Solid-State Mid-Infrared Laser Sources Topics in Applied Physics

Topics in Applied Physics Volume 89 AvailableOnline Topics in Applied Physics is part of the SpringerLink service. For...

Author: Irina T. Sorokina | Konstantin L. Vodopyanov

66 downloads 937 Views 14MB 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

Topics in Applied Physics Volume 89

AvailableOnline Topics in Applied Physics is part of the SpringerLink service. For all customers with standing orders for Topics in Applied Physics we offer the full text in electronic form via SpringerLink free of charge. Please contact your librarian who can receive a password for free access to the full articles by registration at:

http://www.springerlink.com/orders/index.htm If you do not have a standing order you can nevertheless browse through the table of contents of the volumes and the abstracts of each article at:

http:/Iwww.springerlink.com/series/tap/ There you will also find more information about the series.

Springer

Berlin Heidelberg New York Hong Kong London Milan Paris Tokyo

Physics and Astronomy~

ONLINELIBRARY

http://www.springer.de/phys/

Topics in Applied Physics Topics in Applied Physics is a well-established series of review books, each of which presents a comprehensive survey of a selected topic within the broad area of applied physics. Edited and written by leading research scientists in the field concerned, each volume contains review contributions covering the various aspects of the topic. Together these provide an overview of the state of the art in the respective field, extending from an introduction to the subject right up to the frontiers of contemporary research. Topics in Applied Physics is addressed to all scientists at universities and in industry who wish to obtain an overview and to keep abreast of advances in applied physics. The series also provides easy but comprehensive access to the fields for newcomers starting research. Contributions are specially commissioned. The Managing Editors are open to any suggestions for topics coming from the community of applied physicists no matter what the field and encourage prospective editors to approach them with ideas. See also: http://www.springerlink.com/physlbooks/tapl Managing Editors Dr. C l a u s E. A s c h e r o n

Dr. H a n s J. K o e l s c h

Springer-Verlag Heidelberg Topics in Applied Physics Tiergartenstr. 17 69121 Heidelberg Germany Emaih [email protected]

Springer-Verlag New York, Inc. Topics in Applied Physics 175 Fifth Avenue New York, NY 1001o-7858 USA Email: [email protected]

A s s i s t a n t Editor Dr. W e r n e r S k o l a u t Springer-Verlag Heidelberg Topics in Applied Physics Tiergartenstr. 17 69121 Heidelberg Germany Email: [email protected]

Irina T.Sorokina KonstantinL.Vodopyanov (Eds.)

Solid-State Mid-Infrared Laser Sources With 263 Figures and 36 Tables

~

Springer

Dr. Irina T. Sorokina

Photonics Institute Technical University of Vienna

Dr. Konstantin L. Vodopyanov Ginzton Laboratory Stanford University

lO4O Vienna Austria

vodopyan@stanford, edu

Gusshausstr. z71387

Stanford, CA 943o5-4o88

USA

sorokina~tuwien, ac. at

Library of Congress Cataloging in Publication Data Solid-state mid-infrared laser sources/Irina T. Sorokina, Konstantin L. Vodopyanov (eds.). p. cm. - (Topics in applied physics, ISSN o3o3-4z16; v. 89) Includes bilbiographical references and index. ISBN 3-54o-o0621-4 (alk. paper) a. Solid-state lasers, z. Laser materials. I. Sorokina, Irina T., 1963- II. Vodopyanov, Konstantin L., 1953IIl. Series

Physics and Astronomy Classification Scheme (PACS): 4z.72.Ai, 42.55.-f, 42.62.-b, 42.65.Ky, 42.65.Yj ISSN print edition: o3o3-4216 ISSN electronic edition: 1437-o859 ISBN 3-54o-oo621-4 Springer-Verlag Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9,1965, in its current version, and permission for use must always be obtained from Springer-Verlag. Violations are liable for prosecution under the German Copyright Law. Springer-Verlag Berlin Heidelberg New York a member of BertelsmannSpringer Science+Business Media GmbH htt p://www.springer.de ' Springer-Verlag Berlin Heidelberg 2oo3 Printed in Germany The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typesetting: DA-TEX - Gerd Blumenstein www.da-tex.de Cover design: design 6" production GmbH, Heidelberg Printed on acid-free paper

56/3141/mf

5 4 3 21 o

Preface

According to the Encyclopaedia Britannica, the "middle infrared" region covers the portion of the electromagnetic spectrum between 400 and 4000 wave numbers, which corresponds to the wavelength range 2.5-25 gin. This range is of particular interest for many applications, especially for spectroscopy, since the electromagnetic frequencies involved coincide with frequencies of the internal vibrational motion of most molecules (the "molecular fingerprint" region). In our book, we define the "mid-infrared" more loosely to cover a slightly broader range starting from ~ 2 gm wavelength. On the longer wavelength end, we even included two chapters on terahertz wave generation, reflecting the fact that as terahertz waves are pushed further to higher frequencies, they converge with the long-wavelength infrared domain (for example, a frequency of 10 THz corresponds to a wavelength of 30 ~tm). The development of solid-state laser sources in the mid-infrared opens unprecedented possibilities in spectroscopy~ as compared to traditionally used Fourier-transform spectrometers. The obvious advantages of lasers are directionality, coherence, narrow linewidth or small pulse duration, and high spectral brightness. Compact and efficient nfid-infrared laser sources can serve advantageously for remote light detection and ranging (LIDAR) (down to parts-per-billion in volume) of many trace gases and vapors that are important in pollution detection, and atmospheric chemistry. Other opportunities for fixed-wavelength and tunable mid-infrared laser sources include medical applications (e.g. microsurgery, dentistry, keratectomy or non-invasive diagnostics by means of breath analysis), ultrasensitive detection of drugs and explosives (down to one part per trillion) using photoacoustie or cavity ringdown spectroscopy, and free-space communications. Near-infrared laser sources (0.8 1.6 p.m) achieved significant market penetration during the past decade, but the technology for mid-infrared lasers, especially those which operate at room temperature, needed considerable improvement, despite the fact that mid-infrared lasers have existed since the beginning of the laser era. In recent years, however, scientific research in the field of generation of coherent radiation in the mid-infrared experienced a revolution. On the one hand, it was largely connected with the rapid progress in semiconductor technology, and on the other, with exploiting new physical ideas and new material development. The most spectacular example is

VI

Preface

the q u a n t u m cascade laser, pioneered by Lucent Technologies in 1994. Another example is optical parametric oscillators that can operate with > 10 W average power or have up to 95 % quantum conversion efficiency. Periodically poled ferroelectric materials revolutionized nonlinear optics and made it possible to create new, highly efficient tunable devices. We have also seen a gigantic leap forward in fiber laser technology with infrared fiber lasers delivering multi-watt output power. Quite different scientific domains (semiconductor physics, solid state physics, material science, nonlinear optics, laser physics, q u a n t u m optics) are converging to investigate new nfid-infrared sources. This book assembles the results from different scientific communities focused on achieving the goal of creating an efficient and inexpensive solid-state mid-infrared laser source. The first two chapters concentrate on semiconductor lasers: Andrd doullid, Philippe Christol, Alexei Baranov and Aurore Vicet describe the latest advances in the 2 5 gm heterojunction laser diodes based on I I I - V as well as IV VI semiconductor structures. Daniel Hofstetter and Jdrdrne Faist review results on distributed feedback q u a n t u m cascade lasers in the wavelength range around 5 p.m and around 10 ~ n and discuss their applications in spectroscopy and free space communications. The chapter by Cornelia Fischer and Markus Sigrist addresses the nonlinear optical technique known as difference frequency generation and gives some examples of spectroscopic applications. The next two chapters describe in detail the underlying principles and advances in nfid-infrared optical parametric oscillators operating in the pulsed mode (chapter by Konstantin Vodopyanov) and in the continuous-wave and synchronously pumped modes (chapter by

Majid Ebrahimzadeh ).

The current state-of-the-art in mid-infrared fiber lasers is reviewed by

Markus Pollnau and Stuart Jackson. The authors describe lasers at transi-

tions ranging from 1.9 to 4 gm in the rare-earth ions T m 3+, Ho 3+, and Er 3+, their population mechanisms, and their power-scaling methods. The chapter by Irina Sorokina surveys ion-doped crystalline lasers operating in the mid-infrared between 2 and 5 microns. Her review includes rare-earth- and transition-metal-based laser crystals, as well as the color-center lasers, with special emphasis on compact r o o m - t e m p e r a t u r e tunable sources based on vibronic lasers. The chapter by Tasoltan Basiev, Vyacheslav Osiko, Alexander Prokhorov and Evgeny Dianov presents the latest achievements in the field of R a m a n lasers based on a variety of crystals and optical fibers. The next two chapters describe terahertz (THz) coherent sources: Kodo Kawase, Jun-ichi Shikata and Hiromasa Ito review methods of generating widely tunable THz waves using nonlinear optical techniques: parametric generation and difference frequency generation. In the chapter by Hiroshi Takahashi, Hidetoshi Murakami, Hideyuki Ohtake, and Nobuhiko Sarukura, the authors describe generation of THz radiation using ultrafast lasers and semiconductor-based materials. The techniques include semiconductor opti-

Preface

VII

cal switches, optical rectification and difference frequency generation, as well as generation of THz waves based on intersubband transitions in semiconductor quantum wells. The last two chapters discuss applications of mi.d-infrared lasers: Frank Tittel, Dirk Richter and Alan Fried present a large number of spectroscopic laser techniques for sensitive, selective, and quantitative trace gas detection with tunable narrow-linewidth mid-infrared coherent sources. The chapter by Benedikt Jean and Thomas Bende reviews applications of mid-infrared lasers in medicine. These applications exploit strong laser light absorption in tissue due to the presence of natural chromophores and include: gynaecology, otorhinolaryngology, neurosurgery, dermatology, urology, dentistry, ophthalmology, cardio vascular surgery and angioplasty. We believe theft this book will be useful for academics, researchers and engineers in various disciplines who require a broad introduction to the subject and would like to learn more about the state-of-the-art and upcoming trends in mid-infrared coherent source development. Finally, we would like to thank all contributors who have found the time, energy and enthusiasm to write these chapters.

Palo Alto Vienna May 2003

Konstantin Vodopyanov Irina Sorokina

Contents

Mid-Infrared 2 - 5 pxm H e t e r o j u n c t i o n Laser D i o d e s A n d % JoulliS, P h i l i p p e Christol, Alexei N. B a r a n o v a n d A u r o r e Vicet . . . 1 1. 2. 3.

Introduction ........................................................ Historical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . State of the A r t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. T y p i c a l H e t e r o j u n e t i o n Laser S t r u c t u r e . . . . . . . . . . . . . . . . . . . . . . . 3.2. M a x i m u m T e m p e r a t u r e of O p e r a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. M a x i m u m O p e r a t i n g T e m p e r a t u r e in the 2-5 ~tm W a v e l e n g t h D o m a i n . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. A n t i m o n i d e Q u a n t u m Well Laser Diodes for the 2 - 3 p.m Spectral R a n g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. S t r a i n e d G a I n A s S b Alloys a n d Q u a n t m n Wells . . . . . . . . . . . . . . . 4.2. F a b r i c a t i o n of A n t i m o n i d e Q u a n t u m Well Laser Diodes . . . . . . 4.3. A n t i m o n i d e Q W Laser Diodes for the 2.0-2.3 ~tm Spectral R a n g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. G a I n A s S b Q W Laser Diodes E m i t t i n g b e y o n d 2.3 bm~ . . . . . . . . 4.5. C h a r a c t e r i z a t i o n of A n t i m o n i d e - B a s e d Laser Diodes Dedicated to Gas D e t e c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. 3 5 bun I n t e r b a n d T y p e - I I Laser Diodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. I n A s S b / I n A s T y p e - I I M u l t i - q a n t u m Well Laser . . . . . . . . . . . . . . 5.2. I n A s / G a I n S b T y p e - I I I "W" Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. C o n c l u s i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 3 5 5 9 11 13 14 20 23 28 33 40 41 45 48 50

High Performance Quantum Cascade Lasers and Their Applications Daniel Hofstetter a n d J6r6me Faist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. 2.

Introduction ....................................................... G a i n Region Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Basic W o r k i n g P r i n c i p l e s of QC Lasers . . . . . . . . . . . . . . . . . . . . . . . 2.2. T h r e e Q u a n t u m Well G a i n Region . . . . . . . . . . . . . . . . . . . . . . . . . . .

61 61 64 64 66

X

Contents

2.3. G a i n Region with D o u b l e - p h o n o n R e s o n a n c e . . . . . . . . . . . . . . . . . F a b r i c a t i o n Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Surface G r a t i n g with L a t e r a l C u r r e n t h @ c t i o n . . . . . . . . . . . . . . . 3.2. Lasers with I n P R e g r o w t h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. J u n c t i o n Down M o u n t i n g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. M e a s u r e m e n t Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. M e a s u r e m e n t Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. D F B Laser with Lateral C u r r e n t I n j e c t i o n in the 5 lain a n d 10 p.m B a n d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. D F B Lasers with I n P O v e r - G r o w n G r a t i n g . . . . . . . . . . . . . . . . . . . 4.4. High Power J u n c t i o n Down M o u n t e d Lasers . . . . . . . . . . . . . . . . . . 4.5. R o o m T e m p e r a t u r e C o n t i n u o u s Wave O p e r a t i o n of QC Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. A p p l i c a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. P h o t o - A c o u s t i c Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. O p t i c a l D a t a Link Using a QC Laser . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

84 86 86 89 92 92

Mid-IR Difference Frequency Generation Cornelia Fischer and Markus W. Sigrist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

3.

1. 2.

3.

4.

5.

6.

I n t r o d u c t i o n to T u n a b l e M i d - I R Laser Sources . . . . . . . . . . . . . . . . . . . . . Basic Principles of N o n l i n e a r Optics a n d Difference F r e q u e n c y G e n e r a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. F r e q u e n c y C o n v e r s i o n Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. N o n l i n e a r Optical Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. P h a s e M a t c h i n g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Material Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Selection of N o n l i n e a r M e d i u m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. P u m p a n d Signal Laser Sources for Difference F r e q u e n c y G e n e r a t i o n . . . . . . . . . . . . . . . . . . . . . . . . Difference F r e q u e n c y Laser Systems R e p o r t e d in the L i t e r a t u r e . . . 4.1. Difference F r e q u e n c y Laser Sources . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Difference F r e q u e n c y Laser Sources Applied in Gas Sensing . Detailed Description of a M i d - I R D F G Laser Source . . . . . . . . . . . . . . 5.1. E x p e r i m e n t a l Set-Up a n d S y s t e m P e r f o r m a n c e . . . . . . . . . . . . . . 5.2. Nd:YAG P u m p Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. E x t e r n a l C a v i t y Diode Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. T u n i n g with Periodically Poled LiNbO3 . . . . . . . . . . . . . . . . . . . . . 5.5. D e t e c t i o n Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A p p l i c a t i o n s of M i d - I R Difference F r e q u e n c y Laser Systems in Gas Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. G e n e r a l C o n s i d e r a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

66 68 68 70 70 71 71 72 80 82

97 100 100 103 105 118 119 121 122 122 122 125 125 126 127 128 129 131 131

Contents E x a m p l e s of Gas S p e c t r o s c o p y P e r f o r m e d w i t h O u r D F G Laser Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusions a n d O u t l o o k . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

XI

6.2.

132 135 136

Pulsed M i d - I R Optical P a r a m e t r i c Oscillators Konstantin Vodopyanov ..............................................

141

1. 2. 3.

141 142 145 145 146 149 151 151 153 153 154 155 160 161 162 163 164 165 165 166

Introduction ...................................................... P r i n c i p l e of O P O O p e r a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O P O s in t h e 2 5 p.m R a n g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. C o m p a r i s o n of N L O M a t e r i a l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. O P O s B a s e d on P P LN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. O e O s B a s e d on P P K T P a n d P P R T A . . . . . . . . . . . . . . . . . . . . . 3.4. O P O s B a s e d on P P K T A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. O P O s Using C o n v e n t i o n a l P h a s e - m a t c h i n g in O x i d e s . . . . . . . . 4. O P O s in t h e 4 201xm R a n g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. C o m p a r i s o n of N L O M a t e r i a l s S u i t a b l e for ~ > 5~tm . . . . . . . . 4.2. O P O s B a s e d on A G S a n d A G S e . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. O P O s B a s e d on Z G P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. O P O s B a s e d on O t h e r C r y s t a l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. T r a v e l i n g - W a v e O p t i c a l P a r a m e t r i c G e n e r a t o r s ( O P G s ) . . . . . . . . . . . 5.1. O P G s B a s e d on P P LN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. O P G s B a s e d on Z G P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. O P G s B a s e d on G a S e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. O P G s B a s e d on C G A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. N a r r o w - L i n e w i d t h O P O s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. U s i n g I n t r a c a v i t y S p e c t r a l - N a r r o w i n g E l e m e n t s . . . . . . . . . . . . . 6.2. N a r r o w - L i n e w i d t h O p t i c a l P a r a m e t r i c G e n e r a t o r O p t i c a l Parametric Amplifier ( O P G OPA) Systems . . . . . . . . . . . . . . . . . 6.3. O P O s w i t h I n j e c t i o n Seeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4. U s i n g a D o u b l y R e s o n a n t C a v i t y . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. E m e r g i n g N o n l i n e a r O p t i c a l M a t e r i a l s for M i d - I R A p p l i c a t i o n s . . . 8. S u m m a r y a n d C o n c l u d i n g R e m a r k s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

167 168 169 171 173 174

Mid-Infrared Ultrafast and C o n t i n u o u s - W a v e Optical P a r a m e t r i c Oscillators Majid Ebrahimzadeh 1. 2.

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

Introduction ...................................................... Optical Parametric Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. O p t i c a l P a r a m e t r i c G a i n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. O p t i c a l P a r a m e t r i c A m p l i f i c a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . .

179 179 181 183 184

XII

3.

4. 5.

6.

Contents 2.3. M i d - I n f r a r e d P a r a m e t r i c G e n e r a t i o n . . . . . . . . . . . . . . . . . . . . . . . . Mid-Infrared OPO Devices ....................................... 3.1. N o n l i n e a r M a t e r i a l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. L a s e r P u m p S o u r c e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mid-Infrared Nonlinear Materials ................................. Mid-Infrared Ultrafast OPOs ..................................... 5.1. M i d - I n f r a r e d P i c o s e c o n d O P O s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. M i d - I n f r a r e d F e I n t o s e e o n d O P O s . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mid-Infrared Continuous-Wave OPOs .............................

186 188 188 190 191 196 197 201 207

7. S u m m a r y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ............................................................

212 214

Mid-Infrared Fiber Lasers M a r k u s P o l l n a u a n d S t u a r t D. J a c k s o n 1. 2.

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

219

Introduction ...................................................... Fiber Materials ................................................... 2.1. S i l i c a t e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. F l u o r i d e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. C h a l c o g e n i d e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. C e r a m i c s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fiber, Pump, and Resonator Geometries .......................... 3.1. P i b e r D e s i g n s for C l a d d i n g P u m p i n g . . . . . . . . . . . . . . . . . . . . . . . . 3.2. F i b e r - L a s e r R e s o n a t o r s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. T h e r m a l I s s u e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

219 220 221 222 222 222 223 223 225 226

4.

S p e c t r o s c o p i c a n d L a s e r P r o p e r t i e s of R a r e - E a r t h I o n s . . . . . . . . . . . . 4.1. S p e c t r a of R a r e - E a r t h I o n s i n G l a s s e s . . . . . . . . . . . . . . . . . . . . . . . 4.2. I n t r a i o n i c P r o c e s s e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. I n t e r i o n i c P r o c e s s e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. O v e r v i e w o f M i d - I n f r a r e d F i b e r L a s e r s . . . . . . . . . . . . . . . . . . . . . .

226 227 228 230 232

5.

T h u l i u m - D o p e d F i b e r L a s e r s a t 1 . 9 - 2 . 0 g m a n d 2.3 2.5 ~ m 5.1. T h r e e - L e v e l L a s e r s a t 1.9 2.0~xm . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. F o u r - L e v e l L a s e r s a t 2.3 2.51xm . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

233 233 234

6.

H o l n f i u m - D o p e d F i b e r L a s e r s a t 2.1 g m a n d 2.9 Ixm . . . . . . . . . . . . . . . 6.1. T h r e e - L e v e l L a s e r s a t 2.1 g m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. F o u r - L e v e l L a s e r s a t 2.9 g m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E r b i u m - D o p e d F i b e r L a s e r s a t 2.7 2 . S g m . . . . . . . . . . . . . . . . . . . . . . . . 7.1. E x c i t e d - S t a t e A b s o r p t i o n a n d C a s c a d e L a s i n g . . . . . . . . . . . . . . . 7.2. L i f e t i m e Q u e n c h i n g b y P r 3+ C o - D o p i n g . . . . . . . . . . . . . . . . . . . . .

3.

7.

.......

7.3. E n e r g y R e c y c l i n g b y E n e r g y - T r a n s f e r U p c o n v e r s i o n . . . . . . . . . Fiber Lasers at Wavelengths Beyond 3 gm ........................ 8.1. Z B L A N F i b e r L a s e r s a t 3.22 ~tm, 3.45 p.m, a n d 3.95 ~tm . . . . . . 8.2. F u t u r e M i d - I n f r a r e d F i b e r L a s e r s . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. C o n c l u s i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ............................................................ 8.

235 236 237 238 238 239 240 242 243 243 244 245

Contents

XIII

Crystalline Mid-Infrared Lasers Irina T. S o r o k i n a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

255

1. 2. 3.

255 257 260 261 281 283 285 285 286 289 292 294 295 310 314 321 323

Introduction ...................................................... Historical O v e r v i e w . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principles of S o l i d - S t a t e Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. S p e c t r o s c o p i c B a c k g r o u n d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Influence of S p e c t r o s c o p i c P a r a m e t e r s on Laser P r o p e r t i e s . . . 3.3. R e a c h i n g t h e T h r e s h o l d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Laser Design C o n s i d e r a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. P u m p Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. C a v i t y Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. T u n i n g M e t h o d s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Q - s w i t c h e d a n d M o d e - L o c k e d O p e r a t i o n . . . . . . . . . . . . . . . . . . . . 5. M i d - I n f r a r e d Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. T M 2 + - D o p e d Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Color C e n t e r Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. R a r e - E a r t h D o p e d Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. C o n c l u s i o n a n d O u t l o o k . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Crystalline and Fiber R a m a n Lasers T . T . Basiev, V . V . Osiko, A. M. P r o k h o r o v a n d E . M . D i a n o v

........

351

I n t r o d u c t i o n (Historical a n d T h e o r e t i c a l Background} . . . . . . . . . . . . . N a n o s e c o n d S R S B a s e d on B a ( N O 3 ) 2 C r y s t a l s . . . . . . . . . . . . . . . . . . . . LiIO~ R a m a n Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P i c o s e c o n d R a m a n Lasers B a s e d on K G W C r y s t a l s . . . . . . . . . . . . . . . N a n o s e c o n d R a m a n Lasers B a s e d on K G W C r y s t a l s . . . . . . . . . . . . . . Search for N e w SRS M a t e r i a l s and Comparative Spectroscopy Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. N e w B a W O 4 S R S C r y s t a l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. B a W O 4 a n d SrWO4 N a n o s e c o n d R a m a n Lasers . . . . . . . . . . . . . . . . . . . 8.1. M i d - I R B a W O 4 R a m a n Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2. M i d - I R Lasing in a C r 2 + : Z n S e C r y s t a l w i t h B a W O 4 R a m a n Laser P u m p i n g . . . . . . . . . . . . . . . . . . . . . . . . 9. B a W O 4 P i c o s e c o n d R a m a n F r e q u e n c y Shifters . . . . . . . . . . . . . . . . . . . . 10. P b W O 4 S R S Shifters a n d R a m a n Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . 11. D i o d e - P u m p e d C W R a m a n F i b e r Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1. R a m a n F i b e r Lasers B a s e d on G e r m a n o s i l i c a t e F i b e r s . . . . . . . 11.2. P h o s p h o s i l i c a t e F i b e r - B a s e d R a m a n Lasers . . . . . . . . . . . . . . . . . 12. C o n c l u s i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

351 355 361 362 366

1. 2. 3. 4. 5. 6.

368 371 372 374 376 376 379 380 381 382 388 388

XIV

Contents

N a r r o w - L i n e w i d t h Tunable Terahertz-Wave Sources Using Nonlinear Optics K o d o Kawase, J u n - i c h i S h i k a t a a n d H i r o m a s a Ito

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

1. 2. 3.

Introduction ...................................................... T h e o r y of T H z - W a v e P a r a m e t r i c G e n e r a t i o n Using P o l a r i t o n s . . . . Injection-Seeded THz-Wave Parametric Generator (is-TPG) ...... 3.1. E x p e r i m e n t a l S e t u p . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. P o w e r E n h a n c e m e n t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. S p e c t r u m N a r r o w i n g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. W i d e T u n a b i l i t y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. A r r a y e d Silicon P r i s m C o u p l e r for a T H z - W a v e P a r a m e t r i c Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. E x p e r i m e n t a l S e t u p . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. E x p e r i m e n t a l R e s u l t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. T u n a b l e T H z - W a v e G e n e r a t i o n from D A S T C r y s t a l Using D u a l S i g n a l - W a v e P a r a m e t r i c O s c i l l a t i o n of P P L N . . . . . . . . . 5.1. E x p e r i m e n t a l S e t u p . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. E x p e r i m e n t a l R e s u l t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. C o n c l u s i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

397 397 399 402 403 404 406 408 411 412 413 416 417 418 419 420

Mid-Infrared and THz Coherent Sources Using S e m i c o n d u c t o r - B a s e d Materials Hiroshi Takahashi, H i d e t o s h i M u r a k a m i , H i d e y u k i O h t a k e and Nobuhiko Sarukura . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. 2.

Introduction ...................................................... T H z R a d i a t i o n from S e m i c o n d u c t o r - B a s e d M a t e r i a l s . . . . . . . . . . . . . . 2.1. P h o t o c o n d u c t i v e S w i t c h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Q u a n t u m C o n f i n e d S t r u c t u r e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. S e m i c o n d u c t o r Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. N o n l i n e a r O p t i c a l P r o c e s s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. C o n c l u d i n g R e m a r k s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

425 425 426 426 431 437 440 442 442

Mid-Infrared Laser Applications in Spectroscopy F r a n k K. T i t t e l , Dirk R i c h t e r a n d A l a n Fried 1.

2.

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

S o l i d - S t a t e M i d - I R S p e c t r o s c o p i c Laser Sources . . . . . . . . . . . . . . . . . . . 1.1. Class "A" Laser Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Class "B" Laser Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F u n d a m e n t a l s of A b s o r p t i o n S p e c t r o s c o p y for Trace Gas D e t e c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

445 445 447 455 462

Contents Spectroscopic Techniques: Signal E n h a n c e m e n t a n d Noise R e d u c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. B a l a n c e d B e a m a n d B a l a n c e d R a t i o m e t r i c D e t e c t i o n (Noise R e d u c t i o n ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. W a v e l e n g t h a n d F r e q u e n c y - M o d u l a t i o n Spectroscopy (Noise R e d u c t i o n ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Long O p t i c a l P a t h L e n g t h Spectroscopy (Signal E n h a n c e m e n t ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. C a v i t y - E n h a n c e d Spectroscopy M e t h o d s (Signal E n h a n c e m e n t ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. P h o t o a c o u s t i c a n d P h o t o t h e r m a l Spectroscopy (Signal E n h a n c e m e n t ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Mid-hffrared Spectroscopic A p p l i c a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Detailed E x a m p l e s Using Selected Spectroscopic Sources a n d Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. S u m m a r y a n d O u t l o o k . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

XV

3.

Mid-IR Laser Applications in Medicine Benedikt Jean and T h o m a s Bende . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.

2.

3.

4.

Introduction ...................................................... 1.1. A b s o r b e r s a n d N a t u r a l l y O c c u r r i n g C h r o m o p h o r e s . . . . . . . . . . 1.2. Laser Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Laser Tissue I n t e r a c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. P h o t o t h e r m a l Laser Tissue I n t e r a c t i o n . . . . . . . . . . . . . . . . . . . . . . 2.2. I R P h o t o t h e r m a l A b l a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. P h o t o s p a l l a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Collateral T h e r m a l D a m a g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental Photothermal Ablation Using Free E l e c t r o n Lasers (FEL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Collateral T h e r m a l Effects as a F u n c t i o n of ~Vavelength . . . . . 3.2. P h o t o m e c h a n i c a l I n t e r a c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. T h e Influence of Specific A b s o r b e r s in the Target M a t e r i a l .. 3.4. Feedback Laser C o n t r o l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clinical Laser A p p l i c a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. G y n e c o l o g y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. O t o r h i n o l a r y n g o l o g y ( E N T ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. N e u r o s u r g e r y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. D e r m a t o l o g y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Urology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6. D e n t a l Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7. O p h t h a l m o l o g y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8. C a r d i o v a c u l a r Surgery, A n g i o p l a s t y . . . . . . . . . . . . . . . . . . . . . . . . .

466 467 470 472 473 478 480 481 496 497

511 511 512 514 516 516 521 522 523 524 525 525 525 527 529 529 530 530 531 533 533 535 535

XVI

Contents

5. P e r s p e c t i v e o f a T u n e a b l e I R L a s e r S o u r c e . . . . . . . . . . . . . . . . . . . . . . . . References ............................................................

536 538

Index

545

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

Mid-Infrared 2–5 µm Heterojunction Laser Diodes André Joullié, Philippe Christol, Alexei N. Baranov, and Aurore Vicet Centre d’Electronique et de Microoptoélectronique de Montpellier (CEM2), UMR CNRS n◦ 5507, Université de Montpellier II, Sciences et Techniques du Languedoc, case 067, 34095 Montpellier Cedex 05, France {joullie,baranov,vicet}@univ-montp2.fr [email protected] Abstract. High performance mid-infrared (2–5 µm) laser diodes are needed for applications such as high resolution and high sensitivity chemical gas analysis and atmospheric pollution monitoring. The goal is to obtain continuous wave laser emission at room temperature, with output power > 1 mW. Different technologies are under investigation to reach this objective. The GaInAsSb/AlGaAsSb strained multi-quantum-well laser showed striking results in CW operation at and above room temperature and appears as a well established technology for laser emission in the 2.0–2.7 µm wavelength range. Beyond 2.7 µm, the IV–VI lasers based on the PbSe/PbSrSe system, type-II “W” quantum well lasers based on the InAs/GaInSb system, and type-II interband and intersubband III–V cascade lasers, are competitive technologies. They could all operate at room temperature or near room temperature, but only in the pulsed regime. This paper presents an overview of the state of the art of these different mid-infrared systems emitting in the 2–5 µm wavelength range. Two heterojunction laser technologies are detailed: the 2–3 µm GaInAsSb multi-quantum-well laser, and the 3–5 µm type-II and type-II “W” laser diodes.

1

Introduction

Light sources emitting in the mid-infrared (IR) wavelength domain (2–5 µm) arouse a growing interest due to their potential applications in telecommunications and molecular spectroscopy. Optical telecommunications can be achieved through the 2–2.5 µm and 3.5–4 µm high transparency atmospheric windows [1]. Gas detection with a high resolution can be developed because a lot of polluting gases and combustion products have strong absorption lines in the mid-IR region, as shown in Fig. 1: NH3 (2.1 µm), HF (2.5 µm), CH4 (2.35 µm and 3.3 µm), HCHO (3.5 µm), HCl (3.5 µm), N2 O (3.9 µm and 4.5 µm), SO2 (4 µm), CO2 (4.25 µm) and CO (2.3 µm and 4.6 µm) [2]. The use of mid-IR emitters offers rich possibilities for areas such as atmospheric pollution monitoring, industrial process control, leak detection, automotive engine exhaust analysis, drug detection, and for the medical diagnosis of disease. Other potential applications include laser surgery, rangefinding, IR I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 1–61 (2003) c Springer-Verlag Berlin Heidelberg 2003

André Joullié et al.

WAVELENGTH (µm)

-1

-2

LINESTRENGTH [cm / (molecules.cm )]

2

5 4.5

4

3.5

3

2.5

2

1E-16

1E-23

1E-17

CH4 HF CO CO2 HCl H2O

1E-18 1E-19 1E-20 1E-21 1E-22

1E-22 1E-21 1E-20 1E-19 1E-18

1E-23

1E-17

1E-24

2000

1E-24

2500

3000

3500

4000

4500

1E-16 5000

-1

WAVENUMBER (cm ) Fig. 1. Strength of absorption lines for various pollutants in the 2–5 µm wavelength domain (HITRAN data base – 1996). Water spectrum in blue refers to the right y-axis. The spectra of CH4 , HF, CO, CO2 and HCl chemical species refer to the left y-axis

illumination and IR countermeasures. These applications require light emitters having a small spectral width and high optical power and brightness. Semiconductor laser diodes can fulfil these criteria. The goal is to obtain devices operating in continuous mode (CW) at room temperature, promising compact sources more widely applicable than conventional technology such as optical parametric oscillators. For gas analysis tunable diode laser absorption spectroscopy (TDLAS) is classically employed [3]. CW output powers of 1–10 mW and single optical mode emission are required, with the ability to tune the lasing wavelength in a wide spectral domain. For other applications much higher CW output powers (≈ 1 W) are needed. A weak temperature dependence of the laser characteristics is also wanted. The characteristic temperature T0 , which describes the evolution of the current threshold with temperature from the approximate expression Ith = I0 exp(T /T0 ), must be high, typically higher than 100 K. Nowadays, in the 2–5 µm wavelength domain, there is no semiconductor laser diode able to work at room temperature in the continuous regime, with the exception of GaInAsSb injection lasers which could operate CW in the shortest mid-IR wavelength region, from 2 to 2.7 µm [4]. Different material systems are employed for the fabrication of mid-IR lasers: lead salt IV–VI lasers, which are the oldest mid-IR devices, for years the only ones commercially available; lasers based on HgCdTe material, which is widely used in IR detection; antimony based type-I or type-II interband lasers which include in their active zone GaInAsSb/AlGaAsSb or GaInAsSb/GaSb quantum wells

Mid-Infrared 2–5 µm Heterojunction Laser Diodes

3

(QWs); type-II “W” quantum well lasers which possess four layers in each period of the active region (e.g. InAs/GaInAsSb/InAs/AlGaAsSb); III–V GaInAs/AlInAs/InP and GaAs/AlGaAs quantum cascade lasers (QCLs), which employ conduction intersubband radiative transitions; and interband quantum cascade lasers (ICLs) which associate interband transitions and the cascade effect. In this chapter we present an overview of the recent developments of these different mid-IR systems operating in the 2–5 µm wavelength domain. Then we detail heterojunction laser systems which are able to operate at room temperature under electrical pumping in this wavelength range: the 2–3 µm GaInAsSb/AlGaAsSb multi-quantum-well (MQW) laser diodes, and the 3–5 µm type-II and type-II “W” laser diodes.

2

Historical

The first demonstrations of mid-IR emission were in 1963, from an InAs p-n junction at 3.1 µm [5] and from an InSb p-n junction at 5.3 µm [6,7]. Laser emission at much longer wavelengths was reported from PbTe and PbSe p-n junction diodes [8]. During the 1980s, standard mid-IR lasers were exclusively fabricated from narrow gap lead and lead-tin based IV–VI semiconductors PbTe, PbSe, PbS, PbSnTe, PbSnSe and PbSSe. Typical devices were diffused laser diodes, operating at low temperature (4–77 K), which emitted in the wavelength range 4–30 µm, the shortest wavelength being limited by the energy gap of PbS or PbTe, to approximately either 4 µm or 5.5 µm at 77 K [9]. A noticeable improvement in IV–VI laser properties was achieved with double heterostructure (DH) lasers, fabricated using advanced growth techniques such as liquid phase epitaxy (LPE), hot wall epitaxy (HWE) and molecular beam epitaxy (MBE) techniques. In 1990, DH lasers based on lead salt compounds remained the standard for mid-IR lasers. Grown on PbS, PbSe or PbTe substrates, they cover the wavelength range from 3 to 30 µm, and can operate beyond 100 K CW. They used as active layers PbEuSSe for the short wavelength region 3–4 µm [10], PbEuSeTe or PbEuSe for the 4–8 µm range, and PbSnTe or PbSnSe for wavelengths beyond 8 µm [11]. At the same time new IR materials are emerging: the GaInAsSb and InAsPSb III–V solid solutions. AlGaAsSb/GaInAsSb/AlGaAsSb DH lasers grown on GaSb substrate, and InAsPSb/InAsSb/InAsPSb DH lasers grown on InAs substrate, showed excellent performance, at room temperature in the 2.0–2.5 µm range [12,13,14,15,16] and at 80 K in the 3–4 µm range [17,18], respectively. An example of a DH laser grown by LPE on an InAs substrate is shown in Fig. 2. Original structures were also studied, where photons are generated, for the first time, from type-II indirect transitions. They consisted of Ga(In)(As)Sb/GaInAsSb p-n [19] and p-p [20] heterojunctions, grown by LPE, which possess at their interface a triangular type-II quantum well.

4


Fig. 2. Cross-section of the first III–V double heterojunction laser emitting beyond 3 µm [17]

After 1990 substantial progress occurred, mainly due to the perfect mastery of the molecular beam epitaxy technique and its use for the growth of DH and quantum well lasers. That was the case for IV–VI laser diodes based on the PbSe/PbSrSe system, which could present high temperature operation in the 3–5 µm wavelength range [21,22,23], and for II–VI lasers based on HgCdTe solid solution which could emit coherent light under electrical injection for the first time [24,25]. But the most impressive results were obtained with III–V quantum well lasers designed by band structure engineering [26,27]. The aim of band structure engineering is to decrease the non-radiative Auger recombination which is the main loss mechanism in long wavelength lasers, and to imagine new laser structures able to reach high optical efficiency. III–V heterostructures possess strong differences in band gap energy and band discontinuity which allows the formation of deep and confining quantum wells. Besides the growth feasibility of extremely complex epitaxial structures by MBE permits the design and fabrication of high performance devices. In 2000, three III–V laser technologies have emerged in this way for laser emission in the mid-IR: the type-II “W” lasers based on the InAs/GaInSb system [28,29,30,31], the interband cascade laser also based on the type-II InAs/GaInSb system [32,33,34,35], and the intersubband quantum cascade lasers employing the GaInAs/AlInAs system [36,37,38,39,40,41,42]. Room temperature or near room temperature operation was obtained from these different technologies in the 3–5 µm wavelength region. On the other hand, in the 2–3 µm wavelength domain, the more conventional strained MQW laser diodes based on the GaIn(As)Sb/AlGaAsSb system showed striking results in CW operation at high temperature and appear as the definitive established technology for laser emission near 2 µm [4,43,44,45,46,47,48,49].


3 3.1

5

State of the Art Typical Heterojunction Laser Structure

A schematic drawing of a Fabry–Pérot laser diode is shown in Fig. 3. The active zone is sandwiched between two thick (2–3 µm) confining layers, which possess high band gap and low refractive index, and must be lattice-matched to the substrate. Classically, the substrate and the first confinement layer are n-type, the active region is undoped, and the second confining layer has a p-type conductivity, so that the heterojunction lies between the active and one cladding layer. A thin and highly doped layer is grown at the top of the structure for contacting. It is often recommended to insert a layer of graded composition between the substrate and the adjacent confining layer, or generally between each confining layer and the adjacent low band gap material, in order to facilitate carrier injection and reduce the series resistance of the laser diode. In DH lasers the active region consists of a thin layer of a narrow band gap material. The radiative transitions arise from band to band recombinations of hole–electron pairs and the emitted wavelength obeys the equation λ(µm) = 1.24/Eg (eV), which means that in order to cover the 2–5 µm spectral range, active mid-IR materials in a DH laser must have a band gap energy in the 0.25–0.62 eV range. There exists an optimal thickness for the active layer, typically 0.15–0.5 µm, depending the wavelength and the optical confinement factor [50]. In III–V lasers, the used substrates are GaSb or InAs, the confining layers are Alx Ga1−x Asy Sb1−y (x ≈ 0.8–1.0), InAs1−x−y Py Sbx (x + y ≈ 0.5) or InAs/AlSb superlattice (the latter provides a straightforward lattice-match to the substrate and circumvents the need for tellurium doping), and the narrow band gap material is Ga1−x Inx Asy Sb1−y quaternary solid solution (for emission at λ ≈ 2–3 µm) or InAs1−x Sbx alloy (for emission at λ ≈ 3–5 µm). For IV–VI DH lasers, the used substrates are n-PbS, p-PbSe or p-PbTe, and typically the p- and n-type confinement layers and the active

Fig. 3. Schematic drawing of a stripe-contact heterojunction laser diode

6


layers all consist of Pb1−x Eux Sey Te1−y , Pb1−x Srx S or Pb1−x Srx Se materials, with an extremely small Eu or Sr concentration in the active layer and a larger Eu concentration in the confining layers. In QW lasers the active region consists of quantum wells shared in the center of a spacer layer (often called a waveguide layer) which has a lower band gap energy and higher refractive index than those of the two adjacent cladding layers. The role of this spacer layer is to reduce the penetration of light in the cladding layers in order to reduce optical absorption loss. QW lasers can be classified into three categories, depending the nature of the band alignment at the well-barrier interface (Fig. 4): nested (type-I), staggered (type-II) and broken gap (type-III). In the first case electrons and holes are confined into the same material (Fig. 4a), while in the two other cases carriers are spatially separated into the adjacent layers. As a consequence indirect radiative recombinations are generated (Fig. 4b and 4c). Note that the typeIII band alignment is a particular case of the type-II one where the conduction band of the well is placed below the valence band of the barrier (Fig. 4c). The knowledge of conduction and valence band-offsets at the interface of a heterostructure is fundamental for the design of the QW laser. Using band gap values, extracted from standard parameters of III–V compounds (Table 1), and valence band-offset values of III–V unstrained hetero-interfaces cladding a) layer

cladding layer

Active zone

waveguide layer

e1

waveguide layer

hh1

Type-I b)

cladding layer

Active zone

cladding layer waveguide layer

e1

hh1

waveguide layer

hh1

Type-II cladding c) layer

cladding layer

Active zone

e1 waveguide layer

hh1

hh1

waveguide layer

Type-III (Type-II broken gap)

Fig. 4. Schematic representation of type-I (a), type-II (b) and type-III or type-II broken gap (c) quantum wells. The fundamental radiative electron–hole transition e1 –hh1 is shown


7

Table 1. Material parameters of various III–V compound semiconductors Parameters

AlAs

AlSb

a0 (˚ A) Eg 4 K (eV)

5.6622 3.13 (Γ) 2.23 (X)

Eg 300 K (eV)

3.03 (Γ) 2.168(X) 0.30 1.25 0.53 2.47

6.1355 2.384 (Γ) 2.327 ( ) 1.696 (X) 2.300 (Γ) 2.211 ( ) 1.615 (X) 0.67 0.88 0.43 1.38

−5.64

Γ (eV) c11 (1012 dyn/cm2 ) c12 (1012 dyn/cm2 ) hydr. deform. pot. av (eV) hydr. deform. pot. ac ( ) (eV) shear deform. pot. b (eV) m∗e (mo ) Ep (eV) m∗lh (mo ) m∗hh (mo ) m∗sh (mo )

GaAs

GaSb

InAs

InP

InSb

5.6532 6.0954 6.0584 5.8688 6.4794 1.519 0.871 ( ) 0.418 1.418 0.236 0.808 ( ) 1.424

0.761 ( ) 0.725 ( )

0.356 1.345 0.172

0.340 1.18 0.54 1.16

0.761 0.88 0.40 0.79

0.380 0.110 0.810 0.83 1.02 0.66 0.45 0.58 0.36 1.00 1.27 0.36

−6.97

−7.17

−6.85

−5.08 −5.04 −6.17

−1.5

−1.4

−1.7

−2.0

−1.8

0.083 21.10 0.160 1.785 0.29

0.102 18.70 0.162 0.469 0.24

0.0665 22.71 0.090 0.377 0.15

0.042 25.7 0.045 0.222 0.075

0.023 0.073 0.0136 21.11 17 22.49 0.027 0.095 0.014 0.263 0.471 0.244 0.049 0.16 0.02

−1.6

−2.1

(Table 2), calculated from the data of Tsou [51], it is possible to determine the type of band alignment for a number of III–V unstrained heterostructures. Figure 5 presents a schematic view at scale of the band-edge alignments of some unstrained III–V binary systems. We can remark that heterointerfaces between arsenides (or antimonides) are always type-I, whereas interfaces between arsenides and antimonides can be type-I (InSb/AlAs, GaSb/AlGaAsSb), type-II (InAs/AlSb, GaInAsSb/GaSb, InAsSb/InAs) and type-III (InAs/GaSb, InAs/GaInSb). This figure also shows that extremely deep quantum wells (of the order of 1 eV) are realized in the conduction band by associating compounds with type-II or type-III band offsets (e.g. InAs/AlSb, InAs/GaSb). In the case of IV–VI heterojunctions, unfortunately, there is a lack of information on the band-offset values, and for studied systems like PbSnTe/PbTe or PbTe/PbEuSeTe, results have been controversial [21]. Besides it was shown that band offsets are temperature dependent for some IV–VI systems like PbSe/PbEuSe. Type-I band alignment and an offset ratio ∆ Ec /∆ Ev = 0.5 are often predicted, but quantitative data on band offsets are needed for a proper design of the IV–VI laser structure.

8


Table 2. Valence band offsets of unstrained III–V systems

A material

'Ev >0 ∆ Ev (eV)

GaSb/AlSb GaSb/AlAs GaSb/GaAs GaSb/InAs GaSb/InSb InAs/InP InAs/AlAs InAs/GaAs AlSb/AlAs AlSb/InAs InSb/InAs InSb/AlAs InSb/AlSb InSb/GaAs InSb/InP GaAs/InP GaAs/AlAs GaAs/AlSb GaAs/GaP

+0.40 +1.16 +0.68 +0.51 +0.08 +0.45 +0.65 +0.17 +0.76 +0.11 +0.43 +1.08 +0.32 +0.60 +0.88 +0.28 +0.48 −0.28 +0.19

AlSb 1.6 eV (X)

InAs 0.35 eV

GaAs 1.42 eV

AlAs 2.16 eV (X)

0.4 eV

Ev

Band alignment Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type Type

I I I III II I I I I II III I I I I II I I I

InSb 0.17 eV

T = 300 K

Ec

GaSb 0.72 eV

A/B system

Fig. 5. Conduction and valence band edges of unstrained III–V binary compounds showing the conduction and valence band discontinuities at the hetero-interfaces for adjacent systems (This diagram is not transitive)


3.2 3.2.1

9

Maximum Temperature of Operation Threshold Current Limitations

The different contributions to the current in an injection laser are the nonradiative Schockley–Hall–Read (SHR) recombinations on deep levels, the interband spontaneous recombinations and the non-radiative Auger recombinations [52]: J = qd(Rnr + Rsp + RAuger ) .

(1)

In (1) J is the current density, d is the width of the active region, Rnr is the SHR recombination rate, Rsp the radiative recombination rate and RAuger the Auger recombination rate. Equation (1) is described by the phenomenological equation: J = qd(AN + BN 2 + CN 3 ) ,

(2)

where N is the concentration of the injected carriers (supposed much higher than equilibrium carrier concentrations of the active region). A is the nonradiative SHR coefficient, B is the spontaneous emission coefficient and C is the Auger coefficient. The laser threshold is obtained when the modal gain Gmod (N ), which increases with the injected carrier concentration N , becomes equal to the total optical losses αtotal : Gmod (N = Nth ) = Γ Gmax (N = Nth ) = αtotal = αint + αFP .

(3)

Γ is the optical confinement factor, Gmax is the maximum optical gain for an injected carrier concentration N , αint is the internal loss coefficient and αFP is the mirror loss of the Fabry–Pérot cavity. The carrier concentration at threshold Nth is then determined by (3), and the threshold current density Jth is given by: 2 3 + CNth ). Jth = qd(ANth + BNth

(4)

For long wavelength lasers, the Auger contribution is widely dominant at room temperature for two reasons: the Auger coefficient exponentially increases with the wavelength [52,53], the internal losses, and particularly losses due to free carrier absorption [54,55], and equally strongly increases with wavelength. As a consequence the carrier concentration at threshold Nth is raised, and the Auger rate, which varies as N 3 , is dramatically increased. In the free carrier absorption mechanism, the photon disappears by collision with a free carrier, electron or hole, which increases its energy. That increase requires a change in the k wave number with the aid of a diffusion mechanism (impurity, optical phonon). The free carrier absorption mechanism is not simple to quantify. A practical equation is [55]: αfc = K[N (cm−3 )/1017 ][λ(µm)/9]p ,

(5)

10


Table 3. Coefficients for the calculation of the free carrier absorption coefficient using (5), as reported by reference [56] Compound

N (1017 cm−3 )

K

p

GaAs InP GaSb InAs InSb AlSb

1–5 0.4–4 0.5 0.3–0.8 1–3 0.4–4

3 4 6 4.7 2.3 15

3 2.5 3.5 3 2 2

-1

FREE CARRIER ABSORPTION (cm )

where K and p are constants dependent on a given material (Table 3). Figure 6 shows the variation of the free carrier absorption coefficient αfc of different mid-IR materials as a function of the carrier concentration N , for a wavelength λ = 3.5 µm. One can see that the αfc value for AlAsSb material, which is currently used as cladding layer in a mid-IR laser structure, becomes high (≈ 20 cm−1 ) at carrier concentrations around 1018 cm−3 which are typical values for injected carriers at threshold. Conversely, the free carrier absorption coefficient of GaSb, InAs and InAsPSb materials, which are usually employed as waveguide layers, remains reasonably small (≈ 2 cm−1 at N = 1018 cm−3 ). Room temperature Auger coefficients for a number of mid-IR materials are shown in Fig. 7. In this figure are reported the experimental data of Meyer et al. [57], Vurgaftman et al. [31], Joullié [47,58], and Pautrat [59]. The following findings can be drawn: In the 2–5 µm mid-IR wavelength domain, the Auger coefficient of II–VI and III–V compounds strongly increases with temperature (from ≈ 10−28 cm6 /s to ≈ 10−26 cm6 /s). In the wavelength range

100

300 K 3.5 µm 10

Sb 0.84 AlAs 0.16 b .13 P 0.29S 0 InAs 0.58

1

0.1

InAs

1

b GaS

10 17

-3

CARRIER CONCENTRATION (10 cm )

Fig. 6. Free carrier absorption coefficient versus free carrier concentration of different mid-IR materials calculated for a wavelength of 3.5 µm


11

Fig. 7. Auger coefficient versus gap wavelength of mid-IR semiconductor materials; open circles: bulk materials; open boxes: type-I GaInAsSb/AlGaAsSb and InAsSb/AlInAsSb multiple quantum wells; solid circles, solid triangles and solid boxes: type-II staggered GaInAsSb/GaSb multiple quantum wells and type-II broken gap InAs/GaInSb “W” quantum wells. For these materials Auger coefficients were derived from photoconductivity measurements (solid circles), pump probe experiments (solid triangles) and correlation of laser threshold with calculated threshold (solid boxes); open triangles: HgCdTe compound; : lead salts

2.0–2.6 µm type-II staggered GaInAsSb/GaSb quantum wells show Auger coefficients smaller than those of type-I GaInAsSb/AlGaAsSb quantum wells. In the 3–5 µm wavelength region, InAs/GaInSb type-II broken gap “W” quantum wells show smaller Auger coefficients than bulk material (HgCdTe, InAs, InAsSb). This reduction is noticeable (its corresponds to a factor ≈ 5), but it is far from that predicted by calculations [60,61]. Around 4 µm, the Auger coefficient of lead salts is strongly inferior to that of type-I and type-II III–V systems by a factor superior to 10. 3.3 Maximum Operating Temperature in the 2–5 µm Wavelength Domain The maximum temperatures of operation Tmax achieved to date from semiconductor laser diodes emitting in the mid-IR wavelength region in the pulsed regime and in the CW regime are shown in Fig. 8 and Fig. 9. In the pulsed or CW regime, whatever the system, the maximum temperatures of operation Tmax regularly decrease with the wavelength. Below 3 µm, excellent results are obtained with type-I MQW not only in the pulsed regime (Tmax = 180 ◦ C at λ = 2.3 µm) but also in CW mode (Tmax = 130 ◦ C at λ = 2.3 µm). In contrast, RT operation is more difficult to achieve beyond 3 µm. In the pulsed

12


GaInAsSb/AlGaAsSb

GaInAs/InAlAs

PULSED Tmax (K)

400 GaInAsSb/GaSb

InAs/GaInSb

lead salts

300 200 InAsSb InSb/AlInSb

100 HgCdTe

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

WAVELENGTH (µm) Fig. 8. Maximum operating temperature of pulsed semiconductor laser diodes in the mid-IR; open circles: Sb-based DH lasers; open boxes: Sb-based type-I MQW lasers; solid boxes: Sb-based type-II MQW lasers; +: Sb-based type-II interband cascade lasers; ∗ GaInAs/AlInAs inter sub-band quantum cascade lasers; : lead salt lasers; open triangles: HgCdTe lasers

CW Tmax (K)

400

GaInAsSb/AlGaAsSb

300

GaInAsSb/GaSbInAs/GaInSb

lead salts

200

InAsSb

100

GaInAs/InAlAs 0 (b)

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

WAVELENGTH (µm) Fig. 9. Maximum operating temperature of CW semiconductor laser diodes in the mid-IR; open circles: Sb-based DH lasers; open boxes: Sb-based type-I MQW lasers; solid boxes: Sb-based type-II MQW lasers; +: Sb-based type-II interband cascade lasers; ∗: GaInAs/AlInAs inter sub-band quantum cascade lasers; : lead salt lasers; HgCdTe laser diodes do not operate CW

CW OUTPUT POWER (mW/facet)


13

10000 1000 InAs/GaInSb 100

GaInAsSb 300K

10 1 0.1

GaInAs/InAlAs

InAsSb (50K)

Type-I MQW DH

ICL

Type-II MQW

QCL

1.5

2.0

80K

W

2.5

3.0

3.5

4.0

4.5

5.0

5.5

WAVELENGTH (µm) Fig. 10. Output power per facet of CW semiconductor laser diodes in the midIR; open circles: Sb-based DH lasers; open boxes: Sb-based type-I MQW lasers; solid boxes: Sb-based type-II MQW lasers; +: Sb-based type-II interband cascade lasers; ∗: GaInAs/AlInAs inter sub-band quantum cascade lasers; : lead salt lasers; HgCdTe laser diodes do not operate CW

regime, only three systems operate at RT (lead salts, type-II “W” and intersubband cascade lasers) while in the CW regime, the best result is obtained for lead salts lasers with a Tmax value of 223 K at (λ = 4.2 µm). In comparison with Fig. 7, we can remark that the performances of laser diodes are strongly dependent on the Auger recombination. Best performances are realized in the wavelength range where the Auger coefficient values are lower. The maximum output power per facet in the CW regime exhibited by the previous systems is reported in Fig. 10. The same behaviour can be observed with a strong decline of the performances beyond 3 µm. Nevertheless, significant output power (100 mW/facet) is measured at low temperature (T = 80 K) for the III–V systems. Lead salt lasers exhibit weak output powers, and HgCdTe laser diodes did not operate in CW mode at T = 80 K.

4 Antimonide Quantum Well Laser Diodes for the 2–3 µm Spectral Range Semiconductor laser diodes emitting between 2 and 3 µm are strongly needed for molecular spectroscopy, remote sensing, and pollution monitoring. The interest in this spectral range is due to the fact that the absorption of many molecules is much stronger compared with shorter wavelengths and, on the other hand, high performance semiconductor light sources and photodetectors operating at room temperature can be realized. The spectroscopic applications require single mode and single frequency laser diodes which can work in the continuous wave (CW) regime near RT. High power CW and quasi-CW

14


(QCW) lasers are attractive as sources for infrared countermeasure applications. The evolution of the diode lasers for the 2–3 µm spectral range followed the requirements imposed by these two fields of application. Two material systems can be used to fabricate laser diodes emitting beyond 2 µm. The first is the InP based materials employing compressively strained GaInAs quantum wells and lattice-matched GaInAsP barriers grown on InP substrates [62,63,64,65,66,67,68,69]. The devices emitting between 2.0 and 2.07 µm [66] exhibited high performance at RT. Pulsed threshold current density as low as 300 A/cm2 and CW output powers as high as 0.8 W/facet have been obtained [69]. Distributed feedback (DFB) lasers operating at 2.043 µm have been realized with the threshold current as low as 6 mA and the maximum output power of 6 mW [68]. Employing InAs quantum wells with GaInAs barriers the spontaneous emission at RT can be extended to 2.4 µm [70] but to date the longest reported emission wavelength for InPbased diode lasers is 2.2 µm [71]. The second system is GaSb based incorporating GaInAsSb active layers and GaAlAsSb barrier/confining layers grown on GaSb substrates. The first MBE grown strained GaInAsSb QW lasers [43] largely outperformed the double heterostructure lasers and type-II interface QW devices and became the main basis for further progress in RT diode lasers for the 2–3 µm spectral range. 4.1

Strained GaInAsSb Alloys and Quantum Wells

The band gap of Ga1−x Inx Asy Sb1−y lattice matched to GaSb varies between 0.725 eV (GaSb) and 0.29 eV (InAs0.91 Sb0.09 ) at RT but strains in mismatched quantum wells usually employed in laser structures modify significantly the band parameters of the material. Type-I and type-II quantum well structures can be fabricated using strained GaInAsSb alloys for the QWs and GaSb barriers. The energy position of the conduction and the valence bands has been calculated in [72,46] using the Van de Walle model [73] based on the local density functional formalism and the “model-solid” approach allows one to calculate energy level shifts due to strains. The strained band offsets have been obtained by adding the shifts due to strain to the unstrained band offsets interpolated from the experimental data on binary compounds. Results obtained with this semi-empirical method are in good agreement with available experimental data [74,75,76]. Figure 11 shows the calculated valence band offset between GaSb and Ga1−x Inx Asy Sb1−y for the entire range of the alloy compositions [72]. As seen from these data, the more favorable for lasing type-I band alignment between the quaternary alloy and GaSb can be obtained only with small As concentrations. With increasing As content both the conduction and the valence bands of GaInAsSb move down with respect to GaSb band positions, improving electron and worsening hole confinement in the QW, and beyond some As concentration the type-II alignment takes place between


15

Fig. 11. Compositional diagram of the Ga1−x Inx Asy Sb1−y quaternary alloy showing the composition of GaSb-lattice-matched alloys (dotted line) and the composition of alloys having a constant strained valence band offset with GaSb (solid lines) [72]. Positive values of band-offsets correspond to a band alignment!type-I Ga1−x Inx Asy Sb1−y / GaSb band alignment; negative values correspond to a typeII band alignment

Fig. 12. Valence band offset between GaSb and strained Ga1−x Inx Asy Sb1−y alloys as a function of As content [46]

these materials. In Fig. 12 the GaInAsSb/GaSb valence band offset is shown as a function of As content for several In concentrations in the alloy [46]. The calculated band gap of pseudomorphic Ga1−x Inx Asy Sb1−y grown on GaSb is shown in Fig. 13 [46]. In the case of compressively strained material (right parts of the curves) the band gap is formed by the Γ -minimum of the conduction band and the heavy hole valence band; for tensile strains (left parts of the curves) the energy distance between the conduction band and the light hole valence band was taken as the band gap. A very important parameter depending on strains which must be taken into account in designing

16


Fig. 13. Band gap of strained Ga1−x Inx Asy Sb1−y alloys grown on GaSb substrates [46]

Fig. 14. Splitting in the valence band of strained Ga1−x Inx Asy Sb1−y /GaSb alloys [46]

laser structures is the spin–orbit splitting in the valence band. The resonance between the split-off valence bands and the energy of emitted photons results in a significant increase of the Auger recombination rate [77] and enhanced intervalence band absorption [78], which deteriorates laser characteristics. The energy difference between the heavy hole valence band and the spin–orbit split-off band is shown in Fig. 14 as a function of the Ga1−x Inx Asy Sb1−y composition as well as the strain induced splitting between the heavy and light hole bands [46]. In a type-II GaInAsSb/GaSb quantum well the radiative transitions occur between electrons in the QW and holes localized in the GaSb barriers near the interface. The energy of these indirect transitions is smaller than in the case


17

Fig. 15. Emission wavelength as a function of the Ga0.5 In0.5 Asy Sb1−y /GaSb quantum well width [46]

of the direct recombination inside the quantum well and the photon energy can be even smaller than the band gap of the QW material. This gives an additional degree of freedom in developing mid-infrared lasers. As the first approximation one can consider that the photon energy is defined by transitions between the fundamental electron level (E1) in the QW and the valence band of GaSb. The emission wavelength calculated in this manner is represented by dashed curves in Fig. 15 as a function of the Ga0.5 In0.5 Asy Sb1−y /GaSb QW width for different As concentrations [46]. In the case of type-I structures the emission wavelength was calculated for direct e1 –hh1 transitions (solid lines in Fig. 15). The calculations were carried out for QW widths not exceeding the critical thickness of strain relaxation if it was found to be smaller than 100 ˚ A. High In and low As concentrations in the QW are required to obtain the type-I band alignment with GaSb. On the other hand, the QW should be narrow enough in order not to exceed the critical thickness, which limits the emission wavelength in structures with direct recombination inside the well. It was found that the maximum emission wavelength of 2.5–2.6 µm at RT still due to the type-I transitions can be achieved with a 80-˚ A-thick QW of Ga0.5 In0.5 As0.12 Sb0.88 and GaSb barriers (Fig. 15) but in the structures with several quantum wells the critical thickness may be exceeded. In general, the indirect radiative recombination in type-II quantum wells is considered as less efficient compared with type-I structures because of the weaker carrier wavefunction overlap. However, under injection electrons confined in the type-II quantum well attract holes from the barriers due to the Coulomb interaction. The hole density near the QW interfaces grows, which increases the wavefunction overlap and the radiative recombination efficiency [79]. Besides, the non-radiative Auger recombination can be signif-

18


icantly reduced in the optimized type-II QW structures compared with their type-I counterparts [80]. Ga1−x Inx Asy Sb1−y /GaSb multiple quantum well (MQW) lasers with different compositions of the (6–7)-nm-wide wells have been studied in [72]. The In concentration was fixed at x = 0.26 or x = 0.35 while the As fraction increased from 0 to 0.16 thus changing the band alignment from type-I to type-II. The laser emission wavelength at RT varied from 1.98 µm to 2.36 µm increasing with the In and As concentrations. No significant degradation of the laser performance was observed after the change of the band alignment type. The threshold current density of (500–900)-µm-long Fabry–Pérot lasers was as low as 280 A/cm2 for the 4-QW type-I structures emitting near 2.0 µm [72,81] and 305 A/cm2 for the 2.36 µm type-II devices with five QWs [82,72]. Moreover, the threshold current density per quantum well of these type-II lasers (62 A/cm2 ) is among the lowest values reported for any lasers emitting beyond 2 µm. The InAs/GaSb system, which is potentially suitable for longer wavelengths, presents an extreme case of the type-II band alignment since the conduction band of InAs lies below the valence band of GaSb. Unlike in conventional type-II structures, the lowest energy of indirect radiative transitions in the InAs/GaSb QW structures is not limited and can be, in principle, tuned down to zero just by increasing the InAs QW width [83]. In order to emphasize this feature of the InAs/GaSb system and for the sake of simplicity the term “type-III” was used in [84] for this kind of band alignment instead of habitual “type-II broken gap”. The InAs/GaSb QW structures were proposed for use in light emitting diodes and lasers quite long ago but at that time they were investigated by photoluminescence and electroluminescence only at low temperatures [85,86,87]. Type-III MQW diode lasers with the active zone consisting of thin InAs layers sandwiched between 30-nm-thick GaSb layers have been reported in [83,46]. The thickness of the InAs QWs was varied from 2 to 8 monolayers (MLs) (0.6–2.4 nm), the number of InAs layers being 5 or 12. The MQW region was embedded between undoped GaSb layers providing the total thickness of the active zone of 0.6 µm required for optical confinement. During the growth special shutter sequences were used at each InAs/GaSb interface in order to force the formation of “InSb-like” bonds, as superior optical properties of this kind of interface have been demonstrated [88]. The structures exhibited bright room temperature electroluminescence (EL). The peak emission wavelength of spontaneous EL shifted from 1.95 µm to 3.3–3.4 µm at RT as the InAs QW width increased from 2 to 8 monolayers. The measured peak emission wavelengths are compared in Fig. 16 with theoretical values calculated for radiative transitions between the first quantified electron state in the InAs QW and hole state located close to the GaSb valence band extremum. The stimulated emission was obtained at 80 K at 1.78–2.85 µm, the threshold current density increasing from 130 A/cm2 to 1.2 kA/cm2 with the InAs


19

Fig. 16. Emission wavelength of InAs/GaSb quantum well structures versus QW width [84]

Fig. 17. Spontaneous and stimulated emission spectra of structures with different width of InAs/GaSb quantum wells: 1–0.6 nm, 2–1.2 nm, 3–2.4 nm [84]

QW width. The five QW structures demonstrated higher threshold current densities than the twelve QW ones. For structures with the thickest QWs, laser action was observed only on the twelve QW devices. Lasing at room temperature was achieved on twelve QW structures containing 0.6- and 1.2nm-thick quantum wells, threshold current densities for 500-µm-long devices being 1.8 and 7.9 kA/cm2 , respectively (Fig. 17). The observed evolution of laser characteristics was attributed to the change in the wave function overlap decreasing with the InAs QW width and to the small gain by QW in type-III structures. For the 0.6-nm-thick QW structure, the characteristic temperature T0 , was 90 K near RT. The peak output power of lasers fabricated from this wafer reached 200 mW, the external quantum efficiency being 25–30 %. In InAs/GaSb QW structures the potential barrier for hole tunneling to the InAs QW is very high compared with the GaInAsSb/GaSb structures discussed above. The wave function overlap, and thus the efficiency of indirect radiative recombination, depends mainly on the position of the electron quantum level. In thin InAs quantum wells the electron level lies close to the conduction band of GaSb providing with electrostatic interaction strong wavefunction overlap. The structures with 2–3-monolayer thick InAs quan-

20


tum wells and barriers made with a smaller bad gap material like GaInAsSb have been proposed to achieve RT lasing near 3 µm. The carrier confinement in the GaInAsSb quantum wells can be improved by use of GaAl(As)Sb instead of GaSb in the barriers. The valence band offset ∆ Ev for the type-I interfaces GaSb/Ga1−x Alx Sb has been calculated in [89] following the common anion rule: ∆ Ev (x) = 0.42x − 0.04x(1 − x) ≈ 0.41x (in eV) .

(6)

The relationship (7) has been proposed in [90] for the valence band offset between GaSb and the lattice matched Ga1−x Alx SbAs alloy: ∆ Ev (x) = 0.48x (in eV) .

(7)

In order to find the valence band offset in GaInAsSb/GaAl(As)Sb quantum wells the values calculated from (6) or (7) should be added to ∆ Ev in the GaInAsSb/GaSb system presented in Figs. 11, 12, and 14. Estimations made on this basis show that type-I GaInAsSb/GaAl(As)Sb quantum wells providing emission at wavelengths up to ≈ 3 µm can be realized. 4.2

Fabrication of Antimonide Quantum Well Laser Diodes

The GaSb based QW laser structures are usually grown on (100)-oriented nGaSb substrates in solid source MBE systems. As a rule, the MBE machines are equipped with a valved arsenic cracking source producing As2 molecules, which provides better control of relatively small As amounts in the epitaxial layers than As4 cells. On the other hand, both Sb2 and Sb4 sources were successfully employed to grow high quality laser structures. Beryllium is used for p-type doping of the antimonide materials and Te-doped n-type layers are grown with GaTe or Sb2 Te3 sources. The growth temperatures are in the range of 420–480 ◦C but sometimes the first GaAlAsSb cladding layer is grown at 500–530 ◦C. A typical GaSb based QW structure is shown in Fig. 18 [91]. The active zone contains 1–5 GaInAsSb quantum wells and 20–30 nm thick GaAlAsSb barriers, sandwiched between GaAlAsSb spacers (SCH layers in Fig. 18) of the same composition. The QW thickness varies in the 6–20 nm range, 10 nm being the most popular value. The Al concentration in these layers lattice matched with the substrate varied from 0.25 to 0.4 but lasers with GaSb barriers have also been reported [46,72,82,92,93,94,95]. The active zone is usually undoped and is of weak p-type (p = (1 − 5) × 1016 cm−3 ) due to the native acceptors in GaSb based materials. The Al fraction in GaAlAsSb cladding layers is greater than in the barriers in order to provide electrical and optical confinement; its value is usually 0.8–0.9 in the structures with an active zone based on the quaternary alloy and 0.55–0.6 in the type-II structures with GaSb barriers. The graded gap regions before the first cladding layer and after the second one (Fig. 18) serve to reduce the series resistance


21

Fig. 18. Energy band diagram of a GaAlAsSb/GaInAsSb/GaSb laser structure, after [91]

due to the high potential barriers at these interfaces. The growth of the structures is started with an n-GaSb buffer layer and ended with a heavily doped p-GaSb contact layer. The GaInAsSb solid solutions are not miscible in the entire range of their compositions because the sizes of the constituent atoms differ considerably. As a consequence, a solid phase miscibility gap exists in this system at temperatures used for MBE growth (Fig. 19). The stable region in the alloy composition plane is limited by the binodal curves. In this region, the solid phase decomposition can take place without surmounting an activation barrier. In the region between the binodal and spinodal curves, an energy barrier retards the decomposition and the solid is metastable [96]. Calculated spinodal isotherms for GaInAsSb solid solutions are shown in Fig. 19. In most MBE grown GaInAsSb QW laser structures, stable alloys with low As concentrations have been employed. It is necessary to note that the consideration used often to calculate the miscibility gap [96] concerns the solid in thermodynamic equilibrium. The real MBE epitaxial growth conditions may be far away from the equilibrium favoring formation of the unstable alloy. Besides, once the solid is formed, the decomposition process can be largely prevented by the presence of an energy barrier related to the strain produced by forming a lattice-mismatched second phase in a single crystalline alloy [97,98]. For these reasons both metastable and unstable materials can be grown using MBE or MOVPE. However, the metastable nature of the solid leads to clustering of like constituents and/or ordering of the large and small atoms in a planar geometric arrangement. These phenomena lead to major changes in the properties of the semiconductor alloys, which depends significantly on the growth conditions [97]. The Ga0.46 In0.54 Sb0.52 As0.48 alloy

22


Fig. 19. Spinodal isotherms for (solid curves). Gax In1−x As1−y Sby Straight line designates the separation direction. Dashed lines represent compositions for lattice-matching to GaSb (a), InAs (b) and InP (c), after [96]

lattice matched with GaSb having the composition lying in the center of the miscibility gap has been successfully used in double heterostructure diode lasers emitting near 3 µm up to 255 K in the pulsed regime and up to 170 K in CW mode [48]. The metastable GaInAsSb solid solutions have also been employed in high performance QW lasers [48,95]. The majority of results on antimonide QW lasers reported to date have been obtained on devices with cleaved Fabry–Pérot resonators. The cleaved facets are often coated for high reflection (> 95 %) on the backside and for partial antireflection (3 %) on the front facet in order to increase the singleended output power and to protect the AlSb based claddings against oxidation. The ridge waveguide geometry is used to produce single mode and single frequency lasers (Fig. 20, [99]). To obtain a single mode waveguide the ridge must be narrow (5–10 µm) and the etch depth should be carefully controlled. Too deep etching favors lateral waveguide modes and shallow etching increases the device threshold because of the current spreading. Lasers designed for high power emission are wide (100–200 µm) in order to improve heat removal by the laser heatsink.

