assume that he needle at S2 vibrates in phase with the needle at St. This is also obvious from Eq. if at a certain instant. let us consider a point B such that S2B  14. This follows from the fact that the disturbance reaching the point B from the source S2 must have started half a period (= T/2) earlier than the disturbance reaching B from SI. It may be pointed out that the amplitudes of the two vibrations reaching the point B will not really be equal as it is at different distances from SI and S2. we will. etc. Such a point corresponds to destructive interference and is known as a node and corresponds to minimum intensity. assume water waves to produce iisplacelnents which are transverse to the direction of propalation.e. the two amplitudes will be very nearly equal and the resultant intensity will be very nearly zero. 14. Consequently. similarly for the vibrating needle at S2. if the displacement at B due to S1 is given by y1 = a cos cot then the displacement at B due to S2 would be given by y2 =acos(cot 7c)_a cos cot and the resultant y = yt + y2 is zero at all times. the listurbance emanating from the source SI produced a crest at distance p from SI then the disturbance from S2 would also . However.2 Optics where yl (= a cos cot) and y2 (= a cos cot) represent the displacements at the point A due to SI and S2 respectively.ances emanating from SI and S2. This is explicitly hown in Fig. (1). We see that the amplitude at A is twice the amplitude produced by each one of the source. if both ieedles are vibrating. i'or the sake of simplicity. the dashed . for example. Although water waves are not really transverse. two sharp needles vibrating up and down at the points SI and S2 (see Fig. It should be noted that at t = T the displacements produced at the point A 4 ( 4v 2co by each of the source would be zero and the resultant will also he zero.ertain frequency v then circular ripples would have spread )ut from the point SI.2 INTERFERENCE PATTERN PRODUCED ON THE SURFACE OF WATER We consider surface waves emanating from two point sources in a water tank. The wavelength would have been i/v and the crests and troughs would have moved outwards. In a similar manner we may consider a point C such that S2CSIC=) where the phase of the vibrations (reaching from SI and S2) are exactly the same as at the point A. 14. Consequently we will = 2a cos cot (1) Fig. they also reach the lowest posiion at the same time. .urves represent (at the same instant) the positions of the roughs. If there was only one needle (say at St) vibrating with a .14.
if a point P is such that S2P . we have a minimum.2 The actual interference pattern produced from two point sources vibrating in phase in (After Ref. 14. 1. Show that the equation of a nodal line is a hyperbola. Solution: For the sake of generality we find the locus of the point P which satisfies the following equation: SIPS2 P =A (5) Solution. The actual interference pattern produced from two point sources vibrating in phase in a ripple tank is shown in Fig. Example 14.. Also obtain the locus of points which correspond to maxima. Consider a point P such that S2P . a ripple tank .r ) cos ?I 73 acos (cot 3I The intensity is therefore 1/4th of the inte//nsity at the maxima. we have a maximum and if A = (n + 4) A. Example 14. (3) nor Eq. 2. In a similar manner one can calculate the intensity at any other point. (4) will neither be a maximum nor zero. Find the ratio of the intensity at the point P to that at a maximum.Two Beam Interference by Division of Wavefi•ont again have constructive interference. then the disturbances reaching the point P from the two sources will be in phase. In general..2.1 The intensity at the point which neither satisfies Eq. (maxima) (3) 14. by If the disturbance reaching the point P from SI is given yl = a cos cot then the disturbance from S2 would be given by y2 = a cos (cot  23 1 Thus.SIP = X113. the interference will be destructive and the intensity will be minimum. 14. if the point P is such that S2P = a [cos cot + cos (wt  23 \l IJ SIP = (n + 2 J 2. On the other hand. We choose the midpoint of SIS2 as the origin.2 The locus of points which correspond to minima are known as nodal lines.3 because a path difference of A1/3 corresponds to a phase difference of 23 Thus the resultant displacement would be y =yi+y2 n = 0. with the xaxis along S 1S2 and the yaxis perpendicular to it Fig. 9... if A = n2. (minima) (4) = 2a cos (cot . used with permission). then the disturbances reaching the point P from the two sources will be out of phase. the interference will be constructive and the intensity will be maximum.S1P = n2.
Assume D >>> A.S2P > d (S1 P .. the needle at S1 produces a crest at a distance R from it them the needle at S2 would produce a trough at a distance R from S2.S2P equals d on the xaxis only). Thus the points corresponding to minima will be equally spaced with a spacing of A. Generalize the result for an arbitrary phase difference between the vibrations of the two needles.1) to vibrate in phase. minima will occur (on the line y = D) would be given by y2 _ =1 A2 ) Y X02 4(d 2 (6) xn 2 = ± r d2 ^n 2 l An D I\ An _+ d 2 1.3).Clearly the points at which. Let the coordinates of the point P be (x. Example 14.3). 0) and (+1. If the distance between S1 and S2 is d. then the coordinates of the points S1 and S2 are (1.4) the two vibrations will always be out of phase and x Fig. Therefore. y). 0) respectively. 14. 1 1/2 // \\2 S 1 PS2P=^I x+2 +y 2 J J rr [(x  2 + 2 1/2 It may be pointed out that there is no point P for which S1P . n = 0. the resulting equation is an ellipse which we know is impossible. 2. For large values of x and y the curves asymptotically tend to the straight lines Y= (d2 . 1.. Thus if.A2 ) 1/2 02 + )2 + y2 1 and \2 S2P = [(X 1/2 2) + y2 Therefore.d^2 1/2 D ^ // y= D I n+ \ 2) d (10) where we have assumed An << d. 14. 14.4 Till now we have assumed the needles at .14. the curves correspond to minima and when A = nA.A2) 112 A2 "x (9) g =4A2 [x2_xd+^+Y2] where An = (n+ 2) A.4xdA + 2 2A [x 2) +y2 Further at large distances from the origin the equation of the nodal lines would be Y= I d2 . Solution: The two needles Si and S2 vibrate out of phase. (6) also repi sents the locus of all those points for which S1P + S2P = A and obviously in this case A can exceed d..2+(x Example 14. 14. Eq. + l2 1/2 +y2 Solution: The equation of this line would be Y=D (8) y 2 +2D I x 2) J or // l2 1/2 2xdO2 = On squaring. Assume now that the needles vibrate with a phase difference of 2V and obtain the nodal lines. Now.S1 and S2 (see Fig. .3 The nodal curves. we obtain 4x2d2 . The fallacy is a result Of the fact that because of a few squaring operations. _ 2] 2 ] = A or (x+f)2+y2 = 6.4 (see Fig. 14. Then 1/2 (x Optics which is the equation of a hyperbola. Find the points on this line where minimum intensity will occur. at all points on the perpendicular bisector OY (see Fig. (6) that when A > d.3' Consider a line parallel to the xaxis at a 2) distance D from the origin (see Fig.. 1/2 Thus we obtain SIP = X. When A = (n + 2) A. at any instant. it appears from Eq.DId. the curves correspond to maxima.
SIP = nA..e.001 sec.. then the interference pattern will keep on changing.n=0. The positions of the maxima and minima will. the observed intensity will be given by I=410 (cos2 . when S2P . If the phase where (. . for a camera with exposure time 0. for example. that there is a phase difference of 2r/3. i. 0) is changing with time. Thus. depend on the phase difference in the vibration_of the two needles. Clearly. then the resultant displacement would be y2 = 2a cos 0/2 cos (wt + 0/2) (12) The intensity (1) which is proportional to the square of. However. sometimes vibrating out of phase. Assume. Clearly if = ± ir ±32r. z = 0. Notice that one again obtains a stationary interference pattern with nodal lines as hyperbolae..S1B = A/2 which the two vibrations will be in phase and we will have maximum. if the phase difference between the sources SI and S2 (i...e. we will have minimum. We next assume that the two needles are sometimes vibrating in phase.... Let the displacement produced by the sources at SI and S2 be given by = a cos wt y2= a cos (an+0)Jr y1 2 I A. ±4 n.) denotes the time average of the quantity inside the angular brackets. (maxima) y = yl + the interference will be constructive and we will have maxima. The above analysis can easily be generalized for arbitrary phase difference between the two needles..5 Fig.f (t)) =1 ff (t) dt (15) a/2 where 'r represents the time over which the averaging is carried out. the resultant intensity will be zero and we will have minima..the amplitude can be written in the form I = 410 cos2O/2 (13) where I0 is the intensity produced by each one of the source individually. the time average of a time dependent function is defined by the following relation: +r/2 (.. if the interference pattern is viewed by a normal eye. because of the initial phase difference of n the conditions for maxima and minima are reversed. at the point satisfies the equation S2B . if there is a crest at a distance R from SI then there is a crest at a distance R . when = 0.+ will correspond to maxima. ±2 tr..±1. the condition SIP S2P=nA.Two Beam Interference by Division of Wavefront 14..±2.. etc. the intensity will be maximum (= 410).A/6 from S2. this averaging will be over about 1/10th of a second. however. sometimes vibrating with a phase difference of 2r/3. etc. and when S2P . if 0 varies in a random manner in . On the other hand.001 sec. (minima) the interference will be destructive and we will have minima. On the other hand. i. For example. Consequently.4 Waves emanating from two point sources St and B S2 vibrating out of phase..SIP = (n + difference changes with such great rapidity that a stationary interference cannot be observed then the sources are said to be incoherent.3 COHERENCE From the above examples we find that whenever the two needles vibrate with a constant phase difference. 14. a stationary interference pattern is produced.e...) (14) 14. Two sources which vibrate with a fixed phase difference between them are said to be coherent.
. if one (or both) of the two vibrating sources are turned on and off in a random manner (such that the phase difference between the vibrations of the two sources varies rapidly) then the interference phenomenon will not be observed. 14. we will not observe any interference pattern on the screen. The eye can notice intensity changes which last at least for a tenth of a second and hence we will observe a uniform intensity over the screen. We summarise the above results by noting that light beams from two independent sources do not have any fixed relationship***. be easily done by putting two polaroids in front of Sl and S2. however. In the actual experiment a light source illuminates the pinhole S (see Fig. therefore. 14. waves from different atoms would differ in their initial phases. light comes from a large number of independent atoms. Light diverging from this pinhole fell on a bather which.was obtained. The trick lies. if we use two conventional light sources (like two sodium lamps) illuminating two pinholes (see Fig.6 times which are small compared to z. can. however. 14:7) were coherent and on the screen beautiful interference fringes were obtained. For example. discuss the interference pattern produced by light waves. a stationary interference pattern. Young explained the Y 14. 14. This can be understood from the following reasoning: In a conventional light source. light coming out from the holes Sl and S2 will have a fixed phase relationship for a period of about 10 10 sec. Spherical waves emanating from S1 and S2 (see Fig. in the division of a single wavefront into two. hence the interference pattern will keep on changing every billionth of a second. We will discuss this point again in Sec. 14. then cos 2 0/2 will randomly vary between 0 and 1 and < cos 2 0/2 > would be 1/2 [see also Sec. i.6 and also in Chapter 17. the screen.5 If two sodium lamps illuminate two pinholes Sl and S2. these two split wavefronts act as if they emanated from two sources having a fixed phase relationship and. We should mention here that by using two independent laser beams it has been possible to record the interference pattern (see Chapter 17). ** This interference pattern will be a set of dark and bright bands only if the light waves have the same state of polarization: This . Young showed that the fringes on the screen disappear when Sl (or S2).5). However. *** Such sources are termed as incoherent sources. each atom emitting light for about 10 1 0 sec.6 Young's arrangement to produce interference pattern. For such a case 1=21° (16) which implies that if the sources are incoherent then the resultant intensity is the sum of the two intensities and there is no variation of intensity! Thus.4' INTERFERENCE OF LIGHT 'WAVES Till now we have considered interference of waves produced on the surface of water.10 sec. for light waves it is difficult to observe a stationary interference pattern.* Even if the atoms Si I S2 Screen Fig. no interference pattern will be observed on Fig. when these two waveswere allowed to interfere. 14. such a short pulse consists of about a million oscillations. We will now. thus it is almost monochromatic (see Chapter 17). light emitted by an atom is essentially a pulse lasting for only 10. contained two pinholes Si and S2 which were very close to one another and were located equidistant from S.e. _ Thomas Young in 1801 devised an ingenious but simple method to lock the phase relationship between the two sources.6]. Consequently. .14. In order to show that this was indeed an interference effect. Optics were emitting under similar conditions. * Since the optical frequencies are of the order of 10 15 sec 1.6). 14.is covered up. if we have a camera whose time of shutter opening can be made less than 10 10 sec then the film will record an interference pattern**. as such they do not produce any stationary interference pattern.
0 being the foot of the perpendicular from the point S on the screen.7 shows the section of the wavefront on the plane containing S.7 interference pattern by considering the principle of superposition.51) 21 1/2 + [(50)2 + (0. In order to find the shape of the fringe on the screen we assume the origin to be at the point 0 and the zaxis to be perpendicular to the plane of the screen as shown . (17) we obtain S2PSP (19) nAD yn d (20) Thus the dark and bright fringes are equally spaced and the distance between two consecutive dark (or bright) fringes is given by = Yn+i Yn = SIP = O. We would determine the positions of maxima and of minima on the line LL' which is parallel to the yaxis and lies in the plane containing the points S.SIP = + S2 Py SIP (18) Fig. 7. the error involved is about 0.8). Eq.D d Qd AD (21) which is the expression for the fringe width.005 cm Thus if we replace S2P + SIP by 2D. for d 0.8. SI and S2 (Adapted from Ref. In order to determine the shape of the interference pattern we first note that the locus of the point P such that S2P . We will show that the interference pattern (around the point 0) consists of a series of dark and bright lines perpendicular to the plane of Fig. 14.49) 100. 14. D = 50 cm.02 cm. In this approximation. OP = 0. the locus is a hyperbola of revolution obtained by rotating the hyperbola about the axis SIS2 .005%. For an arbitrary point P (on the line LL ' ) to correspond to a maximum we must have S2P  If yn. (18) becomes ynd D Using Eq.. For example. used with permission). 14. Consequently. (17) or (n+1)AD d n^.SIP = A (22) L' Fig. 14. S1 and S2.5 THE INTERFERENCE PATTERN Let SI and S2 represent the two pinholes of the Young's interference experiment.Two Beam Interference by Division of Wavefront 14. Figure 14. d << D then negligible error will be introduced if S2P + SIP is replaced by 2D.5 cm (which corresponds to typical values for a light interference experiment) 2] 1/2 S2P + SIP = [(50)2 + (0. Si and S2 (see Fig. 2.7 Sections of the spherical wavefronts emanating from S. n = 0.. 14..(S1P)2 [D2 + Cyn + d)2] [D2 +(yn = 2)2 2y nd and OP = yn Light Dark Central Light maximum Dark Light Dark Light where S1S2 = d Thus S2P . (S2P)2 . Now.8 Arrangement for producing Young's interference pattern. and by measuring the distance between the fringes he calculated the wavelength..2). is a hyperbola in any plane containing the points Sl and S2 (see Example 14. 1.
. 14.. D) respectively. D) and  2 .. 14.8 in Fig.6 THE INTENSITY DISTRIBUTION Let El and E2 be the electric fields produced at the point P by SI and S2 respectively (see Fig.O2)y2 .8). We consider an arbitrary point P on the plane of the screen (i. the two fields will almost be in the same direction.) 2 12 +DZ The intensity (I) will be proportional to the square of the electric field and will be given by 1= KE2 (27) )] which is the equation of a hyperbola. The electric fields El and E2 will. Thus the shape of the fringes is hyperbolic. 0). consequently. Thus we obtain approximately straight line fringes on the screen. they can also be seen through an eyepiece. The fringes so produced are said to be nonlocalized. y. Thus S2P . It is easy to see that if we had slits instead of the point sources we would have obtained again straight line fringes with increased intensities. have different directions and different magnitudes. The resultant field will be given by E=E l +E2 = i [ E01 cos ( y)2 + D2 +E02 E2 =1E02 cos( S2 Pcot ll (25) SIP .14. (0.10 15 sec1 ) and all the terms depending on cot will vary with extreme rapidity (10 15 times in a second).e. Now cost (cot .cctt ) r ) where i represents the unit vector along the direction of either of the electric fields.6. any detector would record an average value of various quantities. we may write El =1E01 cos 2 .02x2 =A2 [D2 + k(d 2 A2 11 or 2 I = K [ E0 1 cos2 ( 2 7r SIP = cot + ) E02 cost I S2P . It should be emphasized that the fringes are straight lines although the sources SI and S2 are point sources.SIP . where K is a proportionality constant. in general. the speed of light in free space. On rearranging we get y=± d 2. Thus. . if the distances S1P and S2P are very large in comparison to the distance S 1S2.* For an optical beam the frequency is very large (co.  Optics 14.wt) + 4 (23) E01 E02 cos [ 22r (S2P .1/2 (d2 . The yaxis is assumed to be parallel to S 2S1 .B)} = 1 J 1 + cos [2 (cot .(S2P + S1P)] }] the loci are straight lines parallel to the xaxis.0)r a} * Equation (27) will be derived in Chapter 22. they can be photographed by just placing a film on the screen.0)] d t 2 2T _z 2 + 167c T {ism 2 (wt .SIP)] + A (28) (24) cos [2wt . 14. (3' . The coordinates of the points SI and S2 are 10.SIP = [x2 +(y + 2 ) + DZ 2 1/2 [x2 +(Y = 0 (say) or f) 2 +D2 ] v2 and (1. z = 0) (see Fig.854 x 1012 cou12/Nm2) represents the permittivity of free space and c.cot) (T S2 Pcot)] (26) or Hence.^2. 1/2 [x2 For values of x such that x2 << D2 O2 . However.6) Let its coordinates be (x. In free space the constant K will be shown to be equal to EO c2 where e0 (=8.
The actual fringe 02 KE 2 represents the intensity produced by the source I2 = pattern (as it will appear on the screen) is shown in S2 if no light from SI is allowed to fall on the screen. 1. If we now carry out the averaging over time scales which are of the order of 10 8 sec.10(a). (30). (30) we may deduce the following: 0.005 mm (/3 = 5 mm) and d = 0. and the minimum intensity occurs when or .10. ^ S Fig. *** Notice that this variation occurs in times of the order of 1010 sec which is about a million times longer than the times for variation of the intensity due to the terms depending on wt. Thus we are justified in first carrying out the averaging which leads to Eq.11 12 cos3 (30) Thus.. we Imin = ( I I. 14.2.3) and thus S would also vary with time*** in a random way. The wave vectors for the two waves are given by n=0.5 Instead of considering two point sources. Figures 14. 1.025 mm (/3 1 mm) respectively. 14..1 sec.Two Beam Interference by Division of Wavefront where T = co Z(^ 22c x 1015 sec for an optical beam).) A. even for a detector having 1 nsec as the resolution time.11(a).. 14. In general.9.10 (a) and (b) correspond to d = Eq.4).0) 14. thus T/z = 6 x 1014. will be given by I=I1 +I2 + 2.. = (Il + I2) 2 and (32) Example 14. 'r = 0. the intensity minimum is zero. = KE2o1 2 (33) I = 21o + No cos S = 4I° cos2 represents the intensity produced by the source SI if The intensity distribution (which is often termed as the 2 no light from S2 is allowed to fall on the screen.= 5 x 105 cm. . that a detector will record. the resultant intensity is the sum of the intensities produced by each one where of the sources independently and no interference pattern is observed. or S2P  Fig..(S2 PSI P) (31) (c) In the arrangement shown in Fig.9 Notice that when II = I2. as such the maximum and minimum that one can see the hyperbolic nature of the fringe values of I are given by pattern in Fig. 14. (b) If the holes Si and S2 are illuminated by different light sources (see Fig.0) will oscillate between +1 and 1 and its average will be zero as can indeed be shown mathematically. ** For a normal eye. similarly cos2 pattern) is shown in Fig. From Fig. cos2 6/2 SIP = nA. Both figures correspond to D = 5 cm and (a) The maximum and minimum values of cos S are +1 and A.9 The variation of intensity with S. represents the phase difference between the displacements then reaching the point P from SI and S2. S= . 14. The values of the parameters are such 1 respectively. For any practical detector** z <<< 1 and since the quantity between the curly brackets will always be between 2 and +2. S=(2n+1)tr'. 11 ^ 12 and the minimum intensity is not zero. S2P  SIP = C n + 1.. for two incoherent sources. In. we may write (cos2 (wt .n=0.6. then cos S) = 0 and we obtain I= I1 +I2 2 (29) The factor cos (2cwt .I2 )2 consider the superposition of two plane waves as shown in The maximum intensity occurs when 8 =2nrc. Further It I2 = Io (say) and 1 I. 14.. 14. if the distances SIP and S2P are extremely large in comparison to d. T/'c = 6 x 105.2. 14. then the phase difference S will remain constant for about 1010 sec (see discussion in Sec. Thus the intensity.
14.0005 mm Example 14. 14.cot] E2 = E02 cos [k2 • r . we once again consider the.11(b) shows the computer generated interference pattern on the screen LL' for 0 = n16 and A = 5000 A. 14. kl = . [kysin .005 mm and 0.11 (a) The superposition of two plane waves on LL'. The first and second bright circles will correspond to a path difference of 99 9 and 98 A respectively. Similarly. and the fringe pattern will be strictly straight lines with fringe width ' given by 2 sin e Figure 14. 14.10.025 mm respectively (both figures correspond s to D = 5 cm and 2 = 5 x 10 cm).7 We finally consider the interference pattern produced on PP' by the superposition of a plane wave incident normally and a spherical wave emanating from the point 0. for both figures S1 S2 = d = 0.12(b) and (c) show the fringe patterns for = 20 cm and D = 10 cm.8). Figures 14.S2P (a) Fig.14.12(a)]. 14.Sr k sin01 + = constant k cos0 1 and k2 = + y k sin02 + k cos02 where k = 2tt/2.k y sin Of + k z cos 01 . (a) and (b) correspond to d = 0. r .5 A respectively. The radii of the fringes can be calculated by using the formula given in Problem 14. if we further assume E01 = E02 = Eo and 01 = 02 = 0 then the resultant field will be given by will be a circle. on the plane PP'. the locus of the point P for which SIP .10 d = 0. the first and second dark rings in the interference pattern will correspond to a path difference of 99.6 In this example. (b) Computer generated interference pattern on the screen LL' for 01 = e2 = it/6 and ? 5000 A. E1 Thus (for this value of d) the central spot will be bright for all values of D and will correspond to n = 100. (b) pattern produced by 2 point sources S1 and S2 on a plane PP' which is perpendicular to the line joining S1 and S2 [see Fig.05 mm and = 5000 A.wt) Thus the intensity distribution on the photograph plate LL' will be given by E =2E0 4 I0 cos2(k y sin e) .:.wt] = Eol cos [. Thus the electric fields of the two waves are described by the following two equations Eol cos [k1 .0005 mm.10 Computer generated fringe pattern produced by two point sources S1 and S2 on the screen LL' (see Fig. have assumed both electric fields along the same direction (say along the aaxis).5 A and 98. Thus a = A = 0. Obviously.wt] E02 cos.x (mm) (a) k I 10 I 0 x(mm) (b) I 10 Fig. 0. The plane wave will be given by El =E0cos(kzwt+0) cos (kysin 0) cos (kzcos 0.11(a). The fringes are parallel to the xaxis. Example 14.025 mm (/3 = 1 mm) Optics rrvrnnvr10 E E lleurlirer^ .02 +kzcos 02 cot] where we. and 01 and 02 are defined in Fig. Obviously. if 0 represents the centre of the fringe pattern then D S1 OS2O=d= 100.
If r...11 Si S2 D (a) (b) Fig. y << D. On the plane z = D. o D i.1 Moire Fringes We may mention here that Moire fringes can be very effectively used to study the formation of fringe patterns.wt) r where r is the distance measured from the point 0 which is assumed to be the origin.12 (c) (a) St and S2 represent two coherent sources.Two Beam Interference by Division of Wavefront 14..e..6. 14. the resultant field will be given by E=E +E2 E0 cos(kDwt+0) cos +  1 and we would obtain circular interference fringes as shown in Fig. Now. +y2 +D 2 ) 2 =DI1+ .13(b).+p denote the radii of nio' and (m + p)'' bright ring then 2 P [kD+2D (x 2 +y2 )cot] 14. r x2+ 2 2D2 ] Ap E. on the plane PP' (z = D) E2 =  +E0 If we assume that D cos[ 2D (x2 + y2)  0 r= (x 2. and the spherical wave will be given by Ao cos (k r . 14. (b) and (c) show the interference fringes observed on the screen PP' when D = 20 cm and D = 10 cm respectively. 14. the amplitude of the spherical wave (on the plane PP') is the same as the amplitude of the plane wave then 2) = (E 2 +y2 D+ x 2Epcost 4 = D (x 2 +y 2 )20 2D J where we have assumed x. and r. In Fig..14 we have shown the overlapping of two simple patterns from which one can understand the formation of Thus .
where X is the foot of the perpen .S1P = nA. 14. The angles S'AO' and S"AO' are 10° each. Solution: We will first calculate the intensity pattern produced by each source. bright and dark fringes when two plane waves propagate in slightly different directions. it can be easily demonstrated by having a periodic pattern on a transparency and overlapping it with its own photocopy at different angles. 17.16). Light from the source S' will produce constructive interference at the point P if XS2 + S2P .16. almost monochromatic light of the same wavelength 2i. Since the two sources are incoherent.S'S1 = XS2 .15 The Moire pattern produced by two overlapping circular patterns. (To get a clearer fringe pattern. one obtains the hyperbolic fringes as shown in Fig. E. (a) Fig.15. The circular pattern was provided by Dr. You will see clear hyperbolic fringes if you put the pattern at a greater distance from the eye.13). 14. R.14 The Moire Pattern produced by two overlapping straight line patterns. 14. you may have to view the patterns from a greater distance.) In Sec.8 Consider two parallel slits S1 and S2 (perpendicular to the plane of the paper) illuminated by two distant incoherent sources S' and S" as shown in Fig. Example 14. the resultant intensity pattern on the screen will be the sum of the intensities produced by each source on the screen.12 Optics Incident plane wave (a) 14.13 (a) Superposition of a plane wave and a spherical wave emanating from the point 0. Bailey. (b) shows the interference fringes observed on the screen PP'.D 'I Fig. determine the intensity pattern on the screen which is at a distance D from the slits (see Fig. Assuming that both the sources emit In the above equation we have assumed S' to be far away from the slits so that S'S2 . In a classroom. Consider an arbitrary point P on the screen. 17. if one overlaps a circular pattern (on a transparency) with its own copy. 14. Similarly. 14. 14.5 we have shown how the beat phenomenon can be understood by observing the Moire fringes obtained by the overlapping of two patterns of slightly different periods (see Fig.14. (b) (c) Fig.
the entire fringe pattern would shift by half a fringe. A portion of the wavefront from S gets reflected from MIM and illuminates the region AD of the screen. 14. it consists of two plane mirrors which are inclined to each other at a small angle 9 and touching at the point M. in calcite the extraordinary wave travels faster than the ordinary wave. A similar fringe pattern is obtained when E is in the ydirection. Thus the intensity pattern on the screen due to S' would be Is. = Io cos 2 8/2 where S= But XS2 [XS2 + S2P  S1P] = S1S2 sin (XS1S2) SlS2 sin 10° r cos2 Therefore Is. and nA.13 P P 0' S" S2 x D Fig. Thus. cos2 [ tb A. = 1o cos 2 L r (S2P .e. 14. one observes interference Example 14.6). Determine the intensity distribution on the screen.Two Beam Interference by Division of Wavefront 14. is shown in Fig. 22. if the path through H1 contains n wavelengths. a ypolarized wave will propagate as an ewave through H1 (see Sec. 14. then (since the ewave travels faster) the path through H2 dicular drawn from SI on S'S2. 14. Now. Thus there is an additional path of A/2 introduced in one of the beams and the entire fringe pattern shifts by half of a fringe. a ypolarized wave will propagate as an owave through H2. A filter F is usually placed in front of the slits to make the light falling on the slits almost monochromatic. One of the experimental arrangements.. i. known as the Fresnel twomirror arrangement. * This example presupposes the knowledge of halfwave plates which has been discussed in Chapter 22 [see Sec.SIP) + S1 S2 sin 10° 1 will contain (n+ 2) wavelengths. Since these two wavefronts are derived from the same source they are coherent.17). The halfwave plate placed in front of slit S l has its optic axis in the ydirection.18. Thus in the region BC. Assume that the light coming from the source S is polarized along (a) the yaxis and (b) the xaxis. whereas the optic axis of the halfwave plate (placed in front of S2) is along the xaxis. 22. respectively.9 In Young's double slit arrangement two calcite halfwave plates* are placed in front of the slit (see Fig. .d sin 10°}] The resultant intensity would be given by I=Is. Fresnel devised a series of arrangements to produce the interference pattern. Consequently.17 H1 and H2 are halfwave plates placed in front of the slits Sl and S2. = 1o = 10 Similarly Is. The optic axis of H1 and H2 are along y and x directions respectively. Another portion of the wavefront gets reflected from the mirror MM2 and illuminates the region BC of the screen. D +d sin 10°}] . Solution: Let us first consider the case when E is along the yaxis. +Is„ The above example is of practical importance when distant stars (like Betelguse and Rigel) are viewed by means of an interference arrangement. 14.16 Two distant incoherent sources S' and S" illuminate the slits S1 and S2. H2 0 S S1 H1 L Y Z Fig. Furthermore. since the optic axis in H1 is along the yaxis. S represents a narrow slit placed perpendicular to the plane of the paper. maxima and minima occur when S2P SIP is (n + 2 ) A.6].7 FRESNEL'S TWOMIRROR ARRANGEMENT After Young's double hole interference experiment. Similarly.
3 cm. He used a biprism.19 and the prism is assumed to stand perpendicular to the plane of the paper. a = 2 cm. The formation of the fringes can also be understood as being due to the interference of the wavefronts from the virtual sources S1 and S2 of S formed by the mirrors M1 and M2 respectively.1)a. C and L represent the positions of the crosswires and the eyepiece respectively. S represents the slit which is also placed perpendicular to the plane of the paper. where a represents the distance from S to the base of the prism. 14. 6 x 10. then the angle S 1 SS2 is also 0 and the angle S 1MS2 is 20.5. . A. The fringes can be viewed through an eyepiece.8 x 103 Fig.jd1 d2 and D=b1 +b2 Typically for d 0. Since each pair of points S1 and S2 produce (approxi * This method is similar to the displacement method for the determination of the focal length of a convex lens. 14. If n represents the refractive index of the material of the biprism and a the base angle. if the angle between the mirrors is 0. where R is the radius of the circle. In above we have considered here a slit instead of a point source. fringes.b2 D b10. /3 = 0. Once /3 is known. can be determined by using the following relation: (34) It may be mentioned that in order to determine d. Let d2 and b2 be the corresponding distances when the lens is at L2.19) where the images of Si and S2 can be seen at the eyepiece. 14. which was actually a simple prism. Further.14. For a fixed position of the eyepiece there will be two positions of the lens (shown as L1 and L2 in Fig. In order to determine d one introduces a lens between the biprism and the crosswires. S1 and S2 lie on a circle whose centre is at the point M.18 Fresnel's twomirror arrangement. Thus S 1 S2 is 2R0. the distance S 1 S2 is 2a(n . one gets d = 0.I C L Fig. one need not measure the value of a. 14. In fact the distances d and Dcan easily be determined by placing a convex lens between the biprism and the eyepiece. 14. From simple geometric considerations. Light from the sodium lamp illuminates the slit S and interference fringes can be easily viewed through the eyepiece. The base of the prism is shown in Fig. The fringe width (/3) can be determined by means of a micrometer attached to the eyepiece. it can be shown that the points S. LI and L2 represent the two positions of the lens where the slits are clearly seen.8 FRESNEL BIPRISM Fresnel devised yet another simple arrangement for the production of interference pattern.cm. then (n 1)a is approximately the angular devia radians. a (20 ' ) = 5. the base angles of which are extremely small (. Light from the slit S gets refracted by the prism and produces two virtual images S h and S2. These images act as coherent sources and produce interference fringes on the right of the biprism. D = 50 cm.14 Optics R Q P 0 L1 Imo. for n = 1. tion produced by the prism and.20' ).01 cm.* Let di be the distance between the two images when the lens is at the position Li (at a distance b1 from theeyepiece). A.012 cm. The biprism arrangement can be used for the determination of wavelength of an almost monochromatic light like the one coming from a sodium lamp. Then. it can easily be shown that d= . therefore. Thus.19 Fresnel's biprism arrangement.
Thus. Now. 5 x 10 5 cm (greenish yellow) and 4.1)t S20' 0' 0 D(n1)t d D Fig. For example. with white light one gets a white central fringe at the point of zero path difference along with a few coloured fringes on both the sides. will be white. this can easily be done by using white light as a source. For example. then the colour will gradually change to violet. 14. While using a white light source. wavelengths corresponding to 30 x 105 /n (n = 1.20. if we put a red (or green) where v (= n) represents the speed of light in the plate. The point Q which satisfies S2Q . Consequently we will have a line devoid of the violet colour and will appear reddish.20 If a thin transparent sheet (of thickness t) is introduced in one of the beams.10 DISPLACEMENT OF FRINGES We will now discuss the change in the interference pattern produced by introducing a thin transparent plate in the path of one of the two interference beams as shown in Fig. Thus the time required for the light to reach from SI to the point P is given by SIP . the central fringe produced at the point 0 (Fig. in the visx 105/ (n + 2) ible region. . as seen by the unaided eye.19) will be white because all wavelengths will constructively interfere here.tin air. Clearly. the central fringe (which corresponds to equal optical path from SI and S2) is formed at the point 0' where S1O' + (n .) will constructively interfere. (35) shows that by introducing the thin plate the effective optical path increases by (n ..15 mately) straight line fringes.6 x 105 cm (indigo) will be absent. The wavelengths corresponding to the violet and red end of the spectrum are about 4 x 105 cm and 7 x 105 cm respectively.1)t. 14. when the path difference is about 2 x 105 cm the fringe will be red in colour. The colour of such light.SIR = 30 x 105 cm.Two Beam Interference by Division of Wavefront 14.t+ = 1 t [SI P t+nt] c v c _ [SIP + (n . if the point P is such that l 2 J then complete destructive interference will occur only for the violet colour. when the thin plate is introduced. Partial destructive interference will occur for other wavelengths. thus.9 INTERFERENCE WITH WHITE LIGHT We will now discuss the interference pattern when the slit is illuminated by white light. Let t be the thickness of the plate and let n be its refractive index..1) t D/d. wavelengths corresponding to 30 will destructively interfere. slightly below (or above) the point 0 the fringes will become coloured. Eq. 2.3 x 105 cm (violet). the colour soon fading off to white.67 x 10 5 cm (orange). 5. 6 x 105 cm (yellow). the fringe pattern gets shifted by a distance (n . In the visible region these wavelengths will be 7.5 x 105 cm (yellow).5 x 10 5 cm (red). Thus following the white central fringe we will have coloured fringes. at a point R. 14. the wavelengths 6.5 x 105 cm (= 2d ( S2P . In many interference experiments it is necessary to determine the position of the central fringe and. It is easily seen from the figure that light reaching the point P from SI has to traverse a distance t in the plate and a distance SIP .1)t] (35) 1 i will be devoid of the red colour. 4. such that S2R . 14. we will see the interference pattern corresponding to the red (or green) light. the slit will also produce straight line fringes of increased intensity.SIP = 2 x 105 cm I = 'violet filter in front of our eye. Further. as has been discussed above. No other wavelength (in the visible region) will neither constructively nor destructively interfere. In the usual interference pattern with a nearly monochromatic source (like a sodium lamp) a large number of interference fringes are obtained and it is extremely difficult to determine the position of the central fringe. The coloured fringes will soon disappear because at points far away from 0 there will be so many wavelengths (in the visible region) which will constructively interfere that we will observe uniform white illumination. It will correspond to almost constructive interference for the violet colour.. Thus.S1Q = 3. 14.
Solution: The point 0' (see Fig. Solution: Hence A = 0. The above principle enables us to determine the thickness of extremely thin transparent sheets (like that of mica) by measuring the displacement of the central fringe. using the principle of optical reversibility.20) are 0.21.16 Since [see Eq. In the next section. we will get minima (i.2 cm.2 cm.. 14..1 cm.1)t = . 14. for more details see Refs.00' therefore (n . (This is discussed in detail in Problem 14. Due to the introduction of the mica sheet the central fringe gets shifted by 0. and its virtual image S2 to form two coherent sources which produce the interference pattern. According to this principle.00' Optics (36) Thus the fringe pattern gets shifted by a distance A which is given by the following equation: . if white light is used as asource. Example 14. One may thus consider the slit SI * This 14.S1P = nA.14.1x0. Further.. 1. . a light ray that is reflected or refracted will retrace its original path if its direction is reversed.1 cm and 50 cm respectively. even a dielectric surface has very high reflectivity (see Chapter 23). destructive interference). we will show that if there is an abrupt phase change of 7r when light gets reflected by a denser medium. 14. then no such abrupt phase change occurs when reflection takes place at a rarer medium. the displacement of the central fringe is easy to measure. D = 50 cm d0 t _ 0. if S2PSIP=In we will get maxima. Determine the thickness of the mica sheet.1)t = 0. when the point P on the screen is such that S2P .S1 O' _ 13.* principle is consequence of time reversal invariance accordingte which processes can run either way in time. one may introduce a thin mica sheet in the path of the direct beam so that the central fringe appears in the region BC.7x104 cm D(n 1) Example 14.. where it touches the end of the reflector. Alternatively. if the central fringe is observed with white light.58. This implies that the reflected beam undergoes a sudden phase change of r on reflection.111 In an experimental arrangement similar to the one discussed in the above example one finds that by introducing the mica sheet the central fringe occupies the position that was originally occupied by the eleventh bright fringe. It should be noted that at grazing incidence one really need not have a mirror.20) corresponds to the eleventh bright fringe. in the absence of any absorption.2 50 x 0.58t 14. 14.12 PHASE CHANGE ON REFLECTION We will now investigate the reflection of light at an interface between two media using the principle of optical reversibility. 3.D(n1)t d (37) :U2 Fig.e. 2. 3 and 8.10 In a double slit interference arrangement one of the slits is covered by a thin mica sheet whose refractive index is 1. the central fringe cannot be observed on the screen unless the latter is moved to the position L '1L'2.58 =6. 14. On the other hand. it is found to be dark. If the source of light is a sodium lamp (2 = 5893 A) determine the thickness of the mica sheet.21 The Lloyd's mirror arrangement.2) Indeed. n = 0.21). Consequently. As can be seen from Fig. The light directly coming from the slit SI interferes with the light reflected from the mirror forming an interference pattern in the region BC of the screen. The distances S1S2 and AO (see Fig. thus S20'SI O' = 11A= (n . d = 0.11 THE LLOYD'S MIRROR ARRANGEMENT In this arrangement light from a slit SI is allowed to fall on a plane mirror at grazing incidence (see Fig. (19)] S20' .
17 In Chapter 24. i.e. a stationary interference pattern is obtained. PA. therefore. . we may infer from Eq. the colour soon fading off to white. almost straightline interference fringes are formed on some planes and by measuring the fringe width (which represents the distance between 2 consecutive fringes) one can calculate the wavelength. Similarly.ri (38) 14. the distance between the two holes is 0. 14.2 Figure 14. We now reverse the rays and we consider a ray of amplitude at1 incident on medium 1 and a ray of amplitude arl incident on medium 2 as shown in Fig. 5 cm and 190 cm respectively. It will be shown that the coefficients satisfy Stokes' relations. 5 cm. 14.23 For Problem 14. S is a point source emitting waves of frequency 6 x 1014 sec1 . (b) Rays of amplitude arl and at1 incident on a medium of refractive index n1. Thomas Young devised an ingenious but simple method to lock the phase relationship between two sources of light. the phase change on reflection will also be discussed there.A and B represent the two ends of a mirror placed horizontally and LOM represents the screen. (a) Determine the position of the region where L s•_ 0 P A B (a) (b) Fig. 2 = 5 x 10 5 cm and D = 50 cm. 14.22(b). one gets a white central fringe at the point of zero path difference along with a few coloured fringes on both the sides.22(a). 14. The ray of amplitude at1 will give rise to a reflected ray of amplitude at1 r2 and a transmitted ray of amplitude at1 t2 where r2 and t2 are the amplitude reflection and transmission coefficients when a ray is incident from medium 2 on medium 1. Let the amplitude reflection and transmission coefficients be r1 and t1 respectively.Two Beam Interference by Division of Wavefront Consider a light ray incident on an interface of two media of refractive indices nl and n2 as shown in Fig. This is indeed borne out by experiments. Equations (38) and (39) are known as Stokes' relations. (39) that no such abrupt phase change occurs when light gets reflected by a rarer medium.5 mm. Thus. the two rays of amplitudes at1 r2 and anti must cancel each other.22 (a) A ray travelling in a medium of refractive index nl incident on a medium of refractive index n2. one can calculate the thickness of the mica sheet. ♦ On a plane which is normal to the line joining the two coherent point sources. we will calculate the amplitude reflection and transmission coefficients for plane waves incident on a dielectric and also on a conductor. 14.23 represents the layout of Lloyd's mirror experiment. The distances SP.6). the fringes get displaced and by measuring the displacement of fringes.2. ♦ In the Young's double slit interference pattern. 14. If we now introduce a very thin slice of transparent material (like mica) in the path of one of the interfering beams.22(a). Summary Further. What will be the fringe width? 14. ♦ For two coherent point sources. thus ari + ati t2 = a or t1t2 = 1 .. the ray of amplitude arl will give rise to a ray of amplitude ari and a refracted ray of amplitude arl t1 . Problems Since we know from the Lloyd's mirror experiment that an abrupt phase change of 7V occurs when light gets reflected by a denser medium. the fringe pattern is circular. M Fig.1 In the Young's doublehole experiment (see Fig. when these two waves were allowed to interfere. 14. The trick lies in the division of a single wavefront into two. if we use a white light source. these two split wavefronts act as if they emanated from two sources having a fixed phase relationship and. at1 r2 + arl tl = 0 or r2=r1 (39) ♦ In 1801. if the amplitude of the incident ray is a. According to the principle of optical reversibility the two rays of amplitudes ari and at1 t2 must combine to give the incident ray of Fig. then the amplitudes of the reflected and refracted rays would be art and at1 respectively. AB and BO are 1 mm.
Physics. (b) Calculate the thickness of a mica sheet (n = 1.12 In continuation of Problem 14.1) as where a represents the distance from the source to the base of the prism (see Fig.10 assume that d = 0. Hecht and A. 14. John Wiley. I. 1970. `Moire Fringes'. 1973. Reading. Australia. 5.14 Assume a plane wave incident normally on a plane containing two holes separated by a distance d.11 In continuation of the above problem calculate the radii of the first two dark rings for (b) D 10 cm.02)]l = VW .13 Using the expressions for the amplitude reflection and transmission coefficients (derived in Chapter 24). Chapter 52. Mass. . Booth.9 In the double hole experiment using white light.5 In order to determine the distance between the slits in the Fresnel biprism experiment. 14.4 mm and D = 20 cm. AddisonWesley.. valid for D >> d. M. Si 0 . D. D. Proceedings of the Conference and Workshop on the Teaching of Optics (Edited by: G. 14. 14.3 (a) In the Fresnel's biprism arrangement.15 In Problem 14.. 7..5 Inn.16 In the Young's double hole experiment calculate I/Imp where 'l represents the intensity at a point where the path difference is A/5. Ditchburn.19). P. AddisonWesley. If the central fringe gets shifted by 0. [Ans: = 0.18 the fringes will be visible and calculate the number of fringes. Assuming n = 1. [Ans: (a) 2 cm.3995 mm] 14. one puts a convex lens in between the biprism and the eye piece.S2P for the point P to be first dark ring and first bright ring. Geometrical and Physical Optics. Wolf. Can you use Eq. (21) for the fringe width? 14. and for sodium light (A. King. interference fringes are formed using sodium light which predominantly comprises of two wavelengths (5890 A and 5896 A). If d1 and d2 are the distances between the images (of the slits) as measured by the eye piece.39975 mm. A. then the inter ference pattern will consist of exactly parallel straight lines. Reading.8 In the Fresnel's two mirror arrangement (see Fig. AddisonWesley. R.. 0. 8. 2nd Edition. What would happen if D < 4f ? Optics 14. Longman. 40 fringes. Heath & Co. 14. d = 0.4 In the Young's double hole experiment a thin mica sheet (n = 1.O 2) T [4D2 + 4Dd + (d2 . The velocity of light is 3 x 10 1° cm/sec. Chicester (2000). Assume d = 0. if the plane does not lie on the front focal plane. Born and E. Show that if D > 4f one will obtain two positions of the lens where the image of the slits will be formed at the eye piece.2 cm. If S1P .5) is introduced in the path of one of the beams. and 4 cm.1 cm. Chapter 3.18) 'show that the points S. = 5 x 10 5 cm and D = 100 cm.5) which should be introduced in the path of the direct ray so that the lowest fringe becomes the central fringe. Welch. here f is the focal length of the convex lens and D is the distance between the slit and the eye piece. ()pat. R. show that they satisfy Stokes' relations. show that d = 2(n . J. 1 cm. A. 3.5 mm. Modern Physics and Anti Physics. the fringe pattern will be hyperbolae. E. Find the wavelengths (in the visible region) which correspond to constructive and destructive interference. 14. one corresponding to a path difference of 5000 A and the other corresponding to a path difference of 40000 A. However. 14. Bailey and M.14. consider two points on the screen. Optics. Boston.0) (d + A) 0 where the last expression is. 4. E. 998 and n = 997 respectively. W. Vol. What will be the colour of these points? 14. You may assume d = 0. Leighton and M. 1974. I. as observed on the focal plane of the lens.22 cm] 14.7 If one carries out the Young's double hole interference experiment using microwaves of wavelength 3 cm. Mazzolini and G.10 (a) Consider a plane which is normal to the line joining two point coherent sources Si and S2 as shown in Fig. Cambridge University Press. Longhurst. a is the angle of the biprism and n the refractive index of the material of the biprism. (a) D = 20 cm and [Ans: (a) = 0. Feynman. Thus the central (bright) spot will correspond to n = 1000.71 cm and 1.C. D. 14. Baker. (b) For X = 0. Principles of Optics.6 In the Young's double hole experiment. Mass. Graham Smith and T. Calculate the radii of the first. 2000. Optics and Photonics: An Introduction. University of Melbourne. Smith. Calculate the value of S1P . Zajac. then show that y = 20 (d2 . [Ans: 0.5 mm and D = 100 cm. REFERENCES AND SUGGESTED READINGS 1.A. A.71°] 14. (b) In a typical biprism arrangement b/a = 20. show that the fringe width. show that if the plane (containing the holes) lies in the front focal plane of the lens. F. You may assume D = 100 cm. PSSC. second and third bright rings which will correspond to n = 999. S.1 cm. then show that d = c11 d2 . 14. 9. calculate the angle a. R. 1965. B. here b is the distance between the biprism and the screen. Sands. If we place a convex lens behind the slits. and D = 50 cm. Light. 14. will be f A/d where f is the focal length of the lens. and S2 lie on a circle and S I S2 = 2b6 where b = MS and 6 is the angle between the mirrors. Obtain the regions on the screen where the fringe pattern will disappear. (b) 38 µm] 14. R. 1965. Academic Press. London.5. P.14. London (1976). Mass. The Feynman Lectures on Physics.14. calculate the thickness of the mica sheet. Cambridge. = 5893 A) one obtains a fringe width of 0. 2.S2P = A. discuss the nature of the fringe pattern if d = 0. 6.S20 = 800 A.1 cm.
1881 A. Sir James Jeans in The Universe Around Us. Your interferometer has rendered it possible to obtain a nonmaterial standard of length possessed of a degree of accuracy never hitherto attained. light coming out of a pinhole was allowed to fall on two holes. In this chapter we will consider the formation of interference pattern by division of amplitude. for example. W. Micrographia. DIVISION OF AMPLITUDE Following a method suggested by Fizeau in 1868. Morley carried out the famous MichelsonMorley experiment using the Michelson interferometer to detect the motion of the earth with respect to the 'Luminiferous Aether'. 1802 Thomas Young gave a satisfactory explanation of `Newton's rings' based on wave theory. he described his observations of the colours produced in flakes of mica. Michelson invented the "Michelson interferometer". Such studies have many practical applications and also explain phenomena like the formation of beautiful colours produced by a soap film illuminated by white light.1 INTRODUCTION In the previous chapter we discussed the interference pattern produced by division of a wavefront.Chapter Fifteen INTERFERENCE BY. \ 15. and to restore it with absolute infallibility. supposing it were to get lost " 1887 A. the British physicist Robert Hooke described his observations with a compound microscope having a converging objective lens and a converging eye lens. Michelson and E. (1930) Important Milestones 1665 In his treatise. if a plane wave falls on a thin film then the wave reflected from the upper surface interferes with the wave reflected from the lower surface.A. In the same work. Hooke advocated a wave theory for the propagation of light 1704 "Newton's rings" were first observed by Boyle and Hookethey are named after Newton because he had given an explanation using the corpuscular model which was later found to be unsatisfactory. the President of the Royal Swedish Academy of Sciences said. soap bubbles and films of oil on water.A. Cambridge University Press. Professor Michelson has produced what is perhaps the most ingenious and sensational instrument in the service of astronomythe interferometer. He recognised that the colour produced in mica flakes is related to their thickness but was unable to establish any definite relationship between thickness and colour. . He was awarded the 1907 Nobel Prize in Physics "for his optical precision instruments and the spectroscopic and metrological investigations carried out with their aid" Michelson was America's first Nobel prize winner in science and during the presentation ceremony of the Nobel prize. and spherical waves emanating from these two holes produced the interference pattern. for example. "Professor Michelson. By its means we are enabled to ensure that the prototype of the meter has remained unaltered in length.
2 The solid and the dashed lines represent the crests of the waves reflected from the upper surface and from the lower surface of the thin film. if we place a photographic plate at P (see Fig. For an air film between two glass plates (see Fig.2 INTERFERENCE BY A PLANE PARALLEL FILM WHEN ILLUMINATED BY A PLANE . 1. 15. 15. but a phase change of 2T will occur on reflection at the airglass interface and the conditions for maxima and minima will remain the same.2 Optics = 15.15. 15. and as such the conditions for destructive or constructive interference will be given by r V T d Fig.3) no phase change will occur on reflection at the glassair interface. Let the solid and the dashed lines in Fig. then the film will appear to be uniformly illuminated. 15. 1.1). In order to observe the interference pattern without obstructing the incident beam. Further. it will be dark when 2nd = m2 and bright when 2nd = (m + Z) A . the wave reflected from the lower surface of the film traverses an additional optical path of 2nd. Thus. On the other hand. 15. if the film is placed 2nd = m A = (in + 2 destructive interference (la) (lb) J constructive interference where m = 0. . This is because of the fact that the effective wavelength in a medium of refractive index n is Ain. *** In general. it may just be a beam coming out of a laser. with appropriate choice of the refractive indices of media II and • III. II n III 4 Fig.7. then the wave reflected from the upper surface of the film will undergo a sudden change in phase of r (see Sec.1. alternatively. waves reflected from the upper and lower surfaces of the film respectively. 15.1 WAVE If a plane wave is incident normally on a thin* film of uniform thickness d (see Fig.1). 14. in general. However. in air.*** Clearly..1 The normal incidence of a parallel beam of light on a thin film of refractive index n and thickness d. the two amplitudes can be made very nearly equal (see Example 15. Instead of placing the photographic plate. lower surface.. where n represents the refractive index of the material of the film. G denotes a partially reflecting plate and P represents a photographic plate. represents the free space wavelength. m = 0. be slightly different. Such an arrangement also enables us to eliminate the direct beam from reaching the photographic plate P (or the eye).1) then the waves reflected from the upper surface interfere with the waves reflected from the P t . if we try to view the film (from the top) with naked eye.. and as such the interference will not be completely destructive. 2.. Notice that the distance between the consecutive crests inside the film is less than the corresponding distance in medium I. ** Notice that the distance between consecutive crests in the film is less than the corresponding distance in air. It may be noted that the amplitudes of the waves reflected from the upper and lower surfaces will. if the I * Why the film should be thin is explained in Sec.2 represent the positions of the crests*'` (at any particular instant of time) corresponding to the. in this section we will study this interference pattern... the wave reflected from the lower surface of the film will suffer multiple reflections.12) . The effect of such multiple reflections is neglected (see Chapter 16). then the plate will receive uniform illumination...and a. we use a partially reflecting plate G as shown in Fig. The plane wave may be produced by placing an illuminated pinhole at the focal point of a corrected lens.. 15. 2. 15.
The solid and the dashed lines represent the corresponding positions of the crests. 5).5). which is given by (see Fig. We will show in the next section that A = 2n2d cos 0' (4) 15. 15.4). whenever the refractive index of the II medium lies in between the refractive indices of the I and the III media. Once again. We drop a perpendicular BJ from the point B on the lower surface LL' and extend BJ and FD to the point B' where they meet (see Fig. 15. BG.5): A = n 2 (BD + DF) . the II medium is an oil of refractive index 1. For a film placed in air. The latter traverses an additional optical path A. medium is crown glass (n 1.3 THE COSINE LAW where 0' is the angle of refraction.60 and the III medium is flint glass (n = 1. 15. The image formed at the retina will be dark or bright depending on the value of A (see Eq. a phase change of r will occur when reflection takes place at the point B and as such.4 The oblique incidence of a plane wave on a thin film.n 1 BC] = 2n2d cos 0' (6) Let 0 and 0' denote the angles of incidence and refraction respectively. ZJBD = LBDN = ZNDF = 0' where N is the foot of the perpendicular drawn from the point D on BF. the wave reflected from the upper surface of the film interferes with the wave reflected from the lower surface of the film.5 (3) where C is the foot of the perpendicular from the point F on Calculation of the optical path difference between the waves reflected from the upper surface of the film and from the lower surface of the film. and Thus 2 0' LB'DJ = is . We next consider the oblique incidence of the plane wave on the thin film (see Fig. 2 IA JJ minima maxima (2a) (2b) In general. Clearly.[_e'+e'+e'] = BD = BD' and BJ = JB' = d .Interference by Division of Amplitude Glass Air Glass Fig. then the conditions of maxima and minima would be given by Eqs (2a) and (2b).5) it will receive uniform illumination. if we try to view the film with naked eye (at the position E . 15. 15._ Now LBDJ= Fig. 15.n 1 BC Fig.3 (5a) (5b) = (In + 2 l I a. the conditions of destructive and constructive interference would be given by In this section we will show that the wave reflected from the lower surface of the film traverses an additional optical path which is given by the following expression: A [= n2 (BD + DF) . 15. maxima If we place a photographic plate at P (see Fig. 15.66) then a phase change of 7c will occur at both the reflections and the conditions for maxima and minima would be 2nd = (nt+ = mA A = 2n2 d cos 0' = mA minima 15. The solid and dashed lines denote the boundary of the wave reflected from the upper surface and from the lower surface of the film.3 Thin film of air formed between two glass plates.52).4) then only light rays reflected from a small position QR of the film will reach the eye. The eye E receives the light reflected from the region QR. P denotes a photographic plate.see Fig.
