Tunable Laser Optics

TUNABLE LASER OPTICS This Page Intentionally Left Blank T U N A B L E LASER OPTICS Francisco J. Duarte Eastman Kod...

Author: Frank J. Duarte

282 downloads 3456 Views 10MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

TUNABLE LASER OPTICS

This Page Intentionally Left Blank

T U N A B L E LASER OPTICS

Francisco J. Duarte Eastman Kodak Company Research Laboratories Rochester, New York

ELSEVIER ACADEMIC PRESS Amsterdam Boston Heidelberg London New York Oxford Paris San Diego San Francisco Singapore Sydney Tokyo

This book is printed on acid-free paper. @ Copyright 2003, Elsevier Science (USA). All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permissions may be sought directly from Elsevier's Science & Technology Rights Department in Oxford, UK: phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail: [email protected]. You may also complete your request on-line via the Elsevier Science homepage (http://elsevier.com), by selecting "Customer Support" and then "Obtaining Permissions." ACADEMIC PRESS An imprint of Elsevier Science 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA http://www.academicpress.com Academic Press 84 Theobald's Road, London WC1X 8RR, UK http://www.academicpress.com Library of Congress Control Number: 2003108747 International Standard Book Number: 0-12-222696-8

PRINTED IN THE UNITED STATES OF AMERICA 03 04 05 06 07 9 8 7 6 5 4 3 2 1

To my parents, Ruth Virginia and Luis Enrique


Contents

Preface

xiii

Chapter 1

Introduction to Lasers 1.1 Introduction 1 1.1.1 Historical Remarks 2 1.2 Lasers 3 1.2.1 Laser Optics 5 1.3 Excitation Mechanisms and Rate Equations 1.3.1 Rate Equations 5 1.3.2 Dynamics of the Multiple-Level System 1.3.3 Transition Probabilities and Cross Sections 1.4 Laser Resonators and Laser Cavities 14 Problems 20 References 20

11

Chapter 2

Dirac Optics 2.1 Dirac Notation in Optics 23 2.2 Interference 25 2.2.1 Geometry of the N-Slit Interferometer 2.2.2 N-Slit Interferometer Experiment 2.3 Diffraction 32 2.4 Refraction 38 2.5 Reflection 39 40 2.6 Angular Dispersion

29 29

vii

viii

Contents

2.7 Dirac and the Laser Problems 42 42 References

41

Chapter 3

The Uncertainty Principle in Optics 3.1 Approximate Derivation of the Uncertainty Principle 45 3.1.1 The Wave Character of Particles 45 3.1.2 The Diffraction Identity and the Uncertainty Principle 46 49 3.1.3 Alternative Versions of the Uncertainty Principle 49 3.2 Applications of the Uncertainty Principle in Optics 3.2.1 Beam Divergence 50 3.2.2 Beam Divergence and Astronomy 52 3.2.3 The Uncertainty Principle and the Cavity Linewidth Equation 54 Problems 55 References 55

Chapter 4

The Physics of Multiple-Prism Optics 4.1 Introduction 57 4.2 Generalized Multiple-Prism Dispersion 58 4.2.1 Double-Pass Generalized Multiple-Prism Dispersion 60 4.2.2 Multiple Return-Pass Generalized Multiple-Prism Dispersion 62 4.2.3 Single-Prism Equations 64 64 4.3 Multiple-Prism Dispersion and Linewidth Narrowing 4.3.1 The Mechanics of Linewidth Narrowing in Optically Pumped Pulsed Laser Oscillators 65 4.3.2 Design of Zero-Dispersion Multiple-Prism Beam Expanders 67 4.4 Multiple-Prism Dispersion and Pulse Compression 68 72 4.5 Applications of Multiple-Prism Arrays Problems 72 References 73

Contents

Chapter 5

Linear Polarization 5.1 Maxwell Equations 75 5.2 Polarization and Reflection 77 5.2.1 The Plane of Incidence 79 5.3 Polarizing Prisms 79 5.3.1 Transmission Efficiency in Multiple-Prism Arrays 80 5.3.2 Induced Polarization in a Double-Prism Beam Expander 81 5.3.3 Double-Refraction Polarizers 82 5.3.4 Attenuation of the Intensity of Laser Beams Using Polarization 84 5.4 Polarization Rotators 85 5.4.1 Fresnel Rhombs and Total Internal Reflection 85 5.4.2 Birefringent Rotators 86 5.4.3 Broadband Prismatic Rotators 87 Problems 90 References 91

Chapter 6

Laser Beam Propagation Matrices 6.1 Introduction 93 93 6.2 ABCD Propagation Matrices 6.2.1 Properties of A B C D Matrices 95 96 6.2.2 Survey of A B C D Matrices 6.2.3 The Astronomical Telescope 96 6.2.4 A Single-Prism in Space 103 6.2.5 Multiple-Prism Beam Expanders 104 6.2.6 Telescopes in Series 106 6.2.7 Single-Return-Pass Beam Divergence 107 6.2.8 Multiple-Return-Pass Beam Divergence 108 6.2.9 Unstable Resonators 110 6.3 Higher-Order Matrices 111 Problems 114 References 114

ix

Contents

Chapter 7

Pulsed Narrow-Linewidth Tunable Laser Oscillators 7.1 Introduction 115 7.2 Transverse and Longitudinal Modes 116 7.2.1 Transverse-Mode Structure 116 7.2.2 Longitudinal-Mode Emission 118 7.3 Tunable Laser Oscillator Architectures 122 7.3.1 Tunable Laser Oscillators Without Intracavity Beam Expansion 122 7.3.2 Tunable Laser Oscillators with Intracavity Beam Expansion 126 7.3.3 Widely Tunable Narrow-Linewidth External-Cavity Semiconductor Lasers 131 7.3.4 Distributed-Feedback Lasers 134 7.4 Wavelength-Tuning Techniques 136 7.4.1 Prismatic Tuning Techniques 137 7.4.2 Diffractive Tuning Techniques 138 7.4.3 Interferometric Tuning Techniques 139 7.4.4 Longitudinal Tuning Techniques 141 7.4.5 Synchronous Tuning Techniques 142 7.5 Polarization Matching 144 7.6 Design of Efficient Narrow-Linewidth Tunable Laser Oscillators 146 7.6.1 Useful Axioms for the Design of Narrow-Linewidth Tunable Laser Oscillators 147 7.7 Narrow-Linewidth Oscillator-Amplifiers 148 7.7.1 Laser-Pumped Narrow-Linewidth Oscillator-Amplifier Configurations 148 7.7.2 Narrow-Linewidth Master-Oscillator Forced-Oscillator Configurations 150 Problems 152 References 152

Chapter 8

Nonlinear Optics 8.1 Introduction 157 8.2 Generation of Frequency Harmonics 159 8.2.1 Second-Harmonic and Sum-Frequency Generation 8.2.2 Difference-Frequency Generation and Optical Parametric Oscillation 162

159

xi

Contents

8.2.3 The Refractive Index as a Function of Intensity 8.3 Optical Phase Conjugation 167 8.4 Raman Shifting 170 8.5 Applications of Nonlinear Optics 172 Problems 174 References 174

166

Chapter 9

Lasers and Their Emission Characteristics 9.1 Introduction 177 9.2 Gas Lasers 178 9.2.1 Pulsed Molecular Gas Lasers 179 9.2.2 Pulsed Atomic and Ionic Metal Vapor Lasers 9.2.3 Continuous-Wave Gas Lasers 182 9.3 Dye Lasers 184 9.3.1 Pulsed Dye Lasers 184 9.3.2 Continuous-Wave Dye Lasers 187 9.4 Solid-State Lasers 189 9.4.1 Ionic Solid-State Lasers 189 9.4.2 Transition Metal Solid-State Lasers 189 9.4.3 Color-Center Lasers 191 9.4.4 Diode-Laser-Pumped Fiber Lasers 191 9.4.5 Optical Parametric Oscillators 192 9.5 Semiconductor Lasers 193 195 9.6 Additional Lasers References 196

181

C h a p t e r 10

Architecture of N-Slit Interferometric Laser Optical Systems 10.1 Introduction 203 10.2 Optical Architecture of the N-Slit Laser Interferometer 10.2.1 Beam Propagation in the N-Slit Laser Interferometer 206 10.3 An Interferometric Computer 208 10.4 Applications of the N-Slit Laser Interferometer 211 10.4.1 Digital Laser Microdensitometer 211 10.4.2 Light Modulation Measurements 214 10.4.3 Secure Interferometric Communications in Free Space 214

204

xii

Contents

10.4.4 Wavelength Meter and Broadband Interferograms 221 10.5 Sensitometry 222 Problems 224 References 225 Chapter 11

Spectrometry and Interferometry 11.1 Introduction 11.2 Spectrometry

227 227

11.2.1 Prism Spectrometers 228 11.2.2 Diffraction Grating Spectrometers 229 11.2.3 Dispersive Wavelength Meters 231 11.3 Interferometry 233 11.3.1 Two-Beam Interferometers 233 11.3.2 Multiple-Beam Interferometers 236 242 11.3.3 Interferometric Wavelength Meters Problems 247 References 247 Chapter 12

Physical Constants and Optical Quantities 12.1 12.2 12.3 12.4 12.5

Fundamental Physical Constants 249 Conversion Quantities 250 Units of Optical Quantities 250 Dispersion of Optical Materials 250 ~n/o~Tof Optical Materials 251 References 252

Appendix of Laser Dyes Index 267

253

Preface

Since the introduction of the laser, the field of optics has experienced an enormous expansion. For students, scientists, and engineers working with lasers but not specialized in lasers or optics, there is a plethora of sources of information at all levels and from all angles. Tunable Laser Optics was conceived from a utilitarian perspective to distill into a single, and concise, volume the fundamental optics needed to work efficiently and productively in an environment employing lasers. The optics tools presented in Tunable Laser Optics use humble, practical mathematics. Although the emphasis is on optics involving macroscopic low-divergence, narrow-linewidth lasers, some of the principles described can also be applied in the microscopic domain. The style and the selection of subject matter in Tunable Laser Optics were determined by a desire to reduce entropy in the search for information on this wonderful and fascinating subject. The author is grateful to the U.S. Army Aviation and Missile Command (Redstone Arsenal, Alabama) for supporting some of the word discussed in the book. F. J. Duarte

Rochester, New York April, 2003

xiii


Chapter I

Introduction to Lasers

1.1 I N T R O D U C T I O N Lasers are widely applied in academic, medical, industrial, and military research. Lasers are also used beyond the boundary of research, in numerous applications that continue to expand. Optics principles and optical elements are applied to build laser resonators and to propagate laser radiation. Optical instruments are utilized to characterize laser emission, and lasers have been incorporated into new optical instrumentation. Tunable Laser Optics focuses on the optics and optical principles needed to build lasers, on the optics instrumentation necessary to characterize laser emission, and on laser-based optical instrumentation. The emphasis is on practical and utilitarian aspects of relevant optics, including the necessary theory. Though this book refers explicitly to macroscopic lasers, many of the principles and ideas described here are applicable to microscopic lasers. Tunable Laser Optics was written for advanced undergraduate students in physics, nonoptics graduate students using lasers, engineers, and scientists from other fields seeking to incorporate lasers and optics into their work. Tunable Laser Optics is organized into three areas. It begins with an introduction to laser concepts and a series of chapters that introduce the ideas necessary to quantify the propagation of laser radiation and that are central to the design of tunable laser oscillators. The second area begins with a chapter on nonlinear optics that has intra- and extracavity applications. The attention is then focused on a survey of the emission characteristics of most wellknown lasers. The third area includes a chapter on interferometric optical

2

Tunable Laser Optics

instrumentation and by a chapter on instrumentation for measurements on laser characteristics. A set of fairly straightforward problems ends every chapter, to assist the reader in assessing assimilation of the subject matter. Thus, the book begins with an introduction to some basic concepts of laser excitation mechanisms and laser resonators in Chapter 1. The focus then turns to optics principles, with Dirac optics being discussed in Chapter 2 and the uncertainty principle introduced in Chapter 3. The principles of dispersive optics are described in Chapter 4, while linear polarization is discussed in Chapter 5. Next, propagation matrices are introduced in Chapter 6. The optical principles discussed in Chapters 1 through 6 can all be applied to the design and construction of tunable laser oscillators, as described in Chapter 7. Nonlinear optics, with an emphasis on frequency conversion, is outlined in Chapter 8. A brief but fairly comprehensive survey of lasers in the gaseous, liquid, and solid states is given in Chapter 9. Attention is focused on the emission characteristics of the various lasers. For this second area it is hoped that the student will have gained sufficient confidence and familiarity with the subject of laser optics to select an appropriate gain medium and resonator architecture for its efficient use in an applied field. The optics architecture and applications of N-slit laser interferometers are considered in Chapter 10, while optics-based diagnostic instrumentation is described in Chapter 11. The book concludes with an appendix on useful physical constants and optical quantities. It should be emphasized that the material in this book does not require mathematical tools above those available to a third-year undergraduate physics student. Also, perhaps with the exception of Chapter 7, individual chapters can be studied independently.

1.1.1 HISTORICAL REMARKS A considerable amount has been written about the history of the maser and the laser. For brief and yet informative historical summaries the reader should refer to Willett (1974), Siegman (1986), and Silfvast (1996). Here, remarks will be limited to mention that the first experimental laser was demonstrated by Maiman (1960) and that this laser was an optically pumped solid-state laser. More specifically, it was a flashlamp-pumped ruby laser. This momentous development was followed shortly afterwards by the introduction of the first electrically excited gas laser (Javan et al., 1961). This was the He-Ne laser emitting in the near infrared. From a practical perspective, the demonstration of these laser devices also signaled the birth of experimental laser optics, since the laser resonators, or laser optical cavities, are an integral and essential part of the laser.


3

Two publications apparently unrelated to the laser are mentioned next. The first is the description that Dirac gave on interference in his book The Principles of Quantum Mechanics, first published in 1930 (Dirac, 1978). In his statement on interference, Dirac refers first to a source of monochromatic light and then to a beam of light consisting of a large number of photons. In his discussion, it is this beam composed of a large number of undistinguishable photons that is divided and then recombined to undergo interference. In this regard, Dirac could have been describing a high-intensity laser beam with a very narrow linewidth (Duarte, 1998). Regardless of the prophetic value of Dirac's description, his was probably the first discussion in physical optics to include a coherent beam of light. In other words, Dirac wrote the first chapter in laser optics. The second publication of interest is The Feynman Lectures on Physics, authored by Feynman et al. (1965). In Chapter 9 of the volume on quantum mechanics, Feynman uses Dirac's notation to describe the quantum mechanics of stimulated emission. In Chapter 10 he applies that physics to several physical systems, including dye molecules. Notice that this was done just prior to the discovery of the dye laser by Sorokin and Lankard (1966) and Sch~fer et al. (1966). In this regard, Feynman could have predicted the existence of the tunable laser. Further, Feynman made accessible Dirac's quantum notation via his thought experiments on two-slit interference with electrons. This provided the foundations for the subject of Dirac optics, described in Chapter 2, where the method outlined by Feynman is extended to generalized transmission gratings using photons rather than electrons.

1.2 L A S E R S The word laser has its origin in an acronym of the words light amplification by stimulated emission of radiation. Although the laser is readily associated with the spatial and spectral coherence characteristics of its emission, to some the physical meaning of the concept still remains shrouded in mystery. Looking up the word in a good dictionary does not help much. A laser is a device that transforms electrical energy, chemical energy, or incoherent optical energy into coherent optical emission. This coherence is both spatial and spectral. Spatial coherence means a highly directional light beam, with little divergence; spectral coherence means an extremely pure color of emission. An alternative way to cast this idea is to think of the laser as a device that transforms ordinary energy into an extremely well-defined form of energy, both in the spatial and the spectral domains. However, this is only the manifestation of the phenomenon, since the essence of this energy transformation lies in the device called the laser.

4


Physically, the laser consists of an atomic or molecular gain medium optically aligned within an optical resonator or optical cavity, as depicted in Fig. 1.1. When excited by electrical energy or optical energy, the atoms or molecules in the gain medium oscillate at optical frequencies. This oscillation is maintained and sustained by the optical resonator or optical cavity. In this regard, the laser is analogous to a mechanical or radio oscillator but oscillating at extremely high frequencies. For the green color of A = 500 nm, the equivalent frequency is u ~ 5.99 x 1014 Hz. A direct comparison between a laser and a radio oscillator makes the atomic or molecular g~,~ medium equivalent to the transistor and the elements of the optical cavity equivalent to the resistances, capacitances, and inductances. Thus, from a physical perspective the gain medium, in conjunction with the optical cavity, behaves like an optical oscillator (see, for example, Duarte (1990a)). The spectral purity of the emission of a laser is related to how narrow its linewidth is. High-power narrow-linewidth lasers can have linewidths of Au ~ 300 MHz; low-power narrow-linewidth lasers can have Au ~ 100 kHz; and stabilized lasers can yield Au ~ 1 kHz or less. In all the instances mentioned here the emission is in the form of a single longitudinal mode; that is, all the emission radiation is contained in a single electromagnetic mode. In the language of the laser literature, a laser emitting narrow-linewidth radiation is referred to as a laser oscillator or as a master oscillator (MO). High-power narrow-linewidth emission is attained when an M O is used to inject a laser amplifier, or power amplifier (PA). Large high-power systems include several M O P A chains, with each chain including several amplifiers. The difference between an oscillator and an amplifier is that the amplifier simply stores energy to be released upon the arrival of the narrow-linewidth oscillator signal. In some cases the amplifiers are configured within unstable resonator cavities in what is referred to as a forced oscillator (FO). When that is the case, the amplifier is called a forced oscillator and the integrated configuration is referred to as a M O F O system. This subject is considered in more detail in Chapter 7.

Gain medium

M1

--

L

M2 ,

Figure 1.1 Basiclaser resonator. It comprises an atomic or molecular gain medium and two mirrors aligned along the optical axis. The length of the cavity is L, and the diameter of the beam is 2w. The gain medium can be excited optically or electrically.

Introduction to Lasers 1.2.1

5

LASER O P T I C S

Laser optics refers to the individual optics elements that comprise laser cavities, to the optics ensemblies that comprise laser cavities, and to the physics that results from the propagation of laser radiation. In addition, the subject of laser optics includes instrumentation employed to characterize laser radiation and instrumentation that incorporates lasers.

1.3 E X C I T A T I O N EQUATIONS

MECHANISMS

AND

RATE

There are various methods and approaches to describing the dynamics of excitation in the gain media of lasers. Approaches range from complete quantum mechanical treatments to rate equation descriptions (Haken, 1970). A complete survey of energy level diagrams corresponding to gain media in the gaseous, liquid, and solid states is given by Silfvast (1996). Here, a basic description of laser excitation mechanisms is given using energy levels and classical rate equations applicable to tunable molecular gain media. The link to the quantum mechanical nature of the laser is made via the cross sections of the transitions.

1.3.1 RATEEQUATIONS Rate equations are widely applied in physics and in laser physics in particular. Rate equations, for example, can be used to describe and quantify the process of molecular recombination in metal vapor lasers or to describe the dynamics of the excitation mechanism in a multiple-level gain medium. The basic concept of rate equations is introduced using an ideally simplified two-level molecular system, depicted in Fig. 1.2. Here, the pump excitation intensity Ip(t), populates the upper energy level N1 from the ground state No. Emission from the upper state is designated as Iz (x, t, A) since it is a function of position x in the gain medium, time t, and wavelength A. The time evolution of the upper-state, or excited-state, population can be written as

(ON1/Ot) = N0cr0,1/p(t) - Nlcrell(x, t, A)

(1.1)

which has a positive factor, due to excitation from the ground level, and a negative component, due to the emission from the upper state. Here, a01 is the absorption cross section and O"e is the emission cross section. Cross sections have units of cm 2, time has units of seconds, the populations have units of -1 molecules cm -3, and the intensities have units of photons cm-2s .

6


ao.1

ae

No Figure 1.2 Simpletwo-levelenergy systemincluding a ground level and an excited (upper) level.

The pump intensity Ip(t), undergoes absorption due to its interaction with a molecular population No, a process that is described by the equation

(l/c) ( OIp ( t) / Ot) = - Uoao,11p ( t)

(1.2)

where c is the speed of light. The process of emission is described by the time evolution of the intensity It(x, t, )~) given by

(1/c)(OZl(x,t,A)/Ot) + (OZ,(x,t,A)/Ox) - (Nitre- Nocr~,l)Zl(x,t,)~ ) (1.3) In the steady state this equation reduces to

(OI,(x,A)/Ox) ~ (Nitre- NoJo,1)It(x,A)

(1.4)

which can be integrated to yield

Ii(x, )~) - It(O, ,~)e (N'ae-N~

(1.5)

Thus, if Nitre > NoJo,1, the intensity increases exponentially and there is amplification that corresponds to laserlike emission. Exponential terms such as that in Eq. (1.5) are referred to as the gain.

1.3.2 DYNAMICS OF THE MULTIPLE-LEVEL SYSTEM Here, the rate equation approach is used to describe in some detail the excitation dynamics in a multiple-level energy system, relevant to a wellknown tunable molecular laser known as the dye laser. This approach applies to laser dye gain media either in the liquid or the solid state. The literature on rate equations for dye lasers is fairly extensive and it includes the works of Ganiel etal. (1975), Teschke etal. (1976), Penzkofer and


7

Falkenstein (1978), Dujardin and Flamant (1978), Munz and Haag (1980), Haag etal. (1983), Nair and Dasgupta (1985), and Jensen (1991). Laser dye molecules are rather large, with molecular weights ranging from ~ 175 to ~ 8 3 0 u . An energy level diagram for a laser dye molecule is depicted in Fig. 1.3. Usually, three electronic states are considered, So, S1, and $2, in addition to two triplet states, T1 and T2, which are detrimental to laser emission. Laser emission takes place due to S1 ~ So transitions. Each electronic state contains a large number of overlapping vibrationalrotational levels. This plethora of closely lying vibrational-rotational levels is what gives origin to the broadband gain and to the intrinsic tunability of dye lasers. This is because E = hu, where u is frequency. Thus, a AE implies a Au, which also means a change in the wavelength domain, or AA.

i,

s2

v

N2,0 m

0"1,2

$1

~

7"2,1

N2,o

I'

A N1,0

T2

/

_

u

T

0"1,2

\

\

\ks, T \

o'0,1

Oo!1

T/

0"1,2

71

\

Tl,o /

/

/

N1 ,o

/ 7-T,S

So

I

N0,0

/

___LJ

Figure 1.3 Energy level diagram corresponding to a laser dye molecule. It includes three electronic levels (So, $1, and $2) and two triplet levels (T1 and T2). Each electronic level contains a large number of vibrational and rotational levels. Laser emission takes place from S] to So. [Reprinted from Duarte (1995a), copyright 1995, with permission from Elsevier].

8


In reference to the energy level diagram of Fig. 1.3, and considering only vibrational manifolds at each electronic state, a set of rate equations for transverse excitations was written by Duarte (1995a): m

m

m

m

(1.6)

N-ZENs,v+ZZNr,v S=0 v=0 m

T=I v=0 m

(ONl o/Ot) ~ Z No,vffO, lv,olp(t) + E v=0

v(7o,1 1~,oII (x,t,)%)

v=0

- NI,o

m

q- (N2,0/7-2, 1)

m

(1.7)

O'l,2o,~Ip(t) + EO'eo,v[l(x,t,)% ) v--0

v=0

-Jr-~-~ll,2ovll(X,t,)kv),

Jr- (ks, T +7"-1)1,0

)

v=O

(ONTl,o/Ot) ~ NI,oks, T -- (NTI,o/TT,S) -- NT,,o

m

o~2ovIP(t),

-Jr-E

v=0

(1/c)(0Ip(t)/0t) ~ -

) o. 1Tl,2o,vII (x,

t, ,~v)

(1.8)

v=0

( No,o

O'0,1O,v+ N1,0

v--0 m

+Nv', ~E

O'l,2o,v v=0

)

(1.9)

a r1,2O,v Ie(t)

v--O m

(1/c)(OIl(x, t, A)/Ot) + (OIl(x, t, A)/Ox) ~ Nl,o E aeo,vIl(x, t, Av) v:0 m

-

ENo,vJo, lv,olt(x,t,~v) v=0

(1.10)

m

- Ul,o y ~ d 1,2o,vI,(~, t, ~vl

--

v=0 m NTI,o E v=0

O'Tl 1,2o,vll ( X , t, )%)

m

I,(~, t, ~1 = Z i,(x, t, ~vl v=0

(1.11)


9

(1.12)

I i ( x , t , A ) = I~-(x,t,A) + I T ( x , t , A )

Here, Ip(t) is the intensity of the pump laser beam and Ii(x, t, A) is the laser emission from the gain medium. In this notation, as depicted in Fig. 1.3, the subscripts in the populations designate the electronic state and corresponding vibrational level, so Ns, v refers to the population of the S electronic state at the u vibrational level. The absorption cross sections are designated by a subscript S",SPv,,,v, that designates the electronic transition S" ~ S ~ and the vibrational transition u" ---+u~. The subscript of the emission cross section is is designated by ee,v,,. The broadband nature of the emission is a consequence of the involvement of the vibrational manifold of the ground electronic state, represented by the summation terms of Eqs. (1.9), (1.10), and (1.11). The usual approach to solving an equation system as described here is numerical. Since the gain medium exhibits homogeneous broadening, the introduction of intracavity frequency selective optics (see Chapter 7) enables all the molecules to contribute efficiently to narrow-linewidth emission. A simplified set of equations can be obtained by replacing the vibrational manifolds by single energy levels and by neglecting a number of mechanisms including spontaneous decay from $2 and absorption of the pump laser by T1. Thus, Eqs. (1.6)-(1.10) reduce to (1.13)

U = No + N1 + U r

(OXl/Ot)~No(yo,llp(t) @ ( N0(9"'0,1- Xl(Ye - Nl(Tll,2)II(x,t, )k) (1.14) - N1 ( k s , r + T 1,0 -1 )

\

(ONr /Ot) - N l k s , v - Nr~-T, ls -- NvcrT!zll(x, t, A)

(1.15)

(1/c)(OIp(t)/Ot) = - (N0o0,1 + Nl~rl,Z)Ip(t)

(1.16)

(1/c)(OIl(x,t,A)/Ot) + (OII(x,t,A)/Ox) -- (Nlffe - N0ff/,1

-- N1Jl,2 (1.17)

--Nr rlTt2) It(x, t, A) This set of equations is similar to the set of equations considered by Teschke et al. (1976). This type of equation has been applied to simulate numerically the behavior of the output intensity and the gain as a function of the laserpump intensity and to optimize laser performance. Relevant cross sections and excitation rates are given in Tables 1.1 and 1.2.

T u n a b l e Laser Optics

10

Table 1.1 Transition Cross Sections for Rhodamine 6G Symbol (70,1 (70,1 O'0,1

(70,1 (70,1 (71,2 (71,2 (Te (7e (7e

do,1 O-I 0,1 (7l 1,2 (7T 1,2

(7Tl 1,2 (TTl

1,2

Source:

Cross section ( c m 2) 0.34 • 0.34 x 1.66 x 2.86 x 4.5 x ~0.4 x

10 -16 10 -16 10 -16

10- 1 6 10 -16 10 -16

0.4 • 10 -16

1.86 x 10 -16 1.32 x 10-16 1.3 x 10 -16 < 1 . 0 X 10 -17

1.0 1.0 1.0 6.0

A

(nm)

308 337 510 514.5 530 510 530 572 590 600 580

x 10 -17

600 600 530 590

4.0 X 10 -17

600

x 10 -19 x 10 -17 x 10 -17

Reference Peterson (1979) Peterson (1979) Hargrove and Kan (1980) Peterson (1979) Everett (1991) Hammond (1979) Hillman (1990) Hargrove and Kan (1980) Peterson (1979) Everett (1991) Hillman (1990) Everett (1991) Everett ( 1991) Everett (1991 ) Peterson (1979) Everett (1991)

Duarte (1995a).

Table 1.2 Transition Rates and Decay Times for Rhodamine 6G Symbol

Rate (s -1)

ks, v ks, v

2.0 x 107 3.4 x 106 8.2 x 106

kS, T 7-v,s 7-T,S 7-T,S 7-1,0 7-1,0

2.5 x 10- 7 1.1 X 10 -7 0.5 X 10-7 4.8 x 10-9 3.5 x 10-9 ~1.0 x 10-12

7-2,1

Source:

Decay time (s)

Reference Everett (1991) Webb etal. (1970) Tuccio and Strome Webb et al. (1970) Tuccio and Strome Everett (1991) Tuccio and Strome Everett (1991) Hargrove and Kan

(1972) (1972) (1972) (1980)

Duarte (1995a).