Fig. 20. Schematic of a GaSb-based single-frequency ridge waveguide laser, after [99]


23

4.3 Antimonide QW Laser Diodes for the 2.0–2.3 µm Spectral Range Antimonide lasers emitting between 2 and 2.3 µm can be realized on the basis of type-I GaInAsSb quantum wells with large conduction and valence band offsets. The wells can be made of the alloy in the region of stable compositions and its quality is less sensitive to the MBE growth conditions. The first MBE grown GaSb based QW lasers were reported by Choi and Eglash in 1992 [43]. The active zone of the structure consisted of five 10 nm thick Ga0.84 In0.16 Sb0.86 As0.14 quantum wells and six 20 nm thick Ga0.8 In0.2 As0.02 Sb0.98 barriers. Broad stripe lasers 100 µm wide with uncoated facets were fabricated from the grown wafer. They emitted near 2.1 µm with the threshold current density as low as 260 A/cm2 for 2-mm-long lasers and differential quantum efficiency of 70 % for devices 300 µm long. The lasers operated at up to 150 ◦ C in the pulsed regime and their characteristic temperature T0 was 113 K near RT. The maximum output power in the CW regime at 20 ◦ C reached 190 mW/facet with the initial slope efficiency of 0.1 W/A. Large lasers emit in several spatial modes and in multiple frequencies (longitudinal modes), which makes them unsuitable for molecular spectroscopy. The 8-µm-wide ridge waveguide (Fig. 20) single frequency lasers have been fabricated from the same wafer [99]. The threshold current was as low as 29 mA for the lasers 300 µm long. The maximum output power in the coated facets diodes reached 28 mW/facet with the initial slope of the light-current curves corresponding to the quantum efficiency of 23 %. The lasers exhibited single frequency emission at 2.135 µm with the side mode suppression ratio (SMSR) of 20 dB. In the structure reported in [43,99] the thickness of the active zone (0.17 µm) is comparable with the total thickness of the QWs. In such a waveguide the main part of the lasing mode propagates in the doped cladding layers, where free carrier absorption losses are considerable. Free carrier absorption increases with wavelength, and the effect of the cladding layer loss becomes stronger in long wavelength lasers. Separate confinement heterostructures (SCH) with a broadened waveguide were proposed in [91] in order to decrease internal losses in long wavelength QW lasers. In these structures the region containing the QWs is surrounded with SCH layers (spacers) thus increasing the thickness of the waveguide and decreasing the part of the lasing mode propagating in the cladding layers (Fig. 18). Three laser structures emitting near 2 µm at RT with different waveguide thickness were investigated. The threshold current densities for 2-mm-long facet coated devices were in the range of 260–400 A/cm2 . The internal loss was found to decrease from 32 cm−1 to 2 cm−1 in the structures where the total waveguide thickness W was increased from 0.12 to 0.88 µm. The part of the light propagating in the cladding layers drops from 40 % to 4 % in this series of structures while the QW confinement factor is nearly constant (5–9 %) because the mode intensity in the center of the waveguide, where the QWs are located, does not

24


change significantly. A maximum output power of 1.2 W and a maximum CW efficiency of 36 % were observed for a 100-µm-wide and 2-mm-long laser which had the broadest waveguide. The maximum output power and differential quantum efficiency were lower in the structures with thinner waveguide (Fig. 21). In the single quantum well (SQW) structure with an 0.8-µm-thick waveguide the threshold current density was reduced to 115 A/cm2 for uncoated lasers with 2–3 mm cavity lengths [100,101]. These lasers exhibited a CW output power of 1.9 W, a differential quantum efficiency of 53 %, and a QCW output power of 4 W (Fig. 22). Broadened waveguide GaInAsSb/GaAlAsSb SQW lasers with the threshold current density as low as 50 A/cm2 for 3-mm-long devices have been reported in [102,103]. In this structure the strain in the QW was increased to 1.4 %, using the alloy with the In fraction of 0.22, instead of 1 % in the QWs with xIn = 0.19 employed in [91,100,101]. These lasers had emission wavelengths of ≈ 2.05 µm, characteristic temperature of 65 K, internal quantum efficiency of 95 %, and internal loss coefficient of 7 cm−1 . A single ended CW power of 1 W has been obtained for a 100 µm aperture and the maximum power efficiency was 22.5 %. Linear arrays of tapered lasers operating with 1 ms current pulses have been fabricated from the same wafer [104]. Peak power over 3 W was obtained for nine-element arrays at 18.5 A. Up to 1.7 W peak power, within a 65 mrad full angle cone, was measured in the far field using anamorphic collimating lens arrays, fabricated by mass transport in GaP. The efficiency of the carrier confinement in the active zone defined by band offsets between quantum wells and barriers influences considerably the laser performance. The band offsets can be tuned by the QW composition and strain and the composition of barrier layers. GaInAsSb/GaAlAsSb QW lasers with increased Al concentration in the barriers have been reported

Fig. 21. The CW output power characteristics for 2 mm cavity length, coated lasers with different waveguide thicknesses W , emitting at 2 µm, after [91,101]


25

Fig. 22. Output power characteristics of a 2 µm GaAlAsSb/GaInAsSb SCH SQW broadened waveguide facet coated laser with 2 mm cavity and 200 µm aperture: (a) CW regime at 15 ◦ C; (b) QCW regime at 10 ◦ C with 100 µs pulse duration and 100 Hz repetition rate, after [100,101]

in [105,106]. The 0.924-µm-thick active zone of the structures contained two or four strained Ga0.85 In0.15 Sb0.94 As0.06 wells 10.5 nm wide separated by 18-nm-thick barriers with Al fraction of 0.4 compared with xAl = 0.25 in [91,100,101,102,103]. The change in the Al concentration deepens ∆ Ev by about 70 meV up to 154 meV. Threshold current densities for a 2060-µmlong 2-QW laser and 1730-µm-long 4-QW devices were 174 and 225 A/cm2 , respectively. A differential quantum efficiency of 74 % was observed for a 900µm-long 2-QW laser and a characteristic temperature of 140 K was measured

26


for a 4-QW device in the pulsed regime. The improvement of these parameters has been explained by the increased valence band offset. GaInAsSb/GaAlAsSb QW lasers emitting at 2.26 µm near RT have been reported in [107,108,109]. To achieve this emission wavelength the In concentration in 10-nm-thick QWs was increased to xIn = 0.3, the As fraction being yAs = 0.06. The threshold current density for 2-mm-long lasers at 280 K was as low as 100 and 200 A/cm2 for the structures with one and three quantum wells (TQW), respectively. The characteristic temperature of the threshold current was 110 K. Single-ended output powers of 240 mW [107] and 350 mW [109] in the CW regime were obtained for 64-µm-wide TQW lasers with coated facets at 280 K. Single frequency operation required for spectroscopy applications can be achieved in narrow ridge Fabry–Pérot lasers [99] but DFB lasers provide better noise characteristics and a high side mode suppression ratio which is less sensitive to driving conditions, i.e. laser current and temperature. Usually DBF is obtained by a grating, which is patterned with the period of the Bragg wavelength in the laser waveguide. After the patterning the epitaxy is continued in order to complete the laser structure. For InP based lasers the overgrowth technique is well established. For GaSb based DFB lasers complications arise as Al-containing layers are used in the cladding layers of the waveguide. Because of the reactivity of the Al with oxygen high quality overgrowth is very difficult. Gain coupled GaInSb/GaAlAsSb QW DFB lasers fabricated without regrowth have been reported in [110]. In these lasers single frequency DFB emission was obtained by a first-order chromium Bragg grating with a period of 273.4–269.4 nm on both sides of the 4-µm-wide laser ridge. The Bragg grating acts as a periodic absorber for the evanescent part of the laser mode and thereby provides gain coupling. For lasers with 900 µm length threshold currents were around 20 mA and the maximum CW output power was 10 mW at 20 ◦ C. The lasers exhibited single frequency emission near 2 µm with SMSR of 31 dB (Fig. 23). In Fig. 24 the DFB emission wavelength versus temperature is compared with the data obtained on a reference Fabry–Pérot laser without grating. The wavelength shift with increasing temperature is 0.20 nm/K for the DFB mode, while it is 1.22 nm/K for the Fabry–Pérot mode. The fast shift of the Fabry–Pérot emission reflects the temperature dependence of the band gap. For DFB lasers the emission wavelength is less dependent on the temperature as the mode is predominantly determined by the refractive index. Another type of semiconductor laser, which easily provides single frequency emission, is a vertical cavity surface emitting laser (VCSEL). In a VCSEL light propagates vertically rather than laterally through the structure consisting of an active zone embedded between distributed Bragg reflectors (DBRs). With such a small cavity, the gain bandwidth of the device can only support a single longitudinal mode. The very high difference in refractive indexes of GaSb and AlSb permits one to fabricate high qual-


27

Fig. 23. Spectrum of a laterally coupled DFB laser under CW operation. The cavity length is 900 µm and the facets are ascleaved, after [110]

Fig. 24. Temperature variation of laser emission wavelength for a laterally coupled DFB laser with 900 µm cavity length, after [110]

ity DBRs with reflectivity sufficient for VCSEL operation (> 99 %) using less than 20 stacked pairs on the base of these materials. The first electrically pumped VCSEL emitting beyond 1.8 µm has been reported in [111]. The Te-doped bottom and Be-doped top DBRs consisted of 17.5 period and 15 period AlSb0.93 As0.07 − GaSb quarter-wave stacks, respectively. Six 6.5 nm thick Ga0.65 In0.35 Sb0.9 As0.1 quantum wells with 22-nm-thick GaSb barriers were surrounded by 225 nm of GaSb to form a one wavelength long optical cavity. The last p-GaSb quarter-wave layer of the top DBR also served as a contact layer. The 200-µm-diameter VCSELs fabricated from different parts of the wafer emitted between 2.18 and 2.32 µm in the pulsed regime at RT with a threshold current density of 2 kA/cm2 and an output power of 20 mW. Emission spectra of a VCSEL are shown in Fig. 25. In the spontaneous regime the VCSELs emitted up to 1 mW of continuous optical power concentrated in a spectral band as narrow as 10 nm.

28


Fig. 25. RT spontaneous and lasing emission spectra of a GaSbbased VCSEL [110]. (a) spontaneous spectrum measured from the edge of the device (b) spontaneous spectrum measured from the surface (c) lasing emission spectrum measured from the surface. Driving conditions: (a,b) repetition rate: 20 kHz, pulse width: 20 µs (c) repetition rate: 10 kHz, pulse width: 1 µs

4.4

GaInAsSb QW Laser Diodes Emitting beyond 2.3 µm

A series of RT lasers emitting beyond 2.3 µm based on type-II GaInAsSb/ GaSb QW structures has been reported in [46,72,82,92,93,94,95]. The initial structure [71,82] contained five 7-nm-thick Ga0.65 In0.35 Sb0.85 As0.15 quantum wells between 30-nm-thick GaSb barriers, the total thickness of the active zone being 0.3 µm. The wells and the barriers had a type-II band alignment with a valence band offset ∆ Ev of −60 meV (the heavy hole subband for the QW) and a conduction band offset ∆ Ec = 350 meV. The 200-µm-wide lasers emitted between 2.35 and 2.39 µm with a threshold current density of 305 A/cm2 for 820-µm-long devices at 23 ◦ C and T0 = 55 K. Ridge waveguide 8-µm-wide lasers operating in the CW regime at RT were fabricated from a similar wafer [94]. The lasers had threshold currents in the range 60–140 mA and emitted 2–5 mW/facet of optical power in the range 2.34–2.4 µm. Employing a broadened 0.8-µm-thick waveguide and an optimized QW growth procedure internal losses were reduced from 17 to 7.8 cm−1 and the laser performances were considerably improved [95]. The 10 µm ridge lasers operated in the CW regime up to 50 ◦ C with an output power of 20 mW/facet at RT. The internal quantum efficiency was found to be 89 % and the power


29

efficiency reached 20 %. The lasers exhibited single frequency operation in a large range of currents and temperatures. The side mode suppression ratio reached 30 dB (Fig. 26) and was nearly constant between longitudinal mode hops. Hysteresis was observed in the lasing wavelength versus current characteristics (Fig. 27), which indicated autostabilization of the main mode. The high SMSR, comparable with that of DFB lasers, and the observed autostabilization of the lasing mode have been explained by the photorefractive effect due to DX centers in the Te-doped Ga0.4 Al0.6 Sb0.95 As0.05 cladding layer acting as a periodic saturable absorber. The 2.3 µm laser performances have been further improved by using the Ga0.65 Al0.35 Sb0.98 As0.02 alloy in the waveguide instead of GaSb [48,112]. The 0.8-µm-thick active zone consisted of three 10-nm-thick type-I GaInAsSb quantum wells of nearly the same composition, xIn = 0.35, yAs = 0.16, as

Fig. 26. Emission spectrum of a type-II GaInAsSb/GaSb QW laser measured at 10 ◦ C [95]

Fig. 27. Laser emission wavelength versus current (same laser as in Fig. 26) [95]

30


in [95], the Al concentration in cladding layers being increased to xAl = 0.9. Narrow ridge 5-µm-wide Fabry–Pérot diodes with uncoated facets were soldered epi-side down onto copper heatsinks with a lead–tin alloy. At 22 ◦ C and in the CW regime the threshold current was as low as 20 mA for 440-µm long devices increasing to 55 mA for a length of 2.5 mm. The maximum output power reached 45 mW/facet and was limited by catastrophic degradation of the uncoated facets at an optical power density of the order of 1 MW/cm2 . The devices operated in the CW regime at temperatures up to 140 ◦ C (Fig. 28) and up to 180 ◦C in the pulsed regime (1 µs, 10 kHz) despite the high thermal resistance of these narrow ridge devices with non-optimized heatsinks. The threshold current increased very slowly with temperature, the characteristic temperature T0 being as high as 110–130 K near RT and 60– 80 K at 100 ◦C in the CW regime. The differential quantum efficiency also varied slowly with temperature. Its near threshold value decreased from 50– 75 % at RT to 35–50 % at 100 ◦ C depending on the cavity length. The internal quantum efficiency and internal losses were found to be 90 % and 8 cm−1 , respectively [112]. The emission wavelength shifted with temperature from 2.25 µm at RT to 2.43 µm at 135 ◦ C in the CW regime and single frequency operation could be obtained with a side mode suppression ratio up to 20 dB (Fig. 29). The autostabilization of the lasing mode was not observed in these devices, which is due probably to smaller concentration and activation energy of DX-centers in the Te-doped GaAlAsSb alloy with higher Al content [113]. The emission wavelength of the type-II lasers with GaSb barriers was increased to 2.65 µm by use of the Ga0.5 In0.5 Sb0.78 As0.22 alloy in the 7-nmthick wells [46,93]. A pulsed threshold current density of 1.5 kA/cm2 with a characteristic temperature T0 = 45 K has been obtained for 80-µm-wide devices. Total peak output power reached 130 mW with a differential quantum efficiency of 14 %. The internal quantum efficiency and the internal loss were found to be 33 % and 62 cm−1 , respectively. The high internal loss was

Fig. 28. Output power characteristics of a 680-µm-long GaInAsSb/ GaAlAsSb QW ridge waveguide laser at different temperatures [112]


31

Fig. 29. Emission spectra of a GaInAsSb/GaAlAsSb QW ridge waveguide laser at different temperatures [112]

explained by strong free carrier absorption in the cladding layers for this quite narrow, 0.35-µm-thick, waveguide. GaInAsSb/GaAlAsSb QW lasers emitting at 2.78 µm have been reported in [45]. This long emission wavelength was obtained with relatively small concentrations of In and As in four 10-nm-thick quantum wells, xIn = 0.24 and yAs = 0.16, respectively. Pulsed laser operation of stripe lasers 22 µm wide was obtained at 15 ◦ C with a threshold current density of 10 kA/cm2 , maximum power output of 30 mW, and a maximum differential quantum efficiency of 9 %. The lasers operated pulsed up to 60 ◦ C with a characteristic temperature of 58 K over the range of 0–40 ◦C. The same devices exhibited CW operation up to 234 K emitting at 2.7 µm at this temperature with an output power of 1 mW and a differential quantum efficiency of 0.6 % [114]. GaInAsSb/GaAlAsSb double QW lasers operating between 2.3 and 2.7 µm at RT were studied in [115,116,117,118]. The laser structures contained two Ga1−x Inx Sb0.98 As0.02 quantum wells, Ga0.7 Al0.3 Sb0.98 As0.02 barriers and spacers in an undoped 0.8-µm-thick active zone. The In content in the 10–20 nm-thick wells varied from 0.25 to 0.4 while the As concentration was fixed at the low level in order to keep the well composition outside the miscibility gap. The compressive strain in the QWs increased from 1.5 % to 2.3 % with increasing In fraction. Gain guided Fabry–Pérot lasers with 100µm-wide stripe contacts and 1 or 2 mm cavity lengths were characterized. The devices were mounted epi-side down onto water cooled copper heatsinks. The laser emission wavelength shifted from 2.3 to 2.7 µm with increasing In concentration in the wells. All the devices were able to operate in the CW regime at RT. Figure 30 shows differential quantum efficiency (ηd ) and the threshold current density (Jth ) for 2-mm-long diodes emitting at different wavelengths. The threshold current density in the pulsed regime increased only slightly, from 230 to 300 A/cm2 , while the wavelength increased from 2.3 to 2.6 µm. Corresponding values of ηd were independent of wavelength between 2.3 and

32


Fig. 30. Threshold current densities (pulsed and CW) and differential efficiencies (pulsed) for lasers of different wavelength. For uncoated lasers (2.6 and 2.7 µm), the measured values of efficiencies were doubled to account for light emission from both facets, after [4]

Fig. 31. CW output power characteristics for three diode lasers of different wavelength. For 2.6 µm diode lasers, the measured values were doubled to account for light emission from both facets, after [4]

2.6 µm and close to 30 % for all 2-mm-long devices. Figure 31 shows CW output powers of 500, 250, and 160 mW that were obtained for lasers emitting at 2.3, 2.5 and 2.6 µm, respectively. The characteristic temperature T0 near RT decreased with wavelength from 90–100 K for 2.3 µm lasers to 30–40 K for 2.7 µm devices and the upper limit of CW operation decreased from about 100 ◦ C to 30–40 ◦C in the same range. Despite much longer emission wavelengths, antimonide QW lasers emitting between 2 and 2.7 µm exhibit excellent performances comparable or even superior to those of telecommunications lasers benefiting from well established growth and processing techniques of InP based materials. Their remarkable thermal stability (CW operation up to 140 ◦C, CW T0 of 130 K at 2.3 µm and pulsed T0 of 140 K at 2.0 µm) is due to better electron confinement in GaInAsSb/GaAlAsSb quantum wells possessing high band offsets, which reduces carrier leakage at elevated temperatures. The strained GaInAsSb/GaAlAsSb material system renders a high degree of freedom in


33

band engineering of QW structures. By varying the QW and barrier composition the same emission wavelength can be obtained from the wells with different band offsets in the conduction and in the valence band. Analysis of the obtained results allows one to conclude that strong confinement of one sort of carriers (electrons) is more advantageous than moderate confinement of both electrons and holes. The Coulomb attraction prevents current leakage from the QW in the structures with low confinement barriers for one type of carriers and provides efficient recombination even in type-II structures where the holes are located outside of the QW. The less impressive temperature behavior of type-II GaInAsSb/GaSb lasers is due to weak confinement of electrons in the wells with GaSb barriers rather than to absence of hole confinement without injection. The performances of lasers employing unstable GaInAsSb alloys in the quantum wells show the high quality of these MBE grown materials, at least at the small penetration into the miscibility gap. The disadvantage of antimonide laser diodes compared with InP based devices is the use of easily oxidizing high Al content GaAlAsSb cladding layers, which causes rapid degradation of the lasers. This problem can be solved by coating the laser facets and/or encapsulation of the devices in the inert gas environment. The ageing test of a GaInAsSb/GaAlAsSb ridge waveguide laser [48] which is in progress at the Montpellier University (France) shows no significant degradation of the diode with uncoated facets encapsulated in a housing filled with dry nitrogen for more than 13 000 h. 4.5 Characterization of Antimonide-Based Laser Diodes Dedicated to Gas Detection Antimonide-based semiconductor lasers are well adapted to gas detection in the atmosphere. Their emission window, between 2 and 2.8 µm, is centered on a vapor window [119] (at 2.3 µm) where many pollutant gases present absorption lines (Fig. 1). Gas detection requires two main conditions from laser diodes: single frequency emission and tuning. Gas detection with multimode laser diodes is of course possible but it presents more difficulties for the identification of the line and the calibration of the measure. In order to realize a more correct measure, laser diodes without pure single frequency emission suppose the use of a monochromator to select a mode, which makes the optical setup more complicated [120]. The tuning is the property of the laser to exhibit a variation of the emitted wavelength function of injected current or operating temperature. This tuning must be obtained at least for a few GHz to be able to perform gas detection at atmospheric pressure. Gas detection at low pressures (< 50 Torr) requires less tuning, but is more demanding about the spectral width of laser diodes which has to be smaller than the HWHM of the detected line. Of course, industrial applications require a significant lifetime (more than 10 000 h) from the devices.

34


4.5.1 Laser Diode Characterization for Gas Detection Applications The results presented here have been obtained on type-I [49] and type-II [48] strained quantum well laser diodes. These devices exhibited single mode or multimode emission, a function of the injected current and/or the laser temperature. To select a diode for a specified application, it is necessary to know which are its operating conditions. A classical way is to measure the emission spectra at different currents or temperatures in order to build a mode map. Figure 32 shows a typical mode map realized for two lasers. These data give information about the emitted wavelength, the spectral purity and the tuning, but the main drawback is that it takes a long time to acquire. New methods have to be developed to characterize the laser diodes for gas analysis. Fabry–Pérot interferometers are widely used in Tunable Diode Laser Absorption Spectroscopy (TDLAS) to calibrate the tuning of light sources and identify the absorption lines [121]. The interference figure given by the emitted light through such an interferometer depends on the spectral purity of the beam. Figure 33 shows the data for a type-I laser diode maintained at 67 ◦ C, recorded using a bulk germanium interferometer of 2.3 GHz FSR and 2380 Gain tuneability 1 nm/K

20°C CW

2.28

Mode tuneability 0.1 nm/K

2370

III II

I

2.24

Lambda (nm)

Lambda (µm)

IV 2.26

2360

2.22

2350 2.20 30

40

50

60

70

I (mA)

(a)

80

90

100

0

5

10

15

20

25

Temperature (°C)

(b)

Fig. 32. (a) Mode-map realized CW from 30 to 100 mA of a type-I GaInAsSb/AlGaAsSb QW laser diode emitting around 2.27 µm at a fixed temperature (Tlaser = 20 ◦ C). Emitted power is represented in a log scale. One can see four zones: I–Multimode emission; II–Mode hopping area; III–Monomode emission in a large current range giving a 2 cm−1 continuous tuning; IV–Multimode emission. (b) Mode-map of a type II GaInAsSb/GaSb QW laser diode around 2.3 µm as a function of temperature. The gain tunability changes 10 times faster than longitudinal mode tunability

Mid-Infrared 2–5 µm Heterojunction Laser Diodes T = 67°C CW

35

4 3

P(a.u.)

2

1

0

20

40

60

80

100

I(mA)

Fig. 33. Single acquisition of a Fabry–Pérot cartography of a type-I GaInAsSb/ AlGaAsSb QW laser diode realized at 67 ◦ C. One can see four zones. Zone (1) is before lasing, no light is emitted, the laser threshold is 41 mA. (2) and (4) zones show well-contrasted oscillations: the emission is single-frequency. (3) shows nocontrasted oscillations: the emission is multimode 200

Multi frequency

180

Single frequency

160 140 I(mA)

120 100 80 60

Threshold current

40 Spontaneous emission

20 0 14

15

16

17

18

19

20

21

22

T(°C)

Fig. 34. Fabry–Pérot mode-map in the current/temperature space for a type I GaInAsSb/AlGaAsSb QW laser diode. The threshold current is materialized by a white line from 75 mA at 14 ◦ C to 80 mA at 22 ◦ C and correspond to a T0 of 160 K in this temperature range. Dark areas show multimode emission and clear areas show single-frequency emission

a GaInAs photodiode. Because the data are difficult to extract, a numerical recipe is necessary for giving a constant value function of the spectral purity of the emission. By acquiring such data at different temperatures, one can build a Fabry–Pérot cartography (Fig. 34) which gives a convenient way to characterize the spectral quality of the laser [122]. To perform gas detection, it is important to know if a laser diode is suitable to detect a specific species. Keeping the idea of cartography from

36

André Joullié et al. 100

80

I(mA)

60

40

20

0 15

16

17

18

19

20

T(°C)

21

22

23

24

25

Fig. 35. Map in the temperature/current space of the absorption of a type-I laser diode emitting around 2.28 µm. This map has been realized with a tiny calibrated cell containing pure CH4 for a 10 mm-length. The threshold current varies from 40 mA at 15 ◦ C to 45 mA at 25 ◦ C

Fabry–Pérot maps, absorption cartography has been developed. In this procedure, a laser diode is tuned by a current ramp at different temperatures. The emitted laser beam is detected after passing through a gas cell containing a calibrated molecular species. Gas detection is very sensible when realized on a strong absorption line. An example of absorption cartography is given in Fig. 35. In the map, strong absorption lines appear as white lines. The absorption lines are iso-energy lines (at constant wavelength): that gives precious information to localize strong absorption lines and to know how to compensate the temperature effect by the current to follow a given line. Such cartographies can be made for different gas species in order to evaluate if the tested diode is adapted to a gas or another one. These new methods are of great interest for testing and selecting laser diodes for gas analysis. They are all based on the tuning property of the laser diodes. 4.5.2

Tuning Properties of Antimonide-Based Laser Diodes

For a laser cavity of length L, the wavelength λ of the longitudinal mode m is given by: λ=

2nL . m

(8)

The real refractive index n[T, N (T )] is sensitive to the temperature T and to the carrier density N . The carrier density is quasi-constant above the threshold N ≈ Nth but it is temperature dependent. Therefore, the sensitivity of the refractive index to a change of the junction temperature can be given by: ∂n ∂n dN ∆ = + × × ∆T . (9) ∂T ∂N dT


37

From the plasma effect, a change ∆ N of the carrier density induces an adjustment ∆ n of the refractive index given by [123]: r0 λ2 m0 ∆ = ∆N , (10) 2πn me where r0 and m0 are, respectively, the radius and the mass of a free electron and me is the effective mass of the carrier. The temperature dependence of the threshold current can be fitted in a small temperature range by the experimental Pankove law Jth = J0 exp TT0 2 = kNth . By differentiation and with Nth ≈ N at threshold, one obtains:

dN N = . dT 2T0

(11)

As the temperature expansion of the laser length L is negligible compared to the other variations, the temperature tunability of a laser diode is: λ dn r0 λ2 m0 N ∆λ = − . (12) ∆T n dT 2πn me 2T0 One can note that for very low characteristic temperature T0 , it is possible to observe a blue shift of the wavelength. This effect has been previously observed by Baranov et al. on InGaAs QW lasers grown by OMCVD on InAs having a very low characteristic temperature (T0 = 21 K) and emitting around 2.6 µm [124]. As expected in GaInAsSb QW lasers, high values of T0 give a red shift and the first term in the bracket is the most important. The tunability by the current is: ∆λ ∆T ∆λ = . ∆I ∆T ∆I

(13)

T In order to obtain ∆ ∆ I , two experimental techniques can be used. The first one consists in the measurement of the thermal resistance RTh as defined by:

T (I) = RTh PTh (I) ,

(14)

where T (I) is the junction temperature for the injected current I and PTh (I) the thermal power. T (I) can be measured from CW and pulsed measurements of the emitted optical power Pop (I). One can deduce the thermal power PTh (I) from: PTh (I) = I(Vd + Rs I) − Pop (I) ,

(15)

where Vd is the voltage corner of the laser I(V ) characteristic and Rs the series resistance.

38

André Joullié et al. 2.8 20.5°C 24°C 28.5°C 33°C 40°C 44°C

2.6

'Q/'I (GHz/mA)

2.4 2.2 2.0

50°C 55°C 59.5°C 67.3°C 70.7°C 83.1°C

1.8 1.6 1.4

30

40

50

60

70

80

90

100

Average current (mA)

Fig. 36. Current tuning of a type-I GaInAsSb/AlGaAsSb QW laser diode for different temperatures measured using a Fabry–Pérot bulk germanium crystal

As Pop (I) = (I − Ith )ηd (I) hν e , by differentiation one obtain: hν hν dηd dT (I) = RTh Vd − ηd (I) − (I − Ith ) + 2Rs I dI e e dI

(16)

which shows that the tuning is independent of the temperature and linear with the injected current. This is clearly shown in Fig. 36 which presents a typical experimental measurement of the tuning of a GaInAsSb QW laser emitting around 2.3 µm. The experimental data are in agreement with the dn = 1.47 × 10−4 K−1 given by Ghosh [125]. model using the value dT By differentiation of λ(T, I), one obtains for ∆ λ = 0: ∆T ∂λ/∂I =− . ∆I ∂λ/∂T

(17)

As seen in Fig. 37, one can obtain ∆ T /∆ I directly in a map of a gas absorption line (∆ λ = 0) in the current temperature space. 4.5.3

Gas Detection with Antimonide-Based Laser Diodes

Antimonide-based laser diodes are well adapted to TDLAS, especially for detecting species like CH4 , NH3 , H2 O, CO, CO2 or HF. Figure 38 gives the optical scheme of an open path gas analysis demonstrator using a GaInAsSb laser device. It is a 30 cm × 30 cm setup, where the detection is based on 2f wavelength modulation spectroscopy. The diode laser emission goes through a mirror objective [126] which makes a focus point of the laser considered as a point source. One of the main interests of such optical scheme is to provide a geometrical focus point from which the whole setup can be designed. It


39

80 70

'T = 1,4°C 60

'I = 12,7 mA

I (mA)

50 40 30 20

T0=97K

10 0 16

17

18

19

20

21

T (°C)

Fig. 37. Map of a gas absorption line in the current temperature space. One can easily deduce (∆ T /∆ I) = 110 K/A. The black line gives the evolution of the threshold current versus temperature. For this laser the T0 parameter is 97 K

Fig. 38. “Open path” optical setup used for gas analysis

is also less dependent on chromatic aberrations and absorption than transmission systems. The beam is collimated through a lens to a 200 ˚ A pellicle beam splitter. A large part of the light is transmitted and directed to a corner cube, then it comes back to an off-axis parabolic mirror (F = 110 mm, θ = 26◦ ) and it is focused on a GaInAs detector detecting up to 2.6 µm. The signal recorded by this detector is the “open path signal”. 8 % of the incident light is reflected by the splitter, passes through a reference cell filled with a calibrated pure gaseous species near atmospheric pressure and is focused on the same GaInAs detector by an off-axis parabolic mirror (F = 110 mm, θ = 26◦ ). This signal is the “reference signal”. Comparison by correlation

40


800 700

Output power

600 500 400 300 200 0

100 0

0

20

2000

40 60 I (mA)

4000

80

100

6000

8000

10000

12000

14000

Hours Fig. 39. Time evolution of the emitted power of a laser diode. Working conditions are 65 mA CW at 23 ◦ C. The stability is quite good; the relative variations are due to the influence of ambient temperature on the detector which is not stabilized

of the two signals gives, after calibration, the concentration of the analysed species in the atmosphere. Industrial applications need data about the lifetime of the devices. Figure 39 shows, in arbitrary units, the evolution in time of optical power of a type-I GaInAsSb/AlGaAsSb laser diode emitting around 2.27 µm. This diode was maintained at constant temperature (23 ◦ C) and current (65 mA, or 2Ith ). One can see that the emitted power and P (I) characteristics stay relatively constant in time, without any degradation. This laser diode exhibited 15 500 h of CW working [122].

5

3–5 µm Interband Type-II Laser Diodes

To achieve laser emission in the mid-wavelength infrared region (3–5 µm, a wide variety of Sb-based systems, grown on GaSb or InAs substrates, have been investigated. There are InAsSb/InAsPSb [127] and GaInAsSb/ AlGaAsSb [128] double heterostructure lasers, type-I or type-II quantum well lasers with InAsSb/InAs [129,130,131], InAsSb/InAlAsSbg [132,133], InAsSb/InAsP [134,135] or InSb/AlInSb [136] heterostructures. All these traditional systems have presented excellent performance at low temperatures, but laser emission was limited to 225 K in the pulsed regime (Fig. 8) and to 175 K in the CW regime (Fig. 9) [133]. To improve the


41

maximum temperature of operation Tmax , a new laser design has been proposed based on interband type-II broken gap InAs/GaInSb/InAs “W” structures [28]. This system presents several advantages. First, the “W” arrangement induces a phase position of the carriers that increases the electron–hole overlap integral and hence the optical matrix elements, giving values comparable to type-I structures. Next, this structure keeps the advantages of type-II ones like a substantial reduction of Auger recombination [57,60] and internal losses [137], because of the small in-plane electron and hole masses and the elimination of the resonance between the energy gap and the split-off valence band or with any other lower valence subbands. Based on these considerations, interband type-II “W” laser diode seems to be a convenient structure to reach room temperature operation in the mid-infrared wavelength range (3–5 µm). That result was obtained around 3.3 µm at RT in pulsed mode [29], and the laser diodes operated in CW up to 195 K [30]. In this part, we report results of two kinds of non-cascade quantum structures: type-II InAsSb/InAs MQWs grown on InAs substrate and InAs/ GaInSb/InAs “W” structures grown on GaSb substrate. 5.1

InAsSb/InAs Type-II Multi-qantum Well Laser

The InAsSb/InAs system possess a weak type-II band alignment [138,139]. Figure 40 shows the energy-band diagram of the InAs0.92 Sb0.08 /InAs MQW laser structure on an InAs substrate, calculated using a Qc value, the ratio of the strained conduction band-offset to the band gap difference between

Fig. 40. Band diagram of the type-II InAs0.92 Sb0.08 /InAs MQW structure on InAs substrate. On the lower part, fundamental electron (e1 ) and heavy hole (hh1 ) presence probability densities are reported. The calculated fundamental e1 −hh1 optical transition is 3.41 µm at 80 K and 3.86 µm at RT. The overlap wavefunction |fe1 |fhh1 |2 is 42 %, neglecting electrostatic interaction between carriers of opposite charge

42


the strained well and the InAs barrier, equal to 2.30 [139]. This MQW structure is composed by ten 50 ˚ A-thick InAsSb wells and nine 200 ˚ A-thick InAs “barrier” surrounded by 0.5 µm thick InAs waveguide layers to optimize the optical confinement. In Fig. 40, we present only five InAsSb QWs and four InAs barriers for the sake of simplicity. The cladding layers are made from AlAs0.16 Sb0.84 material lattice-matched to InAs, providing good optical and electrical confinement. The electron and heavy hole presence probability densities reported in the lower part of Fig. 40 show that InAsSb quantum wells are not completely decoupled by the 200 ˚ A thick InAs barriers. Consequently, without injection, an important value of the carrier wavefunction overlap |fe1 |fhh1 |2 = 42 % is obtained for this indirect radiative structure. On the other hand, the 2D dimensionality of carrier dispersion energy, which theoretically ensures high differential gain, is not seriously degraded with a first electron miniband of about 6 meV thick. The electron–hole overlap integral increases under injection as a consequence of the Coulomb attraction between the carriers of opposite charges [79]. Such electrostatic interaction generates a modification of the band structure near the type-II heterointerfaces and enhances the optical coupling between electrons and holes. Under typical carrier injection (σ = 1 × 1012 cm−2 ) overlap values as high as 60 % can be obtained, making this type-II structure comparable to a type-I one. In order to evaluate the lasing performances of such type-II structures, the optical gain G(ω) was calculated for different injected carrier densities σ in the active region [140,141]. The optical gain G increases with the injected carrier density σ in the active layer, varying from σ = 0.25 × 1012 cm−2 to σ = 2×1012 cm−2 (Fig. 41). At low temperature (Fig. 41a), it is easy to induce gain because the difference between electron and hole quasi-Fermi energies (Efc − Efv ) reaches the energy transition e1 -hh1 from a carrier injection of σ = 0.2 × 1012 cm−2 . On the other hand at RT (Fig. 41b), it is necessary to inject at least σ = 0.70 × 1012 cm−2 to satisfy the Bernard–Duraffourg condition. However, whatever the temperature, the maximum gain Gmax due to the first energy level transition (e1 -hh1 ) increases with the injected carrier v ) with σ. density like the variation of the Fermi–Dirac distribution (fnc − fm To achieve laser emission, it is necessary to satisfy the following condition: Gmod = αtotal ((3) in Sect. 3.2.1), where Gmod is the modal gain of the structure (the useful part of the gain) and αtotal the total optical losses. The modal gain is obtained from the relation: Gmod = Np Γp Gmax , where Gmax is the peak gain value extracted from the gain curves (Fig. 41), Np is the number of QW periods (Np = 10) and Γp is the optical confinement factor per QW. Γp was found to be 0.33 % for this structure [141]. Figure 42 displays the modal gain versus carrier concentration N3D at T = 100 K and T = 300 K, respectively. The carrier concentration is deduced from the relation N3D = A). σ/Leff where Leff is the effective width of each quantum well (50 ˚

-1

OPTICAL GAIN (cm )

Mid-Infrared 2–5 µm Heterojunction Laser Diodes 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 -100 200

V = 2.0 x 10

12

-2

cm

43

a)

1.75

T = 100 K

1.5

TE mode

1.25 1.0 0.75 0.5 0.25

250

300

350

400

450

500

550

ENERGY (meV)

400

-1

OPTICAL GAIN (cm )

V = 2.0 x 10

300

T = 300 K

12

-2

cm

b)

1.75

TE mode

1.5

200

1.25 1.0

100

0.75

0 0.5

-100 150

0.25

200

250

300 350 400 ENERGY (meV)

450

500

Fig. 41. Gain spectra versus photon energy in a single InAsSb/InAs MQW structure for different carrier injection densities varying from σ = 0.25 × 1012 cm−2 to σ = 2 × 1012 cm−2 . Calculations are made at T = 100 K (a) and T = 300 K (b), respectively

The maximum modal gain which can be reached at low temperature is 55 cm−1 (N3D = 1 × 1018 cm−3 ) but only 11 cm−1 at RT. It is evident that the InAsSb/InAs MQW laser possesses a too poor modal gain to operate at RT. Two combined effects limit the modal gain value: firstly, it is a consequence of insufficient carrier confinement in the InAsSb well, added to a low confinement factor Γp per well; secondly, there is a restrictive injection value because of the relocation of electrons outside the quantum well when the pseudo-Fermi level Efc reaches the top of the quantum well at high injection from N3D = 1 × 1018 cm−3 . The MQW laser structure was grown on (100) oriented n-InAs substrates in a solid-source MBE system [142]. Mesa-stripe Fabry–Pérot laser diodes were fabricated from these QW structures. The devices exhibited, under the

44

André Joullié et al. 60 55 50

-1

Modal Gain (cm )

45

MQW InAsSb/InAs

T=100 K

10 periods

40 35 30 25 20 15

T=300 K

10 5 0 0,0

TE mode

0,5 1,0 18 -3 N3D (10 cm )

1,5

Fig. 42. Modal gain Gmod versus carrier concentration N3D calculated at T = 100 K and T = 300 K. The MQW active region is composed of ten periods

pulsed regime, laser emission at 3.4–3.5 µm with a maximum operating temperature of 220 K [143]. Figure 43 shows the peak optical power versus current in the pulsed regime at 90 K and the emission spectrum recorded at 200 mA is shown in the inset. The threshold current is 130 mA, corresponding to a threshold current density of 200 mA/cm2 and the initial slope efficiency is 47 mW/A/facet. At low temperature, up to 120 K, the threshold current slowly increases, with a characteristic temperature T0 = 110 K. Above 150 K, it rapidly increases, with T0 = 20 K (Jth = 15 kA/cm2 at T = 220 K). The internal loss coefficient was estimated to be 30 cm−1 and then the total optical loss was around 45 cm−1 . This value confirms the impossibility to operate at RT. In the CW regime, ridge waveguide lasers emitted near 3.5 µm up to 130 K (Fig. 44), with output power efficiency of 40 mW/A/facet at 90 K and a characteristic temperature T0 = 40 K. This conventional MQW structure on an InAs substrate showed excellent performance at low temperature but seriously limited above 220 K. To reduce the threshold current and increase the maximum operation temperature, it is possible to optimize the active region design by using a “W” geometry for laser structures grown on an InAs substrate or a GaSb substrate. That was made in references [144] and [145], where the “W” geometry was obtained from the InAs/InAsSb/InAs/InAlAsSb type-II QW system grown by MBE [144], and from the InAsP/InAsSb/InAsP/InAsPSb type-II QW system grown by MOVPE [145]. Despite excellent results exhibited at 80 K (CW operation with output efficiency of 30–60 mW/A/facet), no improvement of the maximum temperature of operation was observed. Growth problems, arising from the solid phase miscibility gap of the quaternary alloys (InAlAsSb waveguide, In(As)PSb cladding layers), inducing crystal imperfections and


45

80

Mesa stripe w = 70 µm L = 900 µm 0.1 µs - 10 kHz

POWER (mW/facet)

70 60 50 40

200 mA

30 20 2,0 2,5 3,0 3,5 4,0 4,5 5,0

10

WAVELENGTH (µm)

0 0,0

0,5

1,0

1,5

2,0

2,5

3,0

CURRENT (A)

Fig. 43. Peak output power versus current for an InAsSb/InAs MQW device having a cavity length of 900 µm, measured at 90 K in the pulsed regime. Inset shows the emission spectrum at 200 mA [143] 100 105 111 116

0,5 CW POWER (mW/facet)

90 K

121 126

0,4 Ridge 10 µm L = 670 µm 0,3

0,2 130 K 0,1

0,0

0

20

40

60 80 100 CURRENT ( mA )

120

140

Fig. 44. CW output power versus current for a ridge laser of length 670 µm, at several temperatures varying from 90 to 130 K [143]

excessive optical losses, have to be solved to obtain the expected increase of Tmax . 5.2

InAs/GaInSb Type-III “W” Laser

InAs/GaInSb heterostructures possess type-III band alignment where electrons and holes are spatially separated. As a consequence, the overlap of the electron and hole wavefunctions is weak, inducing poor radiative efficiency.

46


In order to increase this overlap, a new interband laser design on a GaSb substrate has been proposed by Meyer et al. [28] obtained from band-structure engineering. The designed active region is made of several QW periods, each period including a GaInSb “hole” quantum well sandwiched between a double InAs “electron” quantum well, which forms a conduction band profile in the shape of a “W”. Excellent results have been obtained by the “W” structure design reported in Fig. 45 [29,30]. The active region consisted of five or A)/InAs (15 ˚ A) separated ten W periods of InAs (15 ˚ A)/Ga0.75 In0.25 Sb (27 ˚ by a AlGaAsSb (80 ˚ A) spacer layer. This layer furnishes a two-dimensional dispersion for both electrons and holes with strong confinement in each “W” period. These MQWs are embedded in a 0.6 µm thick AlGaAsSb waveguide layer to maximize the optical confinement factor Γ in the active region, while minimizing free-carrier absorption losses in the cladding layers. The optical cladding layers are made from AlGaAsSb, lattice-matched to GaSb. The fundamental e1 –hh1 optical transition was expected around 3.3 µm at RT. Figure 46 displays the modal gain Gmod versus carrier concentration N3D at T = 100 K and T = 300 K, respectively. Whatever the temperature, comparison with the classical InAsSb/InAs MQW (Fig. 42) shows the superiority of the “W” structure with a Gmod value four times higher. For a reasonable injected carrier concentration N3D = 1×1018 cm−3 , the modal gain can reach 230 cm−1 at T = 100 K and 35 cm−1 at T = 300 K. This last value confirms that the “W” structure can theoretically operate at RT, but only for laser structures having small optical losses with excellent structural quality of the grown layers. Modal gain versus total current density J (3) is shown in Fig. 47. Standard values for Sb-based structures are considered [141]. The non-radiative recombination coefficient A = 108 s−1 can be taken as temperature independent; the radiative coefficient B was taken equal to 7 × 10−11 cm3 /s and AlGaAsSb

GaInSb

conduction band

valence band

InAs InAs 2

II = 57 %

fhh1

2

fe1

2

˚)/InAs (15 A ˚)/ Fig. 45. Schematic band diagram of the InAs(15 ˚ A)/GaInSb (27 A AlGaAsSb (80 ˚ A) “W” laser structure. In the lower part, the fundamental electron 2 are reported. (e1 ) and heavy-hole (hh1 ) presence probability densities fe21 and fhh 1 The carrier wavefunction overlap is 57 %


47

500 450

W structure

400

T = 100 K

-1

Modal gain (cm )

350 300 250 200 150

T = 300 K

100 50 0 0,0

0,5

1,0

1,5 18

2,0

2,5

3,0

-3

N3D (x 10 cm )

Fig. 46. Modal gain Gmod against injected carrier concentration N3D calculated at T = 100 K and T = 300 K. The active region is composed of ten “W” periods 350

-1

MODAL GAIN (cm )

300

InAs/GaInSb W structure

250

T = 100 K

200 150

T = 300 K

100 50 0

10

100

1k

10k

100k 2

TOTAL CURRENT DENSITY (A/cm )

Fig. 47. Modal gain Gmod versus total current density J (A/cm2 ) calculated at T = 100 K and T = 300 K. The active region is composed of ten “W” periods

4 × 10−11 cm3 /s and the Auger coefficient C, which governs the current, was fixed at C = 5 × 10−29 cm6 /s and C = 5 × 10−27 cm6 /s, at T = 100 K and T = 300 K, respectively. The theoretical results confirm the experimental values reported in [29,30] with a threshold current density of 16 kA/cm2 at 300 K. The laser devices were processed into 100 µm wide stripes. In the pulsed regime (0.2–0.5 µs, 200 Hz), the device, with ten “W” periods, operated up to 310 K. At this temperature, the laser emission was observed at 3.27 µm, the output power was 370 µW and the threshold current density Jth was 25 kA/cm2 . On the contrary, in the continuous regime, the best results were

48


Fig. 48. CW output power versus injected current at 180 K and 195 K, after [30]. The CW spectrum at 195 K, with a minimum resolution of 1 nm and an injection current of 3.5 A, is shown in the inset

obtained with five “W” periods. At 80 K the measured threshold current density was 63 A/cm2 , and the CW output power was 140 mW. The CW lasing performances improve when the number of QWs is kept small to limit the threshold, while in the pulsed regime at higher temperature, the ten-period devices outperformed the five ones because significant gain is needed. CW measurements yielded a maximum operating temperature of 195 K (Fig. 48) with a threshold current density of 1.4 kA/cm2 and a characteristic temperature T0 = 38 K [30]. To date, this result is the highest CW Tmax operation temperature for III–V mid-infrared laser diodes (Fig. 9).

6

Conclusion

Different technologies are in competition for the achievement of mid-infrared 2–5 µm laser diodes operating CW at room temperature: conventional GaInAsSb/AlGaAsSb quantum well lasers, lead salt lasers, intersubband GaInAs/InAlAs quantum cascade lasers, type-II InAsSb/InAs multiquantum well lasers, type-III “W” InAs/GaInSb lasers and interband cascade lasers based on the same system. Impressive results have been obtained in the 2.0–2.3 µm wavelength range from GaInAsSb-AlGaAsSb type-I QW lasers: CW output powers of the order of 1 W/facet at 300 K, threshold current densities of 50–150 A/cm2 , characteristic temperatures T0 higher than 100 K, CW operation up to 140 ◦C, high internal efficiencies (up to 95 %) and extremely small internal loss coefficient (as low as ≈ 2 cm−1 ). Beyond 2.3 µm and up to 2.7 µm, the laser characteristics slightly degrade. Above 2.78 µm the laser diodes could not operate at


49

room temperature. Gas detection by TDLAS using GaInAsSb QW lasers operating around 2.3 µm was successfully demonstrated. The used devices are Fabry–Pérot ridge lasers which have single mode emission in a wide range of temperature and injection current. Lead salt lasers presented excellent laser characteristics beyond 3.5 µm: room or near room temperature operation in pulsed conditions and the highest CW Tmax (223 K) in the 3–5 µm spectral range. That is the result of their small non-radiative Auger coefficient (of several 10−28 cm6 /s) which is one order of magnitude smaller than that of III–V compounds. Unfortunately their CW output power remains too small, even at low temperature (of the order of mW), for numerous IR applications. Intersubband quantum cascade lasers employing Ga0.47 In0.53 As/ Al0.48 In0.52 As material lattice-matched to an InP substrate are becoming the dominant technology for IR devices emitting at more than 5 µm (they can now operate in continuous mode at room temperature [146]). For shorter wavelength emission, it is necessary to increase the conduction band offset between the two semiconductor materials. That was obtained from straincompensated GaInAs/AlInAs/InP structures which could emit pulsed at room temperature near 3.5 µm. Nevertheless, in CW conditions the maximum temperature of operation is very low (less than 100 K). Another system was proposed for the realisation of intersubband QC lasers operating in the 3–5 µm wavelength range: the InAs/AlSb system grown on a GaSb substrate [147]. The very high conduction band discontinuity (2.1 eV at Γ points) of this material system allows the design of QC devices at short wavelengths. First results are encouraging, and efforts are being made to improve the design of the QC structure to achieve laser emission at room temperature [148]. Interband heterostructures with type-II (or staggered type-II) and type-III (or broken-gap type-II) band alignment have been considered because they lead to substantial reduction of the Auger recombination rate and internal losses if intervalence resonances are avoided. Type-II interband InAsSb/InAs multi-quantum-well lasers are attractive because of their growth simplicity and excellent performance at 80 K. But they are limited to operation temperatures less than 220 K, and theoretical predictions show that it would be very difficult to increase Tmax with this system. Type-III “W” quantum well laser diodes are basically interesting because they combine strong overlap of electron and hole wavefunctions (and hence large optical matrix elements) with two-dimensional densities of states for both carriers (and hence high differential gain). Mid-infrared “W” lasers based on the InAs/GaInSb type-III system were grown by MBE. They showed the highest operating temperatures of III–V laser diodes in the 3–5 µm wavelength domain (pulsed Tmax = 310 K and CW Tmax = 195 K for an emission at λ = 3.3 µm). Gain modal calculations in this “W” laser structure show

50


that RT operation is feasible, but only with samples having small optical losses, which requires high crystal quality and optimized design. Interband quantum cascade lasers, taking advantage of the type-III band alignment of the InAs/GaInSb system to reuse electrons for sequential photon emissions, have been grown by MBE. Good results have been obtained in the pulsed regime (near room temperature operation and very high peak output power at 80 K: 4 W/facet at emission wavelengths of 3.8–3.9 µm). But like strained intersubband QC lasers the CW operation of IC lasers is difficult to obtain at high temperature (CW Tmax = 127 K at 3.7 µm). It appears finally that GaInAsSb/AlGaAsSb QW laser diodes based on GaSb substrates are the best and already well-established technology for emission wavelengths shorter than 2.7 µm. Beyond 2.7 µm the other technologies under consideration showed encouraging performance, but they must be optimized to achieve the goal of CW mid-infrared emission at room temperature.

References 1. H. A. Gebbie, W. R. Harding, C. Hilsum, A. W. Pryce, V. Roberts: Atmospheric transmission in the 1 to 14µ region, Proc. Soc. A 206, 87 (1951) 1 2. L. S. Rothman, R. R. Gamache, R. H. Tipping, C. P. Rinsland, M. A. H. Smith, D. C. Benner, V. M. Devi, J. M. Flaud, C. Camy-Peyret, A. Perrin, A. Goldman, S. T. Massie, L. R. Brown, R. A. Toth: The HITRAN molecular data base: editions of 1991 and 1992, J. Quantum Spectrosc. Radiat. Transfer 48, 469 (1992) 1 3. H. I. Schiff, G. I. Mackay, J. Bechara: The use of tunable diode laser absorption spectroscopy for atmospheric measurements, in Air Monitoring by Spectroscopy Techniques, M. W. Sigrist (Ed.) (Wiley, New York 1994) 2 4. D. Z. Garbuzov, H. Lee, V. Khalfin, R. Martinelli, J. C. Connolly, G. L. Belenky: 2.3–2.7 µm room temperature CW operation of InGaAsSb-AlGaAsSb broad waveguide SCH-QW diode lasers, IEEE Photon. Technol. Lett. 11, 794– 796 (1999) 2, 4, 32 5. I. Melngailis: Maser action in InAs diodes, Appl. Phys. Lett. 2, 176–178 (1963) 3 6. J. Phelan, A. R. Calawa, R. H. Rediker, R. J. Keyes, B. Lax: Infrared InSb laser diodes in high magnetic fields, Appl. Phys. Lett. 3, 143–145 (1963) 3 7. C. B. Guillaume, P. Lavallard: Laser effect in indium antimonide, Solid State Commun. 1, 148–151 (1963) 3 8. J. F. Butler, A. R. Calawa, R. J. Phelan, T. C. Harman, A. J. Strauss, R. H. Rediker: PbTe diode laser, Appl. Phys. Lett. 5, 75–77 (1964) 3 9. Y. Horikoshi: Semiconductor lasers with wavelengths > 2 µm, in Semiconductors and Semimetals 22, W. T. Tsang (Ed.) (Academic Press, London 1985) pp. 93–151 3 10. D. L. Partin, C. M. Trush: Wavelength coverage of lead-europium-selenidetelluride diode lasers, Appl. Phys. Lett. 45, 193–195 (1984) 3 11. H. Preir: Physics and applications of IV–VI compound semiconductor lasers, Semicond. Sci. Technol. 5, S12–S20 (1990) 3


51

12. C. Caneau, A. K. Srivastava, A. G. Dentai, J. L. Zyskind, M. A. Pollack: Room temperature GaInAsSb/AlGaAsSb DH injection lasers at 2.2 µm, Electron. Lett. 21, 815–817 (1985) 3 13. A. N. Baranov, T. N. Danilova, B. E. Dzhurtanov, A. N. Imenkov, S. G. Konnikov, A. M. Litvak, V. E. Umanskii, Yu. P. Yakovlev: CW lasing in GaInAsSb/GaSb buried channel laser (T = 20◦ , λ = 2.0 µm), Sov. Techn. Phys. Lett. 14, 727–730 (1988) 3 14. A. E. Bochkarev, L. M. Dolginov, A. E. Drakin, P. G. Eliseev, B. N. Sverdlov: Continuous-wave lasing at room temperature in InGaSbAs/GaAlAsSb injection heterostructures emitting in the spectral rang 2.2–2.4 µm, Sov. Techn. Phys. Lett. 14, 727–730 (1988) 3 15. A. N. Baranov, E. A. Grebenschikova, B. E. Dzurthanov, T. N. Danilova, A. N. Imenkov, Yu. P. Yakovlev: Long wavelength lasers based on GaInAsSb solid solutions near the immiscibility boundary (λ ≈ 2.5 µm; T = 300 K), Sov. Tech. Phys. Lett. 14, 798–800 (1988) 3 16. A. N. Baranov, C. Fouillant, P. Grunberg, J. L. Lazzari, S. Gaillard, A. Joullié: High temperature operation of GaInAsSb/AlGaAsSb double heterostructure lasers emitting near 2.1 µm, Appl. Phys. Lett. 65, 616–617 (1994) 3 17. H. Mani, A. Joullié, G. Boissier, E. Tourniè, F. Pitard, A. M. Joullié, C. Alibert: New III–V double heterojunction laser emitting near 3.2 µm, Electron. Lett. 24, 1542–1543 (1988) 3, 4 ˙ Yakovlev: 2.7–3.9 µm 18. A. N. Baranov, A. N. Imenkov, V. V. Sherstnev, Yu.P. InAsSb(P)/InAsSbP low threshold diode lasers, Appl. Phys. Lett. 64, 2480– 2482 (1994) 3 19. A. N. Baranov, B. E. Dzhurtanov, A. N. Imenkov, A. A. Rogachev, Yu. M. Shernyakov, Yu. P. Yakovlev: Generation of coherent radiation in a quantumwell structure with one heterojunction, Sov. Phys. Semicond. 20, 1385–1389 (1986) 3 20. K. D. Moiseev, M. P. Mikhailova, O. G. Ershov, Yu. P. Yakovlev: Longwavelength laser (3.26 µm) with a single isolated p-GaInAsSb/p-InAs type-II heterojunction in the active region, Tech. Phys. Lett. 21, 482–483 (1995) 3 21. G. Bauer, M. Kriechbaum, Z. Shi, M. Tacke: IV–VI quantum wells for infrared lasers, J. Nonlinear Opt. Phys. Mater. 4, 283–312 (1995) 4, 7 22. Z. Shi, M. Tacke, A. Lambrecht, H. B¨ ottner: Midinfrared lead salt multiquantum-well diode lasers with 282 K operation, Appl. Phys. Lett. 66, 2537– 2539 (1995) 4 23. M. Tacke: Lead salt lasers, Philos. Trans. R. Soc. London A 359, 547–566 (2001) 4 24. M. Zandian, J. M. Arias, R. Zucca, R. V. Gil, S. H. Shin: HgCdTe double heterostructure injection laser grown by molecular beam epitaxy, Appl. Phys. Lett. 59, 1022–1024 (1991) 4 25. J. M. Arias, M. Zandian, R. Zucca, J. Singh: HgCdTe infrared diode lasers grown by MBE, Semicond. Sci. Technol. 8, S255–S260 (1993) 4 26. E. P. O’Reilly, A. R. Adams: Band structure engineering in strained semiconductor lasers, IEEE J. Quantum Electron. 30, 366–379 (1994) 4 27. C. R. Ram-Mohan, J. R. Meyer: Multiband finite element modeling of wavefunction engineered electro-optical devices, J. Nonlinear Opt. Phys. Mater. 4, 191–243 (1995) 4

52


28. J. R. Meyer, C. A. Hoffman, F. J. Bartoli, L. R. Ram-Mohan: Type-II quantum well lasers for the mid-wavelength infrared, Appl. Phys. Lett. 67, 757–759 (1995) 4, 41, 46 29. H. Lee, L. J. Olafsen, R. J. Menna, W. W. Bewley, R. U. Martinelli, I. Vurgaftman, D. Z. Garbuzov, C. L. Felix, M. Maiorov, J. R. Meyer, J. C. Connolly, A. R. Stugg, G. H. Olsen: Room-temperature type-II quantum well diode laser with broadened waveguide emitting at λ = 3.30 µm, Electron. Lett. 35, 1743–1745 (1999) 4, 41, 46, 47 30. W. W. Bewley, H. Lee, I. Vurgaftman, R. J. Menna, C. L. Felix, R. U. Martinelli, D. W. Stokes, D. Z. Garbuzov, J. R. Meyer, M. Maiorov, J. C. Connolly, A. R. Stugg, G. H. Olsen: Continuous-wave operation of λ = 3.25 µm broadened-waveguide W quantum-well diode lasers up to T = 195 K, Appl. Phys. Lett. 76, 256–258 (2000) 4, 41, 46, 47, 48 31. B. I. Vurgaftman, C. L. Felix, W. W. Bewley, D. W. Stokes, R. E. Bartolo, J. R. Meyer: Mid-infrared “W” lasers, Philos. Trans. R. Soc. London A 359, 489–503 (2001) 4, 10 32. R. Q. Yang: Infrared lasers based on intersubband transitions in quantum wells, Superlattice Microstr. 17, 77–83 (1995) 4 33. R. Q. Yang, J. D. Bruno, J. L. Bradshaw, J. T. Pham, D. E. Wortman: Highpower interband cascade lasers with quantum efficiency > 450 %, Electron. Lett. 35, 1254–1255 (1999) 4 34. R. Q. Yang, J. D. Bruno, J. L. Bradshaw, J. T. Pham, D. E. Wortman: Interband cascade lasers: progress and challenges, Physica E 7, 69–75 (2000) 4 35. J. D. Bruno, J. L. Bradshaw, R. Q. Yang, J. T. Pham, D. E. Wortman: Lowthreshold interband cascade lasers with power efficiency exceeding 9 %, Appl. Phys. Lett. 76, 3167–3169 (2000) 4 36. J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, A. Y. Cho: Quantum cascade laser, Science 264, 553–556 (1994) 4 37. C. Gmachl, A. Tredicucci, F. Capasso, A. L. Hutchinson, D. L. Sivco, J. N. Baillargeon, A. Y. Cho: High power λ ≈ 8 µm quantum cascade lasers with near optimum performance, Apl. Phys. Lett. 72, 3130–3132 (1998) 4 38. J. Faist, F. Capasso, D. L. Sivco, A. L. Hutchinson, S.-N. G. Chu, A. Y. Cho: Short wavelength (λ ≈ 3.4 µm) quantum cascade laser based on straincompensated InGaAs/AlInAs, Appl. Phys. Lett. 72, 680–682 (1998) 4 39. F. Capasso, C. Gmachl, R. Paiella, A. Tredicucci, A. L. Hutchinson, D. L. Sivco, J. N. Baillargeon, A. Y. Cho, H. C. Liu: New frontiers in quantum cascade lasers and applications, IEEE J. Sel. Topics Quantum Electron. 6, 931–947 (2000) 4 40. F. Q. Liu, Y. Z. Zhang, Q. S. Zhang, D. Ding, Bo Xu, Z. G. Wang, D. S. Jiang, B. Q. Sun: High-performance strain-compensated InGaAs/InAlAs quantum cascade lasers, Semicond. Sci. Technol. 15, L44–L46 (2000) 4 41. A. Y. Cho, D. L. Sivco, H. M. Gmachl, A. Tredicucci, A. L. Hutchinson, S. N. G. Chu, F. Capasso: Quantum devices, MBE technology for the 21st century, J. Crystal Growth 227/228, 1–7 (2001) 4 42. B. Ishaug, W. Y. Hwang, J. Um, B. Guo, H. Lee, C. H. Lin: Continuous-wave operation of a 5.2 µm quantum-cascade laser up to 210 K, Appl. Phys. Lett. 79, 1745–1747 (2001) 4 43. H. K. Choi, S. J. Eglash: High power multiple quantum well GaInAsSb/AlGaAsSb diode lasers emitting at 2.1 µm with low threshold current density, Appl. Phys. Lett. 61, 1154–1156 (1992) 4, 14, 23


53

44. H. K. Choi, G. W. Turner, S. J. Eglash: High-power GaInAsSb-AlGaAsSb multiple quantum well diode lasers emitting at 1.9 µm, IEEE Photon. Technol. Lett. 6, 7–9 (1994) 4 45. H. Lee, P. K. York, R. J. Menna, R. U. Martinelli, D. Z. Garbuzov, S. Y. Narayan, J. C. Connolly: Room temperature 2.78 µm AlGaAsSb/InGaAsSb quantum well lasers, Appl. Phys. Lett. 66, 1942–1944 (1995) 4, 31 46. A. N. Baranov, N. Bertru, Y. Cuminal, G. Boissier, Y. Rouillard, J. C. Nicolas, P. Grech, A. Joullié, C. Alibert: Mid-infrared GaSb-InAs based multiple quantum well lasers, Proc. SPIE 3284, 247–257 (1998) 4, 14, 15, 16, 17, 18, 20, 28, 30 47. A. Joullié: New developments in mid-infrared Sb-based lasers, J. Phys. IV (Paris) 9, Pr2.79–Pr2.96 (1999) 4, 10 48. D. A. Yarekha, G. Glastre, A. Perona, Y. Rouillard, F. Genty, E. M. Skouri, G. Boissier, P. Grech, A. Joullié, C. Alibert, A. N. Baranov: High temperature GaInAsSb/GaAlAsSb quantum well singlemode continuous wave lasers emitting near 2.3 µm, Electron. Lett. 36, 537–539 (2000) 4, 22, 29, 33, 34 49. D. A. Yarekha, A. Vicet, A. Perona, G. Glastre, B. Fraisse, Y. Rouillard, E. M. Skouri, G. Boissier, P. Grech, A. Joullié, C. Alibert, A. N. Baranov: High efficiency GaInAsSb/GaSb type-II quantum well continuous wave lasers, Semicond. Sci. Technol. 15, 390–394 (2000) 4, 34 50. J. Bhan, A. Joullié, H. Mani, A. M. Joullié, C. Alibert: III–V heterostructures for laser emission in 2.55 wavelength region, Materials and Technologies for Optical Communications Proc. SPIE 866, 126–134 (1987) 5 51. Y. Tsou, A. Ichii, E. M. Garmire: Improving InAs double heterostructure lasers with better confinement, IEEE J. Quantum Electron. 28, 1261–1268 (1992) 7 52. G. P. Agrawal, N. K. Dutta: Long wavelength semiconductor lasers, Sanjit Mitra (Ed.) (Van Nostrand Reinhold, New York 1986) 9 53. A. Haug: Auger recombination in quantum well semiconductors: calculation with realistic energy bands, Semicond. Sci. Technol. 7, 1337–1340 (1992) 9 54. B. K. Ridley: Quantum Processes in Semiconductors (Clarendon, Oxford 1988) 9 55. E. Rosencher, B. Vinter: Optoélectronique (Masson, Paris 1998) pp. 265–271 9 56. H. Y. Fan: Semiconductors and Semimetals, R. K. Willardson, A C. Beer (Eds.) (Academic, New York 1967) Vol. 3, Chap. 9 10 57. J. R. Meyer, C. L. Felix, W. W. Bewley, I. Vurgaftman, E. H. Aifer, L. J. Olafsen, J. R. Lindle, C. A. Hoffman, M. J. Yang, B. R. Bennett, B. V. Shanabrook, H. Lee, C. H. Lin, S. S. Pei, R. H. Miles: Auger coefficients in type-II InAs/ Ga1−x Inx Sb quantum wells, Appl. Phys. Lett. 73, 2857–2859 (1998) 10, 41 58. A. Joullié, P. Christol, A. Wilk, P. Grech, M. El Gazouli, F. Chevrier, A. N. Baranov: Recent developments and prospect in mid-infrared (3-5m) Sb-based laser diodes, Quantum Electronics and Photonics 15, mid-IR Workshop, Glasgow (2001) 10 59. J. L. Pautrat: private communication, reported in: Optoélectronique, E. Rosencher, B. Vinter (Eds.) (Masson, Paris 1998) p. 234 10 60. C. H. Grein, P. M. Young, H. Ehrenreich: Theoretical performance of InAs/Inx Ga1−x Sb superlattice-based midwave infrared lasers, J. Appl. Phys. 76, 1940–1942 (1994) 11, 41

54


61. G. G. Zegrya, A. D. Andreev: Mechanism of suppression of Auger recombination proocesses in type-II heterostructures, Appl. Phys. Lett. 67, 2681–2683 (1995) 11 62. S. Forouhar, A. Ksendzov, A. Larsson, H. Temkin: InGaAs/InGaAsP/InP strained-layer quantum well lasers at ≈ 2 µm, Electron. Lett. 28, 945–947 (1992) 14 63. U. Martinelli, R. J. Menna, A. Triano, M. G. Harvey, G. H. Olsen: Temperature dependence of 2 µm strained-quantum-well InGaAs/InGaAsP/InP diode lasers, Electron. Lett. 30, 324–326 (1994) 14 64. M. Ochiai, H. Temkin, S. Forouhar, R. A. Logan: InGaAs-InGaAsP buried heterostructure lasers operating at 2.0 µm, IEEE Photonics Techn. Lett. 7, 825–827 (1995) 14 65. M. Oishi, M. Yamamoto, K. Kasaya: 2.0-µm single-mode operation of InGaAsInGaAsP distributed-feedback buried-heterostructure quantum-well lasers, IEEE Photonics Techn. Lett. 9, 431–433 (1997) 14 66. M. Mitsuhara, M. Ogasawara, M. Oishi, H. Sugiura: Metalorganic molecular beam epitaxy-grown In0.77 Ga0.23 As/InGaAs multiple quantum well lasers emitting at 2.07 µm wavelength, Appl. Phys. Lett. 72, 3106–3108 (1998) 14 67. M. Mitsuhara, M. Ogasawara, M. Oishi, H. Sugiura, K. Kasaya: 2.05-µm wavelength InGaAs-InGaAs distrubuted-feedback multiquantum-well lasers with 10 mW output power, IEEE Photonics Techn. Lett. 11, 33–35 (1999) 14 68. J. Dong, A. Ubukata, K. Matsumoto: Characteristics dependence on confinement structure and single-mode operation in 2-µm compressively strained InGaAs-InGaAsP quantum-well lasers, IEEE Photonics Techn. Lett. 10, 513– 515 (1998) 14 69. J. S. Major, J. S. Osinski, D. F. Welch: High-power 2.0-µm InGaAsP laser diodes, IEEE Photonics Techn. Lett. 5, 594–596 (1993) 14 70. E. Tournié, P. Grunberg, C. Fouillant, A. N. Baranov, A. Joullié, K. H. Ploog: Strained InAs/Ga0.47 /In0.53 As quantum-well heterostructures grown by molecular-beam epitaxy for long-wavelength applications, Solid-State Electron. 37, 1311–1314 (1994) 14 71. J.-S. Wang, H.-H. Lin, L.-W. Sung: Room-temperature 2.2 µm InAs-InGaAsInP highly strained multiquantum-well lasers grown by gas-source molecular beam epitaxy, IEEE J. Quantum Electron. 34, 1959–1962 (1998) 14, 28 72. A. N. Baranov, Y. Cuminal, N. Bertru, C. Alibert, A. Joullié: Strained multiple quantum well lasers grown on GaSb emitting between 2 and 2.4 µm: Proc. SPIE 2997, 2–13 (1997) 14, 15, 18, 20, 28 73. C. G. Van de Walle: Band lineups and deformation potentials in the modelsolid theory, Phys. Rev. B 39, 1871–1879 (1989) 14 74. S. L. Wong, R. T. Warburton, R. J. Nicholas, N. J. Mason, P. J. Walker: A magneto-optical study of coupled quantum wells in strained GaInSb/GaSb, Physica B 184, 106–110 (1993) 14 75. L. Q. Qian, B. W. Wessels: Optical properties of strained-layer Inx Ga1−x Sb/GaSb heterostructures with x ≤ 0.4, J. Vac. Sci. Technol. B 11, 1652–1659 (1993) 14 76. A. N. Baranov, Y. Cuminal, G. Boissier, J. C. Nicolas, J. L. Lazzari, C. Alibert, A. Joullié: Electroluminescence of GaInSb/GaSb strained single quantum well structures grown by molecular beam epitaxy, Semicond. Sci. Technol. 11, 1185–1191 (1996) 14


55

77. A. Haug: Band-to-band Auger recombination in semiconductors, J. Phys. Chem. Solids 49, 599–605 (1988) 16 78. N. A. Gun’ko, G. G. Zegrya, N. V. Zotova, Z. N. Sokolova, N. M. Stus’, V. B. Khalfin: Influence of valence band absorption on the threshold characteristics of long-wavelength InAs lasers, Semicond. 31, 1204–1208 (1997) 16 79. P. Christol, P. Bigenwald, A. Joullié, Y. Cuminal, A. N. Baranov, N. Bertru, Y. Rouillard: Improvement of Sb-based multiquantum well lasers by Coulomb enhancement, IEEE Proc. Optoelectron. 146, 3–8 (1999) 17, 42 80. G. G. Zegrya, A. D. Andreev: Mechanism of suppression of Auger recombination processes in type-II heterostructures, Appl. Phys. Lett. 67, 2681–2683 (1995) 18 81. N. Bertru, A. N. Baranov, Y. Cuminal, G. Almuneau, F. Genty, A. Joullié, O. Brandt, A. Mazuelas, K. H. Ploog: Long-wavelength (Ga,In)Sb/GaSb strained quantum well lasers grown by molecular beam epitaxy, Semicond. Sci. Technol. 13, 936–939 (1998) 18 82. A. N. Baranov, Y. Cuminal, G. Boissier, C. Alibert, A. Joullié: Low-threshold laser diodes based on type-II GaInAsSb/GaSb quantum-wells operating at 2.36 µm at room temperature, Electron. Lett. 32, 2279–2280 (1996) 18, 20, 28 83. G. A. Sai-Halasz, R. Tsu, L. Esaki: A new semiconductor superlattice, Appl. Phys. Lett. 30, 651–653 (1977) 18 84. A. N. Baranov, N. Bertru, Y. Cuminal, G. Boissier, C. Alibert, A. Joullié: Observation of room-temperature laser emission from type-III InAs/GaSb multiple quantum well structures, Appl. Phys. Lett. 71, 735–736 (1997) 18, 19 85. P. S. Kop’ev, N. N. Ledentsov, A. M. Monakhov, A. A. Rogachev: Noninjection light-emitting diode, RU patent no. 2019895 (1991) 18 86. S. V. Ivanov, B. K. Kurinkiev, N. N. Ledentsov, B.Ya. Meltser, A. A. Monakhov, A. A. Rogachev, S. V. Shaposhnikov, P. S. Kop’ev: Middle infrared (λ = 2–5.5 µm) high quantum efficiency luminescence in GaSb/InAs II-type multi-quantum well structures, in Workbook of 6th International Conference on Modulated Semiconductor Structures, Garmisch-Partenkirchen, (1993) p. 1047 18 87. P. S. Kop’ev, B. Ya. Meltser, S. V. Ivanov, A. A. Rogachev, S. V. Shaposhnikov: Photo- and electroluminescence study of type-II InAs/GaSb quantum well heterostructures grown by molecular beam epitaxy, in Proceedings of the VIII International Conference on Molecular Beam Epitaxy, Osaka (1994) p. 542 18 88. J. Schmitz, J. Wagner, F. Fuchs, N. Herres, P. Koidl, J. D. Ralston: Optical and structural investigations of intermixing reactions at the interfaces of InAs/AlSb and InAs/GaSb quantum wells grown by molecular-beam epitaxy, J. Crystal Growth 150, 858–863 (1995) 18 89. A. Mujica, A. Mu˜ nioz: Self consitent tight binding calculations of band offsets in GaAs/Alx Ga1−x As (110) and GaSb/Alx Ga1−x Sb (110) heterojunctions. Theoretical evidence for a new common-anion rule, Solid State Commun. 81, 961–968 (1992) 20 90. J.-L. Lazzari: Etude des hétérostructures GaInAsSb/GaSb et GaAlAsSb/ GaSb et de composants lasers et détecteurs ` a base de GaSb opérant entre 2 et 2.5 µm, Dissertation, Université Montpellier II, Montpellier (1993) 20

56


91. D. Z. Garbuzov, R. U. Martinelly, H. Lee, P. K. York, R. J. Menna, J. C. Connolly, S. Y. Narayan: Ultralow-loss broadened-waveguide high-power 2 µm AlGaAsSb/InGaAsSb/GaSb separate-confinement quantum-well lasers, Appl. Phys. Lett. 69, 2006–2008 (1996) 20, 21, 23, 24, 25 92. N. Bertru, A. N. Baranov, Y. Cuminal, G. Almuneau, F. Genty, A. Joullié, O. Brandt, A. Mazuelas, K. H. Ploog: Long-wavelength (Ga,In)Sb/GaSb strained quantum well lasers grown by molecular beam epitaxy, Semicond. Sci. Technol. 13, 936–940 (1998) 20, 28 93. Y. Cuminal, A. N. Baranov, D. Bec, P. Grech, M. Garcia, G. Boissier, A. Joullié, G. Glastre, R. Blondeau: Room-temperature 2.63 µm GaInAsSb/GaSb strained quantum-well laser diodes, Semicond. Sci. Technol. 14, 283–288 (1999) 20, 28, 30 94. A. Joullié, G. Glastre, R. Blondeau, J.-C. Nicolas, Y. Cuminal, A. N. Baranov, A. Wilk, M. Garcia, P. Grech, C. Alibert: Continuous-wave operation of GaInAsSb-GaSb type-II quantum-well ridge lasers, IEEE J. of Select. Topics in Quant. Electron. 5, 711–715 (1999) 20, 28 95. D. A. Yarekha, A. Vicet, A. Perona, G. Glastre, B. Fraisse, Y.Rouillard, E. M. Skouri, G. Boissier, P. Grech, A. Joullié, C. Alibert, A. N. Baranov: High efficiency GaInAsSb/GaSb type-II quantum well continuous wave lasers, Semicond. Sci. Technol. 15, 283–288 (2000) 20, 22, 28, 29, 30 96. K. Onabe: Unstable region in quaternary In1−x Gax As1−y Sby calculated using strictly regular solution approximation, Jpn. J. Appl. Phys. 21, 964 (1982) 21, 22 97. G. B. Stringfellow: Thermodynamics of semiconductor alloys for optoelectronics, in Materials for Optoelectronics, M. Quillec (Ed.) (Kluwer Academic, Dordrecht 1996) pp. 23–61 G. B. Stringfellow: Immiscibility and spinodal decomposition in III/V alloys, J. Cryst. Growth 65, 454–462 (1983) 21 98. A. N. Baranov, S. G. Konnikov, T. B. Popova, V. E. Umanskii, Yu. P. Yakovlev: Liquid phase epitaxy of GaAlAsSb/GaSb and the effect of strain on phase equilibria, J. Cryst. Growth 66, 547–552 (1984); 21 99. H. K. Choi, S. J. Eglash, M. K. Connors: Single-frequency GaInAsSb/ AlGaAsSb quantum-well ridge-waveguide lasers emitting at 2.1 µm, Appl. Phys. Lett. 63, 3271–3272 (1993) 22, 23, 26 100. D. Garbuzov, R. U. Martinelli, H. Lee, R. J. Menna, P. K. York, L. A. DiMarko, M. G. Harvey, R. J. Matarese, S. Y. Narayan, J. C. Connolly: 4-W quasi-continuous-wave output power from 2 µm AlGaAsSb/InGaAsSb singlequantum-well broadened waveguide laser diodes, Appl. Phys. Lett. 70, 2931– 2933 (1997) 24, 25 101. R. J. Menna, D. Z. Garbuzov, H. Lee, R. U. Martinelli, S. Y. Narayan, J. C. Connolly: High power broadened-waveguide InGaAsSb/AlGaAsSb quantumwell diode lasers emitting at 2 µm, Proc. SPIE 3284, 238–246 (1998) 24, 25 102. W. Turner, H. K. Choi, M. J. Manfra: Ultralow-threshold (50 A/cm2 ) strained single-quantum-well GaInAsSb/AlGaAsSb lasers emitting at 2.05 µm, Appl. Phys. Lett. 72, 876–878 (1998) 24, 25 103. H. K. Choi, G. W. Turner, J. N. Walpole, M. J. Manfra, M. K. Connors, L. J. Missaggia: Low-threshold, high power, high brightness GaInAsSb/AlGaAsSb quantum-well lasers emitting at 2.05 µm, Proc. SPIE 3284, 268–273 (1998) 24, 25


57

104. J. N. Walpole, H. K. Choi, L. J. Missagia, Z. L. Liau, M. K. Connors, G. W. Turner, M. J. Manfra, C. C. Cook: High-power high-brightness GaInAsSbAlGaAsSb tapered laser arrays with anamorphic collimating lenses emitting at 2.05 µm, IEEE Photonics Techn. Lett. 11, 1223–1225 (1999) 24 105. T. C. Newell, L. F. Lester, X. Wu, Y. Zhang, A. L. Gray: Gain and threshold current density characteristics of 2 micron GaInAsSb/AlGaAsSb MQW lasers with increased valence band offset, Proc. SPIE 3284, 258–267 (1998) 25 106. T. C. Newell, X. Wu, A. L. Gray, S. Dorato, H. Lee, L. F. Lester: The effect of increased valence band offset on the operation of 2 µm GaInAsSb-AlGaAsSb lasers, IEEE Photonics Techn. Lett. 11, 30–32 (1999) 25 107. G. Mermelstein, S. Simanowski, M. Mayer, R. Kiefer, J. Schmitz, M. Walther, J. Wagner: Room-temperature low-threshold low-loss continuous-wave operation of 2.26 µm GaInAsSb/AlGaAsSb quantum-well laser diodes, Appl. Phys. Lett. 77, 1581–1583 (2000) 26 108. S. Simanowski, N. Herres, C. Mermelstein, R. Kiefer, J. Schmitz, M. Walther, J. Wagner, G. Weimann: Strain adjustment in (GaIn)(AsSb)/(AlGa)(AsSb) QWs for 2.3–2.7 µm laser structures, J. Cryst. Growth 209, 15–20 (2000) 26 109. S. Simanowski, C. Mermelstein, M. Walther, N. Herres, R. Kiefer, M. Rattunde, J. Schmitz, J. Wagner, G. Weimann: Growth and layer structure optimization of 2.26 µm (AlGaIn)(AsSb) diode lasers for room temperature operation, J. Cryst. Growth 227–228, 595–599 (2001) 26 110. T. Bleubel, M. Muller, A. Forchel: 2–µm GaInSb-AlGaAsSb distributedfeedback lasers, IEEE Photonics Techn. Lett. 13, 553–555 (2001) 26, 27 111. A. N. Baranov, Y. Rouillard, G. Boissier, P. Grech, S. Gaillard, C. Alibert: Sbbased monolithic vertical cavity surface emitting laser operating near 2.2 µm at room temperature, Electron. Lett. 34, 281–282 (1998) 27 112. D. Yarekha: Etude et réalisation de diodes lasers émettant entre 2.2 et 2.4 µm pour application ` a l’analyse de gaz, Dissertation, Université Montpellier II, Montpellier (2001) 29, 30, 31 113. A. Baraldi, C. Ghezzi, A. Parisini, R. Magnanini, L. Tarricone, S. Franchi: Occupancy level of the DX center in Te-doped Alx Ga1−x Sb, J. Appl. Phys. 85, 256–263 (1999) 30 114. D. Z. Garbuzov, R. U. Martinelli, R. J. Menna, P. K. York, H. Lee, S. Y. Narayan, J. C. Connolly: 2.7 µm InGaAsSb/AlGaAsSb laser diodes with continuous-wave operation up to −39 ◦ C, Appl. Phys. Lett. 67, 1346–1348 (1995) 31 115. D. Garbuzov, R. Menna, M. Maiorov, H. Lee, V. Khalfin, L. DiMarco, D. Capewell, R. Martinelli, G. Belenky, J. Connolly: 2.3–2.7 µm room temperature CW-operation of InGaAsSb/AlGaAsSb broad-contact and single-mode ridge-waveguide SCH-QW diode lasers, Proc. SPIE 3628, 124–129 (1999) 31 116. A. D. Andreev, D. V. Donetsky: Analysis of temperature dependence of the threshold current in 2.3–2.6 µm InGaAsSb/AlGaAsSb quantum-well lasers, Appl. Phys. Lett. 74, 2743–2745 (1999) 31 117. D. Garbuzov, M. Maiorov, H. Lee, V. Khalfin, R. Martinelli, J. Connolly: Temperature dependence of continuous wave threshold current for 2.3–2.6 µm InGaAsSb/AlGaAsSb separate confinement heterostructure quantum well semiconductor diode lasers, Appl. Phys. Lett. 74, 2990–2992 (1999) 31 118. M. Maiorov, J. Wang, D. Baer, H. Lee, G. Belenky, R. Hanson, J. Connolly, D. Garbuzov: New room temperature CW InGaAsSb/AlGaAsSb QW ridge

58

119.