15. there would be constructive interference and when 2nd cos 0' = (m+ )A1. However. For example. In order to reduce these losses.n2 n1 + n2 2 r2 (14) (l0a) a. is negative showing that when a reflection occurs at a denser medium a phase change of 7r occurs. respectively. then the amplitudes of the reflected and the transmitted beams are a. 4% of the incident light is reflected. abrupt phase change of 7r occurs at both the reflections. glass (n = 1. for near normal incidence.5+1) 0. (b) corresponds to the case when the beam (propagating in a medium of refractive index n2) is incident on a medium of refractive index n1. * These relations can be derived from electromagnetic theory.12). 15. the refractive index of the film being less than that of the lens. reflected beam and the transmitted beam respectively. ar and a. is incident normally on a medium of refractive index n2. Substituting the above expression for BC in Eq. there would be destructive interference. given by r= t= nl ..n I BC Hence ZCFB = ZCBX = 0 Now or (7) BC = BF sin 0 = sin 0' sin 0 = n1 KF (8) where a.15. *** Since the refractive index of the nonreflecting film is greater than that of air and less than that of the glass. (7) we get A= n2B'F (11a) . Notice that when n2 > n1. Similarly. 15. = 2nI n1 + n2 a. 2 . for a quantitative understanding of the phenomenon.6(b)]. and a. lens surfaces are..5) may be coated with an MgF2 film (see Fig.67 and about 6% of light is reflected.2 lies in reducing the reflectivity of lens surfaces. a.n2 nl +n2 where K is the foot of the perpendicular from B on B'F. # t'a1 (10b) Equations (13) and (14) represent the Stokes' relations (see Sec.6 (a) If a plane wave of amplitude propagating in a medium of refractive index n1. 15.e.4 Optics BD+DF=B'D+DF=B'F 0 = n2B'F . see Eqs. The amplitude reflection and transmission coefficients r and t are.5 n+1) . this we plan to discuss in this section. We all know that in many optical instruments (like a telescope) there are many interfaces and the loss of intensity due to reflections can be severe. when 2nd cos 0' = m2. 15. Thus. then n1 (12) = r r = n2 n2 + n1 t' = 2n2 nl + n2 and _ 41202 1tt'=1_ (ni + n2)2 (13) ni . if we have a large number of surfaces..7) and the film thickness d should be such that*** nl n2 tat rat n2 ni r'a i i Fig.4 NONREFLECTING FILMS One of the important applications of the thin film interference phenomenon discussed in Sec. (67)(72) of Chapter 24 (with 01 = 02 = 0). i.often coated with a A/4n thick `nonreflecting film' .n2KF = n2B'K cos 0' (9) or 0 = 2n2 d which is known as the cosine law. For example.** the reflectivity of crown glass surface in air is n1 2 _ 1 2 (1. For a dense flint glass n = 1. We will now discuss the application of the thin film interference phenomenon in reducing the reflectivity of lens surfaces.(1. 14. are the amplitudes of the incident beam. Consequently. we will assume near normal incidence. therefore. 15.. we will have to assume that when a light beam (propagating in a medium of refractive index n1) is incident normally on a dielectric of refractive index n2 then the amplitudes of the reflected and the transmitted beams are related to that of the incident beam through the following relations* [See Fig. ** In all what follows in this section. the losses at the interfaces can be considerable.04.6(a): ar =' ni n 2 a nl + n2 ` 2nI (llb) nl + n2 It is interesting to point out that if r' and t' are the reflection and transmission coefficients where light propagating in a medium of refractive index n2 is incident on a medium of refractive index n1 [see Fig.
7 If a film (having a thickness of A.Interference by Division of Amplitude 15. to be 5..e. then waves reflected from the upper surface of the film destructively interfere with the waves reflected from the lower surface of the film.5 (1) (2) (5) Air (na = 1) Nonreflecting Film x = 0 d = 4nf (3) (4) of (< n9) x= d Glass (ng) Fig. Such a film is known as a nonreflecting film. nf = 1. We would like to emphasize the following points: (a) Let na. (9)] and of represents the refractive index of the film. 15.na ng . nf and ng be the refractive indices of air.8 Comparison between a glasses lens without antireflective coating (top) and a lens with antireflective coating (bottom).of 2nf a of +na ng +nf of +na nearly equal to unity. 15.nf a of +na ng +nf and 2na ng . nonreflecting film and glass respectively.29 and when ng = 1.38 Figure 15.9x10 5 om 4 x 1. Note the reflection of the photographer in the top lens and the timed reflection in the bottom. Figure adapted from http: / /eii..org/wiki/Optical coating (Photograph taken by Justin Lebar).8 shows a comparison between a grasses lens without antireflective coating (top) and a lens with antireflective coating (bottom).0x10 5 cm 0.5 .na a and o of +na 2na a f +na of . see Eq. Note the reflection of the photographer in the top lens and the tinted reflection in the bottom. Thus. for MgF2. cos 0' 1.nf = of +na ng +nf 4na nf (nf +na ) 2 (17) where we have used the fact that is very respectively (we have assumed nearnormal incidence).4. for na = 1 and zzf = 1.38.97 (nf +na ) 2 On simplification we obtain nf= . 2nfd or = 2 A.Jnang (18) If the first medium is air then na = 1 and with ng = 1. of na of +na or s= 2na ng nf 2nf a of +na ng + n f of +na (16) d 4n f where we have assumed near normal incidence [i. if we assume A.wikipedia. a color photograph appears as Fig.e.7) would be nf . the waves corresponding to rays (2) and (5) should have the same amplitude. respectively. 15. 4na n f = 0. If a is the amplitude of the incident wave then the amplitudes of the reflected and refracted waves (the corresponding rays shown as (2) and (3) in Fig. Now. 14 in the prelim pages. i./4nf and having refractive index less than that of the glass) is coated on the glass. for complete destructive interference.0 x 105 cm (which roughly corresponds to the center of the visible spectrum) we will have d5. (15) Fig. The amplitudes of the waves corresponding to rays (4) and (5) would be 2na ng .66 (dense flint glass) n f should be 1.
2kf d] (27) Assuming the amplitudes of y2 and y5 to be approximately the same.1 Mathematical Expressions for the Reflected Waves It may be worthwhile to carry out a bit of mathematical analysis for the antireflecting film shown in Fig.. (15) A was assumed to be 5000 A. for na = 1. The energy appears mostly in the transmitted beam. in Eq. The minus sign in Eq. the reflectivity would have been about 4%. although the or thick film and not latter will also give destructive interference for the chosen wavelength.2d)] ng . The transmitted wave (shown as 3) would be given by y3 = at1 cos (wt . n n (21) where I 71 I n f + na a (22) r1. In Sec. However.nf [ n f +na ng +nf 2 c of (23) (19) where tl = 2na of +na Thus.2.2.l= (26) where the phase factor is adjusted such that at x = + d we obtain the phase given by Eq. Now for a film. (28) or d= 32. such an effect is automatically taken into account when we solve Maxwell 's equations incorporating the appropriate boundary conditions.. which would be propagating in the negative xdirection] is given by y4 = . (97) of Chapter 24]:* R = r? +r2 +2r1r2 cos 26 (30) yl = a cos (cot .nf ng +nf Ir. (b) The film is nonreflecting only for a particular value of . The reduction of reflectivity is much more pronounced for the dense flint glass. the effect is not serious.2 we will discuss why we should use a 4„ thick film. 4 .atl l r2 It2 cos [cot + kax .na ng .38 and ng = 1.kax).4.2 Rigorous Expressions for Reflectivity In the above section we have considered twobeam interference and have neglected multiple reflections at the lower and upper surfaces. 16..7. The wave (5) would therefore be given by y5 a an = . 2=n f (29) 15. The reflected wave (shown as 2) would therefore be y2 15. there is merely a redistribution of energy. We assume ng > n f> na and that the xaxis is pointing downwards with x = 0 at the upper surface of the film. The displacement associated with the incident wave (propagating in the +x direction) is given by c Thus at x = 0. For example. 15. (21) represents the sudden phase change of rc at x = 0.38 and 1. the reflectivity will be about Optics a is a positive quantity. The effect of multiple reflections will be discussed in Sec. In Sec.36 respectively. For a polychromatic light. I cos [cot + kf (x .4. nf= 1.5% as one goes either to the red or the violet end of the visible spectrum..kf cl) (25) Therefore.4.7 is given by [see Eq. ka = co na (20) = .15. yl = a cos cot. We note that the refractive indices of magnesium fluoride and cryolite thick are 1. the reflectivity rises by about 0.4). In the absence of the film. r2 1 + rig r2 + 2711'2 cos 2S * Equation (30) is actually valid even for oblique incidence with and S defined appropriately (see Sec. the film's nonreflecting property will be falling off when A is greater or less than the above value. 3 g. the wave reflected from the lower surface [wave (4). 15.6 (light crown glass) n f should be 1. This technique of reducing the reflectivity is known as blooming.ca l l r. (24) Thus the displacement at x = d [associated with wave (3)] is y3 = at1 cos (cot .3%.5 the reflectivity will be about 1. 15. (c) As in the case of Young's double slit experiment there is no loss of energy.al rl I cos (wt + kax). for the MgF 2 film on crown glass at 5000 A.. 24.4 we will carry out such an analysis and will show that the reflectivity (at normal incidence) of a dielectric film of the type shown in Fig.22. 24.f 5t 4 .kfx). kf= of . destructive interference (between y2 and y5) would occur if 2kfd = z. however. . (25).
5 and nf= 1. cos 28 = . 15..nf ` Hang +n f \2 (34) 0 2000 4000 A (in Angstroms) (b) 6000 8000 where we have used Eq. however. As an example. In Fig. ng = 1. Elementary differentiation shows us that dR/d8= 0 when sin 28 = 0.2 A. the smallest film thickness is always . (31) represent the Fresnel reflection coefficients at the first and second interface respectively. if we consider A or 6123. as before. ns = 1. the thickness of the film could be 1224. ns = 1. As can be seen. 3A/4nf . and d being the thickness of the film and. the condition cos 28 = 1 implies 28 = .2114/if . Indeed for rl r2 > 0. Notice that both films are antireflecting at 6000 A. and the film is antireflecting (R = 0) when 8 = 7r/2. The film thickness is again 7J4nf where of represents the refractive index of the film. preferred.5 HIGH REFLECTIVITY BY THIN FILM DEPOSITION Another important application of the thin film interference phenomenon is the converse of the procedure just discussed.. for d = . 15.5 (37) 15.7 A../mans consistent with Eq.4%.04 represents the condition for minimum reflectivity and when this condition is satisfied. m = 0...5 A (dashed curve) and of thickness 3673. Thus for antireflecting coating.9(b) we have plotted the reflectivity as a function of wavelength for d = 1224..n fd = (2m + 1)7r . 7c. then from Eq.. implying d =2. let us suppose that we wish to make the film antireflecting at A = 6000 A.5 A (solid curve) with na = 1. which is quite close to the result obtained by using the approximate theory described earlier [see Eq.9 (a) Variation of the reflectivity of a film as a function of 5 (= 27rnf d /2.225. 3 g/2. 37v12. In Fig. (19)]. For na = 1. For example.ns f +ng .n f na +nf and r2 = n f .225. f nano 1. 15.. (18).. Notice that the reflectivity is zero for 8 = 7c/2.225 As expected.2 ns = 1.7 A or 3674.5 and of = Hang =1. 27r. an abrupt phase change of 7r occurs only at the airfilm interface and the beams reflected from the airfilm interface and the filmglass interface constructively interfere. the reflectivity [according to Eq. 57c/2.) for na = 1. represents the free space wavelength. the minimum is broad and the reflectivity small for the entire range of the visible spectrum. 2.Interference by Division of Amplitude 15.. (b) Wavelength variation of the reflectivity for a film of thickness 1224.38.. the film is such that its refractive index is greater than that of the glass.1 (minima) (33) =A nfd (32) 0. R is maximum (= 4%) when 8= 0. the glass surface is coated by a thin film of suitable material to increase the reflectivity./4n f. (26).. A. or d _ A 3A 5A 4nf ' 4nf ' 4nf ' (36) (35) Fig. viz... (31).7 A and 3674. Now. .9 we have plotted the reflectivity as a function of 8 for na = 1 and n f = Jnano = 1.5 and 17f = . consequently.7 where o rl = na . Thus the film is nonreflecting when n f = ... the reflectivity is given by 2 R= rt r2 1 rt r2 _ Hang . (34)] comes out to be 1... 1.
Such a medium is called no + An no An (b) Fig. * This section has been very kindly written by Professor K. and finally reflection from layers N/2 and N are out of phase. rl r2 < 0 (see Eq.1) 2/(2.5.225 In Sec. in addition to the phase difference due to the additional path travelled by the wave reflected from the lower interface. 15.e. then in such a case. r2 = 0.An). 15. reflections from layer 2 and layer (N/2 + 2) are out of phase etc.4 we had shown that a film of thickness A/4nf where A is the free space wavelength and nf is the film refractive index (which lies between the refractive indices of the two surrounding media) acts like an antireflection layer. n f = 2. We now consider a medium consisting of alternate layers of high and low refractive indices of no + An and no .37 and ns = 1. the reflectivity will become [see the analysis in Sec. then the reflectivity will also increase. 15. In Sec. 2 (39) For na = 1.37 2. i.5 we had shown that if the refractive index of the film was smaller (or greater) than both the surrounding media.37 .3715 2. we have ri = . This happens due to the destructive interference occurring between the waves reflected from the top and bottom interfaces.37 + 15 Optics Elementary calculations show that the reflectivity is about 33% which compares well with the value of 35% obtained by using the approximate theory described earlier. (b) If we choose a wayelength (AB + AA.1 4 x 1 x 2.37 .10(a)].1 (maxima) (38) represents the condition for maximum reflectivity.. 15.8 a film of refractive index 2.37 (zinc sulphide) then the reflectivity is (2. In the presence of a glass surface of refractive index 1.6 REFLECTION BY A PERIODIC STRUCTURE* 1' which gives about 35%. It should be noted that if the difference between the refractive indices of the film and the glass is increased. about 16%. Indeed when na < of and n f > ns.) such that reflections from layer 1 and layer (N/2 + 1) are out of phase.15. Thus.5 (light crown glass). then the reflectivity will be zero. each of thickness d = AB / 4no.407.10 (a) Reflection from a periodic structure consisting of alternate layers of refractive indices (no + An) and (no .0.4]: _ 2.37 + 1)2. 15. in such a case a film of thickness AJ4n f would increase the reflectivity rather than reduce it. there would also be an extra phase difference of r between the two reflected waves. (30) to calculate the high reflectivity obtained by thin film deposition. Thyagarajan . 31) and cos 28 = . The maximum value of the reflectivity is given by R = (rl1r1 r2 r2 .0. 15.An of equal thickness d [see Fig.37 + 1 (337) 2 x 2. We can again use Eq.
The above estimation is only an approximation which is valid when N An/no << 1.51 and 1.A( 5750 A) This is referred to as the Bragg condition and is very similar to the Bragg diffraction of Xrays from various atomic layers (see Sec. For reflection from each of the top N/2 layers. One can indeed obtain an where the first term on the LHS is simply the phase difference at AB between reflections 1 and (i + 1) due to the extra path travelled by the latter wave and the second term is that at (AB + AA). A = 2d = 1833 A.9). Thus. As an example. the period of the refractive index variation should be A = 2d = ZB 0 t 0.. If we require a strong reflectivity at A = AB = 5500 A then the required periodicity is A = 5500 2 x 15 A = Fig.. Equation (40) corresponds to the Bragg condition for normal incidence. and if we choose the thickness of each layer to be A. i. the waves reflected from each of the N individual layers are all in phase leading to a strong reflection.5. (Adapted from Ref. we first note that at AB (= 2noA). In order to do this.3 Now if An << no. i.11 that as we move away from the central wavelength (AB = 2noA) the reflectivity of the periodic medium falls off sharply.e.11 The exact variation of reflectivity with wavelength of a 100 layer periodic structure with no = 1.e.50 and An = 0. _rc 2 A2e . 6). Thus the reflectivity at 5500 A should be r R= I i5 l2 I R = 44% (41) should have an additional phase difference of jr. 18.10(b)]. 15. 18. 50 periods) then we may approximate the total resultant amplitude to be 100 x An no 1 1.. then the reflectivity will be zero. 15.1 (40) 5250 5500 . Thus when we move from AB to (AB + AA). we have 2ir no NA AA.e. here we are just trying to obtain a crude estimate of the total reflectivity.11 shows an actual calculated value of the reflectivity as a function of wavelength (using rigorous electromagnetic theorysee Ref.9 a periodic medium and the spatial period of the refractive index variation is given by A = 2d 0. Note that the actual calculation predicts a reflectivity of about 33% which compares well with our crude estimate of 44%! One notices from Fig. 15.5 where An/no is the amplitude reflection coefficient at each interface.. no = 1.see Sec. the waves reflected from the first and 1833 A If the periodic medium is made up of 100 layers (i. The quantity AB is often referred to as the Bragg wavelength. for small reflectivities. d= A 4no 4 (no + An) 4 (no . An = 0.2 and Fig.An) then the reflections arising out of individual reflections from the various interfaces would all be in phase and should result in a strong reflection. If we choose a wavelength (AB + AA) such that the reflections from layer 1 and layer (z + 1) . An = 0. Assuming AA << AB . (The argument is very similar to that used for obtaining the direction of minima in the diffraction pattern of a slit .01. 18.5). = AB = 4no d.49. Thus for strong reflection at a chosen (free space) wavelength AB. 2g NA no 2 ^B NA 2g (AB + AA) no 2 = iv (2 + 1) th layer (42) Figure 15.5. 6) for a periodic medium with no = 1. d = AB / 4no and consisting of 100 layers. there is a reflection from a corresponding lower N/2 layer which is out of phase. and so on up to the reflections from layers and N are out of phase [see Fig. approximate expression for the wavelength deviation AA from AB which will produce a zero reflectivity. If we move away from AB then the individual waves reflected from the various layers will not bean phase and thus the reflectivity reduces.01. The peak reflectivity appears at A. we consider a periodic medium comprising of alternate layers of refractive indices 1.Interference by Division of Amplitude 15. from layer 2 and (i + 2) .01.
15. Figure 15.5).6. the refractive index increases. This is indeed the principle used in white light holography. the doping results in a slightly higher refractive index. Since the fringe width would depend on the angle between the interfering beams. solid line shows the calculated spectrum and the dashed curved show the experimentally measured values of the FBG fabricated at CGCRE. The cladding material is pure silica and the core is usually silica doped with germanium. solid line shows the calculated spectrum and the dashed curve show the experimentally measured values.11. when a germaniumdoped silica core fiber is exposed to ultraviolet radiation (with wavelength around 0. 7.13(c) shows a typical frequency spectrum of the reflected wave.15. Figure 15.001 in the core of the fiber. fiber Bragg gratings etc. the guidance of the light beam takes place because of total internal reflections at the corecladding interface (see Chapters 27 and 29 for details). If the fiber is exposed to a pair of interfering UV beams (see Fig. (c) The spectrum of the reflected wave.o S7. As discussed above.46 and for the periodic structure to be reflecting at A = 1550 nm we must have AB A = "2no = 1550 nm 1.12 A Fiber Bragg Grating (usually abbreviated as FBG) is produced by allowing two beams to produce an interference pattern. 27.13(a) shows the frequency spectrum of the incident polychromatic beam.] I I I 11 I I I 1 1 1 I I I 11'1 11 III A^ 5µm Fig. 15. We will have a very brief discussion on fiber Bragg gratings below. Thus if the incident wave is polychromatic (like white light) the reflected light may have a high degree of monochromaticity. For a silica fiber no = 1.46 0.855 (47) Refractive index grating Core A. . 6 7SS0..24 pm). Kolkata.7). Bandyopadhyay of CGCRI.13. 15. Thus exposing a germanium doped silica fiber to the interference pattern formed between two UV beams leads to the formation of a periodic refractive index variation in the core of the fiber. 766. [Figure courtesy Dr S Bhadra and Dr S.531 pm 2x (46) The corresponding peak reflectivity is given by Rp = tanh2 ( B L) = 0. 15. 15. The refractive index increase can be as large as 0. an interference pattern similar to that shown in Fig. The periodic medium discussed above finds wide applications in high reflectivity multilayer coatings. Kolkata. then we would obtain UV beam which is the Bragg condition.10 or AA AB AB _A Optics O no NA L (43) where we have used Eq.13 (a) The broad spectrum of the light wave incident on the FBG (Fiber Bragg Grating) shown in (b). the refractive index of the germaniumdoped region increases. the corresponding spectrum of the reflected beam is shown in Fig. volume holography. 14.11. 2 S7 Wavelength (nm) (c) Fig. we have AA =11oA (44) which compares very well with the actual value in Fig. this is due to the phenomenon known as photosensitivity which was discovered by Kenneth Hill in 1974. 15. In regions of constructive interference. (40) and L = NA / 2 is the total thickness of the periodic medium. For the example shown in Fig. the reflection from the periodic structure will add up in phase when = AB = 2d no = Ano Bragg condition (45) 15. the period of the grating can be controlled by choosing the angle between the interfering beams (see Example 14.11 (b). We will discuss the optical fiber in Chapter 27 it may suffice here to mention that an optical fiber is a cylindrical structure consisting of a central dielectric core cladded by a material of slightly lower refractive index (see Fig.6 06.12). 15. (a) l IIII111111111 IIIIIIIIIIIIIIIilllll (b) Cladding 76.13(b). Now. We consider a polychromatic beam incident on the fiber as shown in Fig.1 Fiber Bragg Gratings A periodic structure discussed above has a very important application in the working of a fiber Bragg grating (usually abbreviated as FBG).
.17. If such a distributed sensor is put inside a bridge one can measure the strain corresponding to the particular region. peak reflectivity occur at 1544. Each grating has a different period and there Fig. (43)] Q^ i1. 15. Photo courtesy: Dr Tarun Gangopadhyay and Dr. 15. In fact for many newly constructed bridges. Figure 15.15).14. the author is with Dr. Eq.8789 nm respectively. the monochromaticity of the reflected wave) and the peak reflectivity are determined by An and L. For example.6438 nm and 1545. Figure 15.16 shows a typical reflection spectrum and the temperature recorded from the two FBG sensor shown in Fig. India) where the FBG temperature sensors have been installed. fore a specific wavelength at which peak reflectivity occurs.B _ noL 2'a 1+ ((An) L\2 i1 B (48) giving AA = 1 nm. Tarun Gangopathyay and Dr.14 FBG based temperature sensor system on 400 KV power conductor at Subhashgram substation (near Kolkata) of Powergrid Corporation of India. 15. 15. The wavelengths at which peak reflectivity occur are . FBG sensors are put at various places. As can be seen from the above equations that the bandwidth (i. One of the main advantages of the FBG sensor is the fact that several gratings can be written on a single fiber as shown in Fig.15 The substation of Powergrid Corporation of India (near Kolkata. The corresponding bandwidth is given by [cf.Interference by Division of Amplitude 15.e. Kolkata where we have assumed An = 4 x 104 and L = 2 mm. Kamal Dasgupta of CGCRI. In the photograph.18 shows the actual spectrum of the reflected light beam from a fiber on which 6 gratings have been written. CGCRI. Kamal Dasgupta.11 Fig.14 shows the FBG based temperature sensor system on a 400 KV power conductor at an electric power substation (see Fig. 15. Because silica is a dielectric material. FBG based temperature sensors become particularly useful in places where there is high voltage. FBG's are being extensively used as sensors. Because of the extremely small bandwidth of the reflected spectrum. for the two sensors. a small increase in the temperature will increase the period of the grating which will result in an increase of the peak wavelength. Figure 15. Kolkata. each having a slightly different period.
15. Thus for the first grating with 2. for example for the first grating. (48) would give (one has to be careful with the units!) _ AB A 2n 0 0.19.895 nm.01 12 1. Each grating has a length of 1cm. Kolkata.990 nm with 3dB bandwidth of 0. On the photographic plate circular fringes are obtained.5212pm 2x1.150 nm. KA.522 x 106 J giving An = 0.915 nm with 3dB bandwidth of 0. Kamal Dasgupta of CGCRI. in order to observe the film without obstructing the incident beam.915 run. and 1561.230 nm 1537. We will now consider the illumination of the film by a point source of light and. CGCRI.240 nm 1545. assuming L = 0. The wavelengths at which peak reflectivity occur are 1522.03 nm =0. 15. 15. 15.19) such that the distance SK (in Fig. the reflectivity will fall by 50% at A 1521.19) and KS (in Fig.240 nm. we will use a partially reflecting plate G as shown in Fig.20) is equal to SA + AK (in Fig.14.230 nm 1553. Obviously.19). Kolkata. L _ 1.030 rim. Kamal Dasgupta and Dr Tarun Gangopadhyay.19 Light emanating from a point source S is allowed to fall on a thin film of thickness d.01 m and no = 1.] 1522. we get 1522.15. The gratings were fabricated at CGCRI. 15.955 nm. 15.240 nm. (in Fig.950 nm with 3dB bandwidth of 0. the waves reflected from the upper surface of the film will appear to emanate from the point S' where KS' = KS (a) Fig.7 INTERFERENCE BY A PLANE PARALLEL FILM WHEN ILLUMINATED BY A POINT SOURCE In Sec.895 nm with 3dB bandwidth of 0 230 nm.030 nm with 3dB bandwidth of 0.B = 1522.17 (b) (a) The broad spectrum of the light wave incident on a fiber on which 4 gratings have been written as shown in (b). 1529.240 1522 Fig. Eq. The 3dB bandwidth means that. 1529. G is a partially reflecting plate and P represents the photographic plate. A. 1553. each having a slightly different period. 15.990 nm and 1561. once again. Photo courtesy: Dr.20) being normal to the film. (nm) 1565 nm (4.J J_1II_  ' 5 1 41 1 16 1' 12 1 31 irrr rL rr^tI(lam+ 4. 15. 15.46 Further.+1 +^ r¶rrfrr 11111 L_L_.5 nm/division) (49) Fig.522 x 10 6 1 + ( (An) x 0.030 nm.00011. [Figure courtesy Dr. in order to study the interference pattern we may assume the point source S to be right above the film (see Fig. 15. 1.46. 15.46 x 0. 15. 1537. 1545.18 The actual spectrum of the reflected wave from a fiber on which 6 gratings have been written.950 nm. However.12 Optics and 1522. 15. Each grating has a slightly different period because of which each one of them will have peak reflectivity at a different wavelength. Kolkata.910 nm Fig.955 nm with 3dB bandwidth of 0. .01 1.16 A typical reflection spectrum from the two FBG sensor shown in Fig.2 we had considered the incidence of a parallel beam of light on a thin film and had discussed the interference produced by the waves reflected form the upper and lower surfaces of the film.
an abrupt phase change of n occurs. 15.. the photographic plate will be uniformly illuminated.3.20] will be determined by the following relations: A = Cm ` = m2 represents the optical path difference and we have assumed that in one of the reflections. 15. 15. nl and n2 are the refractive indices of media I and II respectively.22).13 where A = [n1 SF + n2(FG + GH) + n I HQ] . are derived from the ray SP ' . For example.e.*** On the other hand. obtain interference fringes. for a given position of the eye. at least for near normal incidence.20) will be very nearly** the same as produced by two point coherent sources S' and S" (which is the double hole experiment of Young discussed in the previous chapter).20).g. if we put a photographic plate P (see Fig. Further. we can obtain A = 2n2d cos 0' (55) i S' 2d/ n2 I S" Fig. The intensity of an arbitrary point Q [in Fig.20 If light emanating from a point source S is incident on a thin film then the interference pattern produced in the region I is approximately the same as would have been produced by two coherent point sources S' and S" (separated by a distance 2d/n2) where d represents the thickness of the film and n2 represents the refractive index of the film. it can be shown that for near normal incidence. Thus. or (1n+ ) X.21(a)]. 15. where KS" = KS + 2d/n2 (50) (see Fig. 15.Interference by Division of Amplitude 15. Now. in general. A = 2n2 d cos 9' A more rigorous calculation shows (see Ref. which focus at a different point 0' on the retina.20)) we will obtain dark and bright concentric rings (see Example 14.20). and the rays P 'M' and Q'R ' . then it can easily be seen that the rings will spread out and in the limit of the point source being taken to infinity (i. ***If the point source is taken far away.n 1SB II is m2. Further.21(b)]. valid even for large angles of incidence. are derived from the single ray SP. 15.* Thus. 15. 15. Instead of looking at the film. not have the same intensity. A = n1SQ + n2(QA + AB) . in general. . and this point will appear to be dark or bright as the optical path difference. The above conditions are rigorously correct. with eye at the position E and the point source at S only a portion of the film around the point B will be visible [see Fig. the waves reflected from the lower surface will appear to emanate from a point which will he displaced from S".e. Such an extended source may be produced by illuminating a ground glass plate by a sodium * This is a consequence of the fact that the image of a point sou rce produced by a plane refracting surface is not perfect. simple geometrical considerations will show that the waves reflected from the lower surface will appear to emanate from the point S". We next consider the illumination by an extended source of light S (see Fig.19) we will. which focus at the point O of the retina. the interference pattern produced in region I (see Fig. 15. if we view the film with naked eye then.. 15. we will be able to see only a very small portion of the film. i. incidence of a parallel beam). Equation (50) is valid only for near normal incidence.. For large angles of incidence. e0 and 9' are defined in Fig.n12sm2 e (53) (eo e` 2 (54) + 2J maxima minima (51a) (51b) S I where the angles 9. ** The fact that this is not identical to the Young's pattern is because of the fact that S" is not a perfect image of the point S. if we put a photographic plate (parallel to the surface of the film (see Fig. then the interference is between the rays which are derived from a single incident ray by reflection from the upper and the lower surfaces of the film [see Fig.20.[n l (SA +AQ)] (52) (see Fig. if the eye is focussed at infinity. 15. e. Since the angles of refraction 9 and 02 (for these two sets of rays) will be different. the rays PM and QR. using a method similar to the one described in Sec. the points 0 and 0' will. 15. Further. 7] A = 2n2 d cos 9' 1nl sinecos0 n22 .6).
lamp. etc.22 (b) Fig. . The eye E is focussed at infinity. Each point on the extended source will produce its own interference pattern on the photographic plate P. (see Fig. Obviously. We will now discuss the interference pattern produced by a film of varying thickness. Since 0' is constant over the circumference of a cone (whose axis is normal to the film and whose vertex is at the eye). 15. rays from all points of the film will reach the eye. In Sec. are known as Haidinger fringes.23(a)]. Such a film may be produced by a wedge which consists of two nonparallel plane surfaces [see Fig. G represents the partially reflecting plate and P represents the photographic plate. S2R'. consequently. 15. In Fig. If the eye is focussed at infinity then parallel light coming in a particular direction reaching the eye would have originated from nearby points of the extended source and the intensity produced on the retina would depend on the value of 2nd cos 0' which is the same for all parallel rays like S 1 Q.n1 SB] is m) . They are also known as fringes of equal inclination because the changes in the optical path are due to the changes in the direction of incidence and hence in the value of 0'. S2Q'. 15.10 we will discuss the Michelson interferometer where such fringes are observed. 15. (b) If the eye is focused for infinity then it receives parallel rays from different directions corresponding to different values of the angles of refraction 0' (and hence different values of the optical path difference).8 INTERFERENCE BY A FILM WITH TWO NONPARALLEL REFLECTING SURFACES Till now we have assumed the film to be of uniform thickness. etc. if we view the film with our eye.) would correspond to a different value of 0' and would focus at a different point on the retina.23(b) the successive positions of the crests (at a particular instant of time) reflected from the upper surface and from the lower surface of the film are shown by solid and dashed lines respectively.22). .15. 15. We first consider a parallel beam of light incident normally on the upper surface of the film [see Fig.14 Optics P A (a) Fig. Light emanating from an extended source illuminates a thin film. The dots in the figure indicate the positions of maxima. (a) if the film is viewed by the naked eye E then the point B will appear to be dark if the optical path [{n 1 SQ + n2 (QA + AB)} .23(a)]. and bright if the optical path is (in + 1) a. 15. Such fringes. these will be displaced with respect to one another. produced by a film of uniform thickness. the eye will see dark and bright concentric rings. 15. 15. However. a photographic plate P will record straight line interference fringes which will be parallel to the edge of the wedge (the edge is the line passing through the point 0 and perpendicular to the plane of the paper). no definite fringe pattern will appear on the photographic plate. Rays emanating in a different direction (like S I R. with the center lying along the direction 0' = 0.21 Light emanating from a point source S is incident on a thin film.
however. (b) The solid and the dashed lines represent the positions of the crests (at a particular instant of time) corresponding to the waves reflected from the upper surface and from the lower surface respectively. On the other hand.AA ') XB' 2.. B L s. m = 0.Interference by Division of Amplitude 15. it when it gets reflected by the upper surface.. the next bright fringe will occur at the point B where 2nBB' 1I m . 15.12n (59) XB' = (A'X) tan 0 A'X=/3= 2nd (60) where /i represents the fringe width and we have assumed 0 to be small. (Notice that the (b) Fig.15 [see Fig. • s" .^ + 2J 11 (58) Thus (a) or But or 2n(BB' . when the wedge angle 0 is very small (which is indeed the case for practical systems) LM + MA = 2AA ' where AA' represents the thickness of the film at A. 60) is. 15.. The fringes will be perpendicular to the plane of the paper. the fringe pattern will be similar to the parallel film case.24 In order to find the distance between two consecutive fringes on the film we note that for the point A to be bright* n(LM + MA) = (m+ 2) A. independent of this condition. E represents the lens of the eye. Fig. The expression . for a point source.. 1. for near normal incidence. * We are assuming here that the beam undergoes a sudden phase change of for the fringe width (Eq.23(a)]. 15. The maxima will correspond to the intersection of the solid and dashed lines. However. (56) Light from a point source illuminating a wedge. the pattern will be very nearly the same as produced by two sources S' and S" (Fig.. Such fringes are commonly referred to as fringes of equal thickness. Thus the condition for the point A to be bright is P l 2nAA' = (in + 21 A. (57) Similarly. 15.e.23 (a) A parallel beam of light incident on a wedge.24). i. 2.
24) then only a small portion of the film (around the point R) would be visible and the point R will be bright or dark as the optical path difference [{SN + n(NL + LR) } . then no definite interference pattern will be observed.** However. the waves reaching K from SI may produce brightness. It is of interest to mention that if we focus the camera on a plane XX' . Since the thickness of the film is not uniform. one exception to this.26 The fringes formed by a wedge will be parallel to the edge 00'. 15. respectively.SR] is (m + 2) A. in order to view the fringes. the fringes are said to be localized. but if the incidence is near normal then the intensity at the point Q will be determined entirely by the thickness of the film at that place. one must focus the camera on the upper surface of the film. and in this sense. if we view the film with a cam ** There is. this is a consequence of the fact that the two surfaces of the film are not parallel. which is slightly above the film. the intensity at the point Q' will be determined by the thickness of the film at Q'. the question now arises as to how thin the film should be. then the light rays reaching the plate G will be epproxirnately parallel and an interference pattern (of low contrast) will be formed on the plate P. Till now we have assumed the film to be `thin'.25) . each point source will produce its own pattern on a photographic plate P. Since the extended source can be assumed to consist of a large number of independent point sources. It should be emphasized that all along we are assuming near normal incidence and the fact that the wedge angle is extremely small. 15. or m. 15. We next consider the illumination by an extended source S as shown in Fig. however. 15.15. the point Q' will be focussed at a different point B' on the retina of the eye. . there should be definite phase + n(AB + BC)) + CQ] . 15. or (rn+ A. This follows from the fact that the light waves reaching the point K from S2 undergo reflection at the points D2 and F2 and the light waves reaching K from SI undergo reflection at the points DI and Fl.25 Localized interference fringes produced by an extended source S. It is left as an exercise for the reader to verify that if the camera is focussed for infinity. The fringes will be straight lines parallel to the edge of the film 00' (Fig. no definite interference pattern will be recorded. when the extended source is taken to a very large distance. 15. One can similarly discuss the case when the eye is focussed for infinity. These assumptions are indeed valid for practical systems.25. ` / minima If we view the film with naked eye (say at the position E . The same phenomenon will also occur if instead of moving the extended source we take the plate P far away from the wedge.[SD + DQ] maxima (61) m + 21 = mA. no definite fringe pattern will be observed.we are assuming near normal incidence. Similarly.16 P Optics era (or with a naked eye) and if the camera is focussed on the upper surface of the film then a particular point on the film will appear dark or bright depending on the fact that whether 2nd is mA. II F. consequently. It may be seen in the figure that interference at the point Q may occur due to light coming from different points on the extended source. Thus. (see Fig. Fringes will be seen only when the eye is focussed on the upper surface of the film.) The intensity of an arbitrary point Q will be determined by the following equations: [SA Q Fig. In order to obtain an interference pattern.see Fig. whereas the waves reaching from S2 may produce darkness.26). however. point S" is not vertically below S'. F2 III Fig.
Further the magnitude of a5 is nearly equal to that of a1. 15. = 5.n ai a=0. If a source remains coherent for a timer then the coherence length (L) will be about c'rc where c is the speed of light in free space.977a a4 l +n n a2 2.1). It a = 0. The distance between two consecutive bright (or dark) fringes is determined by the wedge angle. see Fig. L .10 10 sec.Interference by Division of Amplitude 15. first from the lower surface and then from 0. if we use a white light source no fringes will be visible 2 for A > 2 x 104 cm (see Sec. * Coherence length is defined in Sec. Thus for z^ .129a a3 n + 1 a2 the contrast of the fringes becomes poorer. 15.36 It should be pointed out that interference also occurs in _ 2n _ 2n n1 _ 2 x 1. (2).17 relationship between the waves reflected from the upper surSolution: Let the amplitude of the incident ray be a and let the face of the film and from the lower surface of the film. n 1 . if we are using 1. A laser beam has a very high coherence length and fringes can be visible even 2 6 x 0.28). 13 in the prelim pages.1.9). (This is the reason why the fringes observed in transmis. is the reason why the interference fringes formed in transmission Assuming near normal incidence (6 = 0). 6 x 0.. On the other a5 n2+ 1 a3 2x hand. This Example 15.36)2 ing two reflections. the wavelength of light and by the refractive Air (na = 1) index of the film. are quite different. This is indeed what we see when sunlight falls Glass on a soap bubble or on a thin film of oil on water. when the ray gets reflected at the point B.27) between the directly transmitted a7 a6 a3 a3 n+ 1 n+ l n+ l beam and the beam which comes out of the film after suffer(2.36 x ..9 sion have very poor contrast. It should be mentioned that if the optical path difference between the waves reflected from the upper surface of the film and from the lower surface of the film exceeds a few wavelengths.149a for path differences much greater than 1 m.* For example. show that whereas the have poor contrast.847a = 0. amplitudes of the reflected rays (1) and (5) (Fig.27 In general.27) are nearly equal. If we use a polychromatic source (like an Nonreflecting film incandescent lamp) we will observe coloured fringes. the coherence length is of the order of 1 cm and for 2 2 fringes to be visible A should be much less than 1 cm.) We have seen in the previous section that if light from an extended monochromatic source (like a sodium lamp) is incident normally on a wedge. the two amplitudes We first note that the sign of a5 is opposite to that of a1 which is will be very different and the fringes will have very poor a consequence of the fact that a sudden phase change of 7r occurs contrast (see Example 15. we get pared to the coherence length.874a = 0. the amplitudes of the transmitted rays (4) and (7) are quite COLOURS OF THIN FILMS different. 15. the interference pattern will be Fig.36 o 36 x 0.36 in air. the amplitudes of (2) and (6) many colours and no fringes will be seen (see Sec. Further. On the other hand I a7 I << I a4 I.890 x 10 5 l+n a 0. a. 14.. each fringe of (< ng) representing the locus of constant film thickness (see Fig. 15. the path difference A (= 2nd cos 0') should be small com. . However. but as the value of A increases.1 Consider a film of refractive index 1. etc..36 cm).be denoted by al. whereas the amplitude of (1) and (5) washed out due to the overlapping of interference patterns of are nearly the same.023a the upper surface of the film..Using Eqs (10a) and (10b).36 region III (see Fig.847a a2 l+n a 236 should be pointed out that there is no particular value of A for which the fringes disappear. 14.9).15. if instead of a wedge we have a film of arbitrarily varying thickness we will again observe fringes..129a = 0.153a the Di line of an 2 ordinary sodium lamp (A. then equally spaced dark and bright fringes will be observed. Thus amplitudes of the rays (1)..3 cm. 17. (3).
18 Optics The thickness of the air film is zero at the point of contact 0 and increases as one moves away from the point of contact. 2. we obtain a bright fringe.29. 1. For near normal incidence (and considering points very close to the point of contact) the optical path difference between the two waves is very nearly equal to 2nt. If we allow monochromatic light (such as from a sodium lamp) to fall on the surface of the lens.. 2. 15. m = 0.S..1R. M 2t (65) A B (66) Fig. (63).2„ = t(2R 't) where R represents the radius of curvature of the convex surface of the lens (see Fig. in order to see the fringes with white light.28 A typical fringe pattern produced by an airfilm formed between two glass surfaces (which are not optically flat) and placed in contact with each other. (63) will correspond to minima. Now R = 100 cm and t < 103 cm..29 Q 0 P An arrangement for observing Newton's rings. M represents a travelling microscope. the thickness of the air film will be constant over a circle whose center is at the point of contact O. then the light reflected from the surface AOB interferes with the light reflected from the surface POQ.) Thus. Light from an extended source S is allowed to fall on a thin ' film formed between the planoconvex lens AOB arid the plane glass plate POQ. Sirohi.1. 15. .30).10 NEWTON'S RINGS If we place a planoconvex lens on a plane glass surface. (Photograph courtesy Prof. As mentioned earlier. 1. which implies that the radii of the rings vary as square root of natural numbers. 15. 2nt = I m+ 1) r m = 0. Similarly the condition 2nt m2. thus we 'may neglect t in comparison to 2R to obtain r2 . where n is the refractive index of the film and t the thickness of the film./2. the thickness of the air film will be constant over a circle (whose centre will be at 0) and we will obtain concentric dark and bright rings. Between the two dark rings there will be a bright ring whose radius will be Jm + 2 . 15. the microscope (or the eye) has to be focussed on the upper surface of the film (see the discussion in Sec. whenever the thickness of the air film satisfies the condition Fig. Since the convex side of the lens is a spherical surface. Thus./2. we get r . then (64) r.. ~ 2Rt m or r2 R Substituting this in Eq.7). we obtain a dark fringe and when the thickness is (rai+ 2) A. mA„R. (62) we will have maxima. Each fringe describes a focus of equal thickness of the film.. Let the radius of the mth dark ring be rm and if t is the thickness of the air film where the mth dark ring appears to be formed.15. Whenever the thickness of the airfilm is m. The radii of various rings can easily be calculated. The proper explanation was given by Thomas Young. 15. 15. R. These rings are known as Newton's rings.. * Boyle and Hooke had independently observed the fringes earlier but Newton was the first to measure their radii and make an analysis. Also see `Milestones' in the beginning of this chapter.31).* It should be pointed out that in order to observe the fringes. a thin film of air is formed between the curved surface of the lens (AOB) and the plane glass plate (POQ)see Fig. 15. the film should not be more than few wavelength thick. Thus the rings will become close to each other as the radius increases (see Fig.
second and third dark rings would be approximately 0. 15.134 cm respectively.D»t A (68) 4pR The radius of curvature can be accurately measured with the help of a spherometer and therefore by carefully measuring the diameters of dark (or bright) rings one can experimentally determine the wavelength. like a sodium lamp) is allowed to fall on a glass plate which partially reflects the beam. Light from an extended source (emitting almost monochromatic light. This reflected beam falls on the pianoconvex lensglass plate arrangement and Newton's rings can easily be observed by viewing directly or through a travelling microscope M.Interference by Division of Amplitude 15. the thickness of the air film (where the mth dark ring is formed) is t.30 r. 15. = pAR).rz. Actually. the diameter can be more accurately measured and in terms of the diameters the wavelength is given by the following expression: z z = D»t+^ . dark and bright patches will be seen (Fig. Two consecutive dark fringes will be separated by the air film whose thickness . then.31 and Eq. if the refractive indices of the material of the lens and of the glass plate are different and if the refractive index of the liquid lies in between the two values. (Photograph courtesy Dr. Notice that the spacing between the second and third dark rings is smaller than the spacing between the first and second dark rings. we see a dark spot and when this thickness becomes (m+ z) A/2 we see a bright spot. Normally. one really need not have Equation (69) may be compared with Eq. G. Typically for . Thus while carrying out the experiment one should measure the radii of the mth and the (m+p)th ring ( p = 10) and take the difference in the squares of the radii (r 21+p .19 planoconvex lens.0774 cm. Bose. Further.29.. An important practical application of the principle involved in the Newton's rings experiment lies in the determination of the optical flatness of a glass plate. It may be mentioned that if a liquid of refractive index n is introduced between the lens and the glass plate.28). 15. 0.0774 m cm (67) Fig. (66). (69) would give the radii of the bright rings. and R = 100 cm r.110 cm and 0.» = (mL.. Usually.» = 0.) Newton's rings can easily be observed in the laboratory by using an apparatus as shown in Fig. the radii of the dark rings would be given by r.31 Newton's rings as observed in reflection. the rings would be visible even when a biconvex lens is used. The rings observed with transmitted light are of much poorer contrast.1 = 6 x 105 cm. the central spot will be bright as in Fig. Equation (63) predicts that the central spot should be dark. If a monochromatic light beam is allowed to fall on this combination and the reflected light is viewed by a micro scope. Consider a glass surface placed on another surface whose flatness is known. 15. 15. represents the radius of the mth dark ring. in general. with the presence of minute dust particles the point of contact is really not perfect and the central spot may not be perfectly dark. which is indeed independent of m.Rln) vz (69) Fig./2. The space between the two glass surfaces forms an air film of varying thickness and whenever this thickness becomes m2t. Thus the radii of the first.
Similarly. . R (= 0.32(b)]. Q PA = A/2 QB=A. then the first dark ring collapses to the centre For A = 5.32(c)]. the bright and dark rings of Al superpose on the bright and dark rings of 22 respectively. as the lens is moved upward the rings collapse to the centre. If one carries out this experiment it will be observed that the 200th dark ring will slowly converge to the center and when the lens has moved exactly by 6.4 x 10 3 cm.3 Consider the formation of Newton's rings when two closely spaced wavelengths are present. As the lens moves gradually away from the plate. What will be the effect of the presence of these two wavelengths as the lens is gradually moved away from the plate? What will happen if the sodium lamp is replaced by a white light source? Solution: We will first assume that the lens is in contact with the plane glass plate [see Fig. by measuring the distance between consecutive dark and bright fringes one can calculate the optical flatness of a glass plate. (= 0. 15. the central spot will be dark. The ring which was originally at Q now shifts to Q2. 15. 15. if the lens is moved by 6.32 The rings collapse to the centre as the lens is moved away from the plate.4 x 10 3 cm it has exactly come to the center. we can determine the wavelength. and the radius of this ring. 15. Since the two wavelengths are very close. for example. similarly the ring at R [Fig. Example 15. 1 = 5890 A and a2 = 5896 A).098 cm and OB1 = (. OA 1 =I2tRI / 1/2 = 0.. Consequently. the D1 and D2 lines of sodium (A. the radius of the second dark ring 2will be OB = 12.0566 cm 0.6 x 105 cm) then 2t corresponding to the central spot would be A/2 and instead of the dark spot at the centre we will now have a bright spot. Example 15. RC = 3A/2 B C R1 Qi JO = A/4 Pi Ai = A/2 P al B1= A R1 Ci = 3A12 O (b) Bi C1 P2 0=A/2 Q2B2 =QiB1=QB=A R2C2 =Ri Ci =RC=3A/2 O (c) C2 Fig. Assume the radius of the convex surface to be 100 cm. Now slowly raise the lens vertically above the plate. discuss the ring pattern as seen through the microscope.32(a).890 x 105 cm.aR) = respectively [see Fig. Thus.. .32(a)] collapses to R2 [Fig. OA. This can easily be seen by calculating the radii of the ninth dark and bright ring for each wavelength.32(c)]. 15.15.4 x 105 cm. the first dark ring will form at P where PA = 2/2. will be A"??. in the present case.2 Consider the formation of Newton's rings by monochromatic light of A = 6. 200 rings will collapse to the center. 15.113 cm).20 will differ by M.. Hence if we can measure the distance by which the lens is moved upward and also count the number of dark spots that have collapsed to the centre. For example. The radii of the first and the second dark rings will be \I/2 Optics and the central spot will be dark.080 cm) see Fig. Solution: Since the point of contact is perfect.32(a)]. 15. If the lens is further moved by A/4 (see Fig. If we now raise the lens by +(=1. Assume the point of contact to be perfect.
Interference by Division of Amplitude radius of the ninth bright ring =1 15.1 A. i.230356 cm Thus the rings almost exactly superpose on each other. 15. Thus we should assume all wavelengths between A and A + A A to exist.A2) Thus. Let to be the vertical distance through which the lens has been raised (see Fig. 12(.896 x 105 x 100 = 0. the nearby dark rings for Al.A z) out.21 9 + 2 IAR 0 _ I9. Further. Consequently. if the light consists of two closely spaced wavelengths Al and A2 (like the D1 and D2 lines of sodium) then. To be more specific.896 x 10 3 = 0. will almost fall at the same place as the bright rings for A2 and the interference pattern will be washed out.893 x 10 5 )2 2t .2): LAA (72) .33 In the Newton's rings experiment. if the point J' corresponds to a dark spot for Al then it will also correspond to a dark spot for A2.2 or 2 AA 2 6 x 10$ = 3 x 102 cm This will correspond to m = 500.2t _ 1 (70) A2 Al 2 then around that point the fringe system will completely disappear. The coherence length (L) is related to AA through the following relation (see Sec. Another corollary of the above experiment consists in finding the change in the interference pattern (as we move up the convex lens) when we consider a single line of wavelength but which has a width of AA.2.890x105 x 100 = 0. By finding the approximate height at which the fringes disappear one can calculate AA. if the lens is separated by a distance to ` /= 4(1 I interference fringes will be washed 1. 15. This principle is used in a Michelson interferometer to measure the small wavelength difference AA. However.5 x 5.33) and let to be such that it satisfies the following equation: 2to 2to .5x5. the two ring patterns may produce uniform illumination. if the point J (see Fig. Thus the contrast will be zero and no fringe pattern will be visible. In this way if we continue to move the lens upwards the fringe system will reappear every time the lens is moved up by a distance A1A 2to (= 2 2) .1 2. 15. (70) we get or 2t AlA2 = 1 AIA2 2 1 (5.e. 17. We shall see the effect of the same phenomenon if we slowly raise the convex lens in the upward direction as we had considered in Example 15.230239 cm Similarly. Rewriting Eq. the bright ring for the wavelength Al will fall on the dark ring for the wavelength A2 and conversely. between two closely spaced lines (like the D1 and D2 lines of sodium).236669 cm and radius of the ninth dark ring = J9 x 5. The fringes will reappear when the distance is 2to. It should be pointed out that for complete disappearance of the fringe pattern the intensities of the two lines Al and A2 should be the same.2t1 _ 1 (71) A2 Al Fig.896 x 10 5 cm.1 2 A2 At A2 or to = 4(A1 . for large values of m. for A = 5. radius of the ninth bright ring = V9. if the lens is further moved upwards by a distance to.33) corresponds to a dark spot for Al then it will correspond to a bright spot for A2 andconversely. Now. when the air film thickness t is such that 2t = rnAl = (in + 2 1 2t . then we will have 2t1 .. The fringe pattern will reappear but now with a slightly weaker contrast (see also Chapter 17).236548 cm radius of the ninth dark ring = V9AR = 0. where t1 = 2to. Thus viewing from a microscope we will not be able to see any ring pattern.
then no definite interference pattern will be obtained on a photographic plate placed at the position of the eye.34. 3.36. One of the mirrors (usually M2) is fixed and the other (usually M1) is capable of moving away or towards the glass plate GI along an accurately machined track by means of a screw. 15. 15. GI is a beam splitter. . if we use an extended source.34] separated by a distance (xi .x 2 ).e.11 THE MICHELSON INTERFEROMETER A schematic diagram of the Michelson interferometer is shown in Fig. This can easily be seen from the fact that if xi and x2 are the distances of the mirrors M1 and M2 from the plate GI.34] gets reflected by M2 and gets (partially) reflected by GI and results in the wave shown as (6) in the figure. M2 where m = 0. then on the focal plane we will obtain circular fringes.x2). 1. i. S represents a light source (which may be a sodium lamp) and L represents a ground glass plate so that an extended source of almost uniform intensity is formed.7. and the two resulting beams are made to interfere in the following manner: The reflected wave [shown as (1) in Fig. __ M2 /////////////////// X2 Xi (3) (1) G1 (4) 7'(5) (6) X2 (2) Waves emanating from a point P get partially reflected and partially transmitted by the beam splitter GI. then to the eye the waves emanating from the point P will appear to get reflected by two parallel mirrors [MI and M2' see Fig.35 A schematic of the formation of circular fringes [Adapted from Ref.15. 15. the condition for destructive interference will be 2dcos0=m2. Now.22. Mt and M2 are good quality plane mirrors having very high reflectivity.22 Optics 15. a beam incident on G1 gets partially reflected and partially transmitted. Waves (5) and (6) interfere in a manner exactly similar to that shown in Fig.22 and 15. 15. the circular fringes will look like the ones shown in Fig. the beam reflected from the mirror M2 will undergo an abrupt phase change of 7r (when getting reflected by the beam splitter) and since the extra path that one of the beams will traverse will be 2(x1 .34 Schematic of the Michelson interferometer.35). the mirrors MI and M2 are perpendicular to each other and GI is at 45 ° to the mirror. 15.34] undergoes a further reflection at M1 and this reflected wave gets (partially) transmitted through G1. 15. As discussed in Sec. In the normal adjustment of the interferometer. and d =xi X2 Fig.. if we have a camera focused for infinity. 15. and the angle 0 represents the angle that the rays make with the axis (which is normal to the mirrors as shown in Fig. 15. 7]. The transmitted wave [shown as (2) in Fig. Instead. 2. 15. if the beam splitter is just a simple glass plate. this is shown as (5) in the figure. each circle corresponding to a definite value of 0 (see Figs 15.
if N fringes collapse to the centre as the mirror M1 moves by a distance do.73°.0 mm (c) (d) Fig. 3. the angles at which the dark rings will occur will be [see Fig. corresponding to m = 1000.Interference by Division of Amplitude 15.35).36(a) corresponds to m = 1000. . . 15.13°.IV) A where we have put 0' = 0 because we are looking at the central fringe.13°..15. the fringe pattern will expand..3 mm. say from 0. the fringes will appear to collapse at the centre and the fringes become less closely placed. 5.15.36 Computer generated interference pattern produced by a Michelson interferometer. 6. It may be noted that if d is now slightly decreased.. 495. 999. 5. 15.Thus as we start reducing the value of d.5 the dark central spot in Fig. 7.62°. 496.15 mm. 995. 2(ddo) = (m . 499. (Conversely.. the first dark ring corresponds to m = 999.28°. Thus. 5. 2.14985 mm. 6. the fringe pattern tends to collapse towards the centre.. 997. 498. 996.56°. for A = 6 x 10 5 cm if d = 0. if d is increased. as d decreases. .) Indeed. 15. 4. 497. 998. If we now reduce the separation between the two mirrors so that d = 0. Thus the central dark ring in Fig. 2d= 499.36(b)] 0 = cos 1 (500) = 0°. Similarly. Thus d=0.. etc. 3. .15 mm to 0.28°.. the angles at which the dark rings will occur will be 6= cos 1 (1000) = 0°. then we must have 2d=mA Fig.23 where the angles now correspond to m = 500. the condition for a bright ring would be 2dcos6= (m+2) For example.62°.44°.15 mm d= 0.36(b) (corresponding to m = 500) would disappear and the central fringe will become bright. ..3mm d=0.25°.