For long-pulse or continuous-wave (CW) excitation, the time derivatives approach zero and Eqs. (1.14)-(1.17) reduce to (1.18)


11

Nlks,r - NTTT, lS + NT0.I,T12II(x, )~)

(1.19)

N00.0,1 - -N10.1,2

(1.20)

Oil(x, A)/Ox - (Nl0.e - N00.~,I - N,0.11,2 - NT0.1,TI2)II(x, A)

(1.21)

From these equations some characteristic features of CW dye lasers become apparent. For example, as indicated by Dienes and Yankelevich (1998), from Eq. (1.18) just below threshold, that is, h(x,A) ~ 0,

Ip ~ 0.-1 0,1 (ks, T + 7-1ol)(N1/No)

(1.22)

which means that to approach population inversion using rhodamine 6G under visible laser excitation, pump intensities exceeding 1022 photons cm -2 s -1 are necessary. A problem unique to long-pulse and CW dye lasers is intersystem crossing from N1 into NT. Thus, researchers use triplet-level quenchers, such as O2 and C8H8 (see, for example, Duarte, 1990b), to neutralize the effect of that level. Under those circumstances, from Eq. (1.21), the gain factor can be written as

g__ ( g l ( o . e _ o.1,2)l _ go0./,1)t

(1.23)

From this equation it can be deduced that amplification can occur, in the absence of triplet losses, when the ratio of the populations becomes

( N1/ No ) > a t0,1/( O'e -- 0-z1,2)

(1.24)

From the values of the cross sections listed in Table 1.1, this ratio is approximately 0.1.

1.3.3 TRANSITION PROBABILITIES AND CROSS SECTIONS The dynamics described with the classical approach of rate equations depends on the cross sections, which are measured experimentally and listed here in Table 1.1. The origin of these cross sections, however, is not classical. Their origin is quantum mechanical. Here, the quantum mechanical probability, for a two-level transition, is introduced and its relation to the cross section of the transition outlined. The style adopted here follows the treatment given to this problem by Feynman etal. (1965), which uses Dirac notation. An introduction to Dirac notation is given in Chapter 2.


12

This approach is based on the basic principles of quantum mechanics, described by (~bl~)- Z (~b[/)(j[~) J

(1.25)

and (1.26) For j = 1,2, Eq. (1.25) leads to (qSl~) = (012)(21r + (4~11)(11~)

(1.27)

which can be expressed as (~1r - (~12)C2 + (~bll)C1

(1.28)

C1 - (1 ]~)

(1.29)

C2 = (21~)

(1.30)

where

and

Here, the amplitudes change as a function of time according to the Hamiltonian 2

ih(dCj/dt) -- ~

HjkCk

(1.31)

k

Now, Feynman et al. (1965) define new amplitudes CI and Czi as linear combinations of C1 and C2. However, since

(IIlII) - (II[1)(1111) + (II]2)(2111)

(1.32)

must equal unity, the normalization factor 2 -1/2 is introduced in the definitions of the new amplitudes: 1

Cn = 2-2(C1 + C2)

(1.33)

1 CI

--

2-2(C1 - C2)

(1.34)

For a molecule under the influence of an electric field E, the components of the Hamiltonian are Hll -- E0 + #g~

(1.35)

H22 - E0 - #g~

(1.36)

-A

(1.37)

H12 - H21 -

13


where E = ~o(ei~t-k-e -i~t)

(1.38)

and # corresponds to the electric dipole moment. Expanding the Hamiltonian given in Eq. (1.31) and then subtracting and adding yields i h ( d C i / d t ) = (Eo + A ) C I + # ~ C I I

(1.39)

i h ( d C i i / d t ) = (Eo - A ) C I I + # ~ C I

(1.40)

Assuming a small electric field, solutions are of the form CI = D i e -iEI@t

(1.41)

Cll - Dlle -iEn/ht

(1.42)

EI = Eo + A

(1.43)

E I I = Eo - A

(1.44)

where

and

Hence, neglecting the term (co + coo), because it oscillates too rapidly to contribute to the average value of the rate of change of O I and DII, the following expressions, for D1 and DII, are found: ih(dDi / d t ) -- #'~oDiie -i(~-~~ ih(dDii / d t ) -- Iz'~oDie i(~-~~

(1.45) (1.46)

If at t = 0, DI '~ 1, then integration of Eq. (1.46) yields (Feynman et al., 1965) ]D//I 2 - ( # ~ o T / h ) 2 s i n 2 ( ( ~ - coo)(T/2))/((co - coo)(T/2)) 2

(1.47)

which is the probability for the transition I ~ H. It can be further shown that IDiI 2 -

[D.I 2

(1.48)

which means that the probability for emission is equal to the probability for absorption. This result is central to the theory of absorption and radiation of light by atoms and molecules.


14

Integrating over the sharp resonance with a linewith A~o, using J -- 2eocg~ and replacing # by 3-1/2# (Sargent etal., 1974), the expression for the probability of the transition becomes

IDiiI 2 - (47r2/a)(#2/47reoch2)(oC( co)/A )T

(1.49)

where # is the dipole moment in units of cm, (1/47re0) is in units of Nm2C -2, and J(~c0) is the intensity in units of j s - l m -2. It follows that an expression for the cross section of the transition can be written as

cr = (47rz/ 3 ) (1/4rceoch ) (co/ Aco) # 2

(1.50)

in units of m 2. As indicated by Feynman et al. (1965) for a simple atomic or molecular system, the dipole moment can be calculated from the definition #mn~ =

(mlHIn} = Hmn

(1.51)

where Hmn is the matrix element of the Hamiltonian for a weak electric field. For a simple diatomic molecule, such as I2, the dependence of this matrix element on the Franck-Condon factor (q~,#,,) and the square of the transition moment (IRe 2) is described by Chutjian and James (1969). For the optically pumped I2 lasers, Byer et al. (1972) wrote an expression for the gain of vibrational-rotational transitions of the form

g = aNL

(1.52)

(4rc2/3)(1/4rreoch)(co/Aw)lRelZqv,,v,,(Sj,,/(2J" + 1))NL

(1.53)

or more specifically

g-

where S,, is known as the line strength and J" identifies a specific rotational level. In practice, however, cross sections are mostly determined experimentally, as in the case of those listed in Table 1.1.

1.4 LASER R E S O N A T O R S A N D LASER CAVITIES A basic laser is composed of a gain medium, a mechanism to excite that medium, and an optical resonator and/or optical cavity. These optical resonators and optical cavities, known as laser resonators and laser cavities, are the optical systems that reflect radiation back to the gain medium and determine the amount of radiation to be emitted by the laser. In this section, a brief introduction to laser resonators is provided. In further chapters, this subject is considered in more detail. The most basic resonator, regardless of the method of excitation, is that composed of two mirrors aligned along a single optical axis, as depicted in


15

Fig. 1.1. In this flat-mirror resonator, one of the mirrors is M 00% reflective at the wavelength or wavelengths of interest and the other mirror is partially reflective. The amount of reflectivity depends on the characteristics of the gain medium. The optimum reflectivity for the output coupler is often determined empirically. For a low-gain laser medium this reflectivity can approach 99%, whereas for a high-gain laser medium, the reflectivity can be as low as 20%. In Fig. 1.1 the gain medium is depicted with its output windows at an angle relative to the optical axis. If the angle of incidence of the laser emission on the windows is the Brewster angle, then the emission will be highly linearly polarized. If the windows are oriented as depicted in Fig. 1.1, then the laser emission will be polarized parallel to the plane of incidence. An alternative to the flat-mirror approach is to use a pair of optically matched concave mirrors. Transverse and longitudinal excitation geometries are depicted in Fig. 1.4.

Laser Pumping Geometries

Figure 1.4 (a) Transverse laser excitation. (b) Transverse double-laser excitation. (c) Longitudinal laser excitation. [Reprinted from Duarte (1990a), copyright 1990, with permission from Elsevier].

16


Figure 1.5 (a) Grating-mirror resonator and (b) grating-mirror resonator incorporating an intracavity etalon. [Reprinted from Duarte (1990a), copyright 1990, with permission from Elsevier].

Further, in some resonators the back mirror can be replaced by a diffraction grating, as shown in Fig. 1.5. This is often the case in tunable lasers. These resonators might incorporate intracavity frequency-selective optical elements, such as Fabry-Perot etalons (Fig. 1.5b), to narrow the emission linewidth. They can also include intracavity beam expanders to protect optics from optical damage and to be utilized in linewidth narrowing techniques. Resonators that yield highly coherent, or narrow-linewidth, emission are often called oscillators and are considered in detail in Chapter 7. The transverse-mode structure in these resonators is approximately determined by the ratio (Siegman, 1986) NF -- (wZ/LA)

(1.54)

known as the Fresnel number. Here, w is the beam waist at the gain region, L is the length of the cavity, and A is the wavelength of emission. The lower this number, the better the beam quality of the emission or the closer it will be to a single transverse mode, designated by TEM00. A TEM00 is a clean beam with no spatial structure on it, as shown in Fig. 1.6, and is generally round with a near-Gaussian intensity profile in the spatial domain. Thus, long lasers with relatively narrow beam waists tend to yield single-transversemode emission. As it will be examined in Chapters 4 and 7, an important part of laser cavity design consists in optimizing the dimensions of the beam waist to the cavity length to obtain TEM00 emission and low beam divergences in compact configurations.


17

Figure 1.6 Crosssection of a TEM00laser beam from a high-power narrow-linewidth dispersive laser oscillator. The spatial intensityprofile of this beam is near-Gaussian. [Reprinted from Duarte (1995b), copyright 1995, with permission from Elsevier].

An additional class of linear laser resonators are the unstable resonators. These cavities depart from the flat-mirror design and incorporate curved mirrors, as depicted in Fig. 1.7. These mirror configurations are adopted from the field of reflective telescopes. A widely used design is a variation of the Newtonian telescope known as the Cassegrainian telescope. In this configuration the two mirrors have a high reflectivity. Advantages of unstable resonators include the use of large gain-medium volumes and good transverse-mode discrimination. This topic will be considered further in the context of transfer ray matrices in Chapter 6. For a detailed treatment on the subject of unstable resonators the reader should refer to Siegman (1986). A further class of cavities includes linear and ring laser resonators (Fig. 1.8), developed for CW dye lasers (Hollberg, 1990) and later applied to the generation of ultrashort pulses (Diels, 1990; Diels and Rudolph, 1996). A straightforward unidirectional ring resonator with an 8 shape is illustrated in Fig. 1.8b. In these cavities the oscillation is in the form of a traveling wave that avoids the effect of spatial hole burning that causes the laser to oscillate in more than one longitudinal mode. Linear and ring


18

Figure 1.7 Basicunstable resonator laser cavity. resonators incorporating saturable absorbers are depicted in the ultrashort pulse cavity configurations of Fig. 1.9. In the ring laser, a collision between two counterpropagating pulses occurs at the saturable absorber. This collision causes the two pulses to interfere, thus creating a transient grating that shortens the emission pulse. This effect is known as colliding-pulse-mode (CPM) locking (Fork etal., 1981). The prisms in this cavity are deployed to provide negative dispersion and thus help in pulse compression, as will be described in Chapter 4.

Dye Jet

Dispersive and/or FSE M3

CW Laser v ~ M 1 ~Mp Pump

(a) Dispersive and/or FSE

UDD

M3Lj

M4

CW Laser .. I - / ~ ~ ' ~ " - . . . ~ D _ _ Pum. Dye~jet

.""~ Mp

(b)

Figure 1.8 (a) Linear and (b) unidirectional 8-shape ring dye laser cavities. [Reprinted from Hollberg (1990), copyright 1990, with permission from Elsevier].

19


Pump laser

M1

M4

medium

le absorber Pulse compressor

/

6.0

u

rr

4.0 2.0-

.0 (b)

-0.0100 -0.0050 0.0000 0.0050 0.0100 Screen Axial Distance (meters)

Figure 2.9 (a) Measured interferogram resulting from the interaction of coherent laser emission at A - 632.82 nm and 100 slits, 30 gm wide and separated by 30 gm. The j-to-x distance is 75cm. (b) Corresponding theoretical interefogram from Eq. (2.13). [Reprinted from Duarte (1993), copyright 1993, with permission from Elsevier].

etal., 1965b). His point is well taken. In the discussion related to Fig. 2.9 and its variants, reference was only made to interference. However, what we really have is interference in three diffraction orders. That is, the 0th, or central, order and the +1, or secondary, orders. In other words, there is an interference pattern associated with each diffraction order. Physically, however, this is the same phenomenon. The interaction of coherent

34

Tunable Laser Optics 14.0

12.0

10.0

t~

r

8.0

c

Q 6.0

0 4.0

1\

2.0

0.0

-ls.o

r

'-~d.o

'-slo

'

6

'

s

'~c;o

' ~g.o

Screen Axial Distance (meters) x 10 -a Figure 2.10 Theoretical interfererometric/diffraction distribution using a _, 5 . 0 t/} tO

E 4.0.>_ tr

3.02.01.00.0 -3.0

__j I

I

I

I

I

-1 o oo lo 2o Screen Axial Distance (meters) x 10-3

-2o

3.0

Figure 2.11 Theoretical interferometric distribution produced by a 3 mm aperture illuminated at A = 632.82 nm. The j to x distance is 10 cm.

1.0.

0.8

"~ c--

06'

=> rc

0.4-

0.2

|

-12.5

I

-7.5

|

|

-2.5

0

2.5

|

|

7.5

|

|

12.5

Screen Axial Distance (meters) x 10-3

Figure 2.12 Theoretical interferometric distribution incorporating diffraction-edge effects in the illumination. In this calculation the slits in the array are 30Hm wide and separated by 30~tm, N - 1 0 0 , and the j-to-x distance is 75cm. The aperture-grating distance is 10cm. [Reprinted from Duarte (1993), copyright 1993, with permission from Elsevier].

36


Figure 2.13 Emergence of secondary diffraction (+1) orders as the distance j to x is increased. (a) At a grating-to-screen distance of 5cm, the interferometric distribution is mainly part of a single order. At the boundaries there is an incipient indication of emerging orders. (b) As the distance is increased to 10cm, the presence of the emerging (+1) orders is more visible. (c) At a distance of 25cm, the emerging (+1) orders give rise to an overall distribution with clear "shoulders." (d) At a distance of 75cm, the - 1 , 0, and +1 diffraction orders are clearly established. Notice the increase in the width of the distribution as the j-to-x distance increases from 5 to 75 cm. Slit width is 30 lam, slits are separated by 30 ~tm, N = 100, and A = 632.8 nm.

Dirac Optics

37

Figure 2.13 (Continued).

The intimate relation between interference and diffraction has its origin in the interference equation itself:

N

N

)


38

for it is the COS(~-~m- ~-~j) term that gives rise to the different diffraction orders. From the geometry of Fig. 2.7 we can write sin ff~m -- (~m + (dm/2))/Lm

(2.24)

And for the condition a >> din, we have ILm-k-Lm-ll ,-~ 2Lm. Then using Eqs. (2.21) and (2.24) we have

]Lm - Lm-1] ~ dm sin ff~m

(2.25)

~)m

(2.26)

]lm -- lm-l l ~ dm sin

where Om and (~)m are the angles of incidence and diffraction, respectively. Given that maxima occur at ([lm -- lm-1 Inl -+-ILm - Lm-1 ]nz)ZTr/Av = MTr

(2.27)

then, using Eqs. (2.25) and (2.26), we get

dm(nl sin Om + n2 sin ff~m)(27r/Av) = MTr

(2.28)

where M = 0, 2, 4, 6 , . . . . For nl = n2 we have A = Av, and this equation reduces to the well-known diffraction grating equation dm(sin O m + sin ff~m) = mA

(2.29)

where m = 0, 1, 2, 3 , . . . are the various diffraction orders.

2.4 R E F R A C T I O N An additional fundamental phenomenon in optics is refraction. This is the change in the geometrical path of a beam of light due to transmission from the original medium of propagation to a second medium with a different refractive index. For example, refraction is the bending of a ray of light due to propagation in a glass or crystalline prism. If in the diffraction grating equation dm is made very small relative to a given A, diffraction ceases to occur and the only solution that can be found is for m = 0. That is, under these conditions a grating made of grooves coated on a transparent substrate, such as optical glass, does not diffract and exhibits the refraction properties of the glass. For example, since the maximum value of ( sin ~:)m -+- sin ff~m) is 2 for a 5000-1ines/mm transmission grating, no diffraction can be observed for the visible spectrum. Hence for the condition dm _ h/2

(3.23)

an expression that can be derived using the probability density of the wave function (Feynman et al., 1965).

3.1.3 ALTERNATIVE VERSIONS OF THE UNCERTAINTY PRINCIPLE In addition to Ap Ax ~ h, the uncertainty principle can be expressed in various useful versions. Assuming an independent derivation and using p - hk, it can be expressed in its wavelength-spatial form, A,,~ AN ~ ~2

(3.24)

which can also be stated in its frequency-spatial version, Au Ax ~ c

(3.25)

Using E --mc 2, the uncertainty principle can also be written in its energytime form, AE At ~ h

(3.26)

which, using the quantum energy equation E = hv, can be transformed to its frequency-time version, Au At ~ 1

(3.27)

This result is very important in the area of pulsed lasers and implies that a laser emitting a pulse of a given duration At has a minimum spectral linewidth Au. It also implies that by measuring the width of spectral emission, the duration of that pulse can be determined. The time A t ~ 1/Au is also known as the coherence time. From this time the coherence length can be defined as A x ~ c/Au, which is an alternative form of Eq. (3.25).

3.2 A P P L I C A T I O N S PRINCIPLE

OF THE UNCERTAINTY IN OPTICS

In this section some useful applications of the uncertainty principle in beam propagation and intracavity optics are considered.


50

3.2.1 BEAM DIVERGENCE If the uncertainty principle is assumed to be derived from independent and rigorous methods, then it can be used to derive some useful identities in optics. For example, starting from

Ap A x ~ h the application of p - hk yields Ak Ax ~ 27r

(3.28)

which leads directly to AA ~ A2 / A x For a diffraction-limited beam traveling in the z direction, kx = k sin 0. For a very small angle 0 we can write

kx ,.~ kO

(3.29)

Using A k x A x ~ 27r it is readily seen that the beam has an angular divergence given by

AO ~ A / A x

(3.30)

This equation indicates that the angular spread of a propagating beam of wavelength A is inversely proportional to its original width. That is, narrower beams exhibit a larger angular spread or divergence (see Fig. 3.3).

(a)

-----_.____

(b)

Figure 3.3 Beamdivergence for two different apertures at wavelength A. (a) An expanded laser beam is incident on a microhole of diameter 2w. (b) The same laser beam is incident on a microhole of diameter 4w.

The Uncertainty Principle in Optics

51

(a)

(b)

Figure 3.4 Beam divergence for (a) Al = 450nm and (b) A1 = 650nm. In both cases an expanded laser beam is incident on identical microholes of diameter 2w. Also it states that light of shorter wavelength experiences less beam divergence, which is a well-known experimental fact in laser physics (see Fig. 3.4). Equation (3.30) has the same form as the classical equation for beam divergence (Duarte, 1990), namely, A0-

(A/Trw) (1 + (L~/B) 2 + (AL~/B)2) '/2

(3.31)

where w is the beam waist, L ~ = (Trw2/A) is the Rayleigh length, and A and B are spatial propagation parameters, defined in Chapter 6. For well-chosen experimental conditions, the second and third terms in the parentheses of Eq. (3.31) approach zero, so

AO ,.~ (A/Trw)

(3.32)

and the beam divergence is said to approach its diffraction limit. The equivalence of Eqs. (3.30) and (3.32) is self-evident. The generalized interference equation, Eq. (3.5), was previously used to derive AA ~ AZ/Ax, which subsequently led to an expression for beam divergence. Hence it should also be useful to estimate the diffraction-limited divergence of a coherent beam as it propagates in space. For instance, at A = 632.8 nm, a beam with an original dimension of 2w0 = 200 ~m can be estimated, using, Eq. (3.5), to spread to ~ 1 mm following a propagation of 0.5 m. The same beam waist can be calculated to increase further to ~9.5 mm following propagation of 5 m. This yields a beam divergence of A0 ~ 2 m rad. Alternatively, using Eq. (3.32), one can estimate A0 ~ 2 m rad directly. The beam profiles calculated using the interferometric equation are shown in Fig. 3.5.

52

Tunable Laser Optics (a)

40.0 -

-"~ 30.0 ttl)

(.-

g9 20.0 I1)

rv 10.0 -

0.0 -4.0

-2.0

0.0

2.0

4.0

Screen Axial Distance (m)x 10-3

(b)

40.0 -

~r 3 0 . 0 r B

(D

"~ 20.0 m rr"

10.0 -

0.0 -3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

Screen Axial Distance (m) x 10 -2

Figure 3.5 Beam divergence determined from the generalized interference equation. (a) Beam profile following propagation through a distance of 0.5 m. (b) Beam profile following propagation through a distance of 5 m. Here, w0 = 100 gm and A = 632.8 nm. The divergence is determined by taking the difference at half width, half maximum and dividing by 4.5 m.

3.2.2

BEAM DIVERGENCE AND ASTRONOMY

Consider propagation 95 k m a t A -

of a diffraction-limited laser beam over a distance of

589 n m f o r

guide star a p p l i c a t i o n s in a s t r o n o m y . A m u l t i p l e -


53

prism-grating tunable laser oscillator (Duarte and Piper, 1984; Duarte, 1999) can be designed to provide single-longitudinal-mode oscillation at A - - 5 8 9 n m in a laser with a beam divergence that is nearly diffraction limited. Thus, its divergence is characterized by A0 ~ A/Trw. For a beam waist w ~ 125 lam, the corresponding beam divergence becomes A0 ..~ 1.5 m rad. If the beam of this oscillator were to be propagated over a distance of 95 km it would illuminate a circle nearly 285 m in diameter. This would yield a very weak power density. If, on the other hand, the beam is expanded, without introducing further divergence, by a factor of M ..~ 200, then the new beam divergence becomes A0 ~ A/MTrw. Now, as a consequence of the beam expansion, the beam diameter at the required propagation distance becomes about 1.42 m, which is close to the dimensions needed to achieve the necessary power densities with existing high-power tunable lasers. It should be mentioned that beam magnification factors of about 200 are not difficult to attain. Further, for this type of application the narrow-linewidth oscillator emission is augmented at amplifier stages, where the beam also increases its dimensions, thus reducing the requirements onM. One further application of the beam divergence equation relates to the angular resolution limit of telescopes used in astronomical observations. Reflection telescopes, such as the Newtonian and Cassegrain telescopes, are depicted in Fig. 3.6 and discussed further in Chapter 6. The angular resolution that can be accomplished with these telescopes under ideal conditions is approximately described by the diffraction limit given in Eq. (3.32).

(a)

(b)

Figure 3.6 Reflectiontelescopes used in astronomical observations. (a) Newtonian telescope. (b) Cassegranian telescope.

54


That is, the smallest angular discrimination, or resolution limit, of a telescope with a diameter D = 2w is given by (3.33)

AO ~ 2A/TrD

This equation indicates that the two avenues to increase the angular resolution of a telescope are either to observe at shorter wavelengths or to increase the diameter of the telescope. The latter option has been adopted by optical designers who have built very large telescopes. At A = 500 nm, the angular resolution for a telescope with D - 10 m is A0 ,~ 3.18 x 10-8 rad. For an hypothetical diameter of D -- (1000/70 m the angular resolution at the same wavelength becomes A0 ~ 1.00 x 10.9 rad. In addition to better angular resolutions, large aperture telescopes provide increased signal to noise ratios since the area of collection increases substantially. In the future, this class of telescope might be made available by advances in segmented optics technology.

3.2.3 THE UNCERTAINTY PRINCIPLE AND THE CAVITY LINEWIDTH EQUATION As outlined by Duarte (1992), starting from

[(X[S)[ 2 -- ~

~(rj) 2 if-

2~

q~(rj)

m=j+l

ffJ(rm) coS(am

-- ~')j)

the equation for a diffraction grating, mA = 2d sin O, can be established. Considering two slightly different wavelengths, an expression for the wavelength difference can be written as AA = (2d/m)(sin 01 - sin 02)

(3.34)

For O1 ~ 0 2 ( = 0 ) this equation can be restated as AA ~ (2d/m)AO(1 - ( 3 0 2 / 3 ! ) +

(504/5!)...)

(3.35)

Differentiation of the grating equation leads to (O0/OA) cos 19 = m / 2 d

(3.36)

and substitution into Eq. (3.35) yields /kA ~ A0(00/0A)-I (1 - ( O 2 / 2 ! ) - + - ( O 4 / 4 ! ) . . . ) / c o s O

(3.37)

which reduces to the well-known cavity linewidth equation (Duarte, 1992),


55

m)k ~ mO(OO/O)k) -1

(3.38)

AA ~ A0(V~0) -~

(3.39)

or

where V~0 = (00/0A). This equation has been used extensively to determine the emission linewidth in pulsed narrow-linewidth dispersive laser oscillators (Duarte, 1990). It originates in the generalized N-slit interference equation and incorporates AO, whose value can be determined either from the uncertainty principle or from the interferometric equation itself. This equation is well known in the field of classical spectrometers, where it was introduced using geometrical arguments (Robertson, 1955). In addition to its technical and computational usefulness, Eq. (3.38) and/or Eq. (3.39) illustrate the inherent interdependence between spectral and spatial coherence.

PROBLEMS 1. Calculate the diffraction-limited beam divergence at fullwidth, half maximum ( F W H M ) for (a) a laser beam with a 100-gm radius at A = 590 nm, and (b) a laser beam with a 500-gm radius at A = 590 nm. 2. Repeat the calculations of the first problem for A = 308 nm. Comment. 3. Calculate the dispersive cavity linewidth for a high-power tunable laser yielding a diffraction-limited beam divergence 100gm in radius, at A = 590nm. Assume that an appropriate beam expander illuminates a 3000-1/mm grating deployed in the first order. The grating has a 50-mm length perpendicular to the grooves. 4. (a) For a pulsed laser delivering a 350-MHz laser linewidth, estimate its shortest possible pulse width. (b) For a laser emitting 10-fs pulses, estimate its broadest possible spectral width in nanometers centered around A = 600 nm. 5. Show that Eq. (3.29) can be expressed as Eq. (3.30).

REFERENCES Dirac, P. A. M. (1978). The Principles of Quantum Mechanics, 4th ed. Oxford University Press, London. Duarte, F. J. (1990). Narrow-linewidth pulsed dye laser oscillators. In Dye Laser Principles (Duarte, F. J., and Hillman, L. W., eds.). Academic Press, New York.

56


Duarte, F. J. (1991). Dispersive dye lasers. In High-Power Dye Lasers (Duarte, F. J., ed.) Springer-Verlag, Berlin. Duarte, F. J. (1992). Cavity dispersion equation AA ~ A0(00/0A)-I: a note on its origin. Appl. Opt. 31, 6979-6982. Duarte, F. J. (1997). Interference, diffraction, and refraction, via Dirac's notation. Am. J. Phys. 65, 637-640. Duarte, F. J. (1999). Multiple-prism grating solid-state dye laser oscillator: optimized architecture. Appl. Opt. 38, 6347-6349. Duarte, F. J., and Piper, J. A. (1984). Narrow-linewidth, high-prf copper laser-pumped dyelaser oscillators. Appl. Opt. 23, 1391-1394. Feynman, R. P., Leighton, R. B., and Sands, M. (1965). The Feynman Lectures on Physics, Vol. III. Addison-Wesley, Reading, MA. Haken, H. (1981). Light. North Holland, Amsterdam. Robertson, J. K. (1955). Introduction to Optics: Geometrical and Physical. Van Nostrand, New York.

Chapter 4

The Physics of Multiple-Prism Optics

4.1 I N T R O D U C T I O N Multiple-prism arrays were first introduced by Newton (1704) in his book Opticks. In that visionary volume Newton reported on arrays of nearly isosceles prisms in additive and compensating configurations to control the propagation path and the dispersion of light. Further, he also illustrated slight beam expansion in a single isosceles prism. In his treatise, Newton does provide a written qualitative description of light dispersion in a sequence of prisms, thus laying the foundation for prismatic spectrometers and related instrumentation. In this chapter, a mathematical description of light dispersion in generalized multiple-prism arrays is given. Although the theory was originally developed to quantify the phenomenon of intracavity linewidth narrowing, in multiple-prism-grating tunable laser oscillators (Duarte and Piper, 1982), it is applicable beyond the domain of tunable lasers to optics in general. The theory considers single-pass as well as multiple-pass dispersion in generalized multiple-prism arrays and is useful to both linewidth narrowing and pulse compression. As a pedagogical tool, reduction of the generalized formulae to single-prism calculations are included in addition to a specific design of a particular practical high-magnification multipleprism beam expander. In order to facilitate the flow of information and to maintain the focus on the physics, some of the mathematical steps involved in the derivations are not included. For further details, references to the original work are cited. 57


58

The mathematical notation is consistent with the notation used in the original literature.