120.

121.

122.

123.

124. 125.

126.

127.

128.

129. 130.

131.

132.

133.

134.

André Joullié et al. diode lasers and their application to CO measurements near 2.3 µm, Proc. SPIE 3855, 62–71 (1999) 31 L. S. Rothman, R. R. Gamache, A. Goldman, L. R. Brown, R. A. Toth, H. M. Pickett, R. L. Poynter, J. M. Flaud, C. Camy-Peyret, A. Barbe, N. Husson, C. P. Rinsland, A. H. Smith, The HITRAN database: 1986 edition, Appl. Opt. 26, 4058–4097 (1987) 33 A. Vicet, J. C. Nicolas, F. Genty, E. M. Skouri A. N. Baranov, C. Alibert: Room temperature GaInAsSb/GaSb quantum well laser for tunable diode laser absorption spectroscopy around 2.35 µm, IEE Proc. Optoelectron. 147, 172–176 (2000) 33 R. Mucke, P. Werle, F. D’Amato, A. Lancia: Precise measurement of CO2 concentration using room temperature diode lasers, 5th International Symposium on Gas Analysis by Tunable Diode Lasers, (1998) pp. 83–92 34 A. Vicet, D. A. Yarekha, A. Perona, Y. Rouillard, S. Gaillard, A. N. Baranov: Trace gas detection with antimonide-based quantum-well diode lasers, Spectrochim. Acta Part A 58 2405–2412 (2002) 35, 40 B. Botteldooren, R. Baets: Influence of band-gap shrinkage on the carrierinduced refractive index change in InGaAsP, Appl. Phys. Lett. 54, 1989–1991 (1989) 37 A. N. Baranov, V. Sherstnev, C. Alibert, A. Krier, New III–V semiconductor lasers emitting near 2.6 µm, J. Appl. Phys. 79, 3354–3356 (1996) 37 B. K. Ghosh, L. K. Samanta, G. C. Bhar: The temperature and pressure dependence of refractive index of some III–V semiconductors, Infrared Phys. 26, 111 (1986) 38 J. B. McManus, P. L. Kebabian, M. S. Zahniser: Astigmatic mirror multipass absorption for long-pathlength spectroscopy, Appl. Opt. 34, 3336–3348 (1995) 38 A. Popov, V. Sherstnev, Y. Yakovlev, R. M¨ ucke, P. Werle: High power InAsSb/InAsSbP double heterostructure laser for continuous wave operation at 3.6 µm, Appl. Phys. Lett. 68, 2790–2792 (1996) 40 H. K. Choi, S. J. Eglash, G. W. Turner: Double-heterostructure diode laser emitting at 3 µm with a metastable GaInAsSb active layer and AlGaAsSb cladding layers, Appl. Phys. Lett. 64, 2474–2476 (1994) 40 Y. H. Zhang: Continuous wave operation of InAs/InAsx Sb1−x mid-infrared lasers, Appl. Phys. Lett. 66, 118–120 (1995) 40 S. R. Kurtz, R. M. Biefeld, A. A. Allerman, A. J. Howard, M. H. Crawford: Pseudomorphic InAsSb multiquantum well injection laser emitting at 3.5 µm, Appl. Phys. Lett. 68, 1332–1334 (1996) 40 B. Lane, D. Wu, A. Rybaltowski, J. Diaz, M. Razeghi: Compressively strained multiple quantum well InAsSb lasers emitting at 3.6 µm grown by metalorganic chemical vapor deposition, Appl. Phys. Lett. 70, 443–445 (1997) 40 G. W. Turner, M. J. Manfra, H. K. Choi, M. K. Connors: MBE growth of high power InAsSb/InAlAsSb quantum well diode lasers emitting at 3.5 µm, J. Crystal Growth 175/176, 825–832 (1997) 40 H. K. Choi, G. W. Turner, M. J. Manfra, M. K. Connors: 175 K continuous wave operation of InAsSb/InAlAsSb quantum well diode lasers emitting at 3.5 µm, Appl. Phys. Lett. 68, 2936–2938 (1996) 40 R. M. Biefeld, S. R. Kurtz, A. A. Allerman: The metal-organic chemical vapor deposition growth and properties of InAsSb mid-infrared (3–6 µm) lasers and LEDs, IEEE J. Select. Topics Quantum Electron. 3, 739–748 (1997) 40


59

135. B. Lane, Z. Wu, A. Stein, J. Diaz, M. Razeghi: InAsSb/InAsP strained-layer superlattice injection lasers operating at 4.0 µm grown by metal-organic chemical vapor deposition, Appl. Phys. Lett. 74, 3438–3440 (1999) 40 136. T. Ashley, C. T. Elliott, R. Jefferies, A. D. Johnson, G. J. Pryce, A. M. White, M. Carroll: Mid-infrared In1−x Alx Sb/InSb heterostructure diode lasers, Appl. Phys. Lett. 70, 931–933 (1997) 40 137. W. W. Bewley, I. Vurgaftman, C. L. Felix, J. R. Meyer, C. H. Lin, D. Zhang, S. J. Murry, S. S. Pei, L. R. Ram-Mohan: Role of internal loss in limiting type-II mid-IR laser performance, J. Appl. Phys. 83, 2384–2391 (1998) 41 138. Y. B. Li, D. J. Bain, L. Hart, M. Livingstone, C. M. Ciesla, M. J. Pullin, P. J. Tang, W. T. Yuen, I. Galbraith, C. C. Phillips, C. R. Pidgeon, R. A. Stradling: Band alignment in In(As,Sb)/InAs superlattices, Phys. Rev. B 55, 4589–4595 (1997) 41 139. P. Christol, P. Bigenwald, A. Wilk, A. Joullié, O. Gilard, H. Carrère, F. LozesDupuy, A. Behres, A. Stein, J. Kluth, K. Heime, E. M. Skouri: InAs/InAs(PSb) quantum-well laser structure for the midwavelength infrared region, IEE Proc. Optoelectron. 147, 181–187 (2000) 41, 42 140. D. Ahn, S. L. Chuang: Optical gain and gain suppression of quantum-well lasers with valence band mixing, IEEE J. Quantum. Electron. 28, 13–24 (1990) 42 141. P. Christol, M. El Gazouli, P. Bigenwald, A. Joullié: Performance simulation of 3.3 µm interband laser diodes grown on InAs substrate, Physica E 14, 375–384 (2002) 42, 46 142. A. Wilk, F. Genty, B. Fraisse, G. Boissier, P. Grech, M. El Gazouli, A. Joullié, P. Christol, J. Oswald, T. Simecek, E. Hulicius: MBE growth of InAs/InAsSb/AlAsSb laser structures for mid-infrared lasers, J. Cryst. Growth 223, 341–348 (2001) 43 143. A. Wilk, M. El Gazouli, M. El Skouri, P. Christol, P. Grech, A. N. Baranov, A. Joullié: Type-II InAsSb/InAs strained quantum-well laser diodes emitting at 3.5 µm, Appl. Phys. Lett. 77, 2298–2300 (2000) 44, 45 144. A. Wilk, B. Fraisse, P. Christol, G. Boissier, P. Grech, M. El Gazouli, Y. Rouillard, A. N. Baranov, A. Joullié: MBE growth of InAs/InAsSb/InAlAsSb “W” quantum well laser diodes emitting near 3 µm, J. Crystal Growth 227/228, 586–590 (2001) 44 145. A. Joullié, E. M. Skouri, M. Garcia, P. Grech, A. Wilk, P. Christol, A. N. Baranov, A. Behres, J. Kluth, A. Stein, K. Heime, M. Heuken, S. Rushworth, E. Hulicius, T. Simecek: InAs(PSb)-based “W” quantum well laser diodes emitting near 3.3 µm, Appl. Phys. Lett. 76, 2499–2501 (2000) 44 146. M. Beck, D. Hofstetter, T. Aellen, S. Blaser, J. Faist, U. Oesterle, E. Gini: Continuous-wave operation of quantum cascade lasers, J. Crystal Growth 251, 697–700 (2003) 49 147. C. Becker, I. Prevot, X. Marcadet, B. Vinter, C. Sirtori: InAs/AlSb quantumcascade light-emitting devices in the 3–5 µm wavelength region, Appl. Phys. Lett. 76, 1029–1031 (2001) 49 148. C. Becker, H. Page, M. Calligaro, V. Ortiz, M. Garcia, X. Marcadet, C. Sirtori: GaAs- and GaSb-based quantum cascade lasers: the challenge of the new materials, 5th Int. Conf. Mid-infrared Optoelectronic Materials and Devices (MIOMD-5), Annapolis, MD (2002) 49

Index

Auger recombination, 4, 9–11, 13, 16, 17, 41, 47 band alignment, 6–8 – type-I, 6, 7, 14–18 – type-II, 6, 14–18, 28, 41 – type-III, 6, 18, 19, 45 band offset, 7, 8, 14, 15, 20, 24, 26, 28, 32, 33 characteristic temperature, 2, 19, 23–26, 30–32, 37, 44, 48 confinement factor, 5, 9, 23, 42, 43, 46 critical thickness, 17

internal quantum efficiency, 24, 28, 30 intersubband cascade laser, 4, 13 IV–VI semiconductor, 3, 5, 7 laser – double heterostructure, 3–5, 14, 22, 40 – interband cascade, 3, 4 – intersubband cascade, 4, 13 – lead salt, 2–4, 11, 13 – quantum cascade (QCL), 3 lead salt laser, 2–4, 11, 13 lifetime, 33, 40 miscibility gap, 21, 22, 31, 33, 44

differential quantum efficiency, 23–25, 30, 31 double heterostructure (DH) laser, 3–5, 14, 22, 40 Fabry–Pérot interferometer, 34 Fabry–Pérot laser, 5, 9, 18, 22, 26, 30, 31, 43 free carrier absorption, 9, 10, 23, 31 – loss, 23, 46

non-radiative (SHR) coefficient, 9, 46 optical gain, 9, 42 – maximum gain, 9, 42 – modal gain, 9, 42, 43, 46 optical loss, 9, 42, 44–46 – internal loss, 9, 23, 24, 28, 30, 41, 44 – mirror loss, 9 quantum cascade laser (QCL), 3

GaInAsSb/AlGaAsSb, 2–4, 11, 13–40 GaInAsSb/GaSb, 2, 7, 11, 15, 16, 20, 29

radiative (spontaneous) coefficient, 9, 46 radiative recombination, 6, 9, 17, 19

HgCdTe, 2, 4, 11, 13 series resistance, 5, 20, 37 InAs/AlSb, 5, 7, 49 InAs/GaInSb, 4, 7, 11, 41, 45–48 InAs/GaSb, 7, 18, 19 InAsSb/InAs, 7, 40, 41 InAsSb/InAsP, 40, 44 InAsSb/InAsPSb, 3, 40 interband cascade laser, 3, 4, 12, 13

thermal resistance, 30, 37 tunable diode laser absorption spectroscopy (TDLAS), 2, 34, 38 type-I laser, 2, 11, 14, 17, 18, 20, 23–29, 34, 40–42 type-II laser

Index – “W” laser, 1, 3, 4, 11, 13, 41, 46–47 – interband laser, 1, 2, 4, 41 – intersubband laser, 1, 49 – MQW laser, 4, 13, 28–33, 41–44 type-III laser, 18, 45–48

vertical cavity surface emitting laser (VCSEL), 26, 27

waveguide layer, 6, 10, 42, 46

61

High Performance Quantum Cascade Lasers and Their Applications Daniel Hofstetter and Jérôme Faist Institute of Physics, University of Neuchˆ atel, 2000 Neuchˆ atel, Switzerland [email protected] [email protected] Abstract. This chapter describes our results on distributed feedback quantum cascade lasers in the wavelength range around 5 µm and around 10 µm. We present two different gain region designs; one with three quantum wells and one with a double phonon resonance. Several fabrication techniques are also presented and analysed in terms of fabrication simplicity, performance, yield, and reliability. We will outline typical results for all devices and also show some interesting applications. In light of this, the chapter is organized as follows: We start with a brief introduction; in Sect. 2, the advantages and drawbacks of the different gain regions are outlined; Sect. 3 deals with the fabrication technology which was required to build these lasers; in Sect. 4, we present the measurement results on the devices; and finally, Sect. 5 describes two examples of interesting applications in the fields of optical spectroscopy and optical data transmission. The chapter ends with a brief conclusion and an outlook.

1

Introduction

The development of high-performance mid-infrared light sources has experienced tremendous progress during the last couple of years. Pacemakers of this progress were the appearance and the subsequent improvements of the quantum cascade (QC) laser [1,2,3]. After the pioneering work of Kazarinov and Suris about transport and gain in superlattice structures in 1971 [4], it took 23 years to realize the first working QC laser. During this long period, researchers investigated many different aspects of intersubband transitions [5,6] and the nature of resonant tunneling effects in thin semiconductors [7,8]. They also learned to control the growth of extremely thin semiconductor layers by methods like molecular beam epitaxy (MBE) [9,10,11] and metalorganic vapor phase epitaxy (MOVPE) [12,13]. All these important developments were crucial milestones towards the realization of a midinfrared semiconductor laser based on intersubband transitions. But this kind of preparative work done in the 1970s and the 1980s also paved the way for a rapid development of the QC laser once it was born. Only one year after the demonstration of the QC laser by Faist et al. at Bell Labs in 1994 [1], the first continuous wave (CW) QC laser was operating at cryogenic temperatures [14,15]; and in 1996, room temperature pulsed operation could be I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 61–98 (2003) c Springer-Verlag Berlin Heidelberg 2003

62

Daniel Hofstetter and Jérˆ ome Faist

achieved [2]. From 1997 to 1999, some progress was made to push the highest CW operating temperatures towards higher values, but the relatively large threshold current densities of these devices (5 kA/cm2 at 300 K) prevented a real breakthrough [16,17,18,19]. The magic value of 200 K, which would have allowed operation on a thermo-electric cooler, seemed not to be as easy to reach as people had hoped initially. In 1999, for instance, the record CW operating temperature was 175 K [20]. As the performance of Fabry–Pérottype lasers improved steadily, the gas sensing applications soon called for single-mode lasers based on the distributed feedback (DFB) principle. The performance of these DFB QC lasers went pretty much in parallel with the one of the FP-lasers. Different coupling schemes like index-coupling [21,18] and gain/loss-coupling [22] were tried in order to achieve stable single-mode operation over a wide temperature range. Continuous temperature tuning over more than 1% of the emission wavelength could be demonstrated. Later on, the effort concentrated rather on novel material systems like AlGaAs or AlGaN [23,24,25,26], and on the implementation of the QC laser in systems like gas sensors and telecommunication [27,28,29,30]. During the year 2000, substantial progress was made again on the performance of the lasers themselves, mainly in terms of higher pulsed operating temperatures [31] and high duty cycle operation [32]. This development found its culmination in summer 2001, when the first room temperature CW operated QC laser was published [33]. Rapid progress was also made on GaAs/AlGaAs-based midIR QC lasers [34,35,36], where room temperature pulsed operation has been possible since 2001. Very recently, laser action on far-IR QC lasers at wavelengths close to 70 µm was demonstrated [37,38]. All these examples show that there is still a great deal of development potential behind these devices. For several potential applications, especially in the area of optical sensors for atmospheric trace gases, it is advantageous to operate with single-mode, single frequency lasers. For this purpose, DFB QC lasers have found widespread use and are also well understood [21,22,39]. Although DFB lasers have obvious performance benefits, there are some severe fabrication drawbacks. One of them is the requirement of epitaxial regrowth, which makes fabrication rather complicated and prolonged. This is not only due to the regrowth process itself, but also because of the fact that the material can be tested only after grating fabrication and over-growth. In Sect. 3.1, we will describe a simple method to circumvent some of the typical DFB laser fabrication problems. For certain special applications, the availability of a surface-emitting laser would greatly facilitate the coupling of the light into the optical system. Since, for fundamental reasons (TM polarization), a QC laser cannot be configured as a vertical cavity surface-emitting laser, other methods to achieve surface emission have to be pursued. An additional desired property of a light source for sensors is its single-mode behavior even in pulsed operation and at high power levels. In order to achieve single-mode operation, DFB QC lasers have

High Performance Quantum Cascade Lasers and Their Applications

63

already been extensively investigated and characterized [18,21,22,39]. An elegant method to obtain surface emission is the fabrication of a DFB QC laser utilizing a second-order diffraction grating [40]. This has the additional advantage that the grating with a period of around 3.15 µm can be fabricated by moderate-resolution contact-lithography instead of holography. This contactlithography technique was successfully utilized by the group of the technical University of Vienna [36]. Accordingly, such a device, again with lateral current injection, will be presented in this work. For most potential applications, however, properties like high output power, low power consumption, room-temperature operation, and a large single-mode tuning range are key features. Since continuous-wave operation of QC DFB lasers is not yet possible at room temperature, one has to deal with high average powers under pulsed current injection [18,41,42,43,44,45]. So far, high output powers were mainly a privilege of superlattice QC lasers. However, their major advantage, namely the short lifetime of the lower laser level, was usually obtained at the cost of a slightly worse injection efficiency. We have chosen a combination of two different measures to reach high output powers: first, a new type of gain region was investigated and implemented; second, we mounted our devices junction down to increase the thermal conductance. We will therefore report on QC lasers with a 4 quantum well (QW) gain region which combines the good injection efficiency of the diagonal anticrossed 3 QW design [46] and the short ground state lifetime of the superlattice QC laser [47,48]. But we will also give a brief overview of the performance of junction down mounted devices. Since QC lasers have their strongest output powers in the absorption bands of many gases, it was realized very soon that they are potential candidates for spectroscopic and other applications. With the recent progress in photo-acoustic (PA) spectroscopy such as the advent of highly sensitive microphones and microphone arrays, the construction of resonant multipass PA-cells, or the deployment of Helmholtz-resonator shaped PA-cells, QC lasers are, despite their still limited output power, becoming very interesting light sources for this sensitive technique. Nevertheless, the average QC laser output power was usually not large enough to perform spectroscopy in the ppm or even the ppb range. It is thus necessary to achieve high average power operation under temperature conditions which can be maintained with thermo-electric coolers [41,49,50,51]. We therefore show also results of DFB lasers with lateral current injection which emit up to 13.6 mW average power at −30 ◦C. But there exist more reasons why QC lasers are interesting light sources: their emission lies in the so-called atmospheric window regions around 5 µm and 10 µm. So far, the main application field was infrared (IR) spectroscopy, but thanks to the good transparency of the atmosphere in these wavelength bands even under foggy or rainy conditions, optical data transmission could

64


be a new interesting application field for QC lasers. We thus present also measurements on this topic.

2

Gain Region Designs

In this section, we will first describe the basics of QC laser operation and then briefly discuss the main properties of the two gain region designs which were used for all devices presented in this work. Obviously, the list of possible designs is not complete, and the two cases outlined here are not treated in a very detailed way. 2.1

Basic Working Principles of QC Lasers

Compared to their ‘predecessors’, the lead salt diode lasers, the most striking features of QC lasers are their high output powers and the fact that they can function at temperatures of up to 470 K. In addition, unlike in diode lasers, the fundamental optical transitions are between conduction band sublevels (subbands) rather than between the conduction band and the valence band. Finally, QC lasers are unipolar devices, meaning that only one type of carrier, i.e. electrons, is involved in the process of light generation. The big advantage of these so-called intersubband transitions is that their energy can be tailored in a wide range by adjusting the thickness of the QWs and barriers. In contrast to interband transitions, where the material bandgap defines the transition energy, one can use the same two semiconductor materials for a variety of intersubband transition energies. Of course this nice advantage has its price. The first and most important drawback of intersubband transitions is that the lifetime of an excited state in a QW is extremely short, on the order of 1 ps. This time, which is a factor of 1000 shorter than in typical interband transitions, is controlled by optical phonon scattering, a process which works very efficiently at room temperature. The second point deserving attention is that the optical gain produced by one single intersubband transition is usually not sufficient to reach lasing threshold. While the second issue can be circumvented by using a cascade of up to 35 stages of gain and injector regions, the first point needs a more careful treatment. The fundamental idea for the generation of a population inversion in an intersubband laser structure is to take advantage of a three level system like in Fig. 1. The two lower levels (levels 1 and 2) are separated by exactly one optical phonon energy (34 meV for InP-based materials, 36 meV for GaAs-based materials). This results in a lower state lifetime of about τ21 = 0.5 ps, whereas the lifetime of the upper state (level 3) can be as long as τ32 = 2.0 ps. This trick was at the basis of the first working QC laser, and it has been refined since. Today’s state-of-the-art QC lasers use even a four level system with a double-phonon resonance at the lower laser level. The resulting improvement of the lifetime ratio has led to a considerably lower threshold current


3

Te>>Tl

65

Te=Tl continuum

τ32 minigap 2 1

τ21 miniband

gain region injector region Fig. 1. Basic design features of a generic QC laser cell. One period is divided into a gain region (left half ) and an injector region (right half ). The flow of electrons is indicated by one-head arrows going from left to right; the optical transition by a wavy line. Te is the electron temperature, while Tl denotes the lattice temperature

density at room temperature. A typical stage of a QC laser is composed of a gain region (usually three or four QWs and barriers) and an injector region. The injector region fulfills different tasks; the most important ones are to build a minigap at the energy of the upper lasing level, to provide a miniband aligned with the lower laser levels, and to efficiently inject electrons via a tunnel barrier into the downstream gain region. The injection barrier has a thickness which results in an anticrossing energy splitting of about 15 meV between the injector ground state and the upper laser level. The injector minigap is needed to prevent the injected electrons from directly escaping into the continuum. Finally the miniband helps to efficiently extract the electrons from the gain region. As the carriers move towards the next injection barrier, the injector miniband reduces its width from initially roughly 150 meV to about one half of this value. This narrowing is important in order to cool down the electron distribution within the injector. Quite generally, one can calculate an approximate voltage drop of one photon energy plus roughly two or three phonon energies per period. For a laser operating at 10 µm emission wavelength, one computes 200 mV per period. If the threshold voltage does not exceed 7.6 V, we end up with a maximal number of stages of 38. In a QC laser, the current is obviously, at least in first order, not affected by this cascading scheme. One has therefore multiplied the gain at constant drive current, but at the cost of a relatively high operating voltage. The latter sometimes causes problems, for example when trying to reach CW operation at elevated temperatures or when directly modulating QC lasers at high frequencies. Let us now have a closer look on the two gain region designs which were used throughout this work. They are referred to as the three QW gain region

66


or ‘diagonal-anticrossed’ design and the four QW gain region or ‘doublephonon resonance’ design. 2.2

Three Quantum Well Gain Region

Among the first working QC lasers were devices using double QW gain regions. Their width and barrier thicknesses were designed in order to produce three states named level 1, 2, and 3 as outlined in Sect. 2.1. Level 1 and level 2 are separated by a single phonon energy which makes the lifetime of level 2 very short (on the order of 0.5 ps). Levels 3 and 2 are separated by the desired photon energy and have a high spatial overlap. This results in a vertical transition with a lifetime of the upper lasing level of typically about 1–2 ps. In the three QW gain region design, which is also referred to as diagonal-anticrossed, the main novelty was the introduction of a third, narrow QW just between the injection barrier and the QW pair forming the three lasing states. This narrow well slightly delocalizes the wavefunction of level 3 towards the injection barrier, leading to an increased overlap between level 3 and the ground state of the injector, g. Since the first excited state of this thin well is energetically much higher than levels 1 and 2, their corresponding wavefunctions are not dragged towards the injection barrier; and therefore the overlap argument does not apply for these lower states. Due to the much smaller resulting overlap between level g and levels 1 and 2, injection into the lower laser levels is quite unlikely, while injection into level 3 is very efficient. The main results of the three QW design are thus a good injection efficiency into the upper laser level, a low probability of injecting into states 1 and 2, and a sufficiently high matrix element for the optical transition between the lasing levels 3 and 2. Because of the slightly diagonal nature of this transition, a pronounced Stark shift can be observed at high bias voltages. This effect has been utilized to fabricate Fabry–Pérot lasers with a large tuning range [52]. 2.3

Gain Region with Double-phonon Resonance

In order to achieve a further improvement of the lasing properties, new designs for the gain region have been invented. Taking advantage of the pioneering work of Scamarcio et al. [47] and Tredicucci et al. [53] on superlatticeinterminiband QC lasers,we developed a gain region which combines the good extraction efficiency of those devices and the efficient injection of the three QW gain region [19,46]. This new design is referred to as double-phonon resonance or 4 QW gain region [54], and it works equally well for lasers in the 5 µm and the 10 µm band. In the following, we present a design for a 5.3 µm QC laser, where the central portion of the waveguide consisted of 28 periods; each one of the latter contained a gain region and an injector region, separated by an injection- and an exit barrier. A schematic conduction band diagram of one period of the gain region is shown in Fig. 2.


67

Injection barrier Exit barrier

g

4

MINIGAP 3 2 1

MINIBAND g

4 QW gain region

Injector region

Fig. 2. Schematic conduction band diagram of a portion of the active region and moduli squared of the relevant wavefunctions of a 5.3 µm QC laser with a 4 QW gain region. An electric field of 7.5 × 104 V/m was applied to align the structure. The In0.6 Ga0.4 As/In0.44 Al0.56 As layer sequence of one period of the active region, starting from the injection barrier is as follows: 42/13/14/50/14/44/15/39/24/29/19/26/20/23/21/22/23/21/30/21. Thicknesses are in ˚ A, In0.44 Al0.56 As barrier layers in bold, In0.6 Ga0.4 As QW layers are in roman, doped layers (Si, 4 × 1017 cm−3 ) are underlined

The upper and lower lasing states are the wavefunctions with numbers 4 and 3, respectively; and the ground state of the injector is denoted with the letter g. The gain region was composed of 4 QWs which resulted in three coupled lower states (levels 1, 2, and 3). Each two of the latter (i.e. level 3 to 2, and level 2 to 1) were separated by one phonon energy. This double-phonon resonance yielded a short intersubband electron scattering −1 −1 −1 + τ31 ) = 0.25 ps), and time (τ32 = 0.31 ps, τ31 = 1.24 ps, τlow = (τ32 therefore an efficient extraction of the electrons into the injector region. The upper lasing state exhibits a much longer intersubband electron scat−1 −1 −1 −1 + τ42 + τ41 ) = 1.25 ps (using τ43 = 3.41 ps, tering time of τup = (τ43 τ42 = 3.26 ps, and τ41 = 5.07 ps). The relatively large dipole matrix element, z43 = 2.1 nm, confirms that the lasing transition is mainly a vertical one. Thanks to the thin first well, which reduces the overlap of the injector ground state g with the lower lasing state wavefunctions 1, 2, and 3, the injection efficiency was kept similarly high as in the classical 3 QW design. Finally, the use of strained material makes the barrier height considerably larger (620 meV

68


instead of 520 meV like for unstrained material), which quite efficiently reduces electron tunneling from the upper lasing state into the continuum. In addition, all wells can be made somewhat thicker, which has a positive effect on the interfacial roughness scattering. This is consistent with the relatively narrow linewidth of the luminescence peak measured at 4 K (hν = 244 meV, ∆E = 11 meV) and at 300 K (hν = 233 meV, ∆E = 25 meV).

3

Fabrication Technologies

In this section, we present several key processing steps which were used for the fabrication of the different DFB lasers described in the current work. The aim of these processes was, at least in the beginning, to find simple solutions which would yield acceptable results. Later we went on to more sophisticated fabrication methods which resulted in higher performance. 3.1

Surface Grating with Lateral Current Injection

Epitaxial regrowth is often considered as a costly and risky process. The most obvious method to avoid epitaxial regrowth in DFB lasers consists of fabricating the grating directly on top of the waveguide [39]. However, in order to prevent the top contact metal from introducing a large waveguide loss (which is especially harmful for QC lasers since they are operating in TM mode), one has to either decrease the grating coupling coefficient by utilizing a thicker top cladding layer, choose a metal with small refractive and absorption indices [55], or completely avoid the metal on top of the waveguide [56]. As shown in Fig. 3, such a device consists basically of a waveguide with a semiconductor lower cladding layer and air acting as top cladding. The heavily n-doped InGaAs cap layer, which serves as host layer for the grating, is highly conducting to allow lateral current injection and distribution throughout the device. The most important consequences of such a design are obviously that there is a large refractive index step between semiconductor and air, and that there are low calculated losses of 12 cm−1 . This results in both a high coupling coefficient of the grating and a relatively high net gain of the laser; thus it potentially allows the fabrication of short devices with a low threshold current. This type of DFB laser operating in both the 5 µm and the 10 µm bands will be presented in Sect. 4.2. The growth for this type of QC laser was based on MBE of lattice matched InGaAs/InAlAs layers on top of an n-doped InP (Si, 2 × 1017 cm−3 ) substrate. The growth process started with the lower waveguide layer (In0.53 Ga0.47 As, Si, 6 × 1016 cm−3 , total thickness 1.5 µm), proceeded with an active region (thickness 1.75 µm) and a 2.1 µm thick upper waveguide layer (In0.53 Ga0.47 As, Si, 6 × 1016 cm−3 ) and was terminated with a 0.5 µm thick highly n-doped contact layer on top (In0.53 Ga0.47 As, Si,


69

InGaAs waveguide layers grating and lateral contact layer ZnSe isolation layer

active region

Ti/Au contact layer

InP substrate

(a)

50 µm

(b)

Fig. 3. (a) Schematic cross-section through the laser waveguide showing the exact position of the grating with respect to the active layer and the metal top contact. (b) Scanning electron microscopy picture of a QC DFB laser. Please note that there is no metal in the central top part of the ridge waveguide

2 × 1018 cm−3 ). This contact layer was also the host layer for the grating. The active region, which thus formed the central part of the waveguide, consisted of 35 periods; those were alternating n-doped funnel injector regions and un-doped triple QW gain regions. The laser transition in the latter was diagonal, similar as described in [19]. The layer sequence of the structure, starting from the injection barrier, is as follows: 42/31/9/64/10/60/28/39/10/38/12/37/15/39/17/40. All thicknesses are in ˚ A, In0.52 Al0.48 As barrier layers are in bold, In0.53 Ga0.47 As well layers are in roman, and n-doped layers (Si, 2.5 × 1017 cm−3 ) are underlined. A more detailed description of the structure was published in a paper about wavelength tunable Fabry–Pérot lasers fabricated from material using the same active region [52]. The fabrication of these DFB lasers relied on holographically defining a grating with 1.59 µm period (neff = 3.22), and wet chemical etching of

70


the grating in an HBr:H2 O2 :H2 O solution to a depth of 0.4 µm. We used a 488 nm Ar-ion laser and a 90◦ corner reflector mounted on a rotational stage for the grating exposure. Standard processing techniques were then employed to define ridge waveguides with a width of 35–55 µm (etch depth 4.5 µm) and a length of 1–1.5 mm. 300 nm of ZnSe served as an electrical passivation layer, Ti/Au (10/400 nm) was used as the top contact metal. Thinning, back contacting (Ge/Au/Ag/Au, 12/27/50/100 nm), and cleaving completed the processing. As shown by the schematic cross-section in Fig. 3a and the scanning electron microscopy picture in Fig. 3b, the contact metal covered only the edges (about 5 µm on each side) of the ridge to prevent large absorption losses in the waveguide, but still allow lateral current injection. 3.2

Lasers with InP Regrowth

Although DFB lasers with surface gratings are, among other amenities, quite simple to fabricate, they suffer from their own specific drawbacks. The most important one is the fragile top contact which tends to overheat, and which does not help very much in distributing the heat. In order to improve the thermal conductivity and the modal overlap, the top cladding should consist of the same material as the lower cladding layer. This can be achieved by growing, in a first step, only the active region and the waveguiding layers. In a second epitaxial growth, the InP top cladding is grown, ideally in an MOVPE system. The second growth is performed after a thorough cleaning of the surface in concentrated sulfuric acid. Depending on the emission wavelength of the laser, the typical sequence of these layers might be as follows: a 2.5 µm thick top cladding layer (InP, Si, 1 × 1017 cm−3 ), a 0.85 µm thick contact layer (InP, Si, 7 × 1018 cm−3 ), and a 10 nm thick cap layer (InP, Si, 1 × 1020 cm−3 ). An even further improvement of the thermal properties can be achieved by growing InP around the entire active region (buried heterostructure design). This requires two regrowth steps in the MOVPE. The first one for the top cladding; the second after having etched ridge waveguides into the material. Such a process will be described in Sect. 4.5. 3.3

Junction Down Mounting

In order to fully benefit from the good thermal conductivity of an InP top cladding, it is necessary to mount the lasers with their active area directly on the heatsink material. This technique is called junction down mounting. For all mounting techniques, we thermally evaporated a 2 µm thick In layer on copper heatsinks. A thin layer of soldering flux was then deposited on the surface. Afterwards, the laser chips were pressed onto this surface and the whole sandwich was heated to 160 ◦ C for a couple of seconds. Flux remnants were then removed by acetone and subsequent O2 -plasma cleaning. An important parameter of the whole procedure is the distance between the front facet and the front edge of the copper heatsink. If this distance is too large,


71

the facet cannot be cleaned properly; if the distance is too small or ‘negative’ (overhanging laser), there is a certain risk that the end section of the laser is not correctly cooled; this can cause severe damage of the device at high duty cycle operation.

4

Measurement Results

This section describes the results achieved with the different devices. We will first present the measurement setup, and then continue with the results from DFB lasers with lateral current injection, surface-emitting DFB lasers, devices with InP-regrowth, and junction-down mounted DFB lasers. At the very end of the section, we will briefly outline the results of the experiments with our room temperature CW QC lasers. 4.1

Measurement Setup

For device testing at low duty cycle, we mounted the lasers on small copper heatsinks which were then placed into a temperature controlled N2 flow cryostat. The light of the QC DFB laser was collected by f/0.8 optics and fed into a high resolution Fourier transform IR spectrometer (Nicolet type Magna-IR 860, FTIR), where we detected it using a liquid nitrogen cooled HgCdTe detector. For the measurement of L–I curves, we measured the intensity with a calibrated 250 × 250 µm2 room temperature HgCdTe detector. Typical pulsed operating conditions used throughout this work were 100 ns pulses with a 5–10 kHz repetition frequency. These conditions guaranteed that minimal device heating would occur during the single pulse. A boxcar integrator allowed us to detect the intensity during that period within the pulse where the laser ran stably. When going to elevated duty cycles, mounting was typically junction down, either on copper heatsinks or on diamond platelets soldered onto copper. Testing was done in a Peltier-cooled aluminum box with an antireflection coated ZnSe window. In this box, they were held at a constant temperature between −30 ◦ C and 60 ◦ C. A commercial pulse generator (Alpes Lasers, TPG 128) with power supply (Alpes Lasers, LDD 100) allowed us to deliver arbitrary long current pulses (10–200 ns) at a variable repetition frequency of up to 5 MHz to the laser. For spectral measurements, two different possibilities were investigated; namely a grating spectrometer and FTIR. In the case of testing with a grating spectrometer, the light of the QC DFB laser was collected by an f/1.33 Au-coated parabolic mirror, directed onto a flat mirror to turn the polarization, and finally bounced off a f/3.75 parabolic mirror to enter the 200 µm wide entrance slit of the spectrometer (Jobin-Yvon, dfocal = 0.3 m). Behind the spectrometer, the light was detected with a battery-driven pyro-electric detector. For the measurement of L–I curves, we measured the average power

72


directly in front of the ZnSe window with a calibrated thermopile detector. In the case of the spectral measurement with an FTIR, we used the same optics as already mentioned above. 4.2 DFB Laser with Lateral Current Injection in the 5 µm and 10 µm Band This section is devoted to the results from lasers with lateral current injection. This technique is simple, reliable, and yielded lasers with very good peak powers. It has also been tried for Fabry–Pérot lasers in the 5 µm band. The drawback of this method is clearly the poor thermal management. Edge Emitting DFB Laser at 10 µm

4.2.1

Typical L–I and I–V curves of a 45 µm wide and 1.2 mm long device are shown in Fig. 4. The current pulses were 100 ns long, and a pulse repetition frequency of 5 kHz was used for all temperatures. At low temperatures, we observed a threshold current of 1 A and a maximum output power of 230 mW for 9.7 V bias voltage. The slope efficiency at this temperature was 220 mW/A and a threshold current density of 1.85 kA/cm2 was determined. At room temperature, we still obtained 80 mW optical output power with a slope efficiency of 80 mW/A; however, the threshold current increased to 2.9 A (threshold current density of 5.4 kA/cm2 ), and an operating voltage of 12.5 V was seen. From the increase in threshold current, we were able to derive a characteristic temperature T0 of 204 K.

20

300

Voltage [V]

15

T0 = 204 K

80K

4

100K 150K

2 100

10

200

200 300 Temperature [K]

200K 250K

80K

100 280K

5 300K

Output power [mW]

2

j th [kA/cm ]

6

300K

0

0

1 2 3 Injection current [A]

4

0

Fig. 4. L–I and I–V curves of a 45 µm wide and 1.2 mm long DFB QC laser measured at different temperatures. The inset shows a plot of the threshold current density versus device temperature


73

A spectral measurement below the lasing threshold revealed regular Fabry–Pérot modes with a spacing of 1.7 cm−1 (cavity length: 850 µm). The Bragg reflector’s stopband with a width of 2.5 cm−1 occurred almost at the centre of the spontaneous emission peak, at 995.7 cm−1 (85 K), 994.6 cm−1 (105 K), and at 992.3 cm−1 (150 K). From the stopband-width, we determined the coupling coefficient of the grating to be κ = ∆ λπneff /λ2 = 24 cm−1 ; this number agrees well with a value obtained from an estimation based on the effective refractive index difference of ∆ n = 1.8 × 10−2 between areas with and without the grating layer (κ = π∆ neff /2λ = 28 cm−1 ). A relatively small free carrier absorption loss of 12 cm−1 was calculated for this device, whereas a laser utilizing our standard waveguide design with a 2.2 µm thick InAlAs/InGaAs upper cladding layer and a metal-covered grating would suffer from a waveguide loss of 30 cm−1 . In addition, the refractive index contrast would be reduced by almost two orders of magnitude, namely to a value of ∆ n = 2.3 × 10−4 . Since the partial removal of the contact layer leads to a slight gain variation, a small amount of loss coupling might also be present in this device. Figure 5 shows the lasing spectra at temperatures between 85 K and 300 K. We observed single-mode operation for all temperatures and, in particular, at maximum power for each individual temperature. We determined the linewidth to be on the order of 0.3 cm−1 , which corresponds to the resolution limit of our experimental set-up. The emission wavelength at 85 K was 996 cm−1 ; at room temperature, it decreased to 982 cm−1 . As already mentioned above, the luminescence peak was found in the vicinity of 975 cm−1 for all temperatures. The temperature tuning coefficient of the lasing peak was constant over the entire temperature range, and its magnitude was 1/λ × ∆ λ/∆ T = 6.5 × 10−5 K−1 (∆ ν/∆ T = −0.063 cm−1 /K). These numbers are consistent with what has been reported in the literature [22].

Intensity [a.u.]

1.0 0.8 0.6

300K

85K

0.4 0.2 0 970

980

990

1000

1010

-1

Wavenumbers [cm ] Fig. 5. Lasing spectra of a 45 µm wide and 850 µm long DFB QC laser at different temperatures between 85 and 300 K. All spectra were measured with the maximum possible output power at each individual temperature

74


In a second processing run with a changed grating period of 1.63 µm (instead of 1.59 µm), we were able to perform some measurements regarding the influence of a detuning between the Bragg peak and gain peak. The first series of DFB lasers with a detuning of about 5–10 cm−1 at room temperature (Λ = 1.59 µm) showed high performance at all temperatures. The second series with a detuning of about 30 cm−1 (Λ = 1.63 µm) did not lase at room temperature, and at low temperatures, the output power did not exceed 60 mW. Since we determined a full width at half maximum (FWHM) of the gain peak of ∆ ν = 65 cm−1 , it is clear that the laser performance will degrade rapidly with increasing de-tuning between the Bragg reflection maximum and gain peak. 4.2.2

Surface Emitting 10 µm DFB Laser

The fabrication of these second order DFB lasers started again by holographically defining a grating with 3.15 µm period (neff = 3.22), and wet chemical etching of the grating in a HBr:H2 O2 :H2 O solution to a depth of 0.6 µm (etch rate approximately 100 nm/s). The grating lines run along the dove-tail direction of the crystal in order to achieve a non-rectangular tooth profile and to obtain a sufficiently high first-order Fourier component. This is quite critical for device performance because a symmetric rectangular second-order grating contains no first-order Fourier component. Unfortunately, grating fabrication by holography and wet etching usually results in under-etching and leads to narrow grating lines and large spaces between them. Because a small duty cycle reduces the average refractive index and therefore the overlap factor of the grating layer, the coupling coefficient and the diffraction efficiency of the grating become small, resulting in a limited amount of surface-emitted power. Typical edge emission L–I and current versus voltage (I–V ) curves of a 55 µm wide and 1.125 mm long device showed a threshold current of 1.3 A and a maximum output power of 210 mW from the facet (Fig. 6). The slope efficiency at this temperature was 105 mW/A and a threshold current density of 2.1 kA/cm2 was determined [57]. At room temperature, we obtained 70 mW optical output power from the facet, with a slope efficiency of 70 mW/A. However, the threshold current increased to 3.45 A (threshold current density of 5.6 kA/cm2 ), and an operating voltage of 10.5 V was seen. From the increase in threshold current, we were able to derive a characteristic temperature T0 of 258 K. The higher slope efficiencies at 150 and 200 K (compared to 85 K) are due to the fact that the grating resonance wavelength is tuned at the center of the gain peak at these temperatures. This also explains the higher T0 value of these devices than our previously published DFB lasers [39]. Figure 6 shows a comparison between the L–I-characteristics for both the facet and grating emissions. Because of the unfavorable duty cycle of the grating (about 0.25–0.35 instead of 0.7), there is a bigger fraction of the total energy radiated from the facet than from the grating. Nevertheless, we


75

200 85 K

Output power [mW]

175

150K

150 125 100 75

grating emission facet emission

300 K

85 K 50 150 K 25

300 K

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Injection current [A] Fig. 6. Comparison between L–I curves with the output intensity collected from the facet and the grating at three representative temperatures. The inset shows a scanning electron micrograph of the grating

Far field intensity [a.u.]

1.0

perpendicular parallel

0.8 0.6 0.4 0.2 0.0 -15

-10

-5

0

5

10

15

Far field angle [ °] Fig. 7. Far field distributions of the surface emission from a second-order DFB QC laser measured in two orthogonal directions (along and perpendicular to the waveguide)

obtained an optical output power from the grating of 60 mW at 85 K and 18 mW at 300 K; these values correspond to slope efficiencies of 30 mW/A and 18 mW/A, respectively. In Fig. 7, we present the far field distribution of the grating emission in both directions. In the direction along the waveguide, we observed, due to the wide aperture and the Bragg reflection, a very narrow far field angle of about 1◦ (FWHM), whereas in the other, perpendicular direction, the far field angle was equivalent to the one observed at the corresponding direction of the facet, namely about 14◦ (FWHM). According to [58], a certain amount of gain/loss coupling is always present in

76


a second order DFB laser. This loss coupling should prevent the device from oscillation in both stopband modes. Most likely because of their relatively weak coupling coefficient, we nevertheless observed, on short devices and at high injection currents, lasing action in both stopband modes. This allowed us a relatively precise measurement of the Bragg reflector’s stopband-width; the value was ∆ ν = 1.1 cm−1 . From this figure, we determined the coupling coefficient of the grating to be κ = ∆ λπneff /λ2 = 12 cm−1 ; with neff being the effective refractive index of the propagating mode. This number is of the same order of magnitude as the calculated value obtained from the effective refractive index difference of ∆ neff = 1.8 × 10−2 between areas with and without the grating layer (κ = π∆ neff /2λ = 9 cm−1 ) and the first Fourier component of the grating teeth. The spectral behavior was very similar as presented for the edge emitting DFB laser presented in Sect. 4.2.1. The emission wavelength at 85 K was 1003 cm−1 ; at room temperature, it decreased to 989 cm−1 . The luminescence peak was found in the vicinity of 1000 cm−1 for all temperatures. The temperature tuning coefficient of the lasing peak was constant over the entire temperature range, and its magnitude was 1/λ × ∆ λ/∆ T = 6.1 × 10−5 K−1 (∆ ν/∆ T = −0.06 cm−1 /K). 4.2.3

Improved Average Power Operation at 10 µm

Some fine tuning in the crystal growth and an improved top metallization made these DFB lasers much more robust. This resulted mainly in higher duty cycle operation than what was reported in Sect. 4.2.1 [19,39,57]. The fabrication process started again by holographically defining a grating with 1.6378 µm period (neff = 3.163), and wet chemical etching of the grating to a depth of 0.3 µm in a H3 PO4 :H2 O2 :H2 O solution (4:1:10, etch rate 800 nm/min). The main difference to the former device processing was the thickness of the top contact. Here, we used 1 µm Au instead of only 400 nm. This had a positive effect on the device heating, but also improved the contact problem across the 4 µm high step between the top of the ridge and the contact pad. At 1.5% duty cycle, we observed threshold currents of 4.35 A at −30 ◦ C and 5.8 A at 60 ◦ C; these values are equivalent to threshold current densities of 5.3 kA/cm2 and 7 kA/cm2 , respectively. Figure 8 shows the maximal average output power measured at a duty cycle of 3%. The highest output power was achieved at −30 ◦ C; its value at the thermal roll-over point was 13.6 mW. At room temperature, we still observed 4 mW, and finally, at 60 ◦ C, the value decreased to 1.5 mW. The last number which was smaller than the one for 1.5% duty cycle indicates that, at a temperature of 60 ◦ C, the duty cycle with the best performance is smaller than 3%. Considering the high voltage necessary to achieve these output powers, it becomes clear that the device suffers from excessive heating, in particular at higher temperatures. This hypothesis is confirmed by the emission spectra at maximal output power and different temperatures between −30 ◦ C and 60 ◦ C. A pulse length of 45 ns and a pulse repetition frequency of 667 kHz was


77

Voltage [V]

-30 °C

10

10

5

-30 °C

5

60 °C

0

60 °C

0

2

4 6 Current(A)

8

Average power [mW]

15 S1805g7up 3 % duty cycle junction up

15

0

Fig. 8. Average power and voltage versus current curves of a 1.5 mm long QC DFB laser operated at 3% duty cycle and at different temperatures between −30 ◦ C and 60 ◦ C

used to drive the laser at a duty cycle of 3%. The single emission peak tunes from 968.6 cm−1 at −30 ◦ C to 961.4 cm−1 at 60 ◦ C. The temperature tuning is ∆ ν/∆ T = −0.08 cm−1 /K, a value which is 27% larger than the one we reported earlier for devices running at low duty cycle. Since the temperature tuning of a DFB laser is only due to a temperature-induced refractive index change, we are able to estimate the overheating of the laser at a heatsink temperature of 60 ◦ C. Under the assumption of a temperature difference ∆ T between the active region and the device holder at −30 ◦ C, we find at 60 ◦ C using the usually observed temperature tuning of DFB lasers a much larger temperature difference of ∆ T + 25 K. This is consistent with the observed increase in threshold current from 5.8 A to 6.1 A at this temperature when going from 1.5% to 3% duty cycle. 4.2.4

High Temperature Operation at 5.3 µm

These lasers utilized a gain region based on a double-phonon resonance as outlined in Sect. 2.3. Fabrication relied this time on MBE of straincompensated In0.6 Ga0.4 As/In0.44 Al0.56 As layers on top of an n-doped InP (Si, 2 × 1017 cm−3 ) substrate. The growth process started with the lower waveguide layers (In0.53 Ga0.47 As, Si, 6 × 1016 cm−3 , thickness 0.34 µm), proceeded with an active region (Si, thickness 1.43 µm) and was finished by a thicker set of upper waveguide layers (In0.53 Ga0.47 As, Si, 6 × 1016 cm−3 , thickness 0.5 µm) and a highly n-doped cap layer (In0.53 Ga0.47 As, Si, 2 × 1018 cm−3 , thickness 0.4 µm) on top. This cap layer was also the host layer for the grating, as reported earlier [39,57]. The fabrication process started once more by holographically defining a grating with 0.825 µm period (neff = 3.21), and wet chemical etching of the grating to a depth of 100 nm in a H3 PO4 :H2 O2 :H2 O solution (4:1:10, etch rate 800 nm/min). Wet chemical etching in a HBr:H2 O2 :H2 O solution (1:1:10, etch rate 800 nm/min) was then used to define broad ridge waveguides with

78


a width of 44 µm (etch depth 5 µm) and a length of 3 mm. 300 nm of Si3 N4 served as an electrical passivation layer and Ti/Au (10/1000 nm) was used as the top contact metal. Thinning to a thickness of 150 µm, back contacting (Ge/Au/Ag/Au, 12/27/50/100 nm), and cleaving completed the processing. Typical L–I and I–V curves of a 44 µm wide and 3 mm long device are shown in Fig. 9. The 45 ns long pulses came along with a pulse repetition frequency of 333 kHz; this corresponds to a duty cycle of 1.5%. The emitted peak power drops from 1.15 W at 0 ◦ C via 0.45 W at 60 ◦ C to 92 mW at 120 ◦ C. These maximum power values correspond to slope effciencies of 247 mW/A at 0 ◦ C, 135 mW/A at 60 ◦ C, and finally 45 mW/A at 120 ◦ C. We observed threshold currents of 4.1 A at 0 ◦ C, 5.6 A at 60 ◦ C, and 7.4 A at 120 ◦ C; these values are equivalent to threshold current densities of 3.1 kA/cm2 , 4.2 kA/cm2 , and 5.6 kA/cm2 at the respective temperatures. The characteristic temperature, T0 , which empirically describes the behavior of the threshold current as a function of the temperature, was 203 K. As a comparison, we also fabricated Fabry–Pérot (FP) lasers from the same material, using the same lateral current injection scheme and an identical geometry. For those multimode lasers, we achieved even higher peak powers of 1.76 W at 0 ◦ C (dP/dI = 311 mW/A, jth = 3 kA/cm2 ) and 0.83 W at 60 ◦ C (dP/dI = 258 mW/A, jth = 5.4 kA/cm2 ). T0 was with 190 K comparably high as for the DFB lasers. The excellent performance of both DFB and FP devices demonstrates that lateral current injection with or without surface gratings is a very effective technique for high power QC lasers [54]. In Fig. 10, we present the emission spectra measured at the thermal rollover point for five representative temperatures between 0 ◦ C and 120 ◦C. The same pulse lengths and repetition rates as in Fig. 9 were used for this experiment. The single lasing mode tunes linearly with temperature from 1898 cm−1

Voltage [V]

15

1.2

S1869b2dn junction up 1.5 % duty cycle

0 °C

0.8

10

0 °C 120 °C

0.4

5

Peak power [W]

0

Current density [kA/cm2] 2 4 6

120 °C

0

0

2

4 6 Current(A)

8

0 10

Fig. 9. L–I and I–V curves of a 44 µm wide and 3 mm long QC DFB laser at a wavelength of 5.3 µm. The curves were measured at a duty cycle of 1.5% and at five different temperatures between 0 ◦ C and 120 ◦ C


Wavelength [µm] 5.32 5.30 5.28

5.34

Intensity [a.u.]

10 1

79

5.26

S1869b2dn junction up 1.5 % duty cycle

0.1

0°C

120 °C

0.01 1870

1880

1890

1900 -1

Wavenumbers [cm ] Fig. 10. Emission spectra of a 44 µm wide and 3 mm long QC DFB laser at a wavelength of 5.3 µm. The spectra were collected at the same five temperatures as the L–I–V curves of Fig. 9 and at the thermal roll-over point for each temperature

at 0 ◦ C to 1881 cm−1 at 120 ◦ C. We thus obtained a temperature tuning rate of ∆ ν/∆ T = −0.145 cm−1 /K. Because of the limited dynamical range of our set-up, a correct measurement of the side mode suppression ratio (SMSR) is not possible. Keeping this fact in mind, we understand that the signal to noise ratio (SNR) of the lasing peaks in Fig. 10 is only a lower limit for the SMSR. Due to the power decrease at higher temperatures, these SNR values drop from 30 dB at 0 ◦ C to 20 dB at 120 ◦C. Obviously this does not necessarily mean that the SMSR dropped also by 10 dB. The full width at half maximum linewidth of the emission peaks is on the order of 0.75 cm−1 for all investigated temperatures. It is obvious that thermal chirping dictates this linewidth, similar as outlined in [45]. For this reason, we made experiments with a reduced pulse length (down to 20 ns) and found a linewidth on the order of 0.15 cm−1 . The maximal average output power at three different temperatures was measured up to 60 ◦ C. This experiment was done using 45 ns pulses as well, but with a higher repetition frequency of 1.11 MHz; this results in a duty cycle of 5%. The highest output power was achieved at 0 ◦ C; its value at the thermal roll-over point was 39.5 mW. At room temperature, we still observed 22.5 mW, and finally, at 60 ◦ C, the value decreased to 9.5 mW. As reported in Sect. 4.2.3 on a 10.4 µm DFB laser using the same lateral contact scheme, we observed that the duty cycle yielding the highest thermal rollover power changed with temperature. This is illustrated best when dividing the average thermal roll-over powers at 5% and 1.5% duty cycle for the three available temperatures. At 0 ◦ C, the ratio r = Pmax (5%)/Pmax (1.5%) is 2.28, at 30 ◦ C we find r = 1.94, and finally at 60 ◦ C we obtain r = 1.38. This shows how the laser suffers from overheating at elevated temperatures, which

80


is understandable when taking into account the small area into which current is injected. Another indication for this overheating effect is given by the temperature tuning rate of ∆ ν/∆ T = −0.145 cm−1 /K, which is somewhat larger than what has been published in the literature. In [21], for example, a value of ∆ ν/∆ T = −0.124 cm−1 /K was reported. Finally, the overheating manifests itself also in slightly higher threshold current densities, for example 3.35 kA/cm2 for 5% at 0 ◦ C compared to 3.1 kA/cm2 for 1.5% at 0 ◦ C. Taking into account the T0 value from above, such an increase of the threshold current density corresponds to a temperature increase in the active region of almost 15 K. 4.3

DFB Lasers with InP Over-Grown Grating

Overheating effects like the ones mentioned in Sect. 4.2.3 play an important role for laser performance and ageing. This is particularly critical for DFB lasers comprising a lateral current injection scheme. Because of their small contact area, the current density in the narrow metal stripes reaches a multiple of the one in the entire laser structure. Already at moderate duty cycles of a few percent, this can lead to contact degradation. Furthermore, if high average output powers are pursued, junction down mounting is the way to go. Since a metal layer on a surface grating is detrimental for device performance, lasers with a lateral contact scheme do not allow junction down mounting. Thus, for reaching high output powers, epitaxially regrown DFB lasers are better suited. In the following, we will present two different laser structures. They are both based on double-phonon resonance gain regions; and they both have the grating overgrown with an InP top cladding. Device S1850 was tested only in junction up configuration. Nevertheless, it performed extremely well at high temperatures of up to 150 ◦ C. On the other hand, S1810 (which was mounted junction down) could be driven at high average output powers. The layer structure of S1850 has been published in [32]. For the first of these DFB lasers, L–I and I–V curves as a function of temperature are presented in Fig. 11. This device was 24 µm wide and 1.5 mm long. The curves are obtained at 1.5% duty cycle in order to demonstrate the excellent high temperature capability of the material. The highest temperature for which we could reliably measure the L–I–V curve was 130 ◦ C. The threshold current increase from 0.95 A at −20 ◦ C to 2.2 A at 130 ◦ C corresponds to a characteristic temperature of 179 K. At −20 ◦ C, we observed a slope efficiency of 187 mW/A with a roll-over power of 300 mW; while at 130 ◦C, 62 mW/A and 47 mW were seen. In Fig. 12, we present the emission spectra measured at temperatures between 0 ◦ C and 150 ◦ C and a constant current of 2.5 A. The duty cycle for this series of measurements was 0.1% with 20 ns long pulses and a repetition rate of 50 kHz. Depending on the heatsink temperature, a side mode suppression ratio of at least 20 dB was seen. The emission peak shifted continuously from

High Performance Quantum Cascade Lasers and Their Applications 12

Voltage [V]

300

8

200 4

100

0

1

2

3

Peak power [mW]

400 S1850b2 junction up 1.5% duty cycle

0

81

0

Current [A]

Fig. 11. Series of L–I and I–V curves of a 24 µm wide and 1.5 mm long DFB laser taken from the sample S1850 and at different temperatures between −20 ◦ C and 130 ◦ C. For this particular measurement, a duty cycle of 1.5% (45 ns pulse length/ 333 kHz pulse repetition frequency) was used

Intensity [a.u.]

1000

S1850b2 junction up 0.1 % duty cycle

100 10 1 1100

1105

1110

1115 -1

Wavenumbers [cm ] Fig. 12. Series of emission spectra of a 24 µm wide and 1.5 mm long DFB laser taken from the sample S1850 and at different temperatures between 0 ◦ C and 150 ◦ C (spacing 10 ◦ C). Each spectrum was measured at an injection current of 2.5 A and at 0.1% duty cycle

1114 cm−1 at 0 ◦ C to 1102.4 cm−1 at 150 ◦ C. This is equivalent to a temperature tuning coefficient of 1/λ × ∆ λ/∆ T = −6.95 × 10−5 K−1 . At low temperatures, where the operating current of 2.5 A was much larger than the threshold current, the laser suffered somewhat from chirping. This led to a small shoulder on the long wavelength side (i.e. the small wavenumber side) of the main peak. The high operating temperature of these DFB lasers could be used to investigate the temperature dependence of the thermal conductivity of the S1850 laser material. For this purpose, we measured another two series of

82


emission spectra under slightly different conditions. In the first experiment, the baseline of the electrical pulses was shifted from 0 mA to 50 mA. This is equivalent to having a 50 mA DC bias, on top of which we added a pulsed signal 50 mA smaller than before. For each temperature and a total current of 2.5 A, the wavelength was compared to the one obtained at normal conditions (no DC bias). In the second measurement, we increased the DC bias to 100 mA and measured again the emission wavelength for each temperature. The DC bias-induced heating caused a measurable wavelength shift of the DFB laser. For each given temperature, this wavelength shift (or temperature increase) remained at a constant value of −0.12 cm−1 for 50 mA DC bias (Vdevice = 2.5 V) and −0.36 cm−1 for 100 mA DC bias (Vdevice = 3.5 V). Since a constant temperature increase was observed for all investigated temperatures (when adding a certain DC bias), we can conclude that the thermal conductivity remained unchanged with increasing temperature. From the additional power to be dissipated and the observed temperature increase, we can calculate the thermal resistance using Rth = ∆ T /(Vdevice × IDC ); this yields Rth = 14.7 K/W or (taking into account the area of the laser) gth = 190 W/K cm2 . 4.4

High Power Junction Down Mounted Lasers

Different spectroscopic techniques in the mid-IR wavelength range have suffered from the absence of convenient semiconductor light sources. Lead salt lasers have existed for a long time, but their highest CW operating temperature could not be pushed beyond 220 K [59]. In addition, even at cryogenic temperatures, their output power remained relatively small, in the range of a couple of milliwatts [60]. Since QC lasers can work in continuous wave up to room temperature, they are very interesting light sources for spectroscopic applications. A variety of different experimental configurations and emission wavelengths has been published in the last couple of years, among them PA spectroscopy [29,42] and single-pass or multiple-pass absorption spectroscopy [50]. As mentioned above, the highest power output is possible if a DFB laser with InP top cladding is mounted with its active side directly on a copper holder. This is exactly what we did with S1810. The layer sequence of sample S1810, starting from the injection barrier, was 38/21/7/63/9/59/9/54/22/36/14/34/14/32/18/31/20/31. All thicknesses are in ˚ A, In0.52 Al0.48 As barrier layers are in bold, In0.53 Ga0.47 As well layers are in roman, and n-doped layers (Si, 3 × 1017 cm−3 ) are underlined. Figure 13 shows a series of L–I and I–V curves of a 26 µm wide and 1.5 mm long DFB laser taken from sample S1810. For this particular measurement, a duty cycle of 16% was used. The emitted average power was 71 mW at −30 ◦C and 15 mW at 60 ◦ C. At 1.5% duty cycle, we observed maximal slope efficiencies of 194 mW/A at −30 ◦ C and 93 mW/A at 60 ◦ C. The corresponding threshold currents were 2.3 A and 3.34 A; these values are equivalent to threshold current densities of 5.9 kA/cm2 and 8.55 kA/cm2 at


60

-30 °C

8 40

-30 °C

4

20

60 °C

0

60 °C

0

1

2

3 4 Current(A)

5

6

Average power [mW]

Voltage [V]

80

S1810c6 junction dn 16 % duty cycle

12

83

0

Fig. 13. Series of L–I and I–V curves of a 26 µm wide and 1.5 mm long DFB laser taken from the sample S1810 and at different temperatures between −30 ◦ C and 60 ◦ C. For this particular measurement, a duty cycle of 16% (45 ns pulse length/ 3.6 MHz pulse repetition frequency) was used

the respective temperatures. The characteristic temperature, T0 , which empirically describes the behavior of the threshold current as a function of the temperature, was 243 K. Exact matching between the Bragg and gain peak was achieved at 40 ◦ C; this led to a T0 value which was somewhat better than the one observed with the Fabry–Pérot laser fabricated from the same material (T0FP = 183 K). At 16% duty cycle, the threshold currents were 2.7 A for −30 ◦ C and 4.4 A for 60 ◦ C, giving rise to a slightly lower T0 value of 184 K. From the threshold current values at low and high duty cycle and the T0 at low duty, one can estimate the temperature increase ∆ T in the active region with respect to the copper heatsink. This calculation shows that for higher temperatures, the laser active region heats again faster than the heatsink (∆ T = 39 K at −30 ◦ C and ∆ T = 67 K at 60 ◦ C) [45,57]. This is also illustrated by the slightly elevated temperature tuning coefficient which can be derived from the spectra shown in Fig. 14. These spectra are measured at 16% duty cycle and the maximal output power for each temperature. The laser can be continuously tuned from 1046 cm−1 at −30 ◦C to 1038 cm−1 at 60 ◦ C with a tuning coefficient of 1/λ × ∆ λ/∆ T = −8.5 × 10−5 K−1 . Due to the overheating of the active region, this value is about 22% larger than expected from the temperature-induced refractive index change [21]. The spectral width of about 1 cm−1 is due to the duration of the electrical pulses; for shorter electrical pulses, we observed a proportional linewidth narrowing. The side mode suppression ratio was on the order of 20 dB for the entire investigated temperature range. In Fig. 15, we present an overview of all DFB lasers which have been described in this chapter. The figure maps peak and average output powers as a function of emission wavelength. In the inset, average power as a function of duty cycle is plotted. Filled circles represent peak powers and empty circles stand for average powers. The whole figure shows nicely how the performance

84


Intensity [a.u.]

100 10 60 °C

-30 °C

1 0.1 0.01 1030

1035

1040

1045

1050

1055

-1

Wavelength [cm ] Fig. 14. Series of emission spectra of a 26 µm wide and 1.5 mm long DFB laser taken from the sample S1810 and at different temperatures between −30 ◦ C and 60 ◦ C. Each spectrum was measured at the maximum output power at 16% duty cycle

1 0 0 1 0 1

S 1 8 4 0

S 1 8 6 9

S 1 8 5 0 8 0 A v e ra g e p o w e r [m W ]

O u tp u t p o w e r [m W ]

1 0 0 0

5

S 1 5 3 6

S 1 8 4 0

6 0

S 1 8 4 0

4 0

S 1 8 1 0

S 1 8 6 9

2 0 0

S 1 8 1 0 S 1 8 0 5

S 1 8 5 0

S 1 8 5 0

S 1 8 1 0

S 1 5 3 6 S 1 8 0 5 0

6

1 0

S 1 8 0 5

2 0 D u ty c y c le [% ]

7

3 0

8

S 1 5 3 6

9

1 0

1 1

W a v e le n g th [µ m ] Fig. 15. Map of all presented DFB lasers of this chapter. Peak (filled circles) and average (empty circles) power versus emission wavelength are shown. The inset contains the information of average power versus duty cycle

improved with increasing sample number (or over time). But it becomes also clear that, for example, the highest average powers do not necessarily correlate with the highest possible duty cycles. 4.5 Room Temperature Continuous Wave Operation of QC Lasers As already mentioned in the introduction, room temperature CW operation of a mid-infrared (9 µm) QC laser was achieved in summer 2001 [33]. This major breakthrough was possible thanks to different innovations which needed to


85

be successfully combined. First and foremost, the new types of active region design have led to lasers with room temperature pulsed threshold current densities on the order of 3 kA/cm2 . When processed appropriately in the narrow stripe buried heterostructure technique, this results in devices with pulsed threshold currents of 300 mA and below. The junction down mounting technique mentioned in the previous paragraph was then used together with diamond heatsinks and ZnSe/PbTe-based high-reflection facet coatings. Thanks to the latter, facet reflectivites of about 70% per facet were achieved; this allowed the use of very short laser cavities with a length of only 750 µm. The stripe width of the best devices was 12 µm resulting in relatively small amounts of heat to be dissipated. As shown in the L–I curves of Fig. 16, we observed 395 mA threshold current at 19 ◦ C heatsink temperature. A continuous output power of 13 mW was achieved under these conditions. The laser could be operated up to 39 ◦ C, at which temperature the threshold current increased to 515 mA and the output power dropped to 2 mW. Due to a small defect in the cavity, this laser was single-mode for all investigated temperatures and currents. This is shown in the inset of Fig. 16. The tuning of the lasing mode with temperature and input electrical power allowed us to determine the thermal conductivity of the mounted device. We found values on the order of 574 W/cm2 K. Based on the threshold current increase at higher temperatures and the pulsed threshold current, we were able to model the maximal possible operating temperature. The result of 48 ◦ C was in good agreement with experiment. The main reason for the much better performance of this laser compared to a normal ridge waveguide sample was the reduced thermally induced shear stress at the top corners of the ridge. In the conventional design, we computed a stress 15 19 ° C 24 ° C

6 Intensity [a.u.]

Voltage [V]

S1850d3-bh junction down

3

100 313 K

29 ° C

295 K

10

34 ° C

1 1091

1093

1095

1097

-1

Energy [cm ]

0

0

10

0.2

0.4 Current [A]

5

39 ° C

0.6

CW optical power [mW]

9

0

Fig. 16. L–I curves of a room temperature continuous wave operated QC laser at temperatures between 292 K and 312 K. The inset shows a series of emission spectra at different temperatures up to 313 K

86


of 22 MPa, whereas the much narrower buried heterostructure laser revealed only 3.6 MPa. Due to the large shear stress in case of a ridge waveguide sample, these lasers failed quite systematically at current densities of 4 kA/cm2 . With the buried heterostructure design, however, the maximal current density could be increased to 6 kA/cm2 without device failure. The laser shown in Fig. 16 could be driven at this current density for several hours without any noticeable degradation. When going to even narrower stripes, we hope to lower the threshold current and the temperature-induced shear stress sufficiently to reach operating temperatures on the order of 100 ◦ C.