. will be visible. It is immediately obvious that the beam (5) traverses the glass plate thrice and in order to compensate for this additional path.37 In an actual interferometer there is also a compensating plate G2. 15. then Mi should be moved by 0.A2 is 1/2.. He had found that the red cadmium line (A = 6438. However. then A = 5800 A The above method was used by Michelson for the standardization of the meter.we will have disappearance of the fringe pattern and if it is equal. one introduces a `compensating plate' G2 which is exactly of the same thickness as Gt.4696 A) is one of the ideal monochromatic sources and as such this wavelength was used as a reference for the standardization of the meter.. . Let us assumethat we have a sodium lamp which emits predominantly two closely spaced wavelengths 5890 A and 5896 A. in a typical experiment.37. If we are using a 0. if the distance d is such that 2d_2d _ 1 Al A2 2 (74) and if 2d cos 0' = mAi . However. the maxima of Al will fall on the minima of A2 and conversely.15.. and the fringe system will disappear. The interferometer is first set corresponding to the zero path difference. It can easily be seen that if 2d 2d 1 . the beam splitter Gl consists of a plate (which may be about 1/2 cm thick).13 red cadmium wavelengths. 2. Michelson interferometer can also be used in the measurement of two closely spaced wavelengths. for A. Indeed.A2 Al (75) then interference pattern will again reappear.9). the refractive index of crown glass is 1. 14. 3.. 15.1)t introduced by Gl can be compen sated by moving the mirror Ml by a distance (n . = 6560 A and 4861 A. * The zero path difference is easily obtained by using white light where only a few coloured fringes. one would observe a few coloured fringes around the point corresponding to zero path difference (see Sec. for a white light source it is not possible to simultaneously satisfy the zero pathdifference condition for all wavelengths.then the interference pattern will appear. occur at the same angle as A. lying between A and A + AA. if the source consists of all wavelengths. Thus. In an actual Michelson interferometer. the path difference between any pair of interfering rays (see Fig. the difference between the two positions corresponding to over hundred wavelengths! Thus. in the presence of the compensating plate G2. then no interference pattern will be observed if Fig. if one finds 1000 fringes collapse to the centre as the mirror is moved through a distance of 2. then the maxima corresponding to the wavelength Al will not. o N (73) This provides us with a method for the measurement of the wavelength. If the mirror Ml is moved away (or towards) the plate Gl through a distance d.5330 respectively. if we have a continuous range of wavelengths from 4861 A to 6560 A. the back surface of which is partially silvered and the reflections occur at the back surface as shown in Fig. In general. In fact he defined the meter by the following relation: 1 meter = 1553164.34) will vary so rapidly with wavelength that we would observe only a uniform white light illumination. around d = 0.90 x 102 cm. The compensating plate is not really necessary for a monochromatic source because the additional path 2(n .1)t where n is the refractive index of the material of the glass plate G1. 15. they accuracy is almost one part in 10'. then 2d cos e' = (m+ 2) A2. Instead of two discrete wavelengths. For example. since the refractive index depends on wavelength.2622 cm for A = 6560 A and by 0. if 2d 2d Al .24 Optics A. to 1. 3/2.* Near d = 0.. both the fringe patterns will overlap. For example. 5/2.5244 and 1. in general.5 cm thick crown glass plate as G1.2665 cm for A = 4861 A..
On of equal thickness d is called a periodic medium and the spatial period of the refractive index variation is denoted by A (= 2d). Light of wavelength 5000 A is incident vertically on the wedge and the film is viewed from the top.578 Summary ♦ If a plane wave is incident normally on a thin film of uniform thickness d then the waves reflected from the upper surface interfere with the waves reflected from the lower surface. Calculate the wavelengths (in the visible region) for which the film will be nonreflecting.8 along a line such that a wedge of 0. For An << no.5° is formed. What will be the qualitative difference in the fringe pattern and what will be the new fringe width? 15. if d = 4n0 (where A is the free space wavelength). Problems 15.Interference by Division of Amplitude 2d 2d 1 A+0A 2 2 or 2d> AA (76) 15. For _5 A=6 x 10 cm what will be the radii of the 9th and 10th bright rings? 15.578 =6A (5890 x 107)2 mm 0. the radius of curvature of the curved surface is 50 cm. Thus for strong reflection at a chosen (free space) wavelength AB. Example 15. Repeat the calculations for the thickness of the film to be 45 x 104 cm. Show that both the films will be nonreflecting for a particular wavelength but only the former one will be suitable. The whole apparatus is immersed in an oil of refractive index 1. then the light reflected from the curved surface interferes with the light reflected from the plane surface. the accuracy is almost one part in 109. He had found that the red cadmium line (A. the thickness of the air film will be constant over a circle and we will see concentric dark and bright rings.4 In the Newton's rings arrangement.2). Assume that its thickness is 9 x 10 6 cm. If we allow monochromatic light (such as from a sodium lamp) to fall (almost normally) on the surface of the lens. = 6438. = 5890 A. calculate the wedge angle. ♦ A medium consisting of a large number of alternate layers of high and low refractive indices of no + An and no .4 For a sodium lamp. 15. Assume . the distance traversed by the mirror between two successive disappearances is 0. and that there are 100 interference fringes per centimetre.578 mm. a thin film of air is formed between the curved surface of the lens and the plane glass plate. the period of the refractive index variation should be A = 2d = Ae 2n0 This is referred to as the Bragg condition. The radii of the concentric rings are such that the difference between the square of the radii of successive fringes is very nearly a constant.5 In the Newton's rings arrangement.289 mm. the radius of curvature of the curved side of the pianoconvex lens is 100 cm. the reflections arising out of the individual reflections from the various interfaces would all be in phase and would resultin a strong reflection.38.578 A A+AA = 1 or Solution: 0A c /12 0.6 is in contact with another glass plate of refractive index 1. The radii of the 9th and 16th In this case the fringes will not reappear because we have a continuous range of wavelengths rather than two discrete wavelengths (see Sec. ♦ The Michelson interferometer was used by Michelson for the standardization of the meter. the wave reflected from the upper surface interferes destructively with the wave reflected from the lower surface and therefore the film acts like an antireflection layer._ Why 15. This is the prin ciple of operation of Fiber Bragg gratings. Since the convex side of the lens is a spherical surface. .1 A glass plate of refractive index 1. Calculate the difference in the wavelengths of the Dl and the D2 lines.7.4696 A) is one of the ideal monochromatic sources and as such this wavelength was used as a reference for the standardization of the meter.3 Consider a nonreflecting film of refractive index 1.2 Two plane glass plates are placed on top of one another and on one side a cardboard is introduced to form a thin wedge of air. 17. Calculate the fringe spacing. Assuming that a beam of wavelength 6000 A is incident normally. Indeed. Thus 0.13 red cadmium wavelengths.25 ♦ If we place a pianoconvex lens on a plane glass surface. When the mirror moves through a distance 0. In fact he defined the meter by the following relation: 1 meter = 1553164.289 mm. These rings are known as Newton's rings. the additional path introduced is 0. ♦ Michelson interferometer can also be used in the measurement of two closely spaced wavelengths.578 _ 0. for a film of thickness 2/4nf [where A is the free space wavelength and of is the film refractive index which lies between the refractive indices of the two surrounding media].
[Ans. why?) (Ans. Assume that the radius of curvature of the curved surface is 400 cm.) 15. Mallick. New York. E. Pergamon Press. Longmans Green and Co. Calculate the wavelength. 5015.. Optical Electronics. H. What will be the corresponding values of 9'? Assume 2. Calculate the wavelength. S. Born and E. M. 4. Cave and L. J. 11. A. Wolf. Francon and S. K. Oxford. White. Principles of Optics. Ghatak and K. London. 1989.9 In the Michelson interferometer arrangement. Jaffe. London.. J. 2.. 610. 0. Iliffe Books Ltd. F.6 In the Newton's rings arrangement if the incident light consists of two wavelengths 4000 A and 4002 A calculate the distance (from the point of contact) at which the rings will disappear.= 5 x 105 cm. Berlin. 8. Clarendon Press. (Ans. H.2235 cm. . `Interference in an optical wedge' Amer. Cambridge University Press.08 mm. Interference of Light. 1976. F. (Ans. 4 cm) 15.26 dark rings are 0. 250 fringes cross the field of view. Cook. 3. 1955. Interferometmy. Atlas of Optical Phenomena. 1955. V. Vol 24. Calculate the radii of the dark rings. Oxford.18 cm and 0. Inexpensive Michelson interferometer'. 10. Oppenheim and J. J. M. M.6 if the lens is slowly moved upward. Phys. Fundamentals of Optics. An Introduction to Interferonietry. calculate the height of the lens at which the fringe system (around the center) will disappear. V. Optical Interferometey. Assume 1 = 6000 A. 1956. [Ans. Steel.10 The Michelson interferometer experiment is performed with a source which consists of two wavelengths 4882 A and 4886 A. [Reprinted by Foundation Books. The radii of curvatures of the two surfaces of the upper lens are 50 cm and those of the lower lens are 100 cm.15. if one of the mirrors is moved by a distance 0.2 mm) 15.0447 m cm] 15. 0. 1971.8 An equiconvex lens is placed on another equiconvex lens. 6400 A] 15. Holroyd. 1968. Show that if d is decreased to 4. Amer.298 mm] In the Michelson interferometer experiment.. Phys.. 1971. A. A. 1975. McGrawHill Book Co. New Delhi. W. 9. Academic Press. SpringerVerlag. The waves reflected from the upper and lower surface of the air film (formed between the two lenses) interfere to produce Optics Newton's rings. (flint: The use of Eq. 61. H. 5. Tolansky. Jenkins and H. Cambridge University Press.7 In Problem 15. (66) will give wrong results. Sladkova. Vol 23. calculate the 15. 0. London. REFERENCES AND SUGGESTED READINGS 1. Through what distance does the mirror have to be moved between two positions of the disappearance of the fringes? [Ans. E. 1967. London.997 x 103 cm the fringe corresponding to in = 200 disappears. Interference of Electromagnetic Waves. M. Cagnet. 6. Thyagarajan.] 7. Francon.11 various values of 6' (corresponding to bright rings) for d = 5 x 103 cm. New York 1966.
the FabryPerot interferometer consists of two partially reflecting mirrors (separated by a fixed distance h) placed in air so that nl = n2 = 1. of refractive index nl as shown in Fig..1 INTRODUCTION In the last two chapters... These beams (on either side) interfere to produce an interference pattern at infinity.. their joint effect is a Combination of the Motions belonging to each..When two Undulations . We will show that the fringes so formed are much sharper than those by two beam interference and.h cos 0(1) S = 2^ p = A'o Ao represents the phase difference (between two successive waves emanating from the plate) due to the additional path traversed by the beam in the film (see Sec. In this chapter. we will discuss interference involving many beams which are derived from a single beam by multiple reflections (division of amplitude).1. 4nn.2 MULTIPLE REFLECTIONS . (1). h the film thickness and 20 is the free space wavelength. we have been discussing interference between two beams which are derived from a single beam either by division of wavefront or by division of amplitude. The wave will undergo multiple reflections at the two interfaces as shown in Fig. 15. Let Ao be the (complex) amplitude of the incident wave.FROM A PLANE PARALLEL FILM We consider the incidence of a plane wave on a plate of thickness h (and of refractive index n2) surrounded by a medium * The author found this quotation in Ref. 16..2 eis 2 m 1r2 e (2) . the interferometers involving multiple beam interference have a high resolving power and hence find applications in high resolution spectroscopy. + tl t2 1. coincide either perfectly or very nearly in Direction. 16. therefore. Thus the resultant (complex) amplitude of the reflected wave will be Ar =Ao[r i +t1 t2 r2 e`s (1+rz ets+r2e2`s+. 02 is the angle of refraction inside the film (of refractive index n2).* Thomas Young (1801) Important Milestones Marie Fabry and Jean Perot invented the FabryPerot interferometer which is characterized by a very high resolving power. if a plane wave falls on a plane parallel glass plate. Thus the amplitude of the successive reflected waves will be Ao rl Ao t1 r2 t2 e'8 Ao t1 r2 e2i6 . Let r1 and t1 represent the amplitude reflection and transmission coefficients when the wave is incident from nl towards 112 and let r2 and t2 represent the corresponding coefficients when the wave is incident from n2 towards n1. for example.1) and in Eq. 16. Thus. 1. as we will discuss later.)] where = A0 1 16. then the beam would undergo multiple reflections at the two surfaces and a large number of beams of successively diminishing amplitude will emerge on both sides of the plate..1(a)..
without any loss of generality.. the amplitude of the successive transmitted waves will be is.7). if the reflectors are lossless. Ao t1 t2 r2 e2ts.] =A tit2 1. Ao t1 t2.R)2 At T= Ao (1.12) R=4= r2 is called the coefficient of Finesse. The same intensity distribution is obtained in the two beam interference pattern (see Sec. Thus the resultant amplitude of the transmitted wave will be given by At = Ao ti t2 [1 + r eis + r2 e2i8 +.16. =R = Ao R (a) O P (1.R cos 3)2 + R2 sin2 S . this circle will be bright or dark depending on the value of O.R)2 + 4 R sin2 or.. 16.. F is small and the reflectivity is proportional to sin2 8/2. (b) Any ray parallel to AB will focus at the same point P.r1. we may mention here that we have obtained sin 2 3/2 instead of cos2 3/2 because of the additional phase change of iv in one of the reflected beams. Now. Ao ti t2 r2 e where... Thus the reflectivity of the FabryPerot etalon is given by 2 2 A. 14. then the point P will rotate on the circumference of a circle centred at the point 0. the reflectivity and the transmittivity at each interface are given by (see Sec. 14. Rays incident at different angles will focus at different distances from the point 0 and one will obtain concentric bright and dark rings for an extended source..cos 3)2 + sin 2 S (1.r2 =A e`S 1R o 1. A F sin2 2 2 (3) 1+Fsin2 where "iX O' P' F= 4R (1. One can immediately see that when R << 1.R e`S Z=t1 t2 = 1R Thus the transmittivity T of the film is given by 2 (1.1 (a) Reflection and transmission of a beam of amplitude Ao incident at an angle ei on a film of refractive index n2 and thickness h.R)2 (4) (b) Fig. If the ray AB is rotated about the normal at B.R e`S where we have used the fact that r2 = . we have assumed the first transmitted wave to have zero phase..2 Optics (We are reserving the symbol T for the transmittivity of the FabryPerot etalon. Similarly.R cos S)2 + R2 sin 2 S 4 R sin2 2 2 (1.) Thus Ar (1.R) eiS = r1 1Ao 1.
let T= Thus F sinz 2 for 5 =2m7r ±2 In this section. * In the visible region of the spectrum. In almost all cases. the incident light will be completely transmitted (i. 4S OS =1 (7) The quantity OS represents the FWHM (Full Width at Half Maximum).90 in the blue region). we will discuss the FabryPerot interferometer which is based on the principle of multiple beam interferometry discussed in the last section. if we consider light of a specific wavelength 4. (6) In Fig. . we may neglect the presence of the plates and consider only the reflection (and transmission) by the metallic film. silver is the best metal to coat with (the reflectivity is about 0. 16. m=1.3 THE FABRYPEROT ETALON 2 (5) It is immediately seen that the reflectivity and the transmittivity of the FabryPerot etalon add up to unity. we will obtain a fringe pattern consisting of concentric ringseach bright ring will correspond to a particular value of m. T=1 when 3 =2m7r .2. the transmittivity will be very small. to a very good approximation. 16. For large values of F. Hence. L. T = 1) if the angle of incidence is such that S= 0 n2 Thus the transmission resources become sharper as the value of F increases (see Fig.3 The FabryPerot etalon In a typical experiment. These two plates are kept in such a way that they enclose a plane parallel slab of air between their coated surfaces. for a given wavelength.2 we have plotted the transmittivity as a function of Sfor different values of F..97 in the red region and decrease to about 0. 16. light from a broad source is collimated by a lens and is passed through the FabryPerot etalon as shown in Fig. But beyond the blue region. If the reflecting glass plates are held parallel to each other at a fixed separation. Aluminum is usually employed below 4000 A. 16. The interferometer (as shown in Fig.3) consists of two plane glass (or quartz) plates which are coated on one side with a partially reflecting metallic film* (of aluminum or silver) of about 80% reflectivity. it is given by 4  2(1R) (8) Fig. 16. if the plates are parallel. at the focal plane of the lens L. we have what is known as a FabryPerot etalon. 16. Thus. In order to get an estimate of the width of the transmission resources.Multiple Beam Interferometry or T= 1 1+ F sin 2 16.e. further.3 16.r Fig.3. The transmission resonances become sharper as we increase the value of F. The sharpness of the bright rings (and hence the resolving power of the etalon) will increase with the value of F. the rays will not undergo any deviation. when 02 is slightly different from the value given by the above equation. the value of m is usually large... In fact.S <<< 1 and therefore.2)..2 The transmittivity of a FabryPerot etalon as a function of S for different values of F. the reflectivity falls rapidly.3. 1 hcos02 =2m7c (9) or cos 02 = ma° 2n2 h (10) 2 m7r S (2 m+2). Further. The FWHM (Full Width at Half Maximum) is denoted by O&.
The next two closely spaced rings correspond to in = 39999 for the two wavelengths. 39997. This is shown as the thick curve in Fig. corresponding to m = 40000.0 cm and F = 400.. h = 1 cm and F = 400 (F = 400 implies R = 0.98 A (=A).4 The variation of intensity with 0 for a FabryPerot interferometer with n2 = 1.5 x 106 = 1. 0 (degrees) Fig.=2=4999.2 0.4 . corresponding to .41°.5 The (computer generated) ring pattern as obtained (on the focal plane of a lens) in a FabryPerot etalon with n2 = 1.3. the two uncoated surfaces of each plate are made to have a slight angle between them (1 to 10 minutes see Fig. If we compare the results obtained in Example 16.5. the fringes corresponding to the wavelengths 5000 A and 4999.128°. 0. In order to see this. corresponding to Ao = 5000 A (=Al) and 20 = 4999.98 A (= "2) first ring corresponds respectively to 20 = 5000 A and 7b = 4999. 16.595°./20 (=250 A=2. 39998. 0.e.. each mirror of the etalon has about 90% reflectivity).5. 39999 and 39998 respectively. 0. both corresponding to m = 40000. The actual fringe pattern (as obtained on the focal plane of a lens of focal length 25 cm) is shown in Fig.A= 2. Indeed. The corresponding ring patterns as obtained on the focal plane of the lens is shown in Fig.1 Flatness of the Coated Surfaces In order to have sharp fringes.3 0. In Fig.0 = 5000 A (=A 1) and A = 4999.A.0 = 5000 A and 4999..98 A.5 JJL 0. 16.98 A will start overlapping.0 cm and F = 400. In the figure. corresponding to m = 40000. Now..98A Fig.02 A are quite well resolved by the etalon. we will find that if there is a variation in the spacing by about A/20.5 IL\ 0.4 we have plotted the intensity variation with 0 for X1.16) .4.162°.. for n2 Optics Ao =A t =5000A Equation (9) gives us = cos l (40000) Thus bright rings will form at 02 82 = 0°.425°. Further..98 A we get 02 =cost (40000.. 16.6 For A0 = 5000 A. the coated surfaces should be parallel to a very high degree of accuracy. we assume that in the above example h is increased by X1. 0. from the figure we can see that the two spectral lines having a small wavelength difference of 0.1 0. 0. 0. for Ao = = 4999.1.436°. 16. 0... 16.98 A.5 x 106 cm): h = 1 + 2.4 Example 16. 16. 16.. h = 1.905.. the coated surfaces should be flat within about X1/50 where A is the wavelength of light. On the other hand. Thus bright rings will form at 02 = 0.57°. i.1 and bright rings will form at 02 = 0. h =1. respectively. 16.587°..1 As an example. Thus the coated surfaces should be parallel within a very small fraction of the wavelength..4. the central bright spot and the .1 = 5000 A . we assume an etalon with = 1. 0.70°.3) so that one could avoid the unwanted fringes formed due to multiple reflections in the plate itself. we will have 02 cost ( 40000. 16.16.0000025 cm 0. This is shown as the thin curve in Fig.1. 39999.
16.54 I 1. (11) that the five lines correspond to m = 399998. (um) I 1. 9). In terms of the frequency c v= 0 Equation (9) tells us that transmission resonance will occur when c (11) v == m 2h where m is an integer.55 I 1.53 I 1. we vary the separation h and measure the intensity variation on the focal plane of the lens L as shown in Fig. 16. For h = 10 cm. v0 ± 5v and v0 ± 2 8v as shown in Fig. For a beam incident normally on the interferometer. Thus a frequency spectral width of 7000 MHz (around 6x10 = x 10 14 Hz) implies a wavelength spread of only 0.2 x 10 5 giving .3. v Fig.005 ym (After Ref.AA. The intensity variation is recorded (by a photodetector) on the focal plane of the lens L.6 A beam having a spectral width of about 7000 MHz (around vo = 6 x 10 1t Hz) is incident normally on a FabryPerot etalon with h = 10 cm and n2 = 1. The above equation represents the different (longitudinal) modes of the (FabryPerot) cavity. o = 9 7 x 104 = 1.57 I 1.Multiple Beam Interferometry 1500 MHz 16.um.v .56 A. The output has five narrow spectral lines.6.06 A. ° = ^ = 6 x 10 14 Hz. 16.005 . * For .2 Modes of the FabryPerot Cavity We consider a polychromatic beam incident normally (02 = 0) on a FabryPerot etalon with air between the reflecting plates (n2 = 1) . One can readily calculate from Eq. 16.see Fig.5 +. 1b = 5000 A and a spectral width of 7.7 shows a typical output of a multilongitudinal (MLM) laser diode.06 A.000 MHz would imply v0 = 0.58 I 1. 16. Such an arrangement is 16. the system is called a FabryPerot interferometer. the frequency spacing of two adjacent modes would be given by Sv I 1.59 = 212 = 1500 MHz For an incident beam having a central frequency of v =v0 =6x 1014 Hz and a spectral width* of 7000 MHz the output beam will have frequencies v0. 16. The wavelength spacing between two modes is about 0.8 A scanning FabryPerot interferometer. 399999.8.o v0 Fig. Fig. 16.4 THE FABRYPEROT .INTERFEROMETER If one of the mirrors is kept fixed while the other is capable of moving to change the separation between the two mirrors.7 Typical output spectrum of a FabryPerot multilongitudinal mode (MLM) laser diode. 400001 and 400002 Figure 16.6.52 I 1. 400000. the wavelength spacing between two modes is about 0.
.10 shows variation of intensity at the point P when the incident beam has two frequencies separated by 300 MHz. which will occur when x = 0. 250 nm. The quantity = 800000 ir (i + h J 0 Ovs = 2ho (13) Thus transmittivity resonances will occur for S = 8000007r. 16. 500 nm.5 3= 800.8) for a monochromatic beam incident normally on a scanning FabryPerot interferometer.. is known as the free spectral range (FSR) of the interferometer. if v=vo+ If the incident beam is monochromatic. 16. 547r we will have the same T vs.. 0. n2 2h0 then one can easily show that transmission resonances will occur at the same values of x. 16.000 Jr 6= 800.5 RESOLVING POWER We will first consider the resolving power corresponding to a beam incident normally on a scanning FabryPerot interferometer. The two curves in Fig. 16. Since the separation h is varied. 3. the corresponding values of S will be 800002ir (corresponding to x = 0). we get 1.9. we write it as h = ho + x (12) Optics We may mention here that if the frequency of the incident beam is increased by c/2ho. p = = 1 and cos c 02 = 1. 8000047r.e.. 16. . 16. Obviously. Indeed if v = vo ± p Zhfl .. This will be followed by the case corresponding to the FabryPerot etalon. Thus when the spectrum has widely separated wavelength components. the two frequencies are well resolved. the solid curve corresponds to F = 1000 and the dashed curve corresponds to F = 100.. 8000047r (corresponding to x = 250 nm) etc. Fig.004 jr 0 Fig.. F=100 F= INC). a typical variation of intensity at the point P is shown in Fig. 2. we will have overlapping of orders.. 8000027r. The figure corresponds to the frequency of the incident beam being v=vo=6x1014 Hz For ho = 10 cm. respectively. x curve.16:6 usually referred as a scanning FabryPerot interferometer.. Notice that the transmission to F = 100 and F resonances become sharper if we increase the value of F..9 correspond 1000. i.9 250 500 x (mm) Variation of intensity at the point P with x (see Fig.
2 1 . as discussed in Sec.11.4  =2m2 OS (14) l1 +72 1.Multiple Beam Interferometry .8) when the incident beam has two frequencies separated by 300 MHz.7 1 v=v0 =6x10 14 Hz 3= 800.11 The individual intensity variations II and I2 in the presence of two frequencies vl and intensity variation (II + 12) when the two frequencies are just resolved.v=v0 300 MHz 16.10 Variation of intensity at the point P with x (see Fig. 16. 16.1 Resolving Power of a Scanning FabryPerot Interferometer We consider the presence of two frequencies vl and v2 of equal intensity in the beam incident normally on a scanning FabryPerot interferometer.002 ir 100 0 50 250 300 400 x (nm) Fig. this happens. if the half intensity point occurs at 31/2 16... v2 and the total . For the two frequencies to be just resolved.0. the minimum of the resultant intensity distribution (shown as the dashed curve in Fig.5. 16. we assume that the half intensity point of vl falls on the half intensity point of v2 as shown in Fig. Now.. 16.11) is about 74% of the corresponding maximum value.2 0 30 20 10 0 x (mm) 10 20 30 Fig. When 1.. 16.2. 16.
27cv 1 (29) Using Eq. for the frequency v1.013 = 0. in terms of the wavelength Resolving power For h=1cm. we have dropped the subscript on µ and e. (9)] = 47cv (24) hµ = 2 m7r c where µ = cos 0. let the half intensity point oc. for the frequency v1. dropping the subscript we get Resolving power = v Lv 7r hv c Ov1 11 Dµ1 7rv1 hJµ1 c (31) Or. Ai) (23) 7rh. h = 1 cm. let the half intensity point occur at h = hi + 6 h1 (the corresponding value of S will be 2mnc + lh OS1 . Consider the frequency mum occurs at h = h1 then =2m7r c Let the intensity maximum for v = 61= h=h2 =h1 +Oh 1 Thus S_ 2 4zthlv1 (16) v2 (= vi + Ov 1) occurs at 47r (h1 + Ohl )(v1 +Ov1) c = 2m (17) Using Eqs (16) and (17) and neglecting the second order term Ohl Ov 1 . 16. Thus. (18) we get for the resolving power v1 hl Oh 7r h1 vi c J f As discussed earlier. (27) we get (22) Resolving power = v1 Dv (30) k 7cvl C g' Or.8 Optics then DS = 4 v1. if the mth order intensity maxima for v = v1 and v = v2 (= v1 + Ov1 ) occur at u1 and = µ2 (= µ1 + Dµ1). T = 1 if the angle of incidence is such that [see Eq. 16. neglecting the second order term we get (26) Equation (18) implies that for Lh1 to be positive. thus using Eq.9).Ali =6x105 cm A^ = 0. (16) 4irv1 Sh1 _ 1 c 2 or Ohl = c v1 5 IT 2 (19) (27) Ov1 i Now.. We can now have arguments very similar to that in Sec.3. (24) 4µl =  µ (20) or (21) 4lrv1 hSµ l 1. We once again consider the presence of two wavelengths . for the two frequencies to be just resolved.lF cos 0 A'o Duo . = 2 OS' = 2 c (28) For the two frequencies to be just resolved ^h1 = 2 Ohl = c^ 7CV l c Sµ1 . and for the sake of simplicity.006 A A Or.2 (32) A for F = 80 for F = 360 x 106 . thus using Eq. Now. Ao Thus for F = 360 (R = 0. we get v1 µ µ= Ohl + hi Ov1 = 0 O h1 =  S hl 1 _ 4icv1hµl =2m7r c (25) or Ov1 (18) and 4irh(vl + Avl) (µ1 + Aµ1) = 2 m 7c c Thus. we assume that the half intensity point of v1 falls on the half intensity point of v2 giving 4µ1 = 2 Sµ1 = Using Eq. then S [see Eq. (15) If the intensity maxi 16.1 and 12 of equal intensity.5. iv1 should be negative.5. in terms of the wavelength 7th l/.1 except now h is fixed and it (= cos 0) is varied. + Sµ1 (the corresponding value of S will be cur at = 2 m 7r + ½A S1). X10 = 5000 Resolving power = AT Duo = 1. (8)]. Now.2 Resolving Power of a FabryPerot Etalon We consider light from a broad source incident on a FabryPerot etalon as shown in Fig.16.
But one cannot use every thick coating of metals to increase the reflectivity as the intensity of the beam would be reduced considerably due to absorption in metallic coatings. When the FabryPerot interferometer is used to analyze spectra with closely spaced lines. This is in contrast to that of a grating (say having 25000 grooves) which resolves up to about 0. In the plane P. 24. . But when the spectrum has widely separated wavelength components.004 A Thus a FabryPerot instrument can resolve wavelengths differing by about 103 A. one obtains fringe patterns on either side of the plate. We will not go into the details of the theory of the LummerGehrcke plate but two points may be noted: where we have assumed normal incidence. on one end of which a small rightangled prism of the same material is fixed (see Fig. the lines at the two wavelengths A and A + AA themselves will have a wavelength spread and this restricts the use of such high resolving powers. The difference in wavelength (AAS) which corresponds to a displacement of one order.6 and 16.2). 16. we have considered two monochromatic lines at A and A + AA. ** Beyond the critical angle. 16. Thus. For example. A2 LAS = 2n h (34) P (b) Fig.9 which is found to be inversely proportional to h.12(b). each with a hyperfine structure. then the distance between the adjacent maxima would be greater than the displacement between the system of rings of the spectral lines. 16. When the spectrum is complex consisting of a number of widely separated wavelength components. The spectrograph separates the spectral components and one obtains in the plane P images of the slit. This is in contrast to the resolving power which depends directly on h [see Eqs (31) and (32)].12 (a) A FabryPerot interferometer used in .1 A at A = 5000 A and that of a prism (made of dense flint glass with 5 cm base) which resolves only up to about 1 A at 5000 A. The angle of the prism is chosen in such a way that the rays incident normally on the surface of the prism hit the two surfaces of the plate at an angle slightly less than the critical angle. The above equation gives = 0. (ii) and (iii) may correspond to the lines in the red. 16. then one can separate the different wavelength components by employing the FabryPerot interferometer along with a spectrograph as shown in Fig.conjunction witha spectrograph. Thus we can write OAS 2nhcos 0 (33) This becomes.12(a). (b) The interlaced fringes formed in the plane of the slit are separated by the prism. A LummerGehrcke plate is a plane parallel made of glass (or quartz). all successive reflections will occur at the same (near critical) angle. Notice that the prism suppresses the externally reflected beam. The fringes are approximately straight lines parallel to the plate surfaces. (i). The results in the `overlapping' of orders (see also the discussion at the end of Sec.6 THE LUMMERGEHRCKE PLATE* We saw in Sec. there will emerge from the upper and lower surfaces of the plate a series of waves which would finally interfere to produce interference fringes in the plane P (see Fig. the reflection is total while slightly below the critical angle.13). It must be noted that in the above analysis. the reflectivity is high (see Sec. 16.13). yellow and green regions respectively as observed on plane P. 16. The light emerging from the source S is rendered parallel by the lens LI. 16. for near normal incidence (B = 0). Most of the light will be reflected with a little fraction being transmitted at each reflection. In general. 16. This difficulty can be overcome by the use of the phenomenon of total internal reflection (instead of metallic reflection). The interference pattern formed by the FabryPerot interferometer (marked by FP in the figure) is made to fall on the slit of the spectrograph.** Since the two surfaces are parallel.7 have been very kindly written by Professor Anurag Sharma. this is used in the LummerGehrcke plate which will be discussed in this section.4). then it might happen that the displacement between the rings is greater than the separation between adjacent maxima. each crossed by fringes as shown in Fig.Multiple Beam Interferometry 16.2 that the sharpness of friniges (and hence the resolving power) of a multiple beam interferometer increases as the reflectivity R of the plate increases. is called the spectral range of the interferometer. * Sections 16.
the filtered light will have a finite width. if n = 1. For example. of aluminum or silver) is deposited on a substrate (generally. 16.14). In this way. However. we get a spectrum consisting of different intensity maxima which satisfy the following relation: 2nhcos Or m2 Fig. LummerGehrcke plates were used in high resolution spectroscopy.000 maxima are observed if h = 1 cm. is deposited. Earlier. it has been replaced by the more flexible FabryPerot interferometer. a A/4 thick film of titanium oxide is deposited on a glass substrate. On this is again deposited a X1/4 thick layer of a. the reflectivity of the glass plate increases. Thus if the incident wave is polychromatic (like white light).* Interference filters using this principle can be obtained by modem vacuum deposition techniques. Cover plate >i Metal films Dielectric film Substrate 16.13 The LummerGehrcke plate. In an alldielectric structure. To increase the reflectivity. If. about 23. maxima of different orders are formed in the transmitted light corresponding to wavelengths given by (36) m If h is large. This structure is again 'covered by another metallic film (see Fig. Larger the difference between the refractive indi ces. that is. To protect this film structure from any damage. The materials generally used in interference filters are titanium oxide (n = 2. The larger the reflectivity.05 cm. if we go on reducing h. (35) Now if a FabryPerot interferometer is illuminated with a collimated white light incident normally (Or 0). It was shown in Chapter 15 how dielectric films can be used to enhance the reflectivity of a surface.10 Optics L Fig. we reach a. multilayer structures of alternate higher and lower refractive index materials are used. To overcome this difficulty. By varying the thickness of the dielectric film. Thus. The sharpness of the transmitted spectrum is determined by the resolving power of the formed FabryPerot structure.AlF3) is deposited over this. there are only two maxima in the visible region. corresponding to A = 6000 A (m = 3) and A = 4500 A (m = 4).7 INTERFERENCE FILTERS When a FabryPerot interferometer is illuminated by a monochromatic (uncollimated) beam. 16. To obtain interference filters.13).16. and (b) The number of reflections is also not very large as in the case of the FabryPerot interferometer. a large number of maxima will be observed in the visible region. 16.14 The interference filter. the narrower is the transmitted spectrum. another glass plate is placed over it. In this way. material of higher refractive index. Then a thin layer of a dielectric material such as cryolite (3NaF. They are widely separated and one of them can be masked so as to transmit only one wavelength. and hence by the reflectivity of the surfaces. But it is not possible to increase the thickness of the metallic films indefinitely as absorption will reduce the intensity of the transmitted light. one can filter out any particular wavelength. metallic film (usually. on a glassplate. the space between the reflecting surfaces is a dispersive medium. a glass plate) by vacuum deposition techniques. it is possible to achieve a reflectivity of more than 90% for any particular wavelength (see Sec: 15. the reflected light may have a high degree of monochromaticity. it is possible to filter a particular wavelength out of a white light beam. A thin A_ 2nh * The FabryPerot structure also behaves as a resonator and supports the oscillation of what are known as modes.6 for a more detailed account). it will have a narrow spectrum sharply peaked about one wavelength. for example. the resolving power of the instrument depends on the length of the plate. (a) Unlike in the case of FabryPerot interferometer. greater will be the reflectivity. layers of dielectric materials of appropriate refractive indices are deposited. . (see Fig.3). Such a structure is known as an interference filter. But. the number of reflections depends on the length of the plate and the angle 0. metallic films are replaced by all dielectric structures. a A/4 thick film of a dielectric material whose refractive index is more than that of glass.5 and h = 6 x 1.situation in which only one or two maxima are obtained in the visible region.8) or zinc sulphide (n = 2. 16. Thus a FabryPerot structure is formed between the two glass plates. Then a thin layer of dielectric material with lower refractive index (such as cryolite or magnesium fluoride) is deposited. However.
Vol.. the interferometers involving multiple beam interference have a high resolving power and hence find applications in high resolution spectroscopy. Baumeister and G. x curve.1 Calculate the resolving power of a FabryPerot interferometer made of reflecting surfaces of reflectivity 0. This is the principle used in the FabryPerot interferometer which is characterized by a high resolving power. 223. Pincus. consider now two wavelengths A0 (= 6. M.9 A incident on a FabryPerot etalon with the same parameters as given in the previous problem. then the fringes so formed are much sharper than those by two beam interference and. For R = 1. These beams (on either side) interfere to produce an interference pattern at infinity.2. (c) What would be the value of L.. 16. The transmittivity T = 1 when 8= 2mtt. 16. Francon and S. F = 400 and h0 = 10 cm. Calculate the radii of the first three bright rings corresponding to each wavelength. Berlin. (b) Also calculate the FWHM Oh for which the transmittivity will be half.5. (a) x = 200 nm (m = 333334). (b) Oh = 9.7 Consider a laser beam incident normally on the FabryPerot interferometer as shown in Fig. m = 1..11 (b) Calculate the radii of the first four bright rings. Mallick. then the beam would undergo multiple reflections at the two surfaces and a large number of beams of successively diminishing amplitude will emerge on both sides of the plate. FabryPerot interferometer with n2 = 1. 1975. P. M. Concentric rings are observed on the focal plane of a lens of focal length 20 cm (a) Calculate the reflectivity of each mirror..2.. 4. 82 is the angle of refraction inside the film (of refractive index n2 ). Scientific American.h if F was 200? [Ans. = 4880 A. M. Pergamon Press.c=3x 10 8 m1s. With h0 = 10 cm. Principles of Optics.5 Consider a monochromatic beam of wavelength 6000 A incident normally on a scanning FabryPerot interferometer with n2 = 1 and F = 400.4 Consider now two wavelengths 6000 A and 5999. What will be the corresponding values of m? Will the lines be resolved? 16. the value of F is very large and the transmission resonances become very sharp. (b) Show that if v = (v0 ±p 1500 MHz. so that T = 1/z occurs at the same value of h for both the wavelengths? 16. R. What will be the corresponding values of m? (c) Calculate the angular width of each ring where the intensity falls by half and the corresponding FWHM (in mm) of each ring.3 Consider a monochromatic beam of wavelength 6. 16.lm. Problems 16.0 + AA incident normally on the. What will be the value of DA. SpringerVerlag. `Optical interference coatings'.000 A) and A. The transmittivity of such a film is given by T= 1+F sine 2 where F = 4R2 is known as coefficient of Finesse and (1R) S= 4xcn2h cos 62 A'o represents the phase difference (between two consecutive waves emanating from the film) due to the additional path traversed by the beam in the film. 16. Born and E. .000 A incident (from an extended source) on a FabryPerot etalon with n2 = 1. What will be the corresponding values of 8? n=1 ♦ If a plane wave falls on a plane parallel film. 3.5 nm]. H x Fig.6 In continuation of Problem 16. 1971. h the film thickness and X10 is the free space wavelength. Atlas of Optical Phenomena.15. Baierlein.Multiple Beam Interferometry Summary 16.. 16. Newton to Einstein: the Trail of Light. Oxford. h = 1 cm and F = 200. = 0. Cagnet. 2.2 Calculate the minimum spacing between the plates of a FabryPerot interferometer which would resolve two lines with DA.85 and separated by a distance 1 mm at A.) we will have the same T vs. Wolf. 59. calculate (a) The first three values of x for which we will have unit transmittivity and the corresponding values of m.1 A at A = 6000 A.v=v0 =5x10 14 s t Plot T as a function of x (100 nm < x < 400 nm) for F = 200 and F = 1000. 16. .3. 1992. (a) Assumeh0 =0. December. 500 nm (m = 333335).15 REFERENCES AND SUGGESTED READINGS 1. Cambridge University Press. The distance between the two mirrors is written as h = h0 + x. therefore. If the reflectivity R at each surface is close to unity.8. 1500 MHz is known as the free spectral range (FSR). 1970. Assume the reflectivity to be 0. p = 1.
Optical Interferometry.16.. F. Tolansky. 1967. 10. . Lin. New York. New York. Academic Press. C.. 1989. M. McGrawHill Book Co. London. C. 1976. London. R. 1966. Oxford University Press. Tolansky. Van Nostrand Reinhold. Modern Applications of Physical Optics. 6. Interscience. S. 1948. Fundamentals of Optics. Longmans Green and Co. 1963. W. Jenkins and H. New York. Francon. `Optical communications: Single mode optical Optics fiber transmission systems'. 8. S. 12. Cambridge University Press. White. 11. London. 9. Lin. M. Steel. Multiple Beam Interferometry of Surfaces and Films. 1976. New York.. 7. London. W. Francon. Ed. Cambridge. Optoelectronic Technology and Lightwave Communications Systems. Interferometry. An Introduction to Interferometry. 1955. H. Ditchburn. E. Light. Academic Press.12 5. A.
. have a definite phase relationship if At << z. a definite phase relationship exists for times of the order of zc.e.1010 sec and for the red cadmium line (A. The finite value of the coherence time ac could be due to many factors.5 x 1014 Hz and i^ .oo<t<oo (where c is the speed of light in free space) is referred to as coherence length. For example. in general. at x = 0 we have [see Fig.1 INTRODUCTION In earlier chapters on interference we had assumed that the displacement associated with a wave remained sinusoidal for all values of time. is represents the average duration of the wavetrains. then the wavetrain undergoes an abrupt phase shift of the type shown in Fig. it must be very monochromatic. Coherence is conveniently measured by the path difference between two rays of the same source.1(b).1(a)]. one has about 50. the displacement is sinusoidal for .109 sec. which is known as the temporal coherence of the beam. Thus. the electric fields at times t and t + At will. for example. 17. 17. and it is evident that in order to yield many interference fringes.1 (a) For a perfectly monochromatic beam. Now. Thus the displacement (which we denote by E) was assumed to be given by E = A cos (kx . 17. 17.000 oscillations in the time z^. = 6328 A). Since we will be considering only light waves.1(b). is known as the coherence time of the source and the field is said to remain coherent for times The length of the wavetrain.for ingenious methods of spectral analysis and for the measurement of the diameter of stars. if a radiating atom undergoes collision with another atom. cc . Michelson used it . i.cot + 0) The above equation predicts that at any value of x. The finite coherence time could also be on account of the random motion of atoms or (1) Obviously this corresponds to an idealised situation because the radiation from an ordinary light source consists of finite size wavetrains.oo < t < 00. E =Acos(cot0). a typical variation of which is shown in Fig.. For example. The time duration r.1(b). the corresponding coherence lengths are 3 cm and 30 cm respectively.1010 sec. z^ . and will (almost) never have any phase relationship if At >> ac. (b) For an actual source. = 6438 A). Dennis Gabor in his Nobel Lecture on Holography. in Fig. the displacement remains sinusoidal for o < t < +oo. for the neon line (A. 17. the electric field remains sinusoidal for times of the order of 2c. December 11. at a given point.Light which is capable of interference is called `coherent'. This is called the coherence length and Albert Michelson were the first to understand that it is a reciprocal measure of the spectroscopic line width. For v . the quantity E represents the electric field associated with the light wave. by which they can differ while still Lord Rayleigh giving observable interference contrast. given by L = c cc (2) E zc >i (b) Fig. 1971 17.
ri le and t . »2c c then the waves arriving at P from S1 and S2 will have no fixed phase relationship and no interference pattern will be observed.= 6438 A). the beam reflected from M1 interferes with the beam reflected by M2 which had originated 2d/c seconds earlier. in general. the interference pattern observed around the point P at time t is due to the superposition of waves emanating from S1 and S2 at times t . Obviously.7. if r2 Optics (which is usually a partially silvered plate) and the waves reflected from the mirrors M1 and M2 interfere (see Fig.10). 17. Thus the central fringe (for which r1 = r2) will.4. This point is discussed in greater detail in Sec..c and t .1010 sec.2. for helium * For more details. Now.iao^ Mi r1 «2c c then the waves arriving at P from S1 and S2 will have a definite phase relationship and an interference pattern of good contrast will be obtained. A light beam falls on a beam splitter G ii / H 44 G M2 t' Fig. 17. G represents the beam splitter. there is no definite phase relationship between the two beams and no interference pattern is observed. The coherence time for a laser beam is usually muchlargein comparison to ordinary light sources. the disappearance occurs when the path difference is about a few centimetres giving .12 respectively.3). 17. . 14. S1 r1 then. 2 7w/o.2 due to the fact that an atom has a finite lifetime in the energy level from which it drops to the lower energy level while radiating. It may be mentioned that there is no definite distance at which the interference pattern disappears. Let M2' represent the image of the mirror M2 (formed by the plate G) as seen by the eye. if 2d c 2^ Fig.17. have a good contrast and as we move towards higher order fringes the contrast of the fringes will gradually become poorer. 15. the contrast of the fringes becomes gradually poorer and eventually the fringe system disappears. If the distance M1M2 is denoted by d. then the beam which gets reflected by mirror M2 travels an additional path equal to 2d. For the neon line (A.2 Young's doublehole experiment. On the other hand. 17. On the other hand for the red cadmium line (A. in general. 17.r2/c respectively. On the other hand. where rt and r2 are the distances/ S1P and S2P respectively.. the interference pattern produced by this experimental arrangement was discussed in considerable detail in Sec. The interference pattern observed around the point P at time t is due to the superposition of waves emanating fromand S2 at times t . See Ref. * In order to understand the concept of coherence time (or of coherence length) we consider Young's double hole experiment as shown in Fig. 17. Indeed.109 sec. We next consider the Michelson interferometer experiment (see Sec. if the path difference (r2 . .r1) is large enough such that r2r. c thus interference fringes of good contrast will be observed at P if (r2 = r l )/c << a. M2' represents the image of M2 as formed by G.3 The Michelson interferometer arrangement. = 6328 A). the coherence length is of the order of 30 cm giving etc . Thus. as the distance increases. If the distance d is such that Zd 'cc c then a definite phase relationship exists between the two beams and welldefined interference fringes are observed.
Coherence
17.3 sources.* We next introduce a mm thick glass plate in front of one of the circular holes; there is almost no change in the interference pattern as can be seen from Fig. 17.5(b). Clearly, the extra path introduced by the plate [= (,u  1) t, see Sec. 14.10] is very small in comparison to the coherence length associated with the laser beam. If we repeat the experiment with a collimated mercury arc beam, we would find that with the introduction of the glass plate the interference pattern disappears (Fig. 17.6). This implies that the extra path length introduced by the glass plate is so large that there is no definite phase relationship between the waves arriving on the screen from the two circular apertures.
neon laser, coherence times as large as 50 milliseconds have been obtained Ref. 9]; this would imply a coherence length of 15,000 km! Commercially available heliumneon lasers have z1.  50 nsec implying coherence lengths of about 15 m. Thus using such a laser beam, high contrast interference fringes can be obtained even for a path difference of a few metres. In order to demonstrate the large coherence length of the laser beam we consider an experimental arrangement shown in Fig. 17.4. A parallel beam of light is incident normally on a pair of circular holes. The Fraunhofer diffraction pattern is observed on the focal plane of a convex lens. We first use a helium neon laser beam, the resulting interference pattern is shown in Fig. 17.5(a) which is simply the product of the Airy pattern and the interference pattern produced by two point
(a) Fig. 17.6
(b)
(a) The interference pattern prodcued for the arrangement shown in Fig. 17.4 using a collimated mercury arc. (b) The interference pattern is washed out when mm thick glass plate is introduced in front of one of the holes.
2
Fig. 17.4 A parallel beam of light is incident normally on a pair of circular holes and the Fraunhofer diffraction pattern is observed on the focal plane of a convex lens.
17.2 ' THE LINEWI]DTH
In the Michelson interferometer experiment discussed in the previous section, the decrease in the contrast of the fringes can also be interpreted as being due to the fact that the source is not emitting at a single frequency but over a narrow band of frequencies. When the path difference between the two interfering beams is zero or very small, the different wavelength components produce fringes superimposed on one another and the fringe contrast is good. On the other hand, when the path difference is increased, different wavelength components produce fringe patterns which are slightly displaced with respect to one another, and the fringe contrast becomes poorer. One can equally well say that the poor fringe visibility for a large optical path difference is due to the nonmonochromaticity of the light source. The equivalence of the above two approaches can be easily understood if we consider the Michelson interferometer experiment using two closely spaced wavelengths Al and A2. Indeed in Sec. 15.10 we had shown that for two closely
(a) Fig. 17.5 _ __
(b)
(a) The interference pattern produced for the arrangement shown in Fig. 17.4 using a HeliumNeon laser beam; (b) The interference pattern produced by the same arrangement with al mm thick glass plate in front of one of the holes. [This and the following figure have been adapted from Ref. 16; the author came across the photographs in Ref. 8].
* see Sec. 19.8.
17.4 spaced wavelengths Al and A2 (like the D1 and D2 lines of sodium), the interference pattern will disappear if 2d_2d A2 Al
1
Optics
Lv^
AAL
'rc
(9) = Llc, we ob
2
(3)
where we have disregarded the sign. Since tain Ov  1 zc
where 2d represents the path difference between the two beams. Thus 2d _ AlA2 a,2 _ 2(Al 2,2 ) 2(A 1 A 2 ) (4)
(10)
Instead of two discrete wavelengths, if we assume that the beam consists of all wavelengths lying between A. and A + DA,, then the interference pattern produced by the wavelengths A, and A + z AA will disappear if rA2 2d
2I Z AA) °A
Thus the frequency spread of a spectral line is of the order of the inverse of the coherence time. For example, for the yel low line of sodium (A. = 5890 A),
, rc
1010 s
Ov 10 10 Hz
 A2 (5) we get
v _ c = 3x10 l0 =5 x 10 14 Hz A 5.89 x 10
Further, for each wavelength lying between A, and A. + 2 DA, there will be a corresponding wavelength (lying between A + i AA and A + DA) such that the minima of one falls on the maxima of the other, making the fringes disappear. Thus, for 2d ? DA (6)
Ov v
10 10 5 14 =2x10 5x10
the contrast of the interference fringes will be extremely poor. We may rewrite the above equation in the form DA?2 (7)
The quantity Aviv represents the monochromaticity (or the spectral purity) of the source and one can see that even for an ordinary light source it is very small. For a commercially available laser beam, 're  50 nsec implying Aviv 4 x 10s . The fact that the finite coherence time is directly related to the spectral width of the source can also be seen using Fourier analysis; this is discussed in Sec. 17.6.
17.3 THE SPATIAL COHERENCE
Till now we have considered the coherence of two fields arriving at a particular point in space from a point source through two different optical paths. In this section we will discuss the coherence properties of the field associated with the finite dimension of the source. We consider the Young's doublehole experiment with the point source S being equidistant from S1 and S2 [see Fig. 17.7(a)]. We assume S to be nearly monochromatic so that it produces interference fringes of good contrast on the screen PP'. The point 0 on the screen is such that S10 = S20. Clearly, the point source S will produce an intensity maximum around the point O. We next consider another similar source S' at a distance 1 from S. We assume that the waves from S and S' have no definite phase relationship. Thus the interference pattern observed on the screen PP' will be a superposition of the intensity distributions of the interference patterns formed due to S and S' (see Sec. 17.5). If the separation l is slowly increased from zero, the contrast of the fringes on the screen PP' becomes poorer because of the fact that the
implying that if the contrast of the interference fringes becomes very poor when the path difference is d, then the spectral width of the source would be  A 2/2d. Now, in Sec. 17.1 we had observed that if the path difference exceeds the coherence length L, the fringes are not observed. From the above discussion it therefore follows that the spectral width of the source AA, will be given by Az Az (8) DA  L cz Thus the temporal coherer e a^ of the beam is directly related to the spectral width AA. For example, for the red cadmium line, A. = 6438 A, L = 30 cm ere = 10 9 sec) giving QA (6.438 x 10 5 ) 2
A2 
cTc 0.01 A
3x10 10 x10 9
For the sodium line, A. = 5890 A, L 3 cm (2 = 1010 sec) and AA  0.1 A. Further, since v = c/A, the frequency spread Av of a line would be
Coherence
P where
17.5
a=a 1 +a 2
and we have assumed a >> d, 1. Thus
S'S2  S'S1 = ld a
0
a1
Thus for the fringes to disappear, we must have
, .
a2
2
or
=
S'S2  S'S l
=
la
P' (a)
2d
Now, if we have an extended incoherent source whose linear dimension is . 2tald then for every point on the source, there is a point at a distance of Aa12d which produces fringes which are shifted by half a fringe width. Therefore the interference pattern will not be observed. Thus for an extended incoherent source, interference fringes of good contrast will be observed only when
1
<<
0
da
(12)
T
a
Now, if 0 is the angle subtended by the source at the slits [see Fig. 17.7(b)] then 9 = 1/a and the above condition for obtaining fringes of good contrast takes the form
d <<
On the other hand, if
(b)
(13)
Fig. 17.7 (a) Young's doublehole interference experiment with two independent point sources S and S'. (b) The same experiment with an extended source. interference pattern produced by S' is slightly shifted from that produced by S. Clearly, if
S'S2 
d 8
(14)
the fringes will be of very poor contrast. Indeed, a more rigorous diffraction theory tells us that the interference pattern disappears when*
d = 1.2282.258,3.248,...