4.2 GENERALIZED MULTIPLE-PRISM DISPERSION Multiple-prism arrays are widely used in optics in a variety of applications, such as (a) intracavity beam expanders, in narrow-linewidth tunable laser oscillators, (b) extracavity beam expanders, (c) pulse compressors in ultrafast lasers, and (d) dispersive elements in spectrometers. Even though multiple-prism arrays were first introduced by Newton (1704), a mathematical description of their dispersion had to wait a long time, until their application as intracavity beam expanders in narrow-linewidth tunable lasers (Duarte and Piper, 1982). Generalized multiple-prism arrays are depicted in Fig. 4.1. Considering the mth prism of the arrangements, the angular relations are given by q~l,m -}- q~Z,m ~- Cm -}- Ogm ~)l,m q" ffdZ,rn = OLm

sin 4~l,m =nm sin ~l,m sin ~2,m = n m sin ~2,m

(4.1) (4.2) (4.3) (4.4)

As illustrated in Fig. (4.1), ~)l,m and ~)2,m are the angles of incidence and emergence, and ~l,m and ~2,m are the corresponding angles of refraction, at the mth prism. Differentiating Eqs. (4.3) and (4.4) and using d~bl ,m / dn -- d~b2,m/ dn

(4.5)

the single-pass dispersion following the mth prism is given by (Duarte and Piper, 1982, 1983) ~7)~2,m -- O--~2,m~nm -t- (kl,mk2,m) -1 ( ~ l , m ~ ) ~ n m -Jr-~7Aq~2,(m_l))

(4.6)

where V~ = 0/0A and the following geometrical identities apply:

kl,m -- COS ~)l,m/COSq~l,m ~l,m

= tan q~l,m/nm

(4.7) (4.8) (4.9)

~2,m

= tan

q~2,m/nm

(4.10)

k2,m = cos ~2,m/cos ff)2,m

The kl,m and k2,m factors represent the physical beam expansion experienced at the mth prism by the incidence and the emergence beams, respectively. The generalized single-pass dispersion equation indicates that the cumulative dispersion at the mth prism, namely, Va~z,m, is a function of the geometry of


/

59

__~~/

, ") f~~,m

(a)

- -,/'1,~

7",e,,,

~

(b) .,,

Figure 4.1 Generalized multiple-prism arrays. (a) In additive configuration and (b) in a compensating configuration. [Reprinted from Duarte (1990), copyright 1990, with permission from Elsevier]. the mth prism, the position of the light beam relative to this prism, the material of this prism, and the cumulative dispersion up to the previous prism, ~7A~)Z,(m_l) (Duarte and Piper, 1982, 1983). For an array of r identical isosceles or equilateral prisms deployed symmetrically in an additive configuration so that r = (b2,m, the cumulative dispersion reduces to (Duarte, 1990a) ~7,~q52,r = r~7)~q52,1

(4.11)

It is this simple dispersion equation that applies to the design of multipleprism spectrometers incorporating identical isosceles or equilateral prisms arranged in symmetrical additive configurations. The generalized single-pass dispersion equation ~7Ar m --

O_~2,m~7Anrn.-]-(kl,mk2,m)-1 (~l,rn~TA/'/m

q-- ~7Ar

)

can be restated in a more practical and explicit notation (Duarte, 1989): V)~ t~2,r = Z ( - ' I - 1) ~,~l,m m= 1

+(M1M2) -1

klj =m

k2r

kl,j H k2,j V,Xrtrn

(+l)~2,m m=l

V~l"lm

"=

=1

j=l

(4.12)

60


where M1 - 1zI klj

(4.13)

j=l

M2 - ~ I k2j

(4.14)

j=l

are the respective beam expansion factors. For the important practical case of an array of r right angle prisms, designed for orthogonal beam exit (that is, q~2,m -- ~2,m -- 0), Eq. (4.12) reduces to

VA~b2,r- ~ (-+-l)~{~l,m m= 1

If

in

addition

(OL1 = tY2 = Ol3 . . . . .

the

prisms

klj

)'

~7Anm

(4.15)

=m

have

identical

apex

angles

tYm) and are configured to have the same angle of

incidence (~bl,1 = ~bl,2 = ~bl,3 . . . . . (Duarte, 1985)

~bl,m), then Eq. (4.15) can be written as

(+l)(1/kl,m)m-lvAnm

VA~2,r- tan ~1,1~

(4.16)

m=l

Further, if the angle of incidence for all prisms is Brewster's angle, then the single-pass dispersion reduces to the elegant expression

V~bZ,r- ~

(+l)(1/nm)mV~,nm

(4.17)

m-1

4.2.1 DOUBLE-PASSGENERALIZEDMULTIPLE-PRISM DISPERSION The evaluation of intracavity dispersion in tunable laser oscillators incorporating multiple-prism beam expanders requires the assessment of the double-pass, or return-pass, dispersion. The double-pass dispersion of multiple-prism beam expanders was derived by thinking of the return pass as a mirror image of the first light passage, as illustrated in Fig. 4.2. The returnpass dispersion corresponds to the dispersion experienced by the return light

61

The Physics of Multiple-Prism Optics Symmetric half

I ! I ]

I

el,

Grating or Mirror

~1'1/~.k

~'~'~r-1

? I,~2r

Figure 4.2 Multiple-prismbeam expander geometry in additive configuration and its mirror image. A dispersive analysis through the multiple-prism array and its mirror image is equivalent to a double-pass, or return-pass, analysis. [Reprinted from Duarte and Piper (1982), copyright 1982, with permission from Elsevier]. beam at the first prism. Thus, it is given by O~tl,m/O)~- ~A~ t1,m~ where the prime character indicates return pass (Duarte and Piper, 1982, 1984):

~7A~l,m , ) -1( ~2,m~A?l t t __ O_~tl,m~7~llm_.l..(kt 1,mk2,m m _4_~7Aq~tl,(m+l))

(4.18)

where

kll ,m m COS ffOl,m/COS ! ! ~l,m k~2,m -- COS r ,//)t2,m

(4.19) (4.20)

~ l , m -- tan r 1,m//'/m

(4.21)

o_~2,m - tan r 2,m/nm

(4.22)

Here, ~7lq~],(m+l ) provides the cumulative single-pass multiple-prism dispersion plus the dispersion of the diffraction grating; that is, =

+ V r

(4.23)

where VaOa is the grating dispersion. If the grating is replaced by a mirror, then we simply have the prismatic contribution and VaqStl,(m+l) = Va~bZ,r

(4.24)

Defining ~7Aq~],m --~7(I)p, where the capital phi stands for return pass and P stands for for multiple prism, the explicit version of the generalized

62


double-pass dispersion for a multiple-prism mirror system is given by (Duarte, 1985, 1989)

-1 (-+-l ) ~al,m

VA~p =2M1 M2

klj

m=l

+ 2

=m

kid Hk2d VAnm

(-+-l)~2,m m=l

V~nm

k2j "=

=1

(4.25)

j=l

For the case of r right angle prisms, designed for orthogonal beam exit (that is, q~2,m- ~b2,m-- 0), Eq. (4.25) reduces to

r

C")

VAmp -- 2M1 ~(+l)~_Tfl,m m=l

-1

klj

~7A/"/m

(4.26)

=m

which can also be expressed as (Duarte, 1985)

gTA~p- 2~m=l(+ 1)(j]~ k l j ) = l tan~bl,mV:~nm

(4.27)

If the angle of incidence for all prisms in the array is made equal to the Brewster angle, this equation simplifies further to (Duarte, 1990a) V~,~bF,- 2 ~

(-'}-l)(nm)m-l~7Anm

(4.28)

m=l

4.2.2 MULTIPLERETURN-PAss GENERALIZEDMULTIPLE-PRISM DISPERSION Here we consider a multiple-prism grating or multiple-prism mirror assembly, as illustrated in Fig. 4.3. The light beam enters the first prism of the array; it is then expanded and either diffracted back or reflected back into the multiple-prism array. In a dispersive laser oscillator this process goes forth and back many times, thus giving rise to the concept of intracavity double pass, or intracavity return pass. For the first return pass, toward the first prism in the array, the dispersion is given by Eq. (4.18). If N denotes the


63

Multiple-prismgrating assembly used to perform a dispersive multiple-return-pass analysis. [Reprinted from Duarte and Piper (1982), copyright 1982, with permission from Elsevier].

Figure 4.3

number of passes toward the grating or the reflecting element and 2N denotes the number of return passes toward the first prism in the sequence, we have (VA~)Z,m)N-- Q-~2,mVArtm-[-(kl,mkz,m)-1 (~l,mVArtm -+-(VA~Z,(m-1))N)(4.29) (V,X~I,m)2N -- ~l,mV~nm + (k~l,mk2,m) -Jr- (~7~l,(m+l)) 2N) (4.30) ' ' ' -l( ~2,mV,xnm ' ' For the first prism of the array, (Va~b2,(m-1))u (with N = 3, 5, 7,...) in Eq. (4.29) is replaced by (VaqS'l,1)2u (with N - 1, 2, 3,...). Likewise for the last prism of the assembly, (VaqS'l,(m+l))ZU (with N - 1, 2, 3,...) in Eq. (4.30) is replaced by (VaOa + (Vaq52,r)u) (with N = 1, 3, 5,...). In the case where the grating is replaced by a mirror, this expression becomes simply (Vaq52,r)U (with N = 1, 3, 5,...). Thus, the multiple return-pass dispersion for a multiple-prism grating assembly is given by (Duarte and Piper, 1984) (V:~0)R = (RMV~,Oa + R V ~ p )

(4.31)

where R is the number of return passes. This equation illustrates the very important fact that in the return-pass dispersion of a multiple-prism grating assembly, the dispersion of a grating is multiplied by the factor RM, where M is the overall beam magnification of the multiple-prism beam expander. Once again, if the grating is replaced by a mirror, that is, X7~OG = 0, the dispersion reduces to (V~0)R = RX7A~e

(4.32)

which implies that the multiple-prism intracavity dispersion increases linearly as a function of R.


64

4.2.3 SINGLE-PRISMEQUATIONS Using Eq. (4.6) for m = 1 yields

V~c/)2,1- ~d~Z,lV~nl + (kl,lkz,1)-l(~l,lV:~nl)

(4.33)

which can also be expressed as V~2,1 -- ((sin ~P2,1/cos 4~2,1) + (cos ~2,1/cos q52,1)tan ~Pl,~)V~n

(4.34)

Division by V~n yields the result given for V, qS2,1 by Born and Wolf (1999). For orthogonal beam exit (q52,1 ,-~ ~P2,1 ~ 0) Eq. (4.34) simplifies to ~7Aq~2,1 ~ tan ~31,1V~n

(4.35)

which is the result given by Wyatt (1978). For a single prism designed for orthogonal beam exit and deployed at Brewster's angle of incidence, Eq. (4.17) becomes ~7A~)2,1 = (1 In) V:~n

(4.36)

a result that also follows from Eq. (4.35). For a double-pass analysis, for m = 1, Eq. (4.18), becomes

v~r

' YAH1--~-(k' l ,1 k'2,1 )-I(Q_~, 2,1 VAg/I-I-VA~2,1) 1 ~ ~-{~1,1

(4.37)

which for orthogonal beam exit reduces to Vamp = 2(k1,1 tan ~)I,1)VAn

(4.38)

For incidence at the Brewster angle this equation simplifies to V ~ p = 2V~n

(4.39)

Reduction from Eq. (4.38) to Eq. (4.39) implies that at the Brewster angle of incidence the beam magnification kl,1 = n (4.40) tan ~Pl,1 = (1/n) (4.41)

4.3 MULTIPLE-PRISM DISPERSION A N D LINEWIDTH NARROWING In Chapter 3, the cavity linewidth equation was derived from interferometric principles as (Duarte, 1992) z~-

A0(V~0) -1

(4.42)

65


where A0 is the beam divergence and V~0 is the overall intracavity dispersion. The message from this equation is that for narrow-linewidth emission we need to minimize the beam divergence and to increase the intracavity dispersion. The multiple return-pass cavity linewidth for dispersive oscillator configurations, as illustrated in Fig. 4.4, is given by (Duarte, 200 l a) zSA - AOR(MRVaOa + R~7A(I)p)-1

(4.43)

and the multiple return-pass beam divergence (see Chapter 6) is given by AOR

-

-

(A/Tl'w)(1

q-(L,~/BR)2 +

(ARL~/BR) 2) 1/2

(4.44)

Here, AR and BR are multiple-pass propagation matrix coefficients and L ~ = 7rwZ/A (where w is the beam waist) is known as the Rayleigh length. For an optimized multiple-prism grating laser oscillator, A0R approaches its diffraction limit following a few return passes.

4.3.1 THE MECHANICS OF LINEWIDTH NARROWING IN OPTICALLY PUMPED PULSED LASER OSCILLATORS The factor R in the multiple-return-pass linewidth equation is related to the total intracavity transit time from the beginning of the excitation pulse to the onset of laser oscillation. Thus, R is determined experimentally by measuring the time delay &- between the leading edge of the pump pulse and the leading edge of the laser emission pulse. This perspective is consistent with the mechanics of recording single-pulse interferograms, which provide a laser linewidth at the early stages of oscillation that is broader than at subsequent times of emission. That is, although the linewidth narrowing continues throughout the duration of the laser emission, the time-integrated recording process (using photographic means, for instance) yields information as broad as that recorded at the onset of the laser oscillation. Once &- has been measured, the number of return intracavity passes is given by R = &-/At

(4.45)

where At is the time taken by the emission to cover twice the cavity length (At = 2L/c), so R = &-(c/2L)

(4.46)

In summary, from the beginning of the pulsed laser excitation the R factor contributes to reduce both the beam divergence and the laser linewidth. As indicated by Eq. (4.44) the beam divergence can decrease only toward its

66


Figure 4.4 (a) MPL grating laser oscillator incorporating a (+, + , + ,-) multiple-prism configuration and (b) a ( + , - , + ,-) multiple-prism configuration. (c) HMPGI grating laser oscillator incorporating a compensating double-prism configuration. In this oscillator, the refraction angle identified as 0' corresponds to ft. [Reprinted from Duarte (1995c), copyright 1995, with permission from Elsevier].


67

diffraction limit, while the laser linewidth decreases linearly as a function of R. In turn, R is a finite number that can be as low as R ~ 3 for high-power short-pulse excitation. It is also necessary to remember that Eq. (4.43) is the dispersive cavity linewidth equation; that is, it quantifies the ability of a dispersive oscillator cavity to narrow the frequency transmission window of the cavity. Once the dispersion linewidth limits emission to a single longitudimal mode, the actual linewidth of that mode is determined by the dynamics of the laser. For further information on this subject the reader should refer to Duarte and Piper (1984) and Duarte (2001a).

4.3.2 DESIGN OF ZERO-DISPERSION MULTIPLE-PRISM BEAM EXPANDERS In practice, the dispersion of the grating multiplied by the beam expansion, that is, M(V~O6), amply dominates the overall intracavity dispersion. Thus, it is desirable to remove the dispersion component originating from the multiple-prism beam expander so that

z2~ ~ AOR(MRV~OG) -1

(4.47)

In such designs the tuning characteristics of the laser are those of the grating around a specific wavelength. To illustrate the design of a quasi-achromatic multiple-prism beam expander, a case of practical interest, with M ,~ 100 and a four-prism expander deployed in a compensating configuration similar to that outlined in Fig. 4.4, is considered. The compensating configuration selected is (+, +, + , - ) ; that is, the additive dispersion of the first three prisms is subtracted by the fourth prism, thus yielding zero overall dispersion. Multiple-prism beam expanders deployed intracavity in tunable lasers employ right angle prisms designed for orthogonal beam exit; that is ~2,m ~ ~2,m ~ 0. Thus, we must have Va~p = 0 in Eq. (4.27) so that (Duarte, 1985) (kl,1 + kl,lkl,2 q- kl,lkl,2kl,3) tan ~1,1 -- (kl,lkl,2kl,3)kl,4 tan r

(4.48)

where t a n ~31,1 ~ tan ~1,2 = t a n ~1,3

(4.49)

A particular wavelength of interest in tunable lasers is A = 590 nm. At this wavelenth, for a material such as optical crown glass, n - 1.5167. For M ~ 100 we select kl,1 -- kl,2 = kl,3 -- 4

68


This means that for the first three prisms, ~1,1 = q)l,2 ~--- q~l,3 =

79.01~

and

1/)1,1 - - @1,2 •

@1,3 - -

40.330

which also becomes the apex angle of the first three prisms. Using Eq. (4.48) it is found that for the fourth prism, 4~1,4= 59.39 ~ @1,4 = 34.57 ~ and kl,4--1.62. This yields an overall beam magnification factor of M = 103.48. Normally the beam waist w in a well-designed high-power multiple-prism grating laser oscillator is about 100 gm. This means that the expanded beam at the exit surface of the fourth prism is 2 w M ~ 20.7 mm. Thus, this particular multiple-prism beam expander can comprise three small prisms with a hypotenuse of 17 mm, though the first two can be even smaller. The larger, fourth prism requires a hypotenuse of 27 mm. The thickness of all prisms can be 10 mm. The surface of these prisms are usually polished to yield a surface flatness of A/4 or better. The design of compact solid-state multiple-prism laser oscillators, as illustrated in Fig. 4.4, requires the use of small prism beam expanders with M ~ 40. The exercise is now repeated for r = 2 in Eq. (4.27). Setting V ~ p = 0 we obtain (Duarte, 2000)

tan @1,1 = kl,2 t a n / / ; 1 , 2

(4.50)

For kl,1 --- 30, ~1,1 ~--- 88.560 and ~)1,1 = 41.23 ~ Hence, for the second prism, Eq. (4.50) yields q51,2= 53.05 ~ ~1,2 = 31.80 ~ and kl,2 = 1.41. Therefore, the overall intracavity beam expansion becomes kl,lkl,2 = 42.41. For a beam waist of w = 100 gm this implies 2 w M ,.~ 8.48 mm. These dimensions require the first prism to have a hypotenuse of ~8 mm and the second prism a hypotenuse of ~ 10 mm.

4.4 MULTIPLE-PRISM DISPERSION A N D PULSE COMPRESSION Prismatic pulse compression was demonstrated for the first time by Dietel etal. (1983), yielding a pulse duration of 53 fs. A prism compensating pair was introduced by Diels etal. (1985), obtaining a pulse duration of 85 fs.

Two compensating prism pairs as pulse compressors were demonstrated by Fork etal. (1984), who reported pulses of 65 fs. Generalized dispersion equations applicable to pulse compression were introduced by Duarte and Piper (1982) and Duarte (1987). Various pulse compression configurations are depicted in Fig. 4.5.


69

(a)

(b)

(c)

(d)

(e)

(f) Figure 4.5 Pulse-compression prismatic configurations. (a) Single prism. (b) Double-prism compensating arrangement. (c) Four-prism array comprised of two double-prism compensating configurations. (d) Collinear array integrated by two N-prism compensating configurations. (e) and (f) depict two arrays, each comprising generalized N-prism additive configurations. The groups compensate relative to each other [Reprinted from Duarte (1995c), copyright 1995, with permission from Elsevier].

70


In the laser linewidth equation

=

x0(v 0)

~7~0 is the overall intracavity dispersion. In laser cavities designed for narrow-linewidth oscillation, ~7~0 is very large and is dominated by the dispersion of the grating, which is often multiplied by M. By contrast, in the case of ultrashort-pulse lasers, as a consequence of the uncertainty principle (see Chapter 3), the emission spectrum is broadband, which is associated with a small value for ~7~0. As discussed by Diels (1990), the phase factor in a femtosecond laser pulse depends on the addition of ~72n and ~72p terms. Here, P is the optical path length of the compressor, whose derivative with respect to wavelength is given by ~7~e = ~,~nVnC~VO P

(4.51)

and its second derivative is (Fork et al., 1984)

~72p - (~72n~Tn~5+ (~TAn)2V2nq~)~Tq~P-~-(~7An)2(VnqS)2~7~P

(4.52)

where ~7~n is the prism material dispersion, 272- 02/0~ 2, ~72n is the derivative of the prism material dispersion, ~'~P is the angular derivative of the path length, and V~P is the second angular derivative of the path length. In general, ~7~n is a negative number and 272n is a positive number and

IV2nl >> IV nl

(4.53)

Numerical values for n, ~7~n, and 272n for some well-known materials are given in Table 4.1. In order to compress ultrashort pulses with a large positive chirp, gTZp should be a large negative number (Diels, 1990). From Eqs. (4.6), (4.18), and (4.25) it should be evident that the prismatic dispersion can be made positive or negative at will by choice of design and configuration. For a generalized multiple-prism array ~7n~ and ~72~bcorrespond to ~n~Z,m and ~72q~Z,m, respectively. The first derivative can be obtained by using the identity Vn~2,m -- Vk(P2,rn(~TAnm) -1

(4.54)


71

Table 4.1 Dispersion Characteristics of Prism Materials for Pulse Compression Material

Quartz BK7 F2 SF 10 LaSF9 ZnSe a

n

A (lam)

V~n (btm-')

V 2n (lam-2)

Reference

1.457 1.51554 1.61747 1.72444 1.84629 1.83257 2.586 2.511

0.62 0.62 0.62 0.62 0.62 0.80 0.62 0.80

-0.03059 -0.03613 -0.07357 -0.10873 -0.11189 -0.05201 -0.698 -0.246

0.1267 0.15509 0.34332 0.53819 0.57778 0.18023 5.068 1.163

Fork etal. (1984) Diels (1990) Diels (1990)

Source: Duarte, 1995b. a Calculated using data from Marple, 1964.

and the second derivative is given by Duarte (1987) 2

Vn~2,m --~-ga2,m(VnO2,m)2nm + (kl,mk2,m) -! -1- V2nq52,(m-1)) -~- (~gal,m -]- Vnq52(m-1))

X {(~--~l,mXl,mkl,mVnr

(Xl,mVn~31,m -- ~2,mk2,mVn~32,m - ~l,mkl,mVn~l,mnm) } -~-(k2,m) -1 (Vnr

(4.55)

+ Vnr

where Xl,m =

tan ~l,m

(4.56)

For a single isosceles prism, that is, m - 1, deployed for minimum deviation and Brewster's angle of incidence, Eqs. (2.54) and (2.55) reduce to Vnq52,1 -- 2 V2q52,1 --

(4n-

(4.57) (2/n3))

(4.58)

which are the single-prism results given by Fork et al. (1984). Using P = 2l cos r Duarte (1987) calculated (at A = 620 nm) the required prism separation length to achieve negative dispersion for an isosceles prism sequence, as illustrated in Fig. 4.5c, for prisms made of quartz, LaSF9, and ZnSe. Those calculations

72


demonstrated a significant reduction in the required prism separation as the index of refraction of the prism material increased. For a single-prism material, the required prism separation decreases as the angle of incidence increases until a minimum is reached. A rapid increase is observed following the minimum. The effect of minute angular deviations from the Brewster angle of incidence on ~n~PZ,mand ~72052,mwas quantified by Duarte (1990b). A detailed discussion of pulse compression can be found in Diels and Rudolph (1996).

4.5 APPLICATIONS OF MULTIPLE-PRISM ARRAYS So far, the use of multiple-prism arrays as intracavity beam expanders in narrow-linewidth tunable laser oscillators and as pulse compressors in ultrafast lasers has been examined. However, multiple-prism arrays as onedimensional telescopes have found a variety of alternative applications as follows. Extracavity double-prism beam expander to correct the ellipticity of laser beams generated by semiconductor lasers (Maker and Ferguson, 1989). These expanders make use of the prism pairs first introduced by Brewster (1813). Generalized laser beam shaping devises as discussed by Duarte (1995c). Beam-expanding devices for optical computing (Lohmann and Stork, 1989) One-dimensional beam expanders yielding extremely elongated Gaussian beams, with minimum height to width with ratios of 1:1000, for scanning applications (Duarte, 1987, 1993, 1995a). This includes characterization of transmission and reflection imaging surfaces. Multiple-prism beam expansion for N-slit interferometry (Duarte, 1991, 1993). One-dimensional multiple-prism telescopes in conjunction with digital detectors (Duarte, 1993, 1995a). This includes applications in conjunction with transmission grating in areas of wavelength measurements (Duarte, 1995c) and secure optical communications (Duarte, 2002). One-dimensional beam expansion for application in laser exposing devices known as laser sensitometers (Duarte, 2001b).

PROBLEMS 1. Show that Eq. (4.6) can also be expressed in its explicit form of Eq. (4.12). 2. Show that for orthogonal beam exit, Eq. (4.12) reduces to Eq. (4.15). 3. Show that for prisms with (c~1 = c~2 = c~3 . . . . . C~m)and deployed for Brewster's angle of incidence, Eq. (4.15) can be restated as Eq. (4.17).


73

4. D e s i g n a t h r e e - p r i s m b e a m e x p a n d e r w i t h Vaq~e = 0 f o r M ~ 70. A s s u m e A = 590 n m , n = 1.5167 a n d w = 100 lam. 5. S h o w t h a t f o r m = 1, o r t h o g o n a l b e a m exit, a n d B r e w s t e r ' s a n g l e o f i n c i d e n c e , Eq. (4.18) r e d u c e s to E q . (4.39).

REFERENCES Born, M., and Wolf, E. (1999). Principles of Optics, 7th ed. Cambridge University Press, Cambridge, UK. Brewster, D. (1813). A Treatise on New Philosophical Instruments for Various Purposes in the Arts and Sciences with Experiments on Light and Colours. Murray and Blackwood, Edinburgh. Diels, J.-C. (1990). Femtosecond dye lasers. In Dye Laser Principles (Duarte, F. J. and Hillman, L. W., ed.). Academic, New York, pp. 41-132. Diels, J.-C., and Rudolph, W. (1996). Ultrafast Laser Pulse Phenomena. Academic Press, New York. Diels, J.-C., Dietel, W., Fontaine, J. J., Rudolph, W., and Wilhelmi, B. (1985). Analysis of a mode-locked ring laser: chirped-solitary-pulse solutions. J. Opt. Soc. Am. B, 2, 680-686. Dietel, W., Fontaine, J. J., and Diels, J.-C. (1983). Intracavity pulse compression with glass: a new method of generating pulses shorter than 60 fs. Opt. Lett. 8, 4-6. Duarte, F. J. (1985). Note on achromatic multiple-prism beam expanders. Opt. Commun. 53, 259-262. Duarte, F. J. (1987). Generalized multiple-prism dispersion theory for pulse compression in ultrafast dye lasers. Opt Quantum Electron. 19, 223-229. Duarte, F. J. (1989). Transmission efficiency in achromatic nonorthogonal multiple-prism laser beam expanders. Opt. Commun. 71, 1-5. Duarte, F. J. (1990a). Narrow-linewidth pulsed dye laser oscillators. In Dye Laser Principles (Duarte, F. J. and Hillman, L. W., eds.). Academic Press, New York, pp. 133-183. Duarte, F. J. (1990b). Prismatic pulse compression: beam deviations and geometrical perturbations. Opt. Quantum Electron. 22, 467-471. Duarte, F. J. (1991). Dispersive dye lasers. In High-Power Dye Lasers (Duarte, F. J., ed.). Springer-Verlag, Berlin, pp. 7-43. Duarte, F. J. (1992). Cavity dispersion equation ~ ~ &0(00/0A)-l: a note on its origin. Appl. Opt. 31, 6979-6982. Duarte, F. J. (1993). On a generalized interference equation and interferometric measurements. Opt. Commun. 103, 8-14. Duarte, F. J. (1995a). Interferometric imaging. In Tunable Laser Applications (Duarte, F. J., ed.). Marcel Dekker, New York, pp. 153-178. Duarte, F. J. (1995b). Dye lasers. In Tunable Lasers Handbook (Duarte, F. J., ed.). Academic Press, New York, pp. 167-218. Duarte, F. J. (1995c). Narrow-linewidth laser oscillators. In Tunable Lasers Handbook (Duarte, F. J., ed.). Academic Press, New York, pp. 9-32. Duarte, F. J. (2000). Multiple-prism arrays in laser optics. Am. J. Phys. 68, 162-166. Duarte, F. J. (200 l a). Multiple-return-pass beam divergence and the linewidth equation. Appl. Opt. 40, 3038-3041. Duarte, F. J. (2001 b). Laser sensitometer using multiple-prism beam expansion and a Polarizer. US Patent No. 6236461.

74


Duarte, F. J. (2002). Secure interferometric communications in free space. Opt. Commun. 205, 313-319. Duarte, F. J., and Piper, J. A. (1982). Dispersion theory of multiple-prism beam expander for pulsed dye lasers. Opt. Commun. 43, 303-307. Duarte, F. J., and Piper, J. A. (1983). Generalized prism dispersion theory. Am. J. Phys. 51, 1132-1134. Duarte, F. J., and Piper, J. A. (1984a). Narrow-llinewidth; high prf copper laser-pumped dye laser oscillators. Appl. Opt. 23, 1391-1394. Duarte, F. J., and Piper, J. A. (1984b). Multi-pass dispersion theory of prismatic pulsed dye lasers. Optica Acta 31, 331-335. Fork, R. L., Martinez, O. M., and Gordon, J. P. (1984). Negative dispersion using pairs of prisms. Opt. Lett. 9, 150-152. Lohmann, A. W., and Stork, W. (1989). Modified Brewster telescopes. Appl. Opt. 28, 13181319. Maker, G. T., and Ferguson, A. I. (1989). Frequency-modulation mode locking of a diodepumped Nd:YAG laser. Opt. Lett. 14, 788-790. Marple, D. T. F. (1964). Refractive index of ZnSe, ZnTe, and CdTe. J. Appl. Phys. 35, 539-542. Newton, I. (1704). Opticks. Royal Society, London. Wyatt, R. (1978). Comment on "On the dispersion of a prism used as a beam expander in a nitrogen laser." Opt. Commun. 26, 9-11.