5

Applications

In this last section, we would like to concentrate on the two most interesting applications of the devices presented in this work. While the use of QC lasers for spectroscopy is well known and also established, optical data transmission is a relatively new field of application. The latter became particularly interesting with the demonstration of room temperature continuous wave operation. 5.1

Photo-Acoustic Spectroscopy

As already mentioned in Sect. 4.4, single-mode QC lasers have become very attractive light sources for mid-infrared absorption spectroscopy [18,54]. They have already been used in different experimental configurations such as PA spectroscopy [30], and single-pass or multiple-pass absorption spectroscopy. For PA spectroscopy however, they have been operated at cryogenic temperatures which required quite an advanced experimental setup. The cryogenic operating conditions were necessary because room temperature performance used to be insufficient to result in low detection limits for PA spectroscopy. Recently, high performance DFB QC lasers, whose room temperature average power is sufficiently large for PA spectroscopy, were developed [45]. Up to now, most PA spectroscopy experiments were done with CO- or CO2 -lasers which emit IR radiation around 5 µm and 10 µm [49,61]. However, the discrete emission spectrum of these gas lasers can be problematic, especially if a gas absorbs between two emission lines. In such a case, high pressure CO2 -lasers are a possible but sophisticated solution [62]. Another possibility is the use of a light source based on difference frequency generation between a pump and a signal laser beam in a nonlinear optical medium [63]. The preferred choice, however, are QC DFB lasers operated at or near room temperature. Their emission wavelength tunes continuously over nearly 1% by simply changing the device temperature. If both design and operating conditions are carefully chosen, the emission peak can be swept across the absorption line of a particular gas. The QC lasers used during this work are single-mode DFB lasers with a surface grating and lateral current injection. But since the linewidth of


87

a pulsed laser is a critical parameter, we would like to present first some measurements of the linewidth dependence on electrical pulse-length and device temperature. Measurements using an FTIR with a long path length difference are shown in Fig. 17; they revealed a linewidth minimum of 0.048 cm−1 both at 3 ns and 10 ns, while seeing slightly larger linewidths when going to 5 ns and 15 ns. In a simple picture, the linewidth of a QC laser should be Fourier transform limited for short pulses and dominated by heating for long pulses, so that it can be written as ∆ ν(t) = C1 /t + C2 × t. The constant C1 = 0.0466 ns/cm is obtained by calculating the linewidth of transform limited rectangular pulses, while C2 was fitted to the experimental data. Deviations from the theory can be attributed to two major causes. First of all, we could not control the shape of the electrical pulse with high accuracy. In fact, a measurement of the optical pulse shape suggested that electrical ringing occurred, especially at a pulse duration of 5 ns. As shown in Fig. 17, this resulted in a much larger error bar for this particular point. Secondly, a Fourier transform limited linewidth can be seen only if the laser is operated near the gain maximum, where the linewidth enhancement factor α is close to zero. Depending on the detuning between Bragg peak and gain peak, the α-factor can become as large as 0.5. Based on this effect, Paiella et al. recently demonstrated the operation of a QC laser in the mode-locked regime [64]. Based on these results, we decided to work with pulse widths of 25 ns and a duty cycle of 3–4% for both lasers used in the following measurements. The PA cell used for these experiments is a 12 cm long cylindrical chamber (diameter d = 5 cm) with a 6 cm long buffer volume (d = 12 cm) on each side. The laser enters and leaves the cell through two ZnSe Brewster windows. This whole setup is embedded between two spherical concave mirrors, which allow for 36 passes of the light through the PA cell [65]. With an overall length of

FTIR spectrometer grating spectrometer calculation

0.5

-1

Linewidth [cm ]

0.6

0.4 0.3 0.2 0.1 0

0

20

40 60 Pulse length [ns]

80

Fig. 17. Linewidth versus electrical pulse length for a QC DFB laser emitting at 10.4 µm. For the grating spectrometer data, there is a linewidth saturation at 0.16 cm−1 for pulse lengths below 20 ns

88


70 cm and the number of passes, we end up with a total optical path length of 24 m (15 m inside the PA cell). Because of the limited transmission per pass (96.1%), this results in a power enhancement factor of 19 (instead of 36 for 100% transmission). In the center of the PA cell, a radial microphone array with 16 microphones increases the signal to noise ratio (SNR) of the setup by a factor of 4. By chopping the beam mechanically with the first longitudinal resonance frequency of the cell (f = 1250 Hz), one gains another factor of 70 (i.e. the Q-factor of the resonance) in the strength of the PA signal [66]. As a result of these three measures, we increased the PA signal by a factor of 64 000 compared to a non-resonant single-pass arrangement. The light of the QC laser was collected with a ZnSe lens (f/1.0), collimated and then directed collinearly with a HeNe laser beam for alignment purposes. Within the PA-cell, the laser beam was left collimated; behind the cell, a pyro-electric detector recorded its power in order to normalize the PA signal. The set-up is fully computer-controlled including the control of the laser wavelength, the stabilization of the chopper frequency on the resonance frequency of the PA cell, the data recording, etc. In Fig. 18, several absorption spectra of ammonia diluted in synthetic air in the wavelength range between 965 and 968 cm−1 are presented. Measurements at atmospheric pressure resulted in a relatively noisy spectrum. This is presumably due to coupling of acoustic noise into the cell. For 400 mbar, the best SNR was achieved; at this pressure and concentration, the laser linewidth equals approximately that of the ammonia absorption features. At lower pressure, we observed the convolution of the laser line-shape with the

PA signal [mV/W]

100

10

1

NH3 HITRAN (a.u.) NH3 PA @ 0.6 ppmV / 950 mbar NH3 PA @ 10 ppmV / 400 mbar NH3 PA @ 10 ppmV / 200 mbar

966.0 966.5 967.0 967.5 968.0 -1

Wavenumbers [cm ] Fig. 18. Comparison between the absorption spectra of NH3 based on the HITRAN database and by PA spectroscopy (10 ppmV/400 mbar, 10 ppmV/200 mbar, 600 ppb/950 mbar, all buffered in synthetic air, 300 K)


89

absorption spectrum owing to the smaller pressure broadening of the ammonia lines. Since the microphone responsivity decreases with lower pressure, the PA signal decreased as well; and this resulted again in a smaller SNR of the measurement. A value of 300 ppbV at 400 mbar was identified as the detection limit for this particular configuration (with SNR = 3). Good agreement with HITRAN data was achieved. 5.2

Optical Data Link Using a QC Laser

As demonstrated in Sect. 4, mid-IR QC lasers have reached a level of maturity that allows one to think of applications other than spectroscopy [67]. Today’s state-of-the-art devices can be operated CW above room temperature [33,54]. While most potential applications lie in the field of optical spectroscopy [29], there are also some other interesting areas of deployment. These include, for example, free-space optical data transmission [68]. In contrast to fiber optical telecommunications, this technique has the advantage of not requiring additional cables to be buried in the ground. Especially in urban areas where lots of fiber optical connections already exist, fast free-space optical data links could be quite convenient. QC lasers are very suitable for such applications because their emission wavelength can be chosen in the so-called atmospheric window regions, i.e. around 5 µm and 10 µm. In addition, the fast internal lifetimes in the device should allow for reasonable modulation frequencies of up to 5–10 GHz. Recently, Martini et al. published results of an optical data link using a high-speed modulated, liquid-nitrogen cooled QC laser over a distance of 70 m and under laboratory conditions [27]. They also succeeded in transmitting a video image via a common TV channel frequency. Since this experiment was carried out within the building, one of the main benefits of using QC lasers, namely having an emission wavelength which is barely affected by atmospheric conditions like rain or fog, was not yet demonstrated. In addition, the use of liquid nitrogen cooled equipment on both sides makes the technique less attractive for applications in the field. In order to take full advantage of our existing QC laser technology, we present here an optical data link between two different buildings separated by about 350 m and using a Peltier-cooled QC laser as well as a room temperature HgCdTe detector [28]. On the emitter side, we used a 3 mm long 9.3 µm multimode QC laser mounted in a Peltier-cooled, temperature-stabilized aluminum box (Alpes Lasers SA) and an f/0.8 Ge lens with 37.5 mm diameter to collimate the laser beam. The device was held at a temperature of 258 K, operated at nearly 50% duty cycle, and pulsed at different repetition frequencies. Using a bias-T, the laser was driven simultaneously at a constant current of 2 A (which corresponds to 0.72 × Ith ) and a 10 W radio frequency signal of up to 350 MHz. On the receiver side, we employed a mirror telescope with a diameter of 16 cm and a focal distance of 62.5 cm, a fast room temperature HgCdTe detector, and a 15 dB small signal amplifier for the detection of the incoming signal.

90


Fig. 19. Schematic drawing of the optical data link in the free atmosphere between two University buildings. The picture was taken from the Physics Institute, and the arrow in the photo shows the ‘targeted’ building in a distance of 350 m

As schematically shown in Fig. 19, a 1 mW red semiconductor laser pointer was directed collinearly with the QC laser beam to facilitate alignment. In the first stage, we aligned the two laser beams in the lab; then the two beams were bounced off a steering mirror and directed towards the other building where the telescope was installed. The steering mirror could be tilted and rotated by manual micrometer screws. The angular accuracy of this kind of beam steering was about 3×10−5 rad. In the distance of the building with the telescope, this corresponded to roughly 1 cm. In order to align the infrared beam correctly into the telescope, the voltage corresponding to the optical signal was transmitted to the emitter side via Ethernet. At a temperature of 258 K and for 50% duty cycle, the threshold current of the QC laser used was 2.7 A (jth = 3.0 kA/cm2 ). For the maximal injection current of 3.2 A, we observed an average output power of 14 mW. When repeating the power measurement at the other end of the transmission line and at 300 MHz, we still obtained 1.9 mW average power for clear sky conditions; the measured peak power was thus on the order of 3.8 mW. With foggy conditions, the visibility range dropped to a value below 100 m. However, the average power signal decreased by barely 20%. Since the loss was thus on the order of 8 dB, these numbers show clearly the advantage of working at an emission wavelength in an atmospheric window region.


91

Average power [mW]

10

Signal [mV]

5

0

4 3 2 1 0

100 150 200 250 300 350 Frequency [MHz]

-5

-10

0

25

50 Time [ns]

75

100

Fig. 20. Transmitted signal at a frequency of 300 MHz. The inset shows the frequency response of the transmitted signal with a substantial decrease of power above 330 MHz

As shown in Fig. 20, the typical transmitted signal consisted of a stream of almost sinusoidal pulses with a repetition frequency of up to 330 MHz and a pulse width of 1.5 ns. For 300 MHz, a noise level of roughly 0.25 mW (peak-to-peak value, before amplification) was observed; together with the transmitted peak power mentioned above, this corresponds to a signal to noise ratio of 15. It should√be noted, however, that the detector had a figure of merit of D = 2.2×107 cm2 Hz/W. From √ this value, we can calculate the noise equivalent power (NEP) using D = ∆ f × A/NEP. With A = 0.625× 10−2 cm2 being the detector area and ∆ f = 1 GHz the bandwidth of the detector system, we end up with a NEP (amplitude) of 0.11 mW. The signal to noise ratio is therefore entirely limited by the detectivity of the detector. In a second experiment, whose result is shown as an inset in Fig. 20, we measured the transmitted average power as a function of the pulse repetition rate. It is evident from the figure that the power has a resonance at 325 MHz, and then drops quickly to quite small values. A naive calculation shows that 325 MHz corresponds roughly to the electrical resonance frequency of the laser. The parasitic capacitance defined by the large contact pad area (3 × 0.5 mm2 ) is about 150 pF; together with the resistance of the low impedance line supplying the current to the laser (approximately 4 Ω), we end up with an RC time constant of 600 ps. The maximum modulation frequency is thus about 250 MHz, in fair agreement with the experimental value. In a different experimental configuration, we used the QC laser to optically transmit data between two computers. This data link was set up over

92


a distance of 10 m between two optical tables and put together entirely within one laboratory building. On the emitter side, we used the serial port of the first PC and an RS232/TTL signal converter to produce a TTL modulation signal at 115 kbit/s. This signal was used to electrically gate the continuous stream of laser pulses. At the receiver side, the signal was low-pass filtered, amplified, and brought into rectangular shape again with a comparator. After this, we used a timer logic circuit triggered by the positive slopes of the single laser pulses to revert to the initial gate signal, and the final TTL/RS232 converter made the signal compatible to the serial port of the second computer. By this technique, we were able to communicate optically between the two computers at the standard transmission speed of 9.6 kbit/s, and also at the highest possible speed of 115 kbit/s. When decreasing the laser power, the link still worked successfully down to a measured signal to noise ratio of 3.

6

Conclusions

We believe that this work clearly shows the big application potential of QC lasers. More effort is necessary in order to get reliable continuous wave operation at room temperature and also at wavelengths other than the one initially demonstrated. In any case, we foresee a brilliant future for this rather new type of semiconductor laser. We would like to thank the following people for their contributions to this work: Mattias Beck (University of Neuchˆ atel), Ursula Oesterle, Marc Ilegems (both EPFL) and Emilio Gini and Hans Melchior (both ETHZ) for crystal growth, Thierry Aellen for device processing and cleaving, Daniel Varidel, Jean-Pierre Bourquin, Michel Rochat, and Stéphane Blaser (all University of Neuchˆ atel) for technical support. In addition, the financial support of the Swiss National Science Foundation and the European Science Foundation are gratefully acknowledged.

References 1. J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A. L. Hutchinson, A. Y. Cho: Quantum cascade laser, Science 264, 553–556 (1994) 61 2. J. Faist, F. Capasso, C. Sirtori, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, S. Chu, A. Y. Cho: High power mid-infrared (λ ∼ 5 µm) quantum cascade lasers operating above room temperature, Appl. Phys. Lett. 68, 3680–3682 (1996) 61, 62 3. C. Sirtori, J. Faist, F. Capasso, D. L. Sivco, A. L. Hutchinson, S. Chu, A. Y. Cho: Continuous wave operation of mid-infrared (7.4–8.6 µm) quantum cascade lasers up to 110 K temperature, Appl. Phys. Lett. 68, 1745–1747 (1996) 61 4. R. F. Kazarinov, R. A. Suris: Possibility of the amplification of electromagnetic waves in a semiconductor with a superlattice, Sov. Phys. Semicond. 5, 707–709 (1971) 61 5. T. Ando, A. B. Fowler, F. Stern: Electronic properties of two-dimensional systems, Rev. Mod. Phys. 54, 437–446 (1982) 61


93

6. L. C. West, S. J. Eglash: First observation of an extremely large-dipole infrared transition within the conduction band of a GaAs quantum well, Appl. Phys. Lett. 46, 1156–1158 (1985) 61 7. R. Ferreira, G. Bastard: Evaluation of some scattering times for electrons in unbiased and biased single-and multiple-quantum-well structures, Phys. Rev. B 40, 1074–1086 (1989) 61 8. F. Capasso, K. Mohammed, A. Y. Cho: Resonant tunneling through double barriers, perpendicular quantum transport phenomena in superlattices, and their device applications, IEEE J. Quantum Electron. 22, 1853–1869 (1986) 61 9. A. Cho: Molecular Beam Epitaxy (AIP, Woodbury, NW 1994) 61 10. H. Morkoc, C. Hopkins, C. A. Evans, Jr., A. Y. Cho: Chromium and tellurium redistribution in GaAs and Al0.3 Ga0.7 As grown by molecular beam epitaxy, J. Appl. Phys. 51, 5986–5991 (1980) 61 11. M. Y. Yen, B. F. Levine, C. G. Bethea, K. K. Choi, A. Y. Cho: Molecular beam epitaxial growth and optical properties of InAs1x Sbx in 8–12 µm wavelength range, Appl. Phys. Lett. 50, 927–929 (1987) 61 12. R. R. Saxena, J. E. Fouquet, V. M. Sardi, R. L. Moon: High quality InP layers grown by organometallic vapor phase epitaxy using tertiarybutylphosphine and phosphine, Appl. Phys. Lett. 53, 304–306 (1988) 61 13. T. Y. Wang, K. L. Fry, A. Persson, E. H. Reihlen, G. B. Stringfellow: Atomic steps in thin GaInAs/lnP quantum-well structures grown by organometallic vapor phase epitaxy, J. Appl. Phys. 63, 2674–2680 (1988) 61 14. J. Faist, F. Capasso, C. Sirtori, D. L. Sivco, A. L. Hutchinson, A. Y. Cho: Continuous wave operation of a vertical transition quantum cascade laser above T = 80 K, Appl. Phys. Lett. 67, 3057–3059 (1995) 61 15. C. Gmachl, A. Tredicucci, F. Capasso, A. L. Hutchinson, D. L. Sivco, J. N. Baillargeon, A. Y. Cho: High-power λ ∼ 8 µm quantum cascade laser with near optimum performance, Appl. Phys. Lett. 72, 3130–3132 (1998) 61 16. J. Faist, A. Tredicucci, F. Capasso, C. Sirtori, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho: High-power continuous-wave quantum cascade lasers, IEEE J. Quantum Electron. 34, 336–343 (1998) 62 17. C. Sirtori, J. Faist, F. Capasso, D. L. Sivco, A. L. Hutchinson, A. Y. Cho: Pulsed and continuous-wave operation of long wavelength infrared (λ = 9.3 µm) quantum cascade lasers, IEEE J. Quantum Electron. 33, 89–93 (1997) 62 18. C. Gmachl, F. Capasso, J. Faist, A. L. Hutchinson, A. Tredicucci, D. L. Sivco, J. N. Baillargeon, S. Chu, A. Y. Cho: Continuous-wave and high-power pulsed operation of index-coupled distributed quantum cascade laser at λ ∼ 8.5 µm, Appl. Phys. Lett. 72, 1430–1432 (1998) 62, 63, 86 19. J. Faist, C. Sirtori, F. Capasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho: High-power long-wavelength (λ = 11.5 µm) quantum cascade lasers operating above room temperature, IEEE Photon. Technol. Lett. 10, 1100–1102 (1998) 62, 66, 69, 76 20. C. Gmachl, A. M. Sergent, A. Tredicucci, F. Capasso, A. L. Hutchinson, D. L. Sivco, J. N. Baillargeon, S. Chu, A. Y. Cho: Improved CW operation of quantum cascade lasers with epitaxial-side heat-sinking, IEEE Photon. Technol. Lett. 11, 1369–1371 (1999) 62 21. J. Faist, C. Gmachl, F. Capasso, C. Sirtori, D. L. Sivco, J. N. Baillargeon, A. Y. Cho: Distributed feedback quantum cascade lasers, Appl. Phys. Lett. 70, 2670– 2672 (1997) 62, 63, 80, 83

94


22. C. Gmachl, J. Faist, J. N. Baillargeon, F. Capasso, C. Sirtori, D. L. Sivco, S. Chu, A. Y. Cho: Complex-coupled quantum cascade distributed-feedback laser, IEEE Photon. Technol. Lett. 9, 1090–1092 (1997) 62, 63, 73 23. C. Sirtori, P. Kruck, S. Barbieri, P. Collot, J. Nagle, M. Beck, J. Faist, U. Oesterle: GaAs/Alx Ga1−x As quantum cascade lasers, Appl. Phys. Lett. 73, 3486–3488 (1998) 62 24. C. Gmachl, F. Capasso, E. Narimanov, J. U. N¨ ockel, D. Stone, J. Faist, D. L. Sivco, A. Y. Cho: High-power directional emission from microlasers with chaotic resonators, Science 280, 1556–1564 (1998) 62 25. C. Gmachl, H. M. Ng, A. Y. Cho: Intersubband absorption in degenerately doped GaN/Alx Ga1−x N coupled double quantum wells, Appl. Phys. Lett. 79, 1590–1592 (2001) 62 26. D. Hofstetter, L. Diehl, J. Faist, W. J. Schaff, J. Hwang, L. F. Eastman, C. Zellweger: Midinfrared intersubband absorption on AlGaN/GaN-based highelectron-mobility transistors, Appl. Phys. Lett. 80, 2991–2994 (2002) 62 27. R. Martini, C. Gmachl, J. Falciglia, F. G. Curti, C. G. Bethea, F. Capasso, E. A. Whittaker, R. Paiella, A. Tredicucci, A. L. Hutchinson, D. L. Sivco, A. Y. Cho: High-speed modulation and free-space optical audio/video transmission using quantum cascade lasers, IEE Electron. Lett. 37, 102–103 (2001) 62, 89 28. S. Blaser, D. Hofstetter, M. Beck, J. Faist: Free-space optical data link using Peltier-cooled quantum cascade laser, IEE Electron. Lett. 37, 778–780 (2001) 62, 89 29. D. Hofstetter, M. Beck, J. Faist, M. N¨ agele, M. W. Sigrist: Photoacoustic spectroscopy with quantum cascade distributed-feedback lasers, Opt. Lett. 26, 887– 889 (2001) 62, 82, 89 30. B. A. Paldus, T. G. Spence, R. N. Zare, J. Oomens, F. Harren, D. H. Parker, C. Gmachl, F. Capasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho: Photoacoustic spectroscopy using quantum-cascade lasers, Opt. Lett. 24, 178–180 (1999) 62, 86 31. B. Ishaug, W. Hwang, J. Um, B. Guo, H. Lee, C. Lin: Continuous-wave operation of a 5.2 µm quantum-cascade laser up to 210 K, Appl. Phys. Lett. 79, 1745–1747 (2001) 62 32. D. Hofstetter, M. Beck, T. Aellen, J. Faist, U. Oesterle, M. Ilegems, E. Gini, H. Melchior: Continuous wave operation of a 9.3 µm quantum cascade laser on a Peltier cooler, Appl. Phys. Lett. 78, 1964–1966 (2001) 62, 80 33. M. Beck, D. Hofstetter, T. Aellen, J. Faist, U. Oesterle, M. Ilegems, E. Gini, H. Melchior: Continuous-wave operation of a mid-infrared semiconductor laser at room-temperature, Science 295, 301–305 (2002) 62, 84, 89 34. H. Page, C. Becker, A. Robertson, G. Glastre, V. Ortiz, C. Sirtori: 300 K operation of a GaAs-based quantum-cascade laser at λ ∼ 9 µm, Appl. Phys. Lett. 75, 3529–3531 (2001) 62 35. S. Anders, W. Schrenk, E. Gornik, G. Strasser: Room-temperature emission of GaAs/AlGaAs superlattice quantum-cascade lasers at 12.6 µm, Appl. Phys. Lett. 80, 1864–1866 (2002) 62 36. W. Schrenk, N. Finger, S. Gianordoli, L. Hvozdara, G. Strasser, E. Gornik: GaAs/AlGaAs distributed feedback quantum cascade lasers, Appl. Phys. Lett. 76, 253–255 (2000) 62, 63 37. R. K¨ ohler, A. Tredicucci, F. Beltram, H. E. Beere, E. H. Linfield, A.Giles, G. Davies, D. A. Ritchie, R. C. Iotti, F. Rossi: Terahertz heterostructuresemiconductor laser, Nature 417, 156–159 (2002) 62


95

38. M. Rochat, L. Ajili, H. Willenberg, J. Faist, H. Beere, G. Davies, E. Linfield, D. Ritchie: Low threshold terahertz quantum cascade lasers, Appl. Phys. Lett. 81, 1381–1383 (2002) 62 39. D. Hofstetter, J. Faist, M. Beck, A. M¨ uller, U. Oesterle: Demonstration of highperformance 10.16 µm quantum cascade distributed feedback lasers fabricated without epitaxial regrowth, Appl. Phys. Lett. 75, 665–667 (1999) 62, 63, 68, 74, 76, 77 40. G. A. Evans, D. P. Bour, N. W. Carlson, J. M. Hammer, M. Lurie, J. K. Butler, S. L. Palfrey, R. Amantea, L. A. Carr, F. Z. Hawrylo, E. A. James, J. B. Kirk, S. K. Liew, W. F. Reichert: Coherent, monolithic two-dimensional strained InGaAs/GaAs quantum well laser arrays using grating surface emission, Appl. Phys. Lett. 55, 2721–2723 (1989) 63 41. P. Repond, T. Marty, M. W. Sigrist: A continuously tunable CO2 laser for photoacoustic detection of trace gases, Helv. Phys. Acta 65, 828–829 (1992) 63 42. M. N¨ agele, D. Hofstetter, J. Faist, M. W. Sigrist: Low power quantum-cascade laser photoacoustic spectrometer for trace-gas monitoring, Analytical Sci. 17, 497–499 (2001) 63, 82 43. P. Hess: Principles of photoacoustic and photothermal detection of gases, in Principles and Perspectives of Photothermal and Photoacoustic Phenomena IV, ed. by A. Mandelis (Elsevier, New-York 1992) Chap. 4 63 44. R. K¨ ohler, C. Gmachl, A. Tredicucci, F. Capasso, D. L. Sivco, S. Chu, A. Y. Cho: Single-mode tunable pulsed, and continuous wave quantum-cascade distributed feedback lasers at λ ∼ 4.6–4.7 µm, Appl. Phys. Lett. 76, 1092–1094 (2000) 63 45. D. Hofstetter, T. Aellen, M. Beck, J. Faist: High average power first-order distributed feedback quantum cascade lasers, IEEE Photon. Technol. Lett. 12, 1610–1612 (2000) 63, 79, 83, 86 46. J. Faist, F. Capasso, D. L. Sivco, A. L. Hutchinson, S. Chu, A. Y. Cho: Short wavelength (λ ∼ 3.4 µm) quantum cascade laser based on strained compensated InGaAs/AlInAs, Appl. Phys. Lett. 72, 680–682 (1998) 63, 66 47. G. Scamarcio, F. Capasso, C. Sirtori, J. Faist, A. L. Hutchinson, D. L. Sivco, A. Y. Cho: High-power infrared (8-micrometer wavelength) superlattice lasers, Science 276, 773–776 (1997) 63, 66 48. A. Tahraoui, A. Matlis, S. Slivken, J. Diaz, M. Razeghi: High-performance quantum cascade lasers (λ ∼ 11 µm) operating at high temperature (T ≥ 425 K), Appl. Phys. Lett. 78, 416–418 (2001) 63 49. S. Bernegger, M. W. Sigrist: CO-laser photoacoustic spectroscopy of gases and vapours for trace gas analysis, Infrared Phys. 30, 375–429 (1990) 63, 86 50. B. Mizaikoff, C. S. Murthy, M. Kraft, V. Pustogow, A. M¨ uller, D. Hofstetter, J. Faist, N. Croitoru: Novel mid-infrared gas sensors based on hollow waveguides and quantum cascade lasers, in Proc. PITTCON 2000 (New Orleans, LA 2000) 63, 82 51. R. K¨ astle, M. W. Sigrist: CO-laser photoacoustic spectroscopy on the dimerization of fatty acid molecules, J. Phys. IV (Paris) Colloque C7, 491–494 (1994) 63 52. A. M¨ uller, M. Beck, J. Faist, U. Oesterle: Electrically tunable, roomtemperature quantum-cascade lasers, Appl. Phys. Lett. 75, 1509–1511 (1999) 66, 69

96


53. A. Tredicucci, F. Capasso, C. Gmachl, D. L. Sivco, A. L. Hutchinson, A. Y. Cho: High performance interminiband quantum cascade lasers with graded superlattices, Appl. Phys. Lett. 73, 2101–2103 (1998) 66 54. D. Hofstetter, M. Beck, T. Aellen, J. Faist: High-temperature operation of distributed feedback quantum-cascade lasers at 5.3 µm, Appl. Phys. Lett. 86, 396–398 (2001) 66, 78, 86, 89 55. C. H. Wu, P. S. Zory, M. A. Emanuel: Contact reflectivity effects on thin p-clad InGaAs single quantum-well lasers, IEEE Photon. Technol. Lett. 6, 1427–1429 (1994) 68 56. R. L. Thornton, W. J. Mosby, H. F. Chung: Surface skimming buried heterostructure laser with applications to optoelectronic integration, Appl. Phys. Lett. 59, 513–515 (1991) 68 57. D. Hofstetter, J. Faist, M. Beck, U. Oesterle: Surface-emitting 10.1 µm quantum-cascade distributed feedback lasers, Appl. Phys. Lett. 75, 3769–3771 (1999) 74, 76, 77, 83 58. W. Streifer, R. D. Burnham, D. R. Scifres: Radiation losses and longitudinal mode selection in distributed feedback lasers, IEEE J. Quantum Electron. 12, 737–739 (1976) 75 59. M. Tacke: New developments and applications of tunable IR lead salt lasers, Infrared Phys. Technol. 36, 447–463 (1995) 82 60. G. Springholz, T. Schwarzl, M. Aigle, H. Pascher, W. Heiss: 4.8 µm vertical emitting PbTe quantum-well lasers based on high-finesse EuTe/Pb1−x Eux Te microcavities, Appl. Phys. Lett. 76, 1807–1809 (2000) 82 61. A. Olafsson, M. Hammerich, J. B¨ ulow, J. Henningsen: Photoacoustic detection of NH3 in power plant emission with a CO2 laser, Appl. Phys. B 49, 91–97 (1989) 86 62. P. Repond, M. W. Sigrist: Continuously tunable high pressure CO2 laser for spectroscopic studies on trace gases, IEEE J. Quantum Electron. 32, 1549– 1556 (1996) 86 63. D. Richter, D. G. Lancaster, R. F. Curl, W. Neu, F. K. Tittel: Compact midinfrared trace gas sensor based on difference-frequency generation of two diode lasers in periodically poled LiNbO3 , Appl. Phys. B 67, 347–349 (1998) 86 64. R. Paiella, F. Capasso, C. Gmachl, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho, H. C. Liu: Self-mode-locking of quantum cascade lasers with giant ultrafast optical nonlinearities, Science 290, 1739–1742 (2000) 87 65. D. Herriot, H. Kogelnik, R. Kompner: Off-axis paths in spherical mirror interferometers, Appl. Opt. 3, 523–525 (1964) 87 66. A. Karbach, P. Hess: Photoacoustic signal in a cylindrical resonator: Theory and laser experiments for CH4 and C2 H6 , J. Chem. Phys. 84, 2945–2967 (1986) 88 67. M. Beck, J. Faist, U. Oesterle, M. Ilegems, E. Gini, H. Melchior: Buried heterostructure quantum cascade lasers with a large optical cavity waveguide, IEEE Photon. Technol. Lett. 12, 1450–1452 (2000) 89 68. N. Mustafa, L. Pesquera, C. Cheung, K. A. Stone: Terahertz bandwidth prediction for amplitude modulation response of unipolar intersubband semiconductor lasers, IEEE Photon. Technol. Lett. 11, 527–529 (1999) 89

Index

absorption – loss, 68 ammonia, 88 beam steering, 90 Bragg reflector, 75 broad ridge waveguides, 77 buried heterostructure design, 70 buried heterostructure technique, 85 CO laser, 86 CO2 laser, 86 comparison grating emission/facet emission, 74 de-tuned grating, 73 diamond heatsink, 71, 85 double QW gain region, 66 double-phonon-resonance gain region, 66 duty cycle, 71 edge-emitting laser, 72 electrical passivation, 69 extraction efficiency, 66 far field, 74 Fourier components, 74 Fourier transform IR spectrometer, 71 Fourier transform limit, 87 free carrier absorption – loss, 73 free-space optical data link, 89

Herriot multipass cell, 87 HgCdTe – detectors, 71 high speed modulation, 89 high thermal conductivity, 70 HITRAN database, 88 holography, 69 index coupling, 74 injection efficiency, 66 injector region, 64 interfacial roughness scattering, 68 intersubband electron scattering time, 67 intersubband transition, 64 junction down mounting, 70 laser – CO, 86 – CO2 , 86 – quantum cascade (QCL), 64–68 – surface-emitting, 74 – surface-skimming, 68 lateral current injection, 68 loss coupling, 74 microphone array, 87 microphone responsivity, 88 molecular beam epitaxy growth, 68 overgrown grating, 70 overheating of active region, 80

gain region design, 64 grating coupling coefficient, 68 grating spectrometer, 71

Peltier cooling, 89 photoacoustic spectroscopy (PAS), 87 processing steps, 68

Henry linewidth parameter, 87

quantum cascade laser (QCL), 64–68

98

Index

room temperature CW operation, 84 shear stress, 85 spectroscopy – photoacoustic (PAS), 87 stopband, 73, 76 – -width, 73, 76 strain-compensated material, 67, 77 subthreshold, 72 superlattice – gain region, 66 surface-emitting laser, 74 surface-skimming laser, 68

telecommunication, 89 temperature difference, 80 thermal conductivity, 85 – measurement, 81 thermal management, 70 thermal roll-over, 78 three QW gain region, 66

waveguide loss, 68 wavelength tuning, 76 wet chemical etching, 69

Mid-IR Difference Frequency Generation Cornelia Fischer and Markus W. Sigrist Swiss Federal Institute of Technology (ETH), Institute of Quantum Electronics, Laser Spectroscopy and Sensing Laboratory, Hoenggerberg, HPF, 8093 Z¨ urich, Switzerland {Cornelia.Fischer,Sigrist}@iqe.phys.ethz.ch Abstract. The availability of continuously tunable narrow-band laser sources emitting in the mid-infrared region of the electromagnetic spectrum between 2.5 µm and 15 µm is important for a large variety of applications. One way to achieve both broad tunability and narrow linewidth is frequency conversion in a nonlinear optical material, in particular difference frequency generation (DFG). After a brief introduction to the requirements of tunable mid-infrared sources, we present an overview of the theoretical background of frequency conversion processes. Emphasized are phase-matching issues, as well as material and design considerations to be observed when implementing a DFG laser source. The variety of possible combinations of pump/signal sources and nonlinear crystals is illustrated by numerous reports found in the literature. Finally, design issues are discussed for a practical example of a DFG laser source wherein the experimental arrangement and system performance, as well as tuning characteristics, are treated. Examples of spectroscopic applications reflect the advantages such as broad and continuous tuning range, narrow linewidth, and compact set-up of a DFG source, but also the drawbacks such as relatively low output power and high costs. The chapter is completed by some concluding remarks and a short outlook.

1

Introduction to Tunable Mid-IR Laser Sources

For many applications in the broad field of spectroscopy, e.g., for trace-gas sensing, the availability of a tunable laser source in the mid-IR region of the electromagnetic spectrum would be of great advantage because most small organic and inorganic molecules have their strong fundamental absorption lines in this region [1]. This is illustrated in Fig. 1 where the absorption ranges of some important functional groups of molecules are plotted for the wavelength range between 2 µm and 20 µm (Fig. 1, top), whereas strong fundamental absorption features of a few selected molecules are listed at the bottom of Fig. 1. The requirements for an appropriate laser source are manifold, depending on the application: broad tunability, narrow linewidth, sufficient laser power, room-temperature operation and robustness. First of all, a broad and preferably continuous tuning range, as well as a narrow linewidth, are prerequisites for multicomponent measurements while maintaining high selectivity. This is illustrated in Fig. 2a,b where the I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 97–143 (2003) c Springer-Verlag Berlin Heidelberg 2003

98

2

Cornelia Fischer and Markus W. Sigrist

4

6

8

10

12 14 16 Wavelength in µm

18

alkane alkene alkyne / allene aromatic homocyclic alcohol / phenol ethers organo - sulphur organo - halogen organo - phosphorus CO2 NO2 SO2 O3 NH3 CH4 20 22 24

Fig. 1. Absorption ranges of some important functional groups of molecules (top) and of selected molecules of environmental importance (bottom)

spectra of various substances at different concentrations calculated using the HITRAN database [2] are plotted in the limited wavelength range between 3.2 µm and 3.7 µm. Figure 2a shows the single spectra of four gases. The overlapping absorption features imply the need for tunable narrow-band sources to guarantee selectivity. Figure 2b is the compound spectrum of the same four gases as in Fig. 2a plus an additional two gases at the indicated concentrations. The arrows mark unique absorption lines for each one of the six gases of the mixture. Furthermore depending on the used absorption detection scheme, sufficient laser power (at µW to mW level or above) is needed to achieve high sensitivity. Finally, (near)-room temperature operation and a compact set-up are preferred for potential applications in industry and for in situ measurements in the field. There have been many attempts to meet all or at least some of these requirements. Each system has it advantages, but also its disadvantages. For instance, CO2 lasers offer high output power and a large tuning range (9.2– 10.8 µm for 12 CO2 ). However, the tunability is not continuous since the emission spectrum coincides with the CO2 emission lines [3]. This disadvantage can be overcome by operating the laser at higher gas pressure ( 10 bar) when continuous tunability is achieved, yet at the loss of CW operation [4]. Similarly, CO lasers emitting in the 5–6 µm wavelength range, with overtone operation extendable to 2.7 µm only offer discrete line tunability [5]. Both these well-established gas lasers have successfully been employed in sensing applications, yet their line tunability sometimes limits the applications

Transmission for 1 m pathlength

Mid-IR Difference Frequency Generation

99

1.0000

0.9995

0.9990 NO2 N2O H2CO CH4

0.9985

0.9980

3300

3400 3500 Wavelength in nm

3600

a) Single spectra of 4 different gases

Transmission for 1 m pathlength

1.000

NO2 3.44 ppm

0.995 CH4 357 ppb

0.990

C2H6 357 ppb 0.985

N2O 51.3 ppm H2CO 2.26 ppm

H2O 0.25 % 0.980

3300

3400 3500 Wavelength in nm

3600

b) Compound spectrum of 6 different gases

Fig. 2. (a) Overview spectra of four different trace gases absorbing around 3.3 µm. (b) Compound spectrum of six gases at the concentrations indicated. In addition to the four gases shown in (a), H2 O and C2 H6 are included

because the laser lines do not always overlap with fundamental molecular absorption lines of interest. Quantum cascade lasers (QCL) can be tailored to emit in a wide range of mid-IR wavelengths [6,7]. The wavelength region accessible with QCLs ranges from 3.5 µm to about 17 µm [8], and recent progress has shifted the limit further to the infrared region, currently to 66 µm [9]. With their compact size, operation temperature up to room temperature and average laser power of several mW, these devices might have a promising future in sensing applications [10,11,12]. Currently, however, the tuning range of a single device is limited to a few cm−1 .

100


Lead salt diode lasers are based on ternary IV–VI compounds. Depending on their composition, their wavelength range can be selected throughout the mid-infrared region between 3 µm and 30 µm [13,14,15]. Mode-hop-free tuning, though, is limited to ≈ 1 cm−1 . In addition, these devices require cryogenic cooling for CW operation, and the power is usually around 0.1 mW. In the near-infrared region, external cavity diode lasers offer large continuous tuning ranges (up to 100 nm) and a narrow linewidth. These systems, however, operate only at wavelengths up to approximately 2 µm and are thus only suited to access the much weaker overtone and combination bands of molecular absorption [16]. Interestingly, there exists a gap with respect to available widely and continuously tunable laser sources between about 2.8 µm and 4 µm. There are to date no active laser materials known emitting in that particular wavelength region. As shown in Figs. 1 and 2, though, there are major molecular absorption features in this wavelength range. Thus, other concepts need to be developed if this particular wavelength region is to be accessed.

2 Basic Principles of Nonlinear Optics and Difference Frequency Generation As pointed out in Sect. 1, there are to date no active laser materials with a broad tuning range in the mid-infrared region around 3 µm. A powerful scheme is the use of frequency conversion in a nonlinear optical material. This method is well established and widely used today for various IR wavelength regions of interest. 2.1

Frequency Conversion Processes

In order to describe the different possible frequency conversion processes, the dielectric polarization P of the material in which the frequency conversion is to take place needs to be considered. This material polarization is induced by the electric field of the incident light which distorts the wavefunctions of the outer electrons. This in turn results in a separation of charges which gives rise to a dipole moment. The dipole moment per unit volume is called the polarization of the material. It can be separated into two different parts: a linear part P lin and a nonlinear part P nl : P = P lin + P nl .

(1)

The linear polarization P lin of the material can be written in the form: P lin = ε0 χ(1) E ,

(2)

where ε0 is the permittivity of free space and χ(1) is the linear susceptibility: χ(1) = n2 − 1 ,

(3)


101

where n is the refractive index of the medium [17]. The nonlinear polarization P nl can be expanded as a power series of the electric field E and the nonlinear susceptibilities χ(m) of order m (the socalled electric dipole approximation): P nl = ε0

∞

χ(m) E m .

(4)

m=2

For the power series to converge, the following must hold: χ(1) E χ(2) E 2 χ(3) E 3 . . . .

(5)

Equation (5) is always true if the electric field of the incident light is much weaker than the electric field of the atoms. χ(m) is a tensor quantity and thus depends on the direction of the electric field involved. For example, the χ(2) tensor is represented by its 3 × 3 × 3 tensor elements. In tensor notation, the second-order nonlinear polarization is: (2)

Pi

(2)

= ε0 χijk Ej Ek ,

(6)

where the first index i corresponds to the index of the induced polarization while the second index j and the third index k are related to the indices of the incident electric field. All these indices represent the three axes of space (1, 2, 3). The electric field mentioned in (6) is the total electric field of the incident light. The lowest-order nonlinear susceptibility χ(2) describes second-order nonlinear processes (m = 2) – which only occur in non-centrosymmetric materials – briefly summarized in the following. Second harmonic generation (SHG) as illustrated in Fig. 3:

ω

Nonlinearly Active Material

ω ω + ω

2ω

2ω

Fig. 3. Schematic illustration of second harmonic generation

If the nonlinear optical material is pumped at high intensity, the second harmonic of the incident frequency ω is generated in the material. Two photons of the initial beam are converted into a single photon at twice its frequency (2ω). SHG was the first nonlinear optical effect observed. It was discovered in 1961 [18].

102


Sum frequency generation (SFG), illustrated in Fig. 4:

ω1 ω2

Nonlinearly Active Material ω 1 + ω2

ω3

ω3

Fig. 4. Schematic illustration of sum frequency generation

Sum frequency generation is a process where two incoming photons at different frequencies ω1 and ω2 are mixed in a nonlinear optical material resulting in the sum frequency ω3 of these two frequencies, i.e., ω3 = ω1 + ω2 . SFG is of interest for the generation of higher frequencies, generally in the UV. Difference frequency generation (DFG), illustrated in Fig. 5:

ω3 ω2

Nonlinearly Active Material ω3 − ω2

ω1

ω1

Fig. 5. Schematic illustration of difference frequency generation

The difference frequency ω1 = ω3 − ω2 results from nonlinear mixing of two incoming photons at frequencies ω3 and ω2 . By convention, the laser beam with the highest frequency ω3 is called the pump laser beam, the one with the lowest frequency ω1 is the idler beam, and the remaining one is the signal laser beam at a frequency ω2 . Optical parametric generation (OPG), illustrated in Fig. 6:

ω3

Nonlinearly Active Material ω3

ω1 + ω 2

ω2 ω1

Fig. 6. Schematic illustration of optical parametric generation

Optical parametric generation is a process wherein the incoming light is converted into two beams at different frequencies. OPG is parametric fluorescence and thus always occurs in a nonlinear optical material. The phasematching condition determines which frequencies are generated. Both the


103

DFG and the OPG process are of special interest for the generation of mid-IR wavelengths. In the following, we will focus on the DFG process. Optical rectification (OR): Optical rectification is a process where the incoming beam at frequency ω mixes with itself or another beam at the same frequency resulting in a difference frequency ∆ ω = 0. This process was first observed by Bass et al., in 1962 [19]. For all processes mentioned above, the energy is conserved: hω 1 + h ¯ ¯ ω2 = h ¯ ω3 ,

(7)

where h ¯ = h/(2 π) denotes Planck’s constant. 2.2

Nonlinear Optical Coefficient

Nonlinear optical materials which show second-order nonlinear processes are commonly described by their nonlinear coefficients dijk . These coefficients (2) are proportional to the nonlinear susceptibilities χijk : P (2) = ε0 χ(2) E E = 2 ε0 d E E ⇐⇒ or in tensor notation:

χ(2) = 2 d (2)

χijk = 2 dijk .

(8)

Since d is a tensor quantity, it can be described by its tensor elements. Due to symmetry aspects (intrinsic to the definition of dijk , j and k can be permuted) some of the elements have the same value. Thus, we can write the d-tensor as a 3 × 6 matrix dim . By convention, the index m is given by the numbering scheme outlined in Table 1. Table 1. Numbering convention for nonlinear optical coefficients d j=1

j=2

j=3

k=1

m=1

m=6

m=5

k=2

m=6

m=2

m=4

k=3

m=5

m=4

m=3

An appropriate coordinate system has to be introduced to describe the tensor quantity d. The system chosen depends on the material and is usually related to the crystallographic symmetries of the crystal used. Unfortunately,

104


different coordinate systems can be found in the literature (a,b,c: natural crystallographic axes, X,Y ,Z: rectangular, right-handed crystallographic axes, x,y,z: dielectric axes, 1,2,3: dielectric axes used in nonlinear optics (NLO) for indexing the nonlinear optical coefficient). In the following, we explain the system that was proposed as standard by Roberts in 1992 [20]. Historically, the first nonlinear optical effects were analyzed in optically uniaxial crystals. We will thus start our description of the coordinate system with optically uniaxial crystals for which the nomenclature is easier and more consistent. The existence of a rotation axis (i.e., symmetry axis) is a common feature of optically uniaxial crystals. In crystallography, this highest symmetry axis is labeled the c-axis. As the a,b,c coordinate system is not necessarily rectangular, there is a corresponding rectangular X,Y ,Z frame. The c-axis is usually chosen (and is also proposed to be used as a standard for NLO [20]) to coincide with the Z-axis of the tensor coordinate system. In NLO, the rotation axis of the crystal is usually chosen to coincide with the often cited optical axis (or 3-axis or z-axis) of the crystal. This nomenclature has the advantage that the refractive index for light polarized perpendicularly (ordinarily) with respect to the 3-axis of the coordinate system is always the so-called ordinary refractive index no . For a uniaxial crystal, the ordinary refractive index is independent of the direction of propagation. Light polarized parallel to the optical axis – i.e., extraordinarily polarized light – sees the extraordinary refractive index ne . Unlike no , the extraordinary refractive index ne depends on the direction of propagation. Important is the angle Θ between the optical axis of the crystal and the direction of propagation of the incoming laser beams as it will determine which nonlinear coefficient the light actually sees. This coefficient is called the effective nonlinear coefficient. The 1-axis of the NLO coordinate system proposed to be the standard for NLO [20] is then chosen along the crystal a-axis, the remaining 2-axis needs to be perpendicular to the 1- and 3-axes. The refractive index for a uniaxial crystal is commonly described by the index ellipsoid or indicatrix introduced in Fig. 7. For uniaxial crystals, two different cases are distinguished: positive and negative birefringent crystals. In the case of a positive uniaxial crystal, the extraordinary refractive index is larger than the ordinary index, i.e., ne > no (Fig. 7a). For a negative uniaxial crystal, the opposite is true, i.e., ne < no (Fig. 7b). Mathematically the equation of the ellipse shown as shaded area in Fig. 7 is given by: sin2 Θ cos2 Θ 1 = + . n2 (Θ) n2e n2o

(9)

Equation (9) clearly indicates the dependence of the refractive index n(Θ) on Θ.

Mid-IR Difference Frequency Generation c

105

3 ne > no c

3

ne < no

ne ne

1

no no b

a)

a

no

2

Optical uniaxial positive birefringent crystal

1

no

a

b

b)

2

Optical uniaxial negative birefringent crystal

Fig. 7a,b. The indicatrix giving the optical axis along which the light sees the refractive indices ne or no

Whereas for uniaxial crystals, the above described nomenclature is widely accepted and found in most publications, the situation for biaxial nonlinear crystals is much more complicated and many different nomenclatures are found. The coordinate system often chosen to describe the nonlinear tensor is the following [20]: The different crystal classes are distinguished. For monoclinic crystals, the 2-axis is chosen along the b-axis of the crystal. For orthorhombic crystals, the axes are chosen along the (a,b,c)-axes of the crystal. For triclinic crystals, the axes are chosen arbitrarily. For a centrosymmetric crystal, all tensor elements vanish; for triclinic crystals, all elements are non-zero. For most other symmetries, many terms of the nonlinear tensor are zero which simplifies the calculations. In Table 2, the nonlinear coefficients given in units of pm V−1 (10−12 mV−1 ) of some commonly used materials for frequency conversion are listed. 2.3

Phase Matching

Phase matching (or momentum conservation) is a crucial issue in frequency conversion. The different phase-matching types including quasi-phasematching will be discussed in this section both for positive and negative birefringent nonlinear bulk crystals and periodically poled materials. Other important parameters such as conversion efficiency, walk-off, acceptance angle, and acceptance bandwidth will be introduced.

106


Table 2. Nonlinear optical coefficients dim of some commonly used nonlinear inorganic uniaxial and biaxial crystals for frequency conversion around 1.064 µm or around 10.6 µm Crystal AgGaS2

Nonlinear coefficient dim (pm V−1 )

Wavelengths λ( µm)

d36 = 17.5 [20]

1.064

d36 = 11.2 [20]

10.600

AgGaSe2

d36 = 33 [20]

AgSbS3

d22 = 8.2 [21]

Ba2 NaNb5 O15

d15 = −12.8 [21]

d24 = 12.8 [21]

d32 = −12.8 [21]

d33 = −17.6 [21]

10.600 d31 = 7.8 [21] d31 = −12.8 [21]

10.600 1.064 10.600

Ba2 TiO3

d32 = −13.5 [23]

d33 = −5 [23]

BeSO4

d14 = 0.22 [20]

d25 = 0.22 [20]

d36 = 0.22 [20]

1.064

β-BaB2 O4 (BBO)

d16 = −2.3 [20]

d21 = −2.3 [20]

d22 = 2.3 [20]

1.064

d24 = 0.1 [20]

d31 = 0.1 [20]

d32 = 0.1 [20]

1.064

1.064

CdGeAs2

d36 = 282 [21]

CdS

d15 = −16 [20]

d24 = −16 [20]

d32 = −16 [20]

d33 = 32 [20]

10.600

CdSe

d31 = −18 [20]

d33 = 36 [20]

10.600

CsH2 AsO4 (CDA)

d36 = 0.40 [22]

CsTiOAsO4 (CTA)

d31 = 2.1 [21]

10.600 d31 = −16 [20]

1.064 d32 = 3.4 [21]

d33 = 18.1 [21]

d36 = 83 [20]

GaAs

10.600

1.064 10.600

GaP

d36 = 37 [20]

GaSe

d14 = 105 [21]

InP

d14 = 263 [22]

10.600

KD2 PO4 (KD∗ P)

d36 = 0.37 [20]

1.064

KH2 PO4 (KDP)

d36 = 0.39 [20]

K3 Li2 Nb5 O15

d31 = 5.8 [21]

KNbO3

d15 = −12.8 [20]

d24 = 11.3 [20]

d32 = 11.3 [20]

d33 = −19.5 [20]

d15 = 5.2 [20]

d24 = 2.9 [20]

d32 = 2.9 [20]

d33 = 12 [20]

d15 = 3.6 [20]

d24 = 2.0 [20]

d32 = 2.0 [20]

d33 = 8.3 [20]

KTiOAsO4 (KTA) KTiOPO4 (KTP) LiB3 O5 (LBO)

10.600 d22 = 54 [21]

10.600

1.064 d33 = 10.5 [21]

d15 = 0.85 [20]

d24 = −0.67 [20]

d32 = −0.67 [20]

d33 = 0.04 [20]

1.064 d31 = −12.8 [20]

1.064

d31 = 5.2 [20]

1.064

1.064 1.064 d31 = 3.6 [20]

1.064 1.064

d31 = 0.85 [20]

1.064 1.064

LiIO3

d15 = 4.4 [20]

d31 = 4.4 [20]

d33 = 4.5 [20]

1.064

LiNbO3

d15 = −4.3 [20]

d24 = −4.3 [20]

d31 = −4.3 [20]

1.064

d32 = −4.3 [20]

d33 = −27 [20]

LiTaO3

d22 = 2 [24]

d31 = −1 [24]

d33 = −21 [24]

1.064

ND4 H2 PO4

d36 = 0.52 [22]

NH4 H2 PO4 (ADP)

d14 = 0.47 [20]

RbH2 PO4 (RDP)

d36 = 0.40 [21]

RbTiOAsO4 (RTA)

d31 = 2.3 [23]

Se

d11 = 97 [21]

1.064 1.064

d36 = 0.47 [20]

1.064 1.064

d32 = 3.8 [23]

d33 = 15.8 [23]

1.064 10.600

SiO2 (Quartz)

d11 = 0.3 [20]

1.064

Te

d11 = 598 [21]

10.600

ZnGeP2

d14 = 69 [20]

d25 = 69 [20]

d36 = 69 [20]

10.600


2.3.1

107

Birefringent Phase Matching

The kind of nonlinear process actually observed mainly depends on the phasematching condition which corresponds to the conservation of momentum. The phase-matching condition in a nonlinear optical single crystal is given by: ∆ k = kp − ks − ki = 0 ,

(10)

where ∆ k is the so-called phase mismatch, and kp , ks , and ki denote the wavevectors of the pump, signal, and idler beam, respectively. All the materials showing second-order nonlinear optical activity are non-centrosymmetric, even though the wavevectors do not necessarily depend on the state of polarization of the incident light according to their crystal class. Two different perpendicular polarizations are distinguished according to the crystallographic point group of the material: ordinary (o) and extraordinary (e) polarization. For processes in nonlinear optical crystals, two different types of phase matching (type I and type II) occur. Historically, they originate from SHG experiments [25]. Kaschke and Koch suggest that a third type (type III) is added [26]. Here, we cite the definition of all three types both for SHG/SFG and DFG processes, but will then concentrate on type I and II phase matching. Table 3 summarizes the possible combinations for both positive and negative birefringence. Using |k| = nc/λ, where c is the speed of light and λ the wavelength, the phase-matching condition in the case of collinear wavevector propagation, Table 3. Different phase-matching types for SHG and DFG processes for positive and negative birefringent crystals (λi ≥ λs > λp : wavelengths of the idler, signal, and pump beams, respectively, PM: phase matching, e: extraordinarily polarized, o: ordinarily polarized, in: incident laser beams, out: generated beam) PM type

Birefringence in λi

SHG/SFG out λs

→

λp

DFG in λp

out λs

→

λi

I

positive

e

e

→

o

o

e

→

e

I

negative

o

o

→

e

e

o

→

o

II

positive

o

e

→

o

o

e

→

o

II

negative

e

o

→

e

e

o

→

e

III

positive

e

o

→

o

o

o

→

e

III

negative

o

e

→

e

e

e

→

o

108


type I phase matching, and a positive birefringent crystal can be expressed as follows for a difference frequency generation (DFG) process: ne (ωi , Θ) nDFG no (ωp ) ne (ωs , Θ) = = − . λi λDFG λp λs

(11)

Using (9), one obtains: −1 no (ωp ) ne (ωs , Θ) sin2 Θpm 1 1 − sin2 Θpm − = + , λp λs λi [ne (ωi )]2 [no (ωi )]2

(12)

where Θpm is the so-called phase-matching angle. Phase matching can be realized by different methods which all take advantage of the fact that the refractive index changes as a function of the incident angle (12), temperature or wavelength. Accordingly, one distinguishes between angle, temperature, and wavelength tuning: Angle Tuning: The most common way to reach phase matching in a single crystal is by rotating the crystal until the phase-matching condition (12) is fulfilled. For a uniaxial crystal, the rotation axis is usually chosen perpendicular to the symmetry axis of the crystal. The situation is depicted in Fig. 8. ne

k ne(ω, Θ)

Θ

n(Θ)

no no(ω,0) no(ω,Θ) = no(ω,0) ne(ω,0)

Fig. 8. Angle tuning (by angle Θ) of a uniaxial crystal changes the refractive index the light experiences, and thus the phase-matching condition can be met by angle tuning of the crystal. The corresponding refractive indices no and ne are indicated


109

Temperature Tuning: Another approach to reach phase matching involves changing the temperature of the crystal as the refractive indices not only depend on the angle Θ, but also on the temperature T and the wavelength λ. Only at a well-defined temperature will the phase-matching condition thus be fulfilled and the nonlinear optical process take place. Wavelength Tuning: A third possibility to reach phase matching is to change the wavelength of the pump and/or the signal beam. In contrast to the previous schemes, this, however, requires a tunable pump and/or signal source. 2.3.2

Acceptance Angle and Acceptance Bandwidth

The angle Θ for which phase matching is fulfilled is defined as the angle between the optical axis and the wavevector k of the light. The indicatrix gives the two refractive indices ne and no for each angle Θ as shown above in Fig. 8. For a uniaxial crystal, the refractive index no is independent of Θ whereas ne takes different values for each angle. For biaxial crystals, both refractive indices depend on Θ. In addition to the different types of phase matching (I and II), critical and non-critical phase matching are distinguished. Figure 9 illustrates the difference. ne

ne

acceptance angle

acceptance angle ne(ω) ne(ω)

no(2ω) no(2ω) no no(ω)

a) non-critical phasematching: ne > no

no no(ω)

b) critical phasematching: ne > no

Fig. 9. Phase matching in a uniaxial birefringent crystal for SHG. (a) Situation for non-critical phase matching. Small deviations from exact phase matching do not cause large efficiency losses. (b) In the case of critical phase matching, small deviations from exact phase matching lead to a large phase mismatch

110


In the case of non-critical phase matching, a small angle mismatch (e.g., induced by a small angle tuning of the crystal) will only have a minor effect on the conversion efficiency. In the case of critical phase matching, a small deviation from prefect phase matching will reduce the conversion efficiency considerably. The acceptance angle δΘ is defined as follows: If the phasematching angle Θpm is tuned to Θpm ± δΘ, the intensity of the generated DFG light shall drop to half its peak intensity, i.e., the acceptance angle gives the deviation from the phase-matching angle Θpm for which the idler intensity decays to half its peak value. This deviation is characterized by the phase shift ∆ kL/2 for which sinc2 (∆ kL/2) = 0.5 for Θ = Θpm ± δΘ (L denotes the crystal length). To obtain the acceptance angle δΘ, the phase mismatch ∆ k is expanded as a power series around Θ = Θpm : 2 ∂n(Θ) ω ω 2 ∂ n(Θ) + (Θ − Θpm ) . (13) ∆ k = 2 (Θ − Θpm ) 2 c ∂Θ Θ=Θpm c ∂Θ Θ=Θpm In the case of non-critical phase-matching, the first derivative vanishes at Θ = Θpm . Thus, only the second term needs to be considered. In the case of critical phase matching, the first term of the Taylor expansion is used to determine the acceptance angle. In order to judge how critical the alignment of the system as a whole and the positioning of the crystal in particular are, not only the acceptance angle, as described above, but also the spectral and temperature acceptance bandwidths need to be considered as the phase mismatch (∆ k) depends on both Θ and λ. The spectral acceptance δλ describes in principle the required degree of coherence of the light in order to achieve efficient phase matching. It is defined similarly to the acceptance angle: the phase shift ∆ kL/2 shall be such that sinc2 (∆ kL/2) = 0.5, i.e., the idler intensity will drop to half its peak value at λ = λpm ± δλ where λpm is the phase-matching wavelength. In order to find an expression for the acceptance bandwidth, the phase mismatch is again expanded as a power series, but this time around the phase-matching wavelength λ = λpm . Similar to the acceptance angle and the spectral acceptance bandwidth, there is also a temperature acceptance bandwidth δT . It is derived from the phase-matching temperature Tpm and is defined as the temperature deviation δT for which the intensity of the generated light drops by a factor of two. 2.3.3

Conversion Efficiency

A nonlinear process is commonly characterized by the conversion efficiency η. This parameter is the ratio of the power Pout or intensity Iout at the desired output wavelength (DFG wavelength here) to the incident laser power or intensity, respectively, i.e., the rate of the incident intensity that is converted


111

into the intensity at the desired wavelength. As there are two laser beams incident on the nonlinear crystal for a DFG process, two different conversion efficiencies can be defined:

ηp =

Pout Iout = Pp Ip

(14)

ηs =

Pout Iout = , Ps Is

(15)

where ηp , ηs , Pp , Ps , and Ip , Is are the conversion efficiencies, the powers and the intensities for the pump and the signal laser, respectively. We here assumed that the cross-sectional area A for the pump, signal and idler beams are equal, i.e., Ip = Pp /A and Is = Ps /A. In the case of difference frequency generation (DFG) and of perfect phase matching, the conversion efficiency η is in the limit of low conversion efficiency [27]:

ηp = 2

ωi2 d2eff L2 Is , c3 ε 0 n p n s n i

(16)

ηs = 2

ωi2 d2eff L2 Ip , c3 ε 0 n p n s n i

(17)

and in the case of ∆ k = 0, i.e., for non-perfect phase matching: ω 2 d2 L2 Is ηp = 2 3i eff sinc2 c ε 0 np ns ni ηs = 2

ωi2 d2eff L2 Ip sinc2 c3 ε 0 n p n s n i

∆ kL 2 ∆ kL 2

,

(18)

,

(19)

where ωi is the frequency of the generated light (i.e., the frequency of the idler beam). np , ns , ni indicate the refractive indices of the pump, the signal, and the idler beams, respectively, c is the speed of light, and ε0 the dielectric constant. deff denotes the effective nonlinear coefficient depending on the type of phase matching and other parameters such as the angle Θ as described below. L is the interaction length (i.e., the pathlength of the light within the crystal where the incident beams interact with each other). As mentioned above, two different conversion efficiencies exist for a DFG process: ηp and ηs . ηp describes how much of the pump laser power is converted into DFG radiation, whereas ηs describes how much signal laser power is converted. The effective nonlinear coefficient deff depends on the direction of propagation of the light within the crystal as well as on the direction of the polarization and the crystal orientation. In Table 4, the effective nonlinear

112


Table 4. Effective nonlinear optical coefficient for some crystallographic point groups and both positive and negative birefringent crystals. Θ denotes the angle between the direction of propagation and the optical axis of the crystal, whereas φ is the angle between the 1-axis (or a-axis) of the crystal and the projection of the direction of propagation onto the a–b plane (standard “Euler” angles in the coordinate system described above in Sect. 2.2) Crystal group

Positive birefringent crystals Type II phase matching deff Type I phase matching deff

6, 4

−d14 sin 2Θ

d15 sin Θ

622, 422

−d14 sin 2Θ

0

6mm, 4mm 0 ¯ 6m2 d22 cos2 Θ cos 3φ

d15 sin Θ −d22 cos Θ sin 3φ

3m ¯ 6

d22 cos2 Θ cos 3φ

d15 sin Θ − d22 cos Θ sin 3φ

cos Θ(d11 sin 3φ + d22 cos 3φ)

cos Θ(d11 cos 3φ − d22 sin 3φ)

3

cos2 Θ(d11 sin 3φ + d22 cos 3φ) −d14 sin 2Θ

cos Θ(d11 cos 3φ + d22 sin 3φ) +d15 sin Θ

32 ¯ 4

d11 cos2 Θ sin 3φ − d14 sin 2Θ

d11 cos Θ cos 3φ

sin 2Θ(d14 cos 2φ − d15 sin 2φ)

− sin Θ(d14 sin 2φ + d15 cos 2φ)

¯ 42m

d14 sin 2Θ cos 2φ

−d14 sin Θ sin 2φ

Crystal group

Negative birefringent crystals Type II phase matching deff Type I phase matching deff

6, 4

d31 sin Θ

d14 sin Θ cos Θ

622, 422

0

d14 sin Θ cos Θ

2

6mm, 4mm d31 sin Θ ¯ 6m2 −d22 cos Θ sin 3φ

0 d22 cos2 Θ cos 3φ

3m ¯ 6

d31 sin Θ − d22 sin 3φ cos Θ

d22 cos2 Θ cos 3φ

cos Θ(d11 cos 3φ − d22 sin 3φ)

cos2 Θ(d11 sin 3φ + d22 cos 3φ)

3

cos Θ(d11 cos 3φ − d22 sin 3φ) +d31 sin Θ

cos2 Θ(d11 sin 3φ + d22 cos 3φ) +d14 sin Θ cos Θ

32 ¯ 4

d11 cos Θ cos 3φ

d11 cos2 Θ sin 3φ + d14 sin Θ cos Θ

− sin Θ(d31 cos 2φ + d36 sin 2φ)

cos 2φ(d14 + d36 ) sin Θ cos Θ − sin 2φ(d15 + d31 ) sin Θ cos Θ

¯ 42m

−d36 sin 2φ sin Θ

(d36 + d14 ) cos 2φ sin Θ cos Θ


113

optical coefficients for some crystal groups and both negative and positive birefringent crystals are given. The intensity of the generated light generally scales with the interaction length L (normally corresponding to the crystal length) and is proportional to the conversion efficiency η. For perfect phase matching, it is given as: IDFG = 2

ωi2 d2eff L2 Ip Is . c3 ε 0 n p n s n i

(20)

This equation is valid as long as there is no pump depletion, i.e., as long as the conversion efficiency is small (ηp,s < 0.3), and if the cross-sectional areas of all three beams can be assumed to be equal. As can be seen from (20), the intensity of the generated DFG light scales with both the pump and the signal laser intensity. If the pump intensity is high, the output DFG power will increase. But there is also a trade-off: if the pump intensity is high, other nonlinear processes – in particular optical parametric generation (OPG) and optical parametric amplification (OPA), which have undesired effects on the linewidth of the generated light – become efficient. In the case of critical phase matching, the waves with different polarizations and different wavelengths will travel along different paths within the crystal due to the crystal birefringence. After some distance, denoted as the walk-off distance, the three beams will be completely separated. There will be no further interaction between the three laser beams and thus the intensity of the generated light will not increase further even if the crystal length L is much longer than the walk-off distance la (also called the aperture distance) [28]: √ πw , (21) la = ρ where w is the 1/e power radius of the Gaussian beam spot of the idler or the signal beam (for an actual calculation, the beam spot of the smaller of the two beams is chosen). The parameter commonly used to describe the walk-off effect is the walk-off angle ρ which describes the angle between two beams of different polarization. In the case of DFG and birefringent phase matching, there are thus two different walk-off angles depending on the type of phase matching. For bulk LiNbO3 , this angle is about 2◦ and depends on the direction. For most birefringent materials, the walk-off angle is between 1◦ and 2.5◦ . Equation (17) gives the maximum conversion efficiency which may be achieved assuming that no losses occur, neither within the crystal, nor due to focusing, nor when coupling into the crystal, etc. In practice, there are always losses, and the conversion efficiency taking into account focusing losses

114


is written as [29]: ηp = 2

ωi2 d2eff L2 Is h(µ, ξs , ρ) , n p n s n i c3 ε 0

(22)

ηs = 2

ωi2 d2eff L2 Ip h(µ, ξp , ρ) , n p n s n i c3 ε 0

(23)

2 where µ = ks /kp , ξs,p = L/(ks,p ws,p ) with ws , wp denoting the Gaussian beam spot radius of the signal and pump laser beam, respectively. The function h(µ, ξ, ρ) contains information about the focusing geometry. For optimum focusing and in the case that there is no birefringence, h(µ, ξ, ρ) ∼ 1 which corresponds to ξ = 1 [30]. If birefringence cannot be neglected, h(µ, ξ, ρ) and thus the efficiency η are reduced. The conversion efficiency described by (16) and (17) is often used as the figure of merit for DFG processes. Unfortunately, there also exist other definitions of the figure of merit. For instance, a normalized conversion efficiency in the form:

ηnorm,1 = 2

ωi2 d2eff L2 3 c ε 0 np ns ni

,

(24)

or in the form: ηnorm,2 = 2

ωi2 d2eff 0 np ns ni

c3 ε

(25)

is also a suitable definition of the figure of merit. The advantage of both these definitions of the figure of merit is that they are independent of the incident laser power and in the case of (25) even independent of the crystal length. 2.3.4

Quasi-phase-matching

Owing to the shape of the dispersion relation within a material, it is often difficult or even impossible to fulfill both energy and momentum conservation simultaneously for most materials. The situation is illustrated in Fig. 10a. For normal dispersion (dn/dω > 0, which can also be expressed as dn/dλ < 0), it is impossible to reach phase matching between the fundamental and the second harmonic of the fundamental at twice the frequency. The same holds true for the nonlinear mixing of two arbitrary frequencies as shown in Fig. 10a. For birefringent materials, however, it is possible to fulfill both energy conservation and phase matching under certain conditions (since the refractive index depends on the polarization of the material) as illustrated in Fig. 10b. There are many materials for which the phase-matching condition and energy conservation can simultaneously be fulfilled. Unfortunately, the largest nonlinear coefficient can quite often not be used for the frequency conversion process. An example is the well-known negative birefringent nonlinear


115

n ω normal normaldispersion: dispersion:dn/d dn/dω >> 00

ω

k

n

energy conservation: ω i = ω p - ωs phase-matching: type I, positive birefringent ne(ωi) ωi = no(ωp) ωp - ne(ωs) ωs

kp ki + ks ks ki

normal dispersion

ωi ωs

ωi + ωs= ωp

ω

a)

ωp

ωs ω i

ne(ωi) ne(ωs) no = no(ωp) λ

b)

Fig. 10. (a) Normal dispersion relation (dn/dω > 0): Phase matching (10) and energy conservation (7) cannot be fulfilled simultaneously. (b) For a birefringent material, the phase-matching condition and energy conservation can simultaneously be fulfilled

material LiNbO3 . Its largest nonlinear coefficient is d33 which implies that the polarizations of all three beams involved are the same. In birefringent phase matching, this coefficient unfortunately cannot be used as it is neither compatible with type I, nor with type II phase matching (see Table 3 above). Thus, other approaches are desirable. One way to avoid this conflict is to use periodically poled nonlinear materials and take advantage of quasiphase-matching (QPM) which is illustrated in Fig. 11. If the phase-matching condition is not met (∆ k = 0), after each coherence length lc given by lc = π/∆ k, the newly generated light will destructively interfere with the light generated in the previous coherence length. Thus, after twice the coherence length all generated light will be destroyed (case C in Fig. 11). On the other hand, if the polarization of the material is changed by 180◦ after each coherence length, the nonlinear optical coefficient will change its sign. Thus, the light will constructively interfere with the light from the previous coherence length and a build-up of the generated light is observed (case B in Fig. 11). The figure also shows that the build-up of the generated intensity is lower than in the case of perfect phase matching where ∆ k = 0 (case A in Fig. 11). Case B is called quasi-phase-matching, and as Fig. 11 implies, the length of the domain ΛQPM with constant polarization must be an even multiple of the coherence length lc , i.e., ΛQPM = 2lc . For a quasi-phase-matching process, the phase mismatch ∆ k is: ∆k =

2π = kp − ks − ki = 0 . ΛQPM

(26)

116

Cornelia Fischer and Markus W. Sigrist IDFG IDFG(4lc)

A

(2/π)2 IDFG(4lc) B

C 0

lc

z

lc

polarisation direction

nonlinear material

Fig. 11. Quasi-phase-matching in a periodically poled nonlinear optical material. (A) Idler intensity for a DFG process in a single crystal if the phase-mismatch ∆ k is equal to zero. (B) Idler intensity for quasi-phase-matching. (C) Idler intensity in a single crystal if the phase-matching condition is not met. lc denotes the coherence length

Although the efficiency of the process is lower than in the case of birefringent phase matching (where ∆ k = 0), one of the major advantages is that quasi-phase-matching is not restricted to type I and type II phase matching, but also other processes are possible, in particular the process e + e → e using the nonlinear coefficient d33 which is largest for many nonlinear materials (see Table 2). In addition for this particular process (e + e → e), the walk-off angle (21) is zero as all involved polarizations are the same. Another advantage is that QMP usually operates in the non-critical phase-matching regime. The conversion efficiency of the pump laser beam ηQPM,p for this process is given as:

ηQPM,p = 2

ωi2 3 c ε 0 np ns ni

ωi2 =2 3 c ε 0 np ns ni

2deff mπ

2

dQPM m

L2 Is 2 L2 Is ,

(27)


117

where m is the order of quasi-phase-matching and dQPM the effective nonlinear coefficient for quasi-phase-matching. An analogous expression is obtained for the conversion efficiency of the signal laser beam ηQP M, s simply by replacing Is in (27) by Ip . The order of the QPM process is given by the number of coherence lengths after which the intensity build-up continues, i.e., first-order quasi-phase-matching means that after each coherence length the build-up continues as shown in Fig. 11, third-order QPM means that only after the third coherence length the sign of the nonlinear coefficient is changed and a further intensity build-up will take place. Usually, first-order QPM is employed, but if the coherence length of the material is short, it becomes difficult to produce the fine periodically poled structure. Thus, the period length ΛQPM is extended. In order to achieve any intensity build-up, the period length needs to be an uneven number of coherence lengths lc of the nonlinear material employed. Owing to the fact that the largest nonlinear optical coefficient can be exploited and that in collinear frequency conversion the whole length of the crystal can be used (since the aperture distance la (21) is usually much larger than the coherence length or there occurs no walk-off at all due to the fact that all involved polarizations are the same), much higher intensities can be achieved with QPM than with birefringent phase matching for many materials [31]. Table 5 compares the achieved effective nonlinear coefficients for several commonly used nonlinear crystals. λPM is the wavelength at which the nonlinear coefficient was measured in the case of birefringent phase matching (at Table 5. Comparison of quasi-phase-matching (QPM) and birefringent phase matching (PM) systems for SHG. λPM is the birefringent phase-matching wavelength, d is the corresponding nonlinear optical coefficient, ∆ λPM is the acceptance bandwidth for a 1 cm-long crystal, λQPM the quasi-phase-matching wavelength, dQPM is the effective nonlinear optical coefficient for QPM, and ΛQPM is the period length of the periodically poled structure Crystal

λPM (nm)

dim (pm V−1 )

∆ λPM (nm)

CsLiB6 O10

478

d36

KNbO3 (KN)

858 858 982

d31 −12.8 d32 −11.3 d24 −11.3

0.09 0.09 0.18

d31 d32

3.6 2.0

0.08

d31

0.85

KTiOPO4 (KTP) LiB3 O5 (LBO)

994 1082 552

1.1

λQPM (nm)

dQPM (pm V−1 )

ΛQPM ( µm)

0.5

850

8.3

4.0

1

LiNbO3 (LN)

1086

d31 −4.3

0.04

850

−17

3.0

LiTaO3 (LT)

1058

d31

0.12

850

10

3.8

1

118


room temperature and for non-critical phase matching), λQPM denotes the corresponding wavelength for QPM. The effective nonlinear optical coefficient dQPM is the one for first-order QPM: dQPM =

2 deff . π

(28)

The acceptance bandwidth ∆ λPM is given for a 1 cm-long bulk crystal. Table 5 clearly shows that the effective nonlinear coefficient for QPM is usually higher than for birefringent phase matching. Equation (27) can be used to calculate the maximum achievable DFG intensity IDFG for QPM: IDFG


2 deff mπ

2 L2 Ip Is .

(29)

The DFG power PDFG is then given as: PDFG


2 deff mπ

2 L2 Pp Ps

1 , A

(30)

where A is the cross-sectional area of the beam assuming that A is not altered during the DFG process and that the two incoming beams have the same cross-section. The equations given in this section were all derived for a flat wave approximation. A careful analysis shows that the cross-section A of the laser beam is proportional to the length of the crystal (for most practical cases). Thus, the overall efficiency, DFG intensity, and DFG power is proportional to length L instead of L2 .