(15)
S'S 1
=
2
(11)
the minima of the interference pattern produced by S will fall on the maxima of the interference pattern produced by S' and no fringe pattern will be observed. It can be easily seen that
l2 1/2
Thus as the separation of the pinholes is increased from zero, the interference fringes disappear when d = 1.22./9; if d is further increased the fringes reappear with relatively poor contrast and they are washed out again when d = 2.25.x/9, and so on. The distance
2
=
= A,/9 (Lateral coherence width)
SS2=
a2
+(+ l d
l
a+ 1 ( +lJ d
2a
(16)
_
and
ll2 1/2
gives the distance over which the beam may be assumed to be spatially coherent and is referred to as the lateral coher
ence width.
S'S 1
=
[a2+(_i
=a+
1 rI d i)
l2
Example 17.1 On the surface of the earth, the sun subtends an angle of about 32'. Assume sunlight to be falling normally on a doublehole arrangement of the type shown in Fig. 17.7 and that
* See, for example, Sec. 5.5 of Ref. 7.
17.6 there is a filter in front of S 1 S2 so that light corresponding to A 5000 A is incident on S1S2. What should be the separation between S1 and S2 so that fringes of good contrast are observed on the screen?
Optics
Solution:
0 32' = 180
x 60 rad = 0.01 rad
25
used. In order to overcome this difficulty, Michelson used two movable mirrors M1 and M2 as shown in Fig. 17.9, and thus he effectively got a large value of d. The apparatus is known as Michelson's stellar interferometer. In a typical experiment the first disappearance occurred when the distance M1M2 was about 24 feet, which gave 0 = 1.22 x 5 x 105 radians = 0.02" 24 x 12 x 2.54 for the angular diameter of the star. This star is known as Arctures. From the known distance of the star, one can estimate that the diameter of the star is about 27 times that of the sun.
Thus the lateral coherence length 5 x 10w 10
= 0.005 cm
Therefore if the pin holes are separated by a distance which is small compared to 0.005 cm, interference fringes of good contrast should be observed.
17.4 MICHELSON STELLAR INTERFEROMETER
Using the concept of spatial coherence, Michelson developed an ingenious method for determining the angular diameter of stars. The method is based on the result that for a distant circular source, the interference fringes will disappear if the distance between the pinholes S1 and S2 (see Fig. 17.8) is given by [see Eq. (15)]:
M2
d = 1.22
8
(17) Fig. 17.9 Michelson's stellar interferometer. We should point out that a laser beam is spatially coherent across the entire beam. Thus, if a laser beam is allowed to fall directly on a doubleslit arrangement (see Fig. 17.10), then as long as the beam falls on both the slits, a clear interference pattern is observed on the screen. This shows that the laser beam is spatially coherent across the entire wavefront. Figure 17.11 shows the interference pattern obtained by Nelson and Collins (Ref. 14) by placing a pair, of slits of width 7.5 pm separated by a distance 54.1 pm on the end of the ruby rod in a ruby laser. The interference pattern agrees
where 0 is the angle subtended by the circular source as shown in Fig. 17.8. For a star whose angular diameter is 107 radians, the distance d for which the fringes will disappearwould be 1.22x5x10 5 = 600 cm d 107 where we have assumed A, = 5000 A. Obviously, for such a large value of d, the fringe width will become extremely small. Further, one has to use a big lens, which is not only difficult to make, but only a small portion of which will be
L
T Sj d
S2 V.
Laser beam
Screen
Fig. 17.8 S
is a source of certain spatial extent; S1 and S2 are two slits separated by a distance d which can be
varied: The fringes are observed on the focal plane of the lens L.
Fig. 17.10 If a laser beam falls on a doubleslit arrangement, interference fringes are observed on the screen. This shows that the laser beam is spatially coherent across the entire wavefront.
Coherence
17.7
111•11•
sIlll•s
s =
Fig. 17.11
The doubleslit interference pattern obtained by placing a pair of slits each 7.5 pm wide and separated by a distance of 54.1 pm across the diameter of a ruby rod. (a) shows the actual interference pattern and (b) shows a densitometer trace of the interference pattern. The dots correspond to a theoretical calculation assuming that a plane wave strikes the pair of slits. After Ref. 14. Photograph courtesy: Dr D.F. Nelson.
11
with the theoretical calculation to within 20%. To show that the spatial coherence is indeed due to laser action, they showed that below threshold (of the laser) no regular interference pattern was observed; only a uniform darkening of the photographic plate was obtained. 17.5 OPTICAL BEATS When two tuning forks, one having a frequency of 256 Hz and the other a frequency of 260 Hz, are made to vibrate at the same time, we hear a frequency of about 258 Hz whose intensity varies from zero to maximum and back with a frequency of 4 Hz. This phenomenon is known as beats. It can be easily understood by considering the superposition of two waves having frequencies w and co + Ow:
yl = a sin ( cot + 01) y2 = a sin [(co + ow)t +
02]
11 (a)
1
1
(c) Fig. 17.12
(a) and (b) show typical time variation of r sin + Ow) t] and sin 2 Ocot] re(o. L
(18)
spectively, and (c) shows the time dependence of their product. respectively. In Fig. 17.12(c), we have plotted their product which represents the resultant displacement. Notice that although the envelope has a frequency of 4n (= 2 Ov) [see Fig. 17.12(b)] the intensity repeats itself after every 1/Ov seconds. This waxing and waning of sound is known as beats. The beat phenomenon can be easily understood by observing the Moire fringes obtained by the overlapping of two patterns of slightly different spatial frequency (see Fig. 17.13). Whenever the dark line of one of the patterns falls on the bright region of the other, then the two waves can be considered to be `out of phase' and we have a broad `dark region' which appears periodically. In a similar manner, one can consider the phenomenon of optical beats. For example, let us consider the superposition of two fields E1 and E2 having frequencies w and co + Acv:
El = Eo1
where we are assuming (for the sake of simplicity) that both the waves have the same amplitude. The resultant displacement would be given by Y=Yi + Y2 l 1 (01 + 02 )] = 2a sin [(co Act)) t
+2
+2
[2 (Ow)t + 2 = 2a sin [(a) + 2 Ow) t] sin 12 /wt
xcos
(02  $t)]
(19)
where we have assumed, without any loss of generality, 01 = lc/2 = 02. Figures 17.12(a) and (b) show the time variation of the terms sin (co
+ 2 Ow ) t
and sin (2 Ado) t
sin (wt + 01)
(20)
17.8
Optics
4.
Fig. 17.13 The Moire pattern produced by the overlapping of two patterns, of parallel lines (of slightly different spatial periods) show the beating phenomenon [after Ref. 1 ]. and E2 = E02 sin [( co + Ow) t + ¢2] (21) If we assume that both the fields are linearly polarized in the same direction then in order to calculate the resultant field, we may simply algebraically add El and E2. Thus the resultant would be given by E =E1_+E2 = E01 sin (wt + 01) + E02 sin Now E2 (t) = E01 sin2 (cot + 01) + E02 sin2 ([w + Lw]t + 02) + Eo1E02[cos (2cot + Owt + 01 + 02) (22) + cos (Awt + 02  01)] For optical frequencies, co = 10 15 Hz and therefore the first three terms would vary with extreme rapidity and a detector (like the eye or the photodetector) would observe a time average of the quantity. Now, the time average of the quantity F(t) over a duration of 2T is defined through the following equation:
(F(t))
which is an extremely small quantity in comparison to unity. It is for this reason that the eye does not see any intensity variations. Even for a fast photodetector with response times  109 sec, 1/(coT) 106 which can also be neglected. Returning to Eq. (22), if we carry out an averaging over times which are long compared to 27r/co but short compared to 2 g/Aco then we would obtain (E2(t))
= 2 Eo1 + 2
E
O2 e (26)
+ Eol E02 cos [(Aco)t + 02  01]
([co + Aw]t + 02)
For example, if Ow = 10' Hz and the photodetector resolution is about 109 sec, then the detector will record only the average values of the first three terms on the RHS of Eq. (22); however, it will be able to record the time variation of the last term. This is what is shown in the above equation leading to the familiar phenomenon of beats. As an example, we consider the beating of the D1 and D2 lines of sodium for which = 5890 A (= co = 3.2003 x 10 15 Hz) A2 = 5896A(= w2 = 3.1970 x 10 15 Hz) Thus Ow = 3.3 x 10 12 Hz In order to observe the beating, the detector should have a response time much smaller than 1/Liw, thus the photodetector response time should be 5 1013 sec which is a practical impossibility. Therefore, in order to observe the beats, we must decrease the value of Ow. Indeed the first experiment on optical beats was carried out by Forrester et al. (Ref. 6) in which they used two closely spaced frequencies by splitting a spectral line using a magnetic field (this splitting is known as Zeeman effect). The weaker the magnetic field, the smaller is the value of Ow. In the experiment of Forrester and his coworkers, Al/ was of the order of 10 10 Hz and they were able to observe optical beats. Obviously, in order that the beats occur very slowly (so that we may use photodetectors of much longer response times) zw should be made even smallerbut then we may have the coherence problem. In the above analysis we have assumed the phase 01 and 02 to remain constant in time. Now for an incoherent source, 01 and 02 will randomly change in times 1o9 sec; thus if the detector response time . is > 10$ sec, we will observe the average of the cos [(Aco)t + 02  01] term in Eq. (26). Obviously, the average value of the `cosine term' is zero and we will have (E2 (t)) fEo l +Eo 22
= 2T f F(t)dt
T +T
+T
(23)
Thus (El sine (wt + 0 1 ))
= Eoi 2T Eo1 [2
T
f sin
2coT
e
(wt+0 1 )dt
{sin2(wt+)}±]
=2
Eon [1
2(T sin 2wT cos 201 ] (24)
For averaging times T >> 1/w, the second term inside the brackets would be extremely small and hence can be neglected. Thus we may write ^Eol sin 2 (wt + 01))  Eoi (25)
For example, the eye would respond to changes in times of the order of 0.05 seconds. Thus T  0.05 sec and since co 1015 Hz, we have = 2x 10 14 coT
implying that the resultant intensity will be just the sum of the independent intensities: I =71 +12 (27)
Coherence
17.9
With the advent of laser beams, the beating experiments have become much easier; a typical arrangement (which resembles a Michelson interferometer) is shown in Fig. 17.14. A typical beat note of the experiment of Lipsett and Mandel (Ref. 11) is shown in Fig. 17.15. It was observed that the beat note changed in frequency from about 33 to approximately 21 MHz in a time of about 0.7 psec. The coherence time is . 0.5 ,usec which is consistent with the duration of the spike. We conclude this section by quoting Feynman: "With the availability of laser sources, someone will be able to demonstrate two sources shining on a wall, in which the beats are so slow that one can see the wall get bright and dark".
In Photocell 2 Photocell 3 Laser 2 Photocell 1
17.6 COHERENCE TIME AND LINEWIDTH VIA FOURIER ANALYSIS
That the frequency spread of a line is of the order of the inverse of the coherence time [see Eq. (10)] can also be shown by Fourier analysis. As an example, we consider a sinusoidal displacement of duration cc . Thus we may write
V(x=0,t) =de 'mot
Itl
< 2ze > Zip
(28)
=0
I
t
I
Laser 1 Fig. 17.14
We will assume that cc is long enough so that the disturbance consists of many oscillations. For example, for a 2nsec pulse corresponding to vo = 5 x 10 14 Hz, the number _9 of oscillations will be 5 x 14 x 2 x 10 = 10 6 , i.e. the pulse will consist of about a million oscillations! Now, while discussing the Fourier transform theory (see Sec. 8.4 and Sec. 9.5), we had shown that for a timedependent function f(t), if we define
10
The experimental arrangement of Lipsett and Mandel (Ref. 11) to observe optical beats using two laser beams. then
F(co) = 2^
f f (t) e
imt dt
(29)
f(t)
=
f
F(co)e ta' t dco
(30)
Replacing f(t)
^] 0.5 microseconds
by 1J(x
= 0, t), we may write
H
W(x = 0, t) =
Fig. 17.15
Oscilloscope trace of the sum of the intensities of the laser beams (upper curve) and the intensity of the superposed laser beam (lower curve) [Ref. 11].
f
2^
A(w)e
tmt
dco
(31)
A(co)
The RHS represents a superposition of plane waves with representing the amplitude* of the plane wave corres
* Notice that the integral appearing on the RHS of Eq. (30) is over negative values of co also. However, the displacement (or the electric field) is the real part of yf which is given by (omitting the 2tc factor):
E = Re [t,u(x = 0, t)] = Re
f
I
A(co) l ei(mt+^) dco
= f I A(co) I cos (cot + 0) d co
f
0
I
A(w) I cos (cqt
+_0) d ca +
f I A (w) I cos_(wt _¢_(w)) dw
0
where we have used the relation A(co) = IA(co)I e°. The above equation can always be written in the form
f
0
C(w)cos.[cot +(co)]dw
Thus the amplitudes associated with the negative frequencies contribute essentially to the corresponding positive frequencies.
17.10 ponding to the frequency co. Equation (31) tells us that lp"(x = 0, t) is the Fourier transform of A(co) and therefore using the inverse Fourier transform [see Eq. (29)] we get A(w) +, J iy(x=0,t)etwt dt
Optics
= 2= 1
17.7 COMPLEX DEGREE OF COHERENCE AND FRINGE VISIBILITY IN YOUNG'S DOUBLEHOLE EXPERIMENT
In this section we will introduce the complex degree of coherence and will show how it can be related to the contrast of the fringes in the Young ' s double hole interference experi ment. We refer back to Fig. 17.2. Let 'Y 1 (P, t) and '1 2 (P, t) represent the complex fields at the point P due to the waves emanating from S1 and S2 respectively. The resultant displacement would be given by `I' = `I' 1(P, t) + T2(P, t) (36)
+I T.
ae i(000)t dt
2
1/2
sin {(w_w o) r a  co o)
_ (2) 1'2 a sin{a(521)) [sin  tc (c 1) ] coo where S2 co
0
(32)
Now, the intensity at the point P will be proportional to I'I h I 2 which is given by the following equation: I'I'1 2 ='I`:;'I`t +' {'P+ `I`1`I`2 + T I T; = 1'1' 1 1 2 + IP 2 1 2 + 2 Re ('F KI 'P2 ) Since 'P1 and T2 vary with extreme rapidity, we can observe only the average values of IT 1 1 2 and I'P 2 1 2 . Thus, if we write I1 = (I `I' 1 (P, t)1 2 ) and
I2 =
and a =
1 2
cvo;. In Fig. 17.16 we have
plotted the function
sin [a (0  1)] (521)
(33)
as a function of Cl for a= 200. One can see that the function is sharply peaked at SI = 1 (where it has a value equal to a) and that the first zero on either side occurs at 52 = 1 ± (7r/a). For larger values of the function will become more sharply peaked; the width of the peak being given by
(I 'P 2 (P, t) 12)
+ I2
a
then
I =
I1
+ 2 1/11 12 Re
Y12
(37)
 (38)
OS2 ( coo J
a
(34) where Y1z=
(P, t)T2 (P, t)) [(I(P, t) I2) (1 T2 (P,
t)12 )] "2
or
OcO 7cco o
27r
Td
a
Thus 1 Av Tc
is known as the complex degree of coherence and (...) denotes the time average of the quantity inside the triangular brackets [see Eq. (23)]. The field '{' 1 (P, t) is due to the waves emanating from the point S1 at t where 11 = S 1 P. Thus, (35) T 1 (P, t) will be proportional to 'P (S l ,,t 'l) where T(S l , t) e denotes the field at S1 at time t. Similarly 'F 2 (P, t) will be proportional to
consistent with Eq. (10). The above analysis shows that a wave having a coherence time 2' is essentially a superposi, tion of harmonic waves having frequencies in the region vo Av < v < vo + Av where Av  1/;. We should mention that the condition expressed by Eq. (35) is quite general,in the sense that it is valid for a pulse of arbitrary shape. For example, for a Gaussian pulse having a duration 2'c, the corresponding frequency spread will again be given by Eq. (35) [see Example 10.4].
21
(S2 , t 
'l). Thus c
('P* S1,
t Y12=
J 'P(5
2
t  J)
2
2
`I'
SI,t 
it
c
'h1S2,t
cJ
1
1/2
Since the overall intensity distribution in the fringe pattern does not change with time, we may write
Coherence
I
17.11
f (S2) = sin a(S2  1) (521) 200
100
(b)
Fig. 17.16
(a) A sinusoidal displacement of duration rc; (b) The variation of the function [sin (S2 1)a]/ (S2 1) as a function of S2 for a= 200. Notice that the function is sharply peaked around S2 = 1.
Yu =
(`F* (Sl , t + r) 'Y (S , t)) 2 )(I ,I, (S2,t)1 2 )] 1/2 [(I`1` (Sl,t)1
2
(39)
where A(t) and 0(t) are slowly varying real functions of time. For a perfectly monochromatic beam (i.e., infinite coherence time) A(t) and 0(t) are constants so that `F*(t + ))'I'(t) = A Consequently Y12(r) = et' Thus, for such a case
1 = I1 + 12
where r = (r2  r1 )/c. In order to discuss the effect of temporal coherence, we assume S, S1 and S2 to be of negligible spatial dimensions. Further, if S1 and S2 are equidistant from S, then we may assume that
2 `u'
e
(43) + 2 I1 I cos cot
`I' (S i , t) = `1'(S2 , t) = `1'(t)
Thus, for such a case, (f *(t + r) q' (t)) Yu( r) _ (I
(40)
2
(44)
and the visibility V, which is defined by
`}'(t)12)
(41)
Now, for .an actual field we may write `1`(t) = A(t)e i[wt+op(t)) (42)
V __ I max +I min I max + I min would be given by V 2 11 +
(45)
g7212
(46)
17.12 For Il = 12 we have V = 1 implying that, for a perfectly monochromatic beam, the contrast of the fringes is perfect. On the other hand, for an ordinary light source having 1010 sec, the functions A(t) and fi(t) can be assumed to be constants in times < 1010 sec. Thus, if > 1010 sec, `h(t + v) will have no phase relationship with T(t) and the time average (T*(t + v)`I'(t)) will be zero. Thus, if the path difference S2P  S1P is such that S2 PS1 P >zc (47) c the fringe pattern will not be observed. In general, we may write (48) cm`+m 712 = 1 712 I e` where 1112 I and /3 may be assumed to be constants around the observation point. This gives us I =11 +12 + 2 1112 I712Icosa where a = cor + P. Thus 'max =11 + and Imin =11 +12  2 1 1 12 1 712 Hence the visibility becomes 2 11 12 (52) V= 'max Imin 11 + 12 I 712 1 'max + Imin Thus the visibility (or the contrast) of the fringes is a direct measure of 11121. If 11 = 12 then V = 111z1. In the present case, since S, S1 and S2 have been assumed to be points 11121 depends only on the temporal coherence of the beam. For r <<;, 171 z1 is very close to unity and the contrast of the fringes will be very good; for v» vc, 11121 will be close to zero and the contrast will be extremely poor. It may be noted from Eq. (43) that for a perfectly monochromatic beam 17121 = 1 and a = cur = co(S2P  SIP)/c. In general, it can be shown that 0 < 11121 < 1; 11121 = 0 implies complete incoherence and 1112 1 = 1 implies complete coherence. In practice, if 11121 > 0.88, the light is said to be `almost coherent'. Further, since + r)A(t)e`"+ti)^(r)I) ('I'*(t + v)`P(t)) = e`wi(A(t and for a nearly monochromatic source A(t) and 0(t) are already slowly varying functions of time, the quantity, inside the angular brackets (on the RHS of the above equation) will not vary rapidly with v. Thus, we may write erpei0 (53) 712 = I Yi21 * This section was kindly written by Professor K. Thyagarajan.
I
Optics
where both 17121 and /3 are slowly varying functions of Sl =P S2P (54) c For a more detailed theory of spatial and temporal coherence, you may look in Refs 2, 3, 7 and 20.
17.8 FOURIER TRANSFORM SPECTROSCOPY*
(49)
12
+ 2 7112 1 712
I
(50)
(51)
In the previous section we have shown that the contrast in an interference pattern depends on the relative magnitudes of the optical path difference A, vis a vis the coherence length of the source Lc (= cvc). For a given source, the contrast varies as the optical path difference A is varied, beginning from an extremely good contrast for A << La to a very poor contrast for A >> Lc. Indeed Fizeau in 1862 interpreted the periodic variation in contrast in Newton 's rings under illumination with a sodium lamp as the lens is moved up, as being due to the presence of two lines separated by 6 A (see Example 15.4). Michelson in the years 18901900 performed various experiments with a number of spectral lines. Using the Michelson interferometer he measured visibility as a function of optical path difference and using a mechanical device he himself had built, he could obtain the spectra. It is the purpose of this section to show that from a knowledge of variation of intensity with optical path difference one can obtain the source spectral distribution by a Fourier transformation. The use of the Michelson interferometer for spectroscopy was revived in the 1950s for application, specially for the relatively complex spectra in the infrared region. We will derive expressions for the variation of visibility with optical path difference for a source having a certain spectral distribution and we will show that from the interference pattern one can obtain the spectral intensity distribution of the given source.
17.8.1 Principle of Fourier Transform Spectroscopy
Figure 17.17 shows the arrangement used in a Fourier trans form spectrometer. Light from the given source is collimated and enters the Michelson interferometer and in the transmitted arm we measure the intensity at the focus of the lens as a function of the path difference A. Now, if a monochromatic beam of intensity 4 ° is split into two beams (each of intensity z 70) and are made to interfere, then the resultant intensity is given by (55) I =1°(1 + cos 8)
we have used Eq.vo) represents the Dirac delta function. It (o)lio Thus if I(v) dv represents the intensity emitted by the source between v and v + dv then the intensity at 0 lying between v and v + dv is given by cos 21cv (57) c Hence. we have I(v) dv = l0 6(v . (30) of Chapter 14 with It =I2= 2 to c (63) 1 Hence II(O) and yvary sinusoidally for all values of path difference 0 [see Figs 17.vo) dv (61) where 3(v . Hence f Y(O) = I0 0 S(v vo) cos 27" c dv Fig.. 17. (b) The corresponding cosinusoidal variation of y (A) with A. .(O) =10 (1 + cos 'v0 where (62) represents the phase difference between the interfering beams.17 The arrangement used in a Fourier transform spectrometer.18(a) and (b)] implying that the coherence length of the source is infinite.0 i. 8 ^ t 10 I (v) dv + f 0 I(v) cos 27cvO dv (58) (a) The quantity IT = f 0 I (v) dv = Ir (0) (59) 0 2 6 8 10 represents the total intensity of the source.Coherence 17. = cos (56) and 1. 17.13 amples giving explicit expressions for It (O) and some specific cases. y(o) for (i) Monochromatic Source For a monochromatic source of intensity Io emitting at a frequency vo. the total intensity at 0 corresponding to a path difference 0 is II (v) dv = I (v) dv [1+ I1(z ) 2 f = f = 0 0 It (v) dv 0 i 2 4 i A/ u 6 '. We first consider some ex (a) The variation of the total intensity at 0 as a function of the path difference A for a monochromatic source. (55).18 0 It is the quantity II (i) which is measured as a function of 0 from which y(O) is evaluated.IT I T Aao 27rvO C (b) = 1 f I(v) cos IT dv (60) Fig. and in writing Eq. We define normalized transmission as 7(0) = It (0) .
2 Inversion to Recover I(v) from y(0) In an actual experiment. 0 cos 2^rvA c cos 2^rv'O c dA 2A 12 A ^i + = f (67) 0 2zv 'O 2zvO dA cos c c =1 where A0 ( _ 1= X12) is the average wavelength. (2). Expressing 8v in terms of 5A. To do this we just multiply Eq. Thus = cos [2z (v1 A] c + V2 ) x cos [tat (v1 2c and 1 v2) (65) f 7(A) cos 27"'0 d A c 0 1.v 2 f dA = 5 v I 1 V' c 1 = OS(v .0 c (68) and ±2xi(vv')A/c 2 )) v1. then as we raise the convex lens above the glass plate we should have a periodic appearance of fringes as we had discussed in Example 15.19. cos IT Such a variation of 1.v2) c 28v (69) ( e±2ni(v+v ')0/c dA = 0 (72) .. we have Lc O' = 2Sa.v') (71) Hence the minimum path difference at which the visibility vanishes is given by c. The sinusoidal variation has a period P= 2c (v1 + v2) f 0 dv I (v) J. As a simple consequence of this. (60) has to be inverted to obtain the source spectral distribution I(v) from the measured y(A). A'n 201 1 .(0) = Io{1+cos[21c (v1 x cos [27c v2) A] IT f 0 d0 c f dv I(v) cos 2gv° cos 2nc 0 O]} (66) (V1 + = Now. We now consider a source emitting two monochromatic lines at frequencies vl and v2.8.vl) cos 'v° dv 0 f 0 5(v . we may consider the Newton's rings experiment with a sodium lamp. Thus Optics which corresponds to the coherence length of the source. we measure IM (O) and I. The difference in path difference between two consecutive positions of the disappearance of the fringes is c/Sv = 521. (65) we note that y(A) corresponds to an amplitude modulated sinusoidal variation.17. The modulation amplitude has zeroes at A values given by f cos 27tcvA cos 2^tc 'Q dA 21c (v1 . (70) I(v) dv = and 2 lo[S(v . 17..14 (ii) Source Emitting Two Monochromatic Lines.3 =1 [cos 2nro 1 A + cos 2z c2 c + v2) 17. (60) by cos 2zc ° and integrate over A. Thus Eq. If we assume that the sodium lamp emits two discrete wavelengths Al and X12. From Eq.v2) cos 2itvO c dv consistent with Eq.v1) + 8(v . Writing the two cosine terms in terms of exponentials and using +.v2 )] (64) y(A) = 2 [i• 5(v .(O) and y(A) with A is shown in Fig.v2) A = (m + 2c or 0=(m+ z 2 1 since the integrand is an even function of A. each characterized by an intensity i lo.
n 1(v) = 2IT C y(A) cos 'c 0. we consider a perfectly monochromatic beam of frequency vo incident on the interferometer. this maximum path difference determines the resolution obtainable in the estimated 1(v). (b) The corresponding variation of y(A) with A. (74). the inversion process gives us a finite spectral width due to a finite value of A. 20m (77) Thus although the incident beam is monochromatic. Such an inversion from y(A) to I(v) is usually performed using a computer. The spectrum is peaked at v0 and the first zero appears at Hence 27c(vvo) 1(v) 4IT c f 0 C 0=± . dA = 0 = 41 T 41 T $ 5(v .3 Resolution 17..v0) J Fig. We have seen that for such a case y(A) varies with A as given by Eq. otherwise (75) (a) it r Hence using Eq. is the maximum path difference measured... (74). y(4) cos 27cv' dA c (74) or v=v0 ± C Thus one can obtain the source spectral distribution I(v) from the measured y(A) just by a cosine transformation.. we have 1(v) 4 6 =^ 0 f o cos (27" O I cos r27tcv01 dA ) l 10 = 21 T c f 27r (v .Coherence 2 r 17. 17.nl sinr2'c(v cvo)Om l . (since v and v' are positive).16. Now in the experiment if A. there is a maximum limit to path differences that can be introduced.15 a 00 2 4 &/2o 6 8 10 From Eq. it follows that to obtain I(v) one must measure y(A) for all values of path difference A lying between 0 and co.19 (a) The variation of the total intensity at 0 as a function of the path difference A for a source emitting two monochromatic lines. the first term in the RHS within brackets is negligible and we obtain sin 2ir (vv 0) A .v') I(v) dv 0 c 27t (v . 17.vo) A dA [cos 27r (v + vo) A + cos c c J sin(2n(v+cvo)O. (62). (b) JA0  21 T c 2c (v + vo) JJ+ `` 2c (v . we obtain Since v and v0 are both positive and much much greater than c/0. . To estimate the resolution. Since in an actual experiment. then y(A) would be y(A) = cos =0 2jrvo A c 0 < A < An.8.v0) C (76) I(v ' ) (73) The above estimated source spectrum is similar to that shown in Fig.
.20 shows the source spectral distribution as well as the variation of y(d) with A.c2''2 JJ o = 1 µm. r 2 2 Example 17.. (2).1 where we have used Eq. the higher will be the resolution. c Sv (86) (85)  l0z f e ( v . Thus.27.3 Consider a quasimonochromatic source characterized by a spectral distribution I(v) = 1 Io.v°)2 T2 ei2nvelc dv Hence.... = . the larger the maximum path difference A. then Sv = At A. y(A) = cos (27cvoA/c) much like that for a monochromatic source. J dv = Io Now.0 Example 17. Notice that in this case for path differences A << c/Sv. one must have A << ca = cloy We may thus define the coherence length as Lc =cc= consistent with Eq. If we now use the integral +.10 Re ^ z f e ( r v .1 5v < v < vo + 8v 2 2 otherwise (87) E eaxe+pxdx _ () 1/2 i exp R2 a. (82) Solution: v° +4& IT = Sv Ip. But as the path difference increases. over which yis measured. = 5 cm. +00 to Re e27tiv° A/c fr e1. Hence Sv = c 2Au. I(v) cos 27cvA J ' c dv v0 z fiv sin ( 7r8vA c J 0 _ (88) 0 l^ J r 1(v) cos 2irv° dv = Io cos 2^cv00 c ( 8v) c 7rO/c . then we may use the Rayleigh criterion and define the minimum resolvable frequency separation to be the frequency width from the peak to the first zero in I(v). the contrast will be very poor. (78) +r00 Optics = . this corresponds to 8a. the modulation amplitude of y(A) is reduced. Thus 1(v) dv IT =J 0 Figure 17.2 We consider a quasimonochromatic source characterized by a Gaussian spectral distribution given by I(v) = 1 Ioe (vv0 ) 2 /(sv) 2 . For a quasimonochromatic source Sv/vo << 1. Re a>0 (81) Calculate y(A) and show that again for path differences we would obtain IT = 10 0» c/Sv./n (Sv) (79) J 0 I(v) cos 27ccvAdv 2= I0 exp . Sv =0 1 vo . (81) with a = 'c and = i2jtNc. = 0. For good contrast.17. As an example if A.v o) 2T2 d v (80) where in the last step we have used the condition 1/2= Sv << vo. 2 z2 eizlr^lc d > = v .vo) 2i2 dv / J 0 +00 z f e( v . 221I = 1 Re { e 2^civ °lc l ° exp [. y(A) = exp 2A2 ^22 cos 2^cv0^ c (84) = 10 2'e (vv0) 2 T 2 Here Sv = 1/2 characterizes the width of the spectrum since 1(v) drops to 1/e of the value at v = vo at v = vo ± Sv.vo 32 xow = 3 G Hz 5 A.r I0 J e (v .cos C 2icco^ (83) Hence.16 If the incident source contains two frequencies.v°Y z2 cos 27cv0 dv c + ..
cos . beginning from an extremely good contrast for A << Lc to a very poor contrast for A >> Le. for the red cadmium line (A = 6438 A). is .Coherence 17. 17.12 and 18.4 Calculate the distance at which a source of diameter 1 mm should be kept from a screen so that two points separated by . 3 'x 106] c 17. Michelson developed an ingenious method for determining the angular diameter of stars. angular diameter of the sun is about 30'. Problems = 11(v) cos 2gc A T 0  dv 17. not monotonically reduce to zero.represents the average duration of the wavetrains. Plot y(A) as a function of A width of a circular source is 1.228 0 t v 1  °/s1.0.3 In Sec. (79).88 or better).rvoA) c (89) quency stability.17 ♦ The length of the wavetrain. CalFor A << cISv..109 sec. . Calculate the line width and the fresin (2.3A/9.. 17. ♦ Using the concept of spatial coherence. calculate the distance between two pinholes which would produce a clear interference pattern.018 A. given by Lc = C2^ t (where c is the speed of the light in free space) is referred to as coherence length. 17. ♦ Indeed from a knowledge of variation of intensity with optical path difference one can obtain the source spectral distribution by a Fourier transformation. in this case y(A) does for good coherence (i.e. Thus y(A) where 6 is the angle subtended by the circular source. one For more details on Fourier transform may look up Refs 10.e.20 Spectral distribution and the variation of y(A) with A for a source characterized by Eq. It can be shown that and notice that unlike in the earlier example.30 cm. the interference fringes (formed by two pinholes) will disappear if the distance between the two pinholes is given by d = 1.22 Ate. (90) A= c/8v Assume A = 6328 A.rSvA c [Ans: 0. ♦ Using two laser beams it is possible to observe optical beats.4 we had mentioned that the lateral coherence which represents the coherence length. the contrast of the interference fringes varies as the optical path difference A is varied. Assuming that the spectroscopy. for a visibility of 0.2 Laser linewidths as low as 20 Hz have been obtained. For example. ♦ In the two beam interference pattern. the corresponding coherence length is .r8vA ♦ The coherence time . i. The method is based on the result that for a distant circular source.o Fig. ♦ The lateral coherence width (l0 of an extended incoherent source represents the distance over which the beam may be assumed to be spatially coherent. the coherence width should be S 0. the electric field remains sinusoidal for times of the order of a^. Summary [Ans. y(A) = cos (22rvoAlc) and the contrast vanishes for culate the coherence length and the frequency stability.02 mm] . 17.1 The orange Krypton line (A = 6058 A) has a coherence length of 20 cm. it is given by 0 vo ^v 1 A l" where 0 is the angle subtended by the source at the point of observation.
[Reprinted by Macmillan India. 1975. [Reprinted by Macmillan India. 6. Vol. 1978. Oxford. Pergamon Press. What is the coherence time of the radiation emerging from the source? 17. AddisonWesley. Vol. Thyagarajan and A. assume X10 = 6 x 105 cm and z. N. 2 a>0 Show that the temporal coherence is z. Contemporary Optics.] 18. Wolf. A. Siegman. 1975. 20.105 sec. 1955. 17. Mazzolini and G. M. M. Weltin. J. D. American Journal of Physics. Assume A = 6 x 105 cm. New York. plot the Fourier transform A( co) [as a function of w] and interpret it physically.10 In problem 17. Kandpal. Thompson. M. = 6000 A. explain this phenomenon qualitatively on the basis of coherence length. 15. `Spatial Coherence in the Output of an Optical Maser'. 'Moir€ Fringes'. 199. 7. Encyclopaedia of Modern Optics (Ed: R. L. calculate the angular diameter of the star. Vol. Guenther. Jaseja. Ed. 17. Optics. 739. Assuming A = 6 x 105 cm. Oxford University Press. B. 1989. 17. I. J. E. B. Hecht and A. V.1/z. Born and E. 13. Plenum Press. E. 1961. Sakai. PrenticeHall.. Michelson observed for the star Betelgeuse. Mandel. 1965. `Fourier Spectroscopy' in Progress in Optics. 1981.. A. 653. 12.109 sec. Nature.2. Pergamon Press. K. 2. M. Show that the frequency spread Acv . Journal of Applied Physics. `Light Beats'. Principles of Optics. 10. 17. Smith. 165. Calculate the frequency components predominantly present in the pulse and compare it with the case corresponding to z . Vanasse and . A.t)=Eoexp . 11. Johnson.] E. A. New Delhi. D. Sharma. Miller. G. Thyagarajan. Collins. 2. Lasers. Elsevier (2005) . Ghatak and K. 8. Vol 30. Wolf.6 If we perform the Young's doublehole experiment using white light. Amsterdam.. Assuming that the visible spectrum extends from 4000 to 7000 A. 14. 1956. C. 5. Townes. 192. Vol. Forrester. John Wiley. 9. O. North Holland Pub.5 In a Michelson interferometer experiment. 4. Soc. New York. T. 32. A. 1964. H. Beran and G. New Delhi. Physical Review Letters. J. Vol. 1962.5 mm may be said to be coherent. Klein. Booth. 7. Bailey and M.7 Using the stellar interferometer. New York. J. S. Nelson and R. The distance SS1 = 1 m and S 1 S 2 = 0. R. New York. Vol. Zajac. Co. it is found that for a source S. Optics 17. Photo. Welch. A. Optics and its Uses. 1963. E. the fringes disappear. 16. Assume A. Reading. 24. H. Plenum Press. Diffraction: Coherence in Optics. Forrester.5 mm. `Frequency Stability of HeNe Masers and Measurement of Length'. J. A. 17. 553. F. Theory of Partial Coherence. G. 1970. Mass. T. Lasers: Theory and Applications. Vol. A. `Coherence Time Measurement of light from Ruby Optical Masers'. Inst. 1891. Francon. that the fringes disappear when the distance between the movable mirrors is 25 inches. Ghatak. 19. then only a few coloured fringes are visible. H. Calculate the angular diameter of the hole S which will produce a good interference pattern on the screen. T. Parrent.. 1967. Assume z >> (1/wo). 1966. REFERENCES AND SUGGESTED READINGS 1. 10. A. 3. A.t 2z2 Show that the Fourier transform is given by A(cv) =Eozexp [ 2 (w _ wO) 2 z 2 1 r You will have to use the following integral 11/2 5 exp[ax 2 +/3x]dx =( expF 4 a1. Javan and C.9 Assume a Gaussian pulse of the form 2 e i0ot `P(x=0. Bayvel and J. A. Proceedings of the Conference and Workshop on the Teaching of Optics. Optics. `On Coherence Properties of Light Waves' American Journal of Physics. 1970. VI. H. Lothian. D. Opat. `Photoelectric mixing of Incoherent Light'. J.8 Consider Young's doublehole experiment as shown in Fig. F. Oxford. Ghatak and H. The Ronald Press Co. Physics Demonstrations and Experiments. 17. F. P. Gudmundsen and P.18 a distance of 0. Physical Review. K. 1986. K. Engr. University of Melbourne. S.. Lipsett and L. New York.17. Midwinter). Ed. Meiners.9. 1974. as one of the mirrors is moved away from the equal path length position by a distance of about 5 cm. 99 (6). Vol. Van Nostrand Reinhold. 1963. Coherence. Englewood Cliffs.. G. 4.
spatial frequency filtering.PART D iffract o n Chapters 18. laser focusing. Dennis Gabor received the 1971 Nobel Prize in Physics for discovering the principle of holography. Xray diffraction etc. Chapter 21 is on holography giving the underlying principle and many applications. resolving power of telescopes. 4 . 19 and 20 cover the very important area of 4 iffraction and discuss the principle behind topics like diffraction divergence of laser beams.
.
The best we can do is. In the Fresnel class of diffraction the source of light and the screen are. if the observations are made carefully then one finds that if the width of the slit is not very large compared to the wavelength. roughly speaking. 18. through a narrow opening is usually referred to as diffraction and the intensity distribution on the screen is known as the diffraction pattern. and there is no specific. at a finite distance from the diffracting aperture [see Fig. important physical difference between them. the effects due to diffraction are not readily observed. larger amounts of energy reach the geometrical shadow. Further.1 INTRODUCTION Consider a plane wave incident on a long narrow slit of width b (see Fig..1 If a plane wave is incident on an aperture then according to geometrical optics a sharp shadow will be cast in the region AB of the screen. The diffraction phenomena are usually divided into two categories: (i) Fresnel diffraction and (ii) Fraunhofer diffraction. interfering. interference corresponds to the situation when we consider the superposition of waves coming out from a number of point sources and diffraction corresponds to the situation when we consider waves coming out from an area source like a circular or rectangular aperture or even a large number of rectangular apertures (like the diffraction grating). In the Fraunhofer class of diffraction. However. 18.No one has ever been able to define the difference between interference and diffraction satisfactorily. is to say that when there are only a few sources. it seems that the word diffraction is more often used.2(a)]. We will discuss the phenomenon of diffraction in this chapter and will show that the spreading out decreases with decrease in wavelength. say two. Indeed. Vol. indeed.1). 1 Important Milestones 819 823 Joseph Fraunhofer demonstrated the diffraction of light by gratings which were initially made by winding fine wires around parallel screws Fraunhofer published his theory of diffraction George Airy calculated the (Fraunhofer) diffraction pattern produced by a circular aperture 18. Richard Feynman in Feynman Lectures on Physics. then the light intensity in the region AB is not uniform and there is also some intensity inside the geometrical shadow. the source and the screen are at infinite distances from the aperture. This spreadingout of a wave when it passes S A B s' Fig. since the light wavelengths are very small (A. We should point out that there is not much of a difference. It is just a question of usage. but if there is a large number of them. if the width of the slit is made smaller. between the phenomena of interference and diffraction. this is easily achieved by placing . According to geometrical optics one expects the region AB of the screen SS' to be illuminated and the remaining portion (known as the geometrical shadow) to be absolutely dark. in general. 18.5 x 105 cm). then the result is usually called interference.
18. the field produced by Al will differ in phase from the field produced by A2.2 SINGLESLIT DIFFRACTION PATTERN We will first study the Fraunhofer diffraction pattern produced by an infinitely long slit of width b. Thus. the collimator renders a parallel beam of light and the telescope receives parallel beams of light on its focal plane. Now.3(b)]. if the number of point sources be n. let n go to infinity and A go to zero such that nA tends to b.. the diffraction pattern corresponds to the Fresnel class.1)A (1) .0).3(b)]. are in phase and.. the amplitudes of the disturbances reaching from Al. The two lenses effectively moved the source and the screen to infinity because the first lens makes the light beam parallel and the second lens effectively makes the screen receive a parallel beam of light. therefore. all that one needs is an ordinary laboratory spectrometer. If the diffracted rays make an angle 0 with the normal to the slit then the path difference would be A2A2' = A sin 9 The corresponding phase difference. A plane wave is assumed to fall normally on the slit and we wish to calculate the intensity distribution on the focal plane of the lens L [see Fig. then b = (n . For an incident plane wave. we will. A 3.. A2. the points Al. the Fraunhofer diffraction pattern is not difficult to observe. 0.7 that cos wt +cos (cot +. 18. A 2. This follows from the fact that the optical paths A 1 B 1P and A2'B2P are the same..4 the source on the focal plane of a convex lens and placing the screen on the focal plane of another convex lens [see Fig. 18.+ cos (cot .. We will now calculate the resultant field produced by these n sources at the point P. because of even slightly different path lengths to the point P.. Further. It turns out that it is much easier to calculate the intensity distribution of a Fraunhofer diffraction pattern which we plan to do in this S Optics slit is a source of Huygens' secondary wavelets which interfere with the wavelets emanating from other points. A2.. (b) In the Fraunhofer class both the source and the screen are at infinity. However. P being an arbitrary point (on the focal plane of the lens) receiving parallel rays making an angle 0 with the normal to the slit [see Fig.. Now the difference in the phases of the disturbance reaching from the points A2 and A3 will also be 0 and thus the resultant field at the point P would be given by E = a[cos cot + cos (cot . 18. and let the distance between two consecutive points be A [see Fig.Point source S' (a) 1. In Chapter 20 we will study the Fresnel class of diffraction and will discuss the transition from the Fresnel region to the Fraunhofer region. would be given by 0= A sin 6 (2) chapter.(n . if the field at the point P due to the disturbance emanating from the point Al is a cos wt then the field due to the disturbance emanating from A2 would be a cos (wt .1)0)] (3) 18..18. Since the slit actually consists of a continuous distribution of sources. will be very nearly the same because the point P is at a distance which is very large in comparison to b.f H (b) Fig. at the point P.Asin0 Now. We assume that the slit consists of a large number of equally spaced point sources and that each point on the where 0= . we had shown in Sec. The diffracting aperture is placed on the prism table.2 Li (a) When either the source or the screen (or both) are at finite distances from the aperture. the additional path traversed by the disturbance emanating from the point A2 will be A2A2' where A2' is the foot of the perpendicular drawn from Al on A2B2..+cos [cot (n1)0] sinn l2 cos cot [ 2 (n 1) 0] (4) .2(b)]. Let the point sources be at Al. Thus..0) +.. 11.. 183(a)]. in the final expression.
then the propagation is only in cetain directions where the displacements add up in phase. 0_27r Asin9 =  .3 (a) Diffraction of a plane wave incident normally on a long narrow slit of width b. Notice that the spreading occurs along the width of the slit. the amplitude Ee becomes cos W2 giving rise to cos2 0/2 intensity distribution [cf.0 the product na tends to a finite limit. Notice that if we have a large number of equidistant sources oscillating in phase. (13) of Chapter 14].Fraunhofer Diffraction: I 18. ** We may mention here that in the limit n + oo and a . therefore.6 sin9 A = A sin /3 where* * A = na and (7) Further 2 = nAsin9absin9 2rbsin0 27r b sin 9 (8) A. Eq. 18. Thus E = Eo cos [cot  would tend to zero and we may.5 f F (b) Fig. the slit is assumed to consist of a large number of equally spaced points. for n = 2. (b) In order to calculate the diffraction pattern. write 2 (n 1) 0] (5) Ee where the amplitude Ee of the resultant field would be given by* a sin (n0/2) Ee (6) sin 0/2 In the limitofn * co and0 in such a way that nA b. Thus. we have asinl 1T 2 JJJ 2 sin 7cbsin9 =na . An * Equation (6) represents the amplitude distribution due to the interference of n point sources.
Thus = to r 2 sin cos /3 _ 2 spa /3 = 0 p2 Amplitude distribution Intensity distribution or sin /3[/3 . we differentiate Eq. Since sin 0 cannot exceed unity. Similarly when m = 3. In order to determine the positions of maxima. ± 2. The corresponding phase difference will be 1r and the resultant disturbance will be zero. 18.18. (b) Graphical method for determining the roots of the equation tan (3 = /3 Fig..2. (10) with respect to /3 and set it equal to zero. 18. the first and second quarters cancelling each other and the third and fourth quarters cancelling each other. (minima) We divide the slit into two halves as shown in Fig. i = 1 and I = lo which corresponds to the maximum of the intensity. It is obvious from Eq./3) (9) The positions of minima can directly be obtained by simple qualitative arguments.4 (a) The intensity distribution corresponding to the single slit Fraunhofer diffraction pattern. the slit is divided into six parts and so on. the disturbance from the point B will be cancelled by the disturbance reaching from the point B'. Thus the resultant disturbance due to the upper half of the slit will be cancelled by the disturbances reaching from the lower half and the resultant intensity will be zero. Let us consider the case m = 1. m = ± 1.5.. (12) (14) as the conditions for minima.4(a). . we divide the slit into four parts. Similarly. ± 3..6 Thus E =A s Optics e cos (wt . Clearly the path difference between the disturbances (reaching the point P) emanating from A and A' is sin 0 which in this case is . 18. The first minimum occurs at 0 = ± sin 1 the second minimum at 0 = ± sin 1 etc. Consider two points A and A' separated by a distance b/2.5 The slit is divided into two halves for deriving the condition for the first minimum.tan /3] = 0 (15) (a) _ A . (10) that the intensity is zero when =m7c. (b) .] Substituting the value of /3 one obtains b sin 0 = mA. the integer which is less than (and closest to) 1. the maximum value of m is (b) . In a similar manner when b sin 0 = 2A.1 Positions of Maxima and Minima The variation of the intensity with /3 is shown in Fig.m^0 (11) [When /3 = 0. The angle 0 satisfies the equation b sin 6 = 2i The corresponding intensity distribution is given by I=Io sine /3 132 (10) (13) where to represents the intensity at 0 = 0 18. iii A 37C_/j IF V2r (b) Fig. 18.
328 x 105 cm (After Ref. and are known as the first maximum.595 x 103. or sin 0=1. 17. the first and second maxima occur at /3 = 1. we have given the actual single slit diffraction pattern (as seen on a screen) for the following values of slit widths: 8.1 A parallel beam of light is incident normally on a narrow slit of width 0.015 m In Fig.6.2.46. This is because of the fact that the lengths of the slits were very large compared to their widths.96% of the central maximum.0176. thus.68% and 0. once again. The intersections occur at /3 = 1. The wavelength of the light used is 6.7 Consequently.46 x 2. A screen is placed at a distance of 3 m from the slit. 18. etc. Assuming that the screen is so far away that the diffraction is essentially of the Fraunhofer type. the conditions for diffraction minima are given by sin 0 = m21/b.328 x 105 cm. (2 As in Example 18.43yr and 2.5 x 103) = 0.5 x 10 2 cm and 7.5 x 103 .5 x10 3 b 2x102 Now. Similarly. Thus b sin 0 = 1. a parallel beam of light The condition sin /3 = 0. 3. on substituting the value of Alb. etc. /3 = 2.0088. Notice that since sin (2. Solution _5 A = 5x10 =2. calculate total width of the central maximum. we get 0 = 2.r.Fraunhofer Diffraction: I 18.r.43. thus the total width of the central maximum is approximately given by 2 x 3 x tan (2.0 x 102 cm. 1. The Fraunhofer diffraction pattern is observed on a screen which is placed at the focal plane of a convex lens whose focal length is 20 cm.4(b)]. The other roots can be found by determining the points of intersections of the curves y = /3 and y = tan /3 [see Fig. or /3 = m7r (m ^ 0) correspond to minima.2 Consider.5x103 and 2.437r is about 0.83% of the central maximum respectively.5 x 103 ) = 2. 0.43. used with permission). 3. 18.8 x 103 cm. 1.46. Since [sin (1. (ii) The values of Alb corresponding to the four slit widths are 7.6 The singleslit diffraction patterns corresponding to b = 0.5 x 103 radians.43) x 2.0496. the first minimum occurs at 0 = 2.1.1.2 mm. We may note the following two points: (i) The spreading is only in the direction of the width of the slit.46x Fig.05 cm on the focal plane of the lens.070 cm respectively.437r)1 2 1.76 x 10 2 cm.4999973 x 103 the error in the approximations sin 0 = 0 is about 1 part in a million! These minima will be separated by a distance (5 x 103 . respectively. Example 18.191 x 103.5 x 103) x 20 = 0. Assume A = 5 x 105 cm and that the lens is placed very close to the slit. 0. the intensities of the second and third maxima are about 1.03 x20 = 0.904 x 103.05 cm Example 18. Solution: = 5 x 105 cm) to be incident normally on a long narrow slit of width 0.43x2.2 mm. the maxima will be separated by the distance given by (2. Similarly. The light wavelength used was 6328 A = 6.035 and 0. 18.808 x 103 and 0.5 x 103 and 5 x 103 radians as the angles of diffraction corresponding to the first and second minima respectively. the second maximum. the intensity of the first maximum is about 4.1 and 2.5 x 1. Calculate the distance between the first two minima and the first two maxima on the screen. We assume 0 to be small (measured in radians) so that we may write sin 0 = 0 (an assumption which will be justified by subsequent calculations). The conditions for maxima are roots of the following transcendental equation tan /3 = /3 (maxima) (16) The root /3 = 0 corresponds to the central maximum.46 .
19.0176 cm P. The detailed derivation of the diffraction pattern for a circular aperture is somewhat complicated (see Sec.4 0 (b) 0.. 1. 3.8 0. most of the energy (of the diffracted beam) is contained between the first two minima.827 cm and 0. 18.328x105 cm Ln 0 .0088 cm and 0. In Fig..25 mm respectively.8 x 103 cm. (10) are given in Fig.7).8 8 (degrees) Fig. 18. In a similar manner one can discuss the diffraction of a plane wave by a circular aperture. of the aperture is placed close to the aperture and the Fraunhofer diffraction pattern is observed on the focal plane of the lens.7 The intensity distribution as calculated by using Eq.644 cm.8 0. (20) * Figure 18.8 shows the arrangement for observing the diffraction pattern.e. 3. a plane wave is incident normally on the circular aperture and a lens whose diameter is much larger than that . (10) for b = 0.8 Thus the diffraction angle at which the first minimum occur will be 0 = sin 0 = 7. 18.8 Experimental arrangement for observing the Fraunhofer diffraction pattern by a circular aper ture.76 x 102 cm. b=0.904 x 103. . 0. i^ ^^_. = 6. then the emergent wave spreads out (along the width of the slit) with angular divergence .5 mm and 0.288 cm.4 0.9(a) and (b) we have shown the Airy patterns corresponding to the radius of the circular aperture being 0. for Optics will 18.4 0 . For b >> i1.808 x 103 and 0. the light almost uniformly spreads out from the slit. the first minima occur at distances of 3. Thus the divergence angle (which would contain most of the energy) would be given by (see also Fig.5) and for b = 8. Because of the rotational symmetry of the system. i. 10. 27r asin0 v= A.0. 0. from the central maximum. the diffraction pattern will consist of concentric dark and bright rings.3) (18) For very small values of b. where the angles are measured in radians. 18.6 corresponds to the photographic film being 15 feet away from the slit. 1.328 x 10 5 cm).5 x 102 cm and 7.3 DIFFRACTION BY A CIRCULAR APERTURE In the previous section we have shown that when a plane wave is incident on a long narrow slit (of width b). 18.76 x 102 cm.8 x 103 cm and 1. respectively. we give here the final result: the intensity distribution is given by 0.8 +. Figure 18. 1.7 for b = 8. i focal plane of the lens) is known as the Airy pattern.18.8 (degrees) I where Io C 2Jt (v) 1 2 L v J (19) Fig. We should also mention that in the limit of A 3 0.413 cm. i .4 0. 09 > 0 and the diffraction effects are completely absent.191 x 103. Thus it records the Fraunhofer pattern (see Sec. this diffraction pattern (as observed on the back ^ <0< b (17) (where 0 is measured in radians).0176 cm 2.595 x 103.A/b.* The intensity distributions predicted by Eq.=6.0 x 102 cm.
In Fig. the wavelength of light and 6 the angle of diffraction. Other zeros of J1 (v) occur at v = 3. For those not familiar with Bessel functions. . 18.9 1 mm 1 mm I 1 mm (a) Fig. .. 10. we have v. we may mention that the variation of J1 (v) is somewhat like a damped sine curve (see Fig.11 Fig. 18. 18.r ^ v = a sin BJ A.11 we have plotted the function [2 J1 (v)12 L v J a being the radius of the circular aperture. = 0.5 J1 (v) 3. Fig. 7. On the focal plane of the convex lens v = 2n (x2 +y 2 )' a f (21) [2 J1 (v)] 2 where f is the focal length of the lens.25 mm respectively at the focal plane of a lens of focal length 20 cm (A.s Fraunhofer Diffraction: I 18.174. A.5 pm).016.. to is the intensity at e = 0 (which represents the central maximum) and J 1 (v) is known as the Bessel function of the first order. 18. 18.0 similar to the relation Lt 2 J1(v) = 1 v smx 1 _ Lt xo X 0.5 mm and a = 0. The intensity variation associated with the Airy pattern. (a) and (b) correspond to a = 0.016 0 I \J I.10 The variation of Jl (v) with v.832.10) and although J1 (0) = 0.9 (b) Computer generated Airy patterns. 4 ! I_ i 8 I I 12 [27.832 7.