Chapter 5

Linear Polarization

5.1 MAXWELL EQUATIONS Maxwell equations are of fundamental importance since they describe the whole of classical electromagnetic phenomena. From a classical perspective, light can be described as waves of electromagnetic radiation. As such, Maxwell equations are very useful to illustrate a number of the characteristics of light including polarization. It is customary just to state these equations without derivation. Since our goal is simply to apply them, the usual approach will be followed. However, for those interested, a derivation by Dyson (1990), attributed to Feynman, is available in the literature. M a x w e l l equations in the rationalized metric system are given by

v.s=0

(5.1)

V . E = p/co

(5.2)

c2V • a = oE/ot +j/~o ~7 ><E = - O B / O t

(5.3) (5.4)

(Feynman et al., 1965). These equations illustrate, with succinct beauty, the unique coexistence in nature of the electric field and the magnetic field. The first two equations give the value of the given flux through a closed surface, and the second two equations give the value of a line integral around a loop. In this notation, v = (OlOx, OlOy, O l & )

75

76


E is the electric vector, B is the magnetic induction, p is the electric charge density, j is the electric current density, co is the permittivity offree space, and c is the speed of light. In addition to Maxwell equations, the following identities are useful: j-

erE

(5.5)

D = eE

(5.6)

B = #H

(5.7)

Here, D is the electric displacement, H is the magnetic vector, er is the specific conductivity, e is the dielectric constant (or permittivity), and # is the magnetic permeability. In the Gaussian systems of units, Maxwell equations are given in the form of

v.B=o

(5.8)

V . E = 4rrp

(5.9)

V x H = (1/c)(OD/Ot + 4rrj)

(5.10)

V • E =-(1/c)(OB/Ot)

(5.11)

(Born and Wolf, 1999). It should be noted that many authors in the field of optics prefer to use Maxwell equations in the Gaussian system of units. As explained by Born and Wolf (1999), E, D, j, and p in this system are measured in electrostatic units, and H and B are measured in electromagnetic units. The use of Maxwell equations in this book is consistent with the tradition established by Born and Wolf. For the case of no charges or currents, that is, j - 0 and p - 0, and a homogeneous medium, Maxwell equations and the given identities can be applied in conjunction with the vector identity V x V • E = V V . E - ~72E

(5.12)

to obtain wave equations of the form (Born and Wolf, 1999) V2F

- (

l 2)(02F lOt 2) - 0

(S.13)

This leads to an expression for the velocity of propagation 1~ = C I ( C ~ ) 1/2

(5.14)

Comparison of this expression with the law of refraction, derived in Chapter 2, leads to what is known as Maxwell's formula (Born and Wolf, 1999):

(S.lS)

77

Linear Polarization

Here, n is the refractive index. It is useful to note that in vacuum r __~_l/C0#0

(5.16)

in the rationalized metric system, where #0 is the permeability of free space (Lorrain and Corson, 1970).

5.2 POLARIZATION A N D REFLECTION Following the convention of Born and Wolf (1999), we consider a reflection boundary, depicted in Fig. 5.1, and a plane of incidence established by the incidence ray and the normal to the reflection surface. Here, the reflected component ~ll is parallel to the plane of incidence, and the reflected component ~ x is perpendicular to the plane of incidence. For the case of #l = # 2 = l , Born and Wolf (1999) consider the electric and magnetic vectors as complex plane waves. In this approach, the incident electric vector is represented by equations of the form

E(xi) - _All cos dp(e-in) E(i) _ _Ax(e-i~i)

(5.18)

g~ i) -- -All sin ~(e -in)

(5.19)

(5.17)

where All and A• are complex amplitudes and Ti is the usual plane-wave phase factor, written as Ti = ~ [t - (x sin ~b+ z cos cb)/v]. Using corresponding I

~"--~

I

RII

_

.

~

Rl

T•

Figure 5.1 Reflectionboundary defining the plane of incidence.

78


equations for E and H for transmission and reflection in conjunction with Maxwell's relation, with # = 1, and the law or refraction, Born and Wolf (1999) derive the Fresnelformulae: Q_UII = ((2 sin ~bcos qS)/sin(q5 + ~b) cos(q5 - ~b))Aii ~ • = ((2 sin ~bcos qS)/sin(q5 + ~b))A• ~11 = (tan(q5 - ~b)/tan(q5 + ~b))Aii ~ • = (sin(q~ - ~b)/sin(~5 + ~b))A•

(5.20) (5.21) (5.22) (5.23)

Using these equations the transmissivity and reflectivity can be expressed as d-tl - ((sin 2q5 sin 2~b)/sin2(q5 + ~b) cos2(~b - ~b))

(5.24)

~z

(5.25)

- ((sin 2~b sin 2~b)/sin2(~b + ~b))

~ll -- (tan2( q5 - ~b)/tan2(q5 + @))

(5.26)

~ z - (sin2(q5 - ~b)/sin2(q5 + ~b))

(5.27)

and ~11 + ~--~11= 1 ~• + ~• = 1

(5.28) (5.29)

Using these expressions for transmissivity and reflectivity, the degree of polarization, ~ , is defined as (Born and Wolf, 1999)

+

(s.30)

The usefulness of these equations is self-evident once ~ll is calculated as a function of angle of incidence (Fig. 5.2) for fused silica at 594nm (n = 1.4582). Here we see that ~11 = 0 at 55.5586 degrees. At this angle (4~ + 9) becomes 90 degrees so that tan (~b + 9) approaches infinity thus causing ~11 = 0. This particular ~b, known as the Brewster angle (~@), has a very important role in laser optics. Since at 4~ = 4~ the angle of refraction becomes 9 - (90 - ~b)~ the law of refraction takes the form of tan q~ = n

(5.31)

For orthogonal, or normal, incidence, the difference between the two polarizations vanishes. Using the law of refraction and the appropriate trigonometric identities in Eqs. (5.24)-(5.27), it can be shown that (Born and Wolf, 1999) -((n-

1)/(n + 1)) 2

~_U - (4n/(n + 1)) 2

(5.32) (5.33)

79

Linear Polarization 1.0 0.9

0.8-

0.7.~ 0 . 6 -

o.5-

rr -- 0.40.30.2-

q~= 55.5586 ~

i

100

Figure 5.2

i

200

i

30s

I

400

]

500

i

600

i

700

i

800

)00

Reflection intensity as a function of angle of incidence.

5.2.1 THE PLANE OF INCIDENCE The discussion in the preceding section uses parameters such as ~1t and ~ • In this convention, 1] means parallel to the plane of incidence and • means perpendicular, orthogonal, or normal to the plane of incidence. The plane of incidence is defined, following Born and Wolf (1999), in Fig. 5.1. However, in more explicit terms, let us consider a laser beam propagating on a plane parallel to the surface of an optical table. If that beam is made to illuminate the hypotenuse of a fight angle prism whose triangular base is parallel to the surface of the table, then the plane of incidence is established by the incident laser beam and the perpendicular to the hypotenuse of the prism. In other words, in this case the plane of incidence is parallel to the surface of the optical table. At this stage it should be mentioned that lasers using windows deployed at Brewster's angle, to minimize their losses, emit parallel to their plane of incidence. However, the plane of incidence of the laser can often be orthogonal to an external plane of incidence. When that is the case, and maximun transmission of the laser through external optics is desired, the laser is rotated by ~r/2 about its axis of propagation.

5.3 P O L A R I Z I N G

PRISMS

There are two avenues to induce polarization using prisms. The first involves simple reflection, as characterized by Fresnel's equations and straightforward refraction. This approach is valid for windows, prisms, or multipleprism arrays made from homogeneous optical materials such as optical glass


80

and fused silica. The second approach involves double refraction in crystalline transmission media exhibiting birefrinegence.

5.3.1

T R A N S M I S S I O N EFFICIENCY IN M U L T I P L E - P R I S M ARRAYS

For a generalized multiple-prism array, as shown in Fig. 5.3, the cumulative reflection losses at the incidence surface of the mth prism are given by (Duarte et al., 1990) (5.34)

Zl,m -- Z2,(m-1) -t- (1 - Z2,(m_l))~l, m

I

I

\\ /1,m I I

,m I

L2,m Figure 5.3 Generalized multiple-prism array, in additive configuration, indicating correspondence to cumulative reflection loss factors Ll,m and L2,m [adapted from Duarte (1990), copyright 1990, with permission from Elsevier].

Linear Polarization

81

and the losses at the mth exit surface are given by L2,m = L l,m -Jr- (1 - L l,m) ~2,m

(5.35)

where ~l,m and ~2,m are given by either ~11 or ~ j . In practice, the optics are deployed so that the polarization of the propagation beam is parallel to the plane of incidence, meaning that the reflection coefficient is given by ~11" It should be noted that these equations apply not just to prisms but also to optical wedges and any homogeneous optical element with an input and exit surface used in the transmission domain.

5.3.2

INDUCED POLARIZATION IN A DOUBLE-PRISM BEAM EXPANDER

Polarization induction in multiple-prism beam expander s should be apparent once the Fresnel equations are combined with the transmission Eqs. (5.34) and (5.35). In this section this effect is made clear by considering the transmission efficiency for both components of polarization of a simple double-prism beam expander, as illustrated in Fig. 5.4. This beam expander comprises two identical prisms made of fused silica, with n = 1.4583 at A ~ 590nm and an apex angle of 42.7 ~ Both prisms are deployed to yield identical magnifications and for orthogonal beam exit. This implies that q~l,1 = q~l,2 = 8 1 . 4 9 ~ ~Pl,1 = ~1,1 = 41.70 ~ q~2,1 = ~2,2 = 0, and ~P2,1= ~P2,2= 0. Under these conditions, for radiation polarized parallel to the plane of incidence,

~11

/

Figure 5.4 Double-prismexpander as described in the text.

82


LI,1 = ~1,1 = 0 . 2 9 8 5

L2,1 = Ll,1 L1,2 = L2,1 -+- (1 - L2,1) ~ 1 , 2 - 0.5079 L2,2 -- L1,2

and for radiation polarized perpendicular to the plane of incidence, Ll,1 = ~1,1 = 0.5739 L2,1 = LI,1 L1,2 -- L2,1 nt- (1 - L2,1) ~ 1 , 2 = 0.8285 L2,2 -- L1,2 Thus, for this particular beam expander, the cumulative reflection losses are 50.79% for light polarized parallel to the plane of incidence, they increase to 82.85% for radiation polarized perpendicular to the plane of incidence. This example helps to illustrate the fact that multiple-prism beam expanders exhibit a clear polarization preference. It is easy to see that the addition of further stages of beam magnification lead to increased discrimination. When incorporated in frequency-selective dispersive laser cavities, these beam expanders contribute significantly to laser emission polarized parallel to the plane of propagation.

5.3.3 DOUBLE-REFRACTIONPOLARIZERS These are crystalline prism pairs that exploit the birefringence effect in crystals. In birefringent materials the dielectric constant, ~, is different in each of the x, y, and z directions, so the propagation velocity is different in each direction: a 12b

-

-

-

-

r

1/2

C/(Ey) 1/2

(5.36) (5.37) (S.38)

Since polarization of a transmission medium is determined by the D vector, it is possible to describe the polarization characteristics in each direction. Further, it can be shown that there are two different velocities for the refracted radiation in any given direction (Born and Wolf, 1999). As

83

Linear Polarization Anti-Reflective Coating P

Laser Output

t" Partially Reflective Coating

Figure 5.5 Glan-Thompsonpolarizer [from Duarte (1995a), copyright 1995, with permission

from Elsevier]. a consequence of the law of refraction, these two velocities lead to two different propagation paths in the crystal and give rise to the ordinary and extraordinary ray. In other words, the two velocities lead to double

refraction. Although there are a number of double-refraction prism pairs, here we consider only those that allow straight transmission of the extraordinary ray from the first to the second prism. Of particular interest in this class of polarizers are those known as the Nicol prism, the Rochon prism, the GlanFoucault prism, and the Glan-Thompson prism. According to Bennett and Bennett (1978), a Glan-Foucault prism pair is an air-spaced Glan-Thompson prism pair. As illustrated in Fig. 5.5, in the Glan-type polarizers the extraordinary ray is transmitted from the first to the second prism in the propagation direction of the incident beam. On the other hand, the diagonal surface of the two prisms is predetermined to induce total internal reflection for the ordinary ray. Glan-type polarizers are very useful because they can be oriented to discriminate in favor of either polarization component with negligible beam deviation. Normally these polarizers are made of calcite. Commercially available calcite Glan-Thompson polarizers with a useful aperture of 10mm provide extinction ratios of ~ 5 x 10-s. It should be noted that Glan-type polarizers are used in straightforward propagation applications as well as intracavity elements. For instance, the tunable single-longitudinalmode laser oscillator depicted in Fig. 5.6 incorporates a Glan-Thompson polarizer as output coupler. In this particular polarizer the inner window is antireflection coated while the outer window is coated for partial reflectivity to act as an output coupler mirror. The laser emission from multiple-prism grating oscillators is highly polarized parallel to the

84

Tunable Laser Optics Solid-state gain medium

-,... M

Figure 5.6 Solid-state MPL grating dye laser oscillator, yielding single-longitudinal-mode emission, incorporating a Glan-Thompson polarizer output coupler [Reprinted from Duarte (1995b), copyright 1995,with permission from Elsevier]. plane of incidence by the interaction of the intracavity flux with the multipleprism expander and the grating. The function of the polarizer output coupler here is to provide further discrimination against unpolarized single-pass amplified spontaneous emission. These dispersive tunable laser oscillators yield extremely low levels of broadband amplified spontaneous emission measured to be in the 10-7-10 -6 range (Duarte, 1995b, 1999).

5 . 3 . 4 ATTENUATION OF THE INTENSITY OF LASER BEAMS

USING POLARIZATION A very simple and yet powerful technique to attenuate the intensity of linearly polarized laser beams involves the transmission of the laser beam through a prism pair, such as a Glan-Thompson polarizer, followed by rotation of the polarizer. This is illustrated in Fig. 5.7. In this technique for a ~ 100% laser beam polarized parallel to the plane of incidence, there is almost total transmission when the Glan-Thompson prism pair is oriented as in Fig. 5.7a. As the prism pair is rotated about the axis of propagation, the intensity of the transmission decreases until it becomes zero once the angular displacement has reached 7r/2. With precision rotation of the prism pair, a scale of well-determined intensities can easily be obtained. This has a number of applications, including the generation of precise laser intensity scales for exposing instrumentation used in imaging (Duarte, 2001).

Linear Polarization

85

(a)

(b) Figure 5.7 Attenuationof polarized laser beams using a Glan-Thompson polarizer. (a) Polarizer set for ~ 100% transmission. (b) Rotation of the polarizer about the axis of propagation by 7r/2 yields ,-,0% transmission. The amount of transmitted light can be varied continuously by rotating the polarizer in the 0 < 0 < 7r/2range.

5.4 POLARIZATION ROTATORS Maximum transmission efficiency is always a goal in optical systems. If the polarization of a laser is mismatched to the polarization preference of the optics, then transmission efficiency will be poor. Although the transmission sometimes can be improved or even optimized by the simple rotation of a laser, it is highly desirable and practical to have optical elements to perform this function. In this section we shall consider three alternatives to perform such rotation: rhomboids, half-wave and quarter-wave plates, and prismatic rotators.

5.4.1 FRESNEL RHOMBS AND TOTAL INTERNAL REFLECTION An interesting situation occurs when light propagation from a higher- to a lower-density medium is considered. Under those circumstances, and provided the angle of incidence is greater than a critical value, the light beam will undergo total internal reflection. For propagation under these conditions, the refractive index is less than 1 and the critical value for the angle of incidence (~bc) is greater than the angle resulting in sin r = 1. In other words, 4~c is greater than the angle determined from sin 4~ = n. Setting sin r = ( sin c~)/n and cos r - +i [( sin 2 (9/n 2) - 1]1/2 in the Fresnel equations, Born and Wolf (1999) show that for each component of polarization the incident light is totally reflected. Further, considering E and H with

86


Figure 5.8 DoubleFresnel rhom.

the phase factors applicable to total internal reflection, these authors show that there is a phase difference between the two components of polarization given by tan(6/2) - cos ~(sin 2 q~- rt2)l/2/sin

2 4~

(5.39)

Thus it is possible to control the change in polarization by careful selection of n and the angle of incidence. This phase difference applies to the rotation of linearly polarized light in Fresnel rhombs and double Fresnel rhombs (see Fig. 5.8). A double Fresnel rhomb, comprising two quarter-wave parallelepipeds, becomes a half-wave rhomb, which rotates linearly polarized light by 7r/2. A commercially available Fresnel rhomb of this class, 10 mm wide and 53 mm in length, offers a transmission efficiency close to 96% with broadband antireflection coatings. As indicated by Bennett and Bennett (1978), these achromatic rotators tend to offer a small useful aperture-tolength ratio, and its achromaticity can be compromised by residual birefringence in the quartz.

5.4.2 BIREFRINGENTROTATORS In birefringent uniaxial crystalline materials the ordinary and extraordinary rays propagate at different velocities. Following Born and Wolf (1999), the D vectors for the incident radiation can be written as D(xi) - acoswt

(5.40)

D~ i) - b sin wt

(5.41)

87

Linear Polarization

and the components for the transmitted radiation can be expressed as

(Se)

(5.42)

D(yt) -- b cos(cot + 60)

(5.43)

D (t) - a cos(cot +

Using standard notation it follows that the phase difference, 6 = 6 e - ,50, introduced by transmission in the birefringent material is given by t5 = k(ne - n o ) d

(5.44)

(5 = (2rc//~)(ne - n o ) d

(5.45)

For linearly polarized transmitted light, the condition 6 = (2m + 1)rr/2 must be met, where m is an integer. Hence, it follows that the thickness of a uniaxial crystalline plate is determined by d = ](2m + 1)/(ne - n o ) l ( A / 4 )

(5.46)

This is a useful and important result since, besides its application in the design of r e t a r d a t i o n p l a t e s , it indicates that these devices are wavelength specific, which can be a limitation. For a q u a t e r - w a v e p l a t e , 6 = 7r/2; for a h a l f - w a v e p l a t e , 6 = 7r.

5.4.3 BROADBAND PRISMATIC ROTATORS An alternative to frequency-selective polarization rotators are prismatic rotators. These devices work at normal incidence and apply the principle of total internal reflection. The basic operation of polarization rotation by 7r/2, due to total internal reflection is shown in Fig. 5.9. This operation, however, reflects the beam into a direction that is orthogonal to the original propagation. Furthermore, the beam is not in the same plane. In order to achieve collinear polarization rotation by 7r/2, the beam must be displaced upward and then be brought into alignment with the incident beam while conserving the polarization rotation achieved by the initial double reflection operation. A collinear prismatic polarization rotator that performs this task using seven total internal reflections is depicted in Fig. 5.10. The individual components of this collinear polarization rotator are shown in Fig. 5.11. For high-power laser applications this rotator is best assembled using a high-precision mechanical mount that allows air interfaces between the individual prisms. The useful aperture in this rotator is about 10mm and its physical length is 30mm.


88

I I

i

)L_U_ i/

iI

Figure 5.9

Basic operator for polarization rotation using two reflections.

f

Figure 5.10 Broadband collinear polarization rotator. For high power applications the rotator is integrated using a high-precision optical mount thus allowing for direct contact at the surfaces of the prisms (From Duarte 1992.)

89

Linear Polarization

J

s

/

/

f

s

/ s

N/"

Figure 5.11 Components of the broadband collinear polarization rotator. It should be noted that despite the apparent complexity of this collinear polarization rotator, the transmission efficiency is relatively high using antireflection coatings. In fact, using broadband (425-675 nm) antireflection coatings with a nominal loss of 0.5% per surface, the measured transmission efficiency becomes 94.7% at )~ = 632.8 nm. The predicted transmission losses using Lr - 1 - (1

-

L) r

(5.47)

are 4.9%, with L = 0.5%, as compared to a measured value of 5.3%. Equation (5.47) is derived combining Eqs. (5.34) and (5.35) for the special case of identical reflection losses. Here, r is the total number of reflection surfaces, which is 10 for this particular collinear rotator. A further parameter of interest is the transmission fidelity of the rotator, since it is also important to keep spatial distortions of the rotated beam to a minimum. The integrity of the beam due to transmission and rotation is quantified in Fig. 5.12, where a slight beam expansion of ~3.2% at F W H M is evident (Duarte, 1992).

90


Magnitude:

6000 .m

6 57

j~

(a)

5000

C

E > .m

4000 3000 2000 1000 440

480

520 Pixel Number

560

(b)

6000 5000 C

_

Magnitude: 6 _

~'~

4000

C m 0

~9

3000 2000

1000 440

480

520 Pixel Number

560

Figure 5.12 Transmission fidelity of the broadband collinear polarization rotator. (From Duarte, 1992.)

PROBLEMS 1. Design a single right angle prism made of fused silica to expand a laser beam by a factor of 2 with orthogonal beam exit. Calculate ~11 and ~ • (use n - 1.4583 at A ~ 590 nm).

Linear Polarization

91

2. For a four-prism beam expander with orthogonal beam exit and using fused silica prisms with an apex angle of 41 ~, calculate the overall beam magnification factor M. Also, calculate the overall transmission efficiency for a laser beam polarized parallel to the plane of incidence (use n = 1.4583 at A ~ 590 nm). 3. Calculate the thickness of a half-wave plate using quartz at ,~ ~ 590 nm. For this particular plate estimate the angular deviation from a 7r/2 rotation experienced at )~ ~ 650 nm 4. Use Maxwell equations in the Gaussian system for the j - 0 and p - - 0 case to derive the wave equations VaF

-

(

,/o2)(OaF (oan/ot

/Ot

-

0

-

0

5. Assume normal incidence in the Fresnel equations to obtain Eqs. (5.32) and (5.33).

REFERENCES Bennett, J. M., and Bennett, H. E. (1978). Polarization. In Handbook of Optics (Driscoll, W. G., and Vaughan, W., eds.). McGraw-Hill, New York. Born, M., and Wolf, E. (1999). Principles of Optics, 7th ed. Cambridge University Press, New York. Duarte, F. J. (1990). Narrow-linewidth pulsed dye laser oscillators. In Dye Laser Principles (Duarte, F. J. and Hillman, L. W., eds.) Academic, New York, pp. 133-183. Duarte, F. J. (1992). Beam transmission characteristics of a collinear polarization rotator. Appl. Opt. 31, 3377-3378. Duarte, F. J. (1995a). Narrow-linewidth laser oscillators. In Tunable Lasers Handbook (Duarte, F. J., ed.) Academic, New York, pp. 9-32. Duarte, F. J. (1995b). Solid-state dispersive dye laser oscillator: very compact cavity. Opt. Commun. 117, 480-484. Duarte, F. J. (1999). Multiple-prism grating solid-state dye laser oscillator: optimized architecture. Appl. Opt. 38, 6347-6349. Duarte, F. J. (2001). Laser sensitometer using a multiple-prism beam expander and a polarizer. U.S. Patent no. 6,236,461 B1. Duarte, F. J., Ehrlich, J. J., Davenport, W. E., and Taylor, T. S. (1990). Flashlamp pumped narrow-linewidth dispersive dye laser oscillators: very low amplified spontaneous emission levels and reduction of linewidth instabilities. Appl. Opt. 29, 3176-3179. Dyson, F. J. (1990). Feynman's proof of Maxwell equations. Am. J. Phys. 58, 209-211. Feynman, R. P., Leighton, R. B, and Sands, M. (1965). The Feynman Lectures on Physics, Vol. II. Addison-Wesley, Reading, MA. Lorrain, P., and Corson, D. (1970). Electromagnetic Fields and Waves. Freeman, San Francisco.


Chapter 6

Laser Beam Propagation Matrices

6.1 I N T R O D U C T I O N A powerful approach to characterizing and designing laser optics systems is the use of beam propagation matrices, also known as ray transfer matrices. This is a practical method that applies to the propagation of laser beams with a Gaussian profile. Since most lasers can be designed to yield beams with Gaussian or near-Gaussian profiles, this is a widely applicable method. From a historical perspective, it should be mentioned that propagation matrices in optics have been known for a while. For early references in the subject the reader is referred to Brouwer (1964), Kogelnik (1979), Siegman (1986), and Wollnik (1987). In this chapter the basic principles of propagation matrices are outlined and a survey of matrices for various widely applicable optical elements is given. The emphasis is on the application of the method to practical optical systems. In addition, examples of some useful single-pass and multiple-pass calculations are given. Higher-order matrices are also considered.

6.2 A B C D PROPAGATION MATRICES The basic idea with propagation matrices is that one vector at a given plane is related to a second vector at a different plane via a linear transformation. This transformation is represented by a propagation matrix. This concept is 93

94


] Figure 6.1

Geometry for propagation through distance I in space.

applicable to the characterization of the deviation of a ray, or beam, of light through either free space or any linear optical media. The ray of light is assumed to be a paraxial ray that propagates in proximity and almost parallel to the optical axis (Kogelnik, 1979). To illustrate this idea further, consider the propagation of a paraxial ray of light from an original plane to a secondary plane in free space, as depicted in Fig. 6.1. Here it is noted that, in moving form the original plane to the secondary plane, the ray of light experiences a linear displacement in the x direction and a small angular deviation; that is, X 2 - - X l -3t- 101

(6.1)

02 -- 01

(6.2)

which in matrix form can be stated as

(02) (0 1) (01) X2

1

l

Xl

63,

The resulting 2 • 2 matrix is known as a ray transfer matrix. Here, it should be noted that some authors (Kogelnik, 1979; Siegman, 1986) use derivatives instead of the angular quantities. For a thin lens the geometry of propagation is illustrated in Fig. 6.2. In this case, there is no displacement in the x direction and the ray is concentrated, or focused, toward the optical axis, so

(6.4)

X2--Xl

02 -- - - ( 1 / f ) x l

-+- 01

(6.5)

which in matrix form can be expressed as

(02) ( X2

1

1) (01)

0

Xl

,66,

LaserBeamPropagationMatrices

95

m

Figure 6.2

Thin convex lens.

In more general terms, the X2 vector is related to the X~ vector by a transfer matrix T known as the ABCD matrix so that

X2-- TX1

(6.7)

T - ( AC DB)

(6.8)

where

At this stage, it is useful to consider the dimensions of the components involved in these ray transfer matrices. By inspection, it is found that

oqOz / OqX l

OqX2/ O01( xlo)l 002/001)

(6.9)

This implies that A is a ratio of spatial dimensions and B is an optical length, while C is the reciprocal of an optical length. Consideration of various imaging systems leads to the conclusion that the spatial ratio represented by A is a beam magnification factor (M) and D is the reciprocal of such magnification (1/M). These observations are very useful to verify the physical validity of newly derived matrices.

6.2.1 PROPERTIESOF A B C D MATRICES A very useful property of ABCD matrices is that they can be cascaded, via matrix multiplication, to produce a single overall matrix describing the propagation properties of an optical system. For example, if a linear optical system is composed of N optical elements deployed from left to right, as


96

ra

r3

Figure 6.3

r4

r,

N optical elements in series.

depicted in Fig. 6.3, then the overall transfer matrix is given by the multiplication of the individual matrices in the reverse order; that is, N

1-I Tm - T N . . . T3 T2 T,

(6.10)

m=l

It is easy to see that the complexity in the form of these product matrices can increase rather rapidly. Thus, it is always useful to remember that any resulting matrix must have the dimensions of Eq. (6.9) and a determinant equal to unity; that is

A D - BC = 1

(6.11)

6.2.2 SURVEYOFA B C D MATRICES In Table 6.1 a number of representative and widely used optical components are represented in ray transfer matrix form. This is done without derivation and using the published literature as reference.

6.2.3 THE ASTRONOMICALTELESCOPE The astronomical telescope (Fig. 6.8) is composed of an input lens with focal length f~, an intralens distance L, and an output lens with focal length f2. Following Eq. (6.10), the matrix multiplication proceeds as

Table 6.1 ABCD Ray Transfer Matrices Optical element or system Distance I in free space (Fig. 6.1)

ABCD matrix

Reference

(b :)

Kogelnik (1979) Kogelnik (1979)

Distance I in a medium with refractive index n (Fig. 6.4)

Duarte (1991)

Slab of material with refractive index n (Fig. 6.5) Thin convex (positive) lens of focal lengthf(Fig. 6.2)

Kogelnik (1979)

Thin concave (negative) lens (Fig. 6.6)

Siegman (1986)

Galilean telescope (Fig. 6.7)

Siegman (1 986)

Astronomical telescope (Fig. 6.8)

Siegman ( 1986)

Flat mirror (Fig. 6.9)

(; Y )

Table 6.1 (Continued) Optical element or system

ABCD matrix

Reference

Curved mirror (Fig. 6.10)

Siegman (1986)

Double pass in Cassegrainian telescope (Fig. 6.11)

Siegman (1986)

Flat grating (Fig. 6.12)

O cos 0 ) cos @/

Duarte (1991)

Flat grating in Littrow configuration (Fig. 6.13) Single right angle prism (Fig. 6.14)

Siegman (1986)

( l / n )cos q5/ cos $ cos $1cos li,

Duarte (1989)

Multiple-prism beam expander (Fig. 6.15)

Duarte (1991)

Multiple-prism beam expander (return pass)

Duarte (1991)

99


Figure 6.4

Geometry for propagation through distance l in a region with refractive index n.

-~

Figure 6.5

!

r.-

Slab of material with refractive index n, such as an optical plate.

Figure 6.6

Concave lens.


100

Figure 6.7

Galilean telescope.

Figure 6.8 Astronomical telescope.

Figure 6.9

Flat mirror.


101

Figure 6.10

J

Figure 6.11

Curved mirror.

J

Cassegrainian telescope.

I !

I

Figure 6.12

Flat reflection grating.

102

Tunable Laser Optics s

s

s

s S s S

s

~

Figure 6.13 Flat reflection grating in Littrow configuration.

qo

Figure 6.14

(

1 -1//2

0)(1 1

Single prism.