3

Material Considerations

The implementation of a DFG laser source requires selecting a pump and signal source as well as an appropriate nonlinear medium. Several parameters such as transparency range, nonlinear coefficient, and damage threshold of the nonlinear crystal, as well as wavelength tunability and power of the available lasers, come into play. Obviously, the selection of the lasers and the crystal has to be carefully matched. With the advent of new laser types and a variety of novel nonlinear crystals and particularly periodically poled materials, new combinations have become available. In the first part of this section, we focus on the nonlinear material while the choice of laser sources is emphasized in the second part.


3.1

119

Selection of Nonlinear Medium

Table 6 and Fig. 12 summarize the most important properties of various nonlinear media which might be used for mid-IR difference frequency generation. The nonlinear material selected needs to be transparent for all wavelengths involved, i.e., for the signal, the pump, and the idler wavelength. Furthermore, the nonlinear coefficient should be as high as possible for high conversion efficiencies. In addition, the material has to be able to cope with high intensities (especially if a high output DFG power shall be achieved) as the intensity of the generated light scales with the intensities of the pump and signal source (20). The damage thresholds of some selected materials are given in Table 6. Table 6 implies that a wide variety of crystals is available for frequency conversion. Also, a large wavelength range is covered (0.1–20 µm).

CdGeAs 2

100

Absolute nonlinear coefficient in pmV-1

ZnGeP2 GaSe CdSe

AgGaSe2

PPLN

HgGa2S4

KNbO3 PPRTA

AgGaS2 PPKTA

AgAsS 3

10

PPKTP LiIO3

KTA LiNbO3 KTP

RTA

KDP

1

CLBO LBO ADP BBO SiO2 0.1

1

10

Transparency range in µm

Fig. 12. Transparency range of some selected nonlinear crystals used for mid-IR DFG generation as a function of the absolute value of the nonlinear coefficient. For the periodically poled materials (PPLN, PPRTA, PPKTA, PPKTP), the nonlinear coefficient d33 is given. For n-th order QPM processes, the values given need to be multiplied by 2/(nπ)

120


Table 6. Important properties of different nonlinear crystals commonly used for difference frequency generation. d denotes the nonlinear optical coefficient. The damage threshold is given in GW cm−2 for the indicated pulse duration in ns Crystal

Transparent ( µm)

AgAsS3 AgGaS2 AgGaSe2 AgSbS3 β-BaB2 O4 (BBO)

0.6–13 0.46–13 0.71–19 0.7–14 0.2–2.6

CdGeAs2 CdSe

2.4–18 [33] 0.75–25 [33]

CsTiOAsO4 (CTA) CsLiB6 O10 (CLBO) GaSe KD2 PO4 (KD∗ P)

0.35–5.3 0.18–2.75 0.62–20 0.19–1.6

[33] [33] [33] [33] [33]

[33] [33] [33] [34]

daim (pm V−1 ) d36 d36 d36 d22 d22 d21 d36 d31 d33 d33 d33 d22 d36

10.4 17.5 33 8.2 2.3 −2.3 282 −18 36 18.1 1 54 0.37

[33] [20] [20] [33] [20] [20] [33] [33] [20] [33] [22] [33] [20]

0.19–1.3 [34] KH2 PO4 (KDP) KNbO3 0.4–5.5 [24] KTiOAsO4 (KTA) 0.35–5.3 [33] KTiOAsO4 (PPKTA)

d36 0.39 [20] d33 −19.5 [20] d15 5.2 [20] d33 12 [20]

KTiOPO4 (KTP) KTiOPO4 (PPKTP) HgGa2 S4 HgS LiB3 O5 (LBO)

0.34–3.2 [33]

d15 d33 d36 d11 d31

3.6 8.3 24 50 0.85

[20] [20] [33] [33] [20]

LiIO3 LiNbO3 LiNbO3 (PPLN) LiTaO3 MgO:LiNb3 NH4 H2 PO4 (ADP) RbTiOAsO4 (RTA) RbTiOAsO4 (PPRTA) SiO2 (Quartz) ZnGeP2

0.3–5.8 [33] 0.35–5.5 [34]

d33 d15 d33 d33 d33 d36 d32 d33

4.5 −4.3 −27 −21 17 0.47 4.6 12.1

[20] [20] [20] [20] [35] [20] [33] [33]

0.55–13 [33] 0.62–13 [33] 0.2–3.2 [33]

0.32–5.2 0.40–5 0.18–1.5 0.35–5.8

[34] [33] [33] [23]

0.2–3.1 [34] 0.74–12 [33]

d11 d36

0.3 [20] 69 [20]

Damage threshold [32,33] (GW cm−2 ) (ns) 0.025 0.025

10 10

9.9 5

1.3 10

0.060

10

26

1

0.5 6 8.4 0.1 10

10 0.25 1.3 10 10

0.16 4.6

20 1.3

18 87 0.06 0.5

1.3 0.03 20 10

0.5

60

1

2

a Note: The nonlinear optical coefficients d11 , d22 , and d33 can only be used in combination with quasi-phase-matching


121

The fabrication of large crystals is advantageous for a high conversion efficiency, especially in the case of birefringent phase matching. Costs of the material will also have to be considered as they increase for larger crystals. In addition, the walk-off angle or the aperture distance la (21) needs to be kept in mind. If the walk-off angle is large, the advantage large crystals offer becomes obsolete since the two incident laser beams disperse and do not interact any longer beyond the aperture distance. Since there often exists no material which meets all material requirements (large nonlinear coefficient, right transparency range, low costs, etc.), some sort of trade-off has to be taken into account. Finally, the possibility of the use of quasi-phase-matching needs to be considered before the system is set up. In recent years, many periodically poled materials have become commercially available. Today, the most commonly used periodically poled materials are PPLN, PPRTA, PPKTA, and PPKTP. Note that the period of the domains required for the wavelengths of interest is not always available, but that usually a larger nonlinear coefficient can be taken advantage of, although it is reduced by a factor 2/π. 3.2 Pump and Signal Laser Sources for Difference Frequency Generation Along with the selection of a suitable nonlinear material, laser sources with appropriate properties need to be found. The tuning range of one of the two lasers (either pump or signal laser) needs to be broad enough in order to cover the desired wavelength range for spectroscopic applications. The difference frequency of the two lasers chosen, of course, has to be in the desired wavelength region and all three lasers should be within the transparency range of the crystal. Furthermore, the laser power plays an important role. The intensity and thus the power of the generated DFG light scales with the pump and signal intensity. If the pump intensity is too low, the generated intensity can be limited by pump depletion [36]. However, for high pump intensities, other nonlinear processes like optical parametric generation (OPG) and amplification (OPA) will compete with DFG. Unwanted side effects of such processes are, e.g., an enlarged linewidth of the generated light. As an example, stateof-the-art external cavity diode lasers (ECDL) offer a broad tuning range. Their CW output power at mW level is appreciable, but still OPG/OPA and other nonlinear processes are sufficiently suppressed. The achieved average idler power using an external cavity diode laser as pump source though is usually only at the µW level which might not be sufficient for spectroscopic applications. Despite their limited CW laser power, ECDLs can successfully be employed as signal lasers in a DFG set-up [37]. A final issue to be considered is the DFG laser linewidth which needs to be sufficiently narrow for many applications. For example in gas sensing, the linewidth should be considerably smaller than the width of the molecular

122


absorption line. At atmospheric pressure, molecular absorption linewidths are typically 0.5 cm−1 (15 GHz), thus a DFG linewidth of ∼ GHz would be desirable. Since the DFG laser linewidth is given by a convolution of the linewidth of the pump and the signal laser, both laser sources need to be narrowband. This requirement further limits the choice of pump and signal sources. For instance, the pulse length of pulsed lasers should not be too short since short and ultrashort pulses in the ps and fs regime exhibit too large intrinsic linewidths. In addition, the peak power of ultrashort pulse lasers might exceed the damage threshold of the crystal, especially at high pulse repetition rates.

4 Difference Frequency Laser Systems Reported in the Literature Although various requirements are posed on pump and signal sources as well as on the nonlinear crystal for efficient difference frequency generation, there is still a large variety of possible combinations. The first DFG system reported dates back to 1974 [38]. It involved an Ar+ laser, a dye laser and a LiNbO3 crystal as nonlinear medium. With the availability of new laser types, new and improved nonlinear optical materials, and driven by numerous applications, many DFG laser sources have been reported. The publications can be divided into two different groups. On the one hand, pure systems without presentations of applications have been described. On the other hand, there also exist reports on DFG systems applied, e.g., for spectroscopic measurements. 4.1

Difference Frequency Laser Sources

Selected laser systems based on difference frequency generation that have been reported in literature in recent years are listed in Table 7. Given are the types and the wavelengths of the two incident lasers (pump laser and signal laser), the wavelength range of the generated DFG beam, along with information about the nonlinear crystal and the type of phase matching used. 4.2

Difference Frequency Laser Sources Applied in Gas Sensing

Table 8 lists some systems employed for spectroscopic applications. Given are the types and wavelengths of the two incident lasers and the wavelength range of the generated DFG signal. In addition, the nonlinear material is stated along with the type of phase matching used. Finally, the gases analyzed are listed.


123

Table 7. Selected DFG systems recently reported. Listed are pump and signal laser types with their wavelengths λp and λs , the generated DFG laser wavelength range λDFG , the nonlinear crystal, and the type of phase matching (PM). (ECDL stands for external cavity diode laser, Dye for dye laser, Ti:Sa for Ti:sapphire laser, Nd:YAG for Nd:YAG laser, DL for diode laser, Nd:YLF for Nd:YLF laser, OPG/OPA/OPO for optical parametric generation, amplification, and oscillation, HeNe for HeNe laser, CO for CO laser, FP for Fabry–Pérot, TDL for tunable diode laser) λp (nm)

λs (nm)

λDFG (nm)

540–620

910–1050

1064

Ti:Sa DL

Pump laser

Signal laser

Dye

Ti:Sa

Nd:YAG

CO

Nd:YAG DL

Crystal

PM

Ref.

1200–1900

KTiOPO4 (KTP)

II

[39]

2600–4100

1430–1800

RbTiOAsO4 (RTA)

532

770–820

1500–1730

LiB3 O5 (LBO)

I

[41]

780

∼1500

∼1500

LiNbO3

QPM

[42]

[40]

Ti:Sa

DL

784

1539

∼1597

LiNbO3

QPM

[43]

Nd:YLF

Ti:Sa

523

750–850

1649–1707

KTiOPO4 (KTP)

QPM

[44]

Ti:Sa

Ti:Sa

690–750

910–1050

2300–2800

KTiOPO4 (KTP)

II

[39]

OPA

OPA

1100–1600

1600–2933

2400–12000

GaAsS2

I & II

[45]

Ti:Sa

OPG/OPA

ECDL

HeNe

Ti:Sa

Nd:YAG

Nd:YAG

806

1200–1500

2500–7600

(3)

CaF2 & BaF2

χ

[46]

1500–1580

3391

2689–2958

Ti:LiNbO3

QPM

[47]

770–820

1064

3000–3600

KTiOPO5 (KTP)

II

[41]

OPO

1064

1430–1600

3200–4200

RbTiOAsO4 (RTA)

n.a.

[48]

OPO

OPO

1700–2010

2260–2840

5000–18000

GaAsSe2

I & II

[49]

Ti:Sa

Nd:YAG

1052–1068

1320

5200–5600

GaAs

QPM

[50]

FP-TDL

ECDL

DL

DL

ZnGeP2 OPG

ZnGeP2 OPG

ECDL

ECDL

ZnGeP2 OPG

ZnGeP2 OPG

1290

1504–1589

7143–7407

AgGaSe2

I

[51]

766–786

830–868

6800–12500

AgGaS2

II

[52]

4000–5000

6500–9500

7000–20000

CdGeAs2

I

[53]

778

842

10200

AgGaS2

I

[54]

∼ 4790

∼ 6740

15600–17600

GaAs

QPM

[55]

124


Table 8. DFG systems reported in the literature which were used for spectroscopic applications. Listed are pump and signal laser types with their wavelengths λp and λs , the generated DFG laser wavelength range λDFG , the nonlinear crystal, and the type of phase matching (PM), and the gases detected. (crit. stands for critical PM, Dye for dye laser, Ar+ for argon ion laser, ECDL for external cavity diode laser, Nd:YAG for Nd:YAG laser, DL for diode laser, DFB for distributed feedback laser, GaAlAs for GaAlAs laser, and Ti:Sa for Ti:sapphire laser) Pump laser

Signal laser

Dye

Ar+

λp (nm)

λs (nm)

λDFG (nm)

Crystal

1887–5263

LiIO3 &

PM

Gas Ref. formula H+ 3 , CO N2 O, H2 CO

LiNbO3

[56]

ECDL

DL

775–795

1096

2650–2900

LiNbO3

QPM

H2 O

[57]

ECDL

Nd:YAG 795–825

1064

3160–3670

LiNbO3

I

CH4 , H2 CO NO2 , C2 H6

[58]

Nd:YAG

ECDL

1064

1500–1600 3176–3660

LiNbO3

QPM

CH4

[37]

Nd:YAG

ECDL

1064

1515–1585 3236–3798

LiNbO3

QPM

CH4

[59]

DL

ECDL

1083

DL

DL

∼ 760

ECDL

ECDL

H2 CO ∼ 980

3262–3454

LiNbO3

QPM

O2 , H2 O CH4

[60]

DFB DL 814–870

1083

3280–4400

LiNbO3

QPM

CH4 , CO2 SO2 , HCl

[61]

DL

775–795

1010

3330–3730

LiNbO3

QPM

CH4

[57]

DL

DL

775–795

1010

3330–3734

LiNbO3

QPM

CH4

[62]

ECDL

Nd:YAG 795–825

1064

3350–3630

LiNbO3

crit. PM

CH4 H2 O

[63]

3358

LiNbO3

DL

Nd:YAG 808

1064

QPM

CH4

[64]

Nd:YAG

ECDL

1064.4

1500–1600 3450–3750

KTiOAsO4 QPM (KTA)

CH4

[65]

Nd:YAG

ECDL

1064.4

1500–1600 3450–3750

KTA, KTP QPM RTA

CH4 N2 O

[66]

DL

DL

1083

1561

3531

LiNbO3

QPM

H2 CO

[67]

ECDL

Nd:YAG 839–864

1064

3980–4620

LiNbO3

QPM

CO, CO2 SO2 , N2 O

[68]

DL

Nd:YAG 852

1064

4250

LiNbO3

QPM

CO2

[69]

DL

Nd:YAG 850

1064

4250

LiNbO3

QPM

CO2

[70]

GaAlAs

Nd:YAG 853–865

1064

4310–4629

LiNbO3

QPM

CO, N2 O

[71]

DL

DL

800–807

4487–5071

AgGaS2

noncrit.

CO OCS

[72]

GaAlAs

Nd:YAG 865

1064

4624

LiNbO3

QPM

CO, N2 O

[73]

AgGaS2

CO2 684–691

DL

Dye

668–673

768–785

5010–5230

I

NO

[74]

Ti:Sa

Ti:Sa

700–810

790–910

8000–19000 GaSe

I

C 2 H2

[75]

Ti:Sa

Ti:Sa

780–900 780–900 8800–15000 GaSe

I

C2 H4

[76]


5

125

Detailed Description of a Mid-IR DFG Laser Source

In the following, we present the design, implementation and applications of a compact DFG laser system developed in our laboratory. Most design considerations and system characteristics have been described in detail previously [77]. 5.1

Experimental Set-Up and System Performance

The whole measurement set-up is shown in Fig. 13. The DFG source basically consists of the pump and signal laser and a periodically poled LiNbO3 (PPLN) crystal. The quasi-phase-matching condition is met by controlling the temperature of the PPLN crystal using a specially designed oven. A unique feature of our DFG source is the combination of a pulsed pump laser with a continuous wave (CW) signal laser. In this configuration, OPG/OPA processes are considerably suppressed, especially when the system is run at low output DFG power (below 2 mW). An appreciable average power (up to 14 mW) has been generated, yet at the loss of the narrow linewidth due to OPG/OPA processes competing with the DFG process. For a pure OPG/OPA process, the linewidth was measured to be as wide as 195 GHz compared to only 154 MHz for a pure DFG process. amplifier trigger pulse Si-photodiode CRD BS

ECDL

PAS

(1500 - 1610 nm, cw)

half-wave

L 2 plate

"Signal" laser, 8 mW power

L3

oven

Gefilter

L4

multipass transmission

"Idler", 3.2 - 3.7 µm, 600 W peak power, up to 14 mW average power

Nd:YAG laser

temperature control unit

(1064 nm, pulsed)

L1 "Pump" laser, 5 kW peak power, 8.4 kHz, 6 ns

detector

PPLN

quarter- and half-wave plate

Fig. 13. Set-up of the difference frequency laser source with different detection schemes for gas sensing. ECDL: external cavity diode laser, PPLN: periodically poled LiNbO3 crystal, CRD: cavity ring-down, PAS: photoacoustic spectroscopy, BS: beam-splitter, L1−4 : lenses

126


The pump laser is a diode-pumped pulsed high peak power passively Q-switched Nd:YAG laser in a non-planar ring oscillator (NPRO) configuration [78] and sets the pulse repetition rate and the linewidth of 154 MHz, as well as the pulse duration for the generated DFG signal. The signal laser is a continuous-wave (CW) external cavity diode laser (ECDL) tunable between 1500 nm and 1600 nm which sets the tuning range of the DFG laser source. The DFG process takes place in a periodically poled lithium niobate crystal (PPLN). The crystal has eight different gratings with grating periods Λ between 29.1 µm and 30.4 µm. Thus, the whole tuning range of the ECDL is accessible for the DFG process. The main features of the DFG laser source are summarized in Table 9. Table 9. Characteristics of the difference frequency laser source Characteristic Parameter

Data

Wavelength range

3.2 µm–3.7 µm

Line width (at low DFG power)

154 MHz

Temperature acceptance bandwidth a

17 ◦ C

Average power of the DFG signal

0.2 mW–14 mW

Pulse repetition rate

3 kHz–8 kHz

Pulse length

6 ns

At powers above ∼ 2 mW, OPG/OPA processes dominate resulting in an enlarged linewidth

a

5.2

Nd:YAG Pump Laser

The exact wavelength of the difference-frequency laser source can easily be determined using the equation for energy conservation introduced above in (7). However, this approach requires the wavelengths of the pump and the signal lasers to be known with sufficient accuracy. We used a different method: The wavelength of the CW ECDL was measured with a commercial wavemeter, whereas the DFG wavelength was deduced from spectroscopic measurements on methane using the local maximum of the transmission curve of 100 ppm methane diluted in synthetic air (79.5 % N2 , 20.5 % O2 ) at around 3380.2387 nm. The accurate wavelength of the pulsed Nd:YAG laser could then be derived by using (7) [79]. As Fig. 14 implies, the wavelength shows a distinct dependence on the temperature of the Nd:YAG crystal.


Measured data

1064.57 Nd:YAG wavelength in nm

127

1064.55

1064.53

1064.51

Slope: 0.014 nm / °C

1064.49 22

24

26

28

30

32

34

36

Temperature in °C

Fig. 14. Dependence of the Nd:YAG laser wavelength on the temperature of the Nd:YAG crystal

In general, the emitted wavelength increases as a function of the temperature. There are five distinct leaps indicating mode hops. The sections between two successive leaps show a linear wavelength increase with temperature with a slope of 0.014 nm/ ◦ C. 5.3

External Cavity Diode Laser

The signal laser is an external cavity diode laser with a large and continuous tuning range and with a CW laser power of 5–8 mW. The combination of these two lasers results in a high average power whilst still considerably suppressing other nonlinear processes. As seen in the previous section (Tables 7, 8), diode lasers emitting around 800 nm are often used as one of the laser sources. In those cases, however, the low-power tunable laser around 800 nm is the pump laser and lower DFG intensities will result due to pump repletion [36]. The disadvantage of the high pump laser intensity, though, is that other nonlinear optical processes will take place in the nonlinear crystal, especially for an optimized temperature of the nonlinear crystal. As a matter of fact, we observed second and third harmonic generation of the pump wavelength, sum frequency generation between the pump and signal laser wavelengths, as well as optical parametric generation (OPG) and amplification (OPA). The optical parametric processes generate the same wavelength as the DFG process and result in an enhanced output power, yet also in an enlarged linewidth of the laser source and thus in a lower spectral resolution of the system. Hence, the DFG linewidth can be tailored by adjusting the pump laser power. For the DFG process at low conversion efficiency, the generated DFG intensity is proportional to both the intensity of the pump and the intensity of the signal

128


laser (20). For the OPA process at high conversion efficiency, the intensity scales exponentially with the pump laser electric field. 5.4

Tuning with Periodically Poled LiNbO3

Idler Wavelength in µm

The tuning of the generated difference frequency wavelength is carried out varying two parameters. On the one hand, the wavelength of the signal laser is changed in order to reach different wavelengths. Energy conservation (7) determines which wavelength will be generated if the quasi-phase-matching condition (26) is fulfilled, i.e., if the crystal period length ΛQPM is appropriately chosen. The phase-matching condition is achieved by selecting the right grating period of the PPLN crystal (out of eight periods as given by the crystal manufacturer) and by heating the crystal to the right temperature. Although the temperature acceptance bandwidth of LiNbO3 is rather large (17 ◦ C), the crystal still needs to be heated to roughly the optimized temperature. In Fig. 15a, the optimized temperatures for efficient frequency conversion are shown for five different grating periods of the PPLN. 29.5 µm grating period 29.7 µm grating period 29.9 µm grating period 30.0 µm grating period 30.2 µm grating period

3.6 3.5 3.4 3.3 3.2 3.1 20

30

40

a)

50

60 70 80 90 100 Crystal Temperature in °C

110

120

Idler Wavelength in µm

4.0

130

29° C 47° C 63° C 81° C 97° C 114° C Corresponding calculated Sellmeier curves

3.8 3.6 3.4 3.2 3.0 2.8 28.5 b)

29.0

29.5 30.0 Crystal Period in µm

30.5

31.0

Fig. 15. Tuning characteristics for different grating periods of the PPLN crystal. (a) Temperature tuning for five different grating periods. The overlap of the generated ranges of idler wavelengths from different grating periods is clearly visible. (b) Generated idler wavelength as a function of the grating period for different temperatures. The solid lines represent the calculated Sellmeier curves [80]


129

Figure 15b shows the generated idler wavelength as a function of the grating period for different temperatures. The calculated Sellmeier curves (solid lines) well coincide with the measured data points. Only by selecting the right grating and by varying both parameters – temperature and signal wavelength – is efficient phase matching over a large wavelength range possible. 5.5

Detection Schemes

In view of applications of the described continuously tunable DFG laser source in the area of gas sensing, we implemented different detection schemes into the set-up as shown in Fig. 13 above. To date the system has been mainly operated with two different detection schemes. 5.5.1

Transmission Spectroscopy

Transmission spectroscopy relies on the measurement of both the incident and the transmitted intensity. On the basis of the familiar Beer–Lambert absorption law, the gas concentration can be determined if the molecular absorption cross-section and the absorption pathlength are known [13]. The detection sensitivity can be enhanced by extending the pathlength and reducing the noise originating from the detector or the laser source. Thus, often multipass absorption cells are used. We implemented a Herriot-type multipass cell with a total pathlength of 36 m into our system. The total throughput of the empty cell is about 23 % at wavelengths around 3 µm. The advantage of this detection method is the high sensitivity achievable even at low laser power. As transmission spectroscopy is an indirect absorption measurement, i.e., a small deviation from a large signal is measured, the sensitivity depends on the signal-to-noise ratio of the detectors used to record both the incident and the transmitted intensity. For low incident laser powers, cooled detectors designed for low powers can be used which usually exhibit a better signal-to-noise ratio. Thus, the small deviation of the transmitted power from the incident laser power can be detected more easily . The disadvantage of this indirect detection technique is the distinct wavelength characteristic of the mirror reflectivity. In addition, the adjustment of the cell is rather difficult. 5.5.2

Photoacoustic Spectroscopy

As an alternative, we developed a small single pass resonant photoacoustic cell for trace-gas detection with the described DFG system. The cell is equipped with four microphones and has a resonance frequency which coincides with the pulse repetition rate of the laser (5.7 kHz) [37].

130


Photoacoustics has the great advantage that it is a direct absorption measurement, i.e., the measured signal is directly proportional to the absorbed laser power and thus the absorption coefficient of the gaseous sample [81]. This is in contrast to transmission spectroscopy which is an indirect absorption measurement technique where a small intensity difference between two measured signals yields the absorbed laser power. Furthermore, the detection scheme itself is wavelength independent and operated at room temperature since the signal detection is performed by means of microphones. The disadvantage of the photoacoustic detection scheme is that the signal scales with the laser power, i.e., in the case of low laser power like for our compact DFG laser source where only a few mW average power is achieved, it is essential to use strong absorption lines for sensitive gas monitoring. In addition, the pulse repetition rate needs to be stable enough to take advantage of the resonant detection scheme. 5.5.3

Cavity Ring-Down Spectroscopy

A third detection scheme to be considered is cavity ring-down (CRD) [82]. In this case, the laser light is injected into an external cavity usually consisting of two highly reflecting mirrors with a typical reflectivity of > 99.9 % for the wavelength of interest. The ring-down of the radiation leaking out of the cavity is monitored as a function of time with a fast and sensitive detector. The resulting high sensitivity of the detection method originates from the very long pathlength (up to several km). The advantage of cavity ring-down spectroscopy (CRDS) is that it is again a direct absorption measurement. The transmitted intensity is recorded as a function of time. The decay time then allows the absorption coefficient to be determined. Since the decay time can be determined from each single pulse entering the cavity, the pulse-to-pulse fluctuations (both in intensity and in the time between two consecutive pulse) do not alter the sensitivity of the system nor do they add any additional error to the measurement. The disadvantage of CRDS is the requirement of highly reflecting mirrors which can only be accomplished for a certain wavelength range ∆ λ, typically ∆ λ/λctr ≈ 10%, where λctr refers to the center wavelength of the high mirror reflectivity. This implies that unlike in photoacoustic spectroscopy, measurements in a different wavelength region require a different set of mirrors. A further aspect of mid-IR CRDS, especially of pulsed CRDS, concerns the need for sensitive mid-IR detectors. Despite cryogenic cooling, these detectors are still less sensitive than photomultipliers commonly used in the visible or near-IR wavelength range. This implies that higher powers are advantageous for achieving high sensitivities.


131

6 Applications of Mid-IR Difference Frequency Laser Systems in Gas Sensing Applications of difference frequency laser systems are manifold. Most applications, though, are found in gas sensing. In the first part of this section, we briefly outline different fields of gas sensing applications. In the second part, we present multicomponent trace-gas measurements performed with the DFG laser system introduced in Sect. 5. 6.1

General Considerations

Although other applications of tunable mid-IR laser sources can be envisaged, the main area of application is in chemical sensing, particularly in trace-gas monitoring and analysis. There are several fields of interest. 6.1.1

Environment

Potential applications include atmospheric monitoring of pollutants, of photochemical smog compounds, of volcanic, or greenhouse gases such as methane, analyses of unknown gas samples, e.g., from industrial stack emission, road emissions, etc. 6.1.2

Industry

Potential applications are gas sensing in process industry to optimize certain processes and simultaneously minimize dangerous emissions, e.g., in the semiconductor or chemical production industry. Another issue is the surveillance of working place concentrations. Many working places are subject to potentially dangerous concentrations of hazardous compounds where continuous monitoring is of great importance to ensure the safety of the workers. 6.1.3

Agriculture

Fruit fermenting processes are known to result in the emission of certain gases such as ethene, methanol, ethanol, water, carbon dioxide, etc. [83,84]. The continuous monitoring of the concentrations of these gases can thus help to optimize the storage conditions. Other applications concern sensing of, e.g., ethane emission of plants, detection of wilting of flowers or deterioration of meat. 6.1.4

Medicine

The analysis of the exhaled human breath could be used as a valuable noninvasive diagnostic tool for certain human diseases, e.g., for the early detection of ulcers, colon cancer or diabetes [85].

132


6.2 Examples of Gas Spectroscopy Performed with Our DFG Laser Source The main advantages of the DFG system described above is its large and continuous tuning range between 3.2 µm and 3.7 µm. This wavelength region coincides with the absorption range of many atmospheric trace-gases, e.g., methane, formaldehyde, ethane, carbon monoxide, carbon dioxide, ammonia, etc. The small linewidth (154 MHz) of the system allows for both high resolution and highly selective measurements. The system is thus ideally suited to measure and analyze multicomponent gas samples. In Table 10, some compounds of interest absorbing in the accessible wavelength range of the DFG source are listed, along with their main absorption line and absorption strength. The absorbance is calculated for 100 ppm of the particular gas at a total pressure of 1 bar and room temperature and for an absorption path length of 1 m. The last column gives the maximum permissible working place concentration (MAK). A typical measurement of a two-component gas sample using the photoacoustic detection scheme is shown in Fig. 16. It depicts a recording of methane and ethane around 3.3 µm at room temperature and 1 bar total pressure. The data clearly demonstrates the advantage of both the large continuous tuning range and the narrow linewidth. The absorption lines of the two Table 10. Gases absorbing in the 3.2–3.7 µm range and their absorption strength Compound

Formula

Absorbancea

Wavelength ( µm)

MAKb (ppmV)

Ammonia

NH3

0.011

3.456

Carbonyl sulfide

OCS

0.018

3.410

Ethane

C2 H6

0.383

3.348

Formaldehyde

H2 CO

0.239

3.574

0.5

Methane

CH4

0.181

3.381

10000

0.493

3.270

Methyl bromide

CH3 Br

3.382

20 10000

5

Methyl chloride

CH3 Cl

0.062

3.372

50

Hydrogen chloride

HCl

0.613

3.375

5

Nitrogen dioxide

NO2

0.046

3.458

3

Ozone

O3

0.0027

3.460

0.1

Absorbance A = log(1/T ), T = transmission for a concentration of 100 ppm diluted in synthetic air (20.5 % oxygen, 79.5 % nitrogen) at room temperature and 1 bar total pressure b MAK = maximum permissible working place concentration as regulated by the Swiss Federal Government a


30.0 2.0.10-3

20.0 10.0

0.0

0.0 3.2696

3.2700 3.2704 Wavelength in µm Normised PA signal in mV/W

20.0

10.0

10 ppm methane 9 ppm ethane

4.0.10-3

methane

3.2708 3.0.10-3

30.0 ethane

2.0.10-3

20.0

1.0.10-3

10.0

Absorbance

30.0

40.0

Absorbance

Normised PA signal in mV/W

Normised PA signal in mV/W

40.0

133

0.0 0.0

3.3472

3.3476

3.3480

3.3484

3.3488

Wavelength in µm

0.0

3.27

3.28

3.29

3.30 3.31 Wavelength in µm

3.32

3.33

3.34

3.35

Fig. 16. A measurement of a gas sample containing 9 ppmV ethane and 10 ppmV methane diluted in synthetic air at room temperature and 1 bar total pressure

gases can easily be distinguished and recorded at high resolution. Despite the rather low average DFG power, photoacoustic measurements are feasible. A measurement with the resolution shown in the insets of the figure takes about 40 minutes each. Between the measurements only the temperature of the PPLN crystal needs to be changed. As PPLN has a rather high temperature acceptance bandwidth, the temperature only needs to be optimized within a few ◦ C. From these measurements, the current detection limits can be determined to be about 1 ppmV for both methane and ethane at a signal-to-noise ratio of 3. These detection limits are low enough to use the system for air monitoring at working places and can further be lowered by increasing the average laser power and improving the photoacoustic cell design. Measurements with the multipass cell yield similar results as shown in Fig. 17 for a three-component mixture of NO2 , N2 O, and H2 O vapour. Despite the large tuning range of the system, highly resolved measurements are possible due to the narrow linewidth and the fact that the injected laser intensity plays a minor role for multipass transmission measurements provided the detectors are sensitive enough, i.e., the detector noise level is sufficiently small. The sensitivity can considerably be enhanced by increasing the averaging time for a single measurement point. The detection limits using the 36 m multipass transmission cell are in the ppbV range for many gaseous species, e.g., 7 ppbV for methane, 3.5 ppbV for ethane, 80 ppbV for nitrogen

134


transmission

1.000

0.995

0.990

2888.3 cm-1

0.985 calculated NO2 conc.: 3.44 ± 0.22 ppm averaging time: 5 s 0.980 0.0

0.1

0.2

0.3

relative wavenumber in cm-1

transmission

1.000

0.995

0.990

0.985 0.0

2786.0 cm-1 calculated N2 O conc.: 51.3 ± 4.3 ppm averaging time: 10 s 0.1

0.2

0.3

0.4


transmission

1.00

0.95

2966.0 cm-1

0.90

0.85 0.0

calculated H2 O conc.: 0.250 ± 0.006 % averaging time: 1 s 0.1

0.2

0.3

0.4


Fig. 17. Example of multicomponent trace-gas analysis. The baseline corrected spectra were recorded using the 36 m multipass transmission cell filled with a mixture of three different gases at the indicated concentrations diluted in synthetic air at atmospheric pressure. The large tuning range allows measurements at many different wavelengths, e.g., at 2888.3 cm−1 for NO2 , 2786.0 cm−1 for N2 O, and 2966.0 cm−1 for H2 O vapour


135

dioxide, and 30 ppbV for formaldehyde for an averaging time of 60 s for each data point and a signal-to-noise ratio of 3. The DFG laser source based spectroscopic system presented is thus suitable for trace-gas monitoring. The system performance in terms of achieved detection limits and detection selectivity meets the requirements for most applications by far. The compact room temperature operated system can thus be employed for a whole variety of gas sensing applications as briefly discussed above.

7

Conclusions and Outlook

Mid-infrared laser systems are expected to play an increasing role in the future. Although laser sources emitting at a fixed, desired wavelength are of interest for specific studies, by far more applications require tunable lasers for optical sensing. In order to access the important finger print region of molecular absorption for spectroscopic applications, difference frequency generation represents a powerful and useful tool. The generated average power in the µW to mW range allows measurements with most common detection schemes. The wavelength range accessible can be chosen between, e.g., 2 µm and 19 µm depending on available pump and signal sources and nonlinear crystals. The tuning of a DFG source is straightforward, and the wavelength generated can simply be calculated. The linewidth of the generated coherent light is basically given by the linewidth of the pump and signal laser allowing for narrow linewidths needed for high resolution spectroscopy. The set-up of a DFG laser system can be made compact, especially if diode lasers are used as pump and signal sources. With the advent of ever improved solid-state and diode lasers to be employed as pump and signal sources in a DFG laser set-up, further improvements with respect to performance can be expected. Potential also lies in the development of nonlinear crystals and periodically poled materials with either better quality and higher damage thresholds or even of novel materials with larger nonlinear coefficients and extended transparency ranges enabling higher conversion efficiencies and hence output powers, as well as a further extension of accessible wavelengths. Acknowledgements The authors thank a number of coworkers for their valuable input. Specially acknowledged are Dr. M. Seiter for experimental contributions to this work and Dr. I. Biaggio for critically reading the manuscript.

136


References 1. P. L. Hanst, S. T. Hanst: Gas measurements in the fundamental infrared region, in Air Monitoring by Spectroscopic Techniques Chem. Anal. Ser. 127, M. W. Sigrist (Ed.) (Wiley, New York 1994) pp. 335–470 97 2. L. S. Rothman, C. P. Rinsland, A. Goldman, S. T. Massie, D. P. Edwards, J. M. Flaud, A. Perrin, C. Camy-Peyret, V. Dana, J. Y. Mandin, J. Schroeder, A. McCann, R. R. Gamache, R. B. Wattson, K. Yoshino, K. V. Chance, K. W. Jucks, L. R. Brown, V. Nemtchinov, P. Varanasi: The HITRAN molecular spectroscopic database and HAWKS (HITRAN atmospheric workstation), 1996 edition, J. Quant. Spectrosc. Radiat. Transfer 60, 665–710 (1998) 98 3. W. J. Wittemann: The CO2 Laser, Springer Ser. Opt. Sci. 53 (Springer, Berlin, Heidelberg 1987) 98 4. P. Repond, M. W. Sigrist: Continuously tunable high-pressure CO2 laser for spectroscopic studies on trace gases, IEEE J. Quant. Electr. 32(9), 1549–1559 (1996) 98 5. W. Urban: Physics and spectroscopic applications of carbon monoxide lasers, a review, Infrared Phys. Technol. 36(2), 465–473 (1995) 98 6. J. Faist, F. Capasso, D. L. Sivco, C. Sirtori, A.-L. Hutchinson, A.-Y. Cho: Quantum cascade laser, Science 264, 553–556 (1994) 99 7. F. Capasso, J. Faist, C. Sirtori, A.-Y. Cho: Infrared (4–11 µm) quantum cascade lasers, Solid State Commun. 102, 231–236 (1997) 99 8. F. Capasso, C. Gmachl, D. L. Sivco, A.-Y. Cho: Quantum cascade lasers, Phys. World 12, 27–33 (1999) 99 9. J. Faist: private communication (2002) 99 10. A. A. Kosterev, F. K. Tittel: Chemical sensors based on quantum cascade lasers, IEEE J. Quantum Electron. 38(6), 582–591 (2002) 99 11. D. Hofstetter, M. Beck, J. Faist, M. Naegele, M. W. Sigrist: Photoacoustic spectroscopy with quantum cascade distributed feedback lasers, Opt. Lett. 26(12), 887–889 (2001) 99 12. B. A. Paldus, T. G. Spence, R. N. Zare, J. Oomens, F.-J. M. Harren, D. H. Parker, C. Gmachl, F. Capasso, D. L. Sivco, J. N. Baillargeon, A.-L. Hutchinson, A. Y. Cho: Photoacoustic spectroscopy using quantum-cascade lasers, Opt. Lett. 24 (3), 178–180 (1999) 99 13. H. I. Schiff, G. I. Mackay, J. Bechara: The use of tunable diode laser absorption spectroscopy for atmospheric measurements, in Air Monitoring by Spectroscopic Techniques Chem. Anal. Ser. 127, M. W. Sigrist (Ed.) (Wiley, New York 1994) pp. 239–333 100, 129 14. M. Tacke: New developments and applications of tunable IR lead salt lasers, Infrared Phys. Technol. 36(1), 447–463 (1995) 100 15. P. Werle, F. Slemr, K. Maurer, R. Kormann, R. M¨ ucke, B. J¨ anker: Near- and mid-infrared laser-optical sensors for gas analysis, Opt. Lasers Engin. 37, 101– 114 (2002) 100 16. A. Boschetti, D. Bassi, E. Iacob, S. Iannotta, L. Ricci, M. Scotoni: Resonant photoacoustic simultaneous detection of methane and ethylene by means of a 1.63-µm diode laser, Appl. Phys. B 74, 273–278 (2002) 100 17. J. D. Jackson: Classical Electrodynamics, 3rd ed. (Wiley, New York 1999) 101 18. P. A. Franken, A. E. Hill, C. W. Peters, G. Weinreich: Generation of optical harmonics, Phys. Rev. Lett. 7(4), 118–119 (1961) 101


137

19. M. Bass, P. A. Franken, J. F. Ward, G. Weinreich: Optical rectification, Phys. Rev. Lett. 9(11), 446–448 (1962) 103 20. D. A. Roberts: Simplified characterization of uniaxial and biaxial nonlinear optical crystals: a plea for standardization of nomenclature and conventions, IEEE J. Quant. Elect. 28(10), 2057–2074 (1992) 104, 105, 106, 120 21. V. G. Dmitriev, G. G. Gurzadyan, D. N. Nikogosyan: Handbook of Nonlinear Optical Crystals, (Springer, Berlin, Heidelberg 1993) 106 22. Eksma Co. Lasers, Optics, Electronics, Lithuania: http://www.eksma.lt 106, 120 23. D¨ ohrer Elektrooptik, Germany: http://www.doehrer-elektrooptik.de/ 106, 120 24. Crystal Technology, Inc., California: http://www.crystaltechnology.com/productdata.html 106, 120 25. J. Q. Yao, T. S. Fahlen: Calculations of optimum phase match parameters for the biaxial crystal KTiOPO4 , J. Appl. Phys. 50(1), 65–68 (1984) 107 26. M. Kaschke, C. Koch: Calculation of nonlinear optical polarization and phase matching in biaxial crystals, Appl. Phys. B 49, 419–423 (1989) 107 27. R. L. Byer, R. L. Herbst: Parametric oscillation and mixing, in Nonlinear Infrared Generation, Topics Appl. Phys. 16, Y. R. Shen (Ed.) (Springer, Berlin, Heidelberg 1977) pp. 81–137 111 28. G. D. Boyd, A. Ashkin, J. M. Dziedzic, D. A. Kleinman: Second-harmonic generation of light with double refraction, Phys. Rev. A 137(4), 1305–1320 (1965) 113 29. T. B. Chu, M. Broyer: Intracavity CW difference frequency generation by mixing three photons and using Gaussian laser beams, J. Phys.(Paris) 46, 523–533 (1985) 114 30. G. D. Boyd, A. Ashkin: Theory of parametric oscillator threshold with singlemode optical masers and observation of amplification in LiNbO3 , Phys. Rev. A 146(1), 187–198 (1966) 114 31. L. E. Myers, R. C. Eckardt, M. M. Fejer, R. L. Byer, W. R. Bosenberg, J. W. Pierce: Quasi-phase-matched optical parametric oscillators in bulk periodically poled LiNbO3 , J. Opt. Soc. Am. B 12(11), 2102–2116 (1995) 117 32. Gs¨ anger Optoelektronik GmbH: http://www.linos-photonics.de/en/homefr2.html 120 33. V. G. Dmitriev, G. G. Gurzadyan, D. N. Nikogosyan: Handbook of Nonlinear Optical Crystals, 2nd rev. ed. (Springer, Berlin, Heidelberg 1997) 120 34. J. S. Browder, S. S. Ballard, P. Klocek: Physical property comparisons of infrared optical materials, in Handbook of Infrared Optical Materials, Opt. Engin. Ser. 127, P. Klocek (Ed.) (Marcel Dekker, New York 1991) pp. 155–425 120 35. Caladox Ltd, Scotland, UK: Product Catalog 120 36. M. Seiter, M. W. Sigrist: Trace-gas sensor based on mid-IR difference-frequency generation in PPLN with saturated output power, Infrared Phys. Technol. 41 (5), 259–269 (2000) 121, 127 37. C. Fischer, M. W. Sigrist, Q. Yu, M. Seiter: Photoacoustic monitoring of trace gases by use of a diode-based difference frequency laser source, Opt. Lett. 26(20), 1609–1611 (2001) 121, 124, 129 38. A. S. Pine: Doppler-limited molecular spectroscopy by difference-frequency mixing, J. Opt. Soc. Am. 64(12), 1683–1690 (1974) 122 39. T. Momose, T. Wakabayashi, T. Shida: Tunable-narrow-linewidth continuouswave mid-infrared light generation by difference-frequency mixing, J. Opt. Soc. Am. B 13(8), 1706–1712 (1996) 123

138


40. O. Pfister, M. M¨ urtz, J. S. Wells, L. Hollberg, J. T. Murray: Division by 3 of optical frequencies by use of difference-frequency generation in non-critically phase-matched RbTiOAsO4 , Opt. Lett. 21(17), 1387–1389 (1996) 123 41. J. Hong, A.-D. O. Bawagan, S. Charbonneau, A. Stolow: Broadly tunable femtosecond pulse generation in the near and mid-infrared, Appl. Opt. 36(9), 1894– 1897 (1997) 123 42. M. H. Chou, K. R. Parameswaran, M. M. Fejer, I. Brener: Multiple-channel wavelength conversion by use of engineered quasi-phase-matching structures in LiNbO3 waveguides, Opt. Lett. 24(16), 1157–1159 (1999) 123 43. M.-H. Chou, J. Hauden, M. A. Arbore, M. M. Fejer: 1.5-µm-band wavelength conversion based on difference-frequency generation in LiNbO3 waveguides with integrated coupling structures, Opt. Lett. 23(13), 1004–1006 (1998) 123 44. G. M. Gibson, G. A. Turnbull, M. Ebrahimzadeh, M. H. Dunn, H. Karlsson, G. Arvidsson, F. Laurell: Temperature-tuned difference-frequency mixing in periodically poled KTiOPO4 , Appl. Phys. B 67(5), 675–677 (1998) 123 45. B. Golubovic, M. K. Reed: All-solid-state generation of 100-kHz tunable midinfrared 50-fs pulses in type I and type II AgGaS2 , Opt. Lett. 23(22), 1760–1762 (1998) 123 46. H. K. Nienhuys, P.-C.M. Planken, R.-A. van Santen, H. J. Bakker: Generation of mid-infrared pulses by χ(3) difference frequency generation in CaF2 and BaF2 , Opt. Lett. 26(17), 1350–1352 (2001) 123 47. D. Hofmann, G. Schreiber, C. Haase, H. Herrmann, W. Grundk¨ otter, R. Ricken, W. Sohler: Quasi-phase-matched difference-frequency generation in periodically poled TiLiNbO3 channel waveguides, Opt. Lett. 24(13), 896–898 (1999) 123 48. D. L. Fenimore, K. L. Schepler, D. Zelmon, S. K¨ uck, U. B. Ramabadran, P. Von Richter, D. Small: Rubidium titanyl arsenate difference-frequency generation and validation of new Sellmeier coefficients, J. Opt. Soc. Am. B 13(9), 1935– 1940 (1996) 123 49. K. S. Abedin, S. Haidar, Y. Konno, C. Takyu, H. Ito: Difference frequency generation of 5–18 µm in a AgGaSe2 crystal, Appl. Opt. 37(9), 1642–1646 (1998) 123 50. P. Bravetti, A. Fiore, V. Berger, E. Rosencher, J. Nagle, O. Gauthier-Lafaye: 5.2–5.6-µm source tunable by frequency conversion in a GaAs-based waveguide, Opt. Lett. 23(5), 331–333 (1998) 123 51. B. Sumpf, D. Rehle, K. Kelz, H. D. Kronfeldt: A tunable diode-laser spectrometer for the MIR region near 7.2 µm applying difference-frequency generation in AgGaSe2 , Appl. Phys. B 67(3), 369–373 (1998) 123 52. V. Petrov, C. Rempel, K. P. Stolberg, W. Schade: Widely tunable continuouswave mid-infrared laser source based on difference-frequency generation in AgGaS2 , Appl. Opt. 37(21), 4925–4928 (1998) 123 53. K. L. Vodopyanov, P. G. Schunemann: Efficient difference-frequency generation of 7–20-µm radiation in CdGeAs2 , Opt. Lett. 23(14), 1096–1098 (1998) 123 54. D. Lee, T. Kaing, J. J. Zondy: An all-diode-laser-based, dual-cavity AgGaS2 CW difference-frequency source for the 9–11 µm range, Appl. Phys. B 67(3), 363–367 (1998) 123 55. D. Zheng, L. A. Gordon, Y. S. Wu, R. S. Feigelson, M. M. Fejer, R. L. Byer, K. L. Vodopyanov: 16-µm infrared generation by difference-frequency mixing in diffusion-bonded-stacked GaAs, Opt. Lett. 23(13), 1010–1012 (1998) 123 56. M. G. Bawendi, B. D. Rehfuss, T. Oka: Laboratory observation of hot bands of H+ 3 , J. Chem. Phys. 93(9), 6200–6209 (1990) 124


139

57. K. P. Petrov, A. T. Ryan, T. L. Patterson, L. Huang, S. J. Field, D. J. Bamford: Mid-infrared spectroscopic detection of trace gases using guided-wave difference-frequency generation, Appl. Phys. B 67(3), 357–361 (1998) 124 58. M. Seiter, M. W. Sigrist: On-line multicomponent trace-gas analysis with a broadly tunable pulsed difference-frequency laser spectrometer, Appl. Opt. 38(21), 4691–4698 (1999) 124 59. L. Goldberg, J. Koplow, D. G. Lancaster, R. F. Curl, F. K. Tittel: Midinfrared difference-frequency generation source pumped by a 1.1–1.5 µm dualwavelength fiber amplifier for trace-gas detection, Opt. Lett. 23(19), 1517–1519 (1998) 124 60. U. Gustafsson, J. Sandsten, S. Svanberg: Simultaneous detection of methane, oxygen and water vapour utilising near-infrared diode lasers in conjunction with difference-frequency generation, Appl. Phys. B 71(6), 853–857 (2000) 124 61. D. Richter, M. Erdelyi, R. F. Curl, F. K. Tittel, C. Oppenheimer, H. J. Duffell, M. Burton: Field measurements of volcanic gases using tunable diode laser based mid-infrared and Fourier transform infrared spectrometers, Opt. Lasers Engin. 37 (2), 171–186 (2002) 124 62. K. P. Petrov, A. T. Ryan, T. L. Patterson, L. Huang, S. J. Field, D. J. Bamford: Spectroscopic detection of methane by use of guided-wave diode-pumped difference-frequency generation, Opt. Lett. 23(13), 1052–1054 (1998) 124 63. M. Seiter, D. Keller, M. W. Sigrist: Broadly tunable difference-frequency spectrometer for trace gas detection with noncollinear critical phase matching in LiNbO3 , Appl. Phys. B 67(3), 351–356 (1998) 124 64. K. P. Petrov, S. Waltman, E. J. Dlugokencky, M.-Arbore, M. M. Fejer, F. K. Tittel, L. W. Hollberg: Precise measurement of methane in air using diodepumped 3.4-µm difference-frequency generation in PPLN, Appl. Phys. B 64(5), 567–572 (1997) 124 65. K. Fradkin-Kashi, A. Arie, P. Urenski, G. Rosenman: Mid-infrared difference frequency generation in periodically poled KTiOAsO4 and application to gas sensing, Opt. Lett. 25(10), 743–745 (2000) 124 66. A. Arie, K. Fradkin-Kashi, Y. Shreberk: Frequency conversion in novel materials and its application to high resolution gas sensing, Opt. Lasers Engin. 37(2–3), 159–170 (2002) 124 67. D. Rehle, D. Leleux, M. Erdelyi, F. K. Tittel, M. Fraser, S. Friedfeld: Ambient formaldehyde detection with a laser spectrometer based on difference-frequency generation in PPLN, Appl. Phys. B 72, 947–952 (2001) 124 68. K. P. Petrov, R. F. Curl, F. K. Tittel: Compact laser difference-frequency spectrometer for multicomponent trace gas detection, Appl. Phys. B 66(5), 531–538 (1998) 124 69. D. Mazzotti, P. De Natale, G. Giusfredi, C. Fort, J. A. Mitchell, L. W. Hollberg: Difference-frequency generation in PPLN at 4.25 µm: an analysis of sensitivity limits for DFG spectrometers, Appl. Phys. B 70(6), 747–750 (2000) 124 70. D. Mazzotti, G. Giusfredi, P. Cancio, P. De Natale: High-sensitivity spectroscopy of CO2 around 4.25 µm with difference-frequency radiation, Opt. Lasers Engin. 37, 143–158 (2002) 124 71. K. P. Petrov, L. Goldberg, W. K. Burns,, R. F. Curl, F. K. Tittel: Detection of CO in air by diode-pumped 4.6-µm difference-frequency generation in quasiphase-matched LiNbO3 , Opt. Lett. 21(1), 86–88 (1996) 124

140


72. B. Sumpf, K. Kelz, M. N¨ agele, H. D. Kronfeldt: A CW AgGaS2 difference frequency spectrometer with diode lasers as pump sources, Appl. Phys. B 64(5), 521–524 (1997) 124 73. T. T¨ opfer, K. P. Petrov, Y. Mine, D. Jundt, R. F. Curl, F. K. Tittel: Roomtemperature mid-infrared laser sensor for trace gas detection, Appl. Opt. 36(30), 8042–8049 (1997) 124 74. H. D. Kronfeldt, G. Basar, B. Sumpf: Application of a CW tunable infrared spectrometer based on difference-frequency generation in AgGaS2 for selfbroadening investigations of NO at 5 µm, J. Opt. Soc. Am. B 13(9), 1859–1863 (1996) 124 75. W. Chen, G. Mouret, D. Boucher: Difference-frequency laser spectroscopy detection of acetylene trace constituent, Appl. Phys. B 67(3), 375–378 (1998) 124 76. W. C. Eckhoff, R. S. Putnam, S. Wang, R. F. Curl, F. K. Tittel: A continuously tunable long-wavelength CW IR source for high-resolution spectroscopy and trace-gas detection, Appl. Phys. B 63(5), 437–441 (1996) 124 77. M. Seiter: Novel pulsed difference-frequency laser sources for compact midinfrared trace-gas sensors, No. 13354, ETH Z¨ urich (1999), and in Ser. Quantum Electron. 14, 1–94 (Hartung-Gorre, Konstanz 1999) 125 78. I. Freitag, A. T¨ unnermann, H. Welling: Passively Q-switched Nd:YAG ring lasers with high average output power in single-frequency operation, Opt. Lett. 22(10), 706–708 (1997) 126 79. C. Fischer, E. Wettstein, M. W. Sigrist: Accurate spectroscopic wavelength determination of pulsed lasers, Appl. Phys. B 73(3), 233–237 (2001) 126 80. D.-H.. Jundt: Temperature-dependent Sellmeier equation for the index of refraction, ne , in congruent lithium niobate, Opt. Lett. 22(20), 1553–1555 (1997) 128 81. M. W. Sigrist: Air monitoring by laser photoacoustic spectroscopy, in Air Monitoring by Spectroscopic Techniques, Chem. Anal. Ser. 127, M. W. Sigrist (Ed.) (Wiley, New York 1994) pp. 163–238 130 82. G. Berden, R. Peeters, G. Meijer: Cavity ring-down spectroscopy: experimental schemes and applications, Int. Rev. Phys. Chem. 19(4), 565–607 (2000) 130 83. M. N¨ agele, M. W. Sigrist: Mobile laser spectrometer with novel resonant multipass photoacoustic cell for trace-gas sensing, Appl. Phys. B 70(6), 895–901 (2000) 131 84. H.-S.-M. de Vries, F.-J. M. Harren, L.-A.-C. J. Voesenek, C.-W.-P. M. Blom, E. J. Woltering, H.-C.-P.-M. van der Valk, J. Reuss: Investigation of local ethylene emission from intact Cherry tomatoes by means of photothermal deflection and photoacoustic detection, J. Plant Physiol. 107, 1371–1377 (1995) 131 85. L. G. Sandstrom, S. H. Lundquist, A. B. Petterson, M. S. Shumate: Tunable diode laser spectroscopy at 1.6 and 2 µm for detection of Helicobacter pylori infection using 13 C-urea breath test, in Selected Topics in Quantum Electronics, IEEE J. Quantum Electron. 5(4), 1040–1048 (1999) 131

Index

β-BaB2 O4 (BBO), 106, 120 absorption – fundamental, 97 acceptance angle, 109, 110 acceptance bandwidth, 109 ADP, 106, 120 AgAsS3 , 120 AgGaS2 (AGS), 106, 120 AgGaSe2 (AGSe), 106, 120 AgSbS3 , 106, 120 ammonia, 132 angle tuning, 108 aperture distance, 113 Ba2 NaNb5 O15 , 106 Ba2 TiO3 , 106 BeSO4 , 106 cavity ring-down spectroscopy (CRDS), 130 CdGeAs2 (CGA), 106, 120 CdS, 106 CdSe, 106, 120 CO (carbon monoxide), 132 CO laser, 98 CO2 , 132 coefficient – effective nonlinear optical, 104, 111 – nonlinear optical, 103, 106, 116, 117, 120 coherence length, 115 conversion efficiency, 110 critical phase matching (CPM), 110 crystal, 106, 117, 120, 121 – biaxial, 106 – biaxial nonlinear, 105 – centrosymmetric, 105

– class, 105 – group, 112, 113 – monoclinic, 105 – negative birefringent, 104, 112 – orthorhombic, 105 – positive birefringent, 104, 112 – triclinic, 105 – uniaxial, 104, 106 – uniaxial negative, 104 CsH2 AsO4 (CDA), 106 CsLiB6 O10 (CLBO), 117, 120 CsTiOAsO4 (CTA), 106, 120 damage threshold, 118, 120 detection limit, 133 detection scheme, 129 difference frequency generation (DFG), 102, 107, 108, 122 – system, 122 dipole moment, 100 dispersion, normal, 114 domain length, 115 effective nonlinear coefficient, 104, 111 electric-dipole – approximation, 101 energy conservation, 103, 114, 128 environment, 131 ethane, 132 formaldehyde, 132 frequency conversion process, 100 GaAs, 106 GaP, 106 gas sensing, 122, 131 GaSe, 106, 120 HgGa2 S4 , 120

142

Index

HgS, 120 HITRAN database, 98 idler, 102 index ellipsoid, 104 indicatrix, 104 InP, 106 interaction length, 111 K3 Li2 Nb5 O15 , 106 KD2 PO4 (KD∗ P), 106, 120 KH2 PO4 (KDP), 106, 120 KNbO3 , 106, 117, 120 KTiOAsO4 (KTA), 106, 120 KTiOPO4 (KTP), 106, 117, 120 laser – CO, 98 – CO2 , 98 – ECDL, 100, 121, 126, 127 – external cavity diode, 100, 121, 126, 127 – lead salt, 100 – Nd:YAG, 126 – quantum cascade (QCL), 99 – signal, 102, 125 LiB3 O5 (LBO), 106, 117, 120 LiIO3 , 106, 120 LiNbO3 , 106, 117, 120 – periodically poled (PPLN), 120, 126, 128 LiTaO3 , 106, 117, 120 MAK, 132 medicine, 131 methane, 126, 132 MgO:LiNb3 , 120 momentum conservation, 107 monolithic non-planar ring oscillator (NPRO), 126 ND4 H2 PO4 , 106 Nd:YAG laser, 126 NH4 H2 PO4 , 106, 120 non-planar ring oscillator (NPRO), 126 nonlinear figure of merit (FOM), 114 nonlinear optical coefficient, 103, 106, 116, 117, 120 – effective, 111, 112

optical parametric amplification (OPA), 121, 127 optical parametric generation (OPG), 102, 121, 127 optical rectification (OR), 103 period length, 117 periodic poling, 115 phase matching, 105, 129 – angle, 108 – birefringent, 107 – critical, 109, 110 – first-order quasi-, 117 – non-critical, 109 – non-perfect, 111 – perfect, 111 – quasi-, 114 – type I, 107, 112 – type II, 107, 112 – type III, 107 phase mismatch, 107 polarization – dielectric, 100 – extra-ordinary, 107 – nonlinear, 101 – of light, 107 – ordinary, 107 potassium titanyl phosphate (KTiOPO4 , or KTP), 106, 117, 120 PP KTA, 119, 120 PP KTP, 119–121 PP RTA, 120, 121 PPLN, 120, 121, 126, 128 pump, 102, 125, 126 quartz, 106, 120 quasi-phase-matching (QPM), 114 – first-order, 117 RbH2 PO4 (RDP), 106 RbTiOAsO4 (RTA), 106, 120 refractive index, 101 – extraordinary, 104 – ordinary, 104 Se, 106 second harmonic generation (SHG), 101, 107, 127

Index signal laser, 102, 125 SiO2 , 106, 120 spectral acceptance, 110 spectroscopy, 97 – cavity ring-down (CRDS), 130 – gas, 132 – photoacoustic, 129 – transmission, 129 sum frequency generation (SFG), 102, 107, 127 susceptibility – linear, 100 – nonlinear, 101 telluride, 106 temperature acceptance bandwidth, 110, 128 temperature tuning, 109 third harmonic generation, 127

transparency – range, 118–120 tuning – angle, 108 – temperature, 109 – wavelength, 109 tuning characteristic, 128 walk-off angle, 113 walk-off distance, 113 wavelength tunability, 118 wavelength tuning, 109 wavevector propagation – collinear, 108 working place concentration, 131 – maximum permissible, 132 ZnGeP2 (ZGP), 106, 120

143

Pulsed Mid-IR Optical Parametric Oscillators Konstantin Vodopyanov Ginzton Laboratory, Stanford University, Stanford, CA 94305-4088, USA [email protected] Abstract. Optical parametric oscillators (OPOs) are the laser sources of choice when one needs broad continuous tunability (one octave or more), high peak (> 1 kW) or average (> 1 W) power, and high (> 30 %) conversion efficiency. OPOs use the nonlinear-optical frequency down-conversion mechanism to get a tunable output. This chapter reviews pulsed OPOs in the 2–20 µm region of the spectrum. It contains a description of the principle of OPO operation, reviews existing and emerging nonlinear optical materials suitable for OPO applications and describes different OPO schemes including cascaded OPOs, narrow-linewidth OPOs as well as traveling-wave devices.

1

Introduction

Frequency conversion using optical parametric oscillators (OPOs) is an effective means of generating coherent light at frequencies where lasers perform poorly or are unavailable. OPOs can generate coherent tunable light ubiquitously from the UV to the infrared and even THz domain. In the infrared, they play a particularly important role in the ‘molecular fingerprint’ region of the spectrum, 2–20 µm, where molecular species have their fundamental absorption features, and where we lack broadly tunable lasers similar to dye lasers in the visible, or titanium:sapphire in the near-infrared. The broad tuning range of optical parametric oscillators (in many cases limited just by the optical transparency of the crystals) and efficient power conversion characteristic make them attractive sources in many applications requiring wide tunability and high peak power. Since the first demonstration of optical parametric oscillation by Giordmaine and Miller in 1965, using a LiNbO3 crystal [1], OPOs have made a transition from being a research curiosity to an actual tool in real applications. This transition happened due to many factors including the advent of novel nonlinear-optical materials with low optical losses and high laserdamage thresholds. Periodically structured nonlinear-optical materials revolutionized nonlinear optics by enabling devices with extremely low pumping thresholds to be built. Dramatic improvement in the optical quality of semiconductor materials having intrinsically very high second-order optical nonlinearities and deep mid-IR transparency made it possible to extend the tuning range of existing optical parametric oscillators far beyond 5 µm. Equally important are the advances in high peak power all-solid-state pump lasers, I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 141–180 (2003) c Springer-Verlag Berlin Heidelberg 2003

142

Konstantin Vodopyanov

including microchip and fiber lasers, and new OPO optical designs including cascaded OPO schemes, optical parametric generator–optical parametric amplifier (OPG–OPA) designs, and others. The robustness and compactness of OPOs make them extremely attractive for new applications like laser spectroscopy (both linear and nonlinear), remote chemical sensing, laser radar imaging, trace gas and vapor detection, biology, medicine and free space communications. The focus of this chapter is pulsed optical parametric oscillators, where typically nanosecond-range lasers are used as a pump source.

2

Principle of OPO Operation

OPOs convert monochromatic laser radiation (pump) into a tunable output via a three-wave mixing process with quantum conversion efficiencies which can exceed 50 %. The heart of an OPO is a nonlinear optical (NLO) crystal which is characterized by a NLO coefficient, deff . In the NLO crystal, the pump photon decays into two less energetic photons (signal and idler) so that the sum of their energies is equal to that of the pump photon (Fig. 1). An important further constraint is that the sum of the signal and idler wavevectors (k-vectors) must equal that of the pump – momentum conservation or ‘phase matching’ (PM) condition [2]. The latter condition is never satisfied in the transparency range of isotropic media, where normal dispersion applies, but can be fulfilled in birefringent crystals. Alternatively, it can be fulfilled in ‘quasi-phase-matched’ (QPM) crystals with periodically modulated nonlinearity (a typical example is periodically poled lithium niobate), where the artificially created grating of optical nonlinearity [3] compensates for the wavevector mismatch (Fig. 2). Speaking in a different language, the PM condition is achieved in QPM structures because the relative phase is corrected at regular intervals using a structural periodicity built into the nonlinear medium. Parametric frequency down-conversion in an OPO can be regarded as the inverse process of sum-frequency generation. Alternatively, a NLO crystal can

hw1 w3 =w1+w2

hw3

hw2

k3 =k1+k2

Fig. 1. Optical parametric frequency conversion. In the NLO crystal, the pump photon decays into two less energetic photons (signal and idler) so that the sum of their energies is equal to that of the pump photon

Pulsed Mid-IR Optical Parametric Oscillators

+-

2p/L K1 K3

Period L

143

K2

Fig. 2. In quasi-phase-matched (QPM) crystals with periodically modulated sign of the nonlinearity (e.g. periodically poled lithium niobate), the artificially created grating compensates for the wavevector mismatch

simply be viewed as a catalyst that promotes decay of the pump photon into two smaller photons. This process depends on the instantaneous intensity of the pump laser beam, thus short laser pulses and tight focusing is an advantage. Rotating the crystal (in the case of birefringent crystals) or changing its temperature changes, via the phase matching condition, the ratio between the signal and idler photon energies, and thus tunes the frequency of the output. The easiest way to illustrate the parametric frequency conversion is to consider the case of a short (duration of a few ns or less) intense (power density ≈ 1 GW/ cm2 ) pump pulse. In this case (Fig. 3a), a single pass through a NLO crystal is sufficient to convert a substantial fraction (> 10 %) of the pump into the signal and the idler. This type of single-pass device is called an optical parametric generator (OPG). Essentially, OPG is an amplifier of quantum noise with a gain factor of 1010 or more. For pump pulses with longer pulse duration (> 5 ns) and, correspondingly lower power density, parametric frequency conversion is weaker and one needs a resonator OPO cavity (e.g. a Fabry–Pérot-type cavity) to enhance this process (Fig. 3b). To achieve a narrow linewidth output, one can incorporate a diffraction grating into the OPO cavity (Fig. 3c). Eventually, when a continuous-wave pump is used and tight focusing is necessary, one can use a four-mirror ring OPO cavity with two flat and two concave mirrors (Fig. 3d). A single-pass power parametric gain factor (for zero phase mismatch) is expressed [2,4] as G = Pout /Pin = cosh2 (Γ L), where L is the length of the nonlinear crystal and Γ is the gain increment given by: 2 2 2 deff 2ω1 ω2 Ipump deff 8π Ipump . (1) = Γ2 = 3 3 n ε0 c n3 λ1 λ2 ε0 c Here Ipump is the pump laser intensity (power density), ω1 and ω2 are idler and signal frequencies, λ1 and λ2 are idler and signal wavelengths, deff is the effective nonlinearity and n is the average refractive index. Also, d2eff /n3 is often referred to as the nonlinear optical figure of merit (NLO FOM) of the crystal.

144

Konstantin Vodopyanov hwpump= hwsignal + hwidler

(a) signal

SINGLE-PASS OPTICAL PARAMETRIC GENERATOR

IN idler

pump

OUT

NONLINEAR CRYSTAL

(b) PULSED OPTICAL PARAMETRIC OSCILLATOR

IN OUT

NONLINEAR CRYSTAL OUTPUT MIRROR

INPUT MIRROR

(c) NARROWLINEWIDTH OPTICAL PARAMETRIC OSCILLATOR

IN

NONLINEAR CRYSTAL INPUT MIRROR

OUT

DIFFRACTION GRATING

(d) CONTINUOUS WAVE OPTICAL PARAMETRIC OSCILLATOR

OUT

IN

NONLINEAR CRYSTAL INPUT MIRROR

OUTPUT MIRROR

Fig. 3. Diferent OPO schemes. (a) When short intense pump pulses are used, a single pass through a NLO crystal (OPG) is enough to convert a substantial fraction (> 10 %) of the pump into the signal and the idler. (b) For the pump pulses with smaller intensity, parametric frequency conversion is weaker and one needs a resonator OPO cavity to enhance this process. (c) For a narrow-linewidth OPO applications, one can incorporate a diffraction grating into the cavity. (d) When a CW pump is used and tight focusing is necessary, one can use a four-mirror ring OPO cavity with two flat and two concave mirrors


145

In the low gain limit (Γ L 1), cosh2 (Γ L) ≈ 1 + (Γ L)2 ,

(2)

while in the high gain limit (Γ L 1), cosh2 (Γ L) ≈ 1/4 × e2Γ L .

(3)

As a numerical example, consider the case of OPO based on periodically poled lithium niobate (PPLN) with λ3 = 1.06 µm (pump), λ2 = 1.6 µm (signal), and λ1 = 3.2 µm (idler). The effective nonlinearity of PPLN in this wavelength range is pretty well established and is deff = 14.2 pm/V [5,6]; the average refractive index is n = 2.131 . At the pump intensity Ipump = 1 MW/cm2 and L = 1 cm we get Γ L = 0.35 and a single-pass parametric gain G = cosh2 (Γ L) = 1.13 (13 % increase in power). At higher pump intensity, 100 MW/cm2 , and L = 4 cm, we get Γ L = 14 and a single-pass gain as high as G = cosh2 (Γ L) = 3.5 × 1011 . To achieve a threshold of an OPO resonating at one of the waves (‘signal’ or ‘idler’) with 10–30 % round-trip losses, and with pumping by 10–100 ns pulses, one needs typically Γ L = 0.3–2. This value is determined by two main factors: compensating the round-trip loss in the cavity and build-up of the intracavity intensity from quantum noise (equivalent input power ≈ nW)2 to a detectable limit, over a limited number of round-trips (10–100) in the OPO cavity [7]. The total build-up of intracavity intensity thus amounts to a factor of about 1010 . In contrast, to achieve a threshold of a single-pass OPG, one needs G ≈ 1010 , corresponding to Γ L ≈ 12; naturally, higher intensities must be applied in this case (e.g. by using short pump pulses).