Assume A. the divergence of a laser .8x103 =3.4 DIRECTIONALITY OF LASER BEAMS (b) (a) Fig.016. If f represents the focal length of the convex lens.5 mm and 0. the small gap acting almost like a point source.832. If the gap is large however.02 An interesting application of the above phenomenon is shown in Fig. thus we may say that the angular spread of the beam is approximately given by AO 0. Small. A (25) Fig. Thus the successive dark rings in the Airy pattern (see Fig.61 A..24 mm corresponding to a = 0.wikipedia.a sin g = 3. in this context. 18. 27r a 27ra (23) (22) Optics and only a small fraction of energy will reach the observer. 18.25 mm respectively. ripples emerge in an almost semicircular pattern.832 2.8 x 103 radians 0= 0 = sin 2 x 0. this will result in a greater diffraction divergence Thus the radius of the first dark ring is =20x1. On the other hand.12(a). 17. where D (= 2a) represents the diameter of the aperture.832 A. = 6 x 10 5 cm. if one uses a loudspeaker of larger diameter.f 7. however.18. 18.6x102 cm Similarly.25 mm respectively. A color photograph appears as Fig. A layman would expect that in order to obtain more directionality of sound waves.9) will correspond to v= .13 If an obstacle with a small gap is placed in the tank the ripples emerge in an almost semicircular pattern.12(b)]. 18. If the gap is large however. both figures correspond to A. = 5000A and f = 20 cm.02 18.org/ wiki/Ripple_tank].13 we have shown that if an obstacle with a small gap is placed in the tank the .v x 0. the radius of the second dark ring is 20x7.12 mm and 0. greater directionality is achieved [see Fig. 7. On the other hand. the diffraction is much more limited. Small. Detailed mathematical analysis shows that about 84% of the energy is contained within the first dark ring (see Sec.7x10 cm 2 2. then the Radii of the dark rings =f tan 0.12. we may say that the angular divergence associated with the diffraction pattern can be written in the following general form: A A0 Linear dimension of the aperture (26) Example 18. The Airy patterns shown in Figs 18. 10. An ordinary source of light (like a sodium lamp) radiates in all directions. 18.016A. one should use a loudspeaker of small aperture as shown in Fig.8. . Solution: The first dark ring occurs at _5 1'22 x 6 x 10 1. or sin 0 = 3.12 The directionality of sound waves increases with increase in the diameter of the speaker.9(a) and (b) correspond to a = 0. the diffraction is much more limited. 18.174. the small gap acting almost like a point source.02 cm at the focal plane of a convex lens of focal length 20 cm. in this context. [Figure adapted from http://en.5 mm and 0.10 which represents the intensity distribution corresponding to the Airy pattern.3 Calculate the radii of the first two dark rings of the Fraunhofer diffraction pattern produced by a circular aperture of radius 0.016x6x10_5 =6. Comparing Eqs (18) and (25). means that the size of the obstacle is comparable to the wavelength of the ripples.2). 17 in the prelim pages. means that the size of the obstacle is comparable to the wavelength of the ripples. 18. In Fig. Thus Radius of the first dark ring 0.016f 27ra ' 27ra (24) where we have assumed 6 to be small so that tan 0 = sin 0..3. 7.
5 mm 12. 0 increases with a decrease in the value of wo implying that smaller the initial spot size of the beam greater will be the diffraction divergence. the transverse amplitude distribution is approximately Gaussian. This quantity wo is called the spot size of the beam. greater is the diffraction).073° and w 6. .25 mm. 19. We define the diffraction angle as tan 0 = w(z) . z) = 1+ Y 2 exp . For most laser beams. 18. 18.018° and w 1. In Sec.76 mm 0.14 Diffraction divergence of a Gaussian beam whose phase front is plane at z = 0. the above equation is consistent with Eq.e. for wo = 0. = 0.11 Thus the transverse intensity distribution remains Gaussian with the beamwidth increasing with z. Notice that 0 increases with decrease in wo (smaller the size of the aperture. the intensity distribution is given by 2 z (28) (31) showing that the rate of increase in the width is proportional to the wavelength and inversely proportional to the initial width of the beam. indeed just when the beam is leaving the laser (which we assume to be z 0). for wo = 1 mm 20=0. beam is primarily due to diffraction effects. Then. as = 5000 A 2 mm 3.7 mm 0 2 4 6 z(m) 10 Fig.z 7c wo where we have assumed that the phase front is plane at z = 0. From the above equation it follows that at a distance wo from the zaxis. the wavelength is assumed to be 5000 A.25 mm. y) = a exp x2+ 2 2' w(z) wo Az 7cw0 = Az 7two (30) (27) wo which shows that the width increases linearly with z.. (26).14). the amplitude distribution can be assumed to be given by A(x. 20=0. In order to get some numerical values we assume A. For large values of z we obtain (>> wo/A. From Eq.88 mm at z = 10 m Similarly.5 µm. the amplitude falls by a factor 1%e (i.2(w2 ^^) where Az 7= wo w(z) = w0 [1 7t + 7 2] 1/2 z [1+ 7C2 w4 l oJI 2=W0 (29) (i) For a given value of A0 .Fraunhofer Diffraction: I 18. The figure shows the increase in the diffraction divergence as the initial spot size is decreased from 1 mm to 0. Y. (31) we find that I (x.35 mm at z = 2 m (see Fig.5 we will show that as the beam propagates in the zdirection. the intensity reduces by a factor 1/e 2).).
16 A filament placed at the focal plane of a convex lens. the radius of the beam is about 2 cm.02 radians 09 10 AO =1 JJ ^2 Ic which is independent of z. Thus for most practical sources. as the total energy crossing the entire x y plane will not change with z.18.00001 radians m which is much much smaller than the angular divergence of the beam due to the finite size of the filament. 18 in the prelim pages. (27) one can readily show that I(x. we have what is known as the geometrical optics limit.5. Notice the nondivergence of the beam. From Eq.15 we have shown a 50W laser beam with wavelength 589.17).6328 µm) is approximately Gaussian with wo = 5 µm. We will have a more detailed discussion in Sec.0 there is no spreading of the beam and.3° ^rwoJ Fig.wikipedia. 0 5 x50 18.4 The output from a singlemode fiber operating at the HeNe laser wavelength (2. If the linear dimension of the filament is about 2 mm (placed on the focal plane of a convex lens of focal length 10 cm) then the angular divergence of the beam (due to the finite size of the filament) is approximately given by 2 mm 0 mm = 0. Indeed. The angular spread of the beam is given by (see Fig.y. 18.15 A 50W laser beam with wavelength 589. Thus.0 . On the other hand.org/wiki/Laser .f represents the focal length of the lens and a represents the beam radius or the radius of the aperture of the lens (whichever is smaller). If the diameter of the aperture of the lens is 5 cm then the angular divergence due to diffraction would be cm = 0. o = 0. Usually laser beams are diffraction limited.z)clxdy Optics Thus. If such a diffraction limited beam is allowed to fall on a convex lens then of (33) a (see Fig. In Fig. n r'of 1 a 2 . 18. laser beams are usually diffraction limited.12 (ii) For a given value of wo the value of (9 (and hence the diffraction divergence) decreases with decrease in the value of X1. A color photograph appears as Fig.158 nm propagating through the atmosphere. 18. In Eq. Notice the nondivergence of the beam. 18. Thus Radius of the focussed spot Area of the focussed spot Am = Fig. if we have a tiny filament at the focal plane of a lens. 18. Figure adapted from http://en. Only if the size of the filament is smaller than 10 3 mm would the beam divergence be determined by diffraction.16) (32) f where l is the length of the filament and f the focal length of the lens. the corresponding divergence is 0 = tan ""I I = 2.1 Focusing of Laser Beams As mentioned earlier. as Al) .4.158 nm propagating through the atmosphere. 19. Example 18. A beam is said to be diffraction limited if it diverges only due to diffraction.16). the beam divergence is due to the finite size of the filament rather than by diffraction. (33). if a screen as placed at a distance of about 50 cm from the fiber. the beam will diverge primarily due to the finite size of the filament (see Fig. This is to be expected. 18.
1 8. the area of the focused spot size is = rr(Xo f/a) 2 .854 x 1012 MKS units represents the dielectric permittivity of free space and c = 3 x 108 m/s represents the speed of light in free space. 18. 5). W. then the wave emerging from the lens will get focused to spot of radius = Xo f/a..6 We next consider a 3 MW laser beam (A0 = 6 x 105 cm and beam width 2a = 1 cm) incident on a lens of focal length of 5 cm.19 A test target is vaporized and bursts into flame upon irradiation by a high power continuous wave carbon dioxide laser emitting tens of kilowatts of infrared light [Ref. then 2 Hof = 7 x 106 cm' Area of the focused spot A = a On the retina. the retina will be damaged not only because of high intensities but also because of large ultraviolet content of the sunlight). 19 in the prelim pages. Such high intensities lead to numerous industrial applications of the laser such as welding. Example 18. etc (see e. . cutting materials.(34) Fig.13 Such high electric fields results in the creation of spark in air (see Fig. 18. the electric field strengths are of the order of 10 9 V/m which results in the creation of a spark in the air.Fraunhofer Diffraction: I 18.19. then Area of the focused spot 2 A = z( 4f ) = 106 cm2 = 10 10 m2 a On the focal plane of the lens. If the pupil diameter ( = 2a) is taken to be 2 mm. In Fig. 18. We illustrate the effects of this focusing through some examples. In the following two examples.5 cm. Example 18.I_ 2 eocEo .18). At the focus. A color photograph appears as Fig. the intensity of the beam is related to the electric field amplitude Eo through the following relation [see Eq. (Photograph courtesy Dr. R. Thus laser beams (because of their high directionality) can be focused to extremely small regions producing very high intensities. Terhune).18 Focusing of a 3 MW peak power pulsed ruby laser beam. (34) we readily get Eo=5x109 V/m .17 If a truncated plane wave (of diameter 2a) is incident on an aberrationless lens of focal length f. where so = 8. Laser beam. the intensity will be approximately given by 2x10 3 W =3x106 I P W/m2 A 7 x 101o m2 Such high intensities will damage the retina!!!! So never look into a (seemingly innocent) low power laser beam.I Fig. Substituting I = 3 x 10 16 W/m2 in Eq. Ref.5 We consider a 2 mW laser beam (7b = 6 x 105 cm) incident on the eye whose focal length is given by f 2.g. we will calculate the intensities (at the retina of our eye) when we directly view a 500 W bulb or the sun (Caution: Never look into the sun.: http: / / en.org/wiki/C02_laser ]. the intensity will be approximately given by P 3xl06 W =3x1016 W/m2 10 10 m2 Now. 18. I4 f . (78) of Chapter 23] I=A= Fig. a test target is vaporized and bursts into flame upon irradiation by a high power continuous wave carbon dioxide laser emitting tens of kilowatts of infrared light. hole drilling.wikipedia.
Thus a small focused spot would lead to a small depth of focus.* It can be readily seen that the intensity would drop by about 20% at z=±0. We may mention here that the intensity distribution at the focal plane of the lens is given by Eq.14 Example 18.5 cm from the pupil of the eye.immediately follows that greater the radius of the beam. Thus Area of the pupil of the eye = 7r (1 x 1) mm2 = 3 x 10 6 m2 Power entering eye = (500 W) x nr2 47r R2 Optic's 4x103 W x (2 x 10 4 )2 m2 = 30 kW/m2 7r = 5 x 106 W Radius of image = Radius of source x demagnification =3cmx 500 =1. 6.21 If we look directly at the sun. because a laser beam can be focused to very narrow areas.8 We next calculate the intensity at the retina if we are directly looking at the sun (see Fig. this can damage the retina of the eye! Fig.(f/a)2 (37) fi +f2 I Fig. 18.g. it has found important applications in areas like eye surgery.7 We consider a 6 cm diameter incandescent source (like a 500 W bulb) at a distance of about 5 m from the eye (see Fig. after the focused spot. etc. the intensity along the axis is given by I° I_ [ where w= (36) A (f ) z sin (w/4) w/4 ] 2 (3S) x2. e. Sec.5x10 4 m where we have assumed the image to be formed at a distance of about 2. intensities as high as 30 kW/m2 are produced. (21).52. . one may use a beam expander (see Fig. Thus..5 x 10' )2 m2 eves 5m IFig.35xi03 x7cx106 =4mW The sun subtends about 0. 6.5 of Ref. welding. One usually defines a depth of focus as the distance over which the intensity of the beam (on the axis) decreases by a certain factor of the value at the focal point.5x and The power density in image 180 To summarize. Now The intensity of solar energy on earth = 1. However.5° on the earth. Indeed. Thus. We assume the pupil diameter to be about 2 mm. 18.18. (19) where the parameter v is given by Eq. 18. 18.21). a 2 mW diffraction limited laser beam incident on the eye can produce an intensity of about 106 W/m 2 at the retinathis would certainly damage the retina.2 mm =2x10 4 m and z = 0 represents the focal plane. 18.50. it is very dangerous to look directly into a 2 mW laser beam. From the above discussion it .22) to produce a beam of greater size and hence a smaller focused spot size. On the other hand. Thus The radius of the image of the sun 0. Example 18. whereas it is quite safe to look at a 500 W bulb.22 Two convex lenses separated by a distance equal to the sum of their focal lengths act like a beam expander.35 kW/m 2 Thus the energy entering the eye =1.20 A 500 W bulb at a distance of about 5 m from the eye. Indeed. The power density in image (5 x 106 W) = 70 W/m2 zc x (1. the beam would have a greater divergence and would therefore expand within a very short distance.20). 18. * The derivation of the formulae has been given at many placessee. the smaller will be the size of the focused spot and hence greater will be the intensity at the focused spot.
Fraunhofer Diffraction: I 18. (a) and (b) correspond to the angular separation of the two objects equal to 6VivD and 1. what we have plotted in Figs 18. the resultant intensity distributions are quite complicated (see Fig.15 which is usually referred to as the depth of the focus or focal tolerance. the focal tolerance is about 3 x 103 cm. 18.832) v )) LL v3. In the first case the objects are well resolved and in the second case (according to the Rayleigh criterion) they are just resolved. 18. (I) [2J1 (v)] v 2 Airy patterns .wikipedia.23 The image of two distant objects on the focal plane of a convex lens. as shown in Fig.25 and 18.23 the Airy patterns are shown to be quite far away from each other and. the Airy disc around each of the stars can be seen.25 and 18. Obviously.24 Image of the binary star Zeta Bootis by a 2. they are said to be resolved.26 are v (b) Fig. the solid curves represent the resultant.5 LIMIT OF RESOLUTION Consider two point sources. = 6 x 105 cm.22a.J 8. Since the radius of the first ring is 1. therefore.25 The dashed curves correspond to the intensity distribution produced by two point sources (pro ducing the same intensity at the central spot) independently.56 m telescope aperture. Notice that larger the value of a. 18. 18. 18. f /D the Airy patterns will overlap more for smaller values of D and hence for better resolution one requires a larger diameter of the objective. It is for this reason that a telescope is usually characterized by the diameter of the objective.23).23. As such. f = 10 cm and a = 1 cm. Fig. In Fig.832 (3) [_"Iv)1 2 + [2J1 (v3. the two objects are said to be well resolved. (2) [all (v3. 20 in the prelim pages. the Airy disc around each of the stars can be seen. If the diffraction patterns are well separated. 18.832)] v3.0 12 Fig. In Figs 18. each point source will produce its Airy pattern as schematically shown in Fig. For A.22 A. in each case we have assumed that the two sources produce the same intensity at their respective central spots. the system can be thought of as being equivalent to a circular aperture of diameter D. As discussed in the previous section. followed by a converging lens of focal length f. [Image by Bob Tubbs in http://en. 18. a 40 inch telescope implies that the diameter of the objective is 40".8. 18. such as stars (so that we can consider plane waves entering the aperture) being focused by a telescope objective of diameter D (see Fig. its focal length and the wavelength of light (see Example 18.24 we have shown the image of the binary star Zeta Bootis by a 2. The diameters of the Airy rings will be determined by the diameter of the objective. .56 m telescope aperture.3)./D respectively. 18. smaller will be the focal tolerance. 18.org/wild/Diffraction] A colour photograph appears as Fig. for example. we have plotted the independent intensity distributions and their resultant produced by two distant objects for various angular separations.832 In Fig.26.27).
18.25 and 18. This criterion of limit of resolution is called the Rayleigh criterion of resolution and the intensity distribution corresponding to this is plotted in Fig. Assuming the light wavelength to be 6 x 105 cm. 18. the better will be its resolving power. one finds that the minimum angular separation of two distant objects which can just be resolved will be 1. 18.18.16 Optics separation of 6 A/iVD and as can be seen the two images are clearly resolved. the resultant intensity distribution has only one peak and therefore the two points cannot be resolved at all. we should mention here that since the point sources are independent sources. 18.22x6xl05 . Figure 18.5 x 104 cm the intensity distributions on the line joining the two centers of the Airy patterns. their intensity distributions (Airy patterns) will add. telescope objective whose diameter and focal length are 5 cm and 30 cm respectively. the radius of the first dark ring (of the Airy pattern) will be 1.26 The dashed curves correspond to the intensity distribution produced independently by two distant point objects having an angular separation of 22t. and (c) unresolved. and as can be seen. If we choose this line as xaxis then the parameter v in Figs 18.25(b). The resultant intensity distribution (shown as a solid curve) has only one peak and hence the objects are unresolved.27. (b) just resolved.26 corresponds to DB= 3rD (39) v Fig.22 D x focal length = 1.22A _ 1.22A/D then the central spot of the one pattern falls on the first minimum of the second and the objects are said to be just resolved. the intensity distributions given in Fig. The actual diffraction patterns are shown in rig. if the angular separation of the two objects is 1.22x6x105 x 30 5 = 4. Finally. In order to get a numerical appreciation of the above results we consider a.26 is given by 27ra v=A fx (38) Now.23(a) correspond to two distant point objects having an angular It is immediately obvious that the larger the diameter of the objective. the diameter of the largest telescope objective is about 80" and the corresponding angular separation of the objects that it can resolve is = 0.27 (c) (b) Computer generated intensity distributions corresponding to two point sources when they are: (a) well resolved. 18. For example. This very low Fig./nnD.1. (a) .07 sec of arc.5 x 105 radians D 5 Further.
6 TWOSLIT FRAUNHOFER DIFFRACTION PATTERN In Sec. there is some phase relationship between the waves emanating from the two object points. one often uses blue light (or even ultraviolet light) for the illumination of the object.28 and we have sin a y OP' y' tan i' D/2 y' sir"' D/2 (42) where we have assumed sin i' = tan i'. 18. The angle a' is defined in Fig. Equation (43) also tells us that the resolving power increases with decrease in A. Using Eqs (41) and (42). In the above analysis. 1. The quantity n sin i is the numerical aperture of the optical system and the resolving power increases with increase in the numerical aperture. In the above discussion we have assumed that the two object points produce identical (but displaced) Airy patterns.0 n sin i (43) 18. a larger aperture still provides a larger light gathering power and hence the ability to see deeper in space. (43) will give the correct order for the limit of resolution. (=Ao/n ') represent the wavelength of light in free space and in the medium of refractive index n' respectively. In this section we will study the Fraunhofer diffraction pattern produced by two parallel slits (each of width b) separated by Fig.22A. 18. (40) D 2 x 101 where we have assumed the pupil diameter to be 2 mm. this is justified since the image distance (OP') is large compared to D. For the points P and Q to be just resolved. For example. at a distance of 20 m. It is of interest to note that if we assume that the angular resolution of the human eye is primarily due to diffraction effects then it will be given by De A^ 6 x 10_5 = 3 x 10^ rad. the eye should be able to resolve two points which are separated by a distance 3x 104 x20 =6x 103 m=6mm One can indeed verify that this result is qualitatively valid by finding the distance at which the millimetre scale will become blurred.61 A.5. However. we get 0. therefore. the point Q' should lie on the first dark ring surrounding the point P' and therefore we must have sin a = 1. As such. t Iu ^I' v 18. 18. 14. in general. 18.1 Resolving Power of a Microscope We next consider the resolving power of a microscope objective of diameter D as shown in Fig.28 The resolving power of a microscope objective.22 _ D n'D (41) which represents the smallest distance that the microscope can resolve. we get Y 0. (39) of Chapter 4]. 18.1 Fraunhofer Diffraction: I limit of resolution is never achieved in ground based telescopes due to the turbulence of the atmosphere. accordingly one has to set up a modified criterion for the limit of resolution such that the two maxima stand out. . V y' Q.03 x 108 cm and therefore such a microscope has a very high resolving power. for such a case the intensities will not be strictly additive (see Sec.61 20 y _ n' sin If we now use the sine law n'y' sin i' = ny sin i [see Eq. Thus.3 we had studied the Fraunhofer diffraction pattern produced by a slit of width b and had found that the intensity distribution consisted of maxima and minima. It is for this reason that in some microscopes the space between the object and the objective is filled with an oiland they are referred to as `oil immersion objectives'. the objects are illuminated by the same source and. Ao and A. in actual practice. nevertheless Eq.17 where n and n' represent the refractive indices of the object and image spaces. we have assumed that the two object points are selfluminous so that the intensities can be added up. in an electron microscope the de Broglie wavelength of electrons accelerated to 100 keV is about 0. rays emanating from the points P and Q will produce spherical wavefronts (after refraction through the lens) which will form Airy patterns around their paraxial image points P' and Q'. However. If that is not the case then the two central maxima will have different intensities.6).28. Let P and Q represent two closely spaced selfluminous point objects which are to be viewed through the microscope. Assuming the absence of any geometrical aberrations.
18./3  ^1 )] Fig. (9)] E1 = A Optics which represents the interference of two waves. d sin 0 (44) The intensity distribution will be of the form sin2 p cos2 y p2 (45) sl^/3 cos (cot .. The above equation can be rewritten in the form E=A sin P cosy cos (cot .18 a distance d. If the diffracted rays make an angle 0 with the normal to the plane of the slits. the intensity distribution is a product of two terms. We would find that the resultant intensity distribution is a product of the singleslit diffraction pattern and the interference pattern produced by two point sources separated by a distance d. In order to calculate the diffraction pattern we use a method similar to that used for the case of a single slit and assume that the slits consist of a large number of equally spaced point sources and that each point on the slit is a source of Huygens' secondary wavelets. used with permission]. 18. we have shown the two slit diffraction patterns corresponding to d = 0.070 cm respectively [After Ref. if the slit widths are very small (so that there is almost no variation of the sin 2 /3/P 2 term with 0) then one simply obtains the Young's interference pattern (see Sec. (A 2 . As can be seen.30.30 Fig. 0.328 x 105 cm..0088 cm and 7 = 6.070 cm with represents the phase difference between the disturbances (reaching the point P) from two corresponding points on the slits. A2.0176.035 and 0.0176. we assume that the distance between two consecutive points in either of the slits is A./3) Similarly. The doubleslit Fraunhofer diffraction pattern corresponding to b = 0.2 I) 11 1) where Y1 2 I=410 2. B2. B 2).18. Hence the resultant field will be E=E 1 +E2 =A si^P [cos (wt . which are separated by a distance d. each of amplitude A S"P and differing in phase by 01. the first term (sin2 /3//32 ) represents the diffraction pattern produced by a single slit of width b and the second term (cos 2 y) represents the interference pattern produced by two point sources separated by a distance d. be given by [see Eq. B 1 ). . B 3 . therefore..29]. 0. As before. 14.. 0.6). (in the first slit) and at B1.. the second slit will produce a field E2=A slnp/3 cos (cot /3^ 1 ) at the point P.035 and 0.. Indeed. 17. A 3 . The values of d are 0. The field produced by the first slit at the point P will. In Fig. 18../3) + cos (wt . Let the point sources be at Al. 0.. where ^1 = d sin 0 where ID sin2 /3//3 2 represents the intensity distribution produced by one of the slits. (in the second slit) [see Fig.. then the path difference between the disturbances reaching the point P from two consecutive points in a slit will be A sin 0.2 /3. by corresponding points we imply pairs of points like (A1.29 Fraunhofer diffraction of a plane wave incident normally on a double slit. 18.
2A.19 x 103] The corresponding angles of diffraction will be given by the following equation: b sin 6 = ma. A... 2. The second photo in Fig.0088 cm.328 x 105 = 8.2..4 1 0 (in degrees) Fig. one would need each slit to be so narrow that its width was comparable to the wavelength of the light. 18... the interference maxima are extremely weak 18. _ 1 0. 7E. if d sin 9 = 0. . 2A. it is an idealized pattern that is not likely to occur in real life. Further..0088 cm d= 0..32 is a real photo and shows singleslit diffraction pattern caused by a slit whose width is the same as the widths of the slits used to make the top pattern.5 1 b = 0.035 cm and 0.0088 cm and A. 3A..3ir.32 is a real photo and shows the diffraction pattern formed by a real double slit.3.31 The doubleslit intensity distribution as predicted by Eq. this is because of the fact that at 6=sin ' = sin l or when 7r 37r 57r bJ [6. (45) is shown in Fig. or when.5 0 0...070 cm respectively (b = 0. 18. 18. 18.Fraunhofer Diffraction: I b = 0. To get it.328 x 105 cm.1/Io n b = 0.328 x 105 cm n.27r.19 b sin 0= A..328x10 5 cm . 1 0.. 18.8 x 103 sins [7. = 6.32 is not a real photo..5 A tl A 0. = 6... 27r. 18. in Fig. a maximum may not occur at all if 9 corresponds to a diffraction minimum.07 cm A = 6.412° the first minimum of the diffraction term occurs..5 . The bottom photo in Fig.328 x 10 5 cm)..n/N.. These are usually referred to as missing orders..0088 cm d= 0._1.) (46) = 0.035 cm The interference maxima occur when y = 0.. 18.. 3A A=6.41 °.00719 radians = 0.31. 1 ^. around 0 = 0.1 Positions of Maxima and Minima Equation (45) tells us that the intensity is zero wherever /3= 7r. The intensity distribution as predicted by Eq.e.31 we can see that for b = 0.) (n=1. (m=l. (45) corresponding to d = 0.. and d sin ( 1 l n+2)A. 3. i. The top photo in Fig.I. (47) The actual positions of the maxima will approximately occur at the above angles provided the variation of the diffraction term is not too rapid. For example.6. The width of each slit is fairly big compared to the wavelength of the light..0088 cm and A.
Example 18.18.+ A where the first term represents the amplitude produced by the first slit. d = 7.1)0 1) (48) 1 Fig.7 NSLIT FRAUNHOFER DIFFRACTION PATTERN We next consider the diffraction pattern produced by N parallel slits. but that's not usually possible.972 x 103... 9 0 4 x 1 04 0 Thus the fringe width will be 15x12x2.33 Fraunhofer diffraction of a plane wave incident normally on a multiple slit.54x0. Solution: The interference minima will occur when Eq.5. (see Fig.33). 18.0413cm Fig. J n = 0.= 0 . 18. .31). 2. As can be seen.452 x 103. cos (cot . 4. 3.. 18. the distance between two consecutive slits is assumed to be d.328 x 10 5 cm (see Fig. 3.068 x 103.904x10 3 1n+2 . 5. The width of each slit is fairly big compared to the wavelength of the light.com/html_books/5op/chO5/ chO5.20 r. 18. 18. How many interference minima will occur between the two diffraction minima on either side of the central maximum? In the experimental arrangement corresponding to Fig. when +.30 the screen was placed at a distance of 15'. Calculate the fringe width.164 x 103. you would need each slit to be so narrow that its width was comparable to the wavelength of the light./3) + A %i. each of width b. This is a real photo.904xl0^ = 0. This is not a real photo. the intensity I=7o sin2 /3 sin2 Ny sine=(n+2)=0. = 6. A color photograph appears as Fig.(N..htm1].. 4. 21 in the prelim pages.cm and 2.260 x 103. The angular separation between two interference maxima is approximately given by (see Eq. = 0. etc.e../3) + cos (wt .lightandmatter. we assume that each slit consists of n equally spaced point sources with spacing i. As before.876 x 103. (48) we get E = `4 sn $ [cos (wt .(N . Thus the field at an arbitrary point P will essentially be a sum of N terms: E =A st^/3 3 cos (wt ./3 . 3. the second term by the second slit.1) 1)] (49) =A sin/3 sin where y= 1 = asin9 The corresponding intensity distribution will be (50) sine y /32 where Io sin2/3//32 represents the intensity distribution produced by a single slit. (46) is satisfied.356 x 10 2. i..780 x 103 Thus there will be sixteen minima between the two first order diffraction minima.8 x 103 cm. 2.+ c1) cos (wt .^ 1) sl^/3 cos (wt ./3 + +. [Figure adapted from http://www.. A real photo of a singleslit diffraction pattern caused by a slit whose width is the same as the widths of the slits used to make the top pattern. Rewriting Eq. and the various symbols have the same meaning as in Sec. 18. To get it. 1. 6.0 x 10. i 1 2 3 H Nw Optics 18.9 Consider the case when b = 8.. This idealized pattern is not likely to occur in real life.. 47) 6 . 1.32 1. A diffraction pattern formed by a real double slit.
) (51) This can be easily seen by noting that sin Ny = Lt N cos Ny = + N. (45)]. 2ir . ± 27r.. 1. the above function would become very sharply peaked at y = 0. to the double slit diffraction pattern [see Eq.. Eq. (10)] and for N = 2. 18.. the function vanishes when Asin /3 (52) (53) P where _ 2cb sin 0 _ 7tb d irbm d (54) y = N . the first term 18. As N becomes larger. . (m = 0. In Fig. lr.34 we have given a plot of the function sin2 Ny sine y as a function of y for N = 5 and N = 11. ±2. 2. one obtains intense maxima at y = m7r.Fraunhofer Diffraction: I 18..e. when d sin 0 = mA..34 The variation of the function sin 2 (Ny) / sin2 y with 7 for N = 5 and 12. ± 37r.. ± 7r.1 Positions of Maxima and Minima When the value of N is very large. ±N ±2N which are referred to as secondary minima.. (50) reduces to the single slit diffraction pattern [see Eq.7. One can immediately see that as the value of N becomes very large. Between the two peaks. but p ^ 0. i.21 ( sin ' p) p2 J distribution is a product of two terms. . sin2 Ny sine y (a) 7r 7r 7 (b) Fig. p = ± 1. ymt sin 7 cos y y.. the function would become more and more sharply peaked at 7 = 0. For N = 1.. ..m7c thus. 18.. the resultant amplitude and the corresponding intensity distributions are given by Lt E=N and Z I=NIo sin 22 13 represents the diffraction pattern produced by a single slit and the second term ( S Yy) represents the interference patS°n tern produced by N equally spaced point sources.
Typical diffraction patterns for N = 1.1) minima. therefore. and 4 are shown in Fig. n = 1. d = 0.. = 6. 18. used with permission). (50) it can easily be seen that the intensity is zero when either b sin 0 = nA. 18..22 Optics Fig..35 The multipleslit Fraunhofer diffraction patterns corresponding to b = 0. 3. m cannot be greater than d/A [see Eq. We may mention here two points: (i) A particular principal maximum may be absent if it corresponds to the angle which also determines the minimum of the singleslit diffraction pattern.. Between two such consecutive minima the intensity has to have a maximum.^. (55) (2N 1) A (2N+ 1) A (2N+ 2) A (57) N N ' N ' Thus. between two principal maxima we have (N .36 Equation (55) gives us the minima corresponding to the single slit diffraction pattern. (56) Fig.1°. When N is very large the principal maxima will be much more in tense in comparison to the secondary maxima.. 17.328 x 105 cm.36.0044 cm.:. The angles of diffraction corresponding to Eq.0132 cm and A. 18.. 2. This will happen when dsin0=mA and b sin 0 = A. 1.82°.0044 cm.. (51)]. there will only be a finite number of principal maxima. 2A. d = 0. The number of slits are 1./t quently. or Ny = pn. (56) are dsin0= A 2A NN (N1)A (N+1)A (N+ 2)A N N N The intensity distribution corresponding to the fourslit Fraunhofer diffraction pattern as predicted by Eq. they add up and the resultant field is N times the field produced by each of the slits. p ^ N. 3 and 4 respectively (After Ref. at these maxima the fields produced by each of the slits are in phase and. 0. 18.35 and the intensity distribution as predicted by Eq. 2. . From Eq. Physically. 3. The principle maxima occur at 6 = 0. thus.... (50) corresponding to b = 0. Since I sin 01 1. 3A.328 x 105 cm. (50) for N = 4 is shown in Fig. (58) (59) .275°. 2N.18.55°. the intensity has a large value unless itself is very small.0132 cm and A = 6. Such maxima are known as principal maxima. Notice the (almost) absent third order. 2. 0. these maxima are known as secondary maxima. conses.
for A = 5 x 105 cm. Since the distance between two consecutive grooves is extremely small. However. one of the important requirements of a good quality grating is that the lines should be as equally spaced as possible. (57). For a large value of N. which is also called the grating equation. cos 6. An arrangement which essentially consists of a large number of equidistant slits is known as a diffraction grating.» + 062..» sin A6. This is achieved by ruling grooves with a diamond point on an optically transparent sheet of material. 2. are satisfied simultaneously and is usually referred to as a missing order. Thus. (56).6 we have discussed the diffraction pattern produced by a system of parallel equidistant slits.). detachable from the parent grating... Thus Eq. only the first two orders of the spectrum will be observed.. for A = 6 x 10 5 cm. Rowland's arrangement gave 14. therefore.062i. and.) = mA ± (61) But sin (6. Fig.7. In contrast to ruled gratings. 1. various spectral components appear at different positions..g. the grooves act as opaque spaces.. for example.. if we are using a polychromatic source (e. The zeroeth order principal maximum occurs at 6 = 0 irrespective of the wavelength.» and 6.. 18. If 6... requires a large number of slits (typically about 15.» ) is known as the angular half width of the mth order principal maximum.438 lines per inch.Fraunhofer Diffraction: I 18. sharper spectral lines).. m = 0. Since the exact positions of the principal maxima in the diffraction pattern depend on the wavelength. the principal maxima corresponding to different spectral lines (associated with a source) will correspond to different angles of diffraction. m = 0. (62) (63) which shows that the principal maximum becomes sharper as 18. therefore.. if b sin his close to an integral multiple of A. N increases. Nowadays gratings are also produced holographically.» = sin 6. the number of such minima will be very small..8. the third order spectrum will also be visible. Clearly. the angles of diffraction are different for different wavelengths and.» ± 06.2 Width of the Principal Maxima We have shown above that in the diffraction pattern produced by N slits. 06.54/14438 = 1.. therefore. (64) This relation. for m ^ 0.. 18. holographic gratings have a much larger number of lines/cm. the minima occur at the angles given by Eq. then i ( A0 1.8°). cos A6. (59) does not hold exactly (i. After each groove is ruled. For such a grating.759 x 104 cm.. Thus the grating spectrum provides us with an easily obtainable experimental set up for determination of In Sec.000 per inch). 18.. Thus by measuring the angles of diffraction for various colours one can (knowing . represent the angles of diffraction corresponding to the first minimum on either side of the principal maximum. corresponding to d = 2.36 around 6 = 0.. the movement of the sheet is obtained with the help of the rotation of a screw which drives the carriage carrying it. 1.. the mth order principal maximum occurs at dsin6 = mA. = 062rn which we write as 06... consequently. can be used to study the dependence of the angle of diffraction 6 on the wavelength A. ± A6. However. a large value of N is required... the pitch of the screw must be constant.. An appropriate strength solution of cellulose acetate is poured on the ruled surface and allowed to dry to form a strong thin film. and it was not until the manufacture of a nearly perfect screw (which was achieved by Rowland in 1882) that the problem of construction of gratings was successfully solved. From Eq. white light) then the central maximum will be of the same colour as the source itself.n) = sin 6. (60) Further. we will also have the diffraction minima (see Eq.5). however. ... the maximum value of m would be 2. the machine lifts the diamond point and moves the sheet forward for the ruling of the next groove..8 THE DIFFRACTION GRATING In Sec.23 wavelengths. ± cos 6.known as the. (61) gives us Nd cos 6».1 The Grating Spectrum 18. (63) we see that for narrow principal maxima (i.» + 06I .e. These impressions of a grating are preserved by mounting the film between two glass sheets. the intensity of the corresponding principal maximum will be very weak (see. 18. A good quality grating. (ii) In addition to the minima predicted by Eq. 55). the corresponding diffraction pattern is. Commercial gratings are produced by taking the cast of an actual grating on a transparent film like that of cellulose acetate. 2. d sin (6..6 we have shown that the positions of the principal maxima are given by d sin 6 = rA. » ± A 0.e.grating spectrum. when N is very large. Further.. where one records the interference pattern between two plane or spherical waves (see Example 14.. Even when Eq.
(i) Show that if we use a white light source the second and third order spectra overlap. A parallel beam of white light emerging from L1 falls on the grating and the diffraction pattern is observed on the focal plane of the lens L2. 18.473 = 28.69 x 104 = sin 0.23° 1.000 lines per inch.. Thus Solution: 02.38 If instead of a point source we have a slit in the focal plane of L1 then one will obtain bands on the focal plane of L2.38. If we differentiate Eq. for large 0. (64). Example 18. Equation (65) tells us that z0 is inversely proportional (ii) to d..37 and 18.. is a constant.e.69 x 10.828 = 55.37 we have shown a small hole placed at the focal plane of the lens L1. measured and knowing the value of d. one can calculate the wavelength of different spectral lines. so that for a given m.38 show schematic diagrams of the experimental arrangement for studying the grating spectrum of a polychromatic source.90° and = sin1 3 x 4x 105 = sin 0.38) . The angles of diffraction for various orders of the grating spectrum can be. . O0/&.. cos 0 = 1) we can see that the angle ^0 is directly proportional to the order of spectrum (in) for a given 0A. 18.. In Fig. we would obtain ..10=sin 0.37 Fraunhofer diffraction of a plafie wave incident normally on a grating. it can easily be shown that the dispersion is greater at the red end of the spectrum.cm 15000 Let 6.10 Consider a diffraction grating with 15. However. = sin 1 2xx. ing in different directions.18. and in the focal plane of the lens L2 we will have a band spectrum as shown in Fig. It may be mentioned that the intensity is maximum for the zeroeth order spectrum (where no dispersion occurs) and it falls off as the value of rn increases.2° 1..we would have parallel beams propagat Fig. Figures 18. 18. and therefore smaller the grating element.69 x 10 where we have assumed the wavelengths of the violet and red colours to be 4 x 105 cm and 7 x 105 cm respectively.60 Optics __ na d cos 0 (65) From this result we can deduce the following conclusions: (i) Assuming 0 to be very small (i. The lens L2 is the objective of a telescope and the diffraction pattern is viewed through an eyepiece.24 the value of m) determine the values of the wavelengths.710 = 45. (ii) What will be the angular separation of the D1 and D2 lines of sodium in the second order spectra? (i) The grating element is d = 2' 54 = 1. 18. Such a spectrum is known as a normal spectrum and in this the difference in angle for two spectral lines is directly proportional to the difference in wavelengths.r represent the angles of diffraction for the mth order spectrum corresponding to the violet and red colours respectively. the larger will be the angular dispersion. 18.69 x 104 02 r 2 x7 x 105 = sin 1 1. Since 02r a B3 Fig.as it is indeed the case in a typical laboratory set up .. If instead of a hole we have a slit at the focal plane of L1 (see Fig. and 0.
the resolving power is proportional to the order of the spectrum. and A + AA respectively.39 dil( The Rayleigh criterion for the resolution of two spectral lines.39). We assume that the refractive index decreases with A (which is usually the case) so that 8 also decreases with A. (69) it appears that the resolving power of the grating would increase indefinitely if N is increased. (66) which are exposed to the incident beam (see the derivation in Sec. 2 ) 2 } 112 0^ Fig. N must be at least (5. as N is increase.89 x 105)/(6 x 108) = 1. the larger the resolving power. Further since sin the third order spectrum for the red colour will not be observed. if the principal maximum corresponding to the wavelength A + AA.000. we have for small AA: 03 . 18. From Eq. 18.8.. the smaller the value of AA. falls on the first minimum (on the either side of the principal maximum) of the wavelength A. (d cos 9) A9 = m(AA) or A9 = rr mAA.8.69x10 i1. (ii) Since d sin 9 = mA.40 the points P1 and P2 represent (67) (68) = mN (69) Fig. PI and P2 represent the images corresponding to A. if we are using telescope of angular magnification 10. According to this criterion.69 x 10 ) 2 1/2 = 0.Fraunhofer Diffraction: I 18.0010 radians = 3. In Fig.4) can again be used to define the limit of resolution. Bar > 1. 18. then the two wavelengths A and A + AA are said to be just resolved (see Fig. the two lines will appear to have an angular separation of 34..5A then only the first order spectrum will be seen. 18.40 gives a schematic description of the experimental arrangement for observing the prism spectrum which is determined through the following formula: sin n(A) _ A+8(A) 2 A sin 2 where AA is the separation of two wavelengths which the grating can just resolve. for a given width of the grating D(= Nd).5A. Thus to resolve the D1 and D2 lines of sodium (AA. Further. = 6 A) in the first order.3 Resolving Power of a Prism We conclude this section by calculating the resolving power of a prism. 18.A. d decreases and therefore the maximum value of m also decreases. 18. Thus if d becomes 2. then the two wavelengths A and A + AA will be just resolved if the following two equations are simultaneously satisfied: d sin 0 = m(A + AA) and dsin6=mA+ A.6).40 which implies that the resolving power depends on the total number of lines in the gratingobviously on only those lines The schematic of the experimental arrangement to observe the prism spectrum. however. The Rayleigh criterion (see Sec.2 Resolving Power of a Grating In the case of a grating the resolving power refers to the power of distinguishing two nearby spectral lines and is defined by the following equation: R .47' Thus.7 '. If this common diffraction angle is represented by 0 and if we are looking at the mth order spectrum. Thus R = N (70) where A represents the angle of the prism and 8 the angle of minimum deviation. 18.25 the second and third order spectra will overlap. 2x6x10s 2x6x10 5 1 1. . only first and second order spectra will be seen and if it is further reduced to about 1. Figure 18. 18.
54] =4.i) + cos i = 0 i = S. for t = 2. the wavelength dependence of the refractive index (in the visible region of the spectrum) can be accurately described by the Cauchy formula n=A+. we have 0= or 2 [7r(A+b)] A+S b sin 0 = a = cos 2 where the length a is shown in the figure. 18. Thus sm .41)..18. + AA respectively. (79) (80) (74) i2 2d sin Hence. C = 6.88 x 1022 cm4 For A = 6 x 105 cm we have do d^ =[4. 18.41).5 cm we have R = A = 1000 which is an order of magnitude less than for typical diffraction gratings with 15.26 the images corresponding to A and A. thus we must have AA (71) Optics the negative sign implying that the refractive index decreases with increase in wavelength..e.000 lines. . For 8 to be minimum .i) + sin i] = 0 cos (S . Thus.9 OBLIQUE INCIDENCE Till now we have assumed plane waves incident normally on the grating. for most glasses. 18.i) + sin i] = mA (77) when S = i + 9 is the angle of deviation.i = 0 (78) Now.40. we need ** Data quoted from Ref. more practical to consider the more general oblique incidence case (see Fig. (12)] of A should fall at the central maximum of A + AA. principal maxima will occur when (76) d(sin 0 + sin i) = mA or d[sin (S .30x102 cm1 Thus. (72) we get for the resolving power AS R A tdA (73) i. sin 2 2 Now from Fig. If the angle of incidence is i. at the position of minimum deviation. the grating condition becomes (75) Z = mA.2)*.51375. or di [sin (S . 2 Thus _dn d ^ + ... therefore. the intensity distribution will be similar to that produced by a slit of width b (see Sec.^ + 4 +.. then the path difference of the diffracted rays from two corresponding points in adjacent slits will be d sin 9 + d sin i (see Fig. As an example we consider telescope crown glass for which** A = 1. 2. 18. For the lines to be just resolved the first diffraction minimum [m = 1 in Eq. we differentiate Eq. We are assuming that AA is small so that the same position of the prism corresponds to the minimum deviation position for both wavelengths. It is.27 x 102 + 3. Further t/2 A 2 = a where t is the length of the base of the prism.608 x 1011 cm2.. we must have In order to express AS in terms of AA. In an actual experiment one usually has a slit source (perpendicular to the plane of the paper) forming line images at P1 and P2. (70): 2 Thus AS do AA A+S(A) dA cos 2 2 sin rrA+S(A) 1 dS 1 do _ Acos[ d A 2 d2. 18. The wavelength measurement can be carried out by using the method of minimum deviation as we do for prisms. Since the faces of the prism are rectangular. For experimental setting it is quite difficult to achieve the condition of normal incidence to a great precision and it is easily seen that slight deviations from normal incidence will introduce considerable errors. (81) * Since we have a slit source +5 not consider the diffraction in a direction perpendicular to the plane of the diagram. B = 4. (72) AA b dA Substituting in Eq.
18. Take the reciprocals of these numbers and multiply by the LCM of the denominators. a (211) plane intercepts the three axes at z ..41 Diffraction of a plane wave incident obliquely on a grating. A color photograph appears as Fig. 18. Fig. Indeed.wikipedia. Figure adapted from http://en. Miller indices can also be negative. Obviously. Enclose in parentheses. a (111) plane intercepts all three axes at one unit distance (see Fig. each intercept with an axis being measured in terms of unit cell dimensions (a. the atoms or molecules arrange themselves in a regular threedimensional pattern which can be obtained by a three dimensional repetition of a certain unit pattern.43(a). 18. a (110) plane intercepts the zaxis at co.Fraunhofer Diffraction: I 18.1 A. it is extremely difficult to make slits which are narrow enough for the study of Xray diffraction patterns. Miller indices are universally used as a system of notation for planes within a crystal. 2. They specify the orientation of planes relative to the crystal axis without giving the position of the plane in space with respect to the origin. Xrays have extensively been used to study crystal structures. In an ideal crystal. Find the intercepts (of the plane nearest to the origin) on the three axes and express them as multiple or fractions of the unit cell dimension. 22 in the prelim pages. Malhotra for his help in writing this section. Xrays are also electromagnetic waves whose wavelengths are .43(b)).10 XRAY DIFFRACTION* Visible light is an electromagnetic wave whose wavelength approximately lies between 4000 A and 7000 A. the minus sign is shown above the digit like J.44 shows the planes characterized by the Miller indices (1 11) in a simple cubic lattice.42). This simplest volume which has all the characteristics of the whole crystal and which completely fills space is called the unit cell. WI W2 and W3 W4 represent the incident and reflected * The author is grateful to Professor Lalit K. 18. In Fig. The minimum deviation position can be obtained in a manner similar to that used in the case of a prism and since the adjustments are relatively simpler.42 Skewed planes in a NaCl crystal.org/wiki/ Bragg_diffraction. 18. These indices are based on the intercepts of a plane with the three crystal axes. this provides a more accurate method for the determination of R. one can use it as a threedimensional diffraction grating for studying the diffraction of Xrays.45 the horizontal dotted lines represent a set of parallel crystal planes with Miller indices (hkl). One can think of various identifiable planes in the regular threedimensional periodic arrangement Fig. 11). For example. 18. . To determine the Miller indices of a plane. Similarly. the following procedure is used: 1. 3. Consider a monochromatic beam of Xrays to be incident on a crystal. b or c) along that axis. 18.27 (see Fig. 1 and 1 unit distances (see Fig. Since the interatomic spacings in a crystal are usually of the order of Angstroms. Figure 18.
The crystal is built up by the repetition of this unit cell in three dimensions. where a represents the lattice constant.18.46).m=1. 18. 18.28 Optics or when 2dhkl sin 0 = mA (83) (a) (b) Fig. For solids which crystallize in cubic structures (which are discussed later). 18. This equation is known as Bragg 's law and gives the angular positions of the reinforced diffracted beams in terms of the wavelength A.4 and Fig. we can determine the value of 0 at which Bragg's equation can be satisfied.. (83) is not satisfied. is called the order of diffraction and 0 is known as the glancing angle. When the Bragg condition 2d sin 0 = m2 is satisfied.3.44 Planes characterized by the Miller indices (T 1 1) in a simple cubic lattice. 3. Constructive interference occurs when the condition given by Eq.. points Al. if there is an atom at the centre of each cube (shown as 9. . of the incoming Xrays and of the interplanar spacings dhkl of the crystal planes. 10.. Obviously. 11 and 12 in Fig.43 (a) The (111) plane intercepts all three axes at one unit distance of each axial dimension. 18.46 A body centred cubic (bcc) lattice.. Bl and C1 will also be in phase on W3 W4 if XB1+B1Y=m). 12. where dl.2. 18. Figure 18.1 and 1 unit distances. When the condition expressed by Eq. the secondary wavelets emanating from the points A. wavefronts respectively. Thus knowing the Miller indices.45 shows a simple cubic structure (abbreviated as SC) in which the atoms are at the corners of a cube which forms what is known as a unit cell. (b) The (211) plane intercepts the three axes at %. (83) is satisfied leading to peaks in the intensity distribution. we can find dhkl and from Bragg's law.7).. B and C are in phase on W3 W4 (see Sec. 12.k1 is the interplanar spacing between crystal planes of indices (hkl). the interplanar spacing dhkl between two closest parallel planes with Miller indices (hkl) is given by a dhkl = v1h2 + k2 + 12 (84) Y Fig. arrangement is known 0 0 0 0. m = 1. the waves scattered from different rows will be in phase. In addition. (82) Fig.'0 0 0 0'0 0 0 0 0' Fig.. the. The (110 ) planes are separated by a/. 2. and the waves emanating from the . There are three types of cubic structures: simple cubic. body centred cubic (BCC) and face centred cubic (FCC).. destructive interference occurs and no reinforced beam will be produced.45 Reflection of a plane wave by a set of parallel crystal planes characterized by the Miller indices (hkl ).
then for all positions of the microcrystal. (200). The distance between two adjacent planes characterized by the Miller indices (1 10) is a/ / which can be verified by simple geometry. 18. with reference to the Xray beam. it is clear that there are essentially three methods which can be used so that Bragg's formula can be satisfied: Rotating crystal method A Fixed e Variable (intentional) When one uses monochomatic Xrays. 18.16 A respectively*. diffraction occurs only on planes whose Miller indices when added together total to an even number. where h + k + 1 is an even number. 18. (220). This method can only be employed if single crystals of reasonable size are available. changes from one microcrystal to the other.k1 and a particular value of m.08 A respectively. 18. As the Xray beam passes through such a polycrystalline material. Thus.47) we will have the FCC structure. However for the BCC structure. barium and tungsten crystallize in the BCC form with a = 4.Fraunhofer Diffraction: I 18. (200). If the microcrystal is rotated about the direction of the incident Xray beam. etc.29 Powder method Laue method Fixed Variable Variable (inherent) Fixed as a BCC structure. zinc crystallizes into a hexagonal structure and carbon forms a diamond structure. Just as there are optical missing orders of a diffraction grating. one can still use monochromatic Xrays provided the sample is in the powder form so that there are always enough crystallites of the right orientation available to satisfy the Bragg relation. the orientation of any given set of planes. Hence one rotates the single crystal so that reflection can occur for a discrete set of 0 values.g. For simple cubic structures. corresponding to any given set of planes there will be a large number of crystals for which Bragg's condition will be satisfied. Consequently. reflections from all (hkl) planes are possible. and on the photographic plate one will obtain concentric rings [see Fig. 4. Thus for the BCC structure. A powder will consist of a large number of randomly oriented microcrystals.09 A and 4. Metals like sodium. silver and gold crystallise in the FCC form with the lattice parameter a = 3.47 A face centred cubic (fcc) lattice. e.03 A and 3. the photographic film is put in a cylindrical form surrounding the polycrystalline sample as shown in Fig. In the case of the FCC crystal structure. one will obtain concentric circular rings on the photographic plate. On the other hand. there are structural extinctions of Xray reflection from a crystal. the direction of the diffracted beam will be different.. 18. 18. Consider a set of planes parallel to AB [see Fig.48 (b)]. While using the powder method. Copper. 2dhkl sin 0 = m2. the principal diffracting planes are those whose Miller indices are either all even or all odd. The appearance of the circular rings can be understood as follows. but it will always lie on the surface of the cone whose semivertical angle will be 29. . * Crystal structures other than cubic are also common. etc. (211) (and other similar planes). each ring will correspond to a particular value of d/.49 (a). If this is not the case. From a study of the interplanar spacings one can determine the crystal structure*. From the position of these arcs one can calculate 9 and thus determine the interplanar spacing.1 Experimental Methods of Xray Diffraction From Bragg's law.49 (b) and (c). the principal diffracting planes for a first order diffraction are (110). for each position of the microcrystal. Although a powder camera with an enclosed film strip has been extensively used in the past. for example. However. The glancing angle e is assumed to satisfy the Bragg condition.10. the most important fact is that in all these structures there is a definite periodicity of atoms. the glacing angle will be the same for these sets of planes. each microcrystal is essentially a single crystal. if instead of having an atom at the center of the cube there is an atom at the center of each of the six faces of the cube (see Fig. Further. e I Y I • s. (111).. Each DebyeScherrer ring will produce an arc on the film.48 (a)]. one obtains a pattern as shown in Figs 18. these rings are known as DebyeScherrer rings. 5. and when the film is unrolled. Bragg's formula cannot be satisfied for an arbitrary value of 0.29 A. ss • <  Fig.61 A. modern Xray crystal analysis uses an Xray diffractometer which has a radiation counter to detect the angle and intensity of the diffracted beam.
. . (SC) (87a) Finally there is the Laue method in which the single crystal is held stationary in a beam of white Xrays.22).48 (a) When a monochromatic Xray beam falls on a polycrystalline sample one obtains the DebyeScherrer rings. 5. 18. for a simple cubic lattice. for an FCC lattice. 18.18.I N (86) N=h 2 +k 2 +12 * For more details. 2. higher order reflections are usually rather weak (see also Problem 18. Each set of planes then chooses its own wavelength to satisfy the Bragg relation (see Fig. (85) can be written in the form: sin 0 where Similarly. 8. (b) Fig. Miller indices are either all even or all odd implying =2 . 6. Thus Eq. 12. (b) and (c) represent schematic diffraction patterns for sodium and copper respectively.. 7. for a BCC lattice h + k + 1 must be even implying N=h2 +k2 +12 = 2. you may look up Ref. 22. 3.(BCC) (87b) Finally. Fig. (b) Diffraction from a polycrystalline sample. one obtains Laue spots.49 (a) While using the powder method the photographic film is kept in a cylindrical form as shown in the figure. (83)] to obtain 2a (85) sin 9 = m. 4. 18. 14. 7...50).50 When a polychromatic Xray beam falls on a single crystal. ^h 2 +k 2 +1 2 We restrict ourselves only to first order reflections (m = 1). Now. 18. (84) in the Bragg's law [Eq.30 Incident xray beam Optics d:!iUr y I.u: Microcrystalline Power (a) Xray beam III Polycrystalline Powder Sodium (b) (a) Copper (c) Fig. 18.. In order to calculate the angles of diffraction we substitute Eq. 10. all values of (hkl) are possible implying the following possible values of N: N = 1. 16.. 6. (84). 4. Each set of planes chooses its own wavelength to satisfy the Bragg relation given by Eq. 20.