L)( 1 0 1 -1/fl

0)

(6.12)

1

(note that here and further, for notational succinctness, the lefthand side of the equation, that is, the ABCD matrix, is abstracted). For a well-adjusted telescope, where (6.13)

L-f2+f~

the transfer matrix becomes A B

+ll)

-fl/f2

(6.14)

103


Figure 6.15 Multiple-prismbeam expanders in various configurations [adapted from Duarte (1991), copyright 1991, with permission from Springer-Verlag].

which is the matrix given in Table 6.1. Defining -M

(6.15)

= -(f2/f,)

this matrix can be restated as C

D

=

0

-1/M

(6.16)

For this matrix it can be easily verified that the condition lAD - BC] = 1 holds.

6.2.4 A SINGLE-PRISM IN SPACE For a right angle prism with orthogonal beam exit, preceded by a distance L1 and followed by a distance L2, as shown in Fig. 6.16, the matrix multiplication becomes

TunableLaserOptics

104 ~ i i L 1

--~

,% Figure 6.16 Singleprism preceded by a distance L1 and followedby a distance L2.

kl/nk

112 (0)(0

ILl

i/k)(0

)

(6.17)

where k - cos@/cosq5

(6.18)

Thus, the transfer matrix becomes

Llk + (Lz/k) + l/(nk)

c

~/k

)

(6.19)

which can be restated as A

B

c o)

(k0

1/k)

(6.20)

(t/,,k)

(6.21)

where B - L,k + (L2/k) +

is the optical length of the system. Notice that both L1 and L2 are modified by the dimensionless beam magnification factor k and that l is divided by the dimensionless quantity nk.

6.2.5 MULTIPLE-PRISMBEAM EXPANDERS For a generalized multiple-prism array, as illustrated in Fig. 6.17, the ray transfer matrix is given by (Duarte, 1989, 1991) C

D

-

0

(M1M2) -1


Ca)

105

__m ~ ~ _

Figure 6.17 Generalized multiple-prism array. [Reprinted from Duarte (1990), copyright 1990, with permission from Elsevier].

where

M1 - [ I kl,m

(6.23)

m=l

M2 -- I-I k2,m

(6.24)

m=l

and -2

r

B =M1 M2 Z m=l

Lm

k2,j

kl,j j=l

j=l

(6.25)

For a straightforward multiple-prism beam expander with orthogonal beam exit, cos ~bz,j = 0 and kz,j -- l, SO the equations reduce to

(AB) C

D

-

0

B)

(M1)-1

,626,


106

where

B =

r

M~ Z L m

( n kl,j )

m=l

j=l

-~-M1 Z(lm/?lm) m=l

kl,j

(6.27)

j=l

For a single prism these equations reduce further to M 1 - - k l , 1 and which is the result for the single prism given in Table 6.1.

B = l/(kl,ln),

6 . 2 . 6 TELESCOPES IN SERIES

For some applications it is necessary to propagate TEM0o laser beams through optical systems including telescopes in series, as illustrated in Fig. 6.18. For a series comprising a telescope followed by a free-space distance, followed by a second telescope, and so on, the single-pass cumulative matrix is given by

A=M r

(6.28)

Mr-2m+2Lm+

B-

D 1V1Arr-2m+l 71 l'JTm

(6.29)

m=l

C= 0 D=M

(6.30) (6.31)

-r

where r is the total number of telescopes and BTm is the B term of the mth telescope. This result applies to a series of well-adjusted Galilean or astronomical telescopes or a series of prismatic telescopes.

I

I

"-/-1-~

"-/-a-~ 4"/-r-l-~

Figure 6.18 Seriesof telescopes separated by a distance Lm.

Laser Beam PropagationMatrices

107

6.2.7 SINGLE-RETURN-PAss BEAM DIVERGENCE

It can be shown (Duarte, 1990) that the double-pass or single-return-pass divergence in a dispersive laser cavity can be expressed as

z:~,O-- ()k/71"W)(1 -[-(L~g/B) 2 -Jr-(AL~/B)2) 1/2

(6.32)

L~ = (Trw2/A)

(6.33)

where

is the Rayleigh length and A and B are the corresponding elements of the ray transfer matrix. The double-pass or single-return pass calculation for a narrow-linewidth multiple-prism grating oscillator, of the class illustrated in Fig. 6.19, can be performed using the reflection surface of the output coupler mirror as the reference point. For this purpose the cavity is unfolded about the reflective surface of the output coupler, as shown in Fig. 6.20. Thus, from the grating toward the output coupler and then proceeding from the output coupler to the grating, the matrix multiplication becomes 1 L, /M) l) (0l)("0

(10 0

~Ip)

(~ 01) (01 Lp/np) l (; L31)

(01 L2

(~ fl

/3~5) (0

1)

(0

(0 l)( o 1 L2

1

1 L3

l)

(6.34) So (Duarte et al., 1997) A - a 2 + ((A + ,=.)a + 6A + fl)x + .=,Ax 2

(6.35)

B - (a2 + ~52)A+ (a + ~5)fl+ Ea~5+ ((E + A)(a + ~5)+ 2fl)Ax + EA2x 2 (6.36) C - (a + 6)X + ~-X2 D

-

~2

(6.37)

+ [(A + E)~5+ 6A + fl]X + "ZAx2

(6.38)

where

2(BMp/M) + 2(L3/M 2) A= (L1,/np) + L1 E = 2L2 +

(6.39) (6.40)

For an ideal laser gain medium with a - 6 ~ 1, X ~ 0, a n d / 3 - / 3 , A ,,~ 1 B ,~ 2A + 2/3 + E

(6.41) (6.42)

108


Grating

e~,f -..

Pump beam

// 1/I"

(~"

iI

/~1

U ',, I I

M

Solid-state gainmedium

Figure 6.19 Long-pulsesolid-state MPL grating laser oscillator. (From Duarte et al. 1998.)

which imply that the beam divergence will approach its diffraction limit as the optical length of the cavity B increases. This is in accordance with experimental observations.

6.2.8 MULTIPLE-RETURN-PAss BEAM DIVERGENCE The multiple-return-pass laser linewidth in a dispersive tunable laser oscillator is given by (Duarte, 2001) A A -- A O R ( R M V a O a + RV.xdPp) -1

(6.43)


109 Symmetry plane

Gain medium

'"'--..'

"',0 L3

..-" L2

L 1

Figure 6.20 Unfolded laser cavity for multiple-return-pass analysis.

where M is the overall intracavity beam magnification, R is the number of return passes, ~7~Oc is the grating dispersion, and 2 7 ~ p is the return-pass multiple-prism dispersion. Here, the multiple-return-pass beam divergence is given by A0R -- (A/ww)(1 + (Ly?/BR) 2

-}-(ARL~/BR)2)1/2

(6.44)

where AR and BR correspond to cumulative multiple-return-pass transfer matrix coefficients. For a multiple-return-pass analysis, the cavity is unfolded multiple times and the multiplication described for the single-return pass is performed multiple times. This leads to the following matrix components (Duarte, 2001): AR =(OzAR-1 -}- X~R-1 )(O~ -Jr-X(~ -- n2)) q-- X~R-1 (xL2 -q- 6)

+ XAR-I(o~L2 + fl) B~ =A~A + (~A~_~ + X ~ . _ . ) + (;3 + 6(Z - L2)) + 6 ~ . _ . (xL2 + ~)

(6.45) (6.46)

+ 6AR-1 (c~L2 + fl) where ~R-1 - AAR-1 + BR-1

(6.47)

For a single-return pass, A1 - (c~ + xA) (c~ + X(Z - L2)) + xA(xL2 + 6) + X(c~L2 + fl)

(6.48)

B, - A1A + (c~ + xA) + (fl + 6(E - L2)) + 6A(xL2 + 6) + 6(~L2 + fl) (6.49)

110


which reduce to Eqs. (6.35) and (6.36), respectively. Note that by definition A0 = 1 and B0 = 0. For an ideal gain medium, with little or no thermal lensing, a = 6 ~ 1, X --~ 0, and/3 =/3, so AR ~ 1

(6.50)

BR ~ R(ZA + 2/3 + E)

(6.51)

which are the multiple-pass versions of Eqs. (6.41) and (6.42), respectively. These results mean that, in the absence of thermal lensing, the beam divergence described by Eq. (6.44) will decrease toward its diffraction limit as the number of intracavity passes increases. A discussion on the application of these equations to low-divergence narrow-linewidth tunable lasers has been given by Duarte (2001).

6.2.9 UNSTABLERESONATORS The subject of unstable resonators is a vast subject and has been treated in detail by Siegman (1986). Unstable resonators are cavities that are configured with curved mirrors in order to provide intracavity magnification, which in turn provides good transverse-mode discrimination and good far-field beam profiles. In addition to their application as intrinsic resonators, unstable resonators are widely applied to conFigure the cavities of forced oscillators that amplify the emission from a master oscillator (see Chapter 7). Here, the discussion is limited to resonators incorporating Cassegrainian telescopes. These are reflective telescopes, as illustrated in Fig. 6.11, which are widely used in the field of astronomy. The double-pass or single-return-pass ray transfer matrix for a telescope comprising a concave back reflector and a convex exit mirror is given in Table 6.1. The radius of curvature of the mirrors is considered positive if concave toward the resonator. In addition to the transfer matrix, the following relations are useful (Siegman, 1986): M = -(Rz/R1)

L-

(R1 + R2) 2

(6.52) (6.53)

The condition for lasing in the unstable regime is determined by the inequality ](A + D)/2] > 1

(6.54)

111


In addition to traditional unstable resonators of the Cassegrainian type, multiple-prism-grating oscillators can also meet the conditions of an unstable resonator when incorporating a gain medium that exhibits thermal lensing (Duarte et al., 1997). On the other hand, in the absence of thermal lensing and for an ideal gain medium, c~ = 6 = 1, X = 0, and/3 =/3, the A and D terms can have a value of unity.

6.3 H I G H E R - O R D E R

MATRICES

The description of optical systems using 3 x 3, 4 x 4, and 6 x 6 matrices has been considered by several authors (Brouwer, 1964; Siegman, 1986; Wollnik, 1987). A more recent description of 4 x 4 matrices, by Kostenbauder (1990), uses the notation

OX2/OXl OX2/O01 Ox2/Otl OX2/OlJl O02/ OX l 002/OO01 O02/ Ot l 002/0//1 Otz/OXl Ot2/O01 Otz/Otl Otz/OUl OlQ/ ON1 OlJ2/ O01 0l'2 / 0 t 1 0112/ 0111

(6.55)

which can be written as

A C G 0

B D H 0

0 0 1 0

E F I 1

(6.56)

where the A B C D terms have their usual meaning. For a plane mirror this matrix becomes

1 0 0 0 0

1 0 0

0 0 1 0

0 0 0

1

(6.57)

112


and for a thin concave lens it becomes

If-~l

1

0 1

0 0

0 0

0 0

0 0

1 0 0 1

(6.58)

For a single mth prism, the components of the matrix are given by (Kostenbauder, 1990; Duarte, 1992) Am -- kl,mk2,m

(6.59)

nm --(k2,m/kl,m)(lm/nm)

(6.60)

Cm = O

(6.61)

Om -- (kl,mk2,m) -1

(6.62)

Em -- BmQ-~l,m(Onm/Oll)

(6.63)

Fm = V A~2,m ( O)~/ OtJ)

(6.64)

Gm -- Amrm//~

(6.65)

n m -- Bm( ~gc~2,m/~)(Onm/OV)

(6.66)

Im -- n m ~ l , m ( O F l m / O l j ) --(lm/V2)(Ol~/Oll)

(6.67)

where (fil,m)/rlm

(6.68)

~-g~2,m ~- (tan (fi2,m)/nm

(6.69)

~Ct~ m = ( t a n

V A~)2,m = O~2,m / O)k

(6.70)

Here, Em is a function of the optical length Bm multiplied by the geometrical factor c'~l,m and (Onm/OZ,'). A more identifiable term is Fm, which is a function of the angular dispersion of the prism V~b2,m multiplied by 0A/0u. The third row of the matrix, containing the terms Gm, nm, and Im, provides temporal information on the propagation through the prism, which is useful when evaluating prisms as pulse compressors in ultrafast lasers. The


113

dependence of the temporal component Gm on dispersion can be made explicit by rewriting Eq. (6.65) as (6.71)

Gm = AmVA~2,m(O)~/Ov),~ -1

The dependence of Im on the optical length Bm is made explicit by restating Eq. (6.67) as Im -- B m ( ~ 2 , m ~ l , m / ) k ) ( O n m / O l l )

2 -- (lm/'V2)(Ov/Ol I)

(6.72)

For a generalized multiple-prism array comprising r prisms, the elements become (Duarte, 1992) (6.73)

Ar = M1 M2

r l(Lm

Br -- M1 M2 ~

m=l

n)

kl,j j=l

-2

k2,j j=l

(6.74) Cr = 0

(6.75)

Dr -- (MtM2) -1

(6.76)

Fr = V ~ 2 , r ( 0 ~ / 0 ~ )

(6.77)

Gr --

Fm m=l

k l,j j=l

kz,j

(6.78)

A-1

j=l

where ~7A~2, r is the generalized single-pass multiple-prism dispersion given by (Duarte and Piper, 1982; Duarte, 1988) ~TA~2,r=(MIM2) -1

(--1-1 ) ~q~2,m m=l

Hkl,jHk2,j j=l

~TAFlm

j=l

-1

(6.79)

where V~nm = Onm/OA

(6.80)

The explicit E~, Hr, and Ir terms for a generalized multiple-prism array are a function of Br. These terms are rather extensive and thus are not included in the text.

114


PROBLEMS 1. Derive the ray transfer matrix for the Galilean telescope given in Table 6.1. 2. Derive the single-pass ray transfer matrix corresponding to a multipleprism beam expander, deployed in an additive configuration and comprising four prisms designed for orthogonal beam exit. 3. Perform the matrix multiplication for the intracavity single-return pass of the multiple-prism-grating tunable laser oscillator depicted in Fig. 6.19 that results in Eqs. (6.35) to (6.38). 4. Show that Eqs. (6.48) and (6.49) reduce to the single-pass A and B terms given by Eqs. (6.35) and (6.36). 5. Use the temporal G, H, I components to write down an expression for t2 in a double-prism pulse compressor.

REFERENCES Brouwer, W. (1964). Matrix Methods in Optical Design. Benjamin, New York. Duarte, F. J. (1988). Transmission efficiency in achromatic nonorthogonal multiple-prism laser beam expanders. Opt. Commun. 71, 1-5. Duarte, F. J. (1989). Ray transfer matrix analysis of multiple-prism dye laser oscillators. Opt. Quantum Electron. 21, 47-54. Duarte, F. J. (1990). Narrow-linewidth pulsed dye laser oscillators. In Dye Laser Principles (Duarte, F. J., and Hillman, L. W., eds.). Academic Press, New York, pp. 133-183. Duarte, F. J. (1991). Dispersive dye lasers. In High-Power Dye Lasers (Duarte, F. J., ed.). Springer-Verlag, Berlin, pp. 7-43. Duarte, F. J. (1992). Multiple-prism dispersion and 4 x 4 ray transfer matrices. Opt. Quantum Electron. 24, 49-53. Duarte, F. J. (1999). Multiple-prism-grating solid-state dye laser oscillator: optimized architecture. Appl. Opt. 38, 6347-6349. Duarte, F. J. (2001). Multiple-return-pass beam divergence and the linewidth equation. Appl. Opt. 40, 3038-3041. Duarte, F. J., and Piper, J. A. (1982). Dispersion theory of multiple-prism beam expanders. Opt. Commun. 43, 303-307. Duarte, F. J., Costela, A., Garcia-Moreno, I., Sastre, R., Ehrlich, J. J., and Taylor, T. S. (1997). Dispersive solid-state dye laser oscillators. Opt. Quantum Electron. 29, 461-472. Duarte, F. J., Taylor, T. S., Costela, A., Garcia-Moreno, I., and Sastre, R. (1998). Long pulse narrow-linewidth dispersive solid-state dye-laser oscillator. Appl. Opt. 37, 3987-3989. Kogelnik, H. (1979). Propagation of laser beams. In Applied Optics and Optics Engineering (Shannon, R. R., and Wyant., J. C., eds.). Academic Press, New York, pp. 155-190. Kostenbauer, A. G. (1990). Ray-pulse matrices: a rational treatment for dispersive optical systems. IEEE J. Quantum Electron. 26, 1148-1157. Siegman, A. (1986). Lasers. University Science Books, Mill Valley, CA. Wollnik, H (1987). Optics of Charged Particles. Academic Press, New York.

Chapter 7

Pulsed Narrow-Linewidth Tunable Laser Oscillators

7.1 I N T R O D U C T I O N In this chapter the basics on interference, the uncertainty principle, polarization, and beam propagation are applied to the design, architecture, and engineering of narrow-linewidth tunable lasers. The principles discussed here apply in general to tunable lasers and are not limited by the type or class of gain media. Thus, the gain media assumed here are generic broadly tunable media either in the gas, liquid, or solid state. A narrow-linewidth tunable laser oscillator is defined as a source of highly coherent continuously tunable laser emission, that is, a laser source that emits highly directional radiation of an extremely pure color. Pure emission in the visible spectrum is defined as having a linewidth narrower than A u ~ 3GHz, which translates approximately into AA ~ 0 . 0 0 1 7 n m at 510nm. This criterion is provided by the laser linewidth requirements to + - X 112g + electronic transition excite single vibrorotational levels in the B 3IIou of the iodine molecule at room temperature. Although most of the discussion is oriented toward high-power pulsed tunable lasers, mention of continuous-wave (CW) lasers is also made when appropriate. 115

116


7.2 T R A N S V E R S E A N D L O N G I T U D I N A L M O D E S 7.2.1 TRANSVERSE-MODE STRUCTURE The most straightforward laser cavity is that comprising a gain medium and two mirrors, as illustrated in Fig. 7.1. The physical dimensions of the intracavity aperture relative to the separation of mirrors, or cavity length, determine the number of transverse electromagnetic modes. The narrower the intracavity aperture and the longer the cavity, the lower the number of transverse modes. The single-pass transverse-mode structure in one dimension can be characterized using the generalized interferometric equation introduced in Chapter 2 (Duarte, 1991a, 1993a),

I(XlS>I2-- j--~l ~(rj)2 -+-2 j~l ~(rJ)'=

m=j+l ql(rm) COS(Qm - Qj)

(7.1)

and in two dimensions by (Duarte, 1995a) N

N

L- Z Z z=l y=l

N

N

(rz l Z E q=l p = l

(7.2)

In addition, a useful tool to determine the number of transverse modes is the Fresnel number (Siegman, 1986), (7.3)

NF = wZ/(LA)

The single-pass approximation to estimate the transverse-mode structure assumes that in a laser with a given cavity length most of the emission generated next to the output-coupler mirror is in the form of spontaneous emission and thus highly divergent. Thus only the emission generated at the opposite end of the cavity and that propagates via an intracavity length L contributes to the initial transverse-mode structure.

D=2w

M1

Gain medium

I

M2

Figure 7.1 Mirror-mirror laser cavity. The physical dimensions of the intracavity aperture relative to the cavity length determine the number of transverse modes.


117

In order to illustrate the use of these equations, let us consider a hypothetical laser with a 10-cm cavity emitting at A = 632.8 nm incorporating a 4-mmwide one-dimensional aperture. Using Eq. (7.1), the intensity distribution of the emission is calculated as shown in Fig. 7.2. Each ripple represents a transverse mode. An estimate of this number can be obtained by counting the ripples in Fig. 7.2, which yields an approximate number of 63. This number of ripples should be compared with the Fresnel number, which is 63.21. The wavelength A = 632.8 nm corresponds to the well-known 3S2-2p4 transition of the He-Ne laser. A He-Ne laser with the cavity dimensions just given would be highly unusual. Atomic gas lasers, such as the Ne-Ne laser, are characterized by long cavities and narrow emission beams. Dimensions of a realistic He-Ne laser are a cavity length of L = 40 cm and an emission 60.0 -

50.0 -

40.0 ~

r r

.>_

30.0

.,..,.=

tr

20.0 -

10.0

.J .........

-35.0

'1"'

-25.0

!

-15.0

I

I

-5.0

I

0

!

5.0

a

!

15.0

l

s

25.0

i

-

a

35.0

Screen Axial Distance (meters) x 10 .4 Figure 7.2 Cross section of diffraction distribution corresponding to a large number of transverse modes. [Reprinted from Duarte (1993a), copyright 1993, with permission from Elsevier].

118


50.0 40.0.m c(9

(9

.>_

30.0-

(9

n-" 2 0 . 0 -

10.0 -

0.0 -10.0

I

-5.0

I

0.0

I

5.0

10.0

Screen Axial Distance (m) x 10 .4

Figure 7.3 Cross section of diffraction distribution corresponding to a near-single-transverse

mode (NF

,~

0.25).

beam with w = 0.25 mm. For such dimensions the calculated intensity distribution, using Eq. (7.1), is given in Fig. 7.3. In this case the Fresnel number becomes NF '~ 0.25. The distribution in Fig. 7.3 indicates that most of the emission intensity is contained in a central near-Gaussian distribution. In practice the spatial distribution of the emission for this He-Ne laser is a near-Gaussian distribution characteristic of what is known as TEM00 emission. In general, continuous-wave lasers using gaseous gain media that emit via atomic transitions yield TEM00 emission. Examples of such coherent sources are the He-Ne, He-Cd, and He-Zn lasers. Reducing the transverse-mode distribution to TEM00 emission is the first step in the design of narrow-linewidth tunable lasers. The task of the designer consists in achieving TEMo0 emission in the shortest possible cavity length.

7.2.2 LONGITUDINAL-MODE EMISSION Once single-transverse-mode emission has been established, the task consists in controlling the number of longitudinal modes in the cavity. In a laser with cavity length L, the longitudinal-mode spacing in the frequency domain is given by 6 v - c/2L

(7.4)


119

and the number of longitudinal modes NLM is given by -

Lx.l .

(7.5)

where Au is the measured laser linewidth. Thus, for a laser with a cavity 30 cm long and a measured linewidth of Au = 3 GHz, the number of longitudinal modes becomes NLM ~ 6. If the cavity length is reduced to 10 cm, then the number of longitudinal modes is reduced to NLM ~ 2 and the emission would be called double-longitudinal-mode (DLM) emission. If the cavity length is reduced to 5 cm, then NLM ~ 1 and the laser is said to be undergoing single-longitudinal-mode (SLM) oscillation. These simple examples highlight the advantages of compact cavity designs, provided the active medium can sustain the gain to overcome threshold. An alternative to reducing the cavity and still achieve SLM emission is to optimize the beam divergence and to increase the intracavity dispersion to yield a narrower cavity linewidth that would restrict oscillation to the SLM regime. In this context the linewidth AA ~ A0(V~0) -1

(7.6)

is converted to Au using the identity

2)

(7.7)

and applying the criterion Au _< 6u

(7.8)

to guide the design of the dispersive oscillator. Multiple-longitudinal-mode emission appears complex and chaotic in both the frequency and temporal domains. Double-longitudinal-mode and SLM emission can be characterized in the frequency domain using FabryPerot interferometry or in the temporal domain by observing the shape of the temporal pulsed. In the case of DLM emission, the interferometric rings appear to be double. In the temporal domain, mode beating is still observed when the intensity ratio of the primary to the secondary mode is 100:1 or even higher. Mode beating of two longitudinal modes, as illustrated in Fig. 7.4, can be characterized using a simple wave representation (Pacala et al., 1984), where each mode of amplitudes E1 and E2, with frequencies ~1 and ~2, combine to produce a resulting field E - E1 cos(~lt - klz) + E2 c o s ( ~ 2 t - k2z)

(7.9)

120


Figure 7.4 Mode beating resulting from double-longitudinal-mode oscillation. (a) Temporal pulse. (b) Calculated temporal pulse assuming interference between the two longitudinal modes. (From Duarte et al., 1988.)

For incidence at z = 0 on a square law temporal detector, the intensity can be expressed as

E 2 - ( E 2 -+- E 2 ) / 2

+

(E 2 cos 2w1 t -+- E 2 cos w21)/2

-k E1 E2 cos(col Jr- w2) t + E1 E2 cos(w1 - w2) t

(7.10)


121

Detectors in the nanosecond regime respond only to the first and the last terms of this equation, so Eq. (7.10) can be approximated by E 2 ~ (E~ + E2)/2 + E1E2 cos(~l - ~ 2 ) t

(7.11)

Using this approximation and a non-Gaussian temporal representation derived from experimental data, for the amplitudes of the form E1 (t) -- (a2t 2 + alt + ao)(blt + bo) -1

(7.12)

a calculated version of the experimental waveform exhibiting mode beating can be obtained as shown in Fig. 7.4b. For this particular dispersive oscillator lasing in a double longitudinal mode, the ratio of frequency jitter 5~ to cavity mode spacing A~ ~ (~l - ~2) was represented by a sinusoidal function at 20 MHz. The initial mode intensity ratio is 200:1 (Duarte et al., 1988; Duarte, 1990a). In the case of SLM emission, the Fabry-Perot interferometric rings appear singular and well defined (see Fig. 7.5). Mode beating in the temporal domain is absent, and the pulses assume a near-Gaussian distribution (see Fig. 7.6). These results were obtained in an optimized solid-state multipleprism grating dye laser oscillator for which A u A t ~ 1 (Duarte, 1999).

Figure 7.5 Fabry-Perotinterferogram corresponding to single-longitudinal-mode emission at Au ~ 350 MHz. (From Duarte, 1999.)


122

Figure 7.6 Near-Gaussiantemporal pulse corresponding to single-longitudinal-modeemission. The temporal scale is 1ns/div. (From Duarte, 1999.)

7.3 T U N A B L E

LASER

OSCILLATOR

ARCHITECTURES

Tunable laser oscillators can be configured in a variety of cavity designs. Here, a brief survey of these alternatives is presented. In general, these cavity architectures can be classified into tunable laser oscillators without intracavity beam expansion and tunable laser oscillators with intracavity beam expansion (Duarte, 1991 a). Further, each of these classes can be divided into open- and closed-cavity designs (Duarte and Piper, 1980). It is assumed that oscillators considered in this section are designed to yield narrow-linewidth emission. The architectures are relevant to a variety of tunable gain media and are applicable to both CW and pulsed lasers.

7.3.1 TUNABLELASER OSCILLATORSWITHOUT INTRACAVITY BEAM EXPANSION Tunable laser oscillators without intracavity beam expansion are those laser resonators in which the intrinsic narrow beam waist at the gain region is not expanded using intracavity optics. The most basic of tunable laser designs is that incorporating an output mirror coupler and a tuning grating in Littrow configuration, as illustrated in Fig. 1.5. Tuning is accomplished by slight rotation of the grating. This cavity configuration will yield relatively broad


123

Figure 7.7 Grating-mirror tunable laser cavities. [Reprinted from Duarte (1995b), copyright 1995b, with permission from Elsevier].

tunable emission in a short pulsed laser, such as a high-power pulsed dye laser, but could emit fairly narrow emission if applied to a CW semiconductor laser. A refinement of this cavity consists in inserting one or more intracavity etalons to further narrow the emission wavelength, as illustrated in Figs. 7.7 and 7.8. Such multiple-etalon grating cavities can yield very narrow-linewidth pulsed emission, down to the SLM regime, in high-power pulsed lasers. The introduction of the etalons provides further avenues of wavelength tuning. The main disadvantage of this class of resonator in the pulsed regime is the very high intracavity power flux that can induce optical damage in the grating and the coating of the etalons. An important configuration in the tunable laser oscillators without intracavity beam expansion class that employs only the natural divergence of the intracavity beam for total illumination of the diffractive element is the grazing-incidence grating design (Shoshan et al., 1977; Littman and Metcalf,

Figure 7.8 Mirror-mirrorlaser cavity incorporating intracavity etalons.


124

1978). In these lasers the grating is deployed at a high angle of incidence. The diffracted beam is subsequently reflected back toward the grating by the tuning mirror. A variation on this design is the replacement of the tuning mirror by a grating deployed in Littrow configuration (Littman, 1978). This cavity has been configured as an open cavity, as shown in Fig. 7.9a, or as a closed cavity, as shown in Fig. 7.9b. A further alternative is the inclusion of an intracavity etalon (Saikan, 1978) for further linewidth narrowing. Grazing-incidence grating cavities have the advantage of being fairly compact and are widely used both in the pulsed regime and in the CW regime. A limitation of these cavities is the relatively high losses associated with the deployment of the diffraction grating at a high angle of incidence, as illustrated in Fig. 7.10 (Duarte and Piper, 1981). For tunable laser oscillators without intracavity beam expansion and in the absence of intracavity etalons, the cavity linewidth equation takes the fairly simple form of ,/kA '~ AOR(RVAOG) -1

(7.13)

where V~Oa is the grating dispersion either in Littrow configuration,

Figure 7.9 Grazing-incidence grating cavities. (a) Open cavity. (b) Closed cavity. In these oscillators, the refraction angle identified as 0' corresponds to 9 [Reprinted from Duarte (1990a), copyright 1990, with permission from Elsevier].


Grating efficiency

50

>,

125

4O

O r ~

._~

uJ

30

-

I

20 10 5 I

60

I

70 Incidence Ange (|

I

80

~,

,

, i,

85

,7-'~1

90

Figure 7.10 Grating efficiency curve as a function of angle of incidence at A = 632.8 nm. (From Duarte and Piper, 1981.)