OPOs in the 2–5 µm Range

3 3.1

Comparison of NLO Materials

The most advanced and suitable nonlinear optical materials for the 2–5 µm range are ferroelectric oxides, namely lithium niobate (LiNbO3 ), potassium titanyl phosphate (KTP), potassium titanyl arsenate (KTA) and rubidium titanyl arsenate (RTA). Periodic poling of all four ferroelectrics is now well developed, using electric field poling techniques [3]. Quasi-phase-matching in periodically poled 1

2

We have used the LiNbO3 coefficient d33 = 27 pm/V reported for the second harmonic generation of 1.06 µm and scaled it to appropriate wavelengths via Miller’s rule; 2/π coefficient was added to account for the effect of quasi phasematching. The equivalent input quantum noise power of an OPO is given by Yariv [2] as (photon energy)×(bandwidth). For λ ≈ 3 µm (photon energy = 6.6 × 10−20 J) and for a typical bandwidth of 1 cm−1 = 3 × 1010 Hz, we get a quantum noise power of 2 nW

146


(PP) nonlinear optical materials offers significant advantages over conventional (birefringent phase matching) schemes for the realization of efficient optical parametric oscillators. These advantages include: • collinear propagation of the pump and the generated waves: there is no birefringent (Poynting vector) walk-off in space, hence tight focusing is possible which results in very low OPO thresholds; • large acceptance angle due to so-called non-critical phase-matching (NCPM); • possibility of using the largest nonlinear optical tensor (e.g. d33 ) element to maximize deff . In PP LiNbO3 (PPLN), for example, quasi-phase-matching with all waves polarized parallel to the z-axis yields a gain enhancement over the birefringently phase-matched process of (2d33 /πd31 )2 ≈ 16; • avoiding limitations on tunability imposed by PM condition – due to extra degree of freedom provided by the possibility of engineering different quasiphase-matched grating periods. Therefore, except for a few cases of high-power OPOs, we will mostly consider OPOs based on quasi-phase-matched materials in this section. Table 1 compares linear and nonlinear optical characteristics of the four periodically poled QPM oxide crystals, namely LiNbO3 (LN), KTP, KTA, RTA. The transparency range is given at the ‘zero’ level of transmission, corresponding to an absorption coefficient ≈ 5 cm−1 . The NLO coefficient corresponds to the largest d33 tensor element for the OPO process around 1.06 µm → 1.6 µm + 3.2 µm and includes the 2/π reduction factor [3] due to the QPM effect. Using the relative (with respect to PPLN) NLO figure of merit from this table, and the numerical example of Sect. 2, one can easily estimate the parametric gain in these crystals at a given pump intensity. Table 1. Comparison of linear and nonlinear optical properties of periodically poled oxide crystals in the mid-IR (from [5,6,8,9]) Crystal

PP PP PP PP

3.2

LN KTP KTA RTA

deff Average Transparency d33 range ( µm) (pm/V) (pm/V) refractive index 0.40–5.5 0.35–4.3 0.35–5.3 0.35–5.3

22.3 16.9 16.2 15.8

14.2 10.8 10.3 10.1

2.13 1.8 1.8 1.8

NLO FOM d2eff /n3 with respect to PPLN 1.00 0.95 0.87 0.83

OPOs Based on PP LN

Owing to its highest nonlinear coefficient among the oxides (Table 1), good optical quality, low absorption and well-established poling technology, peri-


147

odically poled lithium niobate (PPLN) is now one of the most widely used nonlinear materials for generating tunable light in the 1–4 µm range. The first OPO based on a PPLN crystal was reported by Myers et al. [6]. The authors used a 15 mm long, 0.5 mm thick PPLN crystal with a 31-µm grating period. A Q-switched diode-pumped 1.064-µm Nd:YAG laser with 7–20 ns pulse duration and 100 Hz–10 kHz repetition rate was used as a pump. The OPO resonator (Fig. 4) was a linear cavity with mirrors selected to resonate at the signal wave as a singly resonant oscillator (SRO). The mirrors with 25-mm radius of curvature were separated by 33 mm. They were highly reflective (98 % and 92 %) at the signal wavelength, and transmissive (70–95 %) for the pump and the idler. The pump laser was focused in the crystal to a spot size of 47 µm. The OPO pump threshold was 12 µJ and the output was continuously tunable from 1.66 to 2.95 µm, with the crystal temperature varying from room temperature to 180 ◦ C. To extend the tuning range of the PPLN OPO, the same team reported an OPO which used multiple grating sections on a single PPLN ‘chip’. The PPLN chip was 26 mm long and consisted of 25 gratings (0.5 mm wide) with QPM periods from 26 to 32 µm in 0.25-µm steps (Fig. 5) [10]. The SRO oscillation threshold was as low as 6 µJ with a 7 ns pump pulse (fluence 0.09 J/cm2 ). For tuning, the PPLN crystal was translated across the beam so that it interacted with different grating sections and a tunable IR output from 1.36 to 4.83 µm was achieved. The OPO tuning curve, as a function of grating period, is shown in Fig. 6. At the repetition rate of 1 kHz and the average pump power 100 mW, the output power at 4 µm amounted to 6 mW. To exploit the PPLN OPO operation in the regime of long pulses and a long-wave pump, Hansson et al. have used a diode-pumped Tm:YAG laser at 2.01 µm with 530 ns duration as an OPO pump [11]. The AR-coated PPLN OPO crystal (50 × 11 × 0.5 mm3 ) was periodically poled in eight stripes of different QPM periods over the full length of the material. Domain period lengths ranged from 25.5 to 28.7 µm. A pump pulse energy of 5.1 mJ generated 0.65 mJ of signal and idler at a 50 Hz repetition frequency. By changing the temperature and grating period, tuning over a wide range of idler wavelengths, 3.26–5.34 µm, was performed. A PPLN OPO driven by a fiber laser was reported by Britton et al. [12]. The authors used a Q-switched erbium fiber laser near 1560 nm, which operated at 300 Hz and produced pulses of up to 150 µJ and 50 ns duration. The

Fig. 4. Setup for the first PPLN-based OPO using a Q-switched 1.064-µm Nd:YAG laser as a pump [6]

148


Fig. 5. Experimental setup for the multigrating QPM OPO. For tuning, the PPLN crystal is translated through the resonator so the pump beam interacts with different grating sections [10]

Fig. 6. OPO tuning as a function of grating period, achieved by translation of the PPLN crystal through 24 different grating sections. Solid line: Theoretical curve [10]

laser output was frequency doubled to get pulses with up to 60 µJ energy at 780 nm, which were used as an OPO pump. The PPLN sample, 20 mm long and 0.5 mm thick, contained a number of different grating periods (19, 19.5, 20, and 20.5 µm) suitable for 780 nm pumped OPO operation in the signal range 950–1450 nm. The OPO cavity had a concave (R = 10 cm) input coupler and a flat output coupler. Both mirrors were reflective at the signal wave (reflectivity 99.5 and 75 % correspondingly) and transmissive for the pump and the idler band. The pump beam was focused to a spot diameter of 80 µm within the PPLN sample. The OPO threshold was ≈ 10 µJ, and the idler wave tunability 1.69–4.45 µm. The quantum conversion efficiency was ≈ 9 %. To get mid-IR pulses of higher energy, one needs larger PPLN apertures, since the limiting factor is the optical damage of the crystal’s surfaces. An OPO based on a 1 mm thick and 20 mm long PPLN with grating periods


149

ranging from 27 to 30 µm has been reported by Bellomonte et al. [13]. The singly resonant cavity was pumped by a Q-switched diode-pumped Nd:YAG laser. At the pump pulse energy of 8.5 mJ and the beamsize of 0.85 mm, the idler energy of 1 mJ was extracted at 3.9 µm, corresponding to the photon conversion efficiency of ≈ 40 %. Thus, PPLN proves to be an ideal material for generating mid-IR radiation in the 2–4 µm range with moderate pulse energies – of up to 1 mJ. However because of the existence of the unwanted photorefractive effect in undoped lithium niobate at room temperature, it needs to be operated at elevated (> 100 ◦C) temperatures. 3.3

OPOs Based on PP KTP and PP RTA

Periodically poled potassium titanyl phosphate KTiOPO4 (PP KTP) and its isomorphs – rubidium titanyl arsenate RbTiOAsO4 (RTA) and potassium titanyl arsenate KTiOAsO4 (KTA), were successfully developed recently as an alternative OPO material to PPLN. Their advantages are high laser damage threshold (several times higher than in PPLN) and small susceptibility to thermal lensing which makes them suitable for high average power optical parametric oscillation. Furthermore, the coercive field of ≈ 2 kV/mm in these crystals is an order of magnitude lower than that in LiNbO3 . This lower field allows thicker (3 mm) samples to be periodically poled. Also, very low sensitivity to photorefractive damage permits operation at room temperature. Hellstr¨ om et al. [9] used a 1 mm thick and 20 mm long PP KTP crystal to demonstrate an efficient nanosecond OPO pumped by a flash-lamp-pumped Q-switched Nd:YAG laser with 5 ns pulse duration, 20 Hz repetition rate and a focused beam waist of 240 µm. The OPO cavities were 25 mm long and consisted of two flat mirrors that were antireflection coated for the pump and reflective in the signal region. Changing the temperature from 10 to 100 ◦ C resulted in tuning from 1.85 to 1.92 µm for the signal wavelength and from 2.51 to 2.39 µm for the idler. A maximum total (signal + idler) output energy of 1.8 mJ was obtained (in both directions) at a pump level of 3.5 mJ with conversion efficiency reaching 50 %. The OPO threshold was ≈ 0.3 mJ. Peltz et al. have reported optical parametric oscillators with high average powers and high pulse energies using 3 mm thick PPKTP and PP RTA [14]. The samples in this study had the useful length of 8 mm (PP KTP) and 7 mm (PP RTA). In high average power experiments, a diode-pumped Nd:YVO4 laser system was used, which produced nearly diffraction limited (beam quality factor M 2 = 1.2) pulses at 1064 nm. The pulse duration was 5.8 ns and the repetition rate 10–20 kHz. The resonator of the PPKTP OPO consisted of two mirrors resonating at the signal wave; one had a radius of curvature 100 mm and the other (the output coupler) was flat. The signal reflectivity of the output coupler was 64 %. Both mirrors were highly transmissive for the pump and idler wavelength; they were placed close to the crystal surfaces so that the cavity length was 15 mm. The pump beam was focused into

150


the OPO cavity to a beam waist radius of 130 µm. The PPKTP sample was kept at room temperature. At a repetition rate of 10 kHz and 4.8 W pump power, corresponding to a peak intensity of 310 MW/cm2 , the OPO generated 830 mW of signal power (at 1.72 µm) and 500 mW of idler power (at 2.76 µm). The slope efficiency was 33 % and the maximum photon conversion efficiency was 28 %. The measured OPO threshold in this configuration was 900 mW (90 µJ per pulse). At an even higher repetition rate, 20 kHz, the pump power was 7.2 W and the total (signal + idler) output power reached 2 W. Similar experiments were made with the PPRTA-based OPO. The laser repetition rate was 20 kHz and the beam waist inside the cavity was 90 µm. The measured pump power at OPO threshold was 4 W. With a maximum pump power of 8 W (peak intensity 540 MW/cm2 ) the OPO generated 1 W of signal (at 1.58 µm) and 0.27 W of idler (at 3.26 µm) radiation. The conversion efficiency for this particular cavity configuration was 16 %. In the high pulse energy OPO experiments by the same authors, thick PPKTP and PPRTA crystals were pumped by Q-switched flash-lamppumped Nd:YAG lasers at a repetition rate of 10 Hz and pulse duration 10 ns [14]. The pump-beam diameter was approximately 2.5 mm. The OPO consisted of the crystal within a plane parallel cavity. The input mirror was highly reflective for the signal wavelength. The output coupler had a signal reflectivity of 80 %. In the PPKTP case, the OPO threshold was 0.1 J/cm2 , the conversion efficiency reached 38 % with total OPO pulse energies of up to 18.3 mJ. In the case of a PPRTA crystal, the OPO signal-pulse energy reached 17 mJ when pumped with 65 mJ pulses. The pump to signal efficiency was 26 %. From the authors’ point of view [14], large aperture PPKTP and PPRTA crystals show very similar OPO output characteristics and can be used for high power/energy applications. The biggest difference between the two materials is in their transparency: RTA has an advantage over KTP at λ > 3 µm, since its cut-off lies farther into the infrared part of the spectrum. Scaling a nanosecond OPO to high pulse energies entails increasing the beam diameters to avoid surface damage. The OPO cavity, however, should be kept short for high efficiency and low threshold. As a result, the cavity Fresnel number (NF = d2 /λL where d is the beam diameter and L is the cavity length) increases which worsens the spatial coherence across the beam. Hansson et al. [15] used an unstable OPO resonator to improve the beam quality of a PP RTA OPO. The unstable resonator effectively filters out OPO modes with high spatial frequency components by a combination of laser-mode magnification and feedback of only the lowest-order spatial modes. The PP RTA crystal was 8 mm long and 3 mm thick; the QPM period was 40.2 µm so that the OPO generated a signal at 1.56 µm and an idler at 3.33 µm when pumped at 1.064 µm. The OPO cavity was formed by a 30-cm radiusof-curvature concave input coupler (Fig. 7) and a 25-cm radius-of-curvature convex output coupler which formed a confocal unstable resonator of 1.2 times magnification when used with an optical path cavity length of 2.5 cm.


151

Fig. 7. PP RTA-based OPO with an unstable resonator cavity [15]

At the pump beam size of 3 mm, the Fresnel number was ≈ 200. With an unstable resonator, the authors have achieved a factor of three improvement in M 2 over a comparable flat–flat resonator (M 2 of 2.3–2.5 was obtained for the idler wave). The output energies were similar in both cases: ≈ 4.5 mJ (signal + idler) at 30 mJ pump. 3.4

OPOs Based on PP KTA

Another member of the KTP family group is the material KTiOAsO4 (KTA). Its optical properties in the mid-infrared region are similar to those of RTA (a cutoff wavelength 5.3 µm and nonlinear optical coefficient d33 = 16.2 pm/V, Table 1). Rosenman et al. [16] reported the fabrication, by low temperature poling, of PP KTA crystals (0.5 mm thick, 10 mm long) and the observation of optical parametric oscillation in the mid-infrared region. Two uncoated samples were tested in OPO devices pumped by a diode pumped Nd:YAG laser. The OPO resonator consisted of two plane mirrors, 15 mm apart. The input coupler was highly transmitting at the pump wavelength and highly reflecting at the signal wavelength. The CaF2 output coupler was highly reflecting at pump and signal wavelengths and highly transmitting at the idler wavelength. The measured signal and idler wavelengths were 1.505 and 3.631 µm, respectively, for the 39 µm QPM period and 1.439 and 4.083 µm, respectively, for the 37.4 µm QPM period. The maximal total conversion efficiency was 30 % at 20 MW/cm2 pump power density. 3.5

OPOs Using Conventional Phase-matching in Oxides

Despite the fact that QPM crystals are more attractive, as compared to bulk crystals using birefringent phase matching, due to their large acceptance

152


angles, higher gain and lower OPO thresholds, the latter may be beneficial for high power applications, when large apertures are required. For example, Wu et al. have achieved 4.1 W of 3.5 µm output from an optical parametric oscillator using a noncritically phase-matched type-II KTA [17]. The authors have used an intracavity OPO configuration – the OPO was pumped within the cavity of a Q-switched diode-pumped Nd:YALO laser operating at 10 kHz. The laser cavity (Fig. 8) was formed by flat mirrors M1 and M2 with high reflectivity (HR) at 1 µm. The KTA OPO consisted of an x-cut (θ = 90◦ , φ = 0◦ ), 7 × 7 × 35 mm3 KTA crystal placed between flat mirrors M3 and M1. The mirror M3 is AR coated for 1 µm and is highly reflective (R > 99 %) at both 1.5 and 3–4 µm. Besides acting as a high reflecting mirror at 1 µm, M1 was also the output coupler for the OPO. It was AR coated at 3-4 µm and had R = 90 % at 1.5 µm. Hence, this KTA OPO was a singly resonant OPO. Furthermore 4.1 W of the output at 3.5 µm, 10.9 W of 1.5 µm and 11.3 W of 1 µm radiation were obtained simultaneously. A KTP OPO with the average power of 53 W has been achieved by Cheung et al. [18]. As a pump laser, the authors used a diode arraypumped Nd:YAG master oscillator–power amplifier system with repetition rate 20 kHz, pulse duration 40 ns and average power 135 W. The degenerate OPO (that is λsignal = λidler = 2.13 µm) used six KTP crystals, 3 × 3 × 6 mm3 in size, cut for type-II phase matching in a walk-off compensated configuration. The pump beam diameter was 0.6 mm and the reflectivity of the OPO outcoupler mirror was only 30 %. The average power of 53 W (in two orthogonal polarizations) has been achieved at 2.13 µm with 43 % conversion efficiency. Vysniauskas et al. [19] demonstrated a high-energy OPO based on a critically phase-matched (CPM) angular-tuned KTA crystal. The OPO was pumped by a Q-switched Nd:YAG laser (Continuum Surelite I) with 5 ns pulse duration and 10 Hz repetition rate. The pump beam had a diameter

Fig. 8. Schematic diagram of the intracavity KTA OPO. Mirrors M1 and M2 form the Nd:YALO laser cavity, while M1 and M3 form the KTA OPO cavity [17]


153

of 4 mm and was double-passed through the 15 mm KTA crystal, for higher efficiency. The OPO was singly resonant for the idler wavelength and had an unstable resonator cavity to improve the spatial beam distribution. The idler beam tunability over the 2.6–3.2 µm range has been achieved with the pulse energy as high as 14–17 mJ at 2.9 µm for a pump pulse energy of 145 mJ. Even higher mid-IR OPO output energies were reported by Mennerat and Kupecek [20]. An OPO was based on a 50 mm long, 15 mm aperture 47◦ -cut CPM LiNbO3 crystal and was pumped at 1.064 µm (15 ns pulse duration). Maximum output energy of 300 mJ (signal + idler) was achieved near degeneracy (λ ≈ 2 µm) for a 600 mJ pump.

4

OPOs in the 4–20 µm Range

4.1

Comparison of NLO Materials Suitable for λ > 5 µm

Table 2 compares linear and nonlinear optical characteristics of the five most promising nonlinear optical crystals, suitable for NLO applications in the λ > 5 µm region. It contains four chalcopyrites: silver gallium sulfide AgGaS2 (AGS), silver gallium selenide AgGaSe2 (AGSe), zinc germanium phosphide ZnGeP2 (ZGP), and cadmium germanium arsenide CdGeAs2 (CGA); and layered chalcogenide gallium selenide GaSe. All five crystals utilize birefringent phase matching. Using relative NLO figures of merit (with respect to PPLN) from this table, and the numerical example of Sect. 2, one can estimate the parametric gain in these crystals at a given pump intensity. The last three crystals in Table 2 – namely ZGP, CGA, and GaSe – exhibit larger NLO figures of merit, as compared to AGS and AGSe. In addition, the former have much higher thermal conductivity (exceeding that of AGS and AGSe by more than an order of magnitude), and higher laser damage thresholds (> 1 J/cm2 ). Table 2. Comparison of linear and nonlinear optical properties of birefringent crystals suitable for the λ > 5 µm region (from [8]) Crystal Transparency deff range ( µm) (pm/V) AGS AGSe ZGP GaSe CGA

0.47–13.0 0.71–19.0 0.74–12.4 0.62–20.0 2.40–18.0

12 33 75 54 236

Average refractive index

NLO FOM d2eff /n3 with respect to PPLN

2.40 2.65 3.13 2.73 3.60

0.5 2.8 8.8 6.8 57.0

154


4.2

OPOs Based on AGS and AGSe

The first OPO based on AGS was reported in 1984 by Fan et al. [21]. A type-I OPO was pumped by a Q-switched Nd:YAG laser. (τ = 18 ns) and was tunable from 1.4 to 4 µm. The photon conversion efficiency was 16 %. Lately, Vodopyanov et al. demonstrated an AGS optical parametric oscillator with a much larger tuning range of 3.9–11.3 µm [22]. The singly resonant angle-tuned OPO was formed by two flat mirrors and used a 20 mm AGS crystal cut at θ = 45.1◦ to the optical z-axis for type-II phase matching. It was pumped by 1.06 µm pulses from a Nd:YAG laser with 12–100 ns pulse duration and yielded up to 0.37 mJ idler wave energy at 15 mJ pump. In the configuration with the recycling of the pump and idler beams, the OPO pump threshold was as low as 85 µJ. The OPO linewidth was ≈ 1 cm−1 and the quantum conversion efficiency reached 22 %. This AGS OPO with its tuning range of 3.9–11.3 µm can be considered to be the longest wavelength OPO pumped at 1 µm. Figure 9 shows the CO absorption spectrum taken with the OPO, which is compared with the FTIR spectrum with a 2 cm−1 resolution. Silver gallium selenide, AgGaSe2 (AGSe), provides broader IR tunability than AGS, but it cannot be pumped at 1 µm, because of phase matching limitations; however it phase-matches with a variety of pump sources in the range 1.3–3 µm. Eckardt et al. have reported the first AGSe OPO [23] which was pumped by a Q-switched Nd:YAG laser at 1.34 µm or a Ho:YLF laser at 2.05 µm. In the latter case, the tunability of 2.65–9.05 µm was achieved with a photon conversion efficiency of 18 %.

wavelength (mm) 4.8

4.7

optical density = -ln(T)

3 2

4.6

4.5

CO absorption L=10cm p=0.83atm FTIRspectrometer res. 2 cm-1

1 0 1cm-1

6

AgGaS2 OPO

5

2164 2166 2168

4 3 2 1 0 2050

2100

2150

2200

wavenumber (cm-1)

Fig. 9. CO absorption spectrum taken with an FTIR spectrometer (2 cm−1 resolution, upper curve) and with the AGS OPO system (lower curve) [22]


155

Raffy et al. [24] have used a 1.064 µm Q-switched Nd:YAG laser to pump two OPOs in series. The first one used a LiNbO3 crystal and was tuned to λ = 1.82 µm. It pumped the second OPO which used an AgGaSe2 crystal and was tunable between 2.15 and 12 µm. The total conversion from the 1.064 µm laser to the mid-IR output was 1.75 %. Chandra et al. [25] reported an AGSe OPO, which provided continuously tunable output in the 6–14 µm range when pumped by the 1.57 µm signal output from a Nd:YAG pumped KTP OPO. The latter used a 20 mm long noncritically phase-matched KTP crystal and had a 4 cm long flat–flat cavity with a gradient reflectivity mirror outcoupler. When pumped with 80 mJ, 10 ns pulses at 1.06 µm (repetition rate of 5 Hz, beam diameter 3 mm), up to 35 mJ of the 1.57 µm output was obtained, with a pulse width of 6 ns and spectral bandwidth of 5.7 cm−1 . The second stage OPO used a 6.5 × 6.5 × 35.3 mm3 type-I AgGaSe2 crystal cut at θ = 54◦ . The resonator was designed to resonate the signal and double pass the 1.57 µm pump. At the pump beam diameter of 6 × 5.4 mm2 , the OPO threshold was 0.02 J/cm2 , or 3 MW/cm2 in power density. The output idler energy reached 1 mJ at 9 µm, which corresponds to 20 % quantum efficiency, with respect to 1.57 µm pump, and 1.25 % overall efficiency of converting 1 µm into 9 µm radiation. Continuous tuning was obtained from 6.1 to 14.1 µm by a 40◦ rotation of the AgGaSe2 crystal. The idler bandwidth was measured to be 5 cm−1 near λ = 9 µm. It should be noted that low thermal conductivity and susceptibility to surface degradation resulting from a low laser damage threshold (0.1– 0.25 J/ cm2 ) severely limit the practical application of AGS and AGSe. 4.3

OPOs Based on ZGP

The very high nonlinear optical coefficient of ZGP – its NLO FOM is almost 9 times that of PPLN (see Table 2) – combined with good optical, mechanical, and thermal properties, favors a variety of nonlinear optical applications, including efficient high-average-power, as well as broadly tunable OPOs. Because of some residual absorption in the near-IR region, ZGP pump wavelength should be chosen at 2 µm or above. Thus, 2-µm holmium or 3-µm erbium lasers are good candidates for this purpose. Using ZGP crystals, mid-IR OPOs with the highest reported average output power at λ > 3 µm have been built. For example, Budni et al. [26] reported a 10 W coherent 3–5 µm light source for IR countermeasure applications based on ZGP. The type-I OPO using a 14 mm ZGP crystal was pumped at 10 kHz by 11 ns, 2.05 µm pulses from a diode-pumped Q-switched Ho,Tm:YLF laser operating at T = 77 K. At the maximum pump drive level of 20.1 W (two diode-pumped amplifier stages were used) incident onto the OPO crystal, the output average power (signal at 3.67 µm plus idler at 4.67 µm) reached 10.1 W, corresponding to a conversion efficiency of 50.2 %.

156


It is important that the operation of this OPO at maximum power levels was well below the crystal’s damage threshold. In the work by Cheung et al. [18] even higher powers in the 3–5 µm range were obtained from a single ZGP OPO. The type-I ZGP OPO was pumped by 25 W of average power at 2.13 µm from a Nd:YAG-laser-pumped KTP OPO (20 kHz, 40 ns), and its signal + idler output, tunable between 3.7 and 4.8 µm, reached an average power of 14 W. Also, more than 20 W of average power was extracted from two independent ZGP OPO stages pumped by two orthogonal polarizations of the KTP OPO. An extremely efficient ZGP OPO emitting simultaneously at 3.8 and 4.65 µm was reported by Budni et al. [27]. The OPO pump source was a 2.09 µm Ho:YAG laser operating at 10 kHz with ≈ 30 ns pulse duration and 9.5 W average power. It was pumped in turn by a continuous-wave Tm:YLF laser at 1.9 µm, end-pumped by 60 W of 793 nm diode laser power. More than 4.2 W of output was achieved from the ZGP OPO working in the 3.8 and 4.65 µm spectral region. Thus, the optical-to-optical conversion efficiency (793 nm laser diode to mid-IR) achieved in this work was > 7 %. Also, as compared to [26], the whole system worked at room temperature. Wu et al. [28] demonstrated a coupled tandem ZGP OPO approach, where 2 µm output of a KTP OPO pumps a ZGP OPO. Not only is the ZGP OPO placed within the oscillator of the first KTP OPO, but also the KTP OPO is placed within the oscillator of a diode-pumped Nd:YALO laser (Fig. 10). The Nd:YALO laser is able to generate ≈ 58 W of linear polarized, 1 µm output power at the repetition rate of 5 kHz and is formed by a high reflector M1 and an output coupler M2. The KTP OPO utilizes four Type-II, 51◦ -cut, 5 mm × 5 mm × 8 mm, diffusion bonded walkoff compensated KTP crystals, placed between mirrors M3 and M1. A ZGP OPO is placed in the arm of s-polarization of 2 µm KTP OPO output and is formed by a type-I, 53◦ cut, 5 mm × 5 mm × 10 mm ZGP, and mirrors M4 and M5. The output from the ZGP OPO was tunable over the range 3–6 µm. The maximum output power ~ 43 cm DBWOC KTP OPO Laser cavity ZGP OPO 1-mm

A-O Q-switch

DBWOC KTP

2-mm

Mid-IR

Nd:YALO

P M2 R=85%@1-mm

Nd:YALO Pump Module

M3 AR@1-mm HR@2-mm

M1 R=20%@2-mm HR@1-mm

M4 AR@2-mm HR@3~5-mm

M5 R=60%@3~5-mm HR@2-mm

Fig. 10. Schematic diagram of the coupled tandem KTP OPO–ZGP OPO. DBWOC KTP stands for Diffusion Bonded Walkoff Compensated KTP [28]


157

was achieved at 4.3 µm and was 2.53 W at approximately 580 W of diode pump power. The M 2 beam quality factor at 4.3 µm was MH2 ≈ 3.0 and MV2 ≈ 5.5, respectively. Allik et al. [29] reported a high-energy-per-pulse ZGP optical parametric oscillator pumped by a 2.8 µm Er,Cr:YSGG laser with a 10 Hz repetition rate, 25 mJ energy and 50 ns duration. A 25 mm long ZGP crystal was cut at θ = 65◦ for type-II phase matching. The OPO yielded idler output in the ‘forward’ direction of 0.7–2.4 mJ per pulse, in the wavelength range 6.9–9.9 µm. The quantum conversion efficiency reached 29 % and the OPO linewidth was typically 4 cm−1 . A ZGP OPO with a wide (3.8–12.4 µm) mid-IR tunability and more than 1 mJ output idler energy was demonstrated by Vodopyanov et al. [30,31]. The pump source was a Q-switched flash-lamp-pumped erbium laser with λ = 2.8 µm (Er,Cr:YSGG) or 2.93 µm (Er,Cr,Tm:YAG), with ≈ 100 ns pulse duration, 10 Hz repetition rate, and 10 mJ energy. The AR-coated ZGP crystal was 20 mm long and was cut for type-I (θ0 = 49.5◦) or for type-II (θ0 = 70◦ ) phase matching. The lowest OPO threshold was obtained in the flat–flat cavity configuration shown in Fig. 11. The front mirror, M1 , was transmissive for the pump and the idler and highly reflective (98 %) for the signal. The gold rear mirror, M2 , was highly reflective (R > 98 %) for the pump, signal, and idler. Thus, the signal wave resonated, while the pump and the idler were double-passed. A dichroic beam splitter (BS) separated the incoming pump beam from the outcoming idler. The pump beamsize was 0.82 mm. The OPO output was continuously tunable, via crystal angle tuning, from 3.8 to 12.4 µm (type-I phase matching) and from 4 to 10 µm (type-II phase matching) with a linewidth of 2–3 cm−1 . Figure 12 shows the OPO temporal pulse shape (at λ ≈ 8 µm) together with the incoming (undepleted) pump pulse. The plot of the OPO idler energy (at λ = 6.6 µm) versus pump pulse energy is shown in Fig. 13. The inset to Fig. 13 shows the far zone beam intensity distribution, measured using a 2D infrared beam profiler. The shape is close to the Gaussian and corresponds to the beam quality factor M 2 ≈ 1.5. Figure 14 represents the OPO output energy as a function of the idler wavelength, at two different pump energies: 5 and 10 mJ. Also, the dashed curve corresponds to the OPO pump threshold dependence. Remarkably, the OPO pump threshold was less than 1 mJ in the whole 4–12 µm range and the quantum conversion efficiency reached ≈ 50 % at 6–8 µm output. Pelouch et al. [32] took advantage of the reduced OPO threshold in a noncritical phase matching scheme. The ZGP crystal was 15 mm long and cut at 90◦ . The OPO pump source was a tunable (2.2–2.7 µm) Cr2+ :ZnSe laser (0.35 mJ, 1000 Hz) pumped by a Q-switched Tm:YALO laser operating at 1.94 µm. With the 2.55 µm pump wavelength, the ZGP OPO signal and idler wavelengths were 4.7 and 5.6 µm correspondingly. The pump threshold was 50 µJ and at the 200 µJ pump energy, 80 µJ of the output (signal + idler) was

158

Konstantin Vodopyanov ZGP

M2

M1 BS

out 3.8-12.4mm

pump

Er:Cr:YSGG-laser l=2.8mm

Fig. 11. Schematic and general view of an erbium-laser-pumped ZGP OPO. M1 and M2 are OPO mirrors, BS is a beamsplitter which separates the pump from the OPO output. A continuous tuning range of 3.8–12.4 µm has been demonstrated with the photon conversion efficiency reaching ≈ 50% at 6–8 µm incoming pump OPO

-100

0

100 ns

Fig. 12. ZGP OPO (λ ≈ 8 µm) temporal pulse shape shown together with the incoming (undepleted) pump laser pulse at 2.8 µm


159

1.4 idler energy (mJ)

1.2 1.0

l=6.6mm

0.8 0.6 5 mrad

0.4 0.2 0.0 0

2 4 6 8 10 input laser energy (mJ)

Fig. 13. ZGP OPO energy curve at the output wavelength λ = 6.6 µm. Solid line: trace to guide the eye. Inset: The far zone beam intensity distribution at 6.6 µm [31]

Fig. 14. OPO output energy as a function of the idler wavelength for two pump energies: 5 and 10 mJ. Dashed curve, the OPO pump threshold dependence. Inset, transmission spectrum of the antireflection-coated L = 2 cm ZGP crystal [30]

achieved. In this system, one can achieve the ZPO OPO tuning by tuning the pump; however the dependence of the idler wavelength on the pump wavelength near 2.2–2.7 µm is very weak. A widely tunable low-threshold tandem OPO based on PPLN and NCPM ZGP crystals has been demonstrated recently by Vodopyanov and Schunemann [33]. A noncritically phase-matched ZGP OPO was pumped by the idler wave output of a PPLN OPO with the focused beam size ≈ 200 µm. The latter was pumped by a diode-pumped Nd:YAG laser (1.6 mJ, 20 ns, 1 kHz). The singly resonant OPO (Fig. 15) contained a 24 mm θ = 90◦ ZGP crystal and was formed by a concave output coupler mirror M1 (highly reflective at the signal and transmissive at the idler and the pump) and a gold reflector mirror M2 which was deposited directly into the polished flat surface of the ZGP crystal (the front surface of the ZGP was AR-coated). Tuning the PPLN

160


ZGP OPO wavelength, mm

Fig. 15. Schematic of the widely tunable NCPM ZGP OPO pumped by the output of a PPLN OPO. M1 and M2 are OPO mirrors, BS is a dichroic beamsplitter which reflects the pump and transmits the OPO output 10

ZGP OPO

9 8 7 6 5 4 3 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 PPLN pump wavelength, mm

Fig. 16. Noncritically phase-matched ZGP OPO tuning curve as a function of the pump wavelength. Solid lines: theoretical curves

OPO wavelength in the range 2.3–3.7 µm resulted in tuning the ZGP OPO output from 3.7 to 10.2 µm. OPO tuning curves versus pump wavelength are shown in Fig. 16; the dotted lines on this figure correspond to theoretical prediction based on known dispersion data [34]. The ZGP OPO idler output (λ ≈ 7 µm) amounted to 25 µJ at 1 kHz corresponding to a photon conversion efficiency of 30 % and total conversion from the 1.064 µm laser to the 7 µm output of 1.5 %. At a smaller pump beam size of 125 µm (at λ ≈ 3.1 µm) which is close to the confocal focusing condition, the OPO pump threshold was remarkably low, ≈ 2 µJ; this is the lowest threshold reported so far for singly resonant OPOs. 4.4

OPOs Based on Other Crystals

A tandem OPO based on a CdSe crystal (deff = 18 pm/V) was demonstrated by Isyanova et al. [35]. The first OPO, using KTA, was pumped by a Q-switched Nd:YLF laser (200 mJ, 30 ns). The second OPO, based on NCPM CdSe, 35 mm long, was pumped by the idler of the first OPO. By


161

tuning of the pump wavelength between 2.7 and 3.5 µm the authors achieved tuning of the CdSe OPO in the 4.5–5.2 µm range (signal) and 8–10.6 µm range (idler). The total conversion efficiency of 1 µm radiation to 10 µm was ≈ 1 %. The GaSe crystal has an exceptionally wide transparency range and high second-order nonlinearity (Table 2); most important, it can be pumped by a 1.06 µm laser. Nevertheless, a GaSe OPO has not been demonstrated yet, partly because of the problems associated with the softness of the crystal, the infeasibility of cutting the crystal at the desired angle, and difficulties with AR coating. However, GaSe became the base for the most widely tunable mid-IR OPG (3.3–19 µm), which we will discuss in Sect. 5.3. CGA, the ‘Holy Grail’ of infrared NLO materials because of its enormous second-order nonlinearity (236pm/V), has proved effective for numerous NLO mid-IR applications, including second harmonic and difference frequency generation [8,36]. Yet, CGA OPO has been demonstrated only quite recently by Zakel et al. [37], and this was a result of improvement in the crystal’s optical properties and the availability of a pump source matching the crystal’s transparency. The pump was the second harmonic (λ = 4.8 µm) of a TEA CO2 laser with 2 Hz repetition rate, pulse duration 85 ns and 5.5 mJ energy. The 6 mm × 6 mm × 13 mm CGA crystal, cut for type-I phase matching (θ = 33◦ ), had absorption coefficients of 0.2 and 0.44 cm−1 for the pump and OPO wavelengths respectively. At the focused pump beam size of 256 µm, the measured OPO threshold was 0.95 mJ (5 MW/cm2 ). The OPO output was tunable between 8.5 and 11 µm (that is within the important 8–12 µm atmospheric transmission window); the total conversion efficiency was 3.5 % and the authors expect higher conversion efficiencies with better quality crystals.

5 Traveling-Wave Optical Parametric Generators (OPGs) As mentioned in Sect. 2, the principle of operation of a travelling-wave (superfluorescent) OPG is based on a single-pass high-gain (> 1010 ) amplification of quantum noise in a nonlinear crystal pumped by intense short laser pulses. The main advantages of the travelling-wave OPG scheme are: • Simplicity of the optical design (no cavity mirrors). • Ability to produce very high peak power outputs (> 1 MW) in the form of a single short pulse. • Broad tunability, restricted only by phase matching conditions and the crystal’s optical transparency. • No ‘build-up’ time – hence the possibility of generating two or more synchronized, independently tunable pulses from different OPGs pumped by the same laser, which is attractive for time-resolved laser spectroscopy. However, higher pump intensities (108 –1010 W/cm2 ) must be applied in this case – for example by using short (few ns to sub-ps) pump pulses. Also,

162


the bandwidths of OPG output can be broad – from a few cm−1 to a few hundreds of cm−1 . 5.1

OPGs Based on PP LN

An elegant PPLN-based OPG pumped by a microchip laser was demonstrated by Zayhowski [38]. The Nd:YAG microchip laser, passively Q-switched with Cr4+ :YAG, was longitudinally pumped with a fiber-coupled 12 W diode-laser array. At pulse repetition rates ≈ 1 kHz the microchip laser produces 157 µJ of energy in a single-frequency TEM00 pulse. The duration of the laser pulse was 0.9 ns and the size of the unfocused laser beam was 94 µm corresponding to the confocal parameter of 5.2 cm. A 13 mm long PPLN was placed next to the laser, without any intervening optics (Fig. 17). The PPLN QPM period was designed to convert 1.064 µm radiation to 1.59 µm signal and 3.22 µm idler radiation. Since the crystal had uncoated parallel faces, it was slightly tilted from the normal, to eliminate the possibility that the PPLN would oscillate, forcing it to act as a single-pass OPG. With an incident pump intensity of 1 GW/cm2 , the PPLN OPG was 1.5 times above threshold and converted 25 % of the incident pump radiation into the signal and idler wavelengths, which corresponds to 26 and 12 µJ output pulses at 1.59 and 3.22 µm, respectively. The estimated bandwidth of the output pulses was ≈ 250 cm−1 . S¨ udmeyer et al. reported recently a very high repetition rate single-pass OPG based on periodically poled lithium niobate. An OPG was directly pumped by the output of a passively mode-locked 1.03 µm Yb:YAG laser with pulse duration of 0.6 ps, and repetition rate of 35 MHz [39]. A 7 mm long, 0.5 mm thick PPLN crystal contained ten QPM gratings with poling periods between 27.1 and 30.4 µm. The laser beam was focused to a 40 µm waist in the middle of the PPLN crystal. By switching between the QPM gratings, the OPG signal output was tunable in the spectral region of 1.38–1.56 µm corresponding to the idler tunability of 3.03–4.06 µm. At 5 W of average pump

Fig. 17. Experimental setup for PPLN-based OPG pumped by a Nd:YAG microchip laser [38]


163

power, the signal wave power was measured to be between 0.25 and 0.5 W and the output idler power was estimated to be up to 0.25 W. 5.2

OPGs Based on ZGP

A schematic diagram of the long-wavelength ZGP OPG, tunable between 3.9 and 10 µm, reported by Vodopyanov et al. [40], is presented in Fig. 18. Single 100 ps pulses from an actively mode-locked, Q-switched and cavitydumped 2.8 µm Er,Cr:YSGG laser [41] were amplified to get an energy of 3 mJ. A double-pass OPG setup used a dichroic beam splitter (HR at the laser wavelength, and HT at the OPG wavelengths) and a back reflecting gold mirror. Both type-I (L = 11 mm) and type-II (L = 40 mm) ZGP crystals were utilized in this experiment. A lens L was used to slightly focus the laser beam into the crystal in such a manner that the laser intensity at the second pass through the crystal was 3–4 times higher than at the first pass, and amounted to a few GW/cm2 . Thus, the first pass served as a superluminescent seed, with the second as a parametric amplifier. The focusing was made to ensure that the parametric gain reaches saturation at the second pass, thus giving maximum conversion efficiency. Compared to a single-pass geometry, the double-pass OPG scheme produces much smaller output beam divergence, corresponding to ≈ 2× diffraction limit. The continuous tuning range achieved was 3.9–10 µm with quantum conversion efficiency 10 %. The OPG linewidth was typically 100 cm−1 for the type-I PM and ≈ 10 cm−1 for the type-II. The single-pass OPG threshold (L = 40 mm crystal) was 90 MW/cm2 , which was 300 times less than the ZGP surface damage threshold. In the work by Petrov et al. [42], a type-II ZGP traveling wave OPG was pumped by the output of a seeded optical parametric amplifier (Ti:sapphirelaser-pumped MgO:LiNbO3 OPA) with λ ≈ 3 µm and 2.7 ps pulse duration. With a two-pass arrangement, 5–11 µm tunability was achieved with µJ output energies, nearly bandwidth-limited pulses and 20 % internal quantum Beam splitter out NLO crystal

L pump 2.8mm 2-3mJ

4- pass Er,Cr:YSGG AMPLIFIER

Er,Cr:YSGG laser l=2.8 mm 100ps 0.7mJ

Fig. 18. Schematic of the double-pass travelling-wave OPG pumped by an erbium laser

164


efficiency. Femtosecond optical parametric generation in a very short (2 mm) ZGP was also reported [43]. In this case, the output of a Ti:sapphire-laserpumped BBO OPG, with λ ≈ 2 µm and 140 fs pulse duration, was used to pump a single-pass type-I ZGP, cut at θ = 52.5◦ . The output wavelengths were around 3.1 µm (signal) and 6.1 µm (idler), the pump threshold was found to be 25 GW/cm2 , and conversion efficiency 2.5 %. 5.3

OPGs Based on GaSe

A GaSe OPG with the record continuous tunability of 3.3–19 µm has been reported by Vodopyanov et al. [40,44]. The OPG was pumped, in a double-pass geometry (Fig. 18), by 100 ps , 3 mJ , λ = 2.8 µm pulses from an Er,Cr:YSGG laser. Despite the fact that GaSe can be cleaved only along the (001) plane (z-cut, θ = 0◦ ), its extremely large birefringence (∆ n ≈ 0.35) allows almost any conceivable three-wave interaction in its transparency range to be phasematched. The GaSe crystal length was 14 mm, and elliptical focusing with an aspect ratio 1:20 was used for the 2.8 µm pump beam, in order to keep the beam size sufficiently large in the walk-off plane (the walk-off amounted to 0.8 mm for the internal angle θ = 12◦ ). The OPG threshold intensity was 1.1 GW/cm2 . Figure 19 shows GaSe tuning curves (as a function of the tilt of the crystal) obtained with the 2.8 µm pump for type-I and type-II phase matching. The output linewidth was typically 10–30 cm−1 . The tuning range of 3.3 to 19 µm is the largest tunability range ever obtained with parametric generators/oscillators. It is also worth noticing that the whole tuning range can be covered with a single z-cut crystal. Figure 20 shows the OPG quantum efficiency as a function of the wavelength. The decline in efficiency at longer wavelengths is due to the increase of linear absorption and decreasing of parametric gain. The quantum conversion efficiency reached 5 % at Ipump ≈ 5 GW/cm2 . 20 18

type I

Wavelength, mm

16

GaSe type II

14 12 10 8 6 4 2

30

40

50

External angle qo

60

70

Fig. 19. OPG tuning curves obtained with the z-cut GaSe and 2.8 µm 100 ps pump [40]

Quantum conversion efficiency (%)


165

10 1 0.1 0.01

GaSe OPG

1E-3 1E-4

5.4

4

6

8 10 12 14 16 18 20 Output wavelength (mm)

Fig. 20. GaSe OPG quantum conversion efficiency as a function of the output wavelength [40]

OPGs Based on CGA

The first results on OPG based on CGA crystal has been reported recently by Vodopyanov et al. [45]. The 7 mm long CGA sample was cut at θ = 33◦ for type-I phase matching. The crystal faces were polished but not antirefection coated; the absorption coefficient was less than 0.1 cm−1 in the whole range of 5–12 µm. As a pump source for the OPG, the authors have used λ = 5 µm pulses from the Free-Electron Laser FELIX in Nieuwegein, Netherlands. The pump wavelength was chosen in such a range so as to minimize the linear and two-photon absorption in CGA. The laser beam consisted of a train of ≈ 100 short pulses (600 fs, 10 µJ), which were separated by intervals of 40 ns. The laser spectral width was 24 cm−1 and the overall repetition rate 10 Hz. The pump laser beam was focused into the CGA crystal with a beam size of 450 µm. An OPG threshold was observed at the incoming peak laser intensity of 3 GW/cm2 . The OPG output was angular-tuned within the range 7–18 µm and the spectral width of the OPG output was typically 250 cm−1 away from degeneracy, and 600 cm−1 near degeneracy (λ ≈ 10 µm); the last two values are close to the acceptance bandwidth for the 7 mm CGA crystal. At the peak pump intensity of 6 GW/cm2 , typical OPG internal photon conversion efficiency was 3 %.

6

Narrow-Linewidth OPOs

In many cases, spectroscopic applications of OPOs require narrow linewidths. For example, gases and vapors at atmospheric pressure have spectral features which are typically a few 0.1 cm−1 wide. Alternatively, at low pressures where Doppler broadening dominates, these features can be even sharper: < 10−2 cm−1 . In the latter case, a single-longitudinal-mode OPO is desirable. The spectral width of a pulsed Fabry–Pérot-type OPO without any intracavity spectral-narrowing elements normally ranges between 1 cm−1 and 100 cm−1 , depending on the phase matching type, output wavelength, proximity to the degeneracy point, crystal length, etc. Primarily, it is determined by the OPO phase matching acceptance bandwidth and decreases as the square root of the number of round-trip passes in an OPO.

166

6.1


Using Intracavity Spectral-Narrowing Elements

In the early work of Brosnan and Byer [46], narrow linewidths (0.08 cm−1 ) have been achieved in the Nd:YAG-laser-pumped mid-IR singly-resonant LiNbO3 OPO with the use of 600 lines per mm diffraction grating and a 2 mm Fabry–Pérot etalon with the finesse of 7 inside the cavity. With addition of the second (reflective) etalon, the OPO output collapsed to a single (< 0.02 cm−1 ) longitudinal mode which was remarkably stable in frequency. Raffy et al. [47] have shown that a LiNbO3 OPO (pumped at 930 nm by a Ti:sapphire laser) can be both tuned and linewidth-narrowed with an electrically tunable Fabry–Pérot interferometer that consists of two mirrors separated by an air layer of a few micrometers. Recently, Schlup et al. [48] reported a nanosecond Nd:YAG-pumped optical parametric oscillator based on periodically poled lithium niobate (20 mm long, 0.5 mm thick) whose optical bandwidth has been narrowed to a single longitudinal mode with less than 250 MHz (≈ 0.01 cm−1 ) linewidth by the use of a grazing incidence (88.6◦ ), 900 lines per mm diffraction grating. The single-mode OPO output could be tuned over 1.48–1.80 µm (signal) and 2.61–3.82 µm (idler), with a threshold pump energy of 1.1 mJ and slope efficiencies of up to 36 % (signal) and 10 % (idler). When pumped with 1.6 mJ pulses, output pulse energies of 170 µJ at the signal and 52 µJ at the idler wavelengths were observed. Richman et al. have developed a PPLN-based pulsed mid-IR ring cavity OPO that resonates the signal wave and uses just one intracavity etalon to restrict lasing to a single longitudinal mode of the resonator cavity [49]. The OPO ring resonator (Fig. 21) consisted of three mirrors. The two dichroic mirrors on either side of the PPLN crystal (25 mm long, 0.5 mm high, and 19 mm wide) transmitted the pump and idler beams and reflected the signal beam. The third mirror outcoupled 10 % of the signal beam. The airspace etalon was made of one convex and one concave mirror, with a 95 % reflective coating for the signal wave and each with a curvature that matches the phase-front curvature of the ring cavity TEM00 mode. The etalon mirror spacing was 357 µm (free spectral range 14 cm−1 ), it had a finesse of 60 and an insertion loss of 30 %. The pump was a narrow-linewidth 1 kHz repetition rate Q-switched laser with injection seeding (Continuum Model HPO-1000). It was nearly diffraction limited in profile, and was focused with a 350 mm focal-length lens to a 240 µm spot size in the PPLN crystal. Both the etalon mirror spacing and the OPO resonator length were adjusted electronically by piezoelectric translators (PZTs) so that it was possible to tune the frequency continuously over 10 cm−1 without using motorized parts. Translation of a multigrating (or fan grating) PPLN wafer can allow access to, and local tuning of, any wavelength from 1.45 to 1.8 µm (signal) and from 2.6 to 4 µm (idler). Up to 18 µJ/pulse in the idler beam and up to 15 µJ/pulse in the signal beam were produced with only 200 µJ pump energy at 1.06 µm. The measured single-mode OPO linewidth was 0.005 cm−1 .


dichroic mirrors Nd:YAG pump laser

PPLN crystal

lens air-space etalon (PZT)

signal output coupler

167

idler output and residual pump

cavity length (PZT)

cavity length (PZT)

Fig. 21. Schematic of the narrow-band tunable OPO (from [28]). The ring resonator consists of three mirrors. The two dichroic mirrors on either side of the PPLN crystal transmit the pump and idler beams and reflect the signal beam. The third mirror outcouples 10 % of the signal beam for spectral characterization. The airspace etalon is made of two lenses, each with a 95 % reflective coating and each with a curvature that nearly matches the phase-front curvature of the ring cavity TEM00 mode. The right-hand dichroic mirror and the third mirror are on a single translation stage driven by a PZT to control the cavity length [49] (Courtesy of B. A. Richman)

At longer wavelengths, a narrow-linewidth (≈ 0.1 cm−1 ) ZGP OPO tunable from 3.7 to 8 µm was demonstrated by Ganikhanov et al. [50]. A type-II ZGP OPO (Fig. 22) was pumped by the output of a LiNbO3 OPO at 2.55 µm, pumped in turn by a Q-switched Nd:YAG laser with a 15 ns pulse duration and 10 Hz repetition rate. Despite the fact that the pump linewidth at 2.55 µm was fairly broad (≈ 15 cm−1 ), the authors demonstrated that one can achieve a narrow-linewidth output at the resonating wavelength in a singly resonant OPO, containing an intracavity diffraction grating, in the Littrow configuration and a Fabry–Pérot etalon. The grating had 240 lines/mm (blazed at 3.2 µm) in the case of the SRO cavity to oscillate on the signal wave, and 120 lines/mm (blazed at 6.4 µm) for oscillating on the idler wave. The etalon was simply a 1 mm thick uncoated plane-parallel silicon plate (finesse ≈ 2.6). The output energies were in the range of 10–200 µJ and the linewidth of 0.1 cm−1 corresponded to ≈ 3 axial OPO cavity modes. 6.2 Narrow-Linewidth Optical Parametric Generator–Optical Parametric Amplifier (OPG–OPA) Systems An interesting configuration for a narrow-bandwidth OPG–OPA laser source was reported by Aniolek et al. [51]. The output of the PPLN OPG was spectrally filtered and amplified in a PPLN optical parametric amplifier (OPA).

168

Konstantin Vodopyanov Broadband pump, 2.55 mm

DIFFRACTION GRATING

etalon (Si-plate)

ZGP CRYSTAL Gold Mirror

OUT

Fig. 22. Schematic of a narrow-linewidth 3.7–8 µm ZGP OPO pumped at 2.55 µm [50]

The pump laser was a passively Q-switched (with Cr4+ :YAG) Nd:YAG microlaser, end-pumped by a fiber-coupled diode laser. It delivered 2.6 ns pulses with up to 650 µJ energy in a single longitudinal and lateral mode, at a 120 Hz repetition rate. The two PPLN crystals were 50 mm long by 11 mm wide, with a thickness of either 0.5 mm (for the OPG) or 1.0 mm (for the OPA stage). Both crystals were poled in stepped fashion along the width dimension, with poling periods ranging from 28.5 to 29.9 µm, in 0.2 µm steps. The pump into the OPG was focused to a waist of 190 µm and the optical parametric generation threshold for the first stage was ≈ 70 µJ. When the first stage was pumped with 100 µJ energy, the signal pulse energy was 8 µJ. The spectrum of the OPG signal output was broad, ≈ 15 cm−1 . Filtering it with a highfinesse (450 maximum at 1.55 µm) tunable air-spaced Fabry–Pérot etalon formed a spectrally narrow seed for the OPA stage. The filtered signal output was combined with the pump beam having 250 µJ energy at 1.064 µm (both beams had a waist of 400–450 µm) in the OPA stage to get the output of the seeded OPA with idler pulse energies of 10 µJ and signal pulse energies of 30 µJ. The output bandwidth of such an OPG–OPA system was estimated to be 0.05 cm−1 by means of cavity ring-down spectroscopy measurements of the room temperature water-vapor line at 2893.8 cm−1 . 6.3

OPOs with Injection Seeding

So far we have regarded passive methods of spectral narrowing which use intra- or extra-cavity frequency selective elements to get a narrow linewidth. Injection-seeding OPOs with narrow-linewidth continuous-wave (CW) lasers form an alternative approach to achieve narrow-linewidth OPO. This technique, however, requires an additional single-longitudinal-mode seed laser source for each spectral region, which makes the whole setup more sophisticated.


169

Baxter et al. [52] reported a seeded narrowband PPLN OPO, pumped at 1.064 µm by a pulsed (≈ 10 ns pulse duration), single-mode Nd:YAG laser. The actively controlled PPLN OPO ring cavity was injection seeded at its resonant signal wavelength of 1.55 µm by a 5 mW CW single-mode externalcavity diode laser. The corresponding idler output was tunable near 3.4 µm and had a linewidth of 130 MHz (0.0043 cm−1 ), close to the limit imposed by the Fourier transform of the pulse duration. For spectroscopic applications, the OPO output was amplified in an OPA stage based on bulk LiNbO3 , so that the combined pulse energy of signal and idler output was typically 2 mJ. Injection-seeded nanosecond OPG–OPA was reported by Wu et al. [53]. For the reason that OPG–OPA has no cavity modes, it amplifies a whole continuum of frequencies inside its phase matching acceptance bandwidth, and there is no need of frequency mode-matching between the seed and OPG– OPA modes. Pump pulses in this work were provided by the third harmonic from a single longitudinal mode Nd:YAG laser which delivered pulses of 2.5 ns width at 355 nm, with 0–160 mJ of pulse energy at repetition rates up to 100 Hz, 5 mm beam diameter and a flat-top intensity profile. The OPG and the OPA stage each consisted of two walk-off-compensated beta-barium borate (BBO) crystals. The OPG–OPA setup was capable of tuning from 410 to 2400 nm without gaps. However, seeding was performed in a much narrower range, using a CW SLM laser with a few mW power tunable around 815 nm. The linewidth of such seeded ns OPG–OPA was measured to be 650 MHz (0.02 cm−1 ), close (within a factor of 4) to the transform-limited bandwidth of a 2.5 ns pump pulse. Since BBO material absorbs at wavelengths greater than 2 µm, a PPLN crystal can be used instead, in order to extend the IR tuning range. With a 1064 nm pump, similar OPG–OPA design and a SLM seed laser at 1.3–1.7 µm, one can achieve 4–5 µm tunability. 6.4

Using a Doubly Resonant Cavity

An attractive idea of how to achieve pulsed single longitudinal mode (SLM) operation with wide tunability in a doubly resonant optical parametric oscillator (DROPO) was exploited recently by Scherrer et al. [54,55]. DRPOs are resonant both at the signal and the idler wavelengths and have smaller pumping thresholds, as compared to singly resonant OPOs, but were traditionally considered to be impractical for spectroscopy because the double resonance condition is satisfied only at particular wavelengths, leading to cluster-hopping effects that preclude continuous frequency tuning. To overcome the problems of frequency clustering, a dual-cavity (or ‘entangled cavity’) DROPO was proposed. A schematic of the OPO is illustrated in Fig. 23. The signal and idler waves oscillate between the pairs of mirrors M1 –M3 and M2 –M4 , respectively. The inner mirrors (M2 , M3 ) are deposited onto the NLO crystal faces whereas the external mirrors (M1 , M4 ) are mounted on two PZT actuators for fine tuning of the lengths. A compact passively Q-switched Nd:YAG laser based on a non-planar ring oscillator delivers single-frequency

170

Konstantin Vodopyanov signal

M2 M1

idler

M3 M4

Fig. 23. Schematic top view of the dual-cavity (entangled cavity) DROPO for SLM operation. M1 – M3 form signal cavity, and M2 –M4 form idler cavity. M1 and M4 were moved with piezoelectric translators (Courtesy of E. Rosencher)

output pump pulses (1.064 µm, 20 µJ, 14 ns) at a 12 kHz repetition rate. The pump beam was focused to a 75 µm spot in the NLO crystal. The crystal was a 6 × 3 × 0.5 mm3 multiple-grating PPLN comprising three QPM grating sections with 27.8, 28.2 and 28.7 µm periods. The output coupling of the idler field is obtained through M4 which has a partial reflectivity (70 %) around 4 µm. The finesses of the signal and idler cavities were 900 and 13 respectively. The tunability of the OPO was investigated by changing the PZT voltages which control translating of mirrors M1 and M4 , whereas the crystal grating was fixed. Discrete tuning by means of mode hops of ≈ 0.33 cm−1 was achieved by changing only one cavity length. The OPO linewidth remained less than 80 MHz (0.0027 cm−1 ), demonstrating stable single-mode operation. Small (few µm) changes of the idler cavity length create a shift of the idler mode distribution with respect to the signal one. Therefore, as depicted in Fig. 24, doubly resonant operations will occur successively at the signal cavity resonances with a spectral separation given by the signal mode spacing (Vernier effect). The same is valid for the signal cavity PZT displacements. By selecting the PPLN grating section and adjusting the crystal temperature, SLM output was achieved in the 3.8–4.1 µm range. Continuous (hop-free) frequency tuning was obtained by changing the lengths of the signal and idler cavities in opposite directions. The OPO pump threshold was as small as 8 µJ (7.5 MW/cm2 ), and the pump–idler conversion efficiency reached 5.2 %. Fig. 24. Schematic diagram of discrete tuning with mode hops. Initially, the exact coincidence is obtained with the mode m0 , then the idler mode structure is shifted to the right by changing the idler PZT voltage, and consequently, the signal frequency jumps to the mode m0 − 1 [54]


171

Table 3. Some linear and nonlinear optical properties of zincblende semiconductors (from [56,57]) Crystal Transparency range (µm) GaAs GaP ZnSe ∗

0.90–17.0 0.51– 7.2 0.51–18.0

deff ∗ (pm/V) 2/π · 94.0 = 60.0 2/π · 37.0 = 23.6 2/π · 26.4 = 16.8

Average refractive index

NLO FOM d2eff /n3 with respect to PPLN

3.30 3.05 2.44

6.20 2.30 0.93

The 2/π factor represents the fact that the crystal can only be quasi-phase-matched

7 Emerging Nonlinear Optical Materials for Mid-IR Applications Zincblende semiconductors (GaAs, GaP, ZnSe and others) are extremely attractive for nonlinear optical frequency conversion, due to their good infrared transparency, high optical nonlinearity, superior thermal conductivity, and large surface damage threshold. Some of these properties are listed in Table 3. Most important is that the technology of growth of these crystals is in a very mature state. The only obstacle preventing them from serving as ideal mid-IR NLO materials is that their cubic crystal structure lacks birefringence. The phase matching problem arising from the optical isotropy of zincblende-type semiconductors can be solved by quasi-phase-matching that require spatial modulation of the optical nonlinearity. Unfortunately, zincblende-type semiconductors are not ferroelectric and no techniques analogous to electric field poling in LiNbO3 exist for inducing a quasi-phasematching domain grating in an already grown crystal – domain reversal should be incorporated into the crystal growth process. Angell et al. [58] succeeded in the growth of alternating 100/111oriented stripes of CdTe on a GaAs substrate using metalorganic chemical vapor deposition. The CdTe layer was shown to be a suitable template to pattern the orientation of subsequently grown wide-band-gap II–VI films of ZnSe or ZnTe which may serve as a waveguide or bulk media for nonlinear optical interaction. An ideal QPM could be accomplished in zincblende crystals by creating a√111/111 pattern, in which the nonlinear coefficient deff varies as ±2d14 / 3. The combination of 111/100 orientations, √ utilized in [58], results in an amplitude modulation of deff , between 2d14 / 3 and 0, which is one fourth as efficient as the ideal case. Recently, an all-epitaxial process for fabrication of orientation patterned (domain-inverted) GaAs structures has been developed [59,60,61], and has been applied for making structures for nonlinear optical applications. The orientation patterned GaAs (OP-GaAs) films were fabricated by a multistep process illustrated in Fig. 25. First, GaAs/Ge/GaAs heteroepitaxy is used to create a sublattice-reversed (SR) or antiphase GaAs layer on a GaAs sub-

172


Fig. 25. Fabrication process of periodically patterned (domain-inverted) GaAs. (a) First, GaAs/Ge/GaAs heteroepitaxy is used to create a sublattice-reversed (SR) GaAs layer on a GaAs substrate. (b) Then the wafer is patterned, using photolithography and wet chemical etching, to create an orientation template with the period determined by a specific QPM process. (c) This template then undergoes two epitaxial growth steps (molecular beam epitaxy and hydride vapor phase epitaxy) to produce thick OP-GaAs film [61]

strate (Fig. 25a), after which this wafer is patterned, using photolithography and wet chemical etching, to create an orientation template (Fig. 25b) with the period determined by a specific QPM NLO process. This template then undergoes two epitaxial growth steps (molecular beam epitaxy and hydride vapor phase epitaxy) to produce thick OP-GaAs film (Fig. 25c). In this fashion, OP-GaAs layers can be grown to a thickness of 0.5 mm and length of up to 2 cm with excellent optical quality. Figure 26 shows the polished cross-section of an OP-GaAs sample with the QPM period of 61 µm. Orientation-patterned GaAs samples with different domain periods ranging between 27 and 212 µm, were characterized by continuous-wave CO2 laser frequency-doubling experiments [60] and mid-IR difference frequency generation between 7.95 and 8.6 µm [62] which have shown conversion efficiencies which were close to theoretically predicted for a given crystal’s length. Lately, Skauli et al. [57] performed an absolute comparison between nonlinear figures of merit of OPGaAs and periodically poled LiNbO3 , using quasi-phase-matched second har-

Fig. 26. Cross-section of an OP-GaAs sample with the QPM period of 61 µm (Courtesy of L. A. Eyres)


173

monic generation using λ ≈ 4 µm radiation, and found the nonlinear optical FOM ratio to be (d2eff /n3 )GaAs /(d2eff /n3 )PPLN = 6.2 ± 0.6. Also the authors demonstrated an efficient 4 µm → 2 µm frequency doubling in OP-GaAs with > 30 % internal efficiency at the energy of pump pulses of only 50 µJ. Thus, all-epitaxially grown OP-GaAs proves to be an extremely efficient material for nonlinear optical applications including OPOs and it may serve as a long-wavelength (λ > 5 µm) substitute of PPLN, but with much superior properties – higher nonlinearity, higher thermal conductivity and higher laser damage threshold.

8

Summary and Concluding Remarks

The state-of-the-art of pulsed mid-IR OPOs can be summarized as follows: OPOs can produce an average power of more than 10 W and energies of more than 100 mJ per pulse; continuous OPO/OPG tunability can be as wide as 3–19 µm (almost three octaves) in a single arrangement without changing any optics or crystals; the OPO linewidth is typically a few cm−1 but it can be reduced to its Fourier-transform limit of 10−2 –10−3 cm−1 if the OPO is forced to work in a single longitudinal mode regime; quantum conversion efficiency from the pump laser to the OPO output can exceed 50 %, and pump thresholds can be as low as 2 µJ in a singly resonant OPO. Periodically poled LiNbO3 , PPLN, can serve as a work horse for the majority of spectroscopic applications in the wavelength range 1.5–4 µm, thanks to the crystal’s robustness, high nonlinearity and affordable cost. Yet, the maximum 1 mm thickness combined with the inherent low-damage threshold of PPLN limits the use of this material in high-energy applications. On the other hand, PP KTP, PP RTA and PP KTA have similar NLO characteristics but are better suited for higher energy-per-pulse applications, due to their larger thickness and larger surface damage threshold; also, RTA and KTA have a longer infrared cut-off (> 5 µm). Because of its robustness and huge nonlinearity, ZGP serves in many midIR high average power OPO applications in the 3–5 µm range; also, it has high potential for biomedical and spectroscopic applications in the 3–10 µm range. For example, the first spectroscopic detection of explosives (TNT, RDX and using ZGP OPO via absorption measurements of fundamental bands between 6 and 8 µm. [63]. CGA seems to be the crystal of choice for long-wave applications (especially in the 8–12 µm window where it has very low absorption) and is expected to have the smallest OPO pumping threshold, due to its highest NLO coefficient. However, CGA needs a compact all-solid-state pump laser source in the 3–5 µm range. In contrast, a GaSe OPO can be directly pumped by a 1.06 µm laser and deliver continuous tunability of 1–18 µm, provided that the AR coating and mechanical hardness problems are solved.

174


Even though it is still at a materials research level, orientation patterned GaAs has proven to be attractive in many senses for future applications: it can be pumped at 1 µm, has a broad transparency range (1–17 µm) and one can expect very low OPO pump thresholds of a few µJ because of the crystal’s high nonlinearity and non-critical phase matching which is a characteristic of all QPM crystals. Now OP-GaAs can be fabricated all-epitaxially with a 20 mm length and 0.5 mm thickness, sufficient for practical bulk nonlinear optical applications. Difference frequency generation of mid-infrared radiation has been demonstrated as well as highly efficient second harmonic generation. In summary, major progress in creating advanced NLO materials and pump lasers has enabled efficient, high power, and broadly tunable optical parametric devices for the mid-IR range of the spectrum to be built. It is important to notice that all these devices work at room temperature. New challenges include further development of artificial QPM materials (GaAs, AlGaAs, InP, GaP, ZnS, ZnSe), and progress towards compact all-solid-state mid-IR OPO devices based on diode-pumped laser technology.

References 1. J. A. Giordmaine, R. C. Miller: Tunable coherent parametric oscillation in LiNbO3 at optical frequencies, Phys. Rev. Lett. 14, 973–976 (1965) 141 2. A. Yariv: Quantum Electronics, 3rd ed. (Wiley, New York 1988) 142, 143, 145 3. M. M. Fejer, G. A. Magel, D. H. Jundt, R. L. Byer: Quasi-phase-matched 2nd harmonic-generation, Tuning and tolerances, IEEE J. Quantum Electron. 28, 2631–2654 (1992) 142, 145, 146 4. R. L. Byer, R. L. Herbst: Parametric Oscillation and Mixing, in Nonlinear Infrared Generation, Y.-R. Shen (Ed.), Topics Appl. Phys. 16 (Springer, Berlin 1977) 143 5. D. A. Roberts: Simplified characterization of uniaxial and biaxial nonlinear optical crystals, A plea for standardization of nomenclature and conventions, IEEE J. Quantum Electron. 28, 2057–2074 (1992) 145, 146 6. L. E. Myers, R. C. Eckardt, M. M. Fejer, R. L. Byer, W. R. Bosenberg, J. W. Pierce: Quasi-phase-matched optical parametric oscillators in bulk periodically poled LiNbO3 , J. Opt. Soc. Am. B 12, 2102–2116 (1995) 145, 146, 147 7. S. J. Brosnan, R. L. Byer: Optical parametric oscillator threshold and linewidth studies, IEEE J. Quantum Elec. 15, 415–431 (1979) 145 8. V. G. Dmitriev, G. G. Gurzadyan, D. N. Nikogosyan: Handbook of Nonlinear Optical Crystals (Springer, Berlin, Heidelberg 1997) 146, 153, 161 9. J. Hellstr¨ om, V. Pasiskevicius, F. Laurell, H. Karlsson: Efficient nanosecond optical parametric oscillators based on periodically poled KTP emitting in the 1.8–2.5-µm spectral region, Opt. Lett. 24, 1233–1235 (1999) 146, 149 10. L. E. Myers, R. C. Eckardt, M. M. Fejer, R. L. Byer, W. R. Bosenberg: Multigrating quasi-phase-matched optical parametric oscillators in periodically poled LiNbO3 , Opt. Lett. 21, 591–593 (1996) 147, 148 11. G. Hansson, D. D. Smith: Mid-infrared-wavelength generation in 2-µm pumped periodically poled lithium niobate, Appl. Opt. 37, 5743–5746 (1998) 147


175

12. P. E. Britton, D. Taverner, K. Puech, D. J. Richardson, P. G. R. Smith, G. W. Ross, D. C. Hanna: Optical parametric oscillation in periodically poled lithium niobate driven by a diode-pumped Q-switched erbium fiber laser, Opt. Lett. 23, 582–584 (1998) 147 13. G. Bellomonte, A. Grisard, L. Becouarn, E. Lallier: High efficiency mid-infrared optical parametric oscillation in thick periodically poled lithium niobate, Conference on Lasers and Electro-Optics Europe, Tech. Digest (Eur. Phys. Soc./ IEEE/ Lasers & Electro-Opt. Soc./ Opt. Soc. Am., 2000) p. 332 149 14. M. Peltz, U. B¨ ader, A. Borsutzky, R. Wallenstein, J. Hellstr¨ om, H. Karlsson, V. Pasiskevicius, F. Laurell: Optical parametric oscillators for high pulse energy and high average power operation based on large aperture periodically poled KTP and RTA, Appl. Phys. B 73, 663–670 (2001) 149, 150 15. G. Hansson, H. Karlsson, F. Laurell: Unstable resonator optical parametric oscillator based on quasi-phase-matched RbTiOAsO4 , Appl. Opt. 40, 5446–51 (2001) 150, 151 16. G. Rosenman, A. Skliar, Y. Findling, P. Urenski, A. Englander, P. A. Thomas, Z. W. Hu: Periodically poled KTiOAsO4 crystals for optical parametric oscillation, J. Phys. D: Appl. Phys. 32, L49–L52 (1999) 151 17. R. F. Wu, K. S. Lai, H. F. Wong, W. J. Xie, Y. L. Lim, E. Lau: Multiwatt mid-IR output from a Nd:YALO laser pumped intracavity KTA OPO, Opt. Express 8, 694–698 (2001) 152 18. E. Cheung, S. Palese, H. Injeyan, C. Hoefer, R. Hilyard, H. Komine, J. Berg, W. Bosenberg: High power conversion to mid-IR using KTP and ZGP OPO, in Advanced Solid-State Lasers, Tech. Digest (Opt. Soc. Am., Washington, DC 1999) p. 358–361 152, 156 19. G. Vysniauskas, D. Burns, E. Bente: Development of a nanosecond high energy KTA OPO system operating 2.9 µm, in Conference on Lasers and ElectroOptics (CLEO-2002), Tech. Digest (Opt. Soc. Am., Washington, DC 2002) p. 333. http://www.photonics.ac.uk/lib/ 152 20. G. Mennerat, P. Kupecek: High-energy narrow-linewidth tunable source in the mid infrared, in Advanced Solid-State Lasers, Tech. Digest (Opt. Soc. Am., Washington, DC 1998) p. 47–49 153 21. Y. X. Fan, R. L. Eckardt, R. L. Byer, R. K. Route, R. S. Feigelson: AgGaS2 infrared parametric oscillator, Appl. Phys. Lett. 45, 313–315 (1984) 154 22. K. L. Vodopyanov, J. P. Maffetone, I. Zwieback, W. Ruderman: AgGaS2 optical parametric oscillator continuously tunable from 3.9 to 11.3 µm, Appl. Phys. Lett. 75, 1204–1206 (1999) 154 23. R. C. Eckardt, Y. X. Fan, R. L. Byer, C. L. Marquardt, M. E. Storm, L. Esterowitz: Broadly tunable infrared parametric oscillator using AgGaSe2 , Appl. Phys. Lett. 49, 608–610 (1986) 154 24. J. Raffy, T. Debuisschert, J. P. Pocholle, M. Papuchon: An AgGaSe2 optical parametric oscillator pumped by a LiNbO3 OPO, Ann. Phys. (Paris) 20, S75–76, (1995) 155 25. S. Chandra, T. H. Allik, G. Catella, R. Utano, J. A. Hutchinson: Continuously tunable, 6–14 µm silver-gallium selenide optical parametric oscillator pumped at 1.57 µm, Appl. Phys. Lett. 71, 584–586 (1997) 155 26. P. A. Budni, L. A. Pomeranz, M. L. Lemons, P. G. Schunemann, T. M. Pollak, E. P. Chicklis: 10 W mid-IR holmium pumped ZnGeP2 OPO, in Advanced Solid State Lasers, Tech. Digest (Opt. Soc. Am., Washington, DC 1998) p. 90–92 155, 156

176


27. P. A. Budni, L. A. Pometanz, M. L. Lemons, C. A. Miller, J. R. Mosto, E. P. Chicklis: Efficient mid-infrared laser using 1.9-µm-pumped Ho:YAG and ZnGeP2 optical parametric oscillators, J. Opt. Soc. Am. 17, 723–8 (2000) 156 28. R. Wu, K. S. Lai, W. E. Lau, H. F. Wong, Y. L. Lim, K. W. Lim, L. C. L. Chia: A novel laser integrated with a coupled tandem OPO configuration, in TOPS Vol. XX, Conference on Lasers and Electro-Optics (CLEO-2002), Tech. Digest (Opt. Soc. Am., Washington, DC 2002), p. 154 156, 167 29. T. H. Allik, S. Chandra, D. M. Rines, P. G. Schunemann, J. A. Hutchinson, R. Utano: Tunable 7–12-µm optical parametric oscillator using a Cr,Er:YSGG laser to pump CdSe and ZnGeP2 crystals, Opt. Lett. 22, 597–599 (1997) 157 30. K. L. Vodopyanov, F. Ganikhanov, J. P. Maffetone, I. Zwieback, W. Ruderman: ZnGeP2 optical parametric oscillator with 3.8–12.4-µm tunability, Opt. Lett. 25, 841–3 (2000) 157, 159 31. K. L. Vodopyanov: OPOs target the longwave infrared, Laser Focus World 37, 225–232 (2001) 157, 159 32. W. S. Pelouch, G. J. Wagner, T. J. Carrig, W. J. Scharpf: Mid-wave ZGP OPOs pumped by a Cr:ZnSe laser, in Advanced Solid State Lasers, Tech. Digest (Opt. Soc. Am., Washington, DC 2002) PD1-1–PD1-3 157 33. K. L. Vodopyanov, P. G. Schunemann: Broadly tunable non-critically phasematched ZGP optical parametric oscillator with a 2-microjoule pump threshold, Opt. Lett. 28, 441–443 (2003) 159 34. D. E. Zelmon, E. A. Hanning, P. G. Schunemann: Refractive-index measurements and Sellmeier coefficients for zinc germanium phosphide from 2 to 9 µm with implications for phase matching in optical frequency-conversion devices, J. Opt. Soc. Am. B 18, 1307–1310 (2001) 160 35. Y. Isyanova, A. Dergachev, D. Welford: Multi-wavelength, 1.5–10-µm tunable, tandem OPO, in Advanced Solid State Lasers, Tech. Digest (Opt. Soc. Am., Washington, DC 1999) p. 548–553 160 36. K. L. Vodopyanov, P. G. Schunemann: Efficient difference frequency generation of 7-20 µm radiation using CdGeAs2 , Opt. Lett. 23, 1096–1098 (1998) 161 37. A. Zakel, S. D. Setzler, P. G. Schunemann, T. M. Pollak, M. Burky, S. Guha: Optical parametric oscillator based on cadmium germanium arsenide, in Conference of Lasers and Electro-Optics (CLEO 2002), Tech. Digest (Opt. Soc. Am., Washington, DC 2002) p. 172 161 38. J. J. Zayhowski: Periodically poled lithium niobate optical parametric amplifiers pumped by high-power passively Q-switched microchip lasers, Opt. Lett. 22, 169–171 (1997) 162 39. T. S¨ udmeyer, J. A. D. Au, R. Paschotta, U. Keller, P. G. R. Smith, G. W. Ross, D. C. Hanna: Novel ultrafast parametric systems: high repetition rate singlepass OPG and fibre-feedback OPO, J. Phys. D: Appl. Phys. 34, 2433–2439 (2001) 162 40. K. L. Vodopyanov, V. Chazapis: Extra-wide tuning range optical parametric generator, Opt. Commun. 135, 98–102 (1997) 163, 164, 165 41. K. L. Vodopyanov: Parametric generation of tunable infrared radiation in ZnGeP2 and GaSe pumped at 3 µm, J. Opt. Soc. Am. B 10, 1723–9 (1993) 163 42. V. Petrov, Y. Tanaka, T. Suzuki: Parametric generation of 1 ps pulses between 5 and 11 µm with a ZnGeP2 crystal, IEEE J. Quantum Electronics 33, 1749–55 (1997) 163


177

43. V. Petrov, F. Rotermund, F. Noack, P. Schunemann: Femtosecond parametric generation in ZnGeP2 , Opt. Lett. 24, 414–416 (1999) 164 44. K. L. Vodopyanov: Mid-IR optical parametric generator with extrawide (3–19 µm) tunability, Applications for spectroscopy of two-dimensional electrons in quantum wells, J. Opt. Soc. Am. B 16, 1579–86 (1999) 164 45. K. L. Vodopyanov, G. M. H. Knippels, A. F. G. van der Meer, J. P. Mafetone, I. Zwieback: Optical parametric generation in CGA crystal, Opt. Commun. 202, 205–208 (2002) 165 46. S. J. Brosnan, R. L. Byer: Optical parametric oscillator threshold and linewidth studies, IEEE J. Quantum Electron. 15, 415–431 (1979) 166 47. J. Raffy, T. Debuisschert, J.-P. Pocholle: Widely tunable optical parametric oscillator with electrical wavelength control, Opt. Lett. 22, 1589–1591 (1997) 166 48. P. Schlup, G. W. Baxter, I. T. McKinnie: Single-mode near- and mid-infrared periodically poled lithium niobate optical parametric oscillator, Opt. Commun. 176, 267–271 (2000) 166 49. B. A. Richman, K. W. Aniolek, T. J. Kulp, S. E. Bisson: Continuously tunable, single-longitudinal-mode, pulsed mid-infrared optical parametric oscillator based on periodically poled lithium niobate, J. Opt. Soc. Am. B 17, 1233–1239 (2000) 166, 167 50. F. Ganikhanov, T. Caughey, K. L. Vodopyanov: Narrow-linewidth middleinfrared ZnGeP2 optical parametric oscillator, J. Opt. Soc. Am. B 18, 818–822 (2001) 167, 168 51. K. W. Aniolek, R. L. Schmitt, T. J. Kulp, B. A. Richman, S. E. Bisson, P. E. Powers: Microlaser-pumped periodically poled lithium niobate optical parametric generator–optical parametric amplifier, Opt. Lett. 25, 557–559 (2000) 167 52. G. W. Baxter, Y. He, B. J. Orr: A pulsed optical parametric oscillator, based on periodically poled lithium niobate (PPLN), for high-resolution spectroscopy, Appl. Phys. B 67, 753–756 (1998) 169 53. S. Wu, V. A. Kapinus, G. A. Blake: A nanosecond optical parametric generator/ amplifier seeded by an external cavity diode laser, Opt. Commun. 159, 74–79 (1999) 169 54. B. Scherrer, I. Ribet, A. Godard, E. Rosencher, M. Lefebvre: Dual-cavity doubly resonant optical parametric oscillators, Demonstration of pulsed single-mode operation, J. Opt. Soc. Am. B 17, 1716–29 (2000) 169, 170 55. C. Drag, A. Desormeaux, M. Lefebvre, E. Rosencher: Entangled cavity optical parametric oscillator for mid-infrared pulsed single-longitudinal-mode operation, Opt. Lett. 27, 1238–1240 (2002) 169 56. D. N. Nikogosyan: Properties of Optical and Laser-Related Materials (Wiley, Chichester, New York 1997) 171 57. T. Skauli, K. L. Vodopyanov, T. J. Pinguet, A. Schober, O. Levi, L. A. Eyres, M. M. Fejer, J. S. Harris, B. Gerard, L. Becouarn, E. Lallier: Measurement of nonlinear coefficient of orientation-patterned GaAs and demonstration of highly efficient second harmonic generation, Opt. Lett. 27, 628–630 (2002) 171, 172 58. M. J. Angell, R. M. Emerson, J. L. Hoyt, J. F. Gibbons, L. A. Eyres, M. L. Bortz, M. M. Fejer: Growth of alternating (100)/(111)-oriented II-VI regions for quasiphase-matched nonlinear optical devices on GaAs substrates, Appl. Phys. Lett. 64, 3107–9 (1994) 171

178


59. C. B. Ebert, L. A. Eyres, M. M. Fejer, J. S. Harris: MBE growth of antiphase GaAs films using GaAs/Ge/GaAs heteroepitaxy, J. Cryst. Growth 201, 187– 193 (1999) 171 60. L. A. Eyres, P. J. Tourreau, T. J. Pinguet, C. B. Ebert, J. S. Harris, M. M. Fejer, L. Becouarn, B. Gerard, E. Lallier: All-epitaxial fabrication of thick, orientation-patterned GaAs films for nonlinear optical frequency conversion, Appl. Phys. Lett. 79, 904–906 (2001) 171, 172 61. S. Koh, T. Kondo, Y. Shiraki, R. Ito: GaAs/Ge/GaAs sublattice reversal epitaxy and its application to nonlinear optical devices, J. Cryst. Growth 227, 183 (2001) 171, 172 62. O. Levi, T. Skauli, T. Pinguet, K. L. Vodopyanov, P. S. Kuo, X. Yu, L. A. Eyres, J. S. Harris, M. M. Fejer, S. E. Bisson, T. J. Kulp, B. Gerard, L. Becouarn, E. Lallier: Continuously tunable long-wave IR frequency generation in orientation patterned GaAs, in Conference on Lasers and Electro-Optics (CLEO-2002), Tech. Digest (Opt. Soc. Am., Washington, DC 2002) p. 416–417 172 63. M. W. Todd, R. A. Provencal, T. G. Owano, B. A. Paldus, A. Kachanov, K. L. Vodopyanov, M. Hunter, S. Coy, J. I. Steinfeld, J. T. Arnold: Application of mid-infrared cavity ringdown spectroscopy to trace explosives vapor detection using a broadly tunable (6–10 µm) optical parametric oscillator, Appl. Phys. B 75, 367–376 (2002) 173

Index

AgGaS2 (AGS), 153, 154 AgGaSe2 (AGSe), 153, 154

nonlinear optical figure of merit (NLO FOM), 143, 155, 173

CdGeAs2 (CGA), 153, 161 CdSe, 160 Cr2+:ZnSe, 157 critical phase matching (CPM), 152

OPG–OPA, 142, 167, 169 OPO cavity, 143–145, 148, 150 optical parametric generator (OPG), 143, 145, 161, 163–165 optical parametric oscillator (OPO), 141–170, 173 orientation patterned GaAs (OPGaAs), 171–174

doubly resonant optical parametric oscillator (DROPO), 169, 170 Er:Cr:YSGG laser, 157, 158, 163, 164 ferroelectric oxide, 145 GaAs, 171 GaP, 171 GaSe, 153, 160, 161, 164, 173 KTiOAsO4 (KTA), 145, 146, 149, 151–153, 160, 173 KTiOPO4 (KTP), 145, 146, 149–152, 155, 156 laser – Er:Cr:YSGG, 157, 158, 163, 164 – Nd:YAG, 147–156, 159, 162, 166–169 LiNbO3 , 141, 145, 146, 149, 153, 155, 163–167 – periodically poled (PPLN), 142, 145–149, 153, 155, 157–162, 166–173 MgO:LiNb3 , 163 Nd:YAG laser, 147–156, 159, 162, 166–169 non-critical phase-matching (NCPM), 146, 152, 157, 159, 160, 174

parametric frequency down-conversion, 142 parametric gain, 143, 145, 146, 153, 164 parametric noise, 143, 145 phase matching, 142 – critical, 152 – non-critical, 146, 152, 157, 159, 160, 174 potassium titanyl phosphate (KTiOPO4 , or KTP), 145, 146, 149–152, 155, 156 PP KTA, 146–151 PP KTP, 146, 149, 150 PP RTA, 146, 149–150 PPLN, 142, 145–149, 153, 155, 157–162, 166–173 quasi-phase-matching (QPM), 142, 143, 145–148, 150, 151, 162, 170–174 RbTiOAsO4 (RTA), 145, 146, 149–151 single longitudinal mode (SLM), 165, 166, 168–170, 173 singly resonant oscillator (SRO), 147, 167 threshold

180

Index

– of an OPO, 145 – of a single-pass OPG, 145

zincblende, 171 ZnGeP2 (ZGP), 153, 155–160, 163–173 ZnSe, 171

Mid-Infrared Ultrafast and Continuous-Wave Optical Parametric Oscillators Majid Ebrahimzadeh School of Physics and Astronomy, University of St Andrews North Haugh, St Andrews, Fife KY16 9SS, Scotland, UK [email protected] Abstract. This chapter provides an overview of optical parametric oscillators (OPOs) in the ultrafast picosecond and femtosecond time-scales as well as the continuous-wave (CW) regime, operating in the mid-infrared spectral range at wavelengths above ∼ 2 µm. The treatment includes a discussion of the basic principles of optical parametric generation and amplification, with emphasis on mid-infrared generation, fundamental material and pump laser requirements, and a review of the pertinent mid-infrared materials and phase-matching schemes. The main OPO device configurations and operating principles are highlighted, and the important recent advances in mid-infrared ultrafast and CW OPOs are summarized.