Fraunhofer Diffraction: I N=h2 +k 2 +1 2 = 3, 4, 8, 11, 12, 16, 19, 20, 24, 27,... (FCC) (87b)
18.31 propagation of a light beam due to the dependence of the refractive index on the intensity of the beam. This leads to the selffocusing (or defocusing) of the beam. In order to physically understand the selffocusing phenomenon we assume the nonlinear dependence of the refractive index on the intensity to be of the form n = no
For a given structure and for given values of A. and a one can now easily calculate the different values of 0. For example, if we consider 2 = 1.540 A and 1.544 A (corresponding to the CuKai and CuKa2 lines) then for sodium (which is a BCC structure with a = 4.2906 A), the various values of 0 are (14.70°, (30.50°, (42.18°, (53.38°, (66.22°, 14.74°), 30.59°), 42.32°), 53.58°), 66.56°), (21.03°, (34.58°, (45.88°, (57.33°, (79.41°, 21.09°), 34.68°), 46.03°), 57.56°), 80.23°), (26.08°, (38.44°, (49.59°, (61.54°, 26.15°), 38.56°), 49.76°), 61.82°),
+
2 n' Eo
(88)
The two values inside the parentheses correspond to the two wavelengths 1.540 A and 1.544 A respectively. Because of the presence of two wavelengths one obtains double lines for each family of planes which become resolvable only at higher scattering angles. Similarly one can consider reflections from other structures (see Problems 18.19, 18.20 and 18.21). Each value of 0 will give rise to a DebyeScherrer ring shown in Figs 18.48 (a), 18.49 (b) and 18.49 (c). Finally, we should mention that the intensity of the diffracted wave depends on the number of atoms per unit area in the plane under consideration. For example, corresponding to the ( 110) and the (2 2 2) planes passing through a BCC lattice, there will be one atom and two atoms, respectively, in an area a2. Thus in the first case the intensity of the diffracted wave will be much more than in the second case. 18.11 THE SELFFOCUSING PHENOMENON* With the availability of intense laser beams, a large number of interesting nonlinear optical phenomena have been investigated. One such nonlinear phenomenon is the effect on the
where no is the refractive index of the medium in the absence of the electromagnetic field, n' is a constant representing the nonlinear effect** and E0 representing the amplitude of the electric field. As an example, we consider the incidence of a laser beam (propagating in the zdirection) having Gaussian intensity distribution in the transverse direction, i.e., we assume E(x, y, z, t) E0 cos (kz with
r2 \
cot)
(89)
Eo = Eoo exp  2
a ^
(90)
where a represents the width of the Gaussian beam and r ( = /x2 + y2 ) represents the cylindrical coordinate. In the absence of any nonlinear effects the beam will undergo diffraction divergence. However, if the beam is incident on a medium characterized by a positive value of n', the intensity distribution will create a refractive index distribution which will have a maximum value on the axis (i.e., at r = 0) and will gradually decrease with r. Indeed, using Eqs (88)(90) we will have n = no +2n'Eoexp

22' a J
(12
2 (no + n' E00 )  i no
(91)
* Based on Ref. 8; for a rigorous account, e.g., Ref. 9. ** This dependence may arise from a variety of mechanisms, such as the Kerr effect, electrostriction, thermal effect, etc. The simplest to understand is the thermal effect which is due to the fact that when an intense optical beam having a transverse distribution of intensity propagates through an absorbing medium, a temperature gradient is set up. For example, if the beam has a Gaussian transverse intensity variation (i.e., of the form exp (r2/a2 ); the direction of propagation being along the zaxis), then the temperature will be maximum on the axis (i.e., r = 0) and will decrease with increase in the value of r. If do/dT > 0, the refractive index will be maximum on the axis and the beam will undergo focusing; on the other hand if dn/dT < 0, the beam will undergo defocusing (see, e.g., Ref. 9). The Kerr effect arises due to the anisotropic polarizability of liquid molecules (like CS 2). An intense light wave will tend to orient the anisotropically polarized molecules such that the direction of maximum polarizability is along the direction of the electric vector; this changes the dielectric constant of the medium. On the other hand, electrostriction (which is important in solids) is the force which a nonuniform electric field exerts on a material medium; this force affects the density of the material, which in turn affects the refractive index. Thus, a beam having nonuniform intensity distribution along its wavefront will give rise to a refractive index variation leading to the focusing (or defocusing) of the beam. For a detailed discussion on electrostriction and Kerr effect, refer to Refs 911.
18.32 where a2 = noa 2 2n'E% (92)
Optics
and in writing Eq. (91) we have expanded the exponential term and have retained only the first two terms. In other words, we are restricting ourselves to small values of r, which is the paraxial approximation. The term n'Eoo is usually very small compared to no; so we may write (after squaring) 2 n2 = no 1( L.) a (93)
I^f'I
Fig. 18.51 When a plane wave is incident on a diverging lens, the transmitted rays diverge making an angle 0 a/f with the axis. nonlinear focusing effects will dominate and the ,beam will undergo selffocusing. Forfd  fn1, the two effects will cancel each other and the beam will propagate without any focusing or defocusing. This is the condition of uniform waveguide like propagation. In order to determine the critical power of the beam we note that the condition fd = L implies 1/2 Zka22 no a 2n'Eoo or E2 _ 1 0o non' 8a2 Now the total power of the beam is given by
We may recall that in Sec. 3.4.1 we had considered propagation in a medium whose refractive index decreased parabolically from the axis and had shown that the beam could undergo periodic focusing (see Fig. 3.25). Indeed we had shown that the medium behaved like a converging lens of focal length 7ra/2 [see Eq. (48) of Chapter 3]. In the present case also because of nonlinear effects (with n' > 0), the medium will act as a converging lens of focal length approximately given by v2 ^ no a (94) .f,l1 n a 2 2 2n' Edo the subscript (nl) signifying that the effect is due to a nonlinear phenomenon. Thus because of nonlinear effects the beam is said to undergo selffocusing; the word self signifies the fact that the beam creates its own refractive index gradient resulting in the focusing of the beam*. Our analysis in Sec. 3.4.1 for the calculation of the focal length was based on ray optics and neglected diffraction effects. Now, in the absence of any nonlinear effects, the beam will spread out due to diffraction and the angle of divergence will approximately be given by (see Fig. 18.14) __ (A o /no) 9d (95) ira Ira where Ao is the free space wavelength. Thus the phenomenon of diffraction can be approximated by a diverging lens of focal length (see Fig. 18.51).
fd 
'o
(98)
p = f velocity x (energy/unit volume) x 2ivr dr
0
=
f no o
Ixl 2 EEo x2grdr
J
l
12 not Eo Eoo
Co
)
f exp  azz
o
22rrdr (99)
ed
a
1 2 ka 2
= 4 nocEO
2 Eooa
(96) (97)
where
2g k=  = no
0
Clearly if fd < fit' the diffraction divergence will dominate and the beam will diverge. On the other hand, if fn1 < fd, the
where e (= nooeo) is the dielectric permittivity of the medium and eo (= 8.85 x 10 12 C2/Nm) is the dielectric permittivity of free space (see Sec. 19.2). Substituting the expression for Eoo from Eq. (98) in Eq. (99) we obtain the following expression for the critical power:
* It should be mentioned that if n' were a negative quantity, the refractive index would have increased as we move away from the axis and the beam would have undergone defocusing. For example, if the refractive index decreases with increase in temperature the beam may undergo what is known as thermal defocusing.
Fraunhofer Diffraction: I
18.33 Optical data storage is a system in which data is stored and retrieved by light, which happens to be a laser. Like majority of consumer products, the lasing takes place through a semiconductor laser diode. The optical system consists of two broad parts: a media or disc and a drive. The discs contain digitized information in the form of spiral track consisting of finite lengths of `pits' and `lands'. An optical drive contains the Optical Pickup Unit (OPU) and servo control system for controlling it, a disc rotation system and associated electronics mounted on a mother board. As shown in Fig. 18.52, the OPU is essentially a scanning optical microscope, which is mounted on a servo motor to scan any desired position of the disc. The OPU is used for both recording and reading of data. The OPU is equipped with a laser diode and a set of optical elements which focus the beam on the disc surface and detect the light reflected back. The spot size of the light is given by the familiar optical concept of NA, (Numerical Aperture) of the OPU lens as shown in Fig. 18.53. Higher the NA smaller is the spot size. Thus, as discussed later, this aspect is utilized for making smaller spot sizes to cram more data in same physical structure, resulting in transition to everincreasing data capacity in the form of DVD and BluRay format. The Pits and Lands are essentially physical features (protrusion) on the disc surface, which are put there through injection moulding. The heights of the pits from the surface are not arbitrary, rather it is fixed, being equal to A/4 where A is the wavelength of the laser used. As shown in Fig. 18.54, bump (pit) height causes a path difference of A/2 relative to land. The optical head reads information by
o P _ 32 (cE0) n' or
(100)
Garmire, Chiao and Townes (Ref. 12) carried out experiments on the selffocusing of a ruby laser beam (A0 = 0.6943 µm) in CS2 and found that the critical power was 25 ± 5 kW. Equation (100) gives us
Pcr
=
3.14 x3x10 8 x885x10  12 x (0.6943 x 106 ) 2 32 2 x 10 2o
(101)
= 6.3 kW
where we have used the following parameters for CS 2: no = 1.6276, n' = 1.8 x 10 11 cgs units = 2 x 1020 mks units. [The mks unit for n' is (meter/volt) 2 .] Although the result is wrong by a factor of about 4, one does obtain the correct order; this is indeed the case for all orderofmagnitude calculations. Thus (i) when P < Pcr, the beam will diverge due to diffraction. (ii) when P = Pcr, the beam will propagate without divergence or convergence. This is the condition for uniform waveguide propagation. (iii) when P > Pcr, we may extrapolate that the beam will undergo focusing, which is indeed borne out by more rigorous analysis. This is known as the selffocusing of the beam. We may mention that a detailed study of the selffocusing phenomenon is of considerable importance in laser induced fusion experiments where there is a nonlinear interaction of the laser beam with the plasma.
18.12 OPTICAL MEDIA TECHNOLOGYAN ESSAY1
Optical Media technology has been around for the last 30 years or so. There were various avtars of this technology beginning from Laser Video Disc (an optical media with Analog recording) to the dramatic breakthrough in the form of Compact Disc ROM, Compact Disc Recordable/ Rewritable, MagnetoOptical disc, DVDROM, Recordable & Rewritable to present day BluRay technology. The Optical storage solution using laser typically provides lowest cost per byte, is rugged, transportable and interchangeable. It has fast random access, is durable and is available in both erasable and nonerasable form.
Basically: A scanning microscope
Spot diameter = 2JNA
N
Fig 18.52 Optical Pickup Unit (OPU) is essentially a scanning optical Microscope, which is mounted on a servo motor to scan any desired position of the disc.
1 This essay has been kindly written by Dr Rajeev Jindal and Mr. Giriraj Nyati of Moser Baer India in Greater Noida, India. Moser Baer has done pioneering work in the manufacture of DVDs.
1834 Disc Rotation
Optics
Constructive Interference  In phase  Land Destructive Interference  Out of Phase  Pit
T r;.
Pit (height AI4) Focused Laser Laser Beam Spot
a
T
Numerical aperture Intensity profile NA=sin 8 Fig 18.53 The spot size of the light is determined by the wavelength and the NA (Numerical Aperture) of the OPU (Optical Pickup Unit) lens. Protective Layer
Pit Lenght
Fig. 18.56 Schematic diagram of reflection from the pit.
A/2

Fig 18.54 The Pits and Lands are essentially physical features (protrusion) on the disc surface, which are put there through Injection Molding. The heights of the pits from the surface are not arbitrary rather it is fixed being equal to .1./4 where A, is the wavelength of the laser used. capturing reflected light as the laser beam travels across the pits and landschanges in the light intensity are interpreted as 0's and 1 ' spolarity of pits can be either dark on bright background or reversed. This is shown in Fig. 18.55. The actual working can be seen through a schematic diagram as
10.000.010000000000 1001001 000000
Fig. 18.57 The Pits and Lands are essentially physical features (protrusion) on the disc surface, which are put there through Injection Molding. The heights of the pits from the surface are not arbitrary rather it is fixed being equal to ?/4 where ? is the wavelength of the laser used. A color photograph appears as Fig. 24 in the prelim pages. given in Fig. 18.56. The actual system and a cut way section is shown in Fig. 18.57 and Fig. 18.58. A CDROM substrate is made of optically clear polycarbonate over which the data marks are made through Injection molding. The inner hole has a diameter of 15 mm while the overall diameter of the disc is 120 mm and thickness is 1.2 mm. The top of the disc is covered with very thin layer of silver or gold to form a reflective layer which reflects back the laser beam so as to be read back. The reflected light is incident on a quadrant photo detector, which converts the light to suitable electrical pulses, which are subsequently processed to extract relevant data. The data capacity of the disc is typically 650700 MB of Digital Data.
Pit Pit Land Pit Land Fig. 18.55 Optical head reads information by capturing reflected light as the laser beam travels across the pits and lands; changes in the light intensity are interpreted as 0's and 1 's. Polarity of pits can be either dark on bright background or reversed.
Fraunhofer Diffraction: I
18.35
Red Laser A. 650 nm Increased Aperture of Objective: NA = 0.60 Thin Substrate: thickness = 0.6 nm Infrared Diod Laser: Wave length I = 780 nm smiple Objective Lens: Na = 0.45 Single Disc Substrate: thickness = 1.2 nm Fig. 18.60 The reduction in the spot size by decreasing the spot size and increasing the numerical aperture. Fig. 18.58 A CDROM substrate is made of optically clear polycarbonate over which the data marks are made through injection molding. The .mer hole has a diameter of 15mm while the overall diameter of the disc is 120mm and thickness is 1.2mm. The top of the disc is covered with very thin layer of Silver or Gold to form a reflective layer which reflects back the laser beam so as to be read back. The reflected light is incident on a quadrant photo detector, which converts the light to suitable electrical pulses, which are subsequently processed to extract relevant data. The data capacity of the disc is typically 650700MB of Digital Data. A color photograph appears as Fig. 25 in the prelim pages. Transition From CD to DVD: Due to the need for higher capacity and better resolution (picture quality), it was decided to go for a media having a higher capacity. Due to the need for backward compatibility, it was decided to keep the physical structure of the disc same viz., a plastic substrate of 120 mm diameter with a 15 mm diameter inner hole. However, as shown in Fig. 18.52, the spot size is proportional to the wavelength and inversely proportional to the NA, so in DVD the wavelength has been reduced to 650 nm and NA increased to 0.65. The result has been a reduced spot size. However, the optical path has been reduced in the process resulting in thinner substrate. This entire process is shown Figs 18.59 and 18.60. However, ever increasing hunger for Scanning Spot 1.7 pm Scanning Spot o6 pm L1
HDDVD
more data has resulted in a new product called BluRay where using precedent argument the NA has been further increased to 0.85 and wavelength reduced to 405 nm (BlueViolet). The resulting increase in capacity is 25 GB/layer as shown in Figs 18.61 and 18.62.
=750nm A.=650nm A.=400nm NA = Q6 NA = Q86 NA = Q46 Fig. 18.61 (a) Infrared diode laser ( A = 780 nm) with a simple objective lens with NA = 0.45. (b) Red laser (A. = 650 nm) with increase aperture objective with NA = 0.60. (c) Blue laser (A, = 405 nm) with further increase inNA = 0.85. A color photo appears as Fig. 26 in the prelim pages.
A
UDDVD
50 GB MultiLayer MultiLaser GB Super Resolation Large NA
Fig. 18.59 The reduction in the spot size by decreasing the spot size and increasing the numerical aperture.
1990 2000 1995 2005 2010 Fig. 18.62 The evolution of optical media technology.
18.36 Summary ♦ Interference corresponds to the situation when we consider the superposition of waves coining out from a number of point sources and diffraction corresponds to the situation when we consider waves coining out from an area source like a circular or rectangular aperture or even a large number of rectangular apertures (like the diffraction grating). ♦ When a plane wave is incident normally on N parallel slits, the Fraunhofer diffraction pattern is given by
I
Optics
= to
sin2 /3 sin2 Ny a2 sin2 y
7cdsin 0
where 7rb sin 0 P= 2 ,y=
A is the wavelength of light, 0 the angle of diffraction, b represents the width of each slit and d the separation between two slits. When N =1, we have the single slit diffraction pattern producing a central maximum at 0 = 0 and minima when b sin 0 = mA; In = ± 1,±2, ... When N> 2, the intensity distribution is the product of the single slit diffraction pattern and the interference pattern produced by N point sources separated by a distance d. For N = 2, we obtain the Youngs double slit interference pattern. For large values of N, the principle maxima occur when y= mac implying dsin0=mk;in=0,1,2,... which is usually referred to as the grating condition. ♦ The resolving power of the grating is given by
R=A =mN
where N represents the total number of lines in the grating. For example, in the first order spectrum (m = 1) of a diffraction. grating with N = 10000, for A = 5000 A we get AA = 0.5 A . • Consider a monochromatic beam of Xrays incident on a crystal. The glancing angle .0 for which we have reinforced diffracted beams is given by 2d1ik1 sin 0 = m2l, where dl,kl is the interplanar spacing between crystal planes having Miller indices (hkl); m = 1,2,3, ... is called the order of diffraction and 0 is known as the glancing angle. The above equation is known as Bragg's law and gives the angular positions of the reinforced diffracted beams. Problems 18.1 A plane wave (A. = 5000 A) falls normally on a long narrow slit of width 0.5 mm. Calculate the angles of diffraction corresponding to the first three minima. Repeat the calculations corresponding to a slit width of 0.1 intr. Interpret physically the change in the diffraction pattern. [Ans: 0.057°, 0.115 0.17°; 0.29°, 0.57°, 0.86°]
18.2 A convex lens of focal length 20 cm is placed after a slit of width 0.6 mm, If a plane wave of wavelength 6000 A falls normally on the slit, calculate the separation between the second minima on either side of the central maximum. [Ans: = 0.08 cm] 18.3 In Problem 18.2 calculate the ratio of the intensity of the principal maximum to the first maximum on either side of the principal maximum. [Ans: 21] 18.4 Consider a laser beam of circular crosssection of diameter 3 cm and of wavelength 5 x 105 cm. Calculate the order of the beam diameter after it has traversed a distance of 3 ktn. [Ans:  14 cm. This shows the extremely high directionality of laser beams.] 18.5 A circular aperture of radius 0.01 cm is placed in front of a convex lens of focal length 25 cm and illuminated by a parallel beam of light of wavelength 5 x 105 cm. Calculate the radii of the first three dark rings. [Ans: 0.76, 1.4, 2.02 mm] Consider a plane wave incident on a convex lens of diam18.6 eter 5 cm amid of focal length 10 cm. If the wavelength of the incident light is 6000 A, calculate the radius of the first dark ring on the focal plane of the lens. Repeat the calculations for a lens of same focal length but diameter 15 cm. Interpret the results physically. [Ans: 1.46 x 104 cm, 4.88 x 105 cm] 18.7 Consider a set of two slits each of width b = 5 x 102 cm and separated by a distance d 0.1 cm, illuminated by a Monochromatic light of wavelength 6.328 x 105 cm. If a convex lens of focal length 10 cm ii placed beyond the double slit arrangement, calculate the positions of the minima inside. the first diffraction minimum. [Ans: 0.0316 mm, 0.094 mm] 18.8 Show that when b = d, the resulting diffraction pattern corresponds to a slit of width 2b. 18.9 Show that the first order and second order spectra will never overlap when the grating is used for studying a light beam containing wavelength components from 4000 A to 7000 A. 18.10 Consider a diffraction grating of width 5 cm with slits of width 0.0001 cm' separated, by a distance of 0.0002 cm. What is the corresponding grating element? How many order would be observable at A, = 5.5 x 10 5 cm? Calculate the width of the principal maximum. Would there be any missing orders? 18.11 For the diffraction grating of Problem 18.10, calculate the dispersion in the different orders. What will be the resolving power in each order? 18.12 A grating (with 15000 lines per inch) is illuminated by white light. Assuming that white light consists of wavelengths lying between 4000 and 7000 A,. calculate the angular widths of the first and the second order spectra. [Hint: You should not use Eq. (65); why?]
Fraunhofer Diffraction: I
18.37
18.13 A grating (with 15,000 lines per inch) is illuminated by sodium light. The grating spectrum is observed on the focal plane of a convex lens of focal length 10 cm. Calculate the separation between the DI and D2 lines of sodium. (The wavelengths of the Dl and D2 lines are 5890 and 5896 A respectively.) [Hint: You may use Eq. (65).] 18.14 Calculate the resolving power in the second order spectrum of a 1 inch grating having 15,000 lines. 18.15 Consider a wire grating of width 1 cm having 1,000 wires. Calculate the angular width of the second order principal maxima and compare the value with the one corresponding to a grating having 5000 lines in 1 cm. Assume A = 5 x 105 cm. 18.16 In the minimum deviation position of a diffraction grating the first order spectrum corresponds to an angular deviation of 30°. If A = 6 x 105 cm, calculate the grating element. 18.17 Calculate the diameter of a telescope lens if a resolution of 0.1 seconds of arc is required at A. = 6 x 105 cm. 18.18 Assuming that the resolving power of the eye is determined by diffraction effects only, calculate the maximum distance at which two objects separated by a distance of 2 m can be resolved by the eye. (Assume pupil diameter to be 2 mm and A = 6000 A.) 18.19 (a) A pinhole camera is essentially a rectangular box with a tiny pinhole in front. An inverted image of the object is formed on the rear of the box. Consider a parallel beam of light incident normally on the pinhole. If we neglect diffraction effects then the diameter of the image will increase linearly with the diameter of the pinhole. On the other hand, if we assume Fraunhofer diffraction, then the diameter of the first dark ring will go on increasing as we reduce the diameter of the pinhole. Find the pinhole diameter for which the diameter of the geometrical image is approximately equal to the diameter of the first dark ring in the Airy pattern. Assume A= 6000 A and a separation of 15 cm between the pinhole and the rear of the box. (b) Figure 18.63 shows the quality of the image formed for various values of the diameter of the pinhole. Discuss qualitatively the fact that the image will get blurred if the diameter of the pinhole is too big or too small. [Ans: (a) 0.47 mm] 18.20 Copper is an FCC structure with lattice constant 3.615 A. An Xray powder photograph of copper is taken. The Xray beam consists of wavelengths 1.540 A and 1.544 A. Show that diffraction maxima will be observed at 9 = (21.64°, 21.70°), (25.21°, 25.28°), (37.05__°, 37.160), (44.94°,
45.09°), (47.55°, 47.71°), (58.43°, 58.67°), (68.20°, 68.58°), (72.29°, 72.76°). 18.21 Tungsten is a BCC structure with lattice constant 3.1648 A.
Fig. 18.63
The image formed in a pinhole camera for different diameters of the pinhole. [Photograph downloaded from the internet by Professor K Thyagarajan; Ref: http://www. cs.berkeley.edu/ daf/book/chapter4.pdf].
observe diffraction maxima at 6 = 20.15°, 29.17°, 36.64°, 43.56°, 50.39°, 57.55°, 65.74° and 77.03°. 18.22 (a) In the simple cubic structure if we alternately place Na and Cl atoms we would obtain the NaCl structure. Show that the Na atoms (and the Cl atoms) independently form FCC structures. The lattice constant associated with each fcc structure is 5.6402 A. Corresponding to the Xray wavelength 1.542 A, show that diffraction maxima will be observed at 6= 13.69°, 15.86°, 22.75°, 26.95°, 28.27°, 33.15°, 36.57°, 37.69°, 42.05° 45.26°, 50.66°, 53.98°, 55.10°, 59.84°, 63.69°, 65.06°, 71.27°, 77.45° and 80.66°. (b) Show that if we treat NaCl as a simple cubic structure with lattice parameter 2.82 A then the maxima at 9 = 13.69°, 26.95°, 36.57°, 45.26°, 53.98°, 63.69° and 77.45° will not be observed. Indeed in the Xray diffraction pattern of NaCl, the maxima corresponding to these angles will be very weak. 18.23 Show that the inth order reflection from the planes characterized by (hkl) can be considered as the same as the first order reflection from the planes characterized by (mh mk ml). 18.24 Calculate the Fraunhofer diffraction pattern produced by a doubleslit arrangement with slits of widths b and 3b, with their centres separated by a distance 6b. 18.25 Consider the propagation of a 1 kW laser beam (A, = 6943 A, beam diameter = 1 cm) in CS 2. Calculate fd and f„t and discuss the defocusing (or focusing) of the beam. Repeat __ the calculations corresponding to a _ 1000 kW beam and discuss any qualitative differences that exist between the two cases. The data for no and n2 are given in Sec. 18.10. 18.26 The values of no and n2 for benzene are 1.5 and 0.6 x 1010 cgs. units respectively. Obtain an approximate expression for the critical power.
Show that in the powder photograph of tungsten (corresponding to an Xray wavelength of 1.542 A) one would
18.38 REFERENCES AND SUGGESTED READINGS 1. M. Born and E. Wolf, Principles of Optics, Pergamon Press,
2. 3. 4. Oxford, 1975. F. A. Jenkin and H. E. White, Fundamentals of Optics, McGrawHill Book Co., New York, 1957. E. Hecht and A. Zajac, Optics, AddisonWesley, Reading, Mass., USA, 1974. A. Nussbaum and R. A. Philips, Contemporary Optics for Scientists and Engineers, PrenticeHall, Englewood Cliffs, NJ, 1976. K. Thyagarajan and A. Ghatak, Lasers: Theory and Applications, Plenum Press, New York, 1981; [Reprinted by Macmillan India Ltd., New Delhi.] A. Ghatak and K. Thyagarajan, Optical Electronics, Cambridge University Press, Cambridge, 1989; [Reprinted by Foundation Books, New Delhi.] A. R. Verma and O. N. Srivastava, Crystallography for Solid State Physics, Wiley Eastern, New Delhi, 1982. M. S. Sodha, `Theory of Nonlinear Refraction: Self Focusing of Laser Beams' Journal of Physics Education, (India), Vol. 1 (2), 13, 1973,
Optics
9. M. S. Sodha, A. K. Ghatak and V. K. Tripathi, SelfFocusing
of Laser Beams in Dielectrics, Plasmas and Semiconductors,
5.
6.
7. 8.
Tata McGrawHill, New Delhi, 1974. ' 10. W. K. H. Panofsky and M. Philips, Classical Electricity and Magnetism, AddisonWesley, Reading, Mass., 1962. 11. W. G. Wagner, H. A. Haus and J. M. Marburger, `Large Scale Selftrapping of Optical Beams in Paraxial Ray Approximation' Physical Review Letters, Vol. 175, 256, 1968. 12. E. Garmire, R. V. Chiao and C. H. Townes, `Dynamics and Characteristics of the Selftrapping of Intense Light Beams' Physical Review Letters., Vol. 16, 347, 1966. 13. J. W. Goodman, Introduction to Fourier Optics, McGrawHill, New York, 1968. 14. C. J. Ball, An Introduction to the Theory of Diffraction, Pergamon Press, Oxford, 1971. 15. J. M. Cowley, Diffraction Physics, North Holland, Amsterdam, 1975. 16. M. Francon, Diffraction, Coherence in Optics, Pergamon Press, Oxford, 1966. 17. H. F. Meiners, Physics Demonstration Experiments, Vol. II, The Ronald Press Co., New York, 1970.
Fourier analysis is a ubiquitous tool that has found application to diverse areas of physics and engineering. Joseph Goodman in the Preface to Introduction to Fourier Optics
19.1 INTRODUCTION
In this chapter, we will present a more general analysis of the far field diffraction of a plane wave by different types of aperture; this is known as Fraunhofer diffraction. We will first derive the formula for what is known as Fresnel diffraction, which will be used in the next chapter. We will then make the far field approximation, which will give us the Fraunhofer diffraction pattern; this will be shown to be the Fourier transform of the aperture function. We will also derive the Fourier transforming property of a thin lens that forms the basis of Fourier optics and of spatial frequency filtering.
19.2 THE FRESNEL DIFFRACTION INTEGRAL
We consider a plane wave (of amplitude A) incident normally on an aperture as shown in Fig. 19.1. Using HuygensFresnel principle, we will calculate the field produced at the point P on a screen SS' which is at a distance z from the aperture. Now, for a spherical wave diverging from the origin, the field distribution is given by e tkr r where r is the_distance from the source (at the origin) to the observation point. We consider an infinitesimal area dg drl (around the point M) on the plane containing the aperture; the field at the point P due to waves emanating from this infinitesimal area will be proportional to Ae tkr dg drl (1) r u
Fig 19.1 A plane wave incident normally on an aperture. The diffraction pattern is observed on the screen SS'. where r = MP. In order to calculate the total field (at the point P), we will have to sum over all the infinitesimal areas (in the aperture) to obtain u(p) = C j if A e ikr d di (2) r where C is a proportionality constant and the integration is over the entire aperture. From a more general theory, one can show that [see, e.g., Refs 14]: C __ ik _ 1 2nc (3)
[see also Sec. 19.3]. We thus obtain kr f e'r do d?1 JJ r u(P
_4
(4)
19.2 If the amplitude and phase distribution on the plane z = 0 is given by A(4, ri) then the above integral is modified to ikr r ( A(4, ti) d4 dr1 d4 (5) u(P) = iA, JJ In writing Eqs (4) and (5) we have made two assumptions: 1. Our analysis has assumed that the screen (in the plane of the aperture) does not affect the field at the point P. This assumption is valid when the dimensions of the aperture are large in comparison to the wavelength. A more accurate analysis would take into consideration the effect of the screen on the field at any point P; this, in general, is a very difficult problem. 2. We have used a scalar theory in which we have represented the field by a scalar function u; this implies that the electric field is in the same direction everywhere. This assumption will be valid when the line joining the point 0 and the observation point P makes a small angle with the axis. The quantity r, which represents the distance between the point M [whose coordinates are (4, 71, 0)] on the plane of the aperture and the point P (whose coordinates are x, y, z) on the screen [see Fig. 19.1] will be given by r = [x4) 2 +(y71)2 +z 2] u2
= z/l+a
Optics
expi z [(x  4)2 + (y  71)2]}d4 chi Fresnel Diffraction Integral The above equation can be rewritten in the form u(x, y, z) iA,z eikz exp{2z (4, 77) x (x2 + y2)} JJ A
itti
(9)
77 2)} exp j ik (42 + J 2z
+vri> cl dr1
(10)
where 27ry 27rx (11) and v = Az Az are known as spatial frequencies. Both Eqs (9) and (10) are usually referred to as the Fresnel diffraction integral. In the next chapter we will use the above, integrals to calculate the Fresnel diffraction pattern. We must mention here that in the Fresnel approximation, we have neglected the terms proportional to a2; this will be justified if it leads to maximum phase change which is much less than r. Thus the Fresnel approximation will be valid when u=
8
[(x )2 + (y 77)2 ] 1 kza << 7c z3 Thus, we must have
z>>
7r
(12)
where
a=
(x4)2 + (y71) 2
z2 z2
(6)
(,T x  ^) 2 +(y . A(
.
Now, for a < 1, we may write (7) a2 + 8If we assume a < < 1 and neglect quadratic and higher order terms in the above expansion, we would get + (x )2 + (y 77)2 (8) 2z 2z Further, in the denominator of Eq. (5) we may safely replace r by z so that we may write r r u(x,y,z)^ iAzeikz JA(4,77) x
Condition for Fresnel Approximation to be valid
(13)
As an example, we consider a circular aperture of radius a; if we observe in a region of dimensions much greater than a, then we may neglect the terms involving 4 and 77 on the right hand side to obtain: (14)
4A,
Thus for a circular aperture of radius 0.1 cm if we observe in a radius of about 1 cm the maximum value of (x2 + y2) will be about 1 cm2 ; if we assume A. = 5 x 105 cm, Eq. (13) would imply z >> 17 cm.
1 For example, for A = 6 x 105 cm, the factor cos
kr
becomes. cos 71 10 5 r)
As the value of r is changed from say 60 cmto 60.00002 cm, the cosine factor will change from + 1 to  0.5. This shows the rapidity with which the exponential factor will vary in the domain of integration, although the change in r is extremely small.
13
ri) (see Sec. z) = i z eikz f +°° ik y2 ikYZ e2z dX e2z dY J We introduce the Fresnel number where X = x . (3) is correct. rl) eit" 4+vti)dg d7) Diffraction JJ Integral ICI < (22) =0 il> 2 for all values of rl.Fraunhofer Diffraction: II and Fourier Optics 19.(9) can be written as u(x. we must have g2 + 1 2 Condition for Fraunhofer ]max (17) z>> Approximation to be valid In this approximation. in addition to the condition given by Eq. y.(10). (10) takes the form u(x. If we now use the integral (see Appendix A) a2 NF = (20) f eax2 + x dx = ^2 7c exp V a ^4a (15) Thus for the Fraunhofer approximation to be valid. we have ended up getting the correct result! The above equation also tells us that the value of C given by Eq. Thus. Y.Y.3 UNIFORM AMPLITUDE AND PHASE DISTRIBUTION We first consider the absence of any aperture. at z = 0 A(4. the function exp {(2 + can be replacd by unity or. Eq. (18) gives the very important result that the Fraunhofer diffraction pattern is the Fourier transform of the aperture function. Eq. rl) = A as it indeed should for a uniform plane wave. u(x. (18) we obtain u(x. z) = Ae`kz e ikz Tc2z tc2z 19. y.5 FRAUNHOFER DIFFRACTION BY A LONG NARROW SLIT (16) We first consider Fraunhofer diffraction of a plane wave incident normally on a long narrow slit of width b (along the c axis) placed on the aperture plane.2 corresponds to a rectangular slit . z) = iz eikz exp z (x2 + y2)} x Fraunhofer A(. then we will have a long narrow slit. This shows that in spite of all the approximations that we have made. l fb/2 .2 Diffraction of a plane wave incident normally on a rectangular aperture. we must have NF << 1 we would get u(x. z) = or. 19. rl) = A for all values of and r) +^ and Eq. For a circular aperture of radius a. (22) in Eq.77.J chi (23) (18) Fig. 9.3 which represents the Fraunhofer diffraction pattern. z) = i e ikz exp ck 2z (x2 + y2) x +b/2 " d e e iv. Figure 19. (17) would become z>> a2 (19) 19.6). . The integral on the right hand side is the two dimensional Fourier transform of the functionA(. we assume z to be so large that inside the integral in Eq.4 THE FRAUNHOFER APPROXIMATION In the Fraunhofer approximation. we will have A(. Substituting Eq. Thus Eq. the maximum phase change should be much less than in Thus.if the slit is very long along the 77axis. (13). 19. For such a case. y.and Y = y .
z = 100 cm and we have assumed A= 5 x 10 5 cm. the intensity is zero except on Because of the Sfun the xaxis.3.3 50 100 150 200 250 300 350 .3. (17) would give us 4cm z>> 15 2x(0.2].5 cm [i. 18. 350 19. (13) would give us 3 )27c5(v) (26) z >> 5(0.. We must remember that Eqs (29) and (31) are valid when both Eqs (13) and (17) are satisfied. (b) The corresponding (computer generated) Fraunhofer diffraction pattern on a screen at a distance of 100 cm from the aperture. if we observe in a region of radius 0. we obtain Aba u(x.01 cm.z e expj( l ZZ(x2 +y2) \s 1( Py sin Y) (29) (a) A square aperture of side 0. A= 5 x 10 5 cm. the figure corresponds to a = b = 0.5 cm.I0 Y2 e iub/2 2i nb sin 0 P A (25) sin e y sin e p2 (31) where _ ub _ irbx 2 Az and sin 0 .b sin/i I(P) . 23 in the prelim pages. z) ikZ 0 Fig.3(a) and 18. z) = Az eikzeX p1 z (x2 + y2 ) 1 (siL The above equation represents the Fraunhofer diffraction pattern by a rectangular aperture. Y.01 cm. The Fraunhofer diffraction of a plane wave incident normally on such a rectangular aperture will be given by u(x. 19. y. y < +2 cm.e.6]. Although in the above. (24) and sin 0 va Y__ 2 J e z" dri stay Az Ira sin0 (30) r e!' cg J y iu _2 u c iu4l +b/2 b/2 e iub/2 . Carrying out the integration as in the previous section. (x2 + y2) < 0. 19. 0 representing the angle of diffraction along z the ydirection. we have assumed that we are observing a region of radius 0.2]. Now. in Sec. .6 FRAUNHOFER DIFFRACTION BY A RECTANGULAR APERTURE We next consider a rectangular aperture (of dimension a x b) [see Fig. Thus u(x. 19. 19.19.3 [see Figs 18. thus the intensity distribution along the xaxis will be = 7 cm Further Eq. y. (25). see also Fig. Thus we may write for the intensity distribution .01) 2 (27) I IoI J 5x10We thus obtain the single slit diffraction pattern as discussed in Sec.4 Now. 19.3 we have shown 5(v) 27t and +b/2 b/2 Optics where /3 is given by Eq. 9. 0 representing the angle of diffraction along z the xdirection. We have chosen z 100 cm and we get the diffraction pattern as shown in Fig.I 4 cm (b) = i^.25 cm2 ] then Eq. The intensity distribution due to a square aperture (a = b) is shown in Fig.25)2 1 4x5x10  lction. we have plotted the diffraction pattern for 2 cm < x. z) +b12 Jb/2 eikz exp Lk (x2+ 2 Y) 2z e14 d +a/2 Ja/2  100 (a) 50 f e iv d i (28) 0 where we have chosen the origin to be at the center of the rectangular aperture [see Fig.
5 We may note that along the xaxis. the diffraction pattern will be of the form of concentric circular rings with their centres at the point 0'... 19.0 cm.4. there is no diffraction along the yaxis (see Sec. we may calculate the intensity distribution only along the xaxis (i. Ref. 2. m = 0.5 cm. 19. . respectively.5 Fig.g.. when y = 0 (34) where 0 is the angle that OP v = 0 and sin 9 = x z makes with the zaxis. sin a sin 0 A 7 [7r sin01 A becomes very sharply peaked around 0 = 0. Further. 1. (18) becomes u(P) = `4 e` kz iAz exp akr 2 a 2..5 cm. 1 .e. 0) on the plane of the circular aperture.. 4. ..Fraunhofer Diffraction: II and Fourier Optics 19..5 Cylindrical coordinates (p.c f e ikpsinBcos¢ p z 00 kasine 2ir (k J dp dO (35) Thus U(P) A eikz i^ z ikr exp 2z 2 kasine 1 (k sin 9)2 o d f ei^cos¢ J ikr2 A ikz d = exp { 2z iA. this is consistent with the positions of the minima in Fig.. 19.e. On the plane of the circular aperture we choose cylindrical coordinates [see Fig.4 Diffraction of a plane wave incident on a circular aperture of radius a. 19.4). the intensity will be zero when = rrbxl = Az mrc. for a * co). 2. V Now. corresponding to m 1. Thus u= x=k sin 0 and Eq. because of the circular symmetry of the system. 1. 57. 3 . Since 0 = 0 implies y = 0. 2.z e sin 9)2 J 0 CJc(C)dC (36) 2 The identities associated with Bessel functions can be found in most books on mathematical physics.5] = p cos ¢ and rf = p sin 0 (33) Fig. at points for which y = 0) and in the final result replace x by Ix2 + y2 . the function asiny = b z=m0.7 FRAUNHOFER DIFFRACTION BY A CIRCULAR APERTURE We consider a plane wave incident normally on a circular aperture as shown in Fig. 19. see. Consequently. e. 19. 3. 19. (32) or x = 0..3.0 cm. For the case of a long narrow slit (i.
19.4).  . we obtain a and Eq. The Fruanhofer pattern by an annular aperture is discussed in Problem 19.6.19.6 where c = k p sin 0 and use has made of the following well known relation 2rz (37) 1 eti5cosO do Jo(b) = 27r 0 Optics _ Jo (v) d Jo(v) dv + (v) dJ1( v) dv (44) 2 dv Thus F(r) = [Jo (v) + J1 (v)] If we further use the relation d g)] = C J0( g) d g Ji then Eq. f1) 03(43. 19.6 0 2 4 v ^ 6 8 f I(a)27ra da 0 The fractional energy contained in a circle of radius r. then The above function is plotted in Fig.7.7 Diffraction of a plane wave incident normally on an array of N identical apertures. where I(a) 27ra da would be proportional to the energy contained in the annular region whose radii lie between a and a + do. (36) becomes u(P) = 221 A ikz ikr 2 e exp i^ 2z (k sin 0)2 r (38) Jo (v) + J? (v) 0 = 1 . (39) for the intensity distribution. 18.5. We have already mentioned that the diffraction pattern (in the plane SS') will consist of concentric rings with their centers at the point 0'. 19. Thus the intensity distribution would be given by (39) I(F) = I [2 '1(v) ] where 10 is the intensity at the point 0' (see Fig.3. one can deduce from the curve that about 84% of light is contained within they circle bounded by the first dark ring and about 91% of the light is contained in the circle bounded by the first two dark rings. This is the famous Airy pattern which has been discussed in Sec. If F(r) represents the fractional energy contained in a circle of radius r.JJ(v) + J?(v) (45) Jo (v) + Ji (v)I o J1 (c) o ^kasin5 2 2Jl(v) rikr2 A tkz expj 2z ^^ca v j i^ze where v = k a sin 0. Clearly z Since v = k a sin 0. 19. 19. 113) where we have used Eq. Now . 1. The Fraunhofer diffraction pattern will be the sin 0 6 (41) ka v (42) 0 1(41. 19. (40) becomes j [2J1(v) ]2 v dv v F(r) = f[2J1 (v)] 2 vdv L v 0 (43). Jlzvv) 1 Ji(v)[ Jo( v ) d dvv) ] 4 Fig.8 ARRAY OF IDENTICAL APERTURES We next consider an array of N identical apertures as shown in Fig.0 ii 0. etc..5  f I(a) 2n6 da F(r) = (40) Fig.
(50) of Chapter 18]. If we compare Eq. _ sin 2 Ny sine y (55) e i[u n +a1j n ] (48) where us = Cif e i(u4'+vi l ') d ' dry (49) is the field produced by a single aperture. 19.4 and 17.8 Diffraction of a plane wave incident normally on an array of N identical equally spaced apertures. we obtain the diffraction pattern produced by a grating [see Eq. Thus.Fraunhofer Diffraction: II and Fourier Optics 19. Let (1. the field distribution on the back focal plane of a corrected lens is the Fourier transform of the field distribution on the front plane. y') exp[i(ux' + vy')]dx' dy' (56) represent the spatial frequencies. (18)]: u= = 1 + e iud + .8. [i(u4 + vn)] di. u 2 (51) represents the intensity distribution produced by N point sources. 03.2. + e iNud = 1.e.. Let 01.. A.2. . y). 03. this is the same result as derived in Sec. 02.9 SPATIAL FREQUENCY FILTERING In the next section. (2> 112). could represent the centers of the (rectangular) aperture.3. (56) with Eq. 18. Further. represent points that are identically situated inside the apertures. n=1. (29) of Chapter 9. 0) Fig. on the plane P1 in Fig.7 N ehd"1)d n=1. if each aperture is circular we obtain the product of the Airy pattern and the interference pattern produced by two point sources [see Figs 17. v) where.3.2. y) represents the filed distribution on the front focal plane of a corrected lens (i. we find that / 01 (0.exp s. we may assume =(n1)d and 71n=0. 19. When N = 2.. y) represents the field distribution on the back focal plane P2 then it is related to g(x. Thus if G(x. 19. 19. Thus N s n=1 n=1.. 02.. we obtain the interference pattern produced by 2 point sources. rl n ) as shown in Fig. if the apertures are rectangular in nature. we consider N equally spaced identical apertures as shown in Fig. As an example. 0) f and v= f (57) 0 3 (2d. respectively. the zaxis represents the optical axis of the lens.5]. 03.7. • • • represent the coordinates of the points O102. the resultant intensity distribution would be given by I = Is Il (50) where IS represent the intensity produced by a single aperture and N n=1 which is the interference pattern produced by N identically placed point sources. we will show that if g(x..3 i(N1)ud sum of the fields produced by the individual apertures and will be given by [see Eq. If each aperture is a long narrow slit. For example. then 01. rl ') represent the coordinates of an arbitrary point in a given aperture with respect to the point (n. X 11). 0) 02 (d. y) through the following relation: G(u.e ind c[55 + f f + .7.9) then on the hack focal plane P2 of the lens. then 0 N icd sin 6 y 2 and x sin 6 = z We therefore obtain N y (53) (54) u C^ n=1 ff ^[n(n+'>+v(^n+^'>] 2 e d' drl' (47) e iu(n1)d where ( c'. represents the wavelength of light and f is the focal length of the lens. 7 13).. On the other hand..N Thus 17 = off f f g(x' . 19. (3. one obtains the Fourier transform of g(x.. drl (46) = exp I where ud 2 i(N 1)ud] sin Ny sin y (52) where each integral represents the contribution from a particular aperture. Without loss of generality.e 1 .
y) = go cos(2irax) (61) = k = Z ^il^ = a^1. 4. then the field to the right of the plane Pt would be proportional to cos cot cos (27ra x) We know that for a plane wave with ky = 0. where sin 0 (59) where 8(u) and 8(v) represent the Dirac delta functions. the field variation is of the form (see Example 11. In the above figure. This important property of a corrected lens forms the basis of the subject of spatial frequency filtering which finds applications in many diverse areas (see.and 0 is the angle that the propagation vector k makes with the zaxis. we have ' an object whose transmittance is proportional to cos (27ra x). say) and go G(u. (67) These plane waves will obviously focus to two points at x = . 19. we have neglected an (unimportant) phase factor on the right hand side [see Sec.27ra x)] (64) where kx = k sin 0. a small hole is placed on the axis (in the plane P2) which filters out the highfrequency components. 19. if we use . If we use the identity cos 0 = we would obtain G(u. y) is a constant (= go. (64) represent two plane waves propagating along directions making angles 0 and + with the zaxis. 19. v) = if exp[i(ux' + vy')]dx dy' Now. i. Since 8(u) = 0 for u ^ 0. (32) of Chapter 9]. we have assumed no ydependence of the field.7). if we place the negative of the photograph shown in Fig. (60) that the intensity is zero at all points excepting at the point x = 0.11].A. This is a good approximation in most cases. e'"x dx = 27r 8(u) [see Eq. Since a represents the spatial frequency associated with the object.fa). 19. on the back focal plane. we would obtain G(u. At z = 0. 2.19. (56) are assumed to be from 00 to +00.g.kxx . These two spots will be lying on the xaxis (where v = 0) at u = ±2ira (ie. If in the plane P1. 14. v) = go 47r2 8(u) 8(v) (60) (58) Thus. at x = ±A. Physically this can be understood from the following consideration: When a plane wave is incident normally on the plane P1 (see Fig. 4 On the plane PI (see Fig. g(x.9). we get G(u.fa and x = +u fa on the xaxis in the plane P2. of a fringe and assume a plane wave to be incident normally on the film. one can infer from Eq. This is to be expected because a plane wave gets focused to a point by a corrected lens 3 . the field becomes cos (wt . k = . as such the limits of integration in Eq. (64). We are familiar with the fact that a general time varying signal can be expressed as a superposition of pure sinusoidal 3 We are assuming a very large dimension of the aperture of the lens. v) = Fig. Ref. We must mention here that in writing Eq. then the field distribution would be proportional to cos 2 (27cax) which is equal to 2 [1 + cos(27rax)].. we will obtain two spots in the plane P2. y = 0. We first consider a plane wave incident normally on the lens. This implies that g(x. we find that the two terms on the RHS of Eq.kxx) (66) Comparing the above equation with Eq.27ra) + 8(u + 27ra)] 8(v) (63) If we now use Eq. one essentially obtains. e. the spatial frequency spectrum of the object.e. 2 [e J 2 i2'tax ' + e i2nax ^e iux'dxi x r e ivy' dy ' f (62) 22r2 [8(u .kkz) (65) = 2 [cos (wt + 27ra x) + cos (cot .. kZ = k cos 0.8 Image plane Optics where a is a constant4..9 The plane P2 is the Fourier transform plane where the spatial frequency components of the object (placed in the plane P1) are displayed. Another interesting example is a onedimensional cosinusoidal field distribution in the object plane i. v) (e` e + e ie) = 4f . ` the time dependence is of the form cos cot. (59). 810). ..11(b) with the yaxis along the length . (56).6): cos (wt .
(see. Notice that in (c) shades of gray appear as well details such as the missing part of the eye glass frame.] does not contain the unwanted high frequency noise (see Fig 19. 19. can be expressed as a superposition of sinusoidal variations and one would get the corresponding (spatial) frequency components on the plane P2. In a similar manner. an image of the form shown in (c) is obtained. we have been able to filter out the spatial frequency a. the filters are to be put on the plane P2. (58) of Chapter 8].9 signals [see Eq. ±AfP. we will obtain. For this reason. 19.9. 19. the frequency spectrum is shown in (b). The subject of spatial frequency filtering finds applications in many other areas like contrast improvement. on the plane P3. A. if we place the plane P2 on the front focal plane of lens L2.10 (a) Low pass filter.. . 19.10(c)]. For an arbitrary object. Thus. f(x. 19. 536 (1969). 9. character recognition. we filter out the low frequency components [see Fig.e. i. y) will become f(x. then it will filter out the high frequency components [see Fig. As a simple application. Since the spots are closely spaced. A plane wave is assumed to be incident on a transparency containing one 2dimensional function g(x. 37. e. we consider a halftone photograph (like that in a newspaper).fa. 5 There will however be an inversion.10(a)]. Refs 2. 4 and 9).e.1 The 4f Correlator The 4f correlator is based on the convolution theorem discussed in Sec.9). On the other hand. on the plane P3. then the field distribution on the plane P3 would be proportional to cos 27rJ3x.f a. y = 0) on the plane P2. This is the basic principle behind spatial frequency filtering.7. y = 0) and (x = A. y) = A cos 2ncax + B cos 27c/3x (68) then one would obtain four spots on the plane P2 (all lying on the xaxis). Since the Fourier transform of the Fourier transform is the original function itself5 [see Chapter 9]. if we put a small stop on the axis. an annular aperture on plane P2 will act as a band pass filter as shown in Fig. 19. (b) high pass filter and (c) band pass filter. American Journal of Physics. Thus. When a pinhole is placed in the Fourier transform plane to block the highfrequency components.g.9) we will obtain the amplitude distribution associated with the object. if the amplitude variation of the object is of the form g(x. Spatial Filtering Experiments for Undergraduate Laboratories. then on its back focal plane (i. if we put a small hole on the plane P2. which consists of a large number of spots of varying shades that produce the image pattern.. 19.10(b).11(c)). etc. the field variation across an arbitrary object (placed on the plane PI). [After R. As another example. y) y) • (c) (b) (a) Fig. an image which (a) (c) Fig.11(a) and allow only the low frequency components to pass through (as shown in Fig. in the plane P3 in Fig. Phillips. 19.Fraunhofer Diffraction: II and Fourier Optics 19. 19. If we now put stops at the points (x = +2. it represents a high frequency noise and the overall image has much smaller frequencies associated with it. the plane P2 is often termed as the Fourier transform plane. This can also be seen by simple ray tracing. these spots will appear at x = ±Afa.11 (a) Shows a photograph consisting of regularly spaced black and white squares of varying sizes. if we put a transparency similar to that shown in Fig.
=exp ik 19.f ^< f . We will first show that the the paraxial approximation is x2 + 2 effect of a thin lens of focal length f is to multiply the incident exp ik(d2+ field distribution by a factor pL given by 2d2 Pz. 19. y)] Correlation ofg(x. y) appear on this plane A diverging spherical wavefront of radius d1 P3 Optics A converging spherical wavefront of radius d 0 ^. v) H(u. the incident spherical wave emerges as another spherical wave of radius d2. we have assumed x..e.g.13). immediately in front of the lens. figure adapted from Ref. y)] formed on this plane n Multiplicative transmission mask H(u. Thus.13 Spherical waves emanating from an object point 0.4]: 1 + d2 = f d1 (70) The negative sign in the exponent refers to the fact that we now have a converging spherical wave.I. v) which is nothing but the convolution of g(x. 4. v)]. The Fourier transform of g(x. I f . is placed on this plane. 2. The phase factor corresponding to the disturbance emanating from the point 0 = exp [ik(d1 + d2)] expl .y) and h(x.10 which is placed on the front focal plane of the first lens as shown in Fig. y) and h(x.10 THE FOURIER TRANSFORMING PROPERTY OF A THIN LENS 2d1 where in writing the last expression. 19.12 The 4f correlator. 19. then exp ik d2+ l l )J x2+ 2 2dyIJ = exp 2 x2+ +ikl 21 1 J d+ [ 2dyJ]PL or PL where d1 and d2 represent the magnitude of the distances of the object and image points from the lens. If the image point I is at a distance d2 from the lens. after refraction through a convex lens. Thus the product G(u. y) to be cross correlated G(u. v) [= FT of h(x. This concept is of considerable use in many applications (see. the phase distribution on the transverse plane P2 at a distance d1 from the point 0 (i. v) H(u. y) [= H(u. A transmission mask containing the Fourier transform of the second function. this is known as the paraxial approximation.. where r is the distance measured from the point O. 19.. Now 2 y2 112 l+x + r=(x2+y2+di)ll2=d1 dl J 1 f I. is simply exp(+ikr).. 19. if pL represents the factor that when multiplied to the incident phase distribution gives the phase distribution of the emergent wave. i.13) is given by 1 d1. 810). v) [= FT of g(x. y) [= G(u. 4. Thus.19. v) lies on the front focal plane of the second lens and therefore on its back focal plane. v)] is formed on the back focal plane of the lens. Ref. then d2 is given by [see Sec. we will obtain the Fourier transform of G(u.see Fig. y).d1 ^.>) Fig. h(x. which under In this section we will derive the Fourier transforming property of a thin lens [see Eq. J (71) . e. (56)]. emerge as spherical waves converging to the image point I.2 \ (dl r 2) ].' Fig. 11. x 2 +y2 ^d + exp(+ikr) ( =exp ikl d1 + x 2 + y2 11 2d1 /IJ Zff 2 2 (x2 + Y )J (69) Consider an object point 0 at a distance d1 from an aberrationless thin lens of focal length f (see Fig. y << we have confined ourselves to a region close to the axis of the lens. Input plane containing the function g(x. 1 v Since the image is formed at I.e.12.