V~OG -- (2 tan O)/A

(7.14)

V,xOG -- m/(d cos O)

(7.15)

or in grazing-incidence configuration, VAOG -- 2(sin O + sin ~ ) / ( A cos O) ~7,~OG -- 2m/(d cos O)

(7.16)

(7.17)

where the multiple-return-pass beam divergence A0R is given by

AOR --(A/Trw) (1 + (L~/BR) 2 + (ARL~/BR) 2) 1/2

(7.18)

and the appropriate propagation terms can be calculated according to the optical architecture of the cavity, as described in Chapter 6. For tunable oscillators incorporating a diffraction grating and an intracavity etalon, the linewidth established via Eq. (7.13) can be used to provide Au via Eq. (7.7) and thus to determine the free spectral range (FSR) of the etalon, which is given by (Born and Wolf, 1999)

F S R - c/2nle

(7.19)

where n is the refractive index of the etalon's material and le is the distance between the reflective surfaces. The expression for the FRS has its origin in A u - c/Ax.


126

The minimum resolvable linewidth or resulting laser linewidth obtainable from the etalon is given by

Ab'FRS

-

-

FRS/~

(7.20)

where ~ is the effective finesse of the etalon. The finesse of the etalon is a function of the flatness of the surfaces (often in the range of A/100 - A/50), the dimensions of the aperture, and the reflectivity of the surfaces. The effective finesse is given by (Meaburn, 1976) OY -~-2 -- ~---~R2 + ~ F 2 --I--~--~A2

(7.21)

where ~ R , ~ F , and ~ A are the reflective, flatness, and aperture finesses, respectively. The reflective finesse is given by (Born and Wolf, 1999) ~g

= ~-x/R/(1 - R)

(7.22)

where R is the reflectivity of the surface. Further details on Fabry-Perot etalons are given in Chapter 11. Multiple-etalon systems are described in detail in the literature (Maeda etal., 1975; Pacala etal., 1984). As implied earlier, these multiple-etalon assemblies are designed so that the FRS of the etalon to be introduced is compatible with the measured laser linewidth attained with the previous etalon or etalons. Although this is a very effective avenue to achieve fairly narrow linewidths, the issue of optical damage due to high intracavity power densities does introduce limitations. The performance of representative tunable laser oscillators without intracavity beam expansion is summarized in Table 7.1.

7.3.2

TUNABLE LASER OSCILLATORS WITH INTRACAVITY

BEAM EXPANSION Equation 7.3 indicates that a key principle in achieving narrow-linewidth emission consists of augmenting the intrinsic intracavity dispersion provided by the diffraction grating. This is accomplished by the total illumination of the diffraction surface of the tuning element. In the case of the grazingincidence grating cavities, this is done by deploying the grating at a high angle of incidence; however, that can be associated with low diffraction efficiencies. An alternative method is to illuminate the diffractive element via intracavity beam expansion. Tunable laser oscillators with intracavity beam expansion are divided into two subclasses: those using two-dimensional beam expansion and those using one-dimensional beam expansion. Initially, intracavity beam expansion was accomplished utilizing twodimensional beam expansion and a diffraction grating deployed in Littrow

127

Pulsed Narrow-Linewidth Tunable Laser Oscillators Table 7.1

Performance of Tunable Laser Oscillators Without Intracavity Beam Expansion Gain medium

Cavity A (nm) configuration

Gas lasers XeC1 XeC1 CO2

GI 3 etalons G!

Liquid lasers Rh 590 GI

Tuning range (nm)

Reference

H1 GHz 4 mJ HI50 MHz 2-5 gJ 117 MHz 140mJ

Sugii et al. (1987) Pacala et al. (1984) Duarte (1985b)

600

300 MHz

Littman (1978)

746-918

Semiconductor lasers GaA1As etalon a GI a

Output r/(%) energy

308 308 10,591

Solid-state lasers Ti : A1203 G!

GaA1As

Au

~1.5 GHz

2 mJ

Harrison and Mooradian (1989) Harvey and Myatt (1991)

4 kHz 780

20@780nm

Kangas et al. (1989)

10kHz

a CW regime.

configuration, as illustrated in Fig. 7.11 (H/insch, 1972). Two-dimensional beam expansion has also been demonstrated using reflection telescopes (Beiting and Smith, 1979; Trebino etal., 1982). Transmission telescopes can be either of the Galilean or astronomical class; the reflection telescope can be of the Cassegrainian type (see Chapter 6). The main advantage of this approach is the significant reduction of intracavity energy incident on the tuning grating, thus vastly reducing the risk of optical damage. The main disadvantages of the two-dimensional intracavity beam expansion is the requirement of expensive circular diffraction gratings, a relatively difficult alignment process, and the need for long cavities. Since the telescopes mentioned here can provide low dispersion, the cavity linewidth equation reduces to /kA ~ AOR(RMV)~OG)

-1

(7.23)

where V~Oc is the grating dispersion either in Littrow or grazing incidence, as given in Eqs. (7.14)-(7.17), and the multiple return-pass beam divergence AOR is given by Eq. (7.18). Certainly, it should be indicated that for a single return pass Eq. (7.23) becomes the original expression derived by H/insch (1972), /k)~ ~ / k O ( M V ~ O G ) -1

(7.24)

128


Figure 7.11 Two-dimensional transmission telescope Littrow grating laser cavity. (From Duarte, 1990a.) One-dimensional intracavity beam expansion uses multiple-prism beam expanders rather than conventional telescopes to perform the beam expansion. Multiple-prism grating tunable laser oscillators are classified as multipleprism Littrow (MPL) grating laser oscillators (Klauminzer, 1978; Kasuya et al., 1978; Wyatt, 1978; Duarte and Piper, 1980) and hybrid multiple-prism grazing-incidence (HMPGI) grating laser oscillators (Duarte and Piper, 1981, 1984a). The MPL and HMPGI grating laser oscillators are depicted in Figs. 7.12 and 7.13, respectively. Both of these oscillator subclasses belong to the closed-cavity class. In these laser oscillators the intracavity beam expansion is onedimensional, thus facilitating the alignment process significantly. In addition, the requirements on the dimensions of the diffraction grating perpendicular

Figure 7.12 Optimizedcompact MPL grating solid-state dye laser oscillator. (From Duarte, 1999.)

129


Solid-state gain medium

Grating

_

,,

_~1 M~

\\x[

! ( '~ ~ ~ "

/

Tuning mirror

Figure7.13 Solid-stateHMPG[ gratingdye laseroscillator.[Reprintedfrom Duarte (1997), copyright 1997,with permissionfrom Elsevier]. to the plane of incidence are reduced significantly. Another advantage is compactness, since these high-power tunable laser oscillators can be configured in architectures requiring cavity lengths in the range of 50-100 mm. The first two MPL grating oscillators illustrated in Fig. 4.4 depict the difference in architecture between two compensating multiple-prism configurations yielding near-zero prismatic dispersion at a given design wavelength. Both these oscillators provide rather large intracavity beam expansion factors, which in practice can be in the range 100 _< M l 2 - ~

~(rj) 2 q- 2 ~

~(rj)

~(rm)COS(am -- ~j)

m=j+l


148

2. The larger the optical length of the cavity, the lower the beam divergence, or A0R --(A/Trw)(1 + (L~/BR) 2 + (ARL~/BR) 2)

1/2

3. For diffraction-limited TEM00 emission, the narrower the beam waist w, the larger the beam divergence, or AO = A/Trw

4. The larger the beam magnification, the larger the intracavity dispersion and the narrower the linewidth, or /XA -- A O R ( R M V , x O G -+- R V A ~ p ) -1

5. Also, from the previous axiom, the lower the beam divergence, the narrower the linewidth. 6. From the second and fourth axioms, the larger the number of intracavity passes, the lower the beam divergence and the narrower the linewidth. 7. The shorter the cavity, the longer the longitudinal-mode spacing, or &, = c / 6 x

These axioms clearly illustrate that some of the design parameters have a competing effect on the overall physics. The task of the designer is to apply these principles in a balanced approach to optimize the beam divergence and linewidth performance in a compact-cavity architecture.

7.7 N A R R O W - L I N E W I D T H OSCILLATOR-AMPLIFIERS The dispersive tunable laser oscillators so far described yield singlelongitudinal-mode emission at very low levels of ASE. However, for many applications high energies or high-average powers are required. For that purpose the exquisite emission from the oscillator must be amplified by one or several amplification stages. A review on this subject and its literature is given by Duarte (1990b). Here, the focus will be on the fundamentals and the performance of various representative systems.

7.7.1 LASER-PUMPEDNARROW-LINEWIDTHOSCILLATOR-AMPLIFIER CONFIGURATIONS Amplification of coherent optical radiation in laser-excited systems is well illustrated by the configurations developed for dye lasers. Two of the most

149


interesting features of these systems is that amplification is performed in a single pass and that several stages of amplification are often employed. As such, given that lasing in these systems occurs in the nanosecond regime, it is important to synchronize the arrival of the oscillator pulse with the excitation of the amplifier. This is arranged by allowing the excitation geometry to delay the pump pulse, as illustrated in Fig. 7.25. An additional aspect important to the design of multistage oscillator-amplifier systems is the geometrical matching of the oscillator or preamplified beam with the focused excitation laser at the corresponding amplifier stage. This helps to maintain the cumulative ASE at low levels. For optimum efficiency, proper distribution of the pump energy is required, with only a fraction (often less than 5 %) of it used to excite the oscillator. Correct polarization matching is also important. The performances of illustrative multistage oscillator-amplifier laser systems are listed in Table 7.3. Bos (1981) reports 6% efficiency at the oscillator, 20% at the preamplifier, and 60% at the amplifiers. The overall gain factor is about 229. The copper-vapor-laser pumped dye laser reported from Lawrence Livermore (Bass et al., 1992) operates at a pulse-repetition frequency of 13.2 kHz and comprises several master-oscillator power-amplifier (MOPA) chains performing at a 50-60% overall conversion efficiency. The master oscillators are of the MPL grating class and incorporate an intracavity etalon, and each MOPA chain includes three to four amplifiers in series. Although most oscillator-amplifiers systems considered here utilize highperformance pulsed master oscillators, the alternative of semiconductor laser oscillators lasing in the CW regime is also available, as demonstrated Tunable oscillator

AC1

F

AC2

F

ACa

BSO

beam

Figure 7.25 Single-pass multiple-stage laser amplifier.

150


Table 7.3 Performance of Laser-Pumped Narrow-Linewidth Oscillator-Amplifier Configurations Oscillator A (nm) configuration

Au

Telescopicc HMPGI

590 440

MPU

~590 0.5-5GHz

Amplification Gain Output Output r/(%) Reference stages energya powerb

320 MHz 650MHz

3 2

229 165 mJ ~700 3.5mJ

3-4 d

55 ~9

Bos (1981) Dupre (1987) 2.5 kW@ 50-60 Glass et al. 13.2 kHz (1992)

a Per pulse. bAverage power. eIncludes intracavity etalon. dAt each of four amplification chains.

by F a r k a s and Eden (1993). These authors used a five-stage dye laser amplification system to produce pulses of 1.2mJ at 786nm, with A v = 118 MHz.

7 . 7 . 2 NARROW-LINEWIDTH MASTER-OSCILLATOR FORCEDOSCILLATOR CONFIGURATIONS The master oscillator (MO) comprises the narrow-linewidth dispersive laser oscillators already discussed. The forced oscillator (FO), on the other hand, is an amplifier stage comprising a gain region within a resonator, as depicted in Fig. 7.26. The resonator of the amplifier stage can be an unstable resonator. Several aspects are rather critical to the efficient frequency locking of these configurations. First, the alignment of the master oscillator relative to the forced oscillator must be concentric. Second, there are stringent require-

Figure 7.26 Master-oscillator forced-oscillator laser configuration. (From Duarte and Conrad, 1987.)

151


ments on the timing of the excitation that impose arrival of the masteroscillator pulse at the onset of the forced-oscillator pulse buildup. Efficient frequency locking also occurs. Optimum lasing is achieved when the emission wavelength of the master oscillator is tuned to the central wavelength of the gain spectrum of the forced oscillator. The performance of some representative master-oscillator forced-oscillator systems is described on Table 7.4. The forced-oscillator cavity depicted in Fig. 7.26 is configured after a Cassegrainian telescope and is cataloged as a confocal u n s t a b l e r e s o n a t o r of the positive branch (Siegman, 1986). Here the radius of curvature of the large concave mirror is R2, and the radius for the small mirror is R1 and has a negative value. The magnification of the resonator is given by (Siegman, 1986) M -

(7.69)

-(Rz/R1)

and its length is L - - (R1 + R2)/2

(7.70)

As discussed by Siegman (1986), the condition for oscillation in the unstable regime is satisfied by I(A + O)/21 > 1

(7.71)

where A and D are the matrix elements introduced in Chapter 6. For the and R 2 - 4m, so M - 2 and resonator depicted in Fig. 7.26, R 1 - - 2 m L - 1 m, and the condition for lasing in the unstable regime is satisfied.

Table 7.4 Performance of Narrow-Linewidth Master-Oscillator Forced-Oscillator Configurations MO configuration

A (nm)

Two etalons

589

346 MHz

MPL

590

O)2

%

Figure 8.2 Optical configuration for sum-frequency generation.

Equation (8.35) illustrates the nonlinear dependence of the frequencydoubled output on the input signal and indicates its relation to the (L Ak/2) parameter. This dependence implies that conversion efficiency decreases significantly as (L Ak/2) increases. The distance

L~ = 21Ak

(8.36)

referred to as the coherence length of the crystal, provides a measure of the length of the crystal necessary for the efficient generation of secondharmonic radiation. Sum-frequency generation is outlined in the third term of Eq. (8.4) and involves the interaction of radiation at two different frequencies in a crystal to produce radiation at a third distinct frequency. This process, illustrated schematically in Fig. 8.2, consists of the normal incidence radiation of ~1 and &2 onto a nonlinear crystal to yield collinear output radiation of frequency ~3 = w1 + ~2. Using the appropriate expressions for Em(z, t) and Pm(z, t) in the wave equation, it can be shown that (Boyd, 1992) Ak = k l -t- k2 - k3

(8.37)

and the output intensity again depends on sincZ(L Ak/2). The ideal condition of phase matching is achieved when Ak=0

(8.38)

and it offers the most favorable circumstances for a high conversion efficiency. When this condition is not satisfied, there is a strong decrease in the efficiency of sum-frequency generation.

8.2.2 DIFFERENCE-FREQUENCYGENERATIONAND OPTICAL PARAMETRICOSCILLATION The process of difference-frequency generation is outlined in the fourth term of Eq. (8.4) and involves the interaction of radiation at two different frequencies in a crystal to produce radiation at a third distinct frequency. This process, illustrated schematically in Fig. 8.3, consists of the normal incidence

163

Nonlinear Optics I., I~

L

._1 -I

wl

~._ oJ2 = coa - co1

(~

%

Figure 8.3 Opticalconfiguration for difference-frequency generation. radiation of a~l and co3 onto a nonlinear crystal to yield collinear output radiation of frequency co2 ~- aJ3 - a)l. Assuming that co3 is the frequency of a high-intensity pump-laser beam, which remains undepleted during the excitation process, then A3 can be considered a constant; using an analogous approach to that adopted in the previous section, it is found that (Boyd, 1992)

i(87rd~2/kl c2)A3A~e iAt:z

(8.39)

dA2 / d z - i( 87rdaj2 /k2c2)A 3A*l eiAl~z

(8.40)

dA3/dz - O

(8.41)

dA1/dz

-

where Ak -- k3 - k2 - kl

(8.42)

If the nonlinear crystal involved in the process of frequency difference is deployed and properly aligned at the propagation axis of an optical resonator, as illustrated in Fig. 8.4, then the intracavity intensity can build to very high values. This is the essence of an optical parametric oscillator (OPO). Early papers on OPOs are those of Giordmaine and Miller (1965), Akhmanov etal. (1966), Byer etal. (1968), and Harris (1969). Recent reviews are given by Barnes (1995) and Orr et al. (1995). In the OPO literature, co3 is known as the p u m p frequency, a~l as the idler frequency, and adZ as the signal frequency. Thus, Eq. (8.42) can be restated as A k -- kp - k s - k1

(8.43)

Crystal

M2

M1

Figure 8.4 Basicoptical parametric oscillator configuration.


164

Equations (8.39) and (8.40) can be used to provide equations for the signal under various conditions of interest. For example, for the case when the initial idler intensity is zero and Ak ~ 0, it can be shown that ,

1

,

As(L)As(L ) ~ -~As(O)As(O)(e "YL+ e-'~L) 2

(8.44)

where As(O) is the initial amplitude of the signal. Here, the parameter 7 is defined as (Boyd, 1992) 7--

(647r2d2co2i~O2kllkslc-41ApI2) 1/2

(8.45)

Equation (8.44) indicates that for the ideal condition of Ak ~ 0, the signal experiences an exponential gain as long as the pump intensity is not depleted. Frequency selectivity in pulsed OPOs has been studied in detail by Brosnan and Byer (1979) and Barnes (1995). Wavelength tuning by angular and thermal means is discussed by Barnes (1995). Considering the frequency difference 03 S =

Cdp - - OdI

and Eq. (8.43), it can be shown that for the case of Ak ,,~ 0 (Orr ,)kS ~ ,)kp(n S -- t11)/(11P -- 111)

(8.46)

et al., 1995), (8.47)

which illustrates the dependence of the signal wavelength on the refractive indices. An effective avenue to change the refractive index is to vary the angle of the optical axis of the crystal relative to the optical axis of the cavity, as indicated in Fig. 8.4. For instance, Brosnan and Byer (1979) report that changing this angle from 45 ~ to 49 ~ in a Nd:YAG laser-pumped LiNbO3 OPO tunes the wavelength from ~2 lam to beyond 4 gm. The angular dependence of refractive indices in uniaxial birefringent crystals is discussed by Born and Wolf (1999). It should be mentioned that the principles discussed in Chapter 4 and 7 can be applied toward the tuning and linewidth narrowing in OPOs. However, there are some unique features of nonlinear crystals that should be considered in some detail. Central to this discussion is the issue of phase matching, or allowable mismatch. It is clear that a resonance condition exists around Ak ,,~ 0, and from Eq. (8.44) it is seen that the output signal from an OPO can experience a large increase when this condition is satisfied. Thus, Ak ~ 0 is a desirable feature. Here it should be mentioned that some authors define slightly differently what is known as allowable mismatch. For instance, Barnes (1995) defines it as

A k = Tr/L

(8.48)

165

Nonlinear Optics

which is slightly broader than the definition given in Eq. (8.36). The discussion on frequency selectivity in OPOs benefits significantly by expanding Ak in a Taylor series (Barnes and Corcoran, 1976) so that A k = Ako + ( O A k / O x ) A x + (1/2!)(OZAk/OxZ)Ax 2 + . . .

(8.49)

Here this process is repeated for other variables of interest 2xk = &ko + (02xk/OO)2xO + (1/2!)(022xk/OO2)2xO 2 + . . . A k = Ako + (OAk/OA)AA + (1/2!)(02Ak/OA2)AA 2 -+-... A k = Ako + ( O A k / O T ) A T +(1/2!)(02Ak/OT2)AT 2 + . . .

(8.50) (8.51) (8.52)

Equating the first two series and ignoring the second derivatives, it is found that (Barnes, 1995) AA-

AO(OAk/OO)(OAk/OA) -1

(8.53)

This linewidth equation shows a dependence on the beam divergence, which is determined by the geometrical characteristics of the pump beam and the geometry of the cavity. It should be noted that this equation provides an estimate of the intrinsic linewidth available from an OPO in the absence of intracavity dispersive optics or injection seeding from external sources. Barnes (1995) reports that for a AgGaSe 2 0 P O pumped by a Er:YLF laser, the linewidth is AA = 0.0214 gm at A = 3.82 gm. Introduction of the intracavity dispersive techniques described in Chapter 7 produce much narrower emission linewidths. A dispersive OPO is illustrated in Fig. 8.5. For this oscillator the multiple-return-pass linewidth is determined by AA -- A O R ( R M V a O a + RVa~p) -~

(8.54)

where the various coefficients are as defined in Chapter 7. It should be apparent that Eq. (8.54) has its origin in

A)k--- A0(00/0)k) -1

(8.55)

which is a simplified version of Eq. (8.53). Hence, we have demonstrated a simple mathematical approach to arrive at the linewidth equation that was derived using geometrical arguments in Chapter 4. Using a dispersive cavity incorporating an intracavity etalon in a LiNbO3 OPO excited by a Nd:YAG laser, Brosnan and Byer (1979) achieved a linewidth of Au = 2.25GHz. Also using a Nd:YAG-pumped LiNbO3 OPO and a similar interferometric technique, Milton et al. (1989) achieved single-longitudinal-mode emission at a linewidth of Au ~ 30 MHz.

166


/ Grating

\

~

i!

iI

\

If

~

"

Crystal

M

mirror

Figure 8.5 Dispersiveoptical parametric oscillator using an HMPGI grating configuration. A further aspect illustrated by the Taylor series expansion is that by equating the second and third series it is found that A O - A T ( O A k / O T ) ( O A k / O 0 ) -1

(8.56)

which indicates that the beam divergence is a function of temperature, which should be considered when contemplating thermal tuning techniques. Chapter 9 includes a section on the emission performance of various OPOs.

8.2.3 THE REFRACTIVE INDEX AS A FUNCTION OF INTENSITY

Using a Taylor series to expand an expression for the refractive index yields n = no + (On/OI)I + (1/2!)(02n/012)I 2 + . . .

(8.57)

Neglecting the second-order and higher terms, this expression reduces to n = no + (On/OI)I

(8.58)

where no is the normal weak-field refractive index, defined in Chapter 12 for various materials. The quantity (On/OI) is not dimensionless and has units that are the inverse of the laser intensity, or W -1 cm 2. Using polarization arguments this derivative can be expressed as (Boyd, 1992) O n l O I - 127rZx3/(nZ(a~)c)

(8.59)

167

Nonlinear Optics

This quantity is known as the second-order index o f refraction and is traditionally referred to as n2. Setting On/OI = n2, Eq. (8.58) can be restated in its usual form as n(~) = n0(~) + n2(~)I(~)

(8.60)

The change in refractive index as a function of laser intensity is known as the optical Kerr effect. For a description of the electro-optical Kerr effect, the reader should refer to Agrawal (1995). A well-known consequence of the optical Kerr effect is the phenomenon of self-focusing. This results from the propagation of a laser beam with a near-Gaussian spatial intensity profile, since, according to Eq. (8.60), the refractive index at the center of the beam is higher than the refractive index at the wings of the beam. This results in an intensity-dependent lensing effect, as illustrated in Fig. 8.6. The phenomenon of self-focusing, or intensity-dependent lensing, is important in ultrafast lasers or femtosecond lasers (Diels, 1990; Diels and Rudoph, 1996), where it gives rise to what is known as Kerr lens mode locking (KLM). This is applied to spatially select the high-intensity modelocked pulses from the background CW lasing. This can be accomplished simply by inserting an aperture near the gain medium to restrict lasing to the central, high-intensity, portion of the intracavity beam. This technique has become widely used in femtosecond laser cavities.

8.3 O P T I C A L

PHASE

CONJUGATION

Optical phase conjugation is a technique that is applied to correct laser beam distortions either intracavity or extracavity. A proof of the distortion correction properties of phase conjugation was provided by Yariv (1977) and is

Optical medium

Figure 8.6 Simplifiedrepresentation of self-focusing due to n = no + n2I in an optical medium due to propagation of a laser beam with a near-Gaussian intensity profile.


168

outlined here. Consider a propagating beam in the +z direction, represented by

E(r, t) - Al (r)e -i(";t-kz) -+-...

(8.61)

and the scalar version of the nonlinear wave equation given in Eq. (8.25), assuming that the spatial variations of e are much larger than the optical wavelength. Neglecting the polarization term one can write

(OZA1/Oz 2) + i2k(OA1/Oz) + ((eco2/c 2) - kZ)A1 - 0

(8.62)

The complex conjugate of this equation is

(02A*l/Oz 2) - i2k(OA*l/OZ ) + ( ( e J / c 2) - k2)A*l - 0

(8.63)

which is the same wave equation as for a wave propagating in the - z direction of the form

E(r, t) - A2(r)e -i(~t+kz) + . . .

(8.64)

A2(r) - aA~(r)

(8.65)

provided

where a is a constant. Here, the presence of a distorting medium is represented by the real quantity e (Yariv, 1977). This exercise illustrates that a wave propagating in the reverse direction of A l(r) and whose complex amplitude is everywhere the complex conjugate of A l(r) satisfies the same wave equation satisfied by A l(r). From a practical perspective this implies that a phase-conjugate mirror can generate a wave propagating in reverse to the incident wave whose amplitude is the complex conjugate of the incident wave. Thus, the wavefronts of the reverse wave coincide with those of the incident wave. This concept is illustrated in Fig. 8.7.

52

E1

Inhomogeneous optical medium

PCM

Figure 8.7 The concept of optical phase conjugation.

169

Nonlinear Optics

Figure 8.8 Basic phase-conjugated laser cavity.

A phase-conjugated mirror (PCM), as depicted in Fig. 8.8, is generated by a process called degenerate four-wave mixing (DFWM), which itself depends on X(3) (Yariv, 1985). This process can be described by considering planewave equations of the form Em(r, t) - Am(r)e -i(~t-kmr) + . . .

(8.66)

where m = 1,2, 3, 4 and k and r are vectors. Using these equations and the simplified equations for the four polarization terms (Boyd, 1992),

E2E~] 3X(3) [ g 22 g 29 nt- 2g2glg~]

e l = 3~(3) [E 21E 1, -+- 2El

(8.67a)

P2

(8.67b)

-

-

P3 - 3X(3)[2E3E1Ef + 2E3E2E~ + 2E1E2E~]

(8.67c)

P4 = 3X(3)[2E4E1Ef + 2E4E2E~ + 2E1E2E~]

(8.67d)

in the generalized wave equation V2Em(z, t) -- - c -2 (e(a~m)VtEm2 (z, t) + 47rVZPm(z, t))

eventually leads to expressions for the amplitudes that show that the generated field is driven only by the complex conjugate of the input amplitude. An issue of practical interest is the representation of a phase-conjugated mirror in transfer matrix notation, as introduced in Chapter 6. This problem was solved by Auyeung etal. (1979), who, using the argument that the reflected field is the conjugate replica of the incident field, showed that the ABCD matrix is given by A

B D)-(0

1

0)

(8.68)

0)1

(8.69)

-1

which should be compared to A

B D)-(0

1

170


for a conventional optical mirror. A well-known nonlinear material suitable as a PCM is CS2 (Yariv, 1985). Fluctuations in the phase-conjugated signal generated by D F W M in sodium was investigated by Kumar e t a l . (1984).

8.4 R A M A N SHIFTING Stimulated Raman scattering (SRS) is an additional and very useful tool to extend the frequency range of fixed-frequency and tunable lasers. Also known as Raman shifting, SRS can be accomplished by focusing a TEM00 laser beam onto a nonlinear medium, such as H2 (as illustrated in Fig. 8.9), to generate emission at a series of wavelengths above and below the wavelength of the laser pump. The series of longer-wavelength emissions are known as Stokes and are determined by (Hartig and Schmidt, 1979) (8.70)

USm = Ue -- muR

where us,, is the frequency of a given Stokes, ue is the frequency of the pump laser, uR is the intrinsic Raman frequency, and m = 1,2, 3, 4 , . . . for successively higher Stokes. For the series of shorter anti-Stokes wavelengths, UASm = Up + m u g

(8.71)

where UASm is the frequency of a given anti-Stokes. It should be noted that us~ and uAS~ are generated by the pump radiation, while these fields, in turn, generate u& and uAS2. In other words, for m = 2, 3, 4 , . . . , us,, and UASm are generated by US(m_,~ and UAS(m_~, respectively. Hence, the most intense radiation occurs for m = 1, with successively weaker emission for m = 2, 3, 4 , . . . , as depicted in Fig. 8.10. For instance, efficiencies can decrease progressively from 37% (first Stokes), to 18% (second Stokes), to 3.5% (third Stokes) (Berik etal., 1985). For the H2 molecule, uR ~ 124.5637663THz (or 4155 cm -1) (Bloembergen, 1967).

m

~A ' S~R Figure 8.9 Opticalconfiguration for H2 Raman shifter. The output window and the dispersing prism are made of CaF:.

171

Nonlinear Optics

-I __cI0-I 100

`&s,

(/)

c::

>~ 10 .2

n"

`&So

'&As,

10-3

10 .4

_`&is` I 300

Figure 8.10

I

400

500

I

600 Z (nm)

I

700

I

800

I

900

Stokes and anti-Stokes emission in H2 for )~e = 500 nm.

Using the wave equation and assuming solutions of the form Es(z, t) - A s ( z ) e -i(~st-ksz) + . . .

(8.72)

Ee(z, t) - A e ( z ) e -i(~pt-kez) + . . .