1

Introduction

For nearly 40 years, optical parametric oscillators (OPOs) have been recognized as versatile sources of coherent light for spectral regions inaccessible to conventional lasers [1]. They are potentially capable of providing tunable radiation across extended spectral regions from a single device, offer the advantages of solid-state design, high efficiency and practical output powers, and can deliver output throughout the entire temporal spectrum from the continuous wave (CW) to ultrafast femtosecond time-scales. Operation of OPOs, however, is critically dependent on the spectral and spatial quality of the pump laser as well as the properties of the nonlinear material. Following the demonstration of the first prototype device in 1965 [1], the development of practical OPOs was for many years held back by the absence of suitable laser sources of sufficient coherence and intensity in desired spectral regions and nonlinear materials with favorable optical properties and high damage threshold. These difficulties compounded to render OPOs unreliable sources of cohereht radiation for many practical applications. As a result, there was a conspicuous decline in OPO research and a long period of quiescence in this field for nearly 20 years until, in the 1980s, the advent of new nonlinear materials once again prompted a major resurgence of interest in these devices. The emergence of a new generation of nonlinear optical crystals such as β-BaB2 O4 (BBO), LiB3 O5 (LBO), and KTiOPO4 (KTP), amongst others, with damage thresholds far exceeding those of the more classical materials and superior linear and nonlinear optical properties provided I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 179–219 (2003) c Springer-Verlag Berlin Heidelberg 2003

180

Majid Ebrahimzadeh

new impetus for research efforts in OPOs. At the same time, the availability of high-power laser sources with improved spectral and spatial coherence enabled the development of OPO devices at unprecedented power levels. More recently, the advent of quasi-phase-matched nonlinear materials such as periodically poled LiNbO3 (PPLN) has led to minimal pump power requirements, bringing the operation of OPOs within the reach of commonly available laser pump sources. The remarkable advances in this field have transformed OPOs from proof-of-principle laboratory prototypes to viable sources of coherent radiation and have now firmly established these devices as an important class of practical tunable light sources for many applications. The operating domain of OPOs, once confined to selected spectral and temporal regimes, now spans across extended spectral regions from the near-ultraviolet (near-UV) to the mid-infrared (mid-IR) and over all time-scales from the CW to the ultrafast femtosecond regime. Figure 1 compares the wavelength coverage of a number of OPO devices developed to date with that of several conventional tunable lasers. In the mid-IR spectral range, where there has traditionally been a famine of practical laser sources, OPOs offer particularly attractive alternatives for the generation of coherent radiation. As is evident from Fig. 1, while extended wavelength regions in the mid-IR still remain inaccessible to lasers, OPOs can offer broad and continuous tuning coverage in this range, often from a single device based on one nonlinear crystal. The aim of this chapter is to provide an overview of OPO devices operating in the mid-IR, from basic

Periodically poled LiNbO3 (PPLN) β-BaB2O4 (BBO)

OPOs

LiB3O5 (LBO)

Alexandrite

Tm3+:YAG 3+

Cr :Mg2SiO4

Dye

Co2+:MgF2 3+

Cr :YAG Ti:sapphire

0.2

1.0

Lasers

Cr2+:ZnSe

2.0

3.0

4.0

5.0

6.0

7.0

Wavelength (µm)

Fig. 1. Comparison of the spectral coverage of prominent conventional tunable lasers with that of a number of OPO devices developed to date. Vertical scale has no significance

Mid-Infrared Ultrafast and CW OPOs

181

operation principles to the important recent developments in the field. The treatment includes a description of the basic principles of optical parametric generation and amplification, design criteria and optimization, material and pump laser selection, and a review of devices. While much of the treatment is applicable to OPO devices of all types, particular emphasis is placed on midIR wavelength generation. In the discussion, we will not be concerned with the background to the origin of the nonlinear optical effects and a discussion of crystal optics, which can be found in other reference texts [2,3,4,5]. Other reviews on OPOs can also be found elsewhere in the literature [5,6,7,8,9,10]. We will adopt MKS units throughout.

2

Optical Parametric Process

Optical parametric generation is a second-order nonlinear process involving the interaction of three optical fields at frequencies ω3 , ω2 and ω1 , such that ω3 = ω2 + ω1 .

(1)

The field at ω3 corresponds to an intense input optical pump field, giving rise to a pair of generated fields at ω2 and ω1 . The generated field at the higher frequency, ω2 , say, is usually referred to as the signal, while the field at the lower frequency, ω1 , is termed the idler, although variations in this nomenclature are also frequently used in the literature. The optical parametric process is a consequence of the second-order nonlinear susceptibility χ(2) in noncentrosymmetric crystalline media. More rigorous treatments of the underlying physical principles responsible for the process can be found elsewhere [11,12]. Here, we restrict our attention to the essential features of the treatment and focus on the main results of the analysis. At a fundamental level, the parametric process can be described by considering Maxwell’s wave equation for the propagation of three optical fields, the pump, signal, and idler, at respective frequencies ω3 , ω2 , and ω1 , in a noncentrosymmetric crystalline medium exhibiting second-order nonlinear susceptibility, χ(2) . The propagation of the three optical fields in such a medium involves the solution of Maxwell’s nonlinear wave equation with the secondorder nonlinear polarization as the source term, namely ∂2E ∂2E ∂ 2 P (2) = µε 2 + µ , 2 ∂z ∂t ∂t2

(2)

where P (2) = ε0 χ(2) E 2 is the second-order polarization and E is the electric field of the propagating wave defined as 1 E(z)ei(kz−ωt) + cc , (3) E= 2 with E(z) representing the complex field amplitude. In writing (2), we have ignored nonlinear polarization terms higher than second-order. We have also

182

Majid Ebrahimzadeh

used a scalar notation for convenience and taken the propagation to be along the z-axis. We have also assumed that the medium is lossless, nonconducting, and nonmagnetic, as is generally true in practice. Since the parametric process involves the interaction of three optical fields, the total field E will comprise three harmonic waves representing the pump, signal, and idler, so that E = E 1 (ω1 ) + E 2 (ω2 ) + E 3 (ω3 ) .

(4)

To simplify the analysis further, it is also assumed that the optical fields are infinite uniform plane waves and monochromatic. The parametric interaction process can then be understood by seeking a solution to the nonlinear wave equation (2). This is done by separating (2) into three components at the three different frequencies ω1 , ω2 , ω3 , each of which must separately satisfy the wave equation. Then, by considering the three separate wave equations at each frequency and assuming that the field amplitudes vary only slowly over distances compared to a wavelength, after some manipulation, we obtain the variations of three field amplitudes with propagation as ∂E1 (z) = iκ1 E3 (z)E2∗ (z)ei∆ kz , ∂z ∂E2 (z) = iκ2 E3 (z)E1∗ (z)ei∆ kz , ∂z ∂E3 (z) = iκ3 E1 (z)E2 (z)e−i∆ kz , ∂z

(5a) (5b) (5c)

where κj = (ωj deff /nj c), with j = 1, 2, 3, n is the refractive index, ∆ k = (2) k3 − k2 − k1 is the phase-mismatch parameter, and deff = χeff /2 is the effective nonlinear coefficient representing the appropriate combination of the nonlinear tensor elements taking part in the parametric process. These are the coupled-wave equations governing the parametric interaction of the pump, signal, and idler in a dielectric medium exhibiting second-order nonlinear susceptibility. The coupled-wave equations are the starting point in the analysis of a wide range of nonlinear optical effects and apply universally to any three-wave mixing process involving the second-order susceptibility, where the frequencies of the fields satisfy (1). These include sum-frequency mixing (ω1 + ω2 → ω3 ), difference-frequency mixing (ω3 − ω2 → ω1 ), second harmonic generation (ω + ω → 2ω), and most importantly in the context of this discussion, optical parametric generation (ω3 → ω2 +ω1 ). We notice from (5) that the amplitudes of the three optical fields are coupled to one another through deff . Physically, this coupling provides the mechanism for the exchange of energy among the interacting fields as they propagate through the nonlinear medium. The direction of energy flow in a given three-wave mixing process depends on the relative phase and the intensity of the input fields.


2.1

183

Optical Parametric Gain

In practice, the parametric process is initiated by a single intense pump field at frequency ω3 at the input to a nonlinear crystal. This field, which is provided by a laser, in turn mixes, through the nonlinear susceptibility, with a signal field at ω2 to give rise to an idler field at ω1 = ω3 − ω2 . The idler field so generated in turn mixes back with the pump to produce additional signal and the regenerated signal remixes with the pump to produce more idler. Under suitable conditions (Sect. 2.2), the process can continue in this way until power is gradually transferred from the strong pump to the initially weak signal and idler fields. The generated signal and idler fields can therefore grow to macroscopic levels by draining power from the input pump field as they propagate through the nonlinear crystal. In the absence of a coherent source of signal and idler beams at the input, the initial supply of photons at ω1 and ω2 for mixing with the input pump is provided by the spontaneous break-up of pump photons through spontaneous parametric fluorescence. This process, also referred to as parametric noise or parametric luminescence, may be viewed to arise from the mixing of the zero-point flux of the electromagnetic field at the signal and idler frequency, quantized within the volume of the crystal, with the incoming pump photons, through the nonlinear polarization [13,14]. The effective zero-point flux at the signal and idler is obtained by allowing one half-photon of energy at both ω2 and ω1 or one photon of energy at either frequency to be present in each black-body mode of the quantizing volume. We can obtain the gain and amplification factor for the growth of the signal and idler fields in the parametric process from the solution of the coupled-wave equations (5). The general solution is beyond the scope of the present discussion and can be found elsewhere [15]. However, if it is assumed that the input pump field does not undergo strong depletion with propagation through the medium, then ∂E3 /∂z = 0 in (5c). The coupled-wave equations are reduced to two, with E3 independent of z in both (5a) and (5b). Subject to the initial condition of no input idler field, E1 (z = 0) = 0, and finite input signal, E2 (z = 0) = 0, the net fractional gain in signal intensity with propagation through the nonlinear crystal is obtained as G2 () =

I2 (z = ) sinh2 [Γ 2 2 − (∆ k/2)2 ]1/2 − 1 = Γ 2 2 , I2 (z = 0) [Γ 2 2 − (∆ k/2)2 ]

(6)

where is the interaction length, I = ncε0 EE ∗ /2 is the intensity or flux (in W/ m2 ), and Γ is the gain factor defined as Γ2 =

8π 2 d2eff I3 (z = 0) . cε0 n1 n2 n3 λ1 λ2

(7)

Here, n and λ are the refractive index and wavelength of the respective waves, I3 (z = 0) is the input pump intensity and deff has units of meters/volt. From the same analysis, an expression similar to (6) may be derived for

184

Majid Ebrahimzadeh

the growth of the idler field from its initial zero value at the input to the nonlinear crystal [11]. The case of non-zero input idler as well as signal can also be treated similarly [9] and results analogous to (6) can be derived for the amplification of the generated fields. It is sometimes useful to express the gain factor in the form Γ2 =

8π 2 d2eff (1 − δ 2 )I3 (z = 0) , cε0 n20 n3 λ20

(8)

where δ is the degeneracy factor defined as 1+δ =

λ0 , λ2

1−δ =

λ0 λ1

(0 ≤ δ ≤ 1) ,

(9)

where λ0 (= 2λ3 ) is the degenerate wavelength, and n0 is the refractive index at degeneracy, with n0 ∼ n1 ∼ n2 . The factor δ is a measure of how close the signal and idler wavelengths are to degeneracy. It is clear from (8) that parametric gain has a maximum value at degeneracy, where δ ∼ 0, and decreases for operation away from degeneracy as δ → 1. 2.2

Optical Parametric Amplification

It can be seen from (6) that the magnitude of the nonlinear gain in the parametric process depends on material parameters such as the refractive index, interaction length, and nonlinear coefficient, the signal and idler wavelengths, as well as the input pump intensity. It is also seen that amplification is strongly dependent on the phase-mismatch parameter ∆ k. For ∆ k = 0, the generated fields experience maximum gain, whereas the growth of the parametric waves is severely undermined by an increase in the magnitude of ∆ k. The general functional dependence of G2 () on ∆ k, as represented by (6), can be found in several reference texts [9,12,14]. In practical devices, however, we are usually interested in maximum gain and amplification for the parametric waves, which occurs for ∆ k = 0. This is achieved through phasematching. Under this condition, (6) is reduced to G2 () = sinh2 (Γ ) .

(10)

Parametric devices often operate under different gain conditions, depending on the magnitude of the gain factor, Γ . In the low-gain regime, corresponding to Γ 1, (6) can be approximated by G2 () ≈ Γ 2 2 .

(11)

On the other hand, in the high-gain regime, Γ 1, and so (6) can be simplified to G2 () ≈

1 2Γ e . 4

(12)


185

Therefore, under the phase-match condition (∆ k = 0), the single-pass intensity gain has a quadratic dependence on Γ in the low-gain limit, whereas it increases exponentially with 2Γ in the high-gain limit. Experimentally, the low-gain limit corresponds to parametric generation when using CW or lowto moderate-peak-power pulsed pump sources. This regime is pertinent to most parametric devices, and in particular to optical parametric oscillators (OPOs). On the other hand, the high-gain limit is valid when pumping with high-intensity, pulsed, and amplified laser sources, and corresponds to optical parametric generator (OPG) and optical parametic amplifier (OPA) devices. To obtain an estimate of the magnitude of parametric gain under different pumping conditions, we can consider the main operating regimes of parametric devices. These are summarized in Table 1, where the gain factor, Γ , and the net single-pass intensity gain, G2 (), have been calculated under the phase-matched condition (∆ k = 0) for different pumping conditions, using (8) and (10). We have taken typical values for pump pulse energy, duration, power, focused beam waist radius (w0 ) and crystal length () appropriate to each mode of operation in practice. We have also assumed a typical nonlinear coefficient of deff ∼ 3 pm/V and have considered degenerate operation at 2 µm (λ2 ∼ λ1 ∼ 2 µm) with n3 ∼ n0 ∼ 1.5, for convenience. We can see from Table 1 that the magnitude of the single-pass parametric gain varies from a mere 0.8 % under continuous-wave (CW) pumping with a relatively powerful 5-W laser to as much as ∼ 3.4 × 1031 when using highenergy ultrashort pump pulses. The latter would correspond to the use of high-intensity, mode-locked and amplified laser systems based on regenerative amplification and cavity-dumping schemes. In the intermediate regime of Table 1. Parametric gain factor and the net single-pass intensity gain for different pumping regimes, calculated from (7) and (9). The calculations are based on typical experimental values for pump laser and nonlinear material parameters for each operating regime and assume phase-matched interaction (∆ k = 0) and near-degenerate operation at ≈ 2 µm CW

Q-switched

Mode-locked

Mode-locked amplified

–

10 mJ

15 nJ

10 µJ

Pump pulse duration –

10 ns

100 fs

200 fs

Peak pump power

5W

1 MW

150 kW

50 MW

Focused waist radius (wo )

20 µm

1 mm

15 µm

15 µm

Peak intensity (I3 )

400 kW/cm2 30 MW/cm2 20 GW/cm2

7 TW/cm2

Crystal length ()

10 mm

1 mm

1 mm

Pump pulse energy

10 mm

Γ

0.09

0.77

1.99

37

G2 ()

0.008

0.72

12.88

3.4 × 1031

186

Majid Ebrahimzadeh

pumping with Q-switched nanosecond pulses (see the chapter by Vodopyanov in this volume), optical gains of ≈ 70 % are available, whereas with ultrashort pulses of relatively low energy, corresponding to typical mode-locked lasers (e.g. Ti:sapphire), single-pass intensity gains on the order of ≈ 1300 % are expected. Given that the initial signal intensity at the crystal input is provided by spontaneous parametric emission, it is clear that any meaningful amplification to macroscopic levels over a single pass of the crystal is not practicable, except in the high-energy ultrashort pump pulse regime where large exponential gains are available. This is the configuration corresponding to OPG and OPA devices. In all other operating regimes, discernable output can only be made available by enclosing the nonlinear material within an optical resonator to provide feedback at the generated parametric wave(s). This is the OPO device configuration, which is the most common architecture for parametric devices. In this treatment, we focus mainly on a description of OPO devices, with particular emphasis on mid-IR generation. More extensive discussions of high-gain parametric devices, namely OPGs and OPAs, can be found elsewhere in the literature [15]. 2.3

Mid-Infrared Parametric Generation

From the preceding discussion, it is clear that providing the phase-matching condition (∆ k = 0) can be satisfied, the efficiency of the optical parametric generation process is critically dependent on the gain factor, Γ , through (10)– (12). It is also clear from (7) and (8) that the gain factor is itself determined by a number of inherent material parameters such as the nonlinear coefficient and refractive indices, as well as the pump, signal and idler wavelengths. In the context of mid-infrared generation, these parameters play an increasingly important role in determining the interaction efficiency because of their dependence on wavelength. This can be best understood by writing (7) in terms of a so-called nonlinear figure of merit, F , for the material as 8π 2 1 I3 (z = 0) , (13) Γ =F cε0 λ1 λ2 with F given by deff F = √ . n1 n2 n3

(14)

The nonlinear figure-of-merit thus incorporates the fundamental material parameters. It is essentially a measure of the strength of parametric generation for a given material in a particular wavelength range and for a given set of controllable parameters including the pump intensity and crystal interaction length. The larger the value of F , the higher the nonlinear gain and hence the efficiency of parametric generation. Inspection of (14) reveals a direct dependence of F on the pump, signal and idler wavelengths through dispersion


187

of refractive indices, n1 , n2 and n3 . However, the F parameter also has an indirect dependence on wavelength through dispersion of deff , as described below. In the context of mid-IR parametric generation, these dependencies play an increasingly important role in influencing the magnitude of F , and hence the efficiency of the parametric process. We can obtain a measure of this influence by considering the wavelength dependence of each term on the right-hand-side of (14) separately. Clearly, one can see from the equation that 1 1 ∼ 3/2 , F ∝ √ n1 n2 n3 n

(15)

where, for the sake of simplicity, we have assumed that n1 ∼ n2 ∼ n3 = n. We can also see from (14) that F ∝ deff , with deff itself related to the effective (2) second-order nonlinear susceptibility by deff ∼ χeff /2. On the other hand, we also know, from Miller’s empirical rule [16], that χ(2) is related to the linear susceptibility of the material, χ(1) , through (16) χ(2) (ω3 , ω2 , ω1 ) = ∆ χ(1) (ω3 ) · χ(1) (ω2 ) · χ(1) (ω1 ) , where ∆ is Miller’s constant. We also know that the linear susceptibility is related to the refractive index of the material by χ(1) = n2 − 1. Hence, we can write the dependence of F on deff in terms of the refractive index as 3

F ∝ (n2 − 1) ∼ n6 ,

(17)

where we have again assumed that n1 ∼ n2 ∼ n3 = n for simplicity. Combining (15) and (17), we find the overall dependence of F on material dispersion as F ∝

1 n3/2

· n6 ∼ n9/2 .

(18)

It is thus seen that the nonlinear figure-of-merit is strongly dependent on the refractive indices of the material. This clearly has important implications in the context of mid-IR parametric generation. Because of normal dispersion, the strong dependence of F on refractive index implies that, for a given nonlinear material, the parametric gain decreases towards longer wavelengths. Hence, for a given pump intensity and crystal length, the parametric generation efficiency is progressively diminished in going from the visible and near-IR spectral range towards the mid-IR, purely as a result of fundamental material properties relating to dispersion. At the same time, we see from (13) that regardless of material parameters, for a given pumping intensity, parametric gain also has a direct dependence on the signal and idler wavelengths through Γ ∝√

1 . λ1 λ2

(19)

188

Majid Ebrahimzadeh

Again, this dependence has important consequences for mid-IR parametric generation. We can clearly see that the magnitude of the nonlinear gain decreases with the increase in the signal and idler wavelengths. Hence, for a given nonlinear material and pumping intensity, parametric generation again becomes increasingly less efficient towards the mid-IR spectral regions due to the wavelength factor.

3

Mid-Infrared OPO Devices

In this treatment, we are concerned with a discussion of mid-IR OPOs operating in the CW and ultrafast (picosecond and femtosecond) time-scales. The principal design criterion for optimum operation of these devices is the same as that for other types of OPOs, namely the maximization of the gain factor, Γ , through suitable choice of nonlinear material and laser pump source. Other issues relate to suitable cavity design and appropriate choice of resonance configuration, optimization of phase-matching and focusing parameters. We will not concern ourselves here with a detailed description of OPOs including the different device architectures, cavity resonance configurations, threshold conditions, and efficiencies, which can be found elsewhere [5,6,7,8,9]. Instead, we concentrate on the fundamental design parameters relating to the nonlinear material and laser pump source pertaining to midIR parametric generation. We also provide a survey of the current status of CW and ultrafast OPOs operating in the mid-IR. For the purposes of the discussion, we refer to the mid-IR spectral region as that which includes the wavelength range from ≈ 2 µm to beyond ≈ 5 µm. 3.1

Nonlinear Material

The selection of nonlinear material for mid-IR OPOs is governed by a number of universal factors that are equally applicable to all types of OPO devices. Clearly, a fundamental material parameter is a broad transparency range in the mid-IR and the ability to be phase-matched in the wavelength range of interest. As highlighted in Sect. 2.3, it is also important for the material to have a large nonlinear figure-of-merit, F , for the attainment of the highest efficiency. This, in turn, means that the material must possess a large effective nonlinear coefficient, deff . Other material requirements include a high optical damage threshold, good optical quality and low absorption loss at the pump, signal and idler wavelengths, mechanical, chemical, and thermal stability, and availability in bulk form and sufficiently large size. In addition to the fundamental material parameters highlighted above, there are also a number of other material requirements that are desirable for optimum operation of OPOs. These include favorable phase-matching geometries, small double-refraction and spatial walkoff and, in the case of femtosecond OPOs, low temporal walkoff and group velocity dispersion. It is also


189

desirable for the material to display large tolerances to possible deviations in the spectral and spatial quality of the pump beam, which result in an increase in the magnitude of the phase mismatch ∆ k from zero. Such deviations arise in practice from the finite spectral bandwidth and divergence of the pump beam. The tolerance of the nonlinear crystal to such effects is measured in terms of the so-called spectral acceptance bandwidth and angular acceptance bandwidth, which can be calculated from the rate of change of phase-mismatch with the wavelength spread (∆ λ) and angular divergence (∆ θ) the pump, using a series expansion of ∆ k [2]. The acceptance bandwidths can then be obtained by solving for the quantities ∆ λ and ∆ θ, using the boundary condition |∆ k| ∼ 2π. For a given crystal length, , the acceptance bandwidths set an upper limit to the maximum allowable pump linewidth and divergence before parametric gain is severely diminished. Equivalently, for a given pump linewidth and angular divergence, the acceptance bandwidths determine the maximum length of the nonlinear crystal that can effectively contribute to the nonlinear gain. Since the crystal length follows an inverse relationship with the maximum allowable pump bandwidth and divergence, for shorter crystal lengths larger deviations in pump beam quality can be tolerated and vice versa. For materials that exhibit temperature-dependent refractive indices and hence a temperature-tuning capability, one can similarly define a temperature acceptance bandwidth, ∆ T , which is a measure of the sensitivity of phase-matching, and thus parametric gain, to changes in the crystal temperature. As a general rule, for the attainment of maximum nonlinear gain and minimum OPO threshold, it is advantageous to use materials which exhibit large spectral, angular, and temperature acceptance bandwidths. In the operating regime of CW and ultrafast OPOs, however, there is an additional material requirement that is generally critical for successful device operation. It arises from the relatively low pumping intensities available from the commonly available (low- to moderate-power) CW and mode-locked lasers compared to, for example, pulsed nanosecond pump lasers. This essential material parameter is non-critical phase-matching (NCPM), which allows phase-matched interaction along a principal optical axis of the material with no spatial walkoff. Given the generally low pump powers available from CW and mode-locked pump lasers, the absence of beam walkoff under NCPM enables tight focusing of the pump beam to achieve the necessary intensities without the deleterious effects of beam walkoff. In the absence of a NCPM capability, tight focusing will generally be precluded by the increase in spatial walkoff, with the result that the OPO will not even reach oscillation threshold. Therefore, NCPM is generally a highly important material property in the context of CW and ultrafast OPOs. This requirement is somewhat less stringent when deploying high-power (multi-watt) pump sources for picosecond OPOs, or in femtosecond OPOs, where the high peak powers and short interaction lengths of the material allow the use of critical phase matching (CPM)

190

Majid Ebrahimzadeh

in collinear or noncollinear schemes (Sect. 5.2). However, even in these cases NCPM still offers the advantage of zero spatial walkoff and hence higher nonlinear gain. Another important advantage of NCPM is that under this geometry the angular acceptance bandwidth of the material is maximized, because of the small sensitivity of refractive indices to beam propagation angle along a principal optical axis. Therefore, the NCPM geometry is also highly beneficial when using pump beams of poor spatial quality. 3.2

Laser Pump Source

As with nonlinear materials, the choice of laser pump source is also governed by a number of universal factors relating to the maximization of nonlinear gain. A principal requirement is set by phase-matching to access the wavelength regions of interest and the pump wavelength should obviously be within the transparency range of the material. The pump source must also be of sufficient intensity, as evident from (7), to provide appreciable nonlinear gain for the OPO to reach operation threshold. This, in turn, necessitates a sufficiently low beam divergence to allow high focused intensities and depth of focus within the nonlinear material. Considerations of the phase-mismatch, ∆ k, place further demands on the spectral and spatial coherence of the pump beam. As can be seen from (6), parametric gain is strongly dependent on the phase-mismatch parameter, ∆ k. Maximum gain occurs for ∆ k = 0, while increases in the magnitude of ∆ k result in severe reductions in nonlinear gain. Despite the use of phase-matching techniques, the attainment of perfect phase-matching (∆ k = 0) is generally not possible in practice, because of finite spectral bandwidth and spatial divergence of the pump beam. This leads to an increase in the magnitude of ∆ k, with the net result that the parametric gain is severely reduced from its peak value at ∆ k = 0. The maximum allowable pump bandwidth and divergence can be calculated from considerations of phase-mismatch [17]. As discussed in Sect. 3.1, such limitations to pump beam quality can be equivalently considered in terms of the spatial and spectral acceptance bandwidths for the nonlinear process. To maintain high parametric gains, it is important to employ laser pump sources of narrow linewidth and low beam divergence. From (19), it is also clear that for a given nonlinear material and pumping intensity, the efficiency of parametric generation is diminished in the mid-IR due to the increasing signal and idler wavelengths, regardless of the particular laser pump source. However, another important factor in mid-IR parametric generation is the quantum defect, defined as the ratio of the input pump photon energy to the generated (signal or idler) photon energy (ω3 /ω1,2 ). Foregoing all other material and pump laser considerations, the quantum defect sets the ultimate limit to the maximum power that can be generated in the parametric waves for a given pump wavelength at 100 % photon conversion efficiency. The larger the wavelength difference between the pump and the parametric waves, the larger the quantum defect, and hence the smaller


191

the maximum generated parametric power. For mid-IR parametric generation, where long signal and idler wavelengths are involved, this implies that for a given input intensity, higher output powers will be available with longer pump wavelengths. Hence, it is desirable to use longer wavelength (near-IR) laser pump sources. In practice, however, it is the availability of the suitable pump laser that will dictate the magnitude of the quantum defect. In the absence of practical laser sources at wavelength > 1.2 µm in the nearIR, the Nd:YAG and the Ti:sapphire laser have become the predominant pump sources for mid-IR OPOs operating in the CW and ultrafast regimes. Other novel pump sources including near-IR semiconductor lasers have also been recently used for CW and ultrafast OPOs operating into the mid-IR (Sects. 5.1, 5.2 and 6).

4

Mid-Infrared Nonlinear Materials

Before the advent of the new generation of nonlinear crystals in the early 1980s, the only promising material candidates for mid-IR wavelength generation were the classical crystals such as AgGaS2 , AgGaSe2 , Ag3 AsS3 , CdSe, LiNbO3 and Ba2 NaNb5 O15 . The main optical characteristics of a number of these materials are summarized in Table 2. As can be seen from the table, particularly attractive properties of most of these materials are broad mid-IR transparency (> 10 µm) and large effective nonlinear coefficients (> 10pm/V). On the other hand, the development of CW and ultrafast mid-IR OPOs based on classical materials has been largely precluded by a general lack of NCPM capability for mid-IR parametric generation when pumped in the near-IR using commonly available pump sources such as the Nd:YAG laser. As a result, the majority of mid-IR OPOs based on these materials have only been successfully operated in the pulsed nanosecond regime, where the requirement for NCPM is not imperative due to the larger available pumping intensities. However, even under nanosecond pulsed pumping, the development of practical devices has in many cases been hampered by the relatively low material damage threshold. The low damage tolerance is in part a consequence of the relatively long short-wavelength absorption cut-off in the classical materials, leading to significant absorption of near-IR pumping radiation, and in part due to the relatively poor optical quality resulting in increased absorption and scattering losses in many cases. The short-wavelength absorption edge of classical materials also makes the use of other pump sources such as the Ti:sapphire laser impractical for mid-IR OPO development. The introduction of novel nonlinear materials based on KTiOPO4 (KTP) and its arsenate analogs, particularly KTiOAsO4 (KTA), RbTiOAsO4 (RTA) and CsTiOAsO4 (CTA) in the 1980s, provided a new class of crystals for parametric generation into the mid-IR. While exhibiting significantly lower effective nonlinear coefficients (≈ 4 pm/V) and shorter mid-IR transmission cutoff (≈ 5 µm) than classical materials (Table 2), they offer the important

192

Majid Ebrahimzadeh

Table 2. Characteristics of a number of prominent near- and mid-infrared nonlinear materials, including the classical crystals as well as the new generation of birefringent and quasi-phase-matched materials. For a more comprehensive survey of nonlinear crystals, see [2] Nonlinear Material

Optical category

Transmission range ( µm)

Effective nonlinear coefficient (pm/V)

Observed damage threshold∗ (GW/cm2 )

≈ 13 ≈ 33 ≈ 15 ≈ 18 ≈ 75 ≈5 ≈ 15

∼ 0.05 ∼ 0.05 ∼ 0.05 > 0.05 > 0.03 ∼ 0.3 ∼ 0.003

≈4 ≈4 ≈4 ≈4 ≈ 12

∼ 0.5 ∼ 0.5 ∼ 0.5 ∼ 0.5 0.01–0.2

≈ 16 ≈8 ≈8 ≈8 ≈8

0.3 — — — —

Classical birefringent phase-matched materials AgGaS2 AgGaSe2 Ag3 AsS3 CdSe ZnGeP2 LiNbO3 Ba2 NaNb5 O15

Uniaxial Uniaxial Uniaxial Uniaxial Uniaxial Uniaxial Biaxial

0.5–13 0.71–18 0.6–13 0.75–20 0.74–12 0.33–5.5 0.37–5

New birefringent phase-matched materials KTiOPO4 KTiOAsO4 RbTiOAsO4 CsTiOAsO4 KNbO3

Biaxial Biaxial Biaxial Biaxial Biaxial

0.35–4.3 0.35–5.3 0.35–5.3 0.35–5.3 0.4–5.5

New quasi-phase-matched materials PPLN PPKTP PPRTA PPKTA PPLT ∗

Uniaxial Biaxial Biaxial Biaxial Uniaxial

0.33–5.5 0.35–4.3 0.35–5.3 0.35–5.3 0.28–5.3

Damage thresholds: λ = 1.06 µm; τ = 10 ns [2]

advantage of NCPM, good optical quality, and increased damage threshold. Equally importantly, a short-wavelength absorption cutoff below ≈ 400 nm in these materials enables the use of readily available laser sources without the onset of pump absorption. These characteristics have permitted use of Nd:YAG and Ti:sapphire lasers as the pump source for successful operation of mid-IR OPOs in both CW and ultrafast picosecond and femtosecond regimes. In CW and picosecond operation, however, these devices are characterized by a mid-IR tuning limit of ≈ 3.5 µm, set by the longest idler wavelength that can be phase-matched under NCPM when pumped with the Nd:YAG or Ti:sapphire lasers (Fig. 2). In femtosecond operation, alternative noncollinear CPM schemes can be employed to achieve wavelength extension beyond 5 µm in these materials, as will be discussed in Sect. 5.2. In addition to KTiOPO4

OPO wavelength (Pm)

Mid-Infrared Ultrafast and CW OPOs 4.0

4.0

4.0

3.5

3.5

3.0

3.0

3.0

2.5

2.5

2.5

2.0

2.0

2.0

1.5

1.5

1.5

1.0 40

1.0 90 0

3.5

(a)

50

60

70

80

deff (pm/V)

θ (deg) 13-plane (φ =0°)

193

1.0 20 40 60 80 90 80 70 60 50 40 φ (deg) 12-plane (θ=90°) θ (deg) 23-plane (φ =90°)

4.0 4.0 4.0 (b) 3.5 3.5 3.5 3.0 3.0 3.0 2.5 2.5 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.0 1.0 1.0 40 50 60 70 80 90 0 20 40 60 80 90 80 70 60 50 40 φ (deg) 12-plane (θ=90°) θ (deg) 23-plane (φ =90°) θ (deg) 13-plane (φ =0°)

Fig. 2. (a) The tuning behavior and (b) corresponding effective nonlinear coefficient under type-II birefringent phase-matching in the principal optical planes of RTA for a Ti:sapphire pump wavelength at 830 nm [67]. Note that under NCPM (θ = 90◦ or φ = 90◦ ), wavelength tuning coverage is limited to ≈ 2.5 µm at 830 nm. The tuning behavior is similar in KTA and CTA. The tuning limit in RTA and other arsenate isomorphs of KTP can be extended to ≈ 3 µm using longer Ti:sapphire pump wavelengths [26] and to ≈ 3.5 µm using the Nd:YAG laser pump wavelength at 1.064 µm

and its arsenate isomorphs, there have also been significant developments in other nonlinear materials for mid-IR generation, most notably KNbO3 (KNB). However, a lack of NCPM in this crystal has confined its utility to femtosecond OPOs using noncollinear CPM schemes (Sect. 5.2). At the same time, a particularly important breakthrough in OPO device technology over the past decade, with major impact on mid-IR OPOs, has been the development of quasi-phase-matched nonlinear materials [18]. Whereas the traditional birefringent phase-matching (BPM) methods expolit the natural birefringence of the optically anisotropic crystal to compensate for dispersion [19,20], quasi-phase-matching (QPM) relies on the periodic reversal of the electric dipole domains in the material along the beam propagation direction to achieve phase-matching. The technique was first proposed four decades ago [12], but its practical implementation was not realized until the recent development of reliable fabrication methods in ferroelectric materials [18]. A practical method to achieve domain reversal is through periodic poling of the ferroelectric material by applying a high electrical field (several kV) across the crystal using patterned electrodes. The period of the domain reversal is determined by the coherence length for the parametric interaction

194

Majid Ebrahimzadeh

in the particular wavelength range of interest – it varies typically from a few µm in the visible to a few tens of µm for mid-IR generation. The end result of the QPM process is that the generated signal and idler waves can undergo quasi-continuous growth as they propagate with the pump through the material. While the growth of optical waves in the QPM process is not monotonic, as is the case in BPM, the overriding advantage of the technique is that it can be freely engineered. Moreover, because QPM does not rely on birefringence, the polarization of optical waves and the direction of propagation can be freely chosen to access the largest diagonal elements in the χ(2) susceptibility tensor, not available in BPM, hence leading to the highest effective nonlinear coefficient in the material. This ability more than compensates for the quasi-continuous nature of amplification in QPM compared with BPM and ensures larger growth of the generated waves. The schematic illustration in Fig. 3 provides a qualitative comparison of parametric gain defined under BPM and QPM conditions. The plots correspond to the low

QPM:

BPM:

'k

'k o ∆kQ

k3 k2 k1

Phase-Match Condition:

'k

k3 k2 k1

Parametric Gain (arbitrary)

k2

Phase-Match Condition:

0

k3 k2 k1 − kG ∆kQ

k1

k3 k2 k1 − kG = 0 k2

k3

k1

kG

k3

QPM

2

K

G2 (QPM) ªdeff (QPM) º » «deff (BPM) ¼ G2 (BPM) ¬

K ~ 15 (in LiNbO3) BPM

4S / "

2S / "

0 kG 0

2S / "

4S / "

'k ∆kQ

Fig. 3. Qualitative comparison of parametric gain, as defined through (5), between BPM and QPM. For simplicity we have assumed collinear interaction. The relative gain is represented by η. In the case of BPM, the phase-mismatch is given by ∆ k and the peak gain is centered at ∆ k = k3 − k2 − k1 = 0. In QPM, the phasemismatch is given by ∆ kQ and is centered at ∆ k = k3 − k2 − k1 = kG . While in QPM, an additional phase-mismatch is induced by the presence of the grating vector, kG , the larger available nonlinear tensor coefficients result in substantially higher parametric gains. In the case of LiNbO3 , the increase in peak gain can be as much as 15 times


195

gain operating regime, pertinent to the large majority of OPO devices in all 2 operating regimes, for which Γ 2 2 (∆ k/2) . Under this condition, the general dependence of parametric gain on ∆ k, as defined through (6), can 2 be approximated by a sin2 (∆ k/2)/(∆ k/2) dependence [6], as depicted in Fig. 3. From the plots, the overriding advantage of QPM over BPM is clearly evident. Because of the flexibility provided by QPM, it is in principle possible to obtain parametric generation in any desired wavelength range throughout the material transparency by fabricating the correct poling period. Another highly important advantage of QPM is the ability to use NCPM geometry by choosing a propagation direction along a principal optical axis during the fabrication process, without any constraints imposed by birefringence as is the case in BPM. As noted earlier, this capability is generally imperative in the context of CW and ultrafast OPOs employing commonly available low- to moderate-power laser pump sources. These characteristics make QPM a highly promising route for development of CW and ultrafast OPOs, particularly in the mid-IR, where other birefringent materials cannot be readily used due to the simultaneous requirements highlighted earlier. To date, several quasi-phase-matched nonlinear materials have been successfully developed, with their main optical characteristics included in Table 2. Of these, periodically poled LiNbO3 (PPLN) has become the most widely established material, because of its relatively mature fabrication and poling technology, commercial availability in long interaction lengths (up to 60 mm), and large effective nonlinear coefficient (deff ∼ 16 pm/V) [18]. Another highly important class of new quasi-phase-matched materials is based on KTP and its arsenate isomorphs, where the application of successful poling methods has led to the availability of periodically poled KTP (PPKTP) and RTA (PPRTA), with progress continuing towards the development of PPKTA. Other promising materials include periodically poled LiTaO3 (PPLT) and KNbO3 (PPKNB). As can bee seen from Table 2, all these materials are viable candidates for mid-IR parametric generation, due to the broad transmission range to > 5 µm, short wavelength absorption cutoff < 500 nm, and substantially larger effective nonlinear coefficients than their birefringent counterparts. Combined with a NCPM capability, these characteristics allow the practical development of CW and ultrafast OPOs for the midIR by the use of commonly available near-IR pump sources, most notably the Ti:sapphire and the Nd:YAG laser, without the onset of pump absorption. By suitable design of the poling period, any desired wavelength within the 1–5 µm spectral range can be accessed under NCPM using such schemes. Wavelength tuning across the entire mid-IR band is generally attainable with a single crystal either through temperature tuning (e.g. in PPLN) when using a fixed-frequency pump source (e.g. Nd:YAG) or pump wavelength tuning (e.g. in PPKTP or PPRTA) with a tunable pump laser (e.g. Ti:sapphire), or by combination of both methods. In materials such as PPLN, where several

196

Majid Ebrahimzadeh

different poling periods can be incorporated onto a single crystal, grating tuning can also be used as an alternative method of obtaining coarse wavelength tuning over extended spectral bands.

5

Mid-Infrared Ultrafast OPOs

The ultrafast OPO operating regime corresponds to the deployment of modelocked laser pump sources, which provide picosecond and femtosecond pulses at high repetition rates of typically > 50 MHz. This is a particularly attractive operating regime for OPOs, because the high peak intensities available from the ultrashort pump pulses greatly aid the attainment of large nonlinear gains (see (7) and (8)) to overcome threshold, even in the singly resonant oscillator (SRO) configuration characterized by the highest operation threshold. Moreover, the low energy fluence associated with ultrashort pulses results in increased material damage tolerance, thereby enabling reliable operation of the OPO at high average powers using high-power input pump sources. On the other hand, unlike CW and pulsed nanosecond devices, operation of ultrafast OPOs is based on the principle of synchronous pumping. Because of the instantaneous nature of nonlinear polarization, optical gain in an OPO is available only in the presence of the pump pulse, during which macroscopic amplification of the parametric waves from noise to coherent output must take place. However, in the case of picosecond and femtosecond pulses, the temporal window of the pulse is too narrow to allow sufficient number of cavity round-trips for the build up of the parametric waves over the pump pulse length, even for practical OPO cavity lengths as short as a few millimeters. To overcome this difficulty, the OPO resonator length is matched to the length of the pump laser, so that the round-trip transit time in the OPO cavity is equal to the repetition period of the pump pulse train. In this way, the resonated parametric pulses experience amplification with consecutive coincidences with the input pump pulses as they make successive transits through the nonlinear crystal. In general, the technique is practical only at relatively high pulse repetition rates (> 50 MHz). At lower repetition frequencies, the required OPO cavity lengths for synchronous pumping become too long and cumbersome to be useful in practice. In general, ultrafast OPOs may be classified into either CW or pulsed (quasi-CW) synchronously pumped devices. In CW oscillators, the input pump radiation comprises a continuous train of ultrashort pulses, whereas in pulsed devices the pump consists of trains of pulses contained within a nanosecond or microsecond envelope. With regard to their operating characteristics, CW ultrafast OPOs may be treated as steady-state devices in the same way as CW OPOs, but with the peak pump pulse intensity determining the nonlinear gain. As such, the steady-state analysis of CW OPOs is similarly applicable to CW ultrafast OPOs by using the peak pulse intensity as the incident pump intensity. On the other hand, the operating dynamics of


197

pulsed ultrafast OPOs is analogous to nanosecond oscillators, where a transient analysis taking account of rise time effects is necessary to adequately describe the device behavior. In pulsed ultrafast OPOs, the nonlinear gain is similarly determined by the peak pulse intensity, but as in nanosecond OPOs rise time effects resulting from the finite duration of the pulse envelope lead to an additional loss mechanism. In either case, however, additional effects pertinent to the operation of ultrafast OPOs including temporal walkoff and group-velocity dispersion also have to be taken into account to fully describe the device behavior. In picosecond OPOs involving pulse durations of > 1 ps, such temporal effects can generally be ignored for practical crystal lengths of up to tens of millimeters. However, they become increasingly important in femtosecond OPOs involving pulses of 100 fs or shorter, as discussed in Sect. 5.2. For most practical applications, the CW synchronously pumped OPO is the most desired device configuration because the output consists of a truly continuous train of identical pulses. In pulsed oscillators, the amplitude, intensity and duration of the output pulses can vary across the pulse envelope, and the output does not generally constitute a truly repetitive pulse train. However, because of the significantly higher peak powers available from pulsed mode-locked than the equivalent CW mode-locked pump lasers, operation of ultrafast OPOs has in the past been more readily attainable under pulsed conditions, particularly in SRO configurations of practical interest. For this reason, most of the early ultrafast OPOs based on birefringent materials have been pulsed oscillators, pumped predominantly by the mode-locked and Q-switched (MLQS) Nd:YAG laser and its variations. In contrast, because of the lower peak intensities available from CW mode-locked pump lasers, early attempts in CW ultrafast OPOs were based on a doubly resonant oscillator (DRO) configuration characterized by low oscillation thresholds. However, the concomitant disadvantages of amplitude and spectral instability of the DRO approach limited the practical utility of such devices. With the availability of high-power CW mode-locked laser sources and novel nonlinear materials in recent years, operation of CW ultrafast OPOs in SRO configurations has become a practical reality, circumventing the need for pulsed pumping or the use of the DRO configuration. We therefore focus on a description of mid-IR ultrafast OPOs in the CW SRO configuration, as this represents the most practical and stable mode of operation for ultrafast picosecond and femtosecond OPOs. 5.1

Mid-Infrared Picosecond OPOs

The majority of CW picosecond OPOs demonstrated to date have been based on the mode-locked Nd:YAG laser or its variations as the pump source. For mid-IR generation, an early attempt was based on the classical nonlinear material AgGaS2 in CPM configuration, pumped by a flashlamp-pumped 100 ps Nd:YAG laser [21]. The operation of this device was, however, confined

198

Majid Ebrahimzadeh

to a single idler wavelength of 5.05 µm in the mid-IR. Moreover, to avoid thermal lensing in the AgGaS2 crystal, the pump light had to be mechanically chopped. An average mid-IR idler power of 6 mW was generated from this OPO for 625 mW of pump power. In the meantime, the introduction of KTP and its arsenate isomorphs provided a new class of materials for mid-IR picosecond pulse generation. By using a high-power 40 ps Nd:YLF laser at 1.053 µm and 76 MHz repetition rate to pump KTP in near-NCPM geometry, mid-IR average powers of up to 800 mW were generated over a wavelength range of 3.22–3.28 µm [22]. This device generated a total average power of 2.8 W in the signal and idler for 15 W of input pump power and the signal pulses were of ≈ 12 ps duration. A similar OPO was demonstrated using temporally compressed 2.2 ps pulses from a Nd:YAG laser at 1.064 µm to pump KTP in near-NCPM configuration [23]. This device provided a mid-IR idler coverage from 3.21 to 3.30 µm with an average output power of ≈ 100 mW at 75 MHz, for 4 W of input pump power. The near-IR signal power from this OPO was ≈ 250 mW with signal pulse durations of ≈ 2.8 ps. More recently, operation of a mid-IR picosecond OPO at unprecedented power levels was achieved by using a diode-pumped, CW mode-locked Nd:YVO4 oscillator-amplifier laser system at 1.064 µm to pump the crystal of KTA in a NCPM geometry [24]. With the pump radiation consisting of 7 ps pulses at ≈ 83 MHz repetition rate and an average power of 29 W, the OPO delivered as much as 6.4 W of idler output at 3.47 µm in the mid-IR. The combined signal and idler output power from this OPO was as high as 21 W, corresponding to an external efficiency in excess of 70 %. A major limitation of KTP and its arsenate isomorphs, when pumped by the Nd:YAG laser, is the restricted tuning capability under NCPM. For the attainment of extended mid-IR tuning, the mode-locked Ti:sapphire laser represents an attractive alternative, because the tuning capability of this laser allows wavelength tuning of the OPO under NCPM, without resort to angle (or temperature) phase matching. This also minimizes intracavity reflection losses caused by crystal rotation, thus allowing maximum efficiency to be maintained across the available tuning range. By using NCPM in KTP and a Ti:sapphire laser, tunable picosecond mid-IR pulses in the 2.3–2.9 µm spectral range were generated by tuning the pump laser over 720–853 nm [25]. Average mid-IR output powers of up to ≈ 200 mW and total powers of ≈ 700 mW were produced for 1.6 W of pump. The signal pulse durations from this OPO were ≈ 1.2 ps at ≈ 82 MHz repetition rate. For wavelength coverage further into the mid-IR, the arsenate isomorphs of KTP, namely KTA and RTA, represent excellent material candidates because of their extended transparency beyond that of KTP, as well as their NCPM capability under Ti:sapphire pump tuning. The use of NCPM in KTA has enabled the generation of mid-infrared picosecond pulses with average powers in excess of 100 mW at 81 MHz in pulses of ≈ 3-ps duration [26]. A total signal and idler output power of up to 400 mW could be obtained from the OPO. The


199

wavelength tuning range of this device was from 2.26 to 3.16 µm, limited by the tuning range of the pump laser, as shown in Fig. 4. Ultimately, however, the mid-IR wavelength tuning range available to picosecond OPOs based on KTP and its arsenates under NCPM is limited to ≈3.5 µm with Nd:YAG or Ti:sapphire pump lasers and extension of the tuning range beyond this limit would require practical picosecond laser sources at longer wavelengths. On the other hand, the advent of periodically poled materials has provided new opportunities for the development of CW picosecond OPOs with enhanced tuning capability further into the mid-IR and improved overall performance characteristics based on Nd:YAG and Ti:sapphire pumping. The phase-matching flexibility provided by QPM has permitted picosecond pulse generation across the transparency range of a number of periodically poled materials to beyond 5 µm under NCPM. At the same time, the large effective nonlinear coefficients (deff ≈ 8–16 pm/V) offered by QPM, combined with the availability of long interaction lengths (up to 60 mm in PPLN), have led to minimal power requirements, enabling the deployment of compact, lowto moderate-power diode-pumped solid-state laser pump sources. In particular, compact all-solid-state OPOs based on PPLN or PPRTA and pumped by mode-locked Ti:sapphire and Nd:YLF lasers have been shown to be versatile sources of picosecond pulses with tunability across the entire range of 1–6 µm, often from a single device [27,28,29]. These devices exhibit average pump power thresholds as low as 10 mW, and can provide total output powers of ≈ 400 mW in pulses of ≈ 1–5 ps, with > 10 mW of mid-IR idler power available beyond 5 µm. Figure 5a,b represents the mid-IR wavelength tuning range of picosecond OPOs based on PPRTA and PPLN pumped by Ti:sapphire and Nd:YLF lasers, respectively [27,28]. The configuration of PPLN OPO based on a semi-monolithic cavity design for enhanced power

k3

k2

k1

Output Wavelength (µm)

3.5

(a)

(b)

(i)

3.0 (ii)

2.5

idler

2.0 1.5

(ii)

signal

(i)

1.0 0.76

0.8

0.84

0.88

0.92

Pump Wavelength (µm)

Fig. 4. (a) Collinear phase-matching and (b) the mid-infrared tuning coverage of KTA picosecond OPO under collinear birefringent type II NCPM along the optical x-axis using Ti:sapphire pump tuning [26]. The solid curves (i) and (ii) represent the calculated tuning range based on two different sets of Sellmeier equations for the material

200

Majid Ebrahimzadeh

Output wavelength (µm)

6 Λ = 29 µm

5

30 µm

31 µm

idler 4

(a)

3

(b)

2

signal 1 0.82

0.84

0.86

0.88

0.90

0.92

0.94

Pump wavelength (µm)

Fig. 5. (a) The mid-infrared tuning range of PPRTA picosecond OPO under type I NCPM along the optical x-axis pumped with a Ti:sapphire laser [27]. Using a single grating period, Λ = 30 µm, an idler tuning range to 5 µm can be obtained under pump tuning at room temperature. The solid curves correspond to the calculated tuning range for three different grating periods, Λ = 29, 30, 31 µm using the appropriate Sellmeier data. (b) The tuning coverage of PPLN picosecond OPO with type-I NCPM along the optical x-axis, pumped by a Nd:YLF laser [28]. Under temperature tuning, mid-IR idler wavelengths out to 6.3 µm can be generated using an amplified laser pump source

extraction in the mid-IR and the corresponding tuning range [28] are shown in Fig. 6. Power scaling of CW picosecond OPOs to unprecedented levels has become possible by using flashlamp-pumped Nd lasers in combination with PPLN. By using a picosecond Nd:YLF pump laser, an average mid-IR output power of 2.4 W in pulses of ≈ 45 ps at ≈ 76 MHz repetition rate were generated in PPLN, with a mid-IR tuning range from ≈ 2.15 to ≈ 2.6 µm [30]. The OPO generated a total output power of ≈ 5 W for 7.4 W of input pump power. Substantial enhancements in the mid-infrared output power and photon conversion efficiency have been shown to be attainable by using intracavity difference-frequency-mixing techniques [31]. In this scheme the generated signal photons from the OPO are further difference-frequency-mixed in a second nonlinear crystal within the OPO cavity to provide additional idler photons. In this way, pump-to-idler photon conversion efficiencies in excess of 100 % can be achieved, yielding enhanced mid-IR output power from the OPO, as shown in Fig. 7. In a separate experiment, the use of 80-ps pulses at 76 MHz from a high-power Nd:YAG laser has permitted the generation of more than 4 W of mid-IR idler power from a PPLN picosecond OPO, with a tunable range over 2.2–2.8 µm [32]. With an average input pump power of 18 W, the OPO provided a combined signal and idler output power of up to 12 W at ≈ 65 % extraction efficiency. Operation of CW picosecond OPOs has recently also been extended to novel mode-locked semiconductor pump lasers operating at GHz repetition rates. By using an InGaAs oscillator-amplifier system delivering 7.8 ps pulses at 2.5 GHz, up to 78 mW of mid-IR idler power


201

Fig. 6. (a) Cavity configuration of the semi-monolithic PPLN OPO for the midIR. L: lens; M: mirror; P: prism. (b) Pump tuning range of the PPLN OPO in the mid-IR idler beam for different grating periods at a fixed temperature of 180 ◦ C. The solid curves are the calculated tuning range using the appropriate Sellmeier data for the material

over a tunable range of 1.99–2.35 µm was generated from a PPLN OPO for 900 mW of input pump power [33]. Figure 8 provides a summary of the performance characteristics of CW picosecond OPOs, including mid-IR devices, developed to date. 5.2

Mid-Infrared Femtosecond OPOs

Optical parametric oscillators operating in the femtosecond time domain represent the newest class of parametric devices. The high peak powers available from femtosecond pulses are particularly attractive for the attainment of large nonlinear gains in OPO devices. However, because of the short temporal duration (≈ 100 fs) and large spectral content (≈ 10 nm) of femtosecond pulses, additional effects such as group velocity dispersion, temporal walkoff, and spectral acceptance bandwidths play an important role in the operation of

8 7 6 5 4 3 2

Majid Ebrahimzadeh

(a)

OPO-DFM

OPO

1 0 0 2 4 6 8 10 12 14 16 incident pump power [W]

photon conversion efficiency [%]

idler power [W]

202

120

idler

(b)

100

OPO-DFM

inferred measured

80

idler

60

inferred measured

OPO

40 OPO-DFM

20 difference

inferred

0 0

2

4

6

8 10 12 14 16

incident pump power [W]

Fig. 7. (a) Idler output power and (b) photon conversion efficiency in a picosecond optical parametric oscillator with intracavity difference-frequency mixing [31]

theses devices. Group velocity dispersion often leads to pulse broadening, while temporal walkoff can degrade nonlinear gain or even modify the temporal characteristics of the output pulses. The large spectral bandwidth can also reduce gain as well as set a lower limit to the minimum attainable pulse duration from the OPO. The influence of temporal walkoff and spectral bandwidth can be minimized by using short crystal lengths of typically 1–2 mm, while the effects of group velocity dispersion can be overcome by the inclusion of dispersion compensation in the OPO cavity, as in conventional femtosecond lasers. The choice of nonlinear crystal for femtosecond OPOs, therefore, requires a trade-off with the crystal length. The need for long interaction length for maximum nonlinear gain, see (10),(11),(12), must be traded against the desire for large phase-matching bandwidth and low temporal walkoff. While crystal lengths of 1–2 mm are too short to provide sufficient gain in picosecond and CW OPOs, the large peak powers available with femtosecond pulses can adequately compensate for this shortfall in gain. The high in-crystal peak intensities (few GW/cm2 ) can also induce additional higher-order nonlinear effects such as self-phase-modulation in femtosecond OPOs. Such effects, which are not generally present in picosecond devices, often lead to spectral broadening and chirping of the output pulses and become more pronounced with longer interaction lengths. The spectral broadening due to self-phasemodulation can, however, be exploited for subsequent compression of output pulses. On the other hand, the use of short interaction lengths provides additional phase-matching flexibility in femtosecond OPOs, where alternative collinear or noncollinear CPM schemes can be used for wavelength generation into the mid-IR. These schemes are generally not available to CW and picosecond OPOs, where lower pumping intensities and longer interaction lengths are involved.


203

KTA (7 ps, 83 MHz)

20.0

PPLN (50 ps, 75 MHz)

10.0 PPLN (45 ps, 75 MHz)

5.0

KTP (12 ps, 75 MHz) LBO (22 ps, 75 MHz)

Average Output Power (Watts)

1.0

CTA (1 ps, 80 MHz)

Mid-Infrared KTP (1 ps, 80 MHz) LBO (22 ps, 75 MHz)

0.5

PPLN (4 ps, 120 MHz) KTA (1-3 ps, 80 MHz) LBO (1-2 ps, 80 MHz) LBO (15 ps, 75 MHz)

LBO (1-2 ps, 100 MHz)

PPRTA (1 ps, 80 MHz) PPLN (0.5 ps, 322 MHz) PPLN (1 ps, 75 MHz)

AgGaS2 (50-80 ps, 75 MHz) LBO (1 ps, 80 MHz)

0.1

LBO (1-2 ps, 125 MHz)

PPLN (8 ps, 2.5 GHz)

KTP (1-2 ps, 125 MHz)

0.02

AgGaS2 (50-80 ps, 75 MHz)

PPLN (4 ps, 120 MHz)

0 0.2

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Spectral Coverage (µm)

Fig. 8. Survey of the performance characteristics of several CW picosecond OPOs demonstrated to date. The pump sources have been based primarily on the modelocked neodymium or the Ti:sapphire laser. The average output powers represent the maximum combined power in the signal and idler beams. The indicated tuning limits correspond to the potential tuning range available to the particular crystal in the given phase-matching geometry and pumping configuration, although the experimental tuning range may have been limited by the available mirrors and crystal coatings. The pulse durations correspond to the measured signal or idler pulse width within the tuning range

The first report of a CW femtosecond OPO was in a crystal of KTP pumped by 170-fs pulses from a mode-locked dye laser at 620 nm [34]. To access the necessary peak pumping intensities, the KTP was pumped at the intracavity focus of the dye laser in a noncollinear CPM geometry. This device produced 220 fs signal pulses at milliwatt average power levels in the nearinfrared. The approach was subsequently extended to an externally pumped oscillator where increased output powers of up to 30 mW were generated [35]. The availability of the mode-locked Ti:sapphire laser soon after provided a new source capable of providing femtosecond pulses at substantially higher

8.0

204

Majid Ebrahimzadeh

powers. This laser has since become the primary pump source for femtosecond OPOs, enabling the development of a KTP-based devices capable of producing total average output powers of up to 750 mW at wavelength < 2 µm in the near-IR using noncollinear CPM or collinear NCPM schemes [36,37,38]. Pulse durations of < 100 fs at repetition rates of 75–85 MHz are routinely available from these devices. For mid-IR pulse generation, the noncollinear CPM scheme was soon extended to other KTP-based OPOs [39], as well as to KTA [40,41], CTA [42,43], RTA [44], and KNB [45], providing femtosecond pulses into the 2–5 µm spectral range using pump and angle tuning schemes. Total average powers of hundreds of milliwatts and mid-IR output powers of tens of milliwatts in 100–200 fs pulses have been obtained from these devices. Figure 9 shows the mid-IR tuning coverage of a Ti:sapphire-pumped femtosecond OPO based on KTA in the noncollinear CPM configuration [41]. By using the NCPM scheme in combination with Ti:sapphire pump tuning, operation of a femtosecond OPO based on RTA has also been demonstrated with power thresholds as low as 50 mW, providing mid-IR pulses in the 2–3 µm spectral range at 50–100 mW of average power [46]. More recently, the introduction of periodically poled nonlinear materials has led to further advances in femtosecond OPOs. Devices based on PPLN, PPRTA, and PPKTP have been shown to be highly versatile sources of femtosecond pulses, offering minimum pump thresholds, high output powers, and vast spectral coverage from a single device. Operation of femtosecond OPOs based on periodically poled materials has been demonstrated over a wide spectral range from 1 to 6.8 µm using noncollinear CPM or collinear NCPM schemes combined with temperature, grating, or angle-tuning [47,48,49,50]. Because of the large nonlinear gains available, these oscillators exhibit pump

(a)

z

(b)

s2 θ3

s1

k3

∆θ 2

ρ

∆θ 1

∆θ k

x

k2

k1

Fig. 9. (a) Noncollinear phase-matching. S and k represent the Poynting vector and the wavevector, respectively. (b) The mid-infrared tuning coverage of KTA femtosecond OPO under noncollinear type-II birefringent CPM in the optical xz -plane with angle tuning [41]. The solid curves represent the calculated tuning range based on appropriate Sellmeier equations for the material for two different Ti:sapphire pump wavelengths


205

PPLN PPLN

55 44

Wavelength (µm)

Wavelength (µm)

power thresholds typically well below 100 mW, enabling the use of all-solidstate Ti:sapphire lasers as the pump source [47,49]. These devices can readily provide practical powers in excess of 100 mW in the 1–4 µm range, with milliwatt level output available in the difficult spectral regions beyond 5 µm. Output pulse durations of 100–200 fs at ≈ 80 MHz repetition rate are typically available. Because of the large nonlinear gain bandwidth of periodically poled crystals, combined with the short interaction lengths used in femtosecond OPOs, devices based on such materials have been shown to be particularly flexible in providing extensive wavelength coverage only through cavity length tuning of the OPO [48,49,50]. This provides a highly convenient method for wavelength tuning and without the need for pump, angle, grating, or temperature tuning. This tuning mechanism occurs because the cavity length detuning introduces a loss at the signal wavelength by reducing the synchronism between the pump and signal pulses. To maintain synchronism and optimize gain, the signal shifts to a more favorable wavelength with a group velocity that satisfies a constant round-trip time. This cavity length tuning, which was observed in the first demonstration of a femtosecond OPO [34], is a useful mechanism for tuning the output wavelength, often by as much as 50 nm in birefringent crystals. In periodically poled materials, the large phase-matching bandwidths can provide cavity length tuning over hundreds of nanometers, limited by the bandwidth of OPO mirrors. The large phase-matching bandwidth of PPLN and PPRTA and the correspond-

1

(a)

33 sinc2('kl/2) 0

22 11 19.5

PPRTA

55 44

1

(b)

33 sinc2('kl/2)

22

0

11 20 20.0

20.5

21 21.0

21.5

22 22.0

Grating period (µm)

22.5

23 23.0

28 28

29 30 31 29 30 31 Grating period (µm)

Fig. 10. Parametric gain bandwidth and cavity length tuning in Ti:sapphirepumped femtosecond OPOs based on (a) PPLN and (b) PPRTA. Phase-matching diagrams representing the variation of normalized gain coefficient with the phasemismatch, ∆ k, are shown as shaded areas [49]. The darker shaded areas, corresponding to regions of significant nonlinear gain, indicate the wide phase-matching bandwidths available in these materials. Broadband wavelength tuning, represented by the experimental data, is achieved for a given grating period at a fixed pump wavelength by simply adjusting the OPO cavity length. The pump wavelengths corresponding to the plots in (a) and (b) are 803 nm and 830 nm, respectively

206

Majid Ebrahimzadeh

ing cavity length tuning of femtosecond OPOs based on these materials are shown in Fig. 10a,b, respectively. An alternative approach to mid-IR femtosecond pulse generation has been the use of a cascaded, two-stage pumping arrangement, where the output of a Ti:sapphire-pumped CTA OPO has been used as the pump for a second oscillator based on the classical nonlinear material AgGaSe2 [51]. The wavelength flexibility of the primary OPO provides a suitable pump wavelength above the absorption edge of the material, which avoids material absorption and at the same time facilitates the required phase-matching condition for mid-IR generation. Using this technique, femtosecond pulses in the 4–8 µm spectral range with average powers of up to 35 mW and pulse durations of 300–640 fs have been obtained at 82 MHz repetition rate. Successful implementation of this technique has, however, necessitated the use of high-power 0.9 KTP (75 fs, 90 MHz)

0.8 0.7

Mid-Infrared

0.6

Average Output Power (Watts)

LBO (40 fs, 80 MHz)

0.5

RTA (58 fs, 80 MHz)

0.4

KNB (60 fs, 75 MHz) CTA (56 fs, 75 MHz) PPLN (60 fs, 81 MHz)

0.3

PPRTA (115 fs, 84 MHz) KTP (62 fs, 76 MHz)

0.2

KTA (85 fs, 100 MHz) PPLN (140 fs, 85 MHz)

BBO (13 fs, 80 MHz)

0.1

AgGaSe2 (230 fs, 82 MHz)

0.08 PPLN (150 fs, 84 MHz)

0.06 0.04 KTP (175 fs, 76 MHz)

0.02 0 0.2

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0


Fig. 11. Summary of the performance characteristics of Ti:sapphire-pumped CW femtosecond OPOs demonstrated to date. The indicated tuning limits correspond to the potential tuning range available to the particular crystal in the given phasematching geometry and pumping configuration, although the experimental tuning range may have been limited by the available mirrors and crystal coatings. The output powers on the vertical scale correspond to the maximum combined signal and idler power generated within the demonstrated tuning range. The pulse durations correspond to the measured signal or idler pulse width within the tuning range


207

input laser pump sources. Figure 11 provides a summary of the performance characteristics of several Ti:sapphire-pumped CW femtosecond OPOs, including mid-IR devices, demonstrated to date.