The field g(x. the field distribution will be given by =f + exp{ia[C 2 2CC + x2 . let g(x. 11) e t(« +vn) f J f where we have neglected the unimportant constant phase factors. 19. If we substitute the above expression for I(x. Equation (81) is the same as Eq. y) placed at the front focal plane of a lens produces a field distribution h(x. (9)] [using to the fact that we are referring to a lens.e.14). and a similar expression for I(x. y) =f f00 Thus g(.r7) 2]}d4di (73) where (74) 2f . y) in Eq. (56) and gives the important result that . 19.2f From plane P3 the field will again undergo Fresnel diffraction and therefore on plane P4. ri) x exp j 2 [(x f = HH(x)Hn( y) (78) C 2 ]}dC 4)2 + (y .4)2 + (z. Obviously the field will undergo Fresnel H(x. (76) u(x.ri) 2 ]} u(x.11 the subscript L on p corresponds Eq. i) g(x. For a plane wave incident along the axis. y) = iAf f f+ f f g(C. y).14). at a distance f from the plane exp{ia(c + a )} P1 (see Fig. where Now.g) 2]dC (80) or H4(x) = e 2 ' ax 9c ia Hn (y).14 A field distribution g(x. 77) H(x. 17) H4 (x) H. 17)d4 d17 (77) P1 (see Fig.24" + Hn. Thus the effect of a thin lens on an inciexp{ia[(x . 17) = diffraction and on plane P2 it Will be given by [using Eq.y) ia h(x. =y 2 f f g (4. (9)] x exp{ia[(C. u(x. y)Ip 2 = i ff exp(ikf) x x exp{ia[(x . y. thus on the plane P3. it will be given by P2 P3 P4 H = exp[2iax4] L exp[ia(C.(y)d4 d17 g(. y) g(4. y) in the plane P4 at the back focal plane of the lens.y)2 + (y . Thus 2 2 2c(x+4)+(x+4) 2 (x+4) 2 +4 2 +x 2 =(Cg)22x1. Y) p4 = 7). and the resultant field again undergoes Fresnel diffraction from plane P3 to P4 to produce the field distribution h(x. (69). (69). (76) we would obtain f+d4 d17 u(x. the effect of a thin lens of focal length f is to multiply the incident field distribution by the factor pL given by Eq. 4.t1) 2 ]}d4dt1 (72) H(x) Now. because it is independent of x and y. We would first like to determine the field distribution on the plane P2 i. then it gets mul tiplied by a phase factor due to the presence of the lens.2) 2]14: d'C I Ip [a ff = f f+^ 2 2 ff g (4.. Y) 4 = f e`kf j l(x. Y)Ip4 = g (4. as shown earlier in this section.  represent the spatial frequencies in the x and y directions respectively.'r)Ip3 x in the above equation.a) 2 ] }4d2 (75) dent field is to multiply the incident phase distribution by a Substituting for ul p3 from Eq. y) first undergoes Fresnel diffraction from plane P1 to P2.g+x 2 +C 2 u(x. y)I p3 = ff erkf exp[ia(x + y )] x where g = x + 4. Now. If we use Eq. y. the emerging disturbance will simply be pL. (79) and a similar expression for 4 2 2i. y) which can be seen to be the paraxial approximation of a converging spherical wave front of radius f. (73) we get factor that is given by Eq.Fraunhofer Diffraction: II and Fourier Optics 19. y) represent the field distribution on the plane I(x.. x e2t a(x4+Yn) d4 di 11) e i(4 +vn)dC d1) where we have used Eq. (70) and neglect the first factor r e`kf5 r u(C. (69).(74) and f f • u=2ax= and v = 2ay = Af (81) Fig. 17) x exp{ia[(x 4)2 + (y .^) 2 + (y . 19. .f we obtain Eq.
5. Calculate the positions of the first three maxima and minima on the xaxis (implying = 0) and also on the yaxis (implying 0 = 0). M.. A.E. assume a convex lens (of focal length 20 cm) placed immediately after the aperture. Summary ♦ If the amplitude and phase distribution on the plane z = 0 is given by A(. 19. 7. http://en. J. y) represent the filed distributions on the front focal plane and on the back focal plane of a corrected lens then G(u. Saleh and M. Klein and T. Goyal and S. Fourier Optics: An Introduction. Steward. New York (1986). Co. the Fraunhofer diffraction pattern is given by. E. 9. t1) 71) 2 }(14 dr) +(y k = 27r/2(. A. J.5 x 105 cm. Tl) e `(4+vn) dg do For a plane wave incident normally on a circular aperture of radius a. Roberts & Co. New York (1970). Mullineux. G.2 cm x 0. Thus on the plane of the one obtains the Fourier transform of g(x. Arfken. 8.3 mm with a screen placed at a distance of 100 cm from the aperture.2 mm x 0. Contemporary Optics. USA (2005). Ghatak. Mathematics in Physics and Engineering. K. 19.3 The' Fraunhofer diffraction pattern of a circular aperture (of radius 0. This important property of a corrected lens forms the basis of the subject of spatial frequency filtering. Fundamentals of Photonics. Problems 19. 3. Wolf. y) and G(x. Macmillan India. y.A. 10. Assume A = 5. [Hint: The integration limits of p in Eq. y). Chua. calculate the area of the patch (on focal plane) which will contain 95%. G.2 In the above problem.V. the zaxis represents the optical axis of the lens. Academic Press. >>> 1 where a represents the aperture of the lens. Mathematical Methods for Physicists. v) = off ff g(x'. New York (1959). E.C. New Delhi (1981). John Wiley. B. Mathematical Physics. W.wikipedia.E. [Ans: 0. Born and E. Englewood. z) = i z e lk ' eXp 12z (x2 + y2) if A(. New York (1991). u = 27 the and off back focal f representlensspatial frequencies. Calculate the positions of maxima and minima in a region 0.. J. Second Edition. Assume a plane wave with A = 5 x 10 5 cm incident normally on the aperture. y.5 mm) is observed on the focal plane of a convex lens of focal length 20 cm. Pearson Education. New York (2004).org/wiki/Fourier_optics. Ghatak and K. Goodman. Seventh Edition. Teich. New Delhi (1985). Dover Publications. Show that both Fresnel and Fraunhofer approximations are satisfied.C. Furtak. We should mention here that in writing the limits in the integral fromoe to +00. Thyagarajan. Cambridge. of the total energy. Ple4. we have assumed the lens to be of infinite extent. Third Edition. Cambridge University Press.4 In the above problem. 19. M. the error involved is usually very small because in almost all practical cases a/1. REFERENCES AND SUGGESTED READINGS 1. rf) then the Fresnel diffraction pattern is given by u(x. num Press. I(F) = i L 2JI(v)12 v J where v = k a sin 8. Calculate the radii of the first and the second dark rings. Introduction to Fourier Optics. z) expj c [(x' where e ikz Optics ♦ If g(x.19. 2. Second Edition.5 Obtain the diffraction pattern of an annular aperture bounded by circles of radii al and a2 (> a 1 ). Academic press.2 cm of the screen. 6.18 mm] 19. Optics. Optics. Hecht. UK (1999). Principles of Optics. I. Singapore (2002). John Wiley.13 mm. Irving and N. reprinted by Macmillan India.1 Consider a rectangular aperture of dimensions 0.12 the field distribution on the back focal plane of a corrected lens 'is the Fourier transform of the field distribution on the front plane. y') exp[i(ux + vy )]dx'dy' v where. (103) must be a1 and a2] f f A(. . ♦ The Fraunhofer diffraction pattern is the Fourier transform of the aperture function and is given by u(x. New York (1978). 0.
2. Fresnel and Arago carried out the experiment to demonstrate the existence of the Poisson spot validating the wave theory..One of your commissioners. Poisson.1. when a >>> A. when the rays penetrate there at incidences which are only a little oblique. M. Marie Cornu developed a graphical approach to study Fresnel diffractionthis came to be known as the Cornu 's spiral. 1. the secondary wavelets emanating t6 *The author found this quotation in Ref. Poisson predicted a bright spot at the center of the shadow of an opaque discthis is usually referred to as the `Poisson spot'.3 we had shown that the beam will undergo diffraction divergence and the angular spreading will be given by 2a Thus. The consequence have been submitted to the test of a direct experiment. In the present chapter we will discuss the Fresnel class of diffraction and also study the transition to the Fraunhofer region. This phenomenon is a manifestation of the fact that when a >>> 2. so that the screen is effectively at an infinite distance from the aperture. In Sec. the intensity at a point R (which is deep inside the geometrical shadow) will be negligible.1 INTRODUCTION In Chapter 18 we had mentioned that the phenomenon of diffraction can be broadly classified under two categories: under the first category comes the Fresnel class of diffraction in which either the source or the screen (or both) are at a finite distance from the diffracting aperture. Using Fresnel's theory. on the other hand. In the second category comes the Fraunhofer class of diffraction (discussed in the last two chapters) in which the wave incident on the aperture is a plane wave and the diffraction pattern is observed on the focal plane of a convex lens. be just as illuminated as if the screen did not exist. there will be almost uniform spreading out of the beam resulting in an (almost) uniform illumination of the screen. 18. if a . 20. The underlying principle in the entire analysis is the HuygensFresnel principle according to which: Each point on a wavefront is a source of secondary disturbance and the secondary wavelets emanating from different points mutually interfere: In order to appreciate the implications of this principle we consider the incidence of a plane wave on a circular hole of radius a as shown in Fig. . had deduced from the integrals reported by the author [Fresnel] the singular result that the centre of the shadow of an opaque circular screen must. 1816 1817 1818 1874 20. Dominique Arago to the French Academy of Sciences* II Important Milestones Augustin Fresnel developed the theory of diffraction using the wave theory of light. and observation has perfectly confirmed the calculation.
A. we are assuming that n is not very large number.. P can be written as m+1 um + a. this will be followed by a more rigorous analysis of the Fresnel class of diffraction and its transition to the Fraunhofer region. where un represents the net amplitude produced by the secondary wavelets emanating from the nth zone. the area of the nth halfperiod zone is given by An = 7C Y. one can have a onetoone correspondence between points in various zones..d (2) 20.20. Since the areas of the zones are approximately equal.. then the aperture almost acts as a point source resulting in a uniform illumination of the screen (see Fig. =2VA. it also depends on an obliquity factor which is proportional to i (1 + cos x) where x is the ...1) (3) W' nth half period zone Fig.1)th circle is known as the nth half period zone.1 Diffraction of a plane wave incident normally on a circular aperture of radius a.2J _d2 = jn Ad or r„n/ 1/2 [1+] L2 (1) Fig.>>> A. this is indeed justified for practical systems using visible light..u4 + . The amplitude produced by a particular zone is proportional to the area of the zone and inversely proportional to the distance of the zone from the point P.2.d (n . We will first introduce the concept of Fresnel halfperiod zones to have a qualitative understanding of the Fresnel diffraction pattern. 20. we make the fol Thus the areas of all the halfperiod zones are approximately equal. d + 3A/2.+ (. then with point P as centre we draw spheres of 'adii d + A/2.^ . the alternate negative and positive signs represent the fact that the resultant disturbances produced by two consecutive zones are pc outofphase with respect to each other. u(P) = u1 . 20. 20.1). Now the resultant disturbance produced by the nth zone will be 1r outofphase with the disturbance produced by the (n .1) A . where we have assumed d.7C r'i1 n[n A. there is a corresponding infinitesimal area surrounding the point Q„ in the (n .2 Construction of Fresnel halfperiod zones. Thus.1)th [or the (n + 1)th] Zone. 20.1)th halfperiod zone such that QnP ..2 from different points on the circular aperture so beautifully interfere to produce (almost) zero intensity in the geometrical shadow and a large intensity inside the circular region (see Fig. the resultant amplitude at the point . ei which corresponds to a phase difference of n. The radius of the pith circle will obviously be given by ll2 [(d+ n.. if a . 12.. d + 2A.12.Q n_1 P A = 2 .3). 20. these spheres will intersect WW' in circles as shown in Fig. In order to determine the field at an arbitrary point P due to the disturbances reaching from different portions of the wavefront.2 FRESNEL HALFPERIOD ZONES Let us consider a plane wavefront WW' propagating in the zdirection as shown in Fig. of course. further.u2 + u3 .2. If PO = d. However. The annular region between the nth circle and (n . Optics lowing construction: from the point P we drop: a perpendicular PO on the wavefront. This can easily be seen from the following consideration: For infinitesimal area surrounding a point Qn in the nth half period zone.
Thus we may write ul >u 2 >u 3 . The resultant amplitude at the .. Ref.. (6) where the last term would either i Urn or (z um1 . (6) will be negative. If we further increase a then u(P) would start decreasing and when the circular aperture contains the first two halfperiod zones (which would happen when a = 2 d) the resultant amplitude (= ul . a. where Io represents the intensity at the point P due to the unobstructed wavefront. (4) then it decreases from 1 to 1/2 as m increases from 1 to a n2 . u 3 . we rewrite Eq. Uq + _ I + .Um 2 .. this would happen when a = X . The intensity would therefore be 4Io.3 the brackets are negative.* Thus we may write u„ = constant An Qn P 20. the amplitudes ul. 2. (m odd) (8) ul um_1 . this implies . As a increases.1). It can be shown that if we use the exact expression for rn.2. this obliquity factor comes out automatically from rigorous diffraction theory. In spite of this. ** If one assumes a form of the obliquity factor as given by Eq.2222 (when m is even) Using Eqs (8) and (10) we may approximately write 'u'(P) and u (P) U2 Um u2 uin_1 _ Ul Um 2 + un` (1+cos%) 2 (4) 2 + 2 (when m is odd) (10) where An represents the area of the nth zone. Since the quantities inside * See. (11)** gives the remarkable result that (12) implying that the resultant amplitude produced by the entire wavefront is only one half of the amplitude produced by the first halfperiod zone.Um) according to m being odd or even. (3) in the form u(P)=U1 2 1 2 U3+ 2J(2 u5+ 2) (9) where the last term would now be (? Uin_ 1 + urn ) when m is odd and um when m is even. e. u2.Fresnel Diffraction angle that the normal to the zone makes with the line QP. the intensity at the point P decreases almost to zero. We rewrite Eq. can be increased from zero onwards. (3) can be approximately summed due to a method by Schuster. In order to obtain the upper limits.i can never be smaller that Z However.. If the obliquity factor is such that un > 2 ( un1 + un+1) (7) 20.n in comparison to ul then the Eq. We assume that the radius of the circular aperture. point P would be ul i which is twice the value of the amplitude for the unobstructed wavefront [see Eq. by increasing the hole diameter. (5) The series expressed by Eq..g. that lun. the intensity at the point P would also increase till the circular aperture contains the first halfperiod zone. 2 00. decrease monotonically because of increased obliquity. inter then the quantities inside the brackets in Eq. Thus. consequently u(P) and u(P) Z +2 < 2 +2 < Ul um . this slight increase in the area is exactly compensated by the increased distance of the zone from the point P.1 Diffraction by a Circular Aperture We may use the above analysis to study the diffraction of a plane wave by a circular aperture. slight shift of the point P on the axis will change the amplitude from + 2 to 2 .. 20.u2) would be almost zero.the changes will occur with such great rapidity that one can only observe the average value which will be u. This.. when m is large. the area of the zones increase with n.2 . (m even) where we have assumed that the amplitude of the fields produced by consecutive zones differ only slightly. however. (3) as u (P) = 2 2 + (1 u2 2+2 22 (when m is odd) (when m is even) + 2 2 J + (a  If we can neglect u. we obtain u (P) > ul and ul um u (P) > ul .(12)]. Let the point P be at a distance d from the circular aperture (see Fig.
2).1) will correspond to a maximum.see Sec. the radii of the first.. if up+r a = . Simeon Poisson.. when d = a2 2n  n = 1. Poisson was a great supporter of the corpuscular theory of light and he said that since the bright spot is against common sense.J(2n + 1) Ad . n = 0. 20. . This spot is known as the Poisson spot. Fresnel and Arago carried out the experiment to demonstrate the existence of the Poisson spot [see Fig..3 (a) When a plane wave is incident normally on an opaque disc. .2. the point P will correspond to a minimum. but such a calculation is fairly cumbersome. Whenever the distance a2 d (2n + 1) . 1.. second and third zones would be 0. the screen is 20 m from the coin and the source of light is also 20 m from the coin [photograph adapted from Ref..4 esting result is once again due to the validity of the HuygensFresnel principle and hence would be valid for sound waves also. 0. As a corollary of the above analysis we can consider a circular aperture of a fixed radius a and study the intensity variation along the axis.20.. [maxima] the aperture will contain an odd number of half period zones and the intensity will be maximum.3(b)]. [minima] the aperture will contain an even number of halfperiod zones and the intensity will be minimum..707 mm and 0.up+2 + . n = 0. 20. 2. This is called the `Poisson spot'. used Fresnel's theory to predict a bright spot at the center of the shadow of an opaque disc. Similarly. the famous mathematician. Fig.3 THE ZONEPLATE A beautiful application of the concept of Fresnel halfperiod zones lies in the construction of the zoneplate which consists of a large number of concentric circles whose radii are proportional to the square root of natural numbers and the alternate annular regions of which are blackened (see K.866 mm respectively. 2].3(a)] and if the disc obstructs the first p halfperiod zones then the field at the point P would be u(P) = up+r . 20. The intensity distribution on a screen SS' at offaxis points can be approximately calculated by using the halfperiod zones.4). on the other hand. (b) The Poisson spot at the center of the shadow of a one cent coin. 20. validating the wave theory. (minima) 20. h K.J2n Ad . (maxima) the point P (see Fig. a = . we should always obtain a bright spot on the axis behind a circular disc.. In order to have a numerical appreciation. the wave theory must be wrong. However.2 Diffraction by an Opaque DiscThe Poisson Spot If instead of the circular aperture we have a circular disc [see Fig. we note that for d = 50 cm and A=5 x 105 cm. We may mention here that it was in 1816 that the French physicist Augustin Fresnel developed the mathematical theory of diffraction using the wave theory of light. from the symmetry of the problem..500 mm. 2. 20. one can deduce that the diffraction pattern has to be in the form of concentric circular rings with their centres at the point P.. 20. (b) Fig. (the more rigorous theory also predicts the same result . 2.. 2. (13) 2 Thus. We may generalize the above result by noting that if Optics 20.. n = 1..4. Let the radii of the circles be Ji K. a bright spot is always formed on an axial point. 1. Shortly afterwards..
5A. We consider a point PI which is at a distance K2/? from the zone plate. Between any two consecutive foci there will be dark points on the axis corresponding to which the first circle will contain an even number of halfperiod zones. 20. The zoneplate can also be used for imaging points on the axis. halfperiod zones. thus the resultant amplitude would be (ul .5 the positions of the brightest and darkest points on the axis.. P3 etc.5(b)]... if we have a point source at S then a bright image will be formed at P.SP = ja 2 + r2 + jb2 (15) + r2 . 200/7 cm. Elementary calculations will show that the zoneplate suffers from considerable chromatic aberrations [see Problem 20.SP = producing an intense maximum.. 4`h. Thus. the point P2 will correspond to a minimum.1 cm For 2 = 5 x 105 cm.(a + b) which would again correspond to a maximum.5 Example 20. where K is a constant and has the dimension of length. (b) Fig.5(b)): r2 + u5 + . (b) Imaging of a point object by a zoneplate.. For the brightest point. Thus. (16) P5 P4 P3 P2 (a) implying that corresponding to P2 the first blackened ring contains the 3rd and 4th halfperiod zones.... the even zones are obstructed and the resultant amplitude at PI [see Fig. then SL + LP . 5`h. . 20. 20. K. If the radius of the first circle is r1.5].. we will calculate the positions of various foci.cm A 5x105 The other focal points will be at distances of 200/3.. The most intense focal point will be at a distance Fig.4 The zone plate.01 = 200. for this point the blackened rings correspond to the 2nd.. the aperture should contain only the first zone and thus we must have (see Fig. 3. e. 6th zones..g. 11 th and 12th zones.1 Assume a plane wave (A = 5 x 105 cm) to be incident on a circular aperture ofradius 0.Fresnel Diffraction 20. if a plane wave is incident normally on a zoneplate. from the zoneplate. ..05)2 = 25 cm 2x5x105 Example 20. P4.1) (0.. the second blackened ring contains the 10`h .5 (a)] will be ul + u3 0. 20. Similarly the darkest point would be at a distance (0. ..5 mm.. . We will calculate (a) For a plane wave incident on a zoneplate. etc. the maximum intensity occurs at the points PI. etc. but it would not be as intense as the point PI. 200/5. Obviously. 2 (18) the point L being on the periphery of the first circle of the zone plate [see Fig. etc. For the point P3 (which is at a distance K2/3?) the first blackened ring contains the 4th. the minima occur at P2.. 20. then the corresponding focal points are at distances K2 K2 K2 (17)..u2 + u3) + (u7 .u8 + u9) + . 20. 6`h . (14) SL + LP .2' Consider a zoneplate with radii r„ = 0. where the point P should be such that (see Fig. Between the points PI and P3 there will be a point P2 (at a distance K2/2X) where the resultant amplitude would be (ul u2)+(u5u6)+..05)2 = OP (5 x 105) Thus OP = 50 cm.
20. The zaxis is normal to the plane of the aperture and the screen SS' is assumed to be normal to the zaxis. 20. Using HuygensFresnel principle we showed that the field produced at the point P on a screen SS' (which is at a distance d from the aperture) is given by _ u(P) 4$f efk d drf (21) where the integration is over the area of the aperture.20. It is obvious from the symmetry of the problem that we will obtain circular fringes on the screen SS' however. we will calculate the variation of intensity only along the zaxis. if the amplitude and phase distribution on the plane z = 0 is given by A(.1 Diffraction of a Plane Wave Incident Normally on a Circular Aperture We assume a plane wave incident normally on a circular aperture of radius a as shown in Fig. 20. 19.(18) becomes (19) (20). 20.1 cm) and.(y=71)2]}aT dn (23) a b f + where f = r12/X represents the focal length. it. Now. the coordinates of an arbitrary point M on the aperture will be (p. e. .4 FRESNEL DIFFRACTIONA MORE RIGOROUS APPROACH In Sec.4. Therefore. A very interesting demonstration experiment of the zoneplate can be carried out by using microwave sources (2 . Refs 2 and 3.z expj z[x4)2'+.7 Fig.kr (24) r pdp dO u(P) A J 20. a . Q is an arbitrary point on the periphery of the aperture. * See.6 A plane wave incident normally on an aperture. Further. 71) x i2. e ikz r f JJ A(. (13) of the previous chapter]. Obviously. 20.will be more convenient to use the circular system of coordinates.6 2 Optics = a[1+ 21 a2 +b [l + Zb2 [' 2 1 J (a + b) u(P) f A(. for the sake of mathematical simplicity. the above integral takes the form u(x. it is very difficult to calculate the actual intensity variation on the screen. Equation (20) resembles the lens law. y. 20. i1) err d4 dr (22) tij 2Ca + bJ Thus Eq. z) r. in the Fresnel approximation [see Eq. In this system.. We had considered a plane wave (of amplitude A) incident normally on an aperture as shown in Fig. Thus lA2TCa e .7. r)) then the above integral is modified to y 00 Fig.7] and a small element area dS surrounding the point M will be p dp d¢.g. having aluminum rings on a perspex sheet of dimension 40 cm x 40 cm. instead of the dark rings. where p is the distance of the point M from the centre 0 and 0 is the angle that OM makes with the iaxis [see Fig.6. Diffraction of a plane wave incident normally on circular aperture of radius a.2 we had given a more rigorous analysis of the diffraction of a plane wave by different types of aperture.
the intensity at the point P will be zero and when the aperture contains an odd number of zones.ul (P) 2 d 2d2 or P or a2 Ad = uo (P) . Now. if u l (P) and u2(P) respectively represent the fields at the point P due to a circular aperture and an opaque disc (of the same radius).(26). Equation (27) tells us that the intensity is zero or maximum when p is an even or odd integer. then u i (P) + u2(P) = uo(P) (29) where to is the intensity associated with the incident plane wave.OP= p 2 n (27) 20. for ul (P) we have used Eq. i.5 0.5 (26) a2 where we have defined p by the following equation k [ Fig. 20.7). Thus.uo (P)[l u2 (P) = uo (P) e `pn e'nn ] (30) (28) where.2 Diffraction by a Circular Disc We next consider the diffraction pattern produced by an opaque disc of radius a (see Fig.e. Once again we will assume that the observation point lie on the axis of the disc. The figure shows that when the (circular) aperture contains an even number of halfperiod zones. the intensity at the point P will be maximum. u2 (P) = uo (P) . From Eq..4. if the aperture contains an even number of halfperiod zones. 20. (24) becomes 2n Ja 2 +d 2 u(P) lA J 0 t 2 f d e ll"' dr d^ 7c / . (29) is known as the Babinet's principle.3). the intensity at the point P will be negligibly small and conversely. Equation (21) tells us that in order to calculate the field we have to carry out an integration over the open region of the aperture.(26) we readily get I(P) = 4 lo sin2 QP .8 The intensity variation on an axial point corresponding to a plane wave incident on a circular aperture of radius a./ 0.Fresnel Diffraction 20. when QP . Thus the intensity at the point P on the axis of a circular disc would be 12 (P) = I U2 (P) 12 = lo (P) which is known as the Fresnel number of the aperture. Obviously.25 1.2. The above equation implies 2 where Q is a point on the periphery of the circular aperture (see Fig.2'5 1. 20.2. when d << a (as is usually the case) p = k d 1+ a where uo(P) represents the field in the absence of any aperture.7 4 Now p2+d2 =r2 Thus p dp = r dr 1 to and Eq. 0 11I 0. Thus. the intensity at the point P will be maximum.td 1 1. In Fig.2. 20. .OP is an even or odd multiple of This can be understood physically by using the concept of Fresnel halfperiod zones discussed in Sec. Eq.75 . 20. if the circular aperture contains an odd number of zones. 20.d a2 (31) which gives us the remarkable result that the intensity at a point on the axis of an opaque disc is equal to the intensity at the point in the absence of the disc! This is the Poisson spot discussed in Sec. (25) The integration is very simple and since k = 2 readily obtain u(P) A e tkd (1 e`Pn ) we 0.8 we have plotted the corresponding intensity variation as a function of the dimensionless parameter /.
(38) one can readily show that (1+iy) exp e w2() (33) ff +C 2 I (x. Notice that 0 increases with decrease in wo (smaller the size of the aperture. we obtain [see Appendix C] 2 y2 u (x.e2R(S +y (47) ' From the above equation it follows that a phase variation of the type exp [. the output of a single mode fiber is very nearly Gaussian. rl) = a exp  2 wo (32) 20 = 0.11) u r=[x +y L 2 z4 7c wo (35) (44) . In order to get some numerical values we assume X. z) = 1 + 2 (x2+ y2) w2 (z) (38) +y x 2+.25 mm as the wavelength is decreased from 5000 A to 500 A. = kz + 2 R(z ) (x2 + y2) 2 4 (36) 2 2 +R2] 1/2 1/2 (45) Thus (37) R(z) = z 1 + yz = z [1+ 2 2° z r =R 1+ x2 +y 2 R2 Thus the intensity distribution is given by to exp I (x. This is to be expected.5 pm. the transverse amplitude distribution is Gaussian. indeed as A . 20. = 0. for small values of z. From Eq. for a given value of wo. We define the diffraction angle as tan 8 = w (z) Z 2two (40) showing that the rate of increase in the width is proportional to the wavelength and inversely proportional to the initial width of the beam. where we have assumed I x I.073° and w = 6. In Fig.e. z) dx dy = 2°Io (43) where _ A.25 mm. for wo = 0. Now. I y I << R. 20. as the total energy crossing the entire xy plane will not change with z. (23) and carrying out the integration.88mmatz=l0m Similarly. Further.20. y.10 we have shown the decrease in diffraction divergence for wo = 0. This quantity wo is called the spot size of the beam. the diffraction effects decrease with X.. greater is the diffraction). 2 + 2R (46) y2 which show that the transverse intensity distribution remains Gaussian with the beamwidth increasing with z which essentially implies diffraction divergence. Therefore. Also. k 2R (x2 +y 2 )] (48) w(z) wo Az Az = ltwo Iwo (39) . the width increases quadratiwe obtain cally with z but for large values of z >> wo/). (41) A(4. (32) in Eq. the amplitude falls by a factor 1/e (i. Substituting Eq. y. From Eq.35 mm at z =2 m (42) implying that the phase front is plane at z = O.z y n w2 [ w(z) = wo[l + y 2]v2 = wo1 + 2 2 ] 1/2 (34) which is independent of z. on the plane z = R (see Fig.2 Optics which shows that the width increases linearly with z. for wo = 1mm 28=0. the field distribution is given by 1 eikr r Now. Thus on the plane z = R. 20. We assume a Gaussian beam propagating along the zdirection whose amplitude distribution on the plane z = 0 is given by 2 +. z) (see Fig. As can be seen from Eq.8 20. y. (35).5 GAUSSIAN BEAM PROPAGATION When a laser oscillates in its fundamental transverse mode. the study of the diffraction of a Gaussian beam is of great importance.018° and w=1. the intensity reduces by a factor 1/e 2 ).9). the phase distribution (corresponding to a spherical wave of radius R) would be given by ik 2 2 eikr eikR. 0 * 0 and there is no diffraction which is the geometric optics limit. for a spherical wave diverging from the origin. Then. (32) it follows that at a distance wo from the zaxis.0.
20. = 5000 A 20.5 mm 0. The figure shows the decrease in divergence as the wavelength is decreased from 5000 A to 500 A.5 mm A. 20.7 mm A.76 mm 0. w0= 0.Fresnel Diffraction A.37 mm 0 2 4 z(m) 6 8 10 Fig.25 mm.0073° 11.25 mm 0.9 2 mm 3.9 Diffraction divergence of a Gaussian beam whose phase front is plane at z = O.25 mm. the wavelength is assumed to be 5000 A.10 Diffraction divergence of a Gaussian beam whose phase front is plane at z = 0. = 5000 A 0.5 mm 12. the initial spot size (w 0) is assumed to be 0.7 mm 2 4 6 z (m) 8 10 Fig.073° 12. The figure shows the increase in the diffraction divergence as the initial spot size is decreased from 1 mm to 0. . 500 A 0.
Os: \\\ \\I Fig.13). We will first give a very approximate theory based on Fresnel halfperiod zones.10 (on the x y plane) represents a diverging spherical wave of radius R.11 A spherical wave diverging from the point O. Thus as the beam propagates. We wish to calculate the intensity variation on the screen LL'. is referred to as the waist of the Gaussian beam.12 Diffraction divergence of a Gaussian beam whose phase front is plane at z = 0. (33) we have assumed z to be large. 20. the fringes (wherever they occui.) will be straight lines parallel to the edge. 40 in the prelim pages). see also Fig. (33) does give the correct field distribution even at z = O. It should be mentioned that although in the derivation of Eq. 20.6 DIFFRACTION BY A STRAIGHT EDGE Let us consider a straight edge MN placed perpendicular to the plane of the paper and parallel to a long narrow slit S (see Fig. Thus. Eq.2 d2 Optics R\ =2 (50) R(z) z 1 + wo z2 (49) where d is the distance between the two mirrors. " Fig.12 we have shown a Gaussian beam resonating between two identical spherical mirrors of radius R (see also Fig.11. From the geometry of the arrangement it is obvious that on the screen there will be no intensity variation along the direction parallel to the length of the edge. 20. the phase front which was plane at z = 0 becomes curved.20. which is shown in Fig. In Fig. For this to happen we must have z= R 20. 40 in the prelim pages.13 (a) Diffraction at a straight edge. The dashed curves represent the phase fronts. L 0 z Fig. The dashed curve represents a section of the spherical wavefront at a distance R from the source. 20. (b) ffter. For the beam to resonate. If we compare the above expression with Eqs (42) and (43) we obtain the following approximate expression for the radius of curvature of the phase front: 4 4 1+ 47c wo . . 20. the phase front must have a radius of curvature equal to R on the mirrors.1. this will be followed by a more rigorous analysis. 20. (b) Halfperiod strips of a cylindrical wavefront. the plane z = 0. where the phase front is plane and the beam has the minimum spot size.
. in the absence of the edge).n=i. . the point Q lies on the line joining S and P. 2. By determining the positions of these maxima and minima one can calculate the wavelength..Fresnel Diffraction 20. etc. n=0. Similarly.05 cm from the edge of the shadow and the second and third maxima will occur at distances of 0. MP=[b 2 +x2] 1/2 2 = 2 2. Even then one can draw the following conclusions: (i) Corresponding to the edge of the geometrical shadow (which is shown as E in Fig. For example. . half of the wavefront is obstructed by the edge... hence the amplitude will be given by u(E) = Zuo (52) (51) we will have a minimum and the resultant amplitude will be _ u2 (2 ul 2)+ 4 (57) In general. we draw halfperiod strips in the following manner: Let GQMG' represent a section of the wavefront. The intensity would be a Io.(12)].2.. an arbitrary point P will correspond to maximum intensity if SM + MP . etc.10 cm..SP = (2n + 1) and a minimum if SM+MPSP=2n.SQP = 2 . Now.2.13(a)). the areas of the halfperiod strips will not be equal and thus the analysis becomes quite difficult. The precise variation of the intensity is difficult to calculate from this analysis...11 20.1. 0.. n= 1..e. the positions of the minima are given by [2n+]1/2.1 Analysis Using Half Period Zones In this section we will give a very approximate theory based on Fresnel halfperiod zones. . The halfperiod strips will be on the surface of the cylindrical wavefront as shown in Fig. The points Ql and Q2 on the wavefront are such that SQ1 +Q1PSQP= 2 SQ2 + Q2P . (a+b)+ . 1/2 [b(a+b) Ai =5x10 2 cm a Thus the first maximum will occur at a distance of 6. For a point P1 such that SM+MP1 SP 1 =2 (56) and for the above parameters they will occur at distance of 0. . a more rigorous theory will be given now. The wavefront emanating from the slit is cylindrical and in order to find the amplitude at an arbitrary point P (on the screen). (59) b 1+2 bx2 =b+ 2b 2 SP=[(a+b)2+x2]v2. unlike the Fresnel halfperiod zones.3. (58) n=1.. when x r (2n + 1) b (a + b) a i/2 A. Thus the intensity will be given by I(E) a 2 2(a+b)b x Thus.0866 cm and 0.07 cm.a (61) is the amplitude produced by the first halfwhere period strip in the lower portion and is the resultant amplitude produced by the upper half of the wavefront [see Eq. for a = b = 25 ctn' and A. 20.112 cm respectively. 20.. 2 SM+MPSP = a+b+ 2b (a+b) 2 ( a+b ) where uo'represents the amplitude that would be produced by the unobstructed wavefront (i. (60) = 4 lo (53) (ii) Let us next assume that the point P satisfies the following relation: SM+MPSQP= 2 (54) Thus only the first halfperiod strip of the lower part of the wavefront contributes and the resultant amplitude would approximately be ul ul 24 3u1 4 3 ul 2 2 3 2uo (55) we' Will have a maximum.13(b). = 5 x 105 cm.6. X 2(a+b) 2 Hence.. However. The distance between two consecutive maxima will decrease as we go away from the edge of the geometrical shadow.2.
00419 0.43380 0.1 ru 2 ) du 2 o Since the integrands are even functions of the Fresnel integrals C(i) and S(r) are odd functions of a : C ('r) = C (z) Further.36546 0.2 More Rigorous Analysis of the Straight Edge Diffraction Pattern Before we discuss the straight edge diffraction pattern. .0) + i nu2 du+ifsin(i u2 )du J S (co)] = S(oo).56562 0.6 2.2 4.49843 0.54310 0.1.20.6.42052 0. as can be seen from the figure.0 2.00000 0.0 4.0.8 Fig. for 'r = 1.5 5 eaz2 dx Va (65) we have f eirzu2/2 du = 11 _ J in/2 Now.49919 0.8 3.24934 0.59335 0.45570 0.4 0.6.63629 0.8 2.12 Optics Table 20.34342 0.8 5.5094 0. we get C (.55496 0.14 gives a parametric representation of the Fresnel integrals and is known as the Cornu's spiral.0 00 0. Once again. S(T) = f sin (2 v l / \ // o o Fresnel Integrals: The Fresnel integrals are defined by the following equations: ti C(z) and (63) S('c) = sin (.56320 0.2 3.77989 0.56363 0. C (0) = S (0) = 0 S (T) = S ('r) (67) (68) 0.3) = To summarize. +.39153 0.19992 0.11054 0. C(r) 0.61969 2. For example.49675 0.0 3. Thus.72284 0.6 0.39748 0. there will be =2 f cost 2 = [C (.4 0. 20.8 4. In this section we will make a more rigorous analysis of the diffraction of a plane wave incident normally on a straight edge (see Fig. We now return to the calculation of the straight edge diffraction pattern which we had qualitatively discussed in Sec.44809 0.38894 0.58795 0.46632 0. 5.62340 0.8 1.71353 0.49631 0.2 2.1.77989 and S(r) = 0.48825 0.(66).2 0.49231 0.5 0.00000 0.43849 0.43826.6 3.43826 0. 20.0 0.6 1. introduce the Fresnel integrals.6 4. we C(r) = f cos (2 v2 ldv.54999 0.71544 0.4 4. * Table adapted from Ref.42965 0. The horizontal and the vertical axes represent C(a) and SO respectively and the numbers written on the spiral are the values of 'r.2 1.58110 0.15).03336 0.1 Table of Fresnel Integrals* 20.43833 0.0 1. 20. since 00 =f o cos 12 nu2 (du l ))l (62) and S (T) = S (z) (64) 0.4 1.46227 0. a more detailed table (with greater accuracy) has been given there.33363 0.46749 0.51619 0.56724 0.14 The Cornu's spiral which is a parametric plot of C(T) and S(T).fi e' = (1 + i) (66) exp f i 2 du nu t Figure 20.4 3. using Eq.54172 0.60572 0. C (T) = c ('r ) and The values of the Fresnel integrals for typical values of r are tabulated in Table 20.63889 0.0 . the Fresnel integrals have the following important properties: C (°O) = S (°°) = 2 .
no variation of intensity along the xaxis and. (69) in Eq.Fresnel Diffraction 20. (21). In order to express the above expression in terms of the Fresnel integrals. 20. without any loss of generality. we have replaced r by its minimum value*.Zuo(1+i)[{_ C(vo)}+i { _ s(vo)}] = 1.uo[{2C(vo)}+i r d71 exp ik d + o z + Zd y) (70). we introduce two dimensionless variables u and v such that 2S(v0)} J (76) * This is justified because in carrying out the integration. If the x and y coordinates of an arbitrary point M on the plane of the straight edge are denoted by and n. On substituting the expression for r from Eq. therefore. (72).5). only a small region around the point r = d contributes. where the origin has been assumed to be on the edge of the geometrical shadow. then r =MP= [ 2 +(71y)2 +d2] 1/2 z v2 (73) uo=Ae`m represents the field at the point Pin the absence of the straight edge. thus f Further. we obtain u(P) _ . we obtain u(P) + i f sinll 2 o dv f sin o J = df f d  2 = [ . (70) becomes +2 exp u (P) exp du uo _ 2 f r l (Z ^ ZJ f 2 vo 2 dv (72) where 2 vo_ _^d y and Fig.S(vo)] Substituting in Eq. Eq. d.15 Diffraction of a plane wave incident normally on a straight edge.C(vo)] + i [2 . . 20. y). we may assume the coordinates of an arbitrary point P (on the screen) to be (0. exp i ^ du = 2 [C(oo) + S(oo)] (74) = d 1+ d + +(^ Y) d2 + f vo z exp i 2 dv = f 0 cost 2 v 2 dvJ cos( 50 vo v2) \\\ r ll J V r2 l v2 dv 11 2 (71 Y)2 2d (69) where d is the distance between the straight edge and the screen. where. in the denominator of the integrand.13/ 1 nu z __ k 2 2d = 2 7t 2 and Ad Y)2 2 7rv 2 = d (71 = (1 y ) Thus we may assume u and v to be defined by the following equations: and (71) With these substitutions. the contribution due to faroff points is small because of the rapid oscillations of the exponential term in the integrand (see also footnote in Sec. In ordet to calculate the intensity distribution we use the Fresnel integrals.
the intensity modulation decreases (see Fig.87 with I = 0. 20.16 The intensity variation corresponding to the straight edge diffraction pattern.16). Similarly.74 with I = 0. For such a point vo would tend to 00 [see Eq.08 with I 1. For example. for given values of A.732. In order to determine the field at an arbitrary point P. we have the following interesting property: let us write [C( r2) . for a. (iv) One could have also studied the intensity variation directly from the Cornu's spiral (see Fig. the first three maxima occur at vo = .2.e. (ii) For a given experimental setup.0 as it should indeed be.2. Thus the where we have[4 intensity on the edge of the geometrical shadow is 1/4th of the intensity that would have been in the absence of the edge [see also Eq.15 I0 maxima Similarly..1. (21).16 represents a universal curve. \ 2 Optics +2 J +iI 2+2 \ / (77) Thus. the determination of the positions of maxima and minima is quite straightforward. 20. and d.15) then y = 0 and hence v0 = 0. .S )y (78) If the point P is such that it lies on the edge of the geometrical shadow (i. i. 20.34 with 11.22 with I 1. on the line LL' (see Fig.034 mm respectively.39 with I .e.1. This is due to the fact that associated with the Cornu 's spiral. 1.e. the amplitude at such a point is the same as that in the absence of the edge. we may use Table 20.S('ti)] = A e`9 Thus `[C(z2) C(r )] = A cos9 minima and [S(r2) .16 from which one can make the following observations:. as we go inside the geometrical shadow. (76) would be given by I(P) Fig.(78).1.84310 vo = 3. (i) Figure 20. Thus..872 Io.14). one simply has to calculate v0 as the observation point moves along the yaxis..644 and 2. 20.14 It is of interest to note that a large value of y corresponds to a point which is very far above the edge of the geometrical shadow.06 vo cm Thus the first three maxima will occur at y 0. This also justifies the value of the constant given by Eq.6. For example. as expected.404 and 1.37 I0 v0 . On the other hand. (73)] and we would obtain u(P) = uo = 121 up [.848 mm respectively. 20. thus 1(P) =2 + 4] = 4 1o (79) I0 used the fact that C(0) = S(0) = 0. 0. (iii) As we go inside the geometrical shadow the intensity monotonically decreases to zero. the first three minima will occur at y 1.(53)].C(ai)] + i[S(s2) . The intensity variation is plotted in Fig.vo = 0.and hence vo cc > oo). 1.S('rl)] = A sine (80) = 2 70 [ { _ c(vO }2 + {2 . 2 2 My 4 6 C(vo) = S(vo) giving u(P) . the first three minima occur at v0 = . These results may be compared with those obtained in Sec. when the point P is deep inside the geometrical shadow we obtain (i.20 lo v0 =3.778 lo and vo 2.20.1 to calculate the RHS of Eq.122. The intensity distribution corresponding to Eq. = 6 x 105 cm and d = 120 cm y= . when y . 20..
20. (70) except that the limits of the 7lintegral will be b/2 and +b/2 (we are assuming the origin to be at the centre of the slit) u(P) = 4 (PC)2 I0 As the value of v0 becomes more and more negative. the value of v0 increases. Thus. 20. Consequently vo will be positive.18. In order to use the Cornu's spiral we rewrite Eq. The lines LL' and MM' represent the edges of the geometrical shadow. the length PC keeps on increasing till the point P reaches P1 which corresponds to v0 = . Thus. the corresponding intensity is I(PI ) _ idf f d +b/2 drl exp b/2 k{d+ 42+y)2 2d =4 (2. We wish to calculate the intensity distribution at an arbitrary point P on the screen SS'. 20.7 DIFFRACTION OF A PLANE WAVE BY A LONG NARROW SLIT AND TRANSITION TO THE FRAUNHOFER REGION We next consider a plane wave incident normally on a long narrow slit (of width b) as shown in Fig.17 Computer generated intensity distribution corresponding to the straight edge diffraction pattern. As the value of vo becomes further negative the length PC starts decreasing till it reaches the point P2. y).34.S(z1)] = Thus the length of the line joining the points P and Q will be A and the angle that the line makes with the abscissa will be O. where yi is the angle that QC makes with the abscissa [see Eq. 20. As we move away from the edge of the geometrical shadow to the illuminated region. we obtain . The intensity is again given by I(P) 20.15 Let the points P and Q on the Cornu's spiral (see Fig. Let the point Q on the spiral (see Fig. Hence in the shadow region the intensity uniformly decreases to zero (see Figs 20.34) 2 10 1.Fresnel Diffraction 20.37 lo (82) Carrying out manipulations similar to that in the previous section. 20.14) correspond to = vo. Thus the point Q keeps on moving on the spiral towards the point C and the length QC decreases uniformly. there will be no variation of the intensity along the xaxis and we may (without any loss of generality) assume the coordinates of the point P to be (0.14. we have 2 0 4 2 C(vo)} + i 2 S (vo)} = (QC) {1 e 1' Fig.17).(80)]. Since the point C in the curve corresponds to z = co. Once again. The intensity at this point is maximum and the length P1 C= 2. the value of v0 becomes negative and the corresponding point P (on the Cornu's spiral) lies in the third quadrant as shown in Fig.22. Thus.16 and 20. u(Q) = 12 a (Q C) elf uo or 1(Q) = (QC) 2 10 (81) We can easily see that as the point of observation moves into the shadow region. the intensity keeps on oscillating with decreasing amplitude about 10 as we move more and more into the illuminated region (see Figs 20.1. It is obvious that PM = [C(72) C(71 )] = A cose A sine and QM = [S(72) .14) correspond to z= 71 and z= Z2 respectively. (76): u = 12Z uo L{ 2  C(vo)} +1 j 2 S (vo)}] Let us first consider a point of observation Q in the geometrical shadow region. The field at the point P will again be given by Eq.16 and 20.17).
as y varies on the screen.16 Optics where we have used Eq.19 The intensity distribution produced by diffraction of a plane wave by a long narrow slit corresponding to vl = 0. when the slit width is very large) the Fig.5 and 5. 1.C (v2 . (64). u(P) 2 f uo u 2 expZ (v2v1) 2 2 J du v f expl r l Z 2 Jdv where 1I Ti ( v2 +v 1) 2 A. (86). (86).. = 5 x 10 5 cm. for A. one obtains vl = 2.0. 20. 20.g. 20. In Figs 20. d and b are known which determine v1.d b 2' v? __ 2 Ad Y Using Eq. e.21 and 20.v )} + {S(v2 +v )S(v2 v )} 2] 1 1 1 (84) For a given system A. and vl __ 2 ^.v )}] l 1 Fig. The dashed curves correspond to Eq. Thus the intensity distribution would be I(P) = 2 2 lo [{C (v2 + vl) .20.0.5..19.e. One can see that for a large value of vl (i. 1.5. the quantity v2 also changes. d = 100 cm and b = 0.d (rl .Y) Fig.22 we have plotted the intensity variation as a function of v2 for vl = 0.1 cm. 20.20 )} (83) The intensity distribution produced by diffraction of a plane wave by a long narrow slit corresponding to vl = 1. The dashed curves correspond to Eq.C (v2 .0.0 respectively. further.18 Diffraction of a plane wave incident normally on a long narrow slit.20..V + i {S (v2 "S (v2 . (66) we obtain u(P) =2 uo( +i)x (v2V1 ` [cos(. .v 2 )+isin(v 2 )]dv f o (v2+v1) f 0 [cos(v 2 )+isin1v 2 )] V2 or u(P) = (12 l) uo [{C (v2 + v1) + vl) . 20.
S(v2 2 cos 7CV2 2 + v1) sin 7r v l vz where 0 represents the angle of diffraction (see Fig. when the observation screen is far away from the aperture) the diffraction pattern is essentially of the Fraunhofer type.5. 20. (86).AB (85) Similarly.22 The intensity distribution produced by diffraction of a plane wave by a long narrow slit corresponding to vl = 5. Similarly S(v) Since C(v2 v2 1 = 2cos( 7rV v2 2 diffraction pattern is similar to that produced by two straight edges.1 sin ( 2 v2) 7r v +f 1' 2 7r v sin 2 2 dv ^v J 2+7r sinl 2v z v 8 6 2 J .vl) L 2 + ^1 sin2(v 2 +v l ) 2 z 7r V2 .f cos2v z dv 0 v t f 1.0 expressions of the Fresnel integrals in the limit of v 3 Now.21 The intensity distribution produced by diffraction of a plane wave by a long narrow slit corresponding to v 1 =1. The dashed curves correspond to Eq. This is indeed what we should have also expected. 2 1 10 v1 = 5. for small values of vl (i. Fig. the value of v2 will also be very large and thus we must look for + vl) S(vz  V ^v z 2 (v2 + vi ) sin (7CV l v2) . Clearly.. d is very large. we may write C(v) = f cos 2 vz dv 0 v = f cos2v z dv.[2 + sin 2 (v2 (v2  vl) 21 y VA .23). 20.e.Fresnel Diffraction 20. On the other hand. 20. 20.17 y d Fig. where we have neglected terms which would be of order 1/v3.^Ad 2dy_ I2d is large and vl is small. In order to show this explicitly we notice that _ 2 vz.0. 1I /II 2 4 = 2f v 1 7rV Iicvdv 2 . V2 Fig. we have + vl) C(v2 .23 In the Fraunhofer region. in the Fraunhofer region since d is very large.
39 (1 = 0. ♦ For a plane wave incident normally on a circular aperture of radius a.87 (1 = 0. For large values of z: . and minima at v0 1. . We define the diffraction angle as tan 0 = w(z) = A 7cw0 z showing that the rate of increase in width is proportional to the wavelength and inversely proportional to the initial width of the beam. then we would always obtain a bright spot on the axis behind the disc..2. • If instead of the circular aperture we have opaque disc.... in the Fraunhofer limit.. this is called the 'Poisson spot'.7. Eq.15 I0) . .3.22 (1= 1. . the variation of the spot size is given by A2z2 ]Y2 w(z) = w 1 + 7c g are known as Fresnel integrals.19 . we have minimum intensity and the circular aperture will contain even number of half period zones. ♦ For a plane wave incident normally on a straight edge. ♦ For a plane wave incident normally on a long narrow slit of width b. ♦ For a Gaussian beam (whose phase front is plane at z = 0). the intensity variation on an axial point P is given by y being the distance from the edge of the geometrical shadow and c(x) = I cos(Z7cu 2 )du 0 x and s(x) = J sin (2 7ru 2 ) du 0 rx 1 where =10 sing p2 P a2 Ad .74 (1= 0. this is characteristic of diffraction.. we have maximum intensity and the circular aperture will contain (with respect to the point P) odd number of Fresnel half period zones and when p = 2. The intensity monotonically goes to zero as we go deep inside the geometrical shadow.20 I0). The quantity p is known as the Fresnel number of the aperture.(86) and one can see that the intensity distribution is almost of the Fraunhofer type for vl 5 0. 0L wp .3.37 I0 ). .v l )} 2] where 2 vl _ 2 b '11 Ad 2' vg _ 11 Ad y and y is the distance from the midpoint of the edges of the geometrical shadow.20.5.3.X is the wave length and d is the distance of the point P from the center of the circular aperture.5.87 l0). Equation (86) shows that the intensity distribution is indeed of the Fraunhofer type (see Sec.6.2). As vl becomes large. we obtain the intensity distribution corresponding to two straight edges and for vl + 0 we get the Fraunhofer diffraction pattern.20.8 .84 l0). When p = 1. In Figs 20.w(z) = =2 =100 Io v2 p2 sing (^c vl V2) Az 7cw0 J (86) sing /3 where Ioo which shows that the width increases linearly with z. J + j 2 . As we move away from the edge of the geometrical shadow to the illuminated region.34 (1 = 1..22 the dashed curves correspond to Eq.2.4. .= 1. The corresponding radius of curvature of the wavefront is given by R(z) = Z =2lov 2l 7rb and /3=IC v 1 v2 = ^ by d and e^d (88) 6 (87) r l + 7 2w4] L represents the diffraction angle.78 I0 ). 18. one obtains maxima at v0 . the intensity variation on a screen (at a distance d from the slit) is given by I= 2 Io [{C(v2+v))C(v2vl)}2 + {S(v2 + v1)  S(v2 .. v0 = dy 2 on a wavefront is a source of secondary disturbance and the secondary wavelets emanating from different points mutually interfere.18 Thus. the intensity variation on a screen (at a distance d from the' straight edge) is given by 2 1= where 10 2l0[{2  Qv. .S(vo)} J 2 Summary ♦ The underlying principle in the theory of diffraction is the HuygensFresnel principle according to which: Each point is the intensity in the absence of the straight edge. (84) becomes I(P) where w0 Optics is the spot size at z = O.08 (I = 1.
. d = 0. 0. .5 If a zoneplate has to have a principle focal length of 50 cm corresponding to A= 6 x 105 cm. 0..027 cm] 20. at y = ± 0. (103) must be al and a2] 20. [Ans: the first two maxima occur at y = 0.22). 20.356. Calculate the wavelength. (maxima).10 In a straight edge diffraction pattern.13 In Sec.0.0 (see Figs 20. .. = 5 x 10 5 cm? [ ( 0.. 0. The most intense point on the axis is at a distance of 200 cm from the aperture.01] 20.2 What would happen if the circular aperture in Problem 20. zones were blackened? 20.. cm] [{+_c(vo)}+i{'. Obtain the diffraction pattern of an annular aperture bounded by circles of radii al and a2 (> a l).7 (a) A plane wave is incident normally on a straight edge (see Fig. calculate the distance between the slit and the screen so that the value of vl would be 0. ±0.zones are blackened. at y = ±0. minima will occur at y = 0.. (b) Minimum intensity will occur when a = 0.1.. = 4480 Al 20. Find approximately the distance in centimeters inside the geometrical shadow where I/10 = 0.01685].. 55.(Minima)] 20.11 Consider a plane wave falling normally on a narrow slit of width 0. Assume A = 5 x 105 cm.5.05.. [Ans: (i) (a) z = 166.1 Consider a plane wave of wavelength6 x 105 cm incident normally on a circular aperture of radius 0. obtain an expression for the radii of different zones. (b) Assume A0 = 5000 A and d=100 cm.3 mm.15.2 mm x .1061 cm.5 mm. 0..23).7 cm.0473 cm and 0.Fresnel Diffraction Problems 20.0775 cm. along the yaxis.110 cm. [Ans. 60 cm] 20. Obtain the positions of the first few maxima and minima in the Fraunhofer diffraction pattern along directions parallel to the length and breadth of the rectangle._s(vo)}1 ^Q p •o Straight edge •R Fig.6 cm.4 Consider a circular aperture of diameter 2 mm illuminated by a plane wave.067.01 cm. what would happen if instead the 1 st.60. Using Table 20. [Ans: (b) I/10 = 1. 3` d ..1 is replaced by a circular disc of the same radius? 20.0.1/10 = 1. What would be its principle focal length for A.1.14 Consider a rectangular aperture of dimensions 0. Calculate the wavelength of light. P (y = 0.3 (i) A plane wave (A. 5`h etc. I/10 = 0.05 cm. 20.] 20.05 cm. Q (y = 1 mm) and R (y = 1 mm) where 0 is at the edge of the geometrical shadow.1. [Hint: The integration limits of p in Eq.83 cm.26.. . 20. one observes that the most intense maximum occurs at a distance of 1 mm from the edge of the geometrical shadow.9 In a straight edge diffraction pattern. [Ans: d = 1. sixth. The first two minima occur at y = 0. If the wavelength of light is 6 x 105 cm. if the wavelength of the light used is 6000 A and if the distance between the screen and the straight edge is 100 cm. = 5 x 105 cm and that the diffraction pattern is produced at the focal plane of a lens of focal length 20 cm..1920. if the distance between the screen and the straight edge is 300 cm. Assume A. calculate the distance between the most intense maximum and the next maximum. [Ans: Along the xaxis.24 . (ii) Repeat the calculations for A. . Calculate the values of z (on the axis) for which maximum intensity will occur. [Ans. Plot the intensity as a function of z and interpret physically. 33. the second.1.42 cm.33 cm. = 5 x 105 cm and discuss chromatic aberration of a zone plate. = 6 x 105 cm) is incident normally on a circular aperture of radius a.1 cm) at y = 0.10. 0.0906 cm. 0. (Maxima). 0.1 cm.01 cm.3 cm. 20.5 and 5.] 20.5 mm). [Ans: y = 0.24. Plot the intensity variation as a function of a and interpret physically. 1. 0. Using Table 20. 5 x 105 cm] 20. write approximately the values of I/IO at the points 0.12 Consider the Fresnel diffraction pattern produced by a plane wave incident normally on a slit of width b.6 In a zoneplate. [Ans: At y = 0. 0. (a) Assume a = 1 mm. Calculate the positions of the first two maxima and minima on a screen at a distance of 50 cm from the edge.. Discuss the transition to the Fraunhofer region.0724 cm and 0. Repeat the analysis for b = D cm. ±0.134 cm. fourth. Calculate the values of a for which minimum intensity will occur on the axial point. cm. 1... (b) Assume z = 50 cm. 0. 1/10 = 0. minima will occur at x = 0. 0. Show that the field at an arbitrary point P is given by u(P) = 121uo Ad y. .3 n) mm.56 cm. 20.8 Consider a straight edge being illuminated by a parallel beam of light with A = 6 x 105 cm.19 where vo = v0 20. d = 100 cm. Calculate the positions of the brightest and the darkest points on the axis. approximately calculate the intensity values (for b = 0.9 we obtained the diffraction pattern of a circular aperture of radius a.033.67 cm.