(8.73)

it can be shown, using the fact that the Stokes polarization depends on X(3)EpE*pEs, that the gain at the Stokes frequency depends on the intensity of the pump radiation, the population density, and the inverse of the Raman linewidth, among other factors (Trutna and Byer, 1980). It is interesting to note that the Raman gain can be independent of the linewidth of the pump laser (Trutna et al., 1979). A detailed description on the mechanics of SRS is provided by Boyd (1992). Stimulated Raman Scattering in H2 has been widely used to extend the frequency range of tunable lasers, such as dye lasers. This technique was first demonstrated by Schmidt and Appt (1972) using room-temperature hydrogen at a pressure of 200 atmospheres. This is mentioned because, though simple, the use of pressurized hydrogen requires stainless steel cells and detailed attention to safety procedures. Using a dye laser with an emission wavelength centered around 563 nm, Wilke and Schmidt (1978) generated SRS radiation in H2 from the eight anti-Stokes (at 198 nm) to the third Stokes (at 2064 nm) at an overall conversion efficiency of up to 50%. Using the second harmonic of the dye laser, the same authors generated from the fourth anti-Stokes to the fifth Stokes, as illustrated on Table 8.2, at an overall conversion efficiency of up to 75%. Using a similar dye laser configuration, Hartig and Schmidt (1979) employed a capillary waveguide H2 cell to generate tunable first, second, and third Stokes spanning the wavelength range from 0.7 gm to 7 gm. Using a dye laser system incorporating a MPL grating oscillator and two stages of amplification, Schomburg et al. (1982) achieved generation up to the thirteenth anti Stokes at 138 nm. Brink and Proch (1982) report on a 70%


172 Table 8.2 Tunable Raman Shifting in Hydrogen

Anti-Stokes A range (nm) )k4 ~ A3 ~ A2 ~ A1 ~

192 (~A4~ 5.8)c 210 (~"~3 ~'~ 7.2) 229 (~A2 ~ 8.9) 251 (6A1~ 10.7)

Tunable lasera A range (nm) 275 < A < 287

Source: Wilkeand Schmidt (1978). a Second harmonic from a dye laser. bApproximate values. cCorresponds to a quoted range of 188.7nm
190mJ

13.2kHz

>2.5kW

>50

Dye

Reference

Coumarin 480 Tang etal. (1987) TBS Tallman and Tennant (1991) Rhodamineb Bass et al. (1992)

Source: Adapted from Duarte (1995b). a Pulse length quoted at 500 ns. bTuning range: 550-650nm (Bass etal., 1992).

The performance of laboratory-size laser-pumped dye lasers can be illustrated by considering the work of Bos (1981), who reported a linewidth of Au ~ 320MHz at 590nm with a telescopic oscillator incorporating an intracavity etalon. Using three stages of amplification, the output energy was 165mJ at overall conversion efficiency of 55% for excitation at 532nm. Employing an H M P G I grating oscillator and two stages of amplification, Dupre (1987) reported 3.5 mJ and Au = 1.2 GHz at 440 nm. The conversion energy efficiency was ~ 9 % for excitation at 355 nm. Copper Vapor Laser excitation of an H M P G I grating oscillator yielded Au ~ 600MHz for A = 575nm, and pulse lengths of 12ns (FWHM) at conversion efficiencies of ~ 5 % (Duarte and Piper, 1984). The average output power was 80 mW at a prf of 8 kHz. Using the same class of multipleprism oscillator and one amplifier stage, Singh e t a l . (1994) reported Au ~ 1.5 GHz and a conversion efficiency of 40% at a prf of 6.5 kHz. The performance of large flashlamp-pumped dye lasers is summarized in Table 9.7. For a comprehensive review of flashlamp-pumped dye lasers, the reader should consult Everett (1991). In general, these lasers have been used to generate large energies in pulses in the microsecond regime. The energy per pulse is such that fairly modest prfs can generate average powers in the kilowatt regime. Besides their intrinsic ability to generate large pulsed energies, flashlamppumped dye lasers have been configured in small-scale laboratory versions designed to yield fairly narrow-linewidth tunable laser emission. For a review on this subject, see Duarte (1995b). A ruggedized MPL grating coaxial flashlamp-pumped dye laser oscillator yielding Au ~ 375 MHz at amplified spontaneous emission levels a few parts in 10-7 was reported by Duarte etal. (1991). This dispersive oscillator provided ~3 mJ per pulse using rhodamine 590 at a concentration of 0.01 mM.

186


Table 9.7 Performance of High-Energy Flashlamp-Pumped Dye Lasers Excitation

Pulse duration

Output energy

Efficiency (%)

Lineara

7 las

40 J

0.4

Transverseb

5 las

140Jr

1.8

10las

400 J

0.8

Coaxial

Dye

Reference

Rhodamine6G at 0.08mM Rhodamine6G at 0.025mM Rhodamine6G at 0.022mM

Fort and Moulin (1987) Klimek et al. (1992) Baltakov et al. (1974)

Source: Adapted from Duarte (1995b). a Employs 12 flashlamps in a linear configuration. bEmploys 16 flashlamps in a transverse configuration. Cyields an average power of 1.4 kW at a prf of 10Hz.

The output from narrow-linewidth oscillators can be amplified using single-stage amplifiers to yield hundreds of millijoules per pulse. Flamant and Maillard (1984) used a two-etalon oscillator to excite a flat-mirror amplifier to attain Au = 346 M H z and a pulse energy of 300 mJ at 590 nm. Using a multiple-prism grating oscillator and a single-stage unstableresonator amplifier, Duarte and Conrad (1987) achieved Au ~ 375 MHz and a pulse energy of 600mJ at 590 nm. In addition to traditional liquid dye lasers there has been considerable research and development activity in the area of solid-state dye lasers. An excellent review on dye-doped solid-state matrices is provided by Costela e t a l . (1998), and the photophysical properties of these solid-state gain materials have been characterized by Holzer et al. (2000). Although most recent activity has been centered on polymeric matrices and hybrid silicatepolymer composite materials, there has also been work reported on crystalline dye lasers (Rifani et al., 1995; Braun et al., 2000). A review of organic lasers, including solid-state dye lasers, aimed at assessing the development of electrically pumped polymer lasers is given by Kranzelbinder and Leising (2000). Work at establishing quasi-CW oscillation using improved hybrid dye-doped polymer-nanoparticle matrices is also in progress. The performance of broadband solid-state dye lasers is summarized in Table 9.8. Optimized multiple-prism grating solid-state oscillators, as described in Chapter 7, have yielded tuning ranges in the range of 550-603 nm with TEM00 laser beams, at beam divergences ~1.5 times the diffraction limit. The emission is in a single longitudinal mode at Au ~ 350 M H z in pulses ~3 ns ( F W H M ) with a near-Gaussian temporal profile. Conversion efficiency is reported at ~ 5 % , and the ASE levels are extremely low at ,-,~10-6

187

Lasers a n d Their E m i s s i o n Characteristics

Table 9.8 Performance of Broadband Solid-State Dye Lasers Excitation source Flashlamp Nd:YAG laser (2u) LPDL c FLPDL e Nd:YAG laser (2u)

Matrix

Dye

PMMA a Rhodamine 590 at 0.11 mM M P M M A b Rhodamine 11B at 0.5 mM HEMA:MMA d Rhodamine 6G at 0.5 mM TEOS c Rhodamine 6G at 2 mM ORMOSIL g Rhodamine 6G at 0.086 mM

Output energy

Efficiency (%)

50 mJ

0.8 mJ

Pacheco et al. (1988) 65

Maslyukov et al. (1995)

40

Duarte et al. (1997)

2.5 mJ 3.5mJ

Reference

Duarte et al. (1993) 35

Larrue et al. (1994)

Source: Adapted from Duarte (1995b). Polymethyl methacrylate. b Modified Polymethyl metracrylate. c Laser-pumped dye laser using Coumarin 152. d2_hydroxyethyl methacrylate:methyl methacrylate e Flashlamp-pumped dye laser using Coumarin 525. fSi(OC2H5)4 g Organically modified silicate. a

(Duarte, 1999). Transverse excitation in the long-pulse regime of a four-prism grating solid-state dye laser oscillator has led to pulses as long as 105 ns (FWHM) and Au ~ 650 MHz at pulsed energies of ~0.4 mJ (Duarte et al., 1998).

9.3.2 CONTINUOUS-~AVEDYw LASERS Dye lasers have had a significant impact in high-resolution spectroscopy and other applications, including laser cooling, given their tunability, excellent TEM00 beam quality, and intrinsic narrow linewidths, which can readily reach a level of a few megahertz. Continuous-wave dye lasers typically use Ar + and Kr + as excitation sources, although in principle they could use any compatible laser yielding TEM00 emission. It should be noted that CW dye lasers have been excited with a variety of lasers, including diode lasers (see, for example, Scheps, 1993). Table 9.9 summarizes the performance of relatively high-power dye lasers, some of which yield SLM oscillation at linewidths in the megahartz regime. Stabilization techniques can produce significant improvements in laser linewidth, as indicated in Table 9.10. Here,

188


Table 9.9 Performance of High-Power CW Dye Lasers Cavity

Lineara Lineara Ring a

Spectral range (nm)

Linewidth

560-650 407-887 g

SLM d SLM a

Output power

Efficiency(%)

Reference

33 W b'~ 33 W ef 5.6W h

30 17 23.3

Anliker etal. (1977) Baving et al. (1982) Johnston etal. (1982)

Source: Adapted from Duarte (1995b). a Under Ar + laser excitation. b Maximun CW power quoted: 52 W for a pump power of 175 W. e Using Rhodamine 6G at 0.7 mM. a Linewidth values can be in the few MHz range. eWithout intracavity tuning prism, quoted output power is 43 W for a pump power of 200 W. fUsing Rhodamine 6G at 0.94mM. g Using 11 dyes. h Using Rhodamine 6G.

t h e use o f a n rf-optical h e t e r o d y n e lock t e c h n i q u e e n a b l e d D r e v e r et al. (1983) to r e a c h a laser l i n e w i d t h o f 100 Hz. A n excellent review o n C W dye lasers is given by H o l l b e r g (1990). A f u r t h e r t o w e r i n g c o n t r i b u t i o n o f the C W dye lasers to the field o f optics a n d lasers w a s their use as t o o l s in the d e v e l o p m e n t o f u l t r a s h o r t - p u l s e lasers t h a t gave origin to the f e m t o s e c o n d lasers. A m o n g the i m p o r t a n t c o n c e p t s d e v e l o p e d in this e n d e a v o r w e r e the g e n e r a t i o n o f b a n d w i t h - l i m i t e d u l t r a s h o r t pulses ( R u d d o c k a n d B r a d l e y , 1976), the c o l l i d i n g - p u l s e m o d e ( C P M ) l o c k i n g t e c h n i q u e ( F o r k e t a l . , 1981), a n d p r i s m a t i c pulse c o m p r e s s i o n (Dietel et al., 1983; F o r k et al., 1984). U s i n g a n e x t r a c a v i t y pulse c o m p r e s s o r ,

Table 9.10 Performance of Frequency-Stabilized CW Dye Lasers a Stabilization method

Linewidth

Frequency drift

Reference

Cavity side locka rf-optical Heterodyne lockb Post lasere

150 kHz a 100 Hz ,

C o C

m,,,=

5000 4000

9

3000

~9

2000

>

205

.m

1000

0

I

. . . . . . . . . . . .

200

400

600

800

Number of Pixels Figure 10.2 Extremely elongated, approximately 1000:1, near-Gaussian laser beam [Reprinted from Duarte (1991), Copyright 1991, with permission from Springer-Verlag].

beam expander (Duarte, 1991, 1993a) since this form of beam expansion does not introduce further focusing variables. Further, the multiple-prism beam expander can be designed to yield zero dispersion at a chosen wavelength of design (Duarte and Piper, 1982; Duarte, 1985), as described in Chapter 4. The expanded laser beam illuminates the N-slit array at j, where N subbeams are created and proceed to propagate toward the detection screen at x, which is configured by a digital detector. Following some spatial displacement after j, the N subbeams, due to divergence mandated by the uncertainty principle, begin to undergo interference. The pattern of interference is recorded by the detector at x. Note that although the detector of choice is a digital detector, such as a photodiode array or a CCD array, photographic detection can also be used. If detection is provided by a very thin line of detectors, then, as explained in Chapter 2, the spatial distribution of the interference signal can be described by N I(X S)I2 -- Z j=l

N

t~(rj) Z

~(rm)ei(~m-~t~)

(lO.1)

m=l

which can be expressed as (Duarte and Paine, 1989; Duarte, 1991) [(xl )l 2 -- = gJ(rj

+ 2 ~ qd(rj) m=j+l ~(rm) COS(f'~m -- ~~j)

Interference in two dimensions is described by (Duarte, 1995)

(10.2)


206

N N N N I(xIs)12 -- Z E ~ ( r z y ) E Z ~(rpq)ei(f~qP-f~Y) z=l y=l q=l p=l

(10.3)

Equation (10.2) has been successfully applied to characterize the interference resulting from the interaction of expanded narrow-linewidth laser beam slit arrays of various dimensions and N in the range of 2 ~.5i

(b)

o.)

rr 1.0 0.5 0.0

~--

|

,

!

-2.5-:2.0-1'.5 '1'.0-0'.5 0J0 0.5 1.0 Distance (meters) x 10 .3

#

l

1.5

2.0

2.5

4 . 0

3.5 ._>, 3.0 t/)

o= 2.5 o

(c)

2.0

2

_~ ~.5 rr 1.0 0.5 0.0, i i i i , - 2 . 5 - 2 . 0 - 1 . 5 - 1 . 0 - 0 . 5 0.0 0.5 1.0 Distance (meters) x 10-3

7"-~c

1.5

2.0 2.5

Figure 10.9 Interferometric alphabet. (a) a, (b) b, (c) c, and (d) z. [Reprinted from Duarte (2002), copyright 2002, with permission from Elsevier].

218

Tunable Laser Optics 4.5 4.0 '

3.5

-~ 3.0 r O

r "-" 2.5

(d)

(I) .m

> 2.0

-~ 1.5

rr

1.0 0.5 0.0

-3.o

-i.o

olo

1.'o

Distance (meters) x 10 -3

z'o

3.0

Figure 10.9 (Continued).

Transmission integrity is demonstrated in Fig. 10.10 for the case of the interferometric character a. To this effect an optically smooth surface with an average thickness of ~ 150 ~tm is introduced at an angle in the optical path to cause a reflection of the character a. It should be noted that insertion of the beam splitter normal to the optical axis produces no measurable spatial optical distortions except a decrease in the intensity of ~8%. The angle of incidence of the interferometric character on the beam splitter was selected to be close to the Brewster angle of incidence to reduce transmission losses while still being able to reflect a measurable fraction of the signal. In the sequence of measurements, Fig. 10.10a shows the undistorted character a. The severe distortions depicted in Figs. 10.10b to 10.10d show the effect of introducing the thin beam splitter in the optical path. Fig. 10.10e depicts the intercepted interferometric character a. Although the severe distortions are no longer present, close scrutiny of the interferogram reveals a decrease by ,-.,3.7% in the intensity of the signal relative to the original character shown in Fig. 10.10a. The intercepted signal is also displaced by approximately 50 ~tm in the frame of reference of the detector due to the refraction induced at the beam splitter. In addition, there is a slight obliqueness in the intensity distribution as determined from the secondary maxima. Hence, by comparison with the original interferometric character or a theoretically generated character, it can be concluded that the integrity of the intercepted character a has been distinctly compromised. Although the measurements just considered were performed over short propagation path lengths in the laboratory (0.1 m and 1 m), Duarte (2002) also discusses propagation over larger distances. Using interferometric

Architecture of N-Slit Interferometric Laser Optical Systems _~

219

8000-

c 6000

(a)

.~ 4000 E 2000 0

460

480 500 520 540 Numberof Pixels

|

560

9000-

"~ 7000 C

(b)

.~ 5000 E

3000 1000 460

480 560 520

540 560

Number of Pixels

2600-

.~ 2200 ~ oO

e-

_~ 18oo~ _=

(c)

.~ ~4oo~ looo~ ~" 600: 2O0

'

|

.....

460

|

|

w

i

480 500 520 540 Number of Pixels

!

560

9000"~ 7000 E

~

-

9 5000 .5

(d)

3000 1000460

480 500 520 540 Number of Pixels

560

Figure 10.10 Interception sequence of the interferometric character a (see text for details). [Reprinted from Duarte (2002), copyright 2002, with permission from Elsevier].

220


>,80001 =9 _c

6000t

(e)

2e~ 4000t rr 0~'-, -~-~, ; , . 460 480 500 520 540 560 Number of Pixels Figure 10.10 (Continued).

calculations, via Eq. (10.2) or (10.3), it can be shown that communication in free space can proceed over long path lengths using visible wavelengths and a detector comprising of a few tiled photodiode arrays. One specific example involves the generation of the interferometric character a using two 1 mm slits separated by 1 mm. For A = 632.82 nm, this arrangement produces an interferometric distribution bound within 10cm for a propagation path length of 100 m. The interferometric character z is produced by an array of 26 slits of 1 mm separated by 1 mm. For A = 632.82 nm, this arrangement produces an interferometric distribution bound within 14 cm for a propagation path length of 100 m. This can be accomplished using two off-the-shelf linear photodiode arrays (each 72 mm long) tiled together. If the dimensions of the slits are increased to 3 mm at A = 441.56 nm, interferometric characters could be propagated over distances of 1000m using four such tiled photodiode arrays (Duarte, 2002). The examples considered here assume a propagation path characterized by a single homogeneous propagation medium, such as vacuum, between j and x. One modification would be the introduction of a distortionless beam expander, such as an optimized multiple-prism beam expander. In Chapter 4 it was shown that these expanders can easily provide beam magnification factors of M ~ 100. Deployment of such a multiple-prism beam expander next to the slit array would reduce the beam divergence significantly, thus reducing the requirements on the dimensions of the digital detectors. Free-space communications of interferometric characters in terrestrial environments would need to account for the inherent atmospheric turbulence present in such surroundings. This would certainly detract from the simplicity of the method. This could still be accomplished by noting that atmospheric distortions are stochastic in nature as compared to systematic distortions introduced by optical interception.

Architecture of N-Slit Interferometric Laser Optical Systems

221

From a technological viewpoint, it is important to emphasize the use of TEM00 lasers with narrow-linewidth emission, since that characteristic is essential in providing well-defined interferometric characters close to their theoretical counterparts. The characters could be changed in real time either by using a tunable laser (Duarte, 1999) or by incorporating precision variable-slit arrays. The use of narrow-bandpass filters could allow transmission during daylight. Quantum cryptography provides secure optical communications guaranteed by the uncertainty principle and has been shown to be applicable over distances of tens of kilometers (Jacobs and Franson, 1996). Interferometric communications using the NSLI provides security using the principles of diffraction, refraction, and reflection. As discussed in Chapter 2, all of these principles have their origin in the principle of interference. As outlined in Chapter 3, the uncertainty principle itself can be formulated from interferometric arguments. Advantages of free-space communications using interferometric characters include a very simple optical architecture and the use of relatively high-power narrow-linewidth lasers, although, in principle, the method also applies to single-photon emission.

10.4.4 WAVELENGTHMETER AND BROADBAND INTERFEROGRAMS Generalized N-slit interference equations, such as Eqs. (10.2) and (10.3), are inherently wavelength dependent, since the interference term is a function of wavelength, as explained in Chapter 2. Thus, it is straightforward to predict that, for a given set of geometrical parameters, the measured interferogram depends uniquely on the wavelength of the laser. This feature can be applied to use of the NSLI as a wavelength meter, as will be explained in Chapter 11. Although emphasis up to now has been placed on the desirability of using narrow-linewidth lasers in conjunction with the NSLI, the scope of the measurements can also be extended to include broadband emission. For broadband emission sources the measured interferogram represents a cumulative interferogram of a series of individual wavelengths of the form depicted in Fig. 10.11. This concept is central to this measurement approach, and it is based on Dirac's dictum on interference (Dirac, 1978). That is, interference occurs between undistinguishable photons only. In other words, blue photons do not interfere with green or red photons. Hence, an interferogram with broad features, as illustrated in Fig. 10.11, is a cumulative signal composed of a series of individual interferograms arising from a series of different wavelengths. Once the central wavelength of emission of the broadband interferometer is determined using a standard spectrometer or suitable wavelength meter, a theoretical cumulative interferometer can be constructed to match the measured signal and determine its bandwidth.

222

Tunable Laser Optics 6 0 0 0 ~-

>.,

5000 -

t,r0

4000

.>__ ..b-,

3000

II

2000

1000 460

480

I

I

I

I

500

520

540

560

Number of Pixels Figure 10.11 Interferogramfrom broadband light source. In principle, for broadband ultrashort pulsed lasers, once the bandwidth of the emission is determined, an approximate estimate of the temporal pulse duration is possible using the time-frequency uncertainty relation AuAt ~ 1. This simple concept is applicable only to ultrashort-pulse lasers emitting pulses and spectral distributions obeying the time-frequency uncertainty limit.

10.5 S E N S I T O M E T R Y Traditional sensitometry and sensitometers are described by Altman (1977). In essence, a sensitometer is an instrument that illuminates an unexposed imaging material to produce a series of exposures at various light intensity levels. A laser sensitometer uses stable lasers yielding TEM00 beams and various optical techniques to produce a scale of exposures that can then be optically characterized to determine the sensitivity of the imaging material. In this section the optical architecture of a multiple-laser sensitometer is described. Laser sensitometers work on the principle of exposing a line by displacing a focused near-Gaussian beam with a beam waist in the range 50 ~tm _< w _< 100 ~tm. This line is exposed on the imaging material, which is deployed at a plane perpendicular to the optical axis and to the plane of propagation. The imaging medium is displaced, orthogonal to the optical axis, at a velocity allowing for an overlap (usually 50%) of the near~ Gaussian beam. The movement of the laser beam provides the temporal component of the exposure. Once an exposure of certain dimensions is produced, usually 10 mm in width, the intensity of the laser beam is adjusted, using electro-optical means, and a new series of line exposures is produced.


223

Eventually a scale of rectangular exposures at different laser intensity levels is rendered. Three optical channels, corresponding to blue, green, and red lasers, converging to a single exposure plane are often employed. An alternative laser sensitometer is depicted in Fig. 10.12. This is a singlechannel multiple-laser multiple-prism sensitometer using polarization to vary the intensity of the laser exposure. In this description the laser sensitometer will be referred to as a polarizer multiple-prism multiple-laser (PMPML) sensitometer. The P M P M L sensitometer (see Fig. 10.12) uses a single optical channel with a principal laser as the first element defining the optical axis and secondary lasers adding their radiation via beam splitters. All lasers are polarized parallel to the plane of incidence. A variable broadband neutral-density filter is inserted as a coarse intensity control prior to the polarizer. The polarizer is a Glan-Thompson prism pair, with an extinction coefficient of 1 x 10 -6 or better, mounted on a high-precision annular rotational stage capable of an angular resolution of 0.001 arc sec. As described in Chapter 5, rotation of this polarizer causes the transmission of the lasers to decrease from nearly full transmission to close to total extinction. For the lasers polarized parallel to the plane of polarization, optimum transmission is accomplished with the Glan-Thompson polarizer deployed as depicted in Fig. 10.12. Following the polarizer, the beams enter a telescope-lens system and a multiple-prism beam expander, as shown in Fig. 10.1. The elongated nearGaussian beams are then propagated through a wide aperture so as to produce a diffractive profile, as depicted in Fig. 10.13. Note that the diffractive profile is wider than the width of the exposures needed, so the intensity variation caused by the "ears" of the profile do not affect the wanted area. The fine variations toward the center of the distribution have a negligible effect.

I

Laser 1

/'1

I

F Telescope

Laser 2

Laser 3

Muliple-prism beam expander

Wide slits

Imaging plane

Figure 10.12 Polarizermultiple-prism laser sensitometer. The telescope expands the beam in two dimensions whilst the multiple-prism beam expander magnifies in only one dimension parallel to the plane of propagation.


224

Figure 10.13 Diffraction profile of the illumination line at A = 532 nm.

Using the method just described, a line exposure is instantaneously provided, thus eliminating the need to displace the laser beams and the associated electromechanical means necessary to accomplish this task. Since the line exposure is horizontal, the imaging material is displaced in a plane orthogonal to the plane of incidence of the instrument. In P M P M L sensitometers the temporal exposures are provided by using the lasers in a pulsed mode. Thus, depending on the lasers, it is possible to vary the exposure time from less than 1 ns to 1000ns. It should also be mentioned that careful selection of the beam profiles of the lasers and their respective distances to the main optical axis enables spatial overlapping of the laser beams to within 1 ~tm at the focal plane with a minimal use of extra beam-shaping optics. As indicated by Duarte (2001), lasers suitable for illumination include modelocked diode-pumped frequency-doubled N d : Y A G lasers and pulsed semiconductor lasers.

PROBLEMS 1. For the telescope, lens, multiple-prism, distance architecture depicted in Fig. 10.1, derive its propagation ray transfer matrix given in Eq. (10.9). 2. Show that in the absence of a convex lens, Eq. (10.9) reduces to Eq. (10.16). 3. Show that in the absence of a convex lens, Eq. (10.15) reduces to Eq. (10.17).


225

4. Design a double-prism beam expander yielding zero dispersion at the wavelength of design and M = 5 for an optical system as depicted in Fig. 10.1. Using a Galilean telescope with Mt---20 and a convex lens with f - 30 cm, calculate the width and the height of the resulting extremely elongated near-Gaussian beam at the focal plane. Assume a TEM00 He-Ne laser at A = 632.82 nm and w0 = 250 ~tm. For the material of the multiple-prism beam expander use fused silica. 5. For a beam splitter made of fused silica and a thickness of 1 mm, determine the lateral displacement from its original path of a TEM00 He-Ne laser beam at A = 632.8 nm immediately following the beam splitter if the angle of incidence is 57 ~.

REFERENCES Altman, J. H. (1977). Sensitometry of black-and-white materials. In The Theory of the Photographic Process (James, T. H., ed.). Eastman Kodak Company, Rochester, NY, pp. 481-516. Boffi, P., Piccinin, D., Mottarella, D., and Martinelli, M. (2000). All-optical free-space processing for optical communications signals. Opt. Commun. 181, 79-88. Dainty, J. C., and Shaw, R. (1974). Image Science. Academic Press, New York. Deutsch, D. (1992). Quantum computation. Phys. Worm 5(6), 57-61. Dirac, P. A. M. (1978). The Principles of Quantum Mechanics, 4th ed. Oxford University Press, London. Duarte, F. J. (1985). Note on achromatic multiple-prism beam expanders. Opt. Commun. 53, 259-262. Duarte, F. J. (1987). Beam shaping with telescopes and multiple-prism beam expanders. J. Opt. Soc. Am. A 4, p. 30. Duarte, F. J. (1991). Dispersive dye lasers. In High-Power Dye Lasers (Duarte, F. J., ed.). Springer-Verlag, Berlin, pp. 7-43. Duarte, F. J. (1993a). On a generalized interference equation and interferometric measurements. Opt. Commun. 103, 8-14. Duarte, F. J. (1993b). Electro-optical interferometric microdensitometer system. U.S. Patent No. 5,255,069. Duarte, F. J. (1995). Interferometric imaging. In Tunable Laser Applications (Duarte, F. J., ed.) Marcel Dekker, New York, pp. 153-178. Duarte, F. J. (1996). Generalized interference equation and optical processing. In Proceedings of the International Conference on Lasers '95 (Corcoran, V. J., and Goldman, T. A., eds.). STS Press, McLean, VA, pp. 615-617. Duarte, F. J. (1999). Multiple-prism grating solid-state dye laser oscillator: optimized architecture. Appl. Opt. 38, 6347-6349. Duarte, F. J. (2001). Laser sensitometer using multiple-prism beam expansion and a polarizer. U.S. Patent no. 6,236,461. Duarte, F. J. (2002). Secure interferometric communications in free space. Opt. Commun. 205, 313-319. Duarte, F. J., and Paine, D. J. (1989). Quantum mechanical description of N-slit interference phenomena. In Proceedings of the International Conference on Lasers '89 (Harris, D. G., and Shay, T. M., eds.). STS Press, McLean, VA, pp. 42-27.

226


Duarte, F. J., and Piper, J. A. (1982). Dispersion theory of multiple-prism beam expanders for pulsed dye lasers. Opt. Commun. 43, 303-307. Jacobs, B. C., and Franson, J. D. (1996). Quantum cryptography in free space. Opt. Lett. 21, 1854-1856. Turunen, J. (1986). Astigmatism in laser beam optical systems. Appl. Opt. 25, 2908-2911. Willebrand, H. A., and Ghuman, B. S. (2001). Fiber optics without fiber. IEEE Spectrum 38(8), 40-45. Yu, S. T. S., and Gregory, D. A. (1996). Optical pattern recognition: architectures and techniques. Proc. IEEE 84, 733-752.

Chapter 11

Spectrometry and Interferometry

11.1 I N T R O D U C T I O N Laser optics employs several diagnostic tools. Characterization of pulsed laser emission requires measurements of the following parameters: energy and/or power, temporal profile of laser pulses, spatial characteristics of the laser beam, wavelength, and linewidth. Characterization of CW laser emission requires measurements of power, spatial characteristics of the laser beam, wavelength, and linewidth. Measurements of laser energy and power employ optoelectric and/or thermoelectric detectors. Determination of temporal profiles requires optoelectric detection and display electronics, for pulses down to the nanosecond regime, and autocorrelators (Diels, 1990; Diels and Rudolph, 1996) or other specialized instrumentation, such as frequency-resolved optical gating (Trebino et al., 1997), for pulses in the femtosecond domain. In this chapter, attention is focused on instrumentation useful in the measurement of laser wavelength and linewidth.

11.2 SPECTROMETRY Spectrometry, in principle, depends on the interaction of a light beam with a dispersive element or dispersive elements and on spatial discrimination following postdispersive propagation. Hence, the higher the dispersive 227


228

power and the longer the optical path of the postdispersive propagation, the higher the wavelength resolution. As described in earlier chapters, dispersion can be provided by either prisms, gratings, or prism-grating combinations. For a detailed treatment on the subject of spectrometry the reader should refer to Meaburn (1976).

11.2.1 PRISM SPECTROMETERS Prism spectrometry has its origin in the experiments reported by Newton (1704) in his book Opticks. Prism spectrometers usually deploy equilateral or isosceles prisms in series to augment the dispersion of the instrument. As described in Chapter 4, for an array of r identical isosceles or equilateral prisms deployed symmetrically in an additive configuration so that q~l, 1 = ~)1,2 . . . . . q~l,m and q~l,m -- q~2,m, the cumulative dispersion is given by (Duarte, 1990) ~7A~)2, r = r~Taq~2,1

(11.1)

V~q52,~ - ((sin ~2,1/ cos 4~2,1)+ (COS~2,1/COS4~2,1)tan~l,1)V~n

(11.2)

where all angular parameters are as described in Chapter 4. These equations indicate that dispersion can be augmented by a combination of two factors: increasing the number of prisms and using prisms with a high material dispersion. For some materials, V:~n is given in Table 4.1. Prism spectrometers assume various configurations. Two of these are considered here. The first configuration, depicted in Fig. 11.1, uses two prisms in series, with a considerable distance between the two prisms to increase the overall angular spread of the emerging beam. Long postdispersive optical paths of up to 3 meters, in conjunction with narrow apertures,

Figure 11.1

Long-optical-path double-prism spectrometer.

229


provide wavelength resolutions in the nanometer range. The second architecture, described by Meaburn (1976), includes a sequence of several prisms in series, as depicted in Fig. 11.2. In this configuration, dispersion is simply augmented by a larger factor r in Eq. (11.1). The same observations about the postdispersive optical path and spatial discrimination are relevant. 11.2.2

DIFFRACTION GRATING SPECTROMETERS

As described in Chapter 2, the diffraction grating equation is given by d(sin O + sin ~) = mA

(11.3)

and the angular dispersion is obtained by differentiating this equation so that

O0/OA = m/(acos O)

(11.4)

00/0A = (sin O + sin ~)/(A cos O)

(11.5)

or alternatively

For a grating deployed in Littow configuration, O = 9 and the grating equation becomes 2d sin O = mA

(11.6)

and the dispersion can be expressed as 00/0A = 2 tan O/A

(11.7)

As with prismatic spectrometers, the three factors that increase wavelength resolution in a diffraction grating spectrometer are dispersion, optical path length, and the dimensions of the aperture at the detection plane. From Eq. (11.4) it is clear that one avenue to increase dispersion is to use gratings deployed at a higher order or to employ gratings with a high groove density.

Figure 11.2 Dispersive assembly of multiple-prism spectrometer. The angle of incidence (~l,m and the angle of emergence 4~2,mare identical for all prisms. In addition to the cumulative dispersion, resolution is determined by the path length toward the exit slit and the dimensions of the slit.

230


Since deployment of gratings at high orders might lead to a decrease of diffraction efficiency, the alternative of using high-density gratings with 3000 lines/ram or more is a practical alternative to increase dispersion. Selection of a particular grating should consider the spectral region of desired operation, since high-density gratings might cease to diffract toward the red end of the spectrum. The electromagnetic theory of diffraction gratings is considered in detail by Maystre (1980). A rudimentary form of grating spectrometer is shown in Fig. 11.3 in order to illustrate the basic concept of spectrometry. Among the most widely used diffraction grating spectrometer configurations is the Czerny-Turner spectrometer, shown in Fig. 11.4, where two mirrors are used to increase the optical path length and thus the resolution. A modified Czerny-Turner spectrometer (Meaburn, 1976) includes two gratings to enhance the dispersion (Fig. 11.5). A simple modification to increase resolution is to increase the optical path directly or to add further stages of reflection. Various spectrometer design alternatives are discussed by Born and Wolf (1999) and Meaburn (1976). It should be noted that the use of curved gratings and curved mirrors to compensate for losses has been prevalent in many designs. Notable among the curved-grating approaches are the Rowland and the Paschen configurations (Born and Wolf, 1999). Modern Czerny-Turner spectrometer designs, with a ~4 m folded optical path, provide resolutions in the 0.01-nm region, in the visible spectrum, when using slits a few micrometers wide. These spectrometers are very useful in providing a first approach to determine the value of a laser wavelength.

'~m ha

Grating

"~2

,kl

~

Figure 11.3 Basicgrating spectrometer. Resolution is determined mainly by the dispersion of the grating and by the optical path length toward the exit slit.


231

sl I 1

s2

I

Figure 11.4 Czerny-Turnerspectrometer. Mirrors M1 and M2 provide the necessary curvature to focus the diffracted beam at the exit slit. Widely used spectrometers of this type often have an optical path length between the entrance slit and the back mirror in the range 25-100cm.

I

M1

G

S21

M2

Figure 11.5 Double-grating Czerny-Turner spectrometer. Mirrors M1 and M2 provide the necessary curvature to focus the diffracted beam at the exit slit. 1 1 . 2 . 3 DISPERSIVE WAVELENGTH METERS Dispersive wavelength meters are in essence dispersive spectrometers with the conventional slit-detector arrangement replaced by a linear photodiode array or C C D detector. Their main function is to determine the wavelength from the emission of tunable lasers. The resolution is determined by the available dispersion, the propagation distance between the dispersive element and the detector array, and the dimensions of the individual pixels at the digital array. Depending on the type of detector, individual pixels vary in size from a few micrometers to ~25 lam in width. The typical width of these digital arrays varies from ~25 m m to ~ 5 0 m m . This type of wavelength meter is used in conjunction with well-known, and preferably stabilized, laser sources for calibration. Morris and McIlrath (1979) used a 40-cm spectrometer incorporating an echelle grating deployed at a high order in conjunction with a photodiode array integrated by 1024 elements (each ~25 m m wide) to determine wavelengths within 0.01 nm in the visible.

232


A wavelength meter based on a high-dispersion prism-grating configuration is illustrated in Fig. 11.6. As discussed in Chapter 4, the dispersion of a diffraction grating can be significantly augmented when deployed with prism assemblies according to (Duarte and Piper, 1982) , , -( ~7A~l, m __ ~ t ~ l ,rn~ A?,lrn _+_( ktl ,mk t2,m ) 1 ~

,

'

)

2 ,m ~ AFlrn -4- ~7Aq~l,(rn+l)

(l 1.8)

where VA~tl,(m+l) = (VAOG q-- VA~2, r)

(1 1.9)

Here, VxOG is the grating dispersion and V~r is the single-pass prism dispersion. For a single prism, as depicted in Fig. 11.6, r = 1 and VAr

= tanr

(11.10)

Using a 600-1ines/mm echelle grating deployed in the third order and a BK-7 optical glass prism (n = 1.512 at A = 568.8 nm) with an apex angle of 41.5 ~ the sensitivity of the system becomes 0.3 ~ nm -1 (Duarte, 1983). For a 1000-mm distance from the prism to the detector, this dispersive combination can measure wavelengths with a 0.01-nm resolution. Absolute wavelength calibration was provided by the spectrum of molecular iodine and the I2 spectral atlas of Gerstenkorn and Luc (1978). Once calibrated, this simple prismgrating wavelength meter configuration offers static operation and can be Grating

1

_., ~R,~M~ ~'R ~, S

~'M

Photodiode ~j' Figure 11.6 Prism-grating wavelength meter. (From Duarte, 1983.)


233

used to characterize either pulsed or CW laser radiation. An improvement in the resolution of this dispersive meter could be attained using an appropriate concave grating.

11.3 INTERFEROMETRY Here some widely applied interferometric configurations in the measurement of wavelength and linewidth are described. Attention is focused on twobeam, and multiple-beam, interferometers. For a detailed treatment on the subject of interferometry, the reader is referred to Born and Wolf (1999) and Steel (1967).

11.3.1

Two-BEAM

INTERFEROMETERS

Two-beam interferometers are optical devices that divide and then recombine a light beam. It is on recombination of the beams that interference occurs. The most well-known two-beam interferometers are the Sagnac interferometer (Fig. 11.7), the Mach-Zehnder interferometer (Fig. 11.8), and the Michelson interferometer (Fig. 11.9). For a highly coherent light beam, such as the beam from a narrow-linewidth laser, the coherence length

Ax~clAu

(11.11)

can be rather large, thus allowing a relatively large optical path length in the two-beam interferometer of choice. Alternatively, this relation provides an

S / J

x

Figure 11.7 Sagnac interferometer. All three mirrors, M1, M2, and M2, are assumed to be identical.

TunableLaserOptics

234

/

Figure 11.8

~X

Mach-Zehner interferometer.

jY

1

/

I Figure 11.9

k

I

Michelson interferometer.

avenue to accurately determine the linewidth of a laser by increasing the optical path length until interference ceases to be observed. Interference in the Sagnac interferometer can be described using Dirac notation via the probability amplitude N

N

(xls) = Z Z (x[j')Q''[l')Q'[s)

(11.12)

j ' = l j=l

where j represents the reflection surface of the beam splitter and the probability amplitude (j'[/') represents the trip around the interferometer, which includes reflections at mirrors M1, M2, and M3. Assuming that (j'[j) = 1

(11.13)


235

Eq. (11.12) reduces to N

(x[s) - ~

(x[/') (/'Is)

(11.14)

j=l

which, for N -

2, becomes

(sis) - (xl2)(2Is) + (Nil)(1 Is)

(11.15)

Here, j - 1 represents the beam splitter in a reflection mode and j - 2 represents the beam splitter in a transmission mode. Interference in the Mach-Zehnder interferometer can be described using N

N

(xls) - ~ ~ (xlk)(k[/) (/'Is)

(11.16)

k=l j = l

for N = 2. Again, if j represents the entrance beam splitter and k the exit beam splitter, we can write

(x[s) = (x]l)(1 [2)(21s ) + (xl l) (1] l) (1]s) + (x]2)(212)(2Is) + (x12) (211)(1 Is) (11.17) Here, j = k = 1 represents the beam splitters in a reflection mode and j = k = 2 represents the beam splitter in a transmission mode. Since (1[1) and (212) lead to x t rather than x, the probability amplitude for this geometry reduces to

(x]s) = (x] l) (112) (2]s) + (x]2) (2] l) (1]s)

(11.18)

For the Michelson interferometer the interference can be characterized using a probability amplitude of the form N

N

f=l

j=l

where j = 1 represents the function of the beam splitter in the reflection mode and j = 2 represents the function of the beam splitter in the transmission mode. In a second-pass approach, j ' = 1 represents the function of the beam splitter in a transmission mode and f = 2 represents the function of the beam splitter in a reflection mode. Since reflection on the first pass (j = 1) and reflection on the second pass ( f = 2) lead back to s, and likewise j = 2 followed by and f = 1, Eq. (11.12) reduces to

(x]s) - (x]2)(2]2)(2[s)+ (xl l) (l l l) (l ls)

(11.19)


236

Here, (111) represents a change at j from reflection to transmission and (212) represents a change from transmission to reflection. It is clear that multiplication of Eqs. (11.15), (11.18), and (11.19)with their respective complex conjugates yields probability equations of an interferometric character.

11.3.2 MULTIPLE-BEAMINTERFEROMETERS An N-slit interferometer, which can be considered a multiple-beam interferometer, was introduced in Chapter 2 and is depicted in Fig. 11.10. In this configuration, an expanded beam of light illuminates simultaneously N slits. Following propagation the N subbeams interfere at a plane perpendicular to the plane of propagation. The probability amplitude is given by N

- Z

j=]

and the probability is N

I(xls)l= -

N

j=l

(11.20)

t~(rm)e i(f~m-f~;)

(rj) m=l

which can also be expressed as (Duarte, 1991, 1993)

I(xls)12 - j~=l ~(rj)2 + 2 ~ t p ( r j ) j=l

t~(rm) COS(Qm -- ~j)

(11.21)

m=j+l

Expressions for two-dimensional and three-dimensional cases are given in Chapter 2. This approach is also applicable to the two-beam interferometer introduced by Hambury-Brown and Twist (1956) (shown in Fig. 11.11) and s

/

Expanded TEMoo laser beam

j

x

N-slit

Diode array

array

Figure 11.10 N-slitinterferometer.


237

M1

I

D1

I M2

I

D2

I~

Figure 11.11 The Hambury-Brown and Twist interferometer. The light from an astronomical source is collected at mirrors M1 and M2 and focused onto detectors D~ and D2. The currents generated at these detectors, il and i2, interferes to produce an interference signal characterized by an equation of the form of Eq. (11.21).

to other multiple-beam interferometers used in astronomical applications (Christiansen et al., 1961). The second multiple-beam interferometer is the Fabry-Perot interferometer, depicted in Fig. 11.12. This interferometer has already been introduced in Chapter 7 as an intracavity etalon. Generally, intracavity etalons are a solid slab of optical glass or fused silica with highly parallel surfaces coated to

r

-I

Optical axis

Optical axis

Figure 11.12 (a) Fabry-Perot interferometer and (b) Fabry-Perot etalon. Dark lines represent coated surfaces. Focusing optics is often included in these interferometers when used in linewidth measurements.


238

increase reflectivity (Fig. l l.12a). These are also known as Fabry-Perot etalons. Fabry-Perot interferometers, on the other hand, are constituted by two separate slabs of optical flats with their inner surfaces coated, as shown in Fig. l l.12b. The space between the two coated surfaces is filled with air or some other inert gas. The optical flats in a Fabry-Perot interferometer are mounted on rigid metal bars with a low thermal expansion coefficient, such as invar. The plates can be moved, with micrometer precision or better, to vary the free spectral range (FSR). These interferometers are widely used to characterize and quantify the laser linewidth. The mechanics of multiple-beam interferometry can be described in some detail by considering the multiple reflection and refraction of a beam incident on two parallel surfaces separated by a region of refractive index n, as illustrated in Fig. 11.13. In this configuration, at each point of reflection and refraction a fraction of the beam, or a subbeam, is transmitted toward the boundary region. Following propagation these subbeams interfere. In this regard, the physics is similar to that of the N-slit interferometer, with the exception that each parallel beam has less intensity due to the increasing number of reflections. Here, for transmission, interference can be described

kj r'

r'

de

(a)

(b)

--~1 D I-~--

Figure 11.13 Multiple-beam interferometer. (a) Multiple internal reflection diagram and (b) detailed view depicting the angles of incidence and refraction.


239

using a series of probability amplitudes representing the events depicted in Fig. 11.13: N

N

N

N

- E Z Z:

(11.22)

m=l l=1 k=l j=l

where j is at the reflection surface of incidence, k is immediately next to the surface of reflection, l is at the second surface of reflection, and m is immediately next to the second surface of reflection, as illustrated in Fig. 11.13. The problem can be simplified considerably if the incident beam is considered a narrow beam incident at a single point j. Propagation of the single beam then proceeds to I and is represented by the incidence amplitude A;, which is a complex number, attenuated by the transmission factor t, so the first three probability amplitudes can be represented by an expression of the form

{l[k) {k[l'){its) - Ait

(11.23)

and Eq. (11.22) reduces to N

{xls} -- Ait ~ {x m} Imil}

(11.24)

m=l

which, using the notation of Born and Wolf (1999), can be expressed as At(p) -- Ait(tt_}_ ttrt2e i, 23 Bra vectors, 23 Brewster's angle, 15, 60, 62, 78, 145 Broadband interferograms, 221-222 Broadband prismatic rotators, 87-90 Cassegrainian telescopes, 17, 53, 127 Cavity linewidth equation dispersion and linewidth narrowing, 64-68 uncertainty principle and, 54-55 Chemical lasers, 195 Closed-cavity design, 122, 128, 133, 134, 194 Coherence length, 49, 162 Coherence time, 49 Colliding-pulse-mode (CPM) locking, 18 Collission, 18 Color-center lasers, 191 Comb, 173 Communications, secure, 21 4-221 Complex number, 23 Continuous-wave (CW) characteristics of, 178 dye lasers, 187-189 excitation, 10-11 gas lasers, 182-184 lasers, 115, 122, 123, 124 Conversion quantities, 250 Copper vapor lasers (CVLs), 181, 185 CO2 lasers, 180, 184

267

268 Cross sections, 5 transition, for Rhodamine 6G, 10 transition probabilities, 11-14 Czerny-Turner spectrometer, 230-231 De Broglie, Louis, 45 Degenerate four-wave mixing (DFWM), 169 Difference-frequency generation, 158, 162-166 Diffraction, 32-38 grating equation, 38, 229 grating spectrometers, 229-231 limit, 51 orders, 38 uncertainty principle and, 46-49 Diffractive tuning techniques, 138-139 Digital laser microdensitometer (DLM), 211-213 Diode-laser-pumped fiber lasers, 191-192 Dirac, P.A.M., 3, 41-42 Dirac optics angular dispersion, 40-41 diffraction, 32-38 interference, 25-32, 41-42 notation, 23-25 reflection, 39-40 refraction, 38-39 Dispersion, multiple-prism double-pass (return-pass), 60-63 generalized, 58-64 linewidth narrowing, 64-68 pulse compression, 68-72 single-pass equation, 58-60 zero-dispersion multiple-prism beam expanders, 67-68 Dispersive wavelength meters, 231-233 Distributed-feedback (DFB) lasers, 134-136 Double-longitudinal-mode (DLM) emission, 119 Double-pass (return-pass) dispersion, 60-63 Double-prism beam expanders, induced polarization in, 81-82 Double-refraction polarizers, 82-84 Duffendack reaction, 182, 184 Dye lasers, 6-11 continuous-wave, 187-189 performance of, 253-268 pulsed, 184-187

lndex

Effective finesse of etalons, 126, 242 Electric dipole moment, 13-14 Electric susceptibility, 157 Electronic states, 7 Emission characteristics in additional lasers, 195-196 in dye lasers, 184-189 in gas lasers, 178-184 in semiconductor lasers, 193-195 in solid-state lasers, 189-193 Equation of refraction. See Snell's law Excimer lasers, 179-180 Excitation mechanisms multiple-level systems, 6-11 rate equations, 5-6 transition probabilities and cross sections, 11-14 External-cavity semiconductor lasers (ECSLs), 131-134 Fabry-Perot etalons, 16, 126, 238, 2"40 Fabry-Perot interferometers, 119, 121,140, 237-238, 240-242, 245 Far-infrared lasers, 196 Feynman, R. P., 3, 23, 32-33, 46 Feynman Lectures on Physics, The

(Feynman), 3 Fizeau configuration, 245 Flashlamp-pumped pulsed dye lasers, 184-187 Flat-mirror resonators, 15 Forced oscillator (FO), 4 configurations, 150-151 Franck-Condon factor, 14, 180 Free-electron lasers (FELs), 195 Free space, secure communications in, 21 4-221 Free spectral range (FSR), 125-126, 142-143, 238, 240-242, 245 Fresnel formulae, 78 Fresnel number, 16, 116, 118 Fresnel rhombs and total internal reflection, 85-86 Gain, 6 Galilean telescopes, 127, 204, 207 Gas lasers, emission in, 178 continuous-wave, 182-184 copper vapor, 181 pulsed atomic, 181 pulsed molecular, 179-180

269

Index

Glan-Foucault prism, 83 Glan-Thompson prism, 83-84 Granularity, 211 Grating equation, 46-47 Littrow configuration, 138 Grating-mirror resonators, 16 Half-wave plate, 87 Hambury-Brown and Twist interferometers, 236-237 He-Cd lasers, 118, 182, 184 He-He lasers, 117-118 Heisenberg's uncertainty principle, 48-49 He-Ne lasers, 182 He-Zn lasers, 118, 182, 184 Higher-order matrices, 111-113 Hybrid multiple-prism grazing-incidence (HMPGI) grating laser oscillators, 128-130, 192 Hybrid telescope grazing-incidence (HTGI) grating configuration, 192 Idler frequency, 163 Interference, 25-32 Interferograms, 221-222 Interferometers multiple-beam, 236-242 two-beam, 233-236 Interferometric computer, 208-211 Interferometric tuning techniques, 139-140 Interferometric wavelength meters, 242-247 Intracavity beam expansion tunable laser oscillators with, 126-131 tunable laser oscillators without, 122-126 Intracavity double pass, 62-63 Intracavity return pass, 62-63 Intrinsic linewidth, 165 Ionic solid-state lasers, 189 Kerr effect, optical, 167 Kerr lens mode (KLM) locking, 167 Ket vectors, 23 Laser cooling, 131, 187, 194 Laser optics, defined, 5 Laser oscillators. See Oscillators Laser-pumped pulsed dye lasers, 184-187

Laser resonators. See Resonators Lasers amplifier, 4 applications, 1 cavities, 14-20 defined, 3-4 historical development of, 2-3 use of word, 3 Law of reflection, 40 Linear polarization Maxwell equations, 75-77 prisms and, 79-85 reflection and, 77-79 rotators and, 85-90 Linear resonators, 17-18 Linewidth narrowing, 64-68 Littrow configuration, 122-123, 124, 126-127 diffractive tuning, 138 external-cavity semiconductor lasers, 131, 132-133 grating equation, 138 interferometric tuning, 140 multiple-prism Littrow (MPL) grating laser oscillators, 128-130 synchronous tuning, 142-143 Longitudinal mode, 118-121 Longitudinal tuning techniques, 141-142 Long-pulse excitation, 10-11 Mach-Zehnder interferometers, 233, 235 Maiman, T. H., 2 Master oscillator (MO), 4 configurations, 150-151 power-amplifier (MOPA) chains, 149 Maxwell equations, 75-77, 159 Maxwell's formula, 76 Michelson interferometers, 233, 235 Microdensitometer, digital laser, 211-213 Microelectromechanical system (MEMS), 195 Miniature lasers, 195 Mode beating, 119 Mode hopping, 142 Modulation measurements, 214 Momentum equation, 45-46 Monochromatic (indistinguishable) photons, 29, 42 Multiple-beam interferometers, 236-242 Multiple double-pass (return-pass) generalized dispersion, 62-63

Index

270 Multiple-level systems, excitation in, 6-11 Mulitple-prism arrays/optics applications, 58, 72 dispersion and linewidth narrowing, 64-68 generalized dispersion, 58-64 introduction of, 57 pulse compression, 68-72 transmission efficiency in, 80-81 zero-dispersion multiple-prism beam expanders, 67-68 Multiple-prism Littrow (MPL) grating laser oscillators, 128-130, 138 Multiple-return-pass beam divergence, 108-110 Narrow-linewidth tunable laser oscillators. See Tunable laser oscillators, narrow-linewidth Nd lasers, 189 Ne-Ne lasers, 117 Newton, I., 57, 228 Newtonian telescope, 53 Nicol prism, 83 Nitrogen lasers, 180 Nonlinear optics applications, 172-174 difference-frequency generation, 158, 162-166 optical parametric oscillation, 162-166 optical phase conjugation, 167-170 Raman shifting, 170-172 refractive index, 166-167 second-harmonic generation, 157, 158, 159-162 second-order nonlinear susceptibilities, 158 sum-frequency generation, 157, 158, 162 N-slit interferometer, 236 applications, 211-222 beam propagation, 206-208 computer, 208-211 experiment, 29-32 geometry of, 29 optical architecture of, 204-208 sensitometry, 222-224 Nuclear-pumped lasers, 196 One-dimensional beam expansion, 128-129 Opened-cavity design, 122, 131-134, 194 Optical clockwork, 172

Optical Kerr effect, 167 Optical materials, 250-252 Optical oscillator, 4 Optical parametric oscillation (OPO), 162-166, 192-193 Optical phase conjugation, 167-170 Optical quantities, 250, 251 Opticks (Newton), 57, 228 Oscillators, 4 See also Tunable laser oscillators, narrow-linewidth linewidth narrowing in pumped pulsed, 65-67 optical, 4, 162-166, 192-193 Particles, wave character of, 45-46 Paschen configuration, 230 Penning reaction, 182, 184 Permeability of free space, 77 Permittivity of free space, 76 Phase-conjugated mirror (PCM), 169 Phase matching, 162 Photonic crystal fiber (PCF), 173 Physical constants, 249 Planck's constant, 45 Plane of incidence, 79 Polarization matching, 144-146 Polarizer multiple-prism multiple-laser (PMPML) sensitometer, 223-224 Power amplifier (PA), 4 Principles of Quantum Mechanics, The (Dirac), 3, 23 Prismatic tuning techniques, 137-138 Prisms polarizing, 79-85 spectrometry, 228-229 Probability amplitude, 23-25 Propagation matrices ABCD, 93-111 higher-order, 111-113 ray transfer matrix, 94, 169, 206-207 Pulse compression, 18 multiple-prism dispersion and, 68-72 Pulsed atomic gas lasers, 181 Pulsed dye lasers, 184-187 Pulsed lasers, 178 Pulsed molecular gas lasers, 179-180 Pump frequency, 163

Index

Quantum cryptography, 221 Quantum energy equation, 45 Quarter-wave plate, 87 Raman shifting, 170-172 Rate equations, 5-6 for multiple-level systems, 6-11 Rayleigh length, 65 Ray transfer matrices. See Propagation matrices Reflection, 39-40 Fresnel rhombs and total internal, 85-86 polarization and, 77-79 telescopes, 127 Refraction, 38-39 double-refraction polarizers, 82-84 Refractive index, 166-167 Resolving power, 48 Resonators description of basic, 4, 14-20 linear, 17-18 unstable, 17, 110-111,151 Retardation plates, 87 Return-pass dispersion, 60-63 multiple beam divergence, 108-110 single beam divergence, 107-108 Rhodamine 6G, 10, 11 Ring resonators, 17-18 Rochon prism, 83 Rotators birefringent, 86-87 broadband prismatic, 87-90 polarization, 85-90 Rowland configuration, 230 Ruby lasers, 189, 190 Ruler, 173 Sagnac interferometers, 233, 234 Second-harmonic generation, 157, 158, 159-162 Second-order index of refraction, 167 Second-order nonlinear susceptibilities, 158 Self-focusing, 167 Sellmeier dispersion equation, 250 Semiconductor lasers, 193-195 Sensitometry, 222-224 Signal frequency, 163 Single-longitudinal-mode (SLM) oscillation, 119, 121 Single-pass dispersion equation, 58-60

271 Single-prism equations, 64 Single-return-pass beam divergence, 107-108 Snell's law, 39 Solid-state lasers, emissions in color-center, 191 diode-laser-pumped fiber lasers, 191-192 ionic, 189 optical parametric oscillators, 192-193 transition metal, 189-190 Spatial coherence, 3 Spatial hole burning, 17 Spectral coherence, 3 Spectrometry, 227 diffraction grating, 229-231 dispersive wavelength meters, 231-233 prism, 228-229 Stimulated Raman scattering (SRS), 170-172 Stokes, 170-171 Sum-frequency generation, 157, 158, 162 Superfluorescence, 20 Superradiant emission, 20 Synchronous tuning techniques, 142-144 TEM00 laser beam, 16, 204 Temporal domain, 119 Ti:sapphire lasers, 189, 190 Transition metal solid-state lasers, 189-190 Transition probabilities, 11-14 Transmission telescopes, 127 Transverse excitation, 15-16 Transverse mode, 116-118 Triplet-level quenchers, 11 Triplet states, 7 Tunable laser oscillator-amplifiers, narrow-linewidth laser-pumped configurations, 148-150 master and forced configurations, 150-151 Tunable laser oscillators, architecture of closed-cavity design, 122, 128, 133, 134, 194 distributed feedback, 134-136 narrow-linewidth external-cavity, 131-134 opened-cavity design, 122, 131-134, 194 with intracavity beam expansion, 126-131 without intracavity beam expansion, 122-126 Tunable laser oscillators, narrow-linewidth defined, 115 efficient design of, 146-148 longitudinal mode, 118-121

Index

272 Tunable laser oscillators, narrow-linewidth polarization matching, 144-146 transverse mode 116-118 tuning techniques, 136-144 Tuning techniques diffractive, 138-139 interferometric, 139-140 longitudinal, 141-142 prismatic, 137-138 synchronous, 142-144 Two-beam interferometers, 233-236, 245 Two-dimensional beam expansion, 126-127

applications of, 49-55 cavity linewidth equation and, 54-55 diffraction identity and, 46-49 Heisenberg's, 48-49 wave character of particles, 45-46 Unstable resonators, 17, 110-111, 151 Vertical cavity surface emitting lasers (VCSELs), 195 Wave character of particles, 45-46 Wave functions, 26-27 Wavelength meters, 221 dispersive, 231-233 interferometric, 242-247

Uncertainty principle alternative versions of, 49

Zero-dispersion multiple-prism beam expanders, 67-68

(continued)

Tunable Laser Optics

Tunable laser optics

Tunable Laser Applications

Phase Conjugate Laser Optics

Laser age in optics

Tunable Lasers Handbook (Optics and Photonics)

Optics of femtosecond laser pulses

Distributed Feedback Laser Diodes and Optical Tunable Filters

Principles of Laser Spectroscopy and Quantum Optics

Tunable Lasers

Tunable Kernel Parameters

Diode Laser Arrays (Cambridge Studies in Modern Optics, No. 14)

Diode Laser Arrays (Cambridge Studies in Modern Optics, No. 14)

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Optics

Tunable diode laser absorption spectroscopy for trace gas measurements with high sensitivity and low drift

Tunable Laser Optics