R. Contemporary Optics.20 20. calculate the area of the patch (on focal plane) which will contain 95% of the total energy. National Bureau of Standards. Fig. A. Principles of Optics. Born and E. New York.83 cm and 4. 20.2 mm.A. Washington (1964). 4. Cambridge University Press. = 6328 A) can be assumed to be Gaussian with plane phase front. 0. . Wolf.25 mm] 20. 20. 2000.27 REFERENCES AND SUGGESTED READINGS 1.15.17 (a) The output of a HeNe laser (A. For wo = 1 mm and wo = 0. calculate the beam diameter atz=20m. Thyagarajan. Baierlein. [Ans: (a) 0. = 5. R = 100 cm and the distance between the 2 mirrors to be 50 cm. Vol.5 µm and wL = 2 µm.15 The Fraunhofer diffraction pattern of a circular aperture (of radius 0. [Ans: 2w = 0. 1992. `Large scale diffraction patterns from circular objects'. 2. (25)] Fig.18 A Gaussian beam is coming out of a laser.25. [Ans: = 5. Abramowitz and I. (b) Repeat the calculation for A= 5000 A and interpret the results physically. Rinard. 44.20. [Ans: 0.27).22 The output of a semiconductor laser can be approximately described by a Gaussian function with two different widths along the transverse (wT) and lateral (WL) directions as I 1ll b3A P b wL x2 WT y2 where x and y represent axes parallel and perpendicular to the junction plane. 70. Discuss the far field of this beam (see Fig.5 mm) is observed on the focal plane of a convex lens of focal length 20 cm. Plenum Press. Handbook of Mathematical Functions with Formulas. Calculate the spot size of the Gaussian beam. What will be the intensity on the axial point P? [Hint: You may use Eq. M. 20. = 1 µm.13 mm.0 cm] 20. Ghatak and K. American Journal of Physics.77 cm. Vol.19 A plane wave of intensity to is incident normally on a circular aperture as shown in Fig. Stegun. 1978. 5. [Ans: wo = 0.4 mm] R 50 cm Fig. Newton to Einstein: The Trail of light. Assume A. After traversing 10 m through vacuum what will be (a) the beam width and (b) the radius of curvature of the phase front. Applied Mathematics Series.16 In Problem 20. R(z) = 1017 cm] Optics 20. the beam width is 1 mm and the phase front is plane.25 20.55 x 10' 3 cm2] 20. 3. Typically wr = 0. M. Calculate the radii of the first and the second dark rings.26 20. Assume A.26). 20.21 Consider a resonator consisting of a plane mirror and a concave mirror of radius of curvature R (see Fig. 20. 20.20 Show that a phase variation of the type yz) r expl ikz+ k(22 R(z) ] represents a diverging spherical wave of radius R. Assume A = 6000 A and that at z = 0. Graphs and Mathematical Tables. Cambridge University Press. 55.5 x 105 cm. 1976. 20. M. P.
1948 .The electron microscope was to produce the interference figure between the object beam and the coherent background.1 INTRODUCTION A photograph represents a twodimensional recording of a threedimensional scene. the image produced by the technique of holography has a true threedimensional form. parallax) of the object scene is lost. These holograms are called Lippmann . Benton invented `Rainbow Holography' for display of holograms in white light. one can change one's position and view a different perspective of the image or one can focus at different distances. . 1971 Important Milestones 1948 1960 1962 1962 1964 1969 Dennis Gabor discovered the principle of holography The first successful operation of a laser device by Theodore Maiman Offaxis technique of holography by Leith and Upatnieks Denisyuk suggested the idea of threedimensional holograms based on thick photoemulsion layers. the threedimensional character (e. 1]. / Leith and Upatnieks pointed out that a multicolour image can be produced by a hologram recorded with three suitably chosen wavelengths. ** Dennis Gabor received the 1971 Nobel Prize in Physics for discovering the principles of holography. because it contained the whole information. in which one not only records the amplitude but also the phase of the light wave. This interference pattern I called a hologram. Dennis Gabor in his Nobel lecture** December 11. The basic technique in holography is the following: In the recording of the hologram. that is to say the nondiffracted part of the illuminating beam. The hologram was then reconstructed with light.g. Gabor's Nobel lecture entitled Holography. Holography is a method invented by Dennis Gabor in 1947. His holograms can be reconstructed in ordinary sun light.from the Greek word holos . 2. This was a vital step to make holography suitable for display applications. Thus one cannot change the perspective of the image in the photograph by viewing it from a different angle or one cannot refocus any unfocussed part of the image in the photograph.the whole. What is recorded is the intensity distribution that prevailed at the plane of the photograph when it was exposed. Since only the intensity pattern has been recorded. as with the object.1971 is nonmathematical and full of beautiful illustrations. The capability to produce images as true as the object itself is what is responsible for the wide popularity gained by holography.. 21. it is reprinted in Ref. Thyagarajan. this is done by using interferometric techniques. Thus.Bragg holograms. in an optical system which corrected the aberrations of the electron optics. the original paper of Gabor appeared in 1948 [see Ref. the phase distribution which prevailed at the plane of the photograph is lost. one superimposes on the object wave another wave called the reference wave and the photo * A portion of this chapter is based on the unpublished lecture notes of Professor K. Because of this. The light sensitive medium is sensitive only to the intensity variations and hence while recording a photograph.
S. The real image can be photographed without the aid of lenses just by placing a light sensitive medium at the position where the real image is formed. 21. in general.) .3 (a) An ordinary photograph of an object. Thus one can move the position of the eye and look behind the objects or one can focus at different distances.2).21. (b) The hologram of the object produced by a method similar to the one as shown in Fig. (c) The reconstructed image as seen by an observer. its hologram and the reconstructed image respectively.1). etc. (Photographs courtesy: Professor R. Photographic plate Reconstruction wave Fig.2 graphic plate is made to record the resulting interference pattern (see Fig. we again illuminate the hologram with another wave. (d) A magnified view of a small portion of the hologram shown in (b). 21. information about the object is coded into the hologram. The virtual image has all the characteristics of the object like parallax. To view the image. this process is termed x o>z Y Optics as reconstruction (see Fig.3(a). Figures 21. The reconstruction process leads.2 Reconstruction process. a hologram has little resemblance with the object. in fact. to a virtual and a real image of the object scene. The reference wave is usually a plane wave. (b) and (c) represent the object. 21. (c) (d) Fig.1. This recorded interference pattern forms the hologram and (as will be shown) contains information not only about the amplitude but also about the phase of the object wave. Unlike a photograph. 21. Sirohi. called the reconstruction wave (which in most cases is identical to the reference wave used during the formation of the hologram). 21.
3). y). We consider the case when the reconstruction wave is identical to the reference wave r(x. then one can see that r(x. y) . y)) . Thus. y) (cos2 {¢(x. y) (see Fig. y) cos {0 (x. then the transmitted field would be given by v(x.2 THEORY If the object is a point scatterer. of the point of observation from the point scatterer and.cot] (1) represents the object wave (which. y. The transmittance of the hologram. y) .cot] cos (27r a x .cot] (4) The photographic plate responds only to the intensity which would be proportional to the time average of [u(x.y) = 2 a2(x.rcot] =A =2 (cos [0(x. y) R(x.cot}) +A2 (cos2 (27r ax . in such a case if R(x. t)) = ( [a(x.. Notice that there is no ydependence because the plane wave has been assumed to have its propagation vector in the xzplane.2). When the photographic plate (which has recorded the above intensity pattern) is developed. y) = A cos [kx sin 9 . y) . y) + 2 AZ] A cos (27r ax_ cot) . 21.cot)) + 2a(x.e. y) =K [I a2 (x. t) =. one obtains a hologram [see Figs 21. A represents a constant. (6) becomes 2 A2 (9) + Aa(x. y) I (x. Any general object can be thought of as being made up of a large number of points and the composite wave reflected by the object would be vectorial sum of these.9). depends on I(x. y) . as mentioned earlier. 17. By a suitable developing process one can obtain a condition under which the amplitude transmittance would be linearly related to I(x. a(x. the field associated with this plane wave would be given by r(x. In such a case we would obtain (omitting the constant K) v(x. the phase distribution associated with it.cot]) = and (cos [0(x.1). y)+ 2 A' R(x. y. 21. y) = (u2 (x.2 co t]) + 2 I (x.2 7r a x] (2) cos (kx sin 9 + kz cos 9 . Thus I(x.. y)+ (cos [0(x. y) A (cos { 0(x. y) .y. y) = K Rr (x. t)] 2. 21. which is contained in 0 (x.27rax] 2 Eq.cot) If r(x.cot + 0) where r represents the distance . The fundamental problem in holography is the recording of this object wave.cot)] 2) (5) From the above relation it is obvious that the phase information of the object wave. y) represents the field at the plane z = 0 due to this reference wave.cot] + A cos [2irax .3 where the angular brackets denote time averaging (see Sec. 19. y) = a2(x. the intensity pattern recorded by the photographic plate would be I(x. y) cos [0(x.z) = Acos [k. Thus.1). k = 21t/). Thus the total field at the photographic plate (which is coincident with the plane z = 0) would be given by u(x. y) . that it is propagating in the xzplane inclined at an angle 9 with the zdirection (see Fig. is due to the superposition of waves from point scatterers on the object) in the plane of the photographic plate which is assumed to be z = 0 (see Fig.cot)) (6) Since (cos2 [0(x. Let us consider the recording process. y) . then the object wave would just be I cos (kr .cot] (3) where a= sin is the spatial frequency (see Sec.cot] = A cos [2. is recorded in the intensity pattern. y) + KA a(x. Let O(x. y) 11 22r ax] (10) where K is a constant. y) cos [0(x.3(b) and (d)]. y) . y).cot} cos (27c ax . for simplicity.cot)) 1 = (cos2 (27rax . y) = L 2 a2 (x.2tc ax]) (8) 1 cos [¢ (x. y). We consider a plane reference wave and assume.Holography* 21. y) represents the field of the reconstruction wave at the hologram plane. The above equation represents the field due to a plane wave inclined at an angle 9 with the zaxis and as can be seen the phase varies linearly with x.rax . y. y) + 27r ax . y) cos [0(x. i. the ratio of the transmitted field to the incident field. Thus. in particular. y) = a(x. y) cos [0(x.cot) + A cos(27cax . y) .cot)) 2 (7) 21.
4 + A2 a(x. We consider each of the three terms separately. y) . For such a wave 0(x. Thus the effect of viewing this wave is the same as viewing the object itself. To determine the effect of the term 4irax. But the resolution in the image decreases as the size of the fragment decreases.cat] A2 a(x. . Thus if the object wave is a diverging spherical wave then the last term represents a converging spherical wave. for example. The reconstructed object wave is traveling in the same direction as the original object wave.. each point of the object illuminates the complete hologram and consequently each point in the hologram receives waves Al cos [kxsin 91 +kzcos 01 wt] (b) Fig. which forms a real image of the object. y). then show that the transmitted light consists of a zeroorder plane wave and two firstorder plane waves. Thus in contrast to the second term. y) . (see also Fig. y) = 0 and the last term would represent a plane wave propagating along a direction 8' = sin l (2 sin 0). Solution: (a) Consider a plane wave withits propagation vector lying in the xzplane and making an angle 91 with the zaxis. (b) Reconstruction of the hologram with another plane wave.11) (b) If we reconstruct the hologram with another plane wave [see Fig. This part of the total field is traveling in the direction of the reconstructed wave. (1) and hence represents the original object wave.21.4(a)]a plane object wave corresponds to a single object point lying far away from the hologram. y) carries a negative sign. y) cos [4n ax . y) cos [O (x. each separate fragment is capable of producing a complete virtual image of the object.1 As an explicit example of the formation and reconstruction of a hologram. y) cos [2nax . Thus the effect of the term 4irax is to rotate the direction of the wave.$ (x. when both the object wave and the reference wave are plane waves.4(b)].* This property can be understood from the fact that for a diffusely reflecting object. the hologram consists of a series of Young's interference fringes having an intensity distribution of the cos t type. this wave forms a real image of the object which can be photographed by simply placing a film (see Fig. this gives rise to a virtual image. the phase term cp(x. when a hologram of a transparency is to be recorded. 14. 21. . A very interesting property possessed by holograms is that even if the hologram is broken up into different fragments. There are cases where this does not hold good. For such a wave.2). The second term is identical (within a constant term) to the RHS of Eq.4 (a) Formation of a hologram.2nax] Acos(2naxcot) =C Optics from the complete object. we consider the case when the object wave is also a plane wave traveling along the zaxis. we consider the simple case when both the object wave and the reference wave are plane waves [see Fig. 21. the two firstorder waves correspond to the primary and conjugate waves. the field is of the form + 2 + A2 a(x.wt] (11) 2 Equation (11) gives the transmitted field in the plane z = O. y). For nondiffusely reflecting objects or for transparencies. Hence the last term on the RHS of Eq. 21.The negative sign represents the fact that the wave has a curvature opposite to that of the object wave.wt] cos [cp(x. (11) represents the conjugate of the object wave propagating along a direction different from that of the reconstruction wave and the object wave. The first term is nothing but the reconstruction wave itself whose amplitude is modulated due to the presence of the term a 2(x. Example 21. Since the waves represented by the three terms are propagating along different directions they separate after traversing a distance and enable the observer to view the virtual image without any disturbance. one makes use of an additional diffusing screen through which the object is illuminated. 21. * This property of a hologram exists only when the object is a diffuse scatterer such that the wave from each scattering point of the object reaches all parts of the hologram plate. (a) Show that for such a case. To study the last term we first observe that in addition to the term 4nax.
21.sin 6. the above expression simplifies to (sin 012A2 cos2sin 62 )] showing that the intensity remains constant along lines parallel to the yaxis with fringe spacing depending on the values of 0. Thus +b/2 E= 2 f (1 + cos 2 as) x b/2 [sin (k r .sin 9) .5).cos (kr . Thus = 2 A sin (kr .s sin 0) .5. Thus the total field in the direction 0 would be given by +b/2 E=A where )3 = 1 kb sin 0 = tcb sin 0 which is of the same form as obtained in Sec.cot) sin (ks sin 0)] ds sin 62 .ks (sin 0 .sin 02)] = 2 (Al .A2 )2 + 2A1 A2 cos 2 k2 sin 12 sin 0 (sin 01 . (b) Before we calculate the transmitted field of the hologram. 21.cot) sin {lc s (sin 0 . +b/2  f sin [k(r . 14. will be given by A2 cos [kx = A J [sin (kr . the intensity distribution is of the cos 2 type (cf. Fig. and 0 is defined in Fig.sin 6. 21. and 02.sin 6i + 2 a) } ds b/2 J sin Fig.cot] ds b12 (12) cos {ks (sin 6 .cot]ds b/2 where 6.sin OA] ] ds +b/2 = ZAsin(kr .cos (k r . represents the angle of incidence of the illuminating plane wave.Holography* 21.cot}] 2) = 2AI `+ 2A2 +Ar A2 cos [kx (sin 0l . In the present case.s sin 6) . The above integral can also be written as +b/2 E If the photographic film is assumed to coincide with the plane z = 0. we first consider a narrow slit of width b being illuminated by a plane wave (see Fig. .cot) cos (ks sin 0) b/2 .) } ds b/2 +b/2 + 2 J cos tics (sin e.cot} + A2 cos {k x sin 02 .sin 62 ) = Ab sin (kr . Further.2.cot) J cos {ks (sin 6 sin 0. + 2a)] (16) 2 (sin 0sin6 1 +2a) . for example.cot] / ksin0 where the second integral is zero because the integrand is an odd function of s. here k = 21r/A. Consider an element ds at a distances from the center of the slit.cot] ds. [b 2 (sin 0 .) } . Thus.11). the hologram has a cos 2as type of variation in transmittance and hence the transmitted field will be of the form +b /2 r (14) E =A J cos2 as sin [kr .cot) cos {lc s (sin 0 .cot] Similarly the field (on the plane of the film) due to a plane wave making an angle 02 with the zaxis.5 A plane wave incident on a narrow slit of width b. then the field distribution on this plane would be given by Al cos [kx sin 01 .sin 0) + 2a)) ds b/2 +b/2 (15) + 2 f cos {ks (sin 6 . Then the amplitude at a far away point P due to this element would be proportional to sin [k (r . 18.cot) E The resultant intensity distribution would be proportional to ( [AI cos {k x sin 01 .5 where A is a constant.cot) sin /3 13 (13) For At = A2.2a)) ds b/2 The above integrations can easily be carried out.sin 6.
6 which becomes more and more sharply peaked around sin e = sin 0i . 14. Let the point source be situated at a distance d from the photographic plate. z = 0. as the size of the hologram becomes larger.co. I Optics The resultant fringe pattern is circular and centered at the origin (see Example 14. The hologram thus formed is essentially a zone plate with the transmittance varying sinusoidally in contrast to the Fresnel zone plate [see Fig.1222 x 10 4 mm apart. y) = A2 2 2d2 +B + AB cos kd+ 2d (x2 + y2) (24) * See. For simplicity assume the reference wave to fall norrrially on the photographic plate.21. . 14. which represent the zeroorder and two first order waves.222 x 10 3 mm. Thus the photographic plate should be able to record fringes as close as 0. the total field at the plane of the photographic plate would be T(x.2 Consider the formation of a hologram with a point object and a plane reference wave (see Fig.13(a)). due to waves emanating from the point object would be given by O(x.. Obtain the interference pattern recorded by the hologram.g.31.7). the film. If the reconstruction source is of the same wavelength and is situated at the same relative position with respect to the hologram as the reference source. = 6328 A (HeNe laser). t) = 0(x. the object and any mirrors used in producing the reference beam must be motionless with respect to one another during exposure. z = 0. Another critical requirement in making holograms is stability of the recording arrangement. Hence if the resolution in the reconstructed image has to be good. one obtains d = 1. 0) on the photographic plate.2a as b . Some of the holographic materials are 649F Kodak or 10E 75 or 8E 75 AgfaGaevert films and plates. thus taking the stability requirements even further. z.depth finds application in studying transient microscopic Thus I(x.cot) + B cos cot r (20) The recorded intensity pattern would be I(x. y.wt) (18) Hence. the maximum path difference between the object wave and the reference wave should not exceed the coherence length. One more requirement which is not so obvious (but is a necessity) is the resolution of the film. During reconstruction. Choose the zaxis to be along the normal from the point source to the plane of the photograph. Refs 312. the source must not be. angular brackets denote time averaging.sin 9i .3 REQUIREMENTS Since holography is essentially an interference phenomenon. t) = B cos wt Thus. (15) in the limit of a large value of b give rise to three plane waves propagating along sin 0 .* The ability to record information about the .('(x.2a and sin 0 = sin 0 + 2a. t) + R (x. z = 0. sin 8 = sin 0i . Thus. we can write +8 2 r=(x2 +y2 +dl) 'h =d+ x 2dy (23) 21. the spatial coherence is important so that the waves scattered from different regions of the object could interfere with the reference beam. It may be worthwhile mentioning here that the reconstruction process has associated with it aberrations similar to that in the images formed by lenses. Thus. if stable interference fringes are to be formed (so that they are recordable). assumed to be coincident with the plane z = 0. thus the spatial frequency is 818 limes/mm. y.4 SOME APPLICATIONS The principle of holography finds applications in many diverse fields. Thus the three integrals in Eq. produce an interference pattern with spacing d = 2 s a Assuming 0 = 15° and 2. A plane wave traveling along a direction . parallel to the zaxis would be given by R (x. the field due to the reference wave at the plane of the photographic plate (z = 0) would be (19) R (x. This requires special kinds of material which tend to be exceedingly slow. y) 2 r If we assume that d >> x. we get =A2 2 (22) + AB coskr I(x.. i. t) = A cos (kr . 21. y. Solution. as before. then the reconstructed image does not suffer from any aberrations. t) 12 ) _ ( 4 cos(kr _ wt)+Bcoso)t r (21) 2) where. broad and must be emitting a narrow band of wavelengths. y. 20. t) = B cos (kz . e. y.13(b) and Sec. y) = ( I T (x. y (which is valid in most practical cases). The field at any point . z = 0. y. the reconstructed image depends both on the wavelength and the position of the reconstructing source.e. Example 21. y. t) = Ar cos (k r wt) (17) where r = (x2 + y2 + d2) is and A represents a constant. Further. Carrying out the above time averaging. Two plane waves making angles +0 and 0 with the axis. In Chapter 17 we had introduced the notion of coherence length. y. certain coherence requirements have to be met with.
they Fig.Holography* 21.6(b)] then we may write O'(x. y) .0(x.wt] ' (26) O'(x. L is the length of the bar. y) _ (2m + 1) Z . y)] (27) where W is the load. y) + A2(x. The photographic plate after development forms the hologram. the object is stressed and the.g.wt] (25) represents the object wave (in the hologram plane) when the object is unstressed [see Fig. y. . y) cos{0' (x. y) . the object superimposed with bright and dark fringes (see Fig. To understand the formation of the fringe pattern.. If we have a bar fixed at one end and loaded at the other and if it results in a displacement S of the end of the bar.7 events. y) cos [O'(x.wt}] 2) A2 (x. one obtains. On reconstruction. Thus. y) cos [0(x. then the event gets frozen into the hologram and hence one can focus through the depth of the reconstructed image and study the phenomenon at leisure. y) . photographic plate is again exposed along with the same reference wave. I is the moment of inertia of crosssection which for a rectangular bar of * The reconstruction process produces other wave components also but as was observed earlier. (b) interfere and produce interference fringes. 13. . y) cos{ 0(x. quantitative study of the fringe pattern produced in the body gives the distribution of strain in the object. if one has to study some transient phenomenon which occurs in a certain volume. whenever sume that the deformation of the object has been such as to (28) O' (x. p. 2. y) . y) . y) . Since (b) the object waves themselves have been reconstructed. y) = 2 m7r. m = 0. When this hologram is illuminated with a reconstruction wave. The ability of the holographic (a) process to release the object wave when reconstructed with a Photographic reconstruction wave allows us to perform interference beplate tween different waves which exist at different times. If a hologram is recorded of the scene. y) . alter only the phase distribution. ** See. y. 21. t) = A(x. 21. one of them corresponds to the unstressed object and the other to the stressed object. y)]. then using ordinary microscopic techniques it becomes difficult to first locate the position and make observation.75. .Thus.6 (a) Recording of the unstressed object wave.0(x. 1. y) to O' (x. e. We will consider here a simple application of the above technique in the determination of the Young's modulus of a material..wt} +A(x. O(x. on reconstruction. then two object waves emerge from the hologram. y) = ([A(x. depending on [O'(x. 2. 1. we as.7). if the two waves would interfere constructively and whenever. (29) where the phase distribution has been assumed to change from 0(x. 21.¢(x. these components travel along different directions. A emulsion to produce the doubly exposed hologram. m = 0.$(x. Then.. t) = A(x. These interference Recording of the stressed object wave on the same fringes are characteristic of the strain suffered by the body. each of the above two object waves emerge from the hologram and what would be observed will be the intensity pattern due to interference of the two waves which would be given by* I(x. the photographic plate is first partially exposed to the object wave and the reference wave. One of the most promising applications of holography lies Cantilever Mirror in the field of interferometry. y. y).6(a)] and if O ' (x. then one can show that** 8 W L3 3Y I (30) [0' (x. Thus. y) cos the two waves interfere destructively. t) represents the object wave when the object is stressed [see Fig. Thus. Here we are concerned only with the object waves. Thus. 21.. in the technique called double exposure holographic interferometry. y) . Ref.
27r or N2 8 .0). The phase change when the cantilever undergoes a displacement S as shown in Fig. In Fig. . Thus if we could determine S for a given load. ♦ If the object wave and the reference wave are plane waves. (30). To view the image. (28)] we can write 8 (cos o1 + coso2) = N . the hologram consists of a series of Young's interference fringes. 21. in general. 21.3). then Y can be determined from Eq.(cos 01 + cos 02) Thus by measuring N. we again illuminate the hologram with another wave. Summary 8 cos02) (31) 8 (cos 01 + cos02) If there are N fringes over the length L of the cantilever. called the reconstruction wave.6 we have shown the undisplaced and displaced positions of the cantilever illuminated by a laser light along a direction making an angle 01 with the zaxis. S. is given by I = ab3/12. From the number of fringes formed. The virtual image has all the characteristics of the object like parallax. thickness 0.21.6(b) would be 0 (8 cos01 = + exposed hologram of an aluminum strip of width 4 cm. 01 and o2 and knowing 2. Figure 21. Y represents the Young's modulus of the material of the rod. The reference wave is usually a plane wave. (Photograph courtesy: Professor R.7 shows the reconstruction of a double ♦ The basic technique in holography is the following : In the recording of the hologram. We observe the cantilever along a direction making an angle 02 with the taxis. This recorded interference pattern forms the hologram and contains information not only about the amplitude but also about the phase of the object wave.. to a virtual and a real image of the object scene. etc. The reconstruction process leads.8 Optics Fig. We will first determine an expression for (0' .7 Interference fringes produced in the measurement of Young's modulus using double exposure interfer ometry. ♦ For a point object and a plane reference wave.) width a and thickness b. one can calculate the Young's modulus (see Problem 21.2 cm and of length 12 cm. the hologram is very similar to a zone plate with the transmittance varying sinusoidally in cpntrast to the Fresnel zone plate. Sirohi. 21. S can be determined. then since a phase difference of 27r corresponds to one fringe [see Eq. one superimposes on the object wave another wave called the reference wave and the photographic plate is made to record the resulting interference pattern.
0. Smith. Flint. Plenum Press. A. 4B2 cos 2 {kd . Upatnieks. Sakher and Ajoy Ghatak. November. A. S. 36.3 Figure 21. P. 1978 (Reprinted by Macmillan. 1951. R. 1971. C. `Holography. Caulfield and S. 1948. Wiley Interscience. 4. A197. Lu. New . theory and applications'. N. American Journal of Physics. F. . Principles of Holography. Advanced Practical Physics for Students. 777. Vol. Ghatak. 1967. 161.. October. Vol. 212. M. Bombay. Vol. 35. by a plane wave traveling along a direction parallel to the zaxis. = 6328 A. K. 1970.2. B. Scientific American. Leith and J. C. p. Academic Press. R. the objects corresponding to the unstrained and strained positions of an aluminum bar of width 4 cm. Assume B = Aid. Proceedings of the Royal Society (London). J. Methernal. `Advances in holography'. Midwinter). Pennington. 40. New York. The Applications of Holography. 218. Contemporary Optics. Lasers: Theory and Applications.) 3.2. Journal of Scientific and Industrial Research. Bayvel and J. L. J. `Photography by Laser'.. 1949. calculate the interference pattern when the incident plane wave makes an angle 0 with the zaxis [see Fig. Holography. 12. New Delhi. 9. T. Collier. `Microscopy by Reconstructed Wavefronts'. 1984. calculate the Young's modulus of aluminum. June. Worsnop and H. 37. `Introduction to holography'. 1968. J. Scientific American.7 corresponds to the reconstruction of a doubly exposed hologram. Vol. D. New York. Elsevier (2005).1 Consider the reconstruction of the hologram as formed in the configuration of Example 21. Encyclopaedia of Modern Optics (Eds. Nature. Harte. K.kx sin 6 + (x2 + y2 ) }] [ Zd 21. Thyagarajan and A. 29. Assume 61 = 62 = 0. Burckhardt and L. Venkateshwarulu. 21. Brown and J. C. 1970. 1981 (Reprinted by 7.Holography* Problems 21. 13. A. Vol.2 cm and length 12 cm. thickness 0. 24. p. Lin. Thyagarajan. 1975. `Acoustical holography'. 13]. Macmillan India Ltd. E. Ghatak and K.7 x 10 11 N/m2] REFERENCES AND SUGGESTED READINGS 1. If the strained position corresponds to a load of 1 gm force applied at the end of the bar. H. Optical Holography. 454.2 In continuation of Example 21. 2. 1969. 11. H. 8. Guenther.York.9 21.) New York. 1965. 6. 1056. Vol. Miller. Givens. 1969. H. l Ans. Asia Publishing House. L.] [Ans. New Delhi. Scientific American. New York. American Journal of Physics. 221. February. [Hint: N represents the number of fringes produced over the length of the cantilever. K. 5. 14. Vol. Show the formation of a virtual and a real image. Plenum Press. Gabor. p. K. D. John Wiley & Sons. `A New Microscopic Principle'. M. assume A. B. A. Vol. `Holography in the undergraduate optics course'. 14. 10. 441.
.
various states of polarization are discussed. etc. In the first chapter of this part. Applications like optical activity. . the wave equation has also been derived which had led Maxwell to predict the existence of electromagnetic waves. total internal reflection. FabryPerot transmission resonances. the. generation and analysis of various forms of polarized light is discussed followed by a detailed analysis of propagation of electromagnetic waves in anisotropic media including first principle derivations of wave and ray velocities.5 E I ect ro m a C haract PART This part consists of three chapters discussing various aspects of the electromagnetic character of light waves. The results directly explain phenomena like Brewster's law. Faraday rotation etc. Reflection and refraction of electromagnetic waves by a dielectric interface has been discussed in chapter 24. have also been discussed. Chapter 23 is a bit mathematical . evanescent waves.starting with Maxwell's equations.
.
t)=acos(kzwt+0 2 ) x(z. . t) = a cos (kz .. The string can also be made to vibrate in the yz plane [see Fig.cot + 0) y(z. Christiaan Huygens Important Milestones 1669 1678 1809 1811 1828 1929 Erasmus Bartholinus discovered double refraction in calcite In the wave theory of light communicated to the Academie des Science in Paris.t)=acos(kzcot+0 1 ) y(z. and these I suppose would spread differently both in the ethereal matter diffused throughout the crystal and in the particles of which it is composed. I wished to try what Elliptical waves.t) = 0 where a represents the amplitude of the wave and 01 is the phase constant to be determined from the boundary condition. an American scientist and inventor. an arbitrary point z = zo will (1) execute a simple harmonic motion of amplitude a. 22.1(b)] for which the displacement would be given by and y(z. 22. If one rotates the end of the string on the circumference of a circle then each point of the string will move in a circular path as shown in Fig. Edwin Land.t) = 0 (2) In general. or rather spheroidal waves. known as a linearly polarized wave.t)=a sin(kzcot+0) (3) so that x2 + y2 is a constant (= a2). would do. 22.1(a)].this prism came to be known as the Nicol Prism. the ycoordinate of the displacement is always zero.. 22. therefore.As to the other emanation which should produce the irregular refraction. Christiaan Huygens gave the theory of double refraction in calcite discovered by Bartholinus Malus showed polarization of light by reflection David Brewster stated 'Brewster's law' William Nicol invented the prism which produced polarized light . such a wave is' known as a circularly polarized wave and the corresponding displacement would be given by and x (z.1 INTRODUCTION If we move one end of a string up and down then a transverse wave is generated [see Fig. the string can be made to vibrate in any plane containing the zaxis. Each point of the string executes a sinusoidal oscillation in a straight line (along the xaxis) and the wave is. It is also known as a plane polarized wave because the string is always confined to the xz plane.:'urther. The displacement for such a wave can be written in the form and x(z.1(a).2. patented Polaroid which is the name of a type of synthetic plastic sheet which is used to polarise light L 22. At any instant the displacement will be a cosine curve as shown in Fig.
We next consider a long narrow slit placed in the path of the string as shown in Fig. then the slit will allow only the component of the displacement. 22. the change in the amplitude of the transmitted wave with the orientation of the slit is due to the transverse character of the wave. (b) Each point on the string rotates on the circumference of the circle.3(a). 22. before we discuss the.2 (a) The displacement corresponding to a circularly polarized wave . Thus. 22.all points on the string are at the same distance from the zaxis.3(b)]. If the length of the slit is along the direction of the displacement then the entire amplitude will be transmitted as shown in Fig. We must mention here that if a longitudinal wave was propagating through the string then the amplitude of the transmitted wave would have been the same for all orientations of the slit. This is because of the fact that the slit allows only the component of the displacement. However. to pass through. an experiment which is.3(a). 22.22. 22. which is along the length of the slit. very similar to the experiment discussed above proves the transverse character of light waves. (b) Fig.1 (a) A linearly polarized wave on a string with the displacement confined to the xz plane. Indeed. 22. if the slit is at right angles to the direction of the displacement then almost nothing will be transmitted to the other side of the slit [see Fig.3 (b) If a linearly polarized transverse wave (propagating on a string) is incident on a long narrow slit. On the other Fig. in principle.4 Optics (a) (b) Fig. . which is along the length of the slit. experiment with light waves we must define an unpolarized wave. hand. (b) A linearly polarized wave on a string with the displacement confined to the yz plane. to pass through.
22. 22. Equations (4) and (5) also show that E and B are at right angles to each other and both the vectors are at right angles to the direction of propagation. If the plane of vibration is changed in a random manner in very short intervals of time. Ey _0.cot). In fact. By = Bo cos (kz .7(b).6(a). where k = c L) v eµ (6) and (7) represents the velocity of the waves. then such a wave is known as an unpolarized wave* . 17. 22. BZ = 0 (5) Let us next consider an ordinary light beam falling on a polaroid P1 as shown in Fig. The direction of the electric vector of the emergent beam will depend on the orientation ofthe polaroid.5) Ex =_E4 cos (kz_wt). 22.7(a)]. If an unpolarized wave falls on a slit Si (see Fig. for the wave to be characterized with a certain frequency v. 1 (8) and Bx = 0. the second slit is said to act as an analyzer. Since EZ = 0 and BZ = 0.4) then the displacement associated with the transmitted wave will be along the length of the slit and a rotation of the slit will not affect the amplitude of the transmitted wave although the plane of polarization of transmitted wave depend on the orientation of the slit.. We will discuss the basic electromagnetic theory in the next chapter where we will show that associated with a plane electromagnetic wave there is an electric field E and a magnetic field B which are at right angles to each other. In general. so that in the short interval it executes a large number of oscillations (see also Sec. For a linearly polarized wave propagating in the zdirection the electric and magnetic fields can be written in the form (see Fig. 22. As will be shown in Sec. 22.4 If an unpolarized wave propagating on a string is incident on a long narrow slit S1 then the transmitted beam will be linearly polarized and its amplitude will not depend on the orientation of Si.5 An xpolarized electromagnetic wave propagating in the zdirection. then the intensity of the emerging wave will depend on the relative orientation of S2 with respect to Si.Polarization and Double Refraction 22. however. the direction of propagation is along the vector (E x B).1). the transmitted wave will be linearly polarized and the slit Si is said to act as a polarizer. c and are the dielectric permittivity and the magnetic permeability of the medium. When such a beam is incident on a polaroid the emergent light is linearly polarized with its electric vector oscillating in a particular direction as shown in Fig. then by rotating the slit S2 we obtain a variation of the transmitted amplitude as discussed earlier. (33) of Chapter 231: Bo = v Eo µ Fig.2 the component of E along a particular direction gets absorbed by the polaroid and the component at right angles to it passes * By a short interval. The electromagnetic theory also tells us that [see Eq. . The transverse character of light waves was known in the early years of the nineteenth century. a polaroid is a plasticlike material used for producing polarized lightit will be discussed in detail in the next section. If this polarized beam falls on another slit S2 (see Fig. the wave is transverse. y Linearly polarized light Fig.4). i.e. the nature of the displacement associated with a light wave was known only after Maxwell had put forward his famous electromagnetic theory. the electric vector (in a plane transverse to the direction of propagation) keeps changing its direction in a random manner [see Fig. 22. Thus. 22. If this polarized wave is allowed to pass through another slit S2. EZ = 0_ (4)_ . we imply times which are short compared to the detection time. 22.5 We once again consider transverse waves generated at one end of a string. this time has to be much greater than 1/v. an ordinary light beam (like the one coming from a sodium lamp or from the sun) is unpolarized. however.
22. if we place another polaroid P2 [see Fig. It essentially consists of a large number of thin copper wires placed parallel to each other as shown in Fig.1 The Wire Grid Polarizer and the Polaroid Unpolarized light (c) Fig. This is due to the fact that the electric field does work on the electrons inside the thin wires and the energy associated with the electric field is lost in the Joule heating of the wires. Clearly. (b) For a linearly polarized wave. for the system to be effective (i. through. then the intensity of the transmitted light will depend on the relative orientation of P2 with respect to P1.e..6(b)]. 22. On the basis of our earlier discussions. Thus the emergent wave is linearly polarized with the electric vector along the xaxis. the. if the position of the eye is as shown in the figure. . the fabrication of such a polarizer for a 3 cm microwave is relatively easy (a) For an unpolarized wave propagating in the + zdirection.8. i.2 PRODUCTION OF POLARIZED LIGHT In this section we will discuss various methods for the production of linearly polarized light waves. (a) Fig. 22.2. The direction of the electric vector of the emergent wave is usually called the pass axis of the polaroid. On the other hand. 22. 22. then by rotating the x Fig. then one will observe no variation of intensity if the polaroid is rotated about the zaxis.7 (b) The physics behind the working of the wire grid polarizer is probably the easiest to understand. the electric (or the magnetic) vector oscillates along a particular direction. also be observed if instead of rotating the polaroid P2 we rotate Pl. However. 22. However. the electric vector (which lies in the xy plane) continues to change its direction in a random manner. displacement associated with a light wave is at right angles to the direction of propagation of the wave. The second polaroid acts as an analyzer.6 Optics P1 Linearly polarized / light Unpolarized light Z (a) Eye polaroid P2 (about the zaxis) one will observe variation of intensity and at two positions there will be almost complete darkness [see Fig. this phenomenon proves the transverse character of light. 22.6(a).. Returning to Fig.8 The wiregrid polarizer. When an unpolarized electromagnetic wave is incident on it then the component of the electric vector along the length of the wire is absorbed.e. for the Ey 'component to be almost completely attenuated) the spacing between the wires should be < 7 . then the emerging beam will be linearly polarized and if we place another Polaroid P2. (sincethe wires are assumed to be very thin) the component of the electric vector along the xaxis passes through without much attenuation. 22. A similar phenomenon will .6 If an ordinary light beam is allowed to fall on a Polaroid.22. The polaroid P1 acts as a polarizer and the transmitted beam is linearly polarized. 22.6(c)].
For the airwater interface. 24.000 wires in about one inch. On the other hand. if an unpolarized beam is incident at this angle. . If we now view through a rotating Polaroid. 22. 22.9(b)]. 10 cm).2 that if the angle of incidence (0) is such that 9=Op =tarit (nt l (9) then the reflection coefficient is zero. one may employ long chain polymer molecules that contains atoms (like iodine) which provide high conductivity along the length of the chain. the sea will appear * For further details. Thus. Nevertheless. the fabrication of a polarizer in which the wires are placed at distances 55 x 10 cm is extremely difficult. The original work of Bird and Parrish was published in 19502 (see Ref. the angle 0p is known as the polarizing angle or the Brewster angle. then the reflected light will be almost polarized (see Fig. see Ref. the electric field parallel to the molecules gets absorbed. However. These long chain molecules are aligned so that they are almost parallel to each other. the reflected and the transmitted rays are at right angles to each other.9(b)]. It will be shown in Sec. the polaroid is usually very effective in producing linearly polarized light.7 because the spacing has to be < 3 cm. 1. nl = 1 and n2 = 1.9(a). Bird and Parrish did succeed in putting about 30. For the airglass interface.5 giving 9p 57°. The aligning of the long chain conducting molecules is not very difficult. 2). As already pointed out. 1. When a light beam is incident on such a polaroid. nl = 1 and n2 = 1. The transmitted beam is partially polarized and if one uses a large number of reflecting surfaces. Because of the high conductivity provided by the iodine atoms.2. 22.33 and the polarizing angle 6p = 53°. 22.9 (a) If a linearly polarized wave (with its E in the plane of incidence) is incident on the interface of two dielectrics with the angle of incidence equal to 8. (= tan 1 n2 /n l ) then the reflection coefficient is zero. instead of long thin wires.Polarization and Double Refraction 22.2 Polarization by Reflection Let us consider the incidence of a plane wave on a dielectric.10). A sheet containing such long chain polymer molecules (which are aligned parallel to each other) is known as a polaroid. Equation (9) is referred to as the Brewster's law and at this angle of incidence. 22. since the light waves are associated with a very small wavelength (5 x.* The details of the procedure for making this wire grating are also discussed in this book. (b) Thus.** 5 5 No reflected wave ni n2 ep Air Glass Air Glass Air Almost polarized (b) Fig. the molecules (aligned parallel to each other) absorb the component of electric field which is parallel to the direction of alignment because of the high conductivity provided by the iodine atoms. then the emergent beam is almost plane polarized with its electric vector in the plane of incidence. The transmitted beam is partially polarized and if this beam is made to undergo several reflections. one would obtain an almost plane polarized transmitted beam [see Fig. it is extremely difficult to fabricate a wire grid polarizer which would be effective for visible light. Thus the aligned conducting molecules act similar to the wires in the wire grid polarizer and since the spacing between two adjacent long chain molecules is small compared to the optical wavelength. We assume that the electric vector associated with the incident wave lies in the plane of incidence as shown in Fig. the component perpendicular to it passes through. Thus if the sunlight is incident on the sea at an angle close to the polarizing angle. if an unpolarized beam is incident at this angle. 22. ** Experimental details of producing the polarizer are given in Ref. then the reflected beam will be linearly polarized with its electric vector perpendicular to the plane of incidence [see Fig. the reflected beam is plane polarized whose electric vector is perpendicular to the plane of incidence.
12 that the two beams have different velocities and as such the corresponding refractive indices will be different. 22. Fig.11 (b) and Fig. figure adapted from http://polarization. If the Polaroid allows the (almost polarized) reflected beam to pass through. 27 in the prelim pages. If one can sandwich a layer of a material whose refractive index lies between the two.com/water/water. this property of selective absorption is known as dichroism. 22.12 we will discuss the phenomenon of double refraction and will show that when an unpolarized beam enters an anisotropic crystal. then the reflected light will be almost polarized.3 Polarization by Double Refraction In Secs 22. 28 in the prelim pages. the glare can be blocked by using a vertical polarizer and one can see the inside of the water [see Fig. 22. we see the glare from water surface [see Fig.htm . .11 shows sunlight incident on a water surface at an angle close to the polarizing angle so that the reflected light is almost polarized. Thus.2. Consequently.5 and 22. then the reflected light will be almost polarized.11 (a) (b) If the sunlight is incident on the water surface at an angle close to the polarizing angle. Figure adapted from http:// polarization. one of the beams gets absorbed quickly and the other component passes through without much attenuation. If we now view through a rotating Polaroid. 22. (a) If the polaroid allows the (almost polarized) reflected beam to pass through. 22.12). each of them being characterized by a certain state of polarization.html .com/water/water. the emergent beam will be linearly polarized (see Fig. 28 in the prelim pages]. if an unpolarized beam is passed through a tourmaline crystal. A crystal like tourmaline has different coefficients of absorption for the two linearly polarized beams into which the incident beam splits up. then for one of the Fig. the sea will appear more transparent when the Polaroid blocks the glare from the reflected light.22. Figure 22.10 If the sunlight is incident on the sea around the polarizing angle. we see the glare from water surface. we could eliminate one of the beams then we would obtain a linearly polarized beam. A color photograph appears as Fig. If by some method. it splits up into two beams. We will show in Secs 22.5 and 22. 22. A color photograph appears as Fig. (b) The glare can be blocked by using a vertical polarizer and one can see the inside of the water. Another method for eliminating one of the polarized beams is through total internal reflection.11 (a)]. A simple method for eliminating one of the beams is through selective absorption.8 Optics more transparent when the Polaroid blocks the reflected light.
14(b)].4 Polarization by Scattering If an unpolarized beam is allowed to fall on a gas. if the incident beam is linearly polarized with its electric vector along the xdirection. 22. used with permission. One of the components gets absorbed quickly and the other component passes through without much attenuation. the incidence will be at a rarer medium and for the other it will be at a denser medium. i.2.3 MALUS' LAW Let us consider a polarizer PI which has a passaxis parallel to the xaxis (see Fig. 22.] beams. the angle of incidence is greater than the critical angle. then there will be no scattered wave in the xdirection. then there will be no scattered light along the xaxis. 22. The Nicol prism. The dashed outline corresponds to the natural crystal which is cut in such a way that the ordinary ray . 22. As such. (b) If a linearly polarized wave (with its E oscillating along the xdirection) is incident on a dipole. [Adapted from Ref.14 22. The ycomponent of the (a) If the electromagnetic wave is propagating along the zdirection. It may be of interest to mention that it was through scattering experiments that Barkla could establish the transverse character of Xrays.15). This principle is used in a Nicol prism which consists of a calcite crystal cut in such a way that for the beam. 3. 22. then the beam scattered at 90° to the incident beam is linearly polarized. . 23.. 68° Fig.9 xpolarized scattered wave (a) Tourmaline crystal Fig.e. Clearly. for which the sandwiched material is a rarer medium.13 22.14).4).12 When an unpolarized beam enters a dichroic crystal like tourmaline. Scattered wave polarized in the xdirection (b) Fig. undergoes total internal reflection at the Canada Balsam layer. Thus this particular beam will be eliminated by total internal reflection. one can carry out an analysis of a scattered wave by allowing it to undergo a further scattering [see Fig. This follows from the fact that the waves propagating in the ydirection are produced by the xcomponent of the dipole oscillations (see Fig. Figure 22. dipole oscillations will produce no field in the ydirection (see Sec. it splits up into two linearly polarized components.13 shows a properly cut calcite crystal in which a layer of Canada Balsam has been introduced so that the ordinary ray undergoes total internal reflection.Polarization and Double Refraction y 22. then the scattered wave along any direction is perpendicular to the zaxis will be linearly polarized. The extraordinary component passes through and the beam emerging from the crystal is linearly polarized. 22. if an unpolarized beam propagating in the zdirection is incident on the polarizer.
.wt + 01) a2 (11) (12) cos (kz . Equation (14) tells us that the resultant is also a linearly polarized wave with its electric vector oscillating along the same axis.02)] 12 represents the amplitude of the wave. we consider the time variation of the resultant electric field at an arbitrary plane perpendicular to the zaxis which we may. it will pass through a maximum and again a minimum before it reaches its original position. In order to find the state of polarization of the resultant field. Equation (10) represents the Malus ' Law. a further rotation will result in a decrease in intensity till the passaxis is parallel to the yaxis. then for the pass axis to be along the xdirection.22. If we further rotate it.cot + 0) (16) (17) where lo represents the intensity of the emergent beam when the pass axis of P2 is also along the xaxis (i.cot + 02) where al and a2 represent the amplitudes of the waves. Thus.15 An unpolarized light beam gets xpolarized after passing through the polaroid PI. without any loss of generality. Thus. the pass axis of the second polaroid P2 makes an angle 0 with the xaxis. 22. then Ex = a1 cos cot (18) (19) and Ey = a2 cos (cot .10 x Optics electric fields associated with the waves can be written in the form El = E2 = al cos (kz . The resultant of these two waves would be given by E=E1 +E2 Fig. The . the long chain molecules must be aligned along the yaxis. this angle will depend on the relative values of a1 and a2. For 0 = n 7r.0) where we have used Eqs (16) and (17) with z = O. the resultant will also be a linearly polarized wave with its electric vector oscillating along a direction making "a certain angle with the xaxis. If the amplitude of the incident electric field is Bo. 22. It should be noted that if the polarizer is a polaroid. where the intensity will be almost zero.cot + 0) + a22 (14) (15) = [a 12 + 2a1 a2 cos(01 . If Ex and Ey represent the x.a1 (independent of t) (21) 22. For 0 = nir. 22.wt) a2 cos (kz . we may have E1 = E2 = a1 cos (kz . if a linearly polarized beam is incident on a polaroid and if the polaroid is rotated about the zaxis. x represents the unit vector along the xaxis and 01 and 02 are phase constants. The intensity of the emerging beam will vary as cos2 O. when 0 = 0). if the polaroid P2 shown in Fig.15). the above equations simplify to Ex = a1 cos cot and Ey = (1)" a2 cos wt (20) from which one obtains Ey Ex + a2 . then the intensity will increase till the passaxis is parallel to the xaxis. We next consider the superposition of two linearly polarized electromagnetic waves (both propagating along the zaxis) but with their electric vectors oscillating along two mutually perpendicular directions.4 SUPERPOSITION OF TWO DISTURBANCES Let us consider the propagation of two linearly polarized electromagnetic waves (both propagating along the zaxis) with their electric vectors oscillating along the xaxis.e. For example. then the intensity of the emergent wave will vary according to the above law. We next consider the incidence of the xpolarized beam on the polaroid P2 whose pass axis makes an angle 0 with the xaxis (see Fig. then the amplitude of the wave emerging from the polaroid P2 will be E0 cos 0 and thus the intensity of the emerging beam will be given by I =1o cos20 (10) which can always be written in the form E= where a a (13) cos (kz .15 is rotated in' the clockwise direction. assume to be z = O. then the electric vector associated with the emergent wave will oscillate along the xaxis.and ycomponents of the resultant field E (= El + E2).
22. oscillating in a direction which is different from the fields of either of the two waves.5 a1 (d) z p Propagation is along zaxiscoming Fig. Thus. e=ir a2 = 0.. For example. 4.r/3 = 2^c LP LEP (h) (i) out of the paper. In the Ex Ey plane. a1 'I e=o a 2 =1.17(c)] and the propagation is in the + zdirection which is coming out of the page. (21). similarly. For 0 ^ nor (n = 0. 4 but in some books the opposite convention is used. The propagation is out of the page. oscillate along a straight line.).. 22. Fig.and ycomponents are given by Eqs (23) and (24).* That the tip of the resultant LP .it* Polarization and Double Refraction 22. We first consider the simple case corresponding to 0 = is/2 with a1 = a2. The resultant is again a linearly polarized wave with its electric vector oscillating in a direction making an angle rp with the xaxis. is again a linearly polarized wave with its electric vector. where the upper and lower signs correspond to n even and 'n odd respectively. . in general. in general. (20) for various values of a2/a 1 .the resultant electric vector does not.11 thick lines shown in the figure.16 is a plot of the resultant field corresponding to Eq. Figure 22.Ex E .. 5. Eq. . 3. 2. the superposition of two linearly polarized electromagnetic waves with their electric fields at right angles to each other and oscillating in phase. for labelling left and right circularly polarized light is consistent with the one used by Ref. 1.17 States of polarization for various values of 0 corresponding. * Our convention.. the angle (0) that this line makes with the Ex axis depends on the ratio a 2/a l . 2.5a1 y (b) LP Ey 0=4n/3 (f) y Ex Ex LEP 0= 3... In fact 0 =tan 1 ±a2 al (22) The condition 0 = nic implies that the two vibrations are either in phase (n = 0.to al = a2 [see Eqs (18) and (19)]. (21) represents a straight line. . .. Thus. The tip of the electric vector oscillates (with angular frequency co) along the Ey Ey Ex= al Ey = a l cos cot (23) sin cot (24) If we plot the time variation of the resultant electric vector whose x.16 The superposition of two linearly polarized waves with their electric fields oscillating in phase along the xaxis and the yaxis. (c) and (g) correspond to right circularly andleft_circularly polarized light respectively.). The equation of the straight line is given by Eq.r/2 (g) 0=. one would find that the tip of the electric vector rotates on the circumference of a circle (of radius al) in the anticlockwise direction [see Fig.) or out of phase (n = 1. Such a wave is known as a right circularly polarized wave (usually abbreviated as a RCP wave). (b) and (d) correspond to right elliptically polarized (REP) light and (f) and (h) correspond to left elliptically polarized (LEP) light. .5. 22.
Eysina Ey = Ex sin a + Ey cos a If we multiply the first equation by cos a and the second equation by sin a and add.0) sin a E2 sin(cot . one may produce an elliptically polarized wave by allowing two linearly polarized waves to propagate through the string. Such a wave is known as a left circularly polarized wave (usually abbreviated as a LCP wave). For 0 ^ m7r/2 (m = 0. similar to the case of an electromagnetic wave. when al ^ a2..17(b)]. it is the magnitude which keeps on oscillating about the zero value with the angular frequency of the wave. we will show that Eqs (18) and (19) represent an elliptically polarized wave. the presence of these fields can be felt by their interaction with a charged particle: In particular. we equate the coefficients of cos cot and sin cot on both sides of the equation to obtain El cos 0 = El sin 0 = and E2 E2 al a2 cos a + a2 cos 0 sin a sin a sin a sin 0 = . we would get E? +. 27r. On the other hand.1.al cos cot sin a + a2 cos(cot ..). E' ( El ) 2 +I E' 2 E2J