6

Mid-Infrared Continuous-Wave OPOs

Of the different types of OPO devices demonstrated to date, practical development of OPOs in the CW regime has been consistently more difficult than pulsed and synchronously pumped devices because of the substantially lower nonlinear gains available under CW pumping. Indeed, while the attainment of operation threshold in the simplest and most practical cavity configuration – the SRO – has been readily attainable in nanosecond pulsed and ultrafast OPOs, in CW operation it has been beyond the reach of existing birefringent materials and available CW laser sources. To overcome the high CW SRO threshold, a variety of techniques have been deployed for CW OPOs based on alternative resonance configurations [6]. These include the use of the traditional DRO configuration, as well other novel resonance schemes such intracavity SRO (IC-SRO), pump-enhanced SRO (PE-SRO), and triply resonant oscillator (TRO). These techniques have successfully brought about major reductions in CW OPO thresholds, from tens of watts in SROs to as low as a few milliwatts in TROs, thus enabling the use of a wide range of nonlinear materials and laser pump sources in CW OPO development. On the other hand, the use of such multiply-resonant cavities places stringent demands on the mechanical stability of the OPO as well as the frequency stability of the laser pump source [6]. Practical implementation of CW OPOs, therefore, generally requires active stabilization techniques for spectral control, tuning, and power stability and the use of stable single-frequency laser sources in many cases. The recent development of novel birefringent and periodically poled nonlinear materials, new innovations in resonance and pumping techniques, and the advent of stable, high-power solid-state laser pump sources have led to important breakthroughs in CW OPOs in the past few years. These developments have also had an important impact on the advancement of devices for the mid-IR spectral range at wavelengths > 2 µm. Because of the NCPM capability, moderate nonlinear optical coefficients, an extended transparency window to 5 µm, the birefringent material KTP and its arsenate isomorphs, KTA and RTA, have been shown to be promising candidates for mid-IR CW OPOs. Configured in IC-SRO cavities and pumped by Ti:sapphire lasers, devices based on KTP and KTA have been shown to be practical sources of CW mid-IR radiation, providing total output powers of up to 1.46 W, with as much as 840 mW of idler power available in the 2.4–2.9 µm spectral range [52,53]. In alternative TRO and PE-SRO resonance configurations, low-threshold operation of mid-IR CW OPOs has also been achieved with the use of semiconductor diode lasers as the pump source [54,55]. By using

208

Majid Ebrahimzadeh

KTP and RTA under NCPM and GaAlAs diode lasers at ≈ 800 nm, pump power thresholds < 100 mW and mid-IR wavelength tuning in the 2.1–2.7 µm range have been demonstrated through the tunability of the diode laser pump source. Because of a lack of significant temperature tuning in KTP and its arsenate analogs, wavelength tuning in CW OPOs based on these materials is generally achieved through static pump tuning while maintaining the NCPM condition. With the available laser pump sources based on CW Ti:sapphire, semiconductor, and Nd:YAG lasers, the potential mid-IR spectral coverage of CW OPOs based on the KTP family of crystals under NCPM conditions is limited to ≈ 3.5 µm. On the other hand, the most important recent advances in CW OPOs, with a major impact on mid-IR devices, have been brought about by the development of periodically poled materials. In particular, the advent of PPLN with high effective nonlinearity and long interaction lengths has brought the threshold of CW SROs within the reach of CW solid-state lasers and enabled the operation of these devices at unprecedented power levels and conversion efficiencies in simple external pumping schemes. Using a 50-mm long PPLN crystal pumped with a high-power diode-pumped Nd:YAG laser, midIR output powers of as much as 3.6 W have been generated from a CW SRO over a spectral range from 3.25 to 3.95 µm, for 13.5 W of input pump power [56]. The configuration of this OPO and the output power characteristics are shown in Figs. 12a and b, respectively. Wavelength tuning was achieved through temperature tuning or by using different grating periods incorporated onto the single crystal. Alternative tuning schemes include the use of PPLN with fanned gratings, where the grating period is progressively

Fig. 12. (a) Cavity configuration of a high-power mid-IR CW SRO based on 50 mm PPLN crystal in a simple external pumping arrangement [56]. (b) Pump depletion and idler output power as a function of input pump power at a mid-IR wavelength of 3.25 µm. The oscillation threshold is 3.6 W and a maximum idler output of ≈ 3.6 W is obtained for 13.5 W of pump power


209

varied across the crystal [57]. This can allow continuous long-range tuning of single-mode output over extensive spectral regions. With the advent of PPLN, the operation of CW SROs in external pumping configuration has also been extended to high-power semiconductor diode pump lasers [58]. Pumped by an InGaAs laser at 925 nm in a master oscillator-power amplifier configuration and a 38 mm PPLN crystal, an SRO operation threshold of 1.7 W and idler output powers of 480 mW were generated over a tuning range of 2.03–2.29 µm, for 2.5 W of pump power. Appropriate synchronization of pump tuning and cavity length scanning results in a wide, continuous, single-frequency tuning range. More recently, high-power fiber lasers have been successfully used to drive CW PPLN-based SROs [59]. Using an Yb-doped CW fiber laser tunable over 1.031–1.100 µm and delivering 8.3 W of pump power into a 40 mm PPLN crystal, mid-IR idler powers of up to 1.9 W over a spectral range of 2.98–3.7 µm have been generated, with a SRO power threshold at 3.5 W. At the same time, the use of intracavity pumping schemes in combination with PPLN have for the first time permitted CW SRO operation with minimal operation threshold based on commonly available, low-power, diodepumped solid-state lasers [60]. Using a Nd:YVO4 laser pumped by a 1 W diode laser, practical mid-IR output powers of 70 mW over a spectral range from 3.16–4.02 µm have been generated from an IC-SRO in a simple, compact, all-solid-state design, with a diode pump power threshold of only 310 mW. The cavity configuration and the tuning range of this device are shown in Figs. 13a and b. Operation of mid-IR CW OPOs based on PPLN has also been achieved in PE-SRO configurations. Pumped by a diode-pumped singlefrequency miniature Nd:YAG laser, 140 mW of idle power over the range 2.29– 2.96 µm has been obtained from a PE-SRO, for 800 mW of pump power. The device had an external pump power threshold of 250 mW and wavelength tuning was available by a combination of temperature and grating period tuning. More recently, operation of a mid-IR PPLN PE-SRO pumped by a CW single-frequency Ti:sapphire laser and tunable from 4.07 to 5.26 µm was also reported [61]. By using a twin-cavity arrangement, mode-hop-free tuning of the single-frequency idler over 10.8 GHz was demonstrated by fine tuning the pump laser over 12.3 GHz. The high optical nonlinearity of PPLN and availability in long interaction lengths has also permitted the development of PE-SRO devices pumped directly by single-stripe semiconductor diode lasers. Operation of a device based on a 50 mm PPLN crystal and pumped at ≈ 810 nm with a singlemode extended-cavity AlGaAs laser has been achieved with a power threshold of only around 25–30 mW [62]. This device could provide up to 4 mW of unidirectional idler power in mid-IR, with tuning coverage from 2.58 to 3.44 µm. The pump tuning behavior and power characteristics of this OPO are shown in Fig. 14a and b, respectively. Using alternative DRO configurations, similar CW OPOs based on PPLN have also been demonstrated by the use of solitary, single-stripe diode lasers [63,64]. These devices have been

210

Majid Ebrahimzadeh

Fig. 13. (a) Configuration of a low-pump-threshold CW IC-SRO based on a 50 mm PPLN crystal with a 1 W diode laser as the primary pump source [60]. (b) Wavelength coverage of the CW PPLN IC-SRO with temperature and grating tuning. The solid curves are the calculated tuning range based on appropriate Sellmeier relations for PPLN. The grating periods corresponding to the curves range from 28.5 µm at the lower end of the signal tuning range to 29.9 µm at the higher end of the signal tuning range (and vice versa over the idler tuning range)

shown to operate with pump powers of less than 20 mW and can provide practical mid-IR output powers of up to 5 mW over a spectral range from 2.2 to 3.7 µm. It is important to note that the potential tuning range available to all PPLN-based CW OPOs under Ti:sapphire, Nd:YAG, or direct diode laser pumping extends across a continuous range of ≈ 1.5–5 µm, limited by the increasing absorption in the material beyond ≈ 5 µm. By suitable choice of OPO mirror and crystal coatings, it is possible to access this entire tuning range with one device incorporating a single PPLN crystal.


211

Idler power (mW)

Wavelength (Pm)

Λ=22.0 µm

3.0 Λ=21.6 µm

2.5 1.2

(a)

3

0.8 (λp ≈ 810 nm)

2

(b)

0.6 0.4

1

Signal power (mW)

1.0 Λ=21.8 µm

3.5

0.2

1.1

0

1.0 806

808

810

812

814

Pump wavelength (nm)

816

30

40 50 60 Pump power (mW)

0.0

Fig. 14. (a) Pump tuning range of CW PE-SRO based on a 50 mm PPLN crystal pumped directly by an AlGaAs laser diode at a fixed temperature of 129 ◦ C for different grating periods. The mid-IR idler coverage is from 2.6 to 3.42 µm. The solid curves are the calculated tuning range based on appropriate Sellmeier relations for PPLN. (b) Output power in the signal (at 1.16 µm) and idler (at 2.68 µm) as a function of input diode pump power. The PE-SRO pump power threshold is ≈ 25 mW

At the present time substantial progress continues in the development of CW OPOs based on PPLN. These devices have now been developed to the point where they can provide widely tunable, stable, single-frequency coherent radiation at practical power levels in the mid-IR. However, effective frequency control and power stabilization of PPLN devices, particularly under high-intensity pumping regimes in SRO configurations, is impeded by the photorefractive effect, thermal phase-mismatching and green-induced infrared absorption, and so practical operation of these devices requires careful control of these parameters. In this context, alternative periodically poled materials, particularly phosphates and arsenates of KTP, represent attractive material candidates for mid-IR CW OPOs, because of the absence of such detrimental effects, an equally extended transparency range to > 5 µm, and lower coersive fields for poling. However, the shorter available interaction lengths (typically < 20 mm) and lower effective nonlinear coefficients lead to higher operation thresholds than in equivalent PPLN devices. This precludes SRO operation in external pumping configurations even with a high-power multiwatt laser pump source, and so other approaches based on the DRO, PE-SRO and IC-SRO are the most viable route to the development of mid-IR CW OPOs based on these materials [65,66]. At the present time, the KTP family of periodically poled materials also offer more limited tuning flexibility than PPLN, because of the lack of significant temperature tuning and the difficulty in incorporating several different grating periods onto a single crystal due to limited apertures. However, the continuing advances in poling and fabrication technology are expected to pave the way for the practical development of stable and widely tunable CW OPOs for the mid-IR based on these materials. Figure 15 provides a summary of the spectral coverage and power

212

Majid Ebrahimzadeh 5 PPLN

4 3

CW Output Power (Watts)

2

KTA

1

KTP

Mid-Infrared

0.8 LBO

0.6

KTP, PPKTP

PPLN

0.4 0.2 0.1

LBO

PPRTA

PPLT

MgO: LiNbO3

PPKTP

0.08

PPLN

0.06 PPLT

0.04

PPLN

0.02 0 0.2

1.0

2.0

3.0

4.0

5.0

6.0


Fig. 15. Survey of a number of CW OPOs demonstrated to date. The indicated tuning limits correspond to the potential tuning range available to the particular crystal in the given phase-matching geometry and pumping configuration, although the experimental tuning range may have been limited by the available mirrors and crystal coatings. The output powers on the vertical scale correspond to the maximum combined signal and idler power generated within the demonstrated tuning range

performance of several CW OPOs, including mid-IR devices, developed to date.

7

Summary

This chapter has provided a description of a particular class of solid-state laser technology for the mid-IR based on OPOs. The treatment has included a summary of the basic principles of parametric generation, with emphasis on mid-IR generation, and a survey of the most recent developments in OPO devices operating in the high-repetition-rate ultrafast as well as pure CW regime. The current status of ultrafast OPO technology permits access to mid-IR wavelengths to ≈ 6 µm in picosecond and up to ≈ 8 µm in femtosecond operation, using a variety of pumping configurations and phase-matching schemes. In picosecond operation, mid-IR average powers in excess of 6 W and pulse durations from typically < 1 to > 50 ps have been generated from OPO devices. The operation of picosecond OPOs has also been extended to diode-pumped all-solid-state architectures and even direct diode-pumping


213

using novel semiconductor laser sources. In the femtosecond regime, mid-IR average output powers of up to a few hundred milliwatts can be obtained in pulse durations from typically < 100 to > 200 fs. In CW operation, OPOs have been shown to be capable of providing mid-IR output powers of up to ≈ 4 W and wavelength coverage to > 5 µm. At the same time, innovations in pumping schemes, novel cavity designs and resonance configurations have led to substantial reductions in power thresholds, bringing the practical operation of CW OPOs within the reach of diode-pumped solid-state lasers as well as semiconductor and fiber pump lasers. A key element in the rapid advancement of ultrafast and CW OPOs into the mid-IR has been the recent development of new nonlinear materials, replacing the more traditional nonlinear crystals. In particular, the advent of PPLN, PPRTA and other periodically poled materials, as well as birefringent materials such as KTP and its arsenate isomorphs have greatly aided the recent advances in this area. On the other hand, the development of practical ultrafast and CW OPOs for mid-IR wavelengths beyond 5 µm remains difficult due to the absorption cutoff in the exiting materials, and further progress in this area is critically dependent on new advances in material science. A particularly important route will be the advancement of QPM technology to the classical nonlinear crystals with extended transparency well beyond 5 µm. This will provide yet another new class of periodically poled materials with further enhanced nonlinearities and flexible NCPM capability for wavelength generation into the 5–20 µm spectral range. Other recently developed mid-IR materials such as LiInS2 , which exhibit a short-wavelength cutoff below ≈ 500 nm, offer a promising route for femtosecond pulse generation in the 5–10 µm range using Ti:sapphire pump lasers. Another important factor will be the advancement of practical tunable CW and mode-locked laser sources for wavelengths significantly longer than 1 µm (e.g. Cr:YAG, Cr:ZnSe, etc.). This would allow the development of CW and ultrafast midIR OPOs based on a number of existing classical materials (e.g. ZnGeP2 , AgGaSe2 , etc.) in birefringent phase-matching, by avoiding pump absorption as well as enabling flexible NCPM schemes to be used. In combination with temperature- or pump-tuning, such an approach could provide tunable generation at wavelengths longer than 5 µm. The use of longer pump wavelengths will also enable the attainment of higher output power efficiencies in the mid-IR by reducing the quantum defect between the pump and the generated photons. Other potential approaches include the use of cascaded, two-stage pumping arrangements using two OPOs in series. Here, the tunable output in a suitable wavelength range from the first OPO would be used to provide convenient phase-matching in a second OPO based on a classical material to obtain wavelength extension beyond 5 µm. While such a technique has already been demonstrated using high-power nanosecond and femtosecond pump sources, its extension to picosecond and CW OPOs has not yet been demonstrated due to the substantially lower peak intensities available.

214

Majid Ebrahimzadeh

To make such an approach practicable with moderate pumping intensities in the CW, picosecond, and femtosecond regime, a promising route would be the deployment of intracavity pumping schemes, where the second OPO crystal is placed internal to the cavity of the first OPO to access higher circulating pumping intensities. Other emerging technologies for mid-IR OPOs include integrated waveguide devices based on conventional nonlinear crystals as well as isotropic semiconductors based on GaAs technology using novel phase-matching techniques. The potential for the advancement of CW and ultrafast OPOs further into the mid-IR therefore remains strong and the deployment of the different techniques and technologies highlighted above will continue to pave the way for further progress in this area.

References 1. J. A. Giordmaine, R. C. Miller: Tunable coherent parametric oscillation in LiNbO3 at optical frequencies, Phys. Rev. Lett. 14, 973–976 (1965) 179 2. V. G. Dimitriev, G. G. Gurzadyan, D. N. Nikogosyan: Handbook of Nonlinear Optical Crystals (Springer, Berlin, Heidelberg 1991) 181, 189, 192 3. M. Ebrahimzadeh, A. I. Ferguson: Novel nonlinear crystals, In Principles and Applications of Nonlinear Optical Materials (Chapman and Hall, London 1993) pp. 99–142 181 4. C. L. Tang: Nonlinear optics, In Handbook of Optics (McGraw-Hill, New York 1995) pp. 38.3–38.26 181 5. R. L. Sutherland: Handbook of Nonlinear Optics (Marcel Dekker, New York 1996) 181, 188 6. M. Ebrahimzadeh, M. H. Dunn: Optical parametric oscillators, In Handbook of Optics, Vol. 4 (Optical Society of America, Washington, D. C. 2000) 181, 188, 195, 207 7. S. E. Harris: Tunable optical parametric oscillators, Proc. IEEE 57, 2096–2113 (1969) 181, 188 8. R. L. Byer: Optical parametric oscillators, In H. Rabin and C. L. Tang (Eds.), Treatise in Quantum Electronics (Academic Press, New York 1973) pp. 587–702 181, 188 9. R. G. Smith: Optical parametric oscillators, In A. K. Levine and A. J. DeMaria (Eds.), Lasers (Marcel Dekker, New York 1976) pp. 189–307 181, 184, 188 10. M. H. Dunn, M. Ebrahimzadeh: Parametric generation of tunable light from continuous-wave to femtosecond pulses, Science 286, 1513–1517 (1999) 181 11. R. W. Boyd: Nonlinear Optics (Academic Press, New York 1992) 181, 184 12. J. A. Armstrong, N. Bloembergen, J. Ducuing, P. S. Pershan: Interaction between light waves in a nonlinear dielectric, Phys. Rev. 127, 1918–1939 (1962) 181, 184, 193 13. W. H. Louisell, A. Yariv, A. E. Siegman: Quantum fluctuations and noise in parametric processes, Phys. Rev. 124, 1646 (1961) 183 14. Y. R. Shen: Principles of Nonlinear Optics (Wiley, New York 1984) 183, 184 15. R. Danielius, A. Piskarskas, A. Stabinis, G. P. Banfi, P. Di Trapani, R. Righini: Traveling-wave parametric generation of widely tunable, highly coherent femtosecond light pulses, J. Opt. Soc. Am. B 10, 2222–2231 (1995) 183, 186


215

16. R. C. Miller: Optical second harmonic generation in piezoelectric crystals, Appl. Phys. Lett. 5, 17 (1964) 187 17. M. Ebrahimzadeh, A. J. Henderson, M. H. Dunn: An excimer-pumped β-BaB2 O4 optical parametric oscillator tunable from 354 nm to 2.370 µm, IEEE J. Quantum Electron. 26, 1241–1252 (1990) 190 18. L. E. Myers, R. C. Eckardt, M. M. Fejer, R. L. Byer, W. R. Bosenberg, J. W. Pierce: Quasi-phase-matched optical parametric oscillators in bulk periodically poled LiNbO3 , J. Opt. Soc. Am. B 12, 2102–2116 (1995) 193, 195 19. J. A. Giordmaine: Mixing of light beams in crystals, Phys. Rev. Lett. 8, 19–20 (1962) 193 20. P. D. Maker, R. W. Terhune, M. Nissenoff, C. M. Savage: Effects of dispersion and focusing on the production of optical harmonics, Phys. Rev. Lett. 8, 21–22 (1962) 193 21. E. C. Cheung, K. Koch, G. T. Moore: Silver thiagallate, singly-resonant optical parametric oscillator pumped by a continuous-wave mode-locked YAG laser, Opt. Lett. 19, 631 (1994) 197 22. Ch. Grasser, D. Wang, R. Beigang, R. Wallenstein: Singly resonant optical parametric oscillator of KTiOPO4 synchronously pumped by the radiation from a continuous-wave mode-locked Nd:YLF laser, J. Opt. Soc. Am. B 10, 2218– 2221 (1993) 198 23. J. Chung, A. E. Siegman: Singly resonant continuous-wave mode-locked KTiOPO4 optical parametric oscillator pumped by a Nd:YAG laser, J. Opt. Soc. Am. B 10, 2201–2210 (1993) 198 24. B. Ruffing, A. Nebel, R. Wallenstein: All-solid-state CW mode-locked picosecond KTiOAsO4 (KTA) optical parametric oscillator, Appl. Phys. B 67, 537 (1998) 198 25. A. Nebel, C. Fallnich, R. Beigang, R. Wallenstein: Noncritically phase-matched continuous-wave mode-locked singly resonant optical parametric oscillator synchronously pumped by a Ti:sapphire laser, J. Opt. Soc. Am. B 10, 2195–2200 (1993) 198 26. S. French, M. Ebrahimzadeh, A. Miller: High-power, high-repetiton-rate picosecond optical parametric oscillator for the near- to mid-infrared, Opt. Lett. 21, 131–133 (1996) 193, 198, 199 27. G. T. Kennedy, D. T. Reid, A. Miller, M. Ebrahimzadeh, H. Karlsson, G. Arvidsson, F. Laurell: Broadly tunable mid-infdrared picosecond optical parametric oscillator based on periodically poled RbTiOAsO4 , Opt. Lett. 23, 503–505 (1998) 199, 200 28. L. Lefort, K. Peuch, G. W. Ross, Y. P. Svirko, D. C. Hanna: Optical parametric oscillation out to 6.3 µm in periodically poled lithium niobate under strong idler absorption, Appl. Phys. Lett. 73, 1610–1612 (1998) 199, 200 29. M. Ebrahimzadeh, P. J. Phillips, S. Das: Low-threshold, mid-infrared optical parametric oscillation in periodically poled LiNbO3 synchronously pumped by a Ti:sapphire laser, Appl. Phys. B 72, 793 (2001) 199 30. K. Finsterbusch, R. Urschel, H. Zakarias: Fourier-transform-limited, high-power picosecond optical parametric oscillator based on periodically poled lithium niobate, Appl. Phys. B 70, 741–746 (2000) 200 31. M. E. Dearborn, K. Koch, G. T. Moore, J. C. Diels: Greater than 100% photonconversion efficiency from an optical parametric oscillator with intracavity difference-frequency mixing, Opt. Lett. 23, 759–761 (1998) 200, 202

216

Majid Ebrahimzadeh

32. C. W. Hoyt, M. Sheik-Bahae, M. Ebrahimzadeh: High-power picosecond optical parametric oscillator based on periodically poled lithium niobate, Opt. Lett. 27, 1543 (2002) 200 33. A. Robertson, M. E. Klein, M. A. Tremont, K.-J. Boller, R. Wallenstein: 2.5-GHz repetition-rate singly resonant optical parametric oscillator synchronously pumped by a mode-locked diode oscillator amplifier system, Opt. Lett. 25, 657 (2000) 201 34. D. C. Edelstein, E. S. Wachman, C. L. Tang: Broadly tunable high repetition rate femtosecond optical parametric oscillator, Appl. Phys. Lett. 54, 1728–1730 (1989) 203, 205 35. G. Mak, Q. Fu, H. M. van Driel: Externally pumped high repetition rate femtosecond infrared optical parametric oscillator, Appl. Phys. Lett. 60, 542–544 (1992) 203 36. Q. Fu, G. Mak, H. M. van Driel: High-power, 62 fs infrared optical parametric oscillator synchronously pumped by a 76-MHz Ti:sapphire laser, Opt. Lett. 17, 1006–1008 (1992) 204 37. W. S. Pelouch, P. E. Powers, C. L. Tang: Ti:sapphire-pumped, high-repetitionrate femtosecond optical parametric oscillator, Opt. Lett. 17, 1070–1072 (1992) 204 38. J. M. Dudley, D. T. Reid, M. Ebrahimzadeh, W. Sibbett: Characteristics of a noncritically phase-matched Ti:sapphire-pumped femtosecond optical parametric oscillator, Opt. Commun. 104, 419–430 (1994) 204 39. S. W. McCahon, S. A. Anson, D.-J. Jang, T. F. Boggess: Generation of 3–4 mm femtosecond pulses from a synchronously pumped, critically phase-matched KTiOPO4 optical parametric oscillator, Opt. Lett. 22, 2309–2311 (1995) 204 40. P. E. Powers, S. Ramakrishna, C. L. Tang, L. K. Cheng: Optical parametric oscillation with KTiOAsO4 , Opt. Lett. 18, 1171–1173 (1993) 204 41. D. T. Reid, C. McGowan, M. Ebrahimzadeh, W. Sibbett: Characterization and modeling of a noncollinearly phase-matched femtosecond optical parametric oscillator based on KTA and operating beyond 4 µm, IEEE J. Quantum Electron. 33, 1–9 (1997) 204 42. P. E. Powers, C. L. Tang, L. K. Cheng: High-repetition-rate femtosecond optical parametric oscillator based on CsTiOAsO4 , Opt. Lett. 19, 37–39 (1994) 204 43. G. R. Holtom, R. A. Crowell, L. K. Cheng: Femtosecond mid-infrared optical parametric oscillator based on CsTiOAsO4 , Opt. Lett. 20, 1880–1882 (1995) 204 44. P. E. Powers, C. L. Tang, L. K. Cheng: High-repetition-rate femtosecond optical parametric oscillator based on RbTiOAsO4 , Opt. Lett. 19, 1439–1441 (1994) 204 45. D. E. Spence, S. Wielandy, C. L. Tang: High average power, high repetition rate femtosecond pulse generation in the 1–5 µm region using an optical parametric oscillator, Appl. Phys. Lett. 68, 452–454 (1996) 204 46. D. T. Reid, M. Ebrahimzadeh, W. Sibbett: Noncritically phase-matched Ti:sapphire-pumped femtosecond optical parametric oscillator based on RbTiOAsO4 , Opt. Lett. 20, 55–57 (1995) 204 47. K. C. Burr, C. L Tang, M. A. Arbore, M. M. Fejer: Boradly tunable mid-infrared femtosecond optical parametric oscillator using all-solid-state-pumped periodically poled lithium niobate, Opt. Lett. 22, 1458–1460 (1997) 204, 205


217

48. D. T. Reid, Z. Penman, M. Ebrahimzadeh, W. Sibbett, H. Karlsson, F. Laurell: Broadly tunable infrared femtosecond optical parametric oscillator based on periodically poled RbTiOAsO4 , Opt. Lett. 22, 1397–1399 (1997) 204, 205 49. D. T. Reid, G. T. Kennedy, A. Miller, W. Sibbett, M. Ebrahimzadeh: Widely tunable near- to mid-infrared femtosecond and picosecond optical parametric oscillators using periodically poled LiNbO3 and RbTiOAsO4 , IEEE J. Sel. Top. Quantum Electron. 4, 238–248 (1998) 204, 205 50. P. Loza-Alvarez, C. T. A. Brown, D. T. Reid, W. Sibbett, M. Missey: Highrepetition-rate ultrashort-pulse optical parametric oscillator continuously tunable from 2.8 to 6.8 µm, Opt. Lett. 24, 1523–1525 (1999) 204, 205 51. S. Marzenell, R. Beigang, R. Wallenstein: Synchronously pumped femtosecond optical parametric oscillator based on AgGaSe2 tunable from 2 µm to 8 µm, Appl. Phys. B 69, 423–428 (1999) 206 52. F. G. Colville, M. H. Dunn, M. Ebrahimzadeh: Continuous-wave, singly resonant intracavity parametric oscillator, Opt. Lett. 22, 75–79 (1997) 207 53. T. J. Edwards, G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh, F. G. Colville: High-power, continuous-wave, singly resonant intracavity optical parametric oscillator, Appl. Phys. Lett. 72, 1527–1529 (1998) 207 54. M. Scheidt, B. Beier, R. Knappe, K.-J. Boller, R. Wallenstein: Diode-laserpumped continuous-wave KTP optical parametric oscillator, J. Opt. Soc. Am. B 12, 2087–2094 (1995) 207 55. M. Scheidt, B. Beier, K.-J. Boller, R. Wallenstein: Frequency-stable operation of a diode-pumped continuous-wave RbTiOAsO4 optical parametric oscillator, Opt. Lett. 22, 1287–1289 (1997) 207 56. W. R. Bosenberg, A. Drobshoff, J. I. Alexander, L. E. Myers, R. L. Byer: 93% pump depletion, 3.5 W continuous-wave, singly resonant optical parametric oscillator, Opt. Lett. 21, 1336–1338 (1996) 208 57. P. E. Powers, T. J. Kulp, S. E. Bisson: Continuous tuning of a continuous-wave periodically poled lithium niobate optical parametric oscillator by use of a fanout grating design, Opt. Lett. 23, 159–161 (1998) 209 58. M. E. Klein, D. H. Lee, J.-P. Meyn, K.-J. Boller, R. Wallenstein: Singly resonant continuous-wave optical parametric oscillator pumped by a diode laser, Opt. Lett. 24, 1142–1144 (1999) 209 59. P. Gross, M. E. Klein, T. Walde, K.-J. Boller, M. Auerbach, P. Wessels, C. Fallnich: Fiber-laser-pumped continuous-wave singly-resonant optical parametric oscillator, Opt. Lett. 27, 418–420 (2002) 209 60. D. J. M. Stothard, M. Ebrahimzadeh, M. H. Dunn: Low-pump-threshold, continuous-wave, singly resonant optical parametric oscillator, Opt. Lett. 23, 1895–1897 (1998) 209, 210 61. G. A. Turnbull, D. McGloin, I. D. Lindsay, M. Ebrahimzadeh, M. H. Dunn: Extended mode-hop-free tuning using a dual-cavity, pump-enhanced optical parametric oscillator, Opt. Lett. 25, 341–343 (2000) 209 62. I. D. Lindsay, C. Petridis, M. H. Dunn, M. Ebrahimzadeh: Continuous-wave pump-enhanced singly-resonant optical parametric oscillator pumped by an external-cavity diode laser, Appl. Phys. Lett. 78, 871–873 (2001) 209 63. I. D. Lindsay, G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh: Doubly-resonant continuous-wave optical parametric oscillator pumped by a single-mode laser diode, Opt. Lett. 23, 1889–1891 (1998) 209

218

Majid Ebrahimzadeh

64. A. J. Henderson, P. M. Roper, L. A. Borschowa, R. D. Mead: Stable, continuously tunable operation of a diode-pumped doubly resonant optical parametric oscillator, Opt. Lett. 25, 1264–1266 (2000) 209 65. T. J. Edwards, G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh: Continuouswave, singly-resonant optical parametric oscillator based on periodically poled RbTiOAsO4 , Opt. Lett. 23, 837–839 (1998) 211 66. T. J. Edwards, G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh: Continuouswave, singly-resonant optical parametric oscillator based on periodically poled KTiOPO4 , Opt. Exp. 6, 58–63 (2000) 211 67. D. T. Reid, M. Ebrahimzadeh, W. Sibbett: Efficient femtosecond pulse generation in the visible in a frequency-doubled optical parametric oscillator based on RbTiOAsO4 , J. Opt. Soc. Am. B 12, 1157–1163 (1995) 193

Index

angular acceptance bandwidth, 189 coherence length, 193 coupled-wave equations, 182 critical phase matching (CPM), 189 degeneracy factor, 184 gain factor, 183 idler, 181 KNbO3 , 193 non-critical phase-matching (NCPM), 189 nonlinear figure of merit (FOM), 186 optical optical 185 optical 185 optical

parametic amplifier (OPA), 185 parametric generator (OPG),

parametric fluorescence, 183 parametric luminescence, 183 parametric noise, 183 periodic poling, 193 phase matching, 184 – critical, 189 – non-critical, 189 phase mismatch – parameter, 182 PP KTA, 192 PP KTP, 192–195 PP RTA, 192 PPLN, 191–212 quasi-phase-matched material (QPM), 193 signal, 181 spectral acceptance – bandwidth, 189 synchronous pumping, 196

parametric oscillator (OPO), pump field, 181

temperature acceptance bandwidth, 189

Mid-Infrared Fiber Lasers Markus Pollnau1 and Stuart D. Jackson2 1

2

Advanced Photonics Laboratory, Institute for Biomedical Imaging, Optics and Engineering, Swiss Federal Institute of Technology 1015 Lausanne, Switzerland [email protected] Optical Fibre Technology Centre, Australian Photonics CRC. The University of Sydney 206 National Innovation Centre, Australian Technology Park Eveleigh NSW 1430, Australia [email protected]

Abstract. The current state of the art in mid-infrared fiber lasers is reviewed in this chapter. The relevant fiber-host materials such as silicates, fluorides, chalcogenides, and ceramics, the fiber, pump, and resonator geometries, and the spectroscopic properties of rare-earth ions are introduced. Lasers at transitions ranging from 1.9 to 4 µm occurring in the rare-earth ions Tm3+ , Ho3+ , and Er3+ and their population mechanisms are discussed on the basis of the fundamental spectroscopic properties of these ions. Continuous-wave, fundamental-mode power levels ranging from a few mW near 4 µm up to ≈ 10 W near 2 µm have been demonstrated in recent years. Power-scaling methods and their limitations, the possibilities to optimize the population mechanisms and increase the efficiencies of these lasers, as well as the prospects of future mid-infrared fiber lasers in a number of rare-earth ions at transitions in the wavelength range beyond 3 µm and extending to 5 µm are described.

1

Introduction

Since the introduction of the double-clad fiber more than a decade ago and with the recent technological advances in the fields of fiber fabrication and beam-shaped high-power diode lasers, the performance of diode-pumped fiber lasers has steadily improved. Today, fiber lasers can compete with their corresponding bulk crystalline systems in certain applications, especially when transverse-fundamental-mode, continuous-wave (CW) laser operation at output powers in the milliwatt to multiwatt range is required. The increased recent interest in fiber lasers emitting at mid-infrared wavelengths between 2 and 3 µm primarily relates to the high potential of these wavelengths for applications in laser microsurgery. Due to the high absorption of water in the spectral region at 2.7–3.0 µm, high-quality laser cutting or ablation has been demonstrated in biological tissues. In addition, laser wavelengths near 2 µm could be suitable for tissue welding. A number of other potential laser applications in the mid-infrared spectral region, e.g. environmental trace-gas I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 219–255 (2003) c Springer-Verlag Berlin Heidelberg 2003

220

Markus Pollnau and Stuart D. Jackson

detection, are currently becoming increasingly important. In all these applications fiber lasers may find their niches. The high development costs of fabricating fibers with sufficiently low losses in the mid-infrared spectral region has impeded the necessary research efforts in the field of mid-infrared fiber lasers. The currently available fiber materials that are suitable as host materials for specific rare-earth-doped fiber lasers in the spectral region 2–5 µm will be introduced in Sect. 2. More than any other idea, the invention of the double-clad fiber geometry has accelerated the output-power scaling and hence the success of fiber lasers. The various aspects of the fiber, pump, and resonator geometries will be described in Sect. 3. A significant number of spectroscopic investigations has led to a better understanding of the population mechanisms of rare-earth-doped laser systems. The fundamental spectroscopic properties of rare-earth ions in solid-state host materials will be reviewed in Sect. 4. Equipped with this general information, the performance of the most important mid-infrared fiber laser transitions in the wavelength range 2–3 µm can be understood in detail. Sect. 5 will be devoted to the Tm3+ fiber lasers at 1.9 and 2.3 µm, whereas the Ho3+ fiber lasers at 2.1 and 2.9 µm will be discussed in Sect. 6. An impressive example of the variety of population mechanisms and operational regimes in a single system is the Er3+ 2.7 µm fiber laser transition that will be investigated in Sect. 7. At wavelengths beyond 3 µm, it becomes increasingly difficult to find suitable host materials for actively doped laser systems. This statement holds true for glass fibers in the same way as for crystalline materials. The prospects of future midinfrared fiber lasers in this wavelength range will be discussed in Sect. 8. Besides general introductions to the different topics of lasers [1,2] that include many aspects relevant also to mid-infrared fiber lasers, a comprehensive introduction to the field of rare-earth-doped fiber lasers can be found in [3].

2

Fiber Materials

The choice of the fiber material involves a number of considerations: the maximum phonon energy, the environmental durability, the draw ability, the rare-earth solubility, and the purity of the starting materials. The maximum phonon energy of the glass sets the overall infrared transparency range of the fiber and the multiphonon relaxation rates which influence the quantum efficiency. The multiphonon relaxation rates for the common fiber glasses as a function of the energy gap between energy levels are shown in Fig. 1. The optical transparency range relates to both the size of the band gap and also the infrared absorption cut-off, hence to the vibrational frequency ν of the anion–cation bonds of the glass. For an ordered structure, (1) ν = (1/2π) k/M ,

-1

Multiphonon Relaxation Rate (s )

Mid-Infrared Fiber Lasers 10

12

10

11

10

10

221

-1

Borate (1400 cm ) -1

Phosphate (1200 cm )

9

10

8

10

-1

Silicate (1100 cm )

7

10

-1

Germanate (900 cm )

6

10

-1

Tellurite (700 cm )

5

10

4

-1

10

ZBLA (500 cm )

3

10

2

10

1

10

-1

GLS (350 cm )

2000

3000

4000

5000

6000

7000

-1

Energy Gap (cm )

Fig. 1. Calculated and measured multiphonon relaxation rates as a function of the energy gap between energy levels for glasses with different maximum phonon energies. (Data taken from [4,5])

where M = m1 m2 /(m1 + m2 ) is the reduced mass for two bodies m1 , m2 vibrating with an elastic restoring force k. While for disordered structures like glass, this is not an accurate expression, nevertheless, it does highlight the important contributions to the glass transparency. The relative cation– anion bond strength is intimated by the field strength Z/r2 , where Z is the valence state of the cation or anion and r is the ionic radius. Generally, glasses composed of large anions and cations with low field strengths display high transparency in the mid-infrared spectral region. The important physical properties of the popular glasses used for optical fibers are shown in Table 1. Table 1. Properties of popular fiber materials Fiber material

Max. phonon energy (cm−1 )

Silica ZBLAN GLS

1100 [4] 550 [7] 425 [5]

2.1

Propagation losses Infrared transparency (λ at minimum) (dB/km) ( µm) < 2.5 < 6.0 < 8.0

0.2 (1.55 µm) 0.05 (2.55 µm) 0.5 (3.50 µm)

Thermal conductivity (W/K m) 1.38 [6] 0.7–0.8 [8] 0.43–0.5 [9]

Silicates

This glass is perhaps the most important material used for optical fiber production [3,10], however, the maximum phonon energy is high (≈ 1100 cm−1 ) and has so far limited the emission wavelength of mid-infrared fiber lasers using this material to ≈ 2.2 µm [11]. Silica is robust and involves the very effec-

222


tive modified chemical vapor deposition (MCVD) technique for fiber fabrication. Reducing the OH− content in the glass, which has two main absorption peaks in the range 1.3–2.0 µm [12], improves the near-to-mid-infrared utility. Rare-earth ions such as Nd3+ and Er3+ which have high field strengths have low solubility in silicate glass which can lead to clustering and micro-scale phase separation. 2.2

Fluorides

The use of fluoride glasses, especially the heavy-metal fluorides [13,14], as host materials for mid-infrared fiber lasers has found wide acceptance. The most common form of heavy-metal fluoride glass is the fluorozirconate (ZrF4 ) composition and the most widespread fluoride fiber material is ZBLAN [15], a mixture of 53 mol.% ZrF4 , 20 mol.% BaF2 , 4 mol.% LaF3 , 3 mol.% AlF3 , and 20 mol.% NaF. Since it can be readily drawn into single-mode optical fiber [16] it is particularly important to mid-infrared fiber lasers [17]. The large atomic weight of the zirconium atom combined with relatively weak bonding provides a maximum phonon energy for ZBLAN of ≈ 550 cm−1 and allows for high infrared transparency up to ≈ 6 µm. Multiphonon relaxation, however, becomes significant for transitions at wavelengths longer than ≈ 3 µm. Compared to silica, ZBLAN has a lower damage threshold and a lower level of inhomogeneous spectral-line broadening (Sect. 4.1) because the rareearth ion is placed in sites of a less perturbed network. The crystal-field strength is also comparatively weaker [18]. An overview of the spectroscopic properties of rare-earth ions doped into ZBLAN has been given in [7]. 2.3

Chalcogenides

Chalcogenides are composed of the chalcogen elements S, Se and Te [19,20,21]. They are environmentally durable, have a low toxicity and have reasonably large glass forming regions. When the rare-earth ions are doped into these glasses [22], the radiative transition probabilities and, therefore, the absorption and emission cross-sections are high as a result of the high refractive index (≈ 2.6) of the glass and the high degree of covalency of the rare-earth ion with the surrounding medium. Maximum phonon energies of 300–450 cm−1 produce low rates of multiphonon relaxation, see Fig. 1, and therefore high quantum efficiencies. The low thermal conductivity, see Table 1, is however an important factor to be considered in the design of chalcogenide-based lasers. Of the large number of rare-earth chalcogenides studied for luminescent emission, the most important glasses are the sulfide glasses GaLaS (GLS) [23] and GeGaS [24] because of the reasonably high rare-earth solubility. 2.4

Ceramics

Studies into the use of ceramics as host materials for the rare earths have recently made a lot of progress [25]. These ceramics are composed of nano-

Mid-Infrared Fiber Lasers

223

crystallites of materials such as Y3 Al5 O12 (YAG) and can be produced in a simple and cost-efficient process at relatively low temperatures. This allows the fabrication of materials with very high melting points [26] that are difficult to grow by other techniques such as the Czochralski method [27]. This class of materials is also available in fiber geometry [28]. Ceramic fibers combine the characteristics of crystalline materials such as high absorption and emission cross-sections, large thermal conductivity, and even the possibility of doping with transition-metal ions [28] with the convenience of guiding the pump and signal light in a fiber. Currently, the losses of these fibers are comparatively high, but further improvement can be expected.

3

Fiber, Pump, and Resonator Geometries

The light oscillating in a fiber-laser resonator can be either free running or deliberately modulated depending on whether CW or pulsed output, respectively, is desired. Consequently, a large number of techniques for pulsed operation including Q-switching and mode locking of fiber lasers have been explored. These techniques have been investigated intensively for the common laser transitions at 1 µm in Nd3+ and Yb3+ and at 1.5 µm in Er3+ , and are usually described in combination with these lasers. The small fiber size limits the peak power through the damage-threshold intensity (propagating power per core area) and, hence, crystalline lasers in bulk geometries or optical parametric processes are often preferred when high-energy short pulses are needed. This argument accounts especially for mid-infrared ZBLAN-based fiber lasers, because these fibers possess a lower damage threshold compared to silica fibers. The description of mid-infrared fiber lasers is, therefore, confined to CW operation and specific techniques for pulsed operation of fiber lasers are not discussed in this chapter. In an analogous way to the optical excitation of bulk gain media, doped optical fibers can be either end pumped (core pumped) or side pumped (cladding pumped). The former method is less scalable since it relies on the use of expensive high-beam-quality pump sources because core areas are usually < 100 µm2 . On the other hand, the larger cladding area (> 104 µm2 ) allows for high-power diode-array pumping [29,30,31,32,33]. The obvious simplicity of the core-pumping method negates further explanation and we will concentrate on the cladding-pumping technique: one of the most important developments in fiber-laser technology. 3.1

Fiber Designs for Cladding Pumping

In the design of fibers for cladding pumping, the core of the fiber is generally made to guide a single-transverse LP01 mode. The shape of the multimode pump cladding, see Fig. 2, however, remains somewhat flexible and can be shaped with a number of considerations in mind. The pump cladding,

224

Markus Pollnau and Stuart D. Jackson Jacket

Outer cladding

(a)

(b)

Pump cladding

(c)

Core

(d)

Fig. 2. Principal double-clad fiber geometries which include (a) circular shaped pump cladding with axially positioned core, (b) circular shaped pump cladding with off-axially positioned core, (c) rectangular shaped pump cladding and (d) D-shaped pump cladding

which in turn is surrounded by a low-refractive-index transparent polymer or glass, provides a high numerical aperture (NA) of 0.3–0.55 for the pump cladding. There are three main double-clad-fiber layouts: circular, circular with offset core, and rectangular as shown schematically in Fig. 2. Maximum pump-light absorption sees the core near the outer edge of the circular pump cladding [34] because a portion of the launched light is skew to the fiber axis and produces an inner caustic and never crosses the central region of the pump cladding. Scrambling these skew rays by bending [35] or by using a graded and slightly elliptical pump cladding [36] increases the pumpabsorption efficiency as does spatially varying refractive-index fluctuations in inhomogeneous pump claddings [37]. Inner caustics can be avoided by rectilinearly shaping the pump cladding [38] which has the ancillary advantage of matching the shape of diode-array output. The overall absorption coefficient of the fiber is reduced by the ratio of the core area to the area of the pump cladding [34]. The propagation losses for the rectangular-shaped pump cladding are higher and the effective numerical aperture lower as compared to the circular shape [39]; however, in certain cases higher dopant concentrations can provide shorter fiber lengths that also lead to reduced nonlinear effects. A D-shaped or truncated circular pump cladding [40], see Fig. 2d, is also effective while being easier to make than rectangular preforms. The circular-multimode pump cladding may also have the gain medium distributed in a ring around the edge of the pump cladding either discretely or continuously in multi-core [41] and M-profile [42] arrangements, respectively. The effective absorption coefficient is now further increased while maintaining high-beam-quality output. A large-mode-area core [43] can also increase the effective absorption coefficient of the fiber. Recently, double-clad pump schemes have been demonstrated also with holey fibers [44]. These structures offer the additional advantage of singlemode guiding over a broad spectral range [45].


3.2

225

Fiber-Laser Resonators

Typical free-running fiber-laser resonators are shown schematically in Fig. 3. In the simplest resonator, see Fig. 3a, the pump light passes through a dichroic mirror that is highly reflective for the oscillating laser light. Fresnel reflection at the cleaved output end facet of the fiber can provide sufficient feedback for laser oscillation; however, with an output-coupler mirror – and pump retroreflector – placed at the output end of the fiber the optical efficiency can be maximized. In an alternative arrangement, the pump light can be launched into the output end of the fiber, see Fig. 3b. A dichroic mirror oriented at 45◦ to the fiber axis extracts the laser output and a broadband highly reflecting mirror is placed at the rear fiber end. To scale the output power, each end of the fiber can be pumped, see Fig. 3c. Periodic V-grooves [46] or prism coupling [47] along the fiber to distribute the pump access allow one to further scale the output power and are useful for pumping fiber ring resonators. Spectrally combining the output from a number of separate fiber lasers is also a promising power-scaling technique [48,49,50]. The highest reported fiberlaser output powers of 110 W in a singly Yb3+ -doped fiber [51] and 150 W in a Nd3+ ,Yb3+ -codoped fiber [52] have been obtained using arrangements as shown schematically in Fig. 3c. Bragg gratings can substitute the fiber-butted mirror if spectrally well-defined output is required. (a)

Output

Pump Fiber M (b)

Pump

M Fiber M Output

(c)

M Pump

Pump Fiber M Output

Fig. 3. Schematic diagram of resonators used for free-running fiber lasers with (a) a single-end co-propagating pump, (b) a single-end counter-propagating pump and (c) dual end pumps. M represents the mirror

226

3.3


Thermal Issues

As higher pump powers become available from laser-diode systems, it is generally recognized that thermal and thermo-optical issues set limitations to the power scalability of end-pumped bulk-laser systems. Owing to the unfavorable temperature dependence of thermal and thermo-optical parameters [53], the large heat load in the crystal leads, firstly, to a significant temperature increase in the rod, secondly, to strong thermal lensing with pronounced spherical aberrations, and ultimately, to rod fracture in a high-average-power end-pumped system. Due to its geometry, the fiber provides potentially high pump- and signalbeam intensities without the drawbacks of significant thermal and thermooptical effects. Its large surface-area-to-volume ratio means that the heat generated from multiphonon relaxation in the core is dissipated effectively by radiation and convection from the outer surface of the fiber. This is especially true for single-clad, core-pumped single-mode fibers where this ratio is highest [54]. Double-clad fibers have a relatively smaller surface-area-tovolume ratio and thermal issues need to be taken into account [6,55,56]. Thermal management will be required when very high output powers are desired. In particular, for high-power mid-infrared operation, thermal management may be very important because of the decreased quantum efficiency and the consequently higher amount of heat dissipation.

4 Spectroscopic and Laser Properties of Rare-Earth Ions The structure of a glass is less well defined as compared to a crystalline material. The local variation of the chemical environment of active ions in a glass has a number of consequences. Most important, the active ions may undergo chemical reactions during the fabrication process and be incorporated in the host in several oxidation states with different spectroscopic properties. Oxidation states other than the desired one may act as impurities that introduce undesired optical effects such as parasitic pump absorption, the reabsorption of oscillating laser light, the lifetime quenching of the laser ion, and the trapping of the excitation energy. A stable oxidation state of the optically active ion is thus highly desirable. The necessity of a stable oxidation state excludes a number of transition-metal ions from the list of suitable dopants in glass environments. This is one of the possible reasons why examples of transitionmetal-ion-doped lasers in glass hosts are rare. On the other hand, most of the rare-earth ions prefer to stabilize in the trivalent oxidation state and are, therefore, suitable candidates as glass and fiber dopants. This chapter will, therefore, concentrate on the rare-earth ions as active dopants of fiber lasers.


4.1

227

Spectra of Rare-Earth Ions in Glasses

The optical transitions of lanthanide (rare-earth) ions in the visible and infrared spectral region occur within the 4f subshell. This subshell is shielded by the outer 5s and 5p subshells and the influence of the host material is relatively small compared to, e.g., the 3d transitions in transition-metal ions. The electronic structure of trivalent rare-earth ions derives from the perturbation of the 4f energy level in the central-field approximation by the noncentrosymmetric electron–electron interaction, the spin–orbit interaction, and the crystal-field splitting (Stark effect); see the example of the energy-level scheme of Er3+ in Fig. 4. The spin–orbit multiplets are commonly denoted by their 2S+1 LJ terms in Russell–Saunders coupling, although the 4f electrons of lanthanide ions exhibit intermediate coupling and the total angular momenta J of the spin–orbit multiplets are linear combinations of the total orbital angular momenta L and total spins S. Single crystal-field (Stark) transitions between two spin–orbit multiplets cannot be distinguished in glasses at ambient temperature, because inhomogeneous spectral-line broadening occurs due to the local variation of the ligand electric field. Also homogeneous (lifetime) broadening mechanisms are relevant in a number of glasses. This spectral-line broadening makes glasses the preferred hosts when broadband,

4 11

4f Er3+

2 4 4

H S

F

4

I

4

F3/2 4

F7/2 F5/2 2 H11/2 4

S3/2

4

F9/2

4

I9/2

4

I11/2

4

I13/2

CentralNonSpinField Centrosym. Orbit Approx. Splitting Splitting 4

I15/2

CrystalField Splitting

Fig. 4. Energy-level scheme of trivalent erbium indicating the splitting of the 4f 11 configuration in the centralfield approximation by the noncentrosymmetric electron–electron interaction, the spin–orbit interaction, and the Stark splitting by the local electric field of the host material (indicated only for selected spin–orbit multiplets)

228


continuous tunability of lasers is desired. On the other hand, the spectralline broadening leads to lower absorption and emission cross-sections for the same transition in glasses compared to single-crystalline hosts. The reduced cross-sections lead to generally higher pump threshold of laser transitions in glasses, a fact that is compensated in fiber geometry because a high pump confinement is achieved over the whole fiber length. 4.2

Intraionic Processes

Generally, the probability of an allowed electric-dipole transition is seven orders of magnitude larger than that of an allowed magnetic-dipole transition. Since electric-dipole transitions within the 4f subshell are parity forbidden, the intensities of radiative transitions in rare-earth ions are weak and the radiative lifetimes of the emitting states are long, typically in the ms range. Mixing of the 4f states with higher-lying (typically 5d) electronic states of opposite parity at ion sites without inversion symmetry, however, means that electric-dipole transitions become partially allowed and are usually the dominant transitions between 4f electronic states. The oscillator strengths f and integrated absorption and emission cross-sections σ of these spin–orbit multiplet-to-multiplet transitions can be calculated with the help of the semiempirical Judd–Ofelt theory [57,58]. If the degree of inhomogeneous spectralline broadening is relatively small and the absorption and emission spectra remain structured, as is the case for ZBLAN, the cross-sections σ(λ) at individual wavelengths that are relevant to pump absorption and stimulated emission of narrow laser lines must be determined experimentally. Besides ground-state absorption (GSA), excited-state absorption (ESA) of pump photons, see Fig. 5a, can play a significant role in fiber lasers, specifically in the case of high-intensity core pumping. An experimental example will be given later in Sect. 7.1. Since the absorption increases exponentially with the absorption coefficient α(λP ) = N σ(λP ), ESA becomes relevant for the population dynamics of a laser when (a) the ESA and GSA cross-sections σ(λP ) are comparable at the pump wavelength λP and (b) the population density N of the excited state in which the second pump-absorption step originates becomes a significant fraction of the density of ions in the ground state, i.e., a large degree of ground-state bleaching must be present for ESA to play a significant role. A radiative transition from an excited state i to a lower-lying state j is characterized by the radiative rate constant Aij . If the decay occurs to several lower-lying states, the overall radiative rate constant Ai is the sum of all individual rate constants. The branching ratio of each radiative transition is defined as βij = Aij /Ai . Radiative decay of excited states is in competition with nonradiative decay by interaction with vibrations of the host material, called multiphonon relaxation. The rate constant of a multiphonon relaxation process decreases exponentially with the energy gap to the next lower-lying state and with the order of the process, i.e., the number of phonons required


(a)

229

(b) 2

2 Pump (ESA) 1

1 Pump (GSA)

Pump

0

0 Ion

Donor Ion

Acceptor Ion

(c)

(d)

2

2 Laser

1

Acceptor Ion

Laser

1 Pump

Pump

0

0 Sensitizing Ion A

Laser Ion B

Laser Ion A

(e)

(f)

2

2

1

1 Pump

Pump

0

Quenching Ion B

Pump

0 Donor Ion

Acceptor Ion

Donor Ion

Acceptor Ion

Fig. 5. Intra- and interionic processes in fiber lasers: (a) excited-state absorption (ESA); (b) energy migration; (c) sensitization and (d) quenching of a laser ion by an ion of a different type; (e) cross-relaxation and (f ) energy-transfer upconversion

230


to bridge the energy gap [59,60]. This fact is illustrated in Fig. 1 for different glasses. The rate constant of multiphonon relaxation increases with host temperature. The measurable luminescence lifetime τi of an excited state i is the inverse of the sum of the overall radiative rate constant Ai and the rate constant of multiphonon relaxation, Wi . The radiative quantum efficiency is defined as η = Ai /(Ai + Wi ). The influence of multiphonon relaxations is stronger in oxides as compared to fluorides because of the smaller atomic mass m2 of the anion and the larger elastic restoring force k, see (1), due to stronger covalent bonds in oxides [3], both resulting in larger maximum phonon energies in oxides. A brief example: The luminescence lifetime of the 4 I11/2 upper laser level of the erbium 3 µm laser (Sect. 7) is partly quenched by multiphonon relaxation. Typically, nonradiative decay becomes dominant if five or less phonons are required to bridge the energy gap. With an energy gap between the 4 I11/2 and the next lower lying 4 I13/2 levels of ≈ 3400–3500 cm−1 , radiative decay prevails for phonon energies below ≈ 600 cm−1 , roughly the maximum phonon energy of ZBLAN, see Table 1. Fluorides are, therefore, preferred over oxides as host materials for most of the mid-infrared laser transitions. Like absorption, the strength of a stimulated-emission process is characterized by the emission cross-section σ(λL ) of the laser transition. From a simple analysis, for one resonator round-trip of oscillating laser photons, the product τ σ(λL ) with τ the luminescence lifetime of the upper laser level, is identified as a “figure of merit” for a possible laser transition. The larger this product, the lower is the expected pump threshold of the laser transition. This “figure of merit”, however, does not take into account the numerous parasitic effects that can occur in the population dynamics of a laser system, such as pump ESA, reabsorption of laser photons, and energy-transfer processes. It is often these parasitic processes that lead to surprising performance characteristics – as likely in the negative as in the positive sense – and make the interpretation of rare-earth-doped solid-state lasers challenging. Examples will be discussed in Sects. 5–7. 4.3

Interionic Processes

In addition to intraionic excitation and decay mechanisms, radiative energy transfer due to reabsorption of emitted photons by other active ions in the sample and nonradiative energy-transfer processes due to multipole– multipole or exchange interactions between neighboring active ions can occur. Radiative energy transfer leads to an increase in the luminescence lifetime. Among the nonradiative energy-transfer processes, most common is the electric dipole–dipole interaction, which can occur as a direct [61] or phononassisted [62] energy transfer. A direct energy transfer requires spectral resonance between the involved emission and absorption transitions whereas an indirect transfer can also be nonresonant, i.e., an existing energy gap between the emission and absorption transitions involved in the transfer is bridged by


231

one or several phonons. A process that leads to phonon emission has typically a higher probability than a process requiring phonon absorption. Since the electrostatic field of an electric dipole decreases with distance r as r−3 , the probability of an energy transfer between two such dipoles exhibits a strong distance sensitivity of r−6 . Therefore, nonradiative energy transfer occurs predominantly between neighboring active ions. An obvious possibility of an energy-transfer process is shown in Fig. 5b. An excited ion transfers its excitation to a nearby ion of the same type. If this process occurs consecutively between a number of similar ions and the energy is thus transferred over a larger distance, it is called energy migration. Quenching of the luminescence lifetime of an excited state by energy transfer to impurities is often accelerated by energy migration among the excited donor ions [63]. Figure 5 displays further energy-transfer processes that typically occur in rare-earth-doped solid-state lasers. Rare-earth ions of a different type can be deliberately co-doped into the host material in order to influence the laser properties of the lasing ions. Efficient excitation by means of absorption and energy transfer of pump light from sensitizing ions to the upper laser level of the lasing ions, see Fig. 5c, can be exploited when the lasing ions do not sufficiently absorb the pump light at the desired pump wavelength or the dopant concentration of the lasing ions is limited because, e.g., the laser transition terminates in the ground-state multiplet. Similarly, the transfer of excitation from the lower laser level of the lasing ions to nearby quenching ions, see Fig. 5d, is desirable when the lifetime of the lower laser level is extremely long. The low relaxation rate from the lower laser level would otherwise lead to accumulation of excitation in this level, which can result in self-terminating laser behavior and/or bleaching of the ground-state population density and, consequently, decreased GSA. Energy-transfer processes that have both ions in excited states before or after the energy transfer are shown in Fig. 5e,f. In the former case, an excited ion transfers part of its excitation to a nearby ion in its ground state. This process is called cross-relaxation. Its rate increases with the average number of non-excited neighboring ions, i.e., with dopant concentration. Therefore, cross-relaxation leads to concentration quenching of the measured luminescent decay time of the initial excited state involved in the transfer process. In the inverse process – called energy-transfer upconversion (ETU) – excitation is transferred from one to another excited ion, see Fig. 5f. After the absorption of two low-energy pump photons, ETU leads to a single excitation of higher energy and a single high-energy photon may be emitted from the second excited state. In the presence of fast energy migration among the active ions, the excitation is spatially diffused and all these energy-transfer processes can be described by rate-equation analysis using a rate term W Nd Na that comprises a macroscopic energy transfer probability W and the population densities Nd

232


and Na of the initial states of the donor and acceptor ions, respectively [64]. This model, however, is usually not applicable at low dopant concentrations where energy migration is weak. In addition, a number of authors reported on active-ion clusters in fiber materials, see, e.g., [65,66,67,68]. This or any other non-uniform distribution of active ions complicates the analysis of the influence of energy-transfer processes on the performance of rare-earth-doped fiber lasers, as ions within such clusters are more susceptible to interionic processes than isolated ions. In the simplest approach, this distinction defines two different classes of ions that exhibit different population dynamics [69]. 4.4

Overview of Mid-Infrared Fiber Lasers

The relevant parameters of the fiber lasers with the highest output powers reported as of the end of 2001 for each mid-infrared transition are summarized in Table 2. As can be seen, most of the realized mid-infrared fiber lasers used ZBLAN as the host material, because silica glasses, on the one hand, have maximum phonon energies that are too high and hence a transparency that is too low in the wavelength region beyond ≈ 2.2 µm and chalcogenide glasses, on the other hand, are only at the beginning of their career as fiber-laser host materials. The transitions included in Table 2 will be discussed in detail in the following sections of this chapter. Table 2. Mid-infrared fiber lasers Ion Tm3+ Ho3+ Tm3+ Er3+ Ho3+ Ho3+ Er3+ Ho3+ a)

Fiber host

λPump (nm)

λLaser ( µm)

Transition

Silica ZBLAN ZBLAN ZBLAN ZBLAN ZBLAN ZBLAN ZBLAN

787 805 790 792 1150 532 653 890

1.9 2.1 2.3 2.7 2.9 3.22 3.45 3.95

F4 → 3 H6 I7 → 5 I8 3 H4 → 3 H5 4 I11/2 → 4 I13/2 5 I6 → 5 I7 5 S2 → 5 F5 4 F9/2 → 4 I9/2 5 I5 → 5 I6 3

5

Output power 14 W 8.8 W 22 mW 1.7 W 1.3 W 11 mW 8 mWb) 11 mWb)

Slope Eff. a) 46 % 36 % 7% 17 % 30 % 2.8 % 3% 3.7 %

Ref. [70] [71] [72] [73] [74] [75] [76] [77]

The values of the slope efficiency are given versus incident, launched, or absorbed pump power as stated by the authors and are not necessarily comparable to each other b) Operation with the fiber cooled below ambient temperature


233

5 Thulium-Doped Fiber Lasers at 1.9–2.0 µm and 2.3–2.5 µm The use of the Tm3+ ion for mid-infrared fiber laser applications has been widespread mainly as a result of the convenient absorption band near 0.79 µm which allows for AlGaAs diode laser pumping. The primary luminescent transitions of Tm3+ relevant to mid-infrared laser emission are the 3 F4 → 3 H5 transition at ≈ 2.3 µm and the 3 H4 → 3 H6 ground-state transition at ≈ 1.9 µm, see the energy-level scheme in Fig. 6. The 3 F4 level is excited by the 0.79 µm pump wavelength. 5.1

Three-Level Lasers at 1.9–2.0 µm

The first explorations into fiber lasers utilizing the 1.9 µm ground-state transition related to the dye-laser pumping at 797 nm of a Tm3+ -doped silica fiber laser [79]. Overlap of the main absorption band with the emission wavelength of AlGaAs diode lasers quickly resulted in diode-laser pumping of these fiber lasers based on either silica [80] or fluoride [81] glass hosts. A useful phononassisted cross-relaxation process, (3 F4 , 3 H6 ) → (3 H4 , 3 H4 ), can transform one absorbed pump photon into two excitations in the 3 H4 upper laser level of the 2 µm transition [82], see Fig. 6. The efficient room-temperature CW operation of Tm3+ -doped diode-pumped bulk crystalline lasers [83] is largely due to this self-quenching process. This process is highly dependent upon the overall concentration of Tm3+ ions and competition by multiphonon relaxation from the 3 F4 level. High concentrations of Tm3+ in low-phonon-energy glasses enable full exploitation of this beneficial phenomenon. The significantly stronger multiphonon relaxation and corresponding shorter lifetime of the 3 F4 level in silica (≈ 20 µs) means that the cross-relaxation process is significantly weaker in silica as compared to fluoride glass. The large degree of Stark splitting of the 3 H6 ground state provides the 3 H4 → 3 H6 transition with a very broad emission spanning ≈ 400 nm in 3

τ3 = 1.5 ms

F4 Laser 2.3 µm NR

CR

3

H5 NR

3

τ1 = 6.8 ms

H4 GSA

Laser 2.0 µm

3

H6 Tm3+

Fig. 6. Partial energy-level scheme of Tm3+ displaying the measured lifetimes when doped into fluoride glass [78]. NR and CR represent nonradiative decay and cross-relaxation, respectively

234


many hosts and represents one of the broadest luminescent transitions available from any rare-earth ion. The broad emission spectrum produces a large degree of wavelength tunability [84]. Since the Tm3+ 1.9 µm transition can be favorably operated in silica fiber (with its higher peak-power damage threshold compared to ZBLAN fiber), pulses in the range of 190–500 fs have been obtained in additive-pulse [85] or passive [86] mode-locking from this broad emission spectrum. The smaller emission cross-section and the three-level nature of the laser transition usually resulted in relatively higher pump thresholds as compared to the standard Nd3+ -doped fiber lasers. Reabsorption from the ground state of the Tm3+ ions has to be overcome because the groundstate multiplet is the lower laser level. Reducing the population of the higher Stark levels of the ground state by way of cooling the fiber causes emission at shorter wavelengths. Tunability to longer wavelengths can be obtained by variation of the fiber length because of the increased level of reabsorption by the ground state with longer lengths of fiber [87]. Early power-scaling experiments involved the use of the convenient 1.064 µm Nd3+ :YAG laser which pumped the short wavelength side of the 3 H5 level [88]. Pumping the long wavelength side of the 3 H5 level with a high-power 1.319 µm Nd3+ :YAG laser also yields efficient output [89]. Inband pumping of the transition at 1.57 µm in silica [90] and at 1.58–1.60 µm in fluoride glass [91,92] has also been demonstrated. Whilst theoretical modeling of Tm3+ -doped silica fiber lasers [93] confirms that inband pumping is the most efficient pump method for silica based fiber lasers because of the high Stokes efficiency, nevertheless, the wide availability of high-power AlGaAS diode lasers means that diode-cladding-pumped systems in both standing-wave [87,70] and ring-resonator [94] arrangements are perhaps the most practical ways of producing high output power, see Fig. 7. Currently, the Tm3+ -doped silica fiber laser is probably the most mature of the midinfrared fiber-laser systems primarily because of the robustness and convenience offered by the silica glass host. While less efficient than Tm3+ -doped fluoride fiber lasers, the comparatively higher pump threshold relevant to Tm3+ -doped silica fiber lasers is easily provided for by currently available high-power diode-laser pump sources. The maximum output power from highpower Tm3+ -doped fiber lasers is still currently an order of magnitude lower than comparable diode-pumped crystalline systems [95]. 5.2

Four-Level Lasers at 2.3–2.5 µm

The increased quantum efficiency of the 3 F4 level when Tm3+ ions are doped into a ZBLAN host offers a greater range of emission wavelengths. Besides lasers at shorter wavelengths such as ≈ 1.47 µm (3 F4 → 3 H4 ) and ≈ 0.8 µm (3 F4 → 3 H6 ), the mid-infrared four-level CW laser at ≈ 2.3 µm (3 F4 → 3 H5 ) [96,97,72] is of interest, see Fig. 6. Deliberately designing the fiber to have a relatively low Tm3+ -ion concentration reduces cross-relaxation and hence severe lifetime quenching of the 3 F4 level. The lifetime of the lower

Mid-Infrared Fiber Lasers 3+

14

1.8 wt.% Tm

12

3.6 mol.% Tm , 0.4 mol.% Ho

3+

Output Power (W)

235

3+

3+

2.2 wt.% Tm 10 8 6 4 2 0 0

10

20

30

40

50

Incident Pump Power (W)

Fig. 7. Measured output powers from diode-cladding-pumped fiber lasers using 1.8 wt.% Tm3+ -doped silica [86], 2.2 wt.% Tm3+ -doped silica [70], and 3.6 mol.% Tm3+ , 0.4 mol.% Ho3+ -doped fluoride glass [71]

laser level of the 3 F4 → 3 H5 transition is quite short and leads to a low pump threshold and broad tunability which can extend from 2.25 µm to 2.5 µm [78]. Simultaneous lasing on the 3 H4 → 3 H6 transition at 1.9 µm produces a two-color fiber laser [98]. Applications requiring highly efficient output or multi-mid-infrared-wavelength output will benefit from the use of Tm3+ -doped ZBLAN fibers.

6

Holmium-Doped Fiber Lasers at 2.1 µm and 2.9 µm

The use of the Ho3+ ion as the active dopant for fiber lasers opens up a number of very useful mid-infrared transitions. In this section, we will concentrate on the 5 I7 → 5 I8 ground-state transition at ≈ 2.1 µm and the 5 I6 → 5 I7 transition at ≈ 2.9 µm, see the energy-level scheme in Fig. 8. One of the significant shortcomings of Ho3+ , however, is the lack of GSA transitions [99] that overlap with convenient high-power pump sources. As a result, many of the early demonstrations of Ho3+ -doped room-temperature crystalline CW lasers [82] involved sensitizing with Tm3+ in order to access the convenient absorption bands and the practical cross-relaxation process Tm3+ provides. Energy migration amongst the Tm3+ ions and a suitable Tm3+ :Ho3+ concentration ratio ensures that efficient energy transfer to Ho3+ takes place [100,101], see Fig. 8. The results of a recent spectroscopic study of Ho:YAG [102] suggest that the cross-relaxation process (5 I5 , 5 I8 ) → (5 I7 , 5 I7 ) and the related ETU process will probably be important in some highly Ho3+ -doped fibers; however, the cross-relaxation process (5 I6 , 5 I8 ) → (5 I7 , 5 I7 ) and its related ETU process seem to have a negligible effect on the overall population dynamics.

236

Markus Pollnau and Stuart D. Jackson 5

S2

Laser 3.2 µm

5

F5

5

I4

3

F4

3

H5

5

I5

Laser CR 3.9 µm ET

3

H4

5

I6

Laser 2.9µm

5

I7

Laser 2.1 µm

GSA

3

H6

6.1

5

Tm3+

Ho3+

I8

Fig. 8. Partial energy-level scheme of Ho3+ with Tm3+ sensitizer. ET represents energy transfer

Three-Level Lasers at 2.1 µm

As is the case with the ground-state transition of the Tm3+ ion discussed above, the 5 I7 → 5 I8 ground-state transition has been the most widely explored laser transition of Ho3+ . The first fiber laser configuration making use of this transition employed ZBLANP glass (a variant of ZBLAN) and argonion pumping [103]. A year later, this was followed by a demonstration of an argon-ion-pumped Ho3+ -doped silica fiber laser [104]. In both cases, the fiber was singly doped with Ho3+ , the output power < 1 mW, and each needed a relatively high pump power to reach laser threshold. Improvements in the output power and efficiency have been made recently with Yb3+ -doped silica fiber laser pumping of the 5 I6 level [105]; however, the output power has only increased to 280 mW. As mentioned above, a practical method of efficiently generating laser emission on the 5 I7 → 5 I8 transition is to co-dope Ho3+ laser ions with Tm3+ sensitizer ions. The first demonstration of a fiber laser operating with the Tm3+ , Ho3+ system occurred in 1991 [106] when 250 mW was generated at a slope efficiency of 52 % from a Ti:sapphire-pumped fluoride fiber laser. A year later [107], this work was followed by an increase in the Tm3+ concentration to improve the cross-relaxation and resulted in a higher slope efficiency being obtained. Demonstration of a Tm3+ , Ho3+ -doped silica fiber laser soon followed [108,109]; however, owing to lower Tm3+ concentrations and strong multiphonon relaxation of the 3 F4 level which forces weaker cross-relaxation, significantly lower slope efficiencies were measured, especially when pumped at 1.064 µm [11]. To date, the highest output power from a fiber laser operating on the 5 I7 → 5 I8 transition has been produced by a diode-cladding-pumped Tm3+ , Ho3+ -doped fluoride fiber laser [71], see


237

Fig. 7. In an analogous way to recent demonstrations in bulk laser systems [110,111], tandem-pumping Ho3+ with a separate Tm3+ laser operating at 1.9 µm may also prove effective in fibers, because it similarly exploits the cross-relaxation process between Tm3+ ions but avoids any ETU between Ho3+ ions in the 5 I7 upper laser level and excited Tm3+ ions [112]. 6.2

Four-Level Lasers at 2.9 µm

As a result of the infrared absorption cut-off of silica and the strong multiphonon relaxation quenching of mid-infrared transitions of rare-earth ions in this host, four-level fiber lasers operating on the 5 I6 → 5 I7 transition at ≈ 2.9 µm have to date only involved fluoride glass as the host material. The lifetime of the 5 I7 level is longer than the lifetime of the 5 I6 level and hence the 2.9 µm transition can be self-terminating. The first demonstration of a fiber laser using this transition [113] produced only ≈ 13 mW when pumped at a wavelength of 640 nm. High-power cascade lasing at 2.9 µm and 2.1 µm has been employed recently to extend the output power and to remove bottlenecking at the 5 I7 level [114]. With direct pumping of the upper laser level, 1.3 W of output power has been measured from this laser scheme [74]. Sensitizing Ho3+ with Yb3+ ions, see the energy-level scheme in Fig. 9, in order to exploit the more favorable absorption features of Yb3+ has been used in diode-pumped crystalline lasers for the generation of both 2.1 µm [115] and 2.9 µm [116] output. When sensitizing with Yb3+ ions, Ho3+ -doped silica fiber lasers at 2.1 µm may produce significantly more output because Yb3+ can be doped to quite high levels in silica thus ensuring strong absorption and sufficient energy transfer to Ho3+ . Similarly, a Yb3+ -sensitized Ho3+ -doped ZBLAN fiber will also be diode-laser pumpable and may efficiently provide high-power 2.9 µm output without the costly requirement of an intermediate laser system. Initial spectroscopic results look encouraging [117]. ET

5

I5

2

F5/2

5

τ2 = 3.5 ms GSA

I6

Laser 2.9 µm

5

I7

τ1 = 12 ms

Laser 2.1 µm 2

F7/2

5

Yb3+

Ho3+

I8

Fig. 9. Partial energy-level scheme of Ho3+ with Yb3+ sensitizer displaying the measured lifetimes of Ho3+ when doped into fluoride glass host [114]

238

7


Erbium-Doped Fiber Lasers at 2.7–2.8 µm

The first observation of coherent emission near 3 µm from erbium ions was reported in 1967 [118]. In 1983, the first CW lasing near 3 µm was obtained [119]. The first erbium 3 µm fiber laser was demonstrated in 1988 [120]. Although the lifetime of the 4 I13/2 lower laser level exceeds that of the 4 I11/2 upper laser level, CW lasing can be obtained on this four-level-laser transition in ZBLAN without employing special techniques to depopulate the 4 I13/2 lower laser level, because the lower laser level is not fed significantly by luminescent decay or multiphonon relaxation from the upper laser level [121]. In addition, the Stark splitting of the laser levels contributes to population inversion, because the laser transition occurs between a low-lying Stark component of the upper and a high-lying Stark component of the lower laser level [122]. During the relaxation oscillations at the onset of lasing, a red-shift of the lasing wavelength is often observed in erbium 3 µm laser systems [123,124,125], because the excitation energy is accumulated in the long-lived 4 I13/2 lower laser level and the character of the lasing process changes from four-level to three-level lasing [122]. For the same reason, the tunability range of a 3 µm CW laser [126] is narrowed and red-shifted with increasing pump power. Depending on the erbium concentration and fiber geometry, the 3 µm fiber laser has been operated in a number of different regimes that will be discussed in the following sections. 7.1

Excited-State Absorption and Cascade Lasing

Pump ESA has a major influence on the performance of low-doped, corepumped erbium 2.7 µm ZBLAN fiber lasers. Pumping at 980 nm directly into the upper laser level provides the highest Stokes efficiency of ηSt = λp /λl = 35 % [127]. However, ESA at 980 nm from the 4 I11/2 upper laser level [128] is detrimental to lasing and must be avoided. Experimentally, the best pump wavelength [129] is near 792 nm. This wavelength is at the peak of ESA from the 4 I13/2 lower laser level [130], as can be seen from the measured GSA and ESA cross-sections in Fig. 10a. The reason for the strong influence of this ESA is that depletion of the lower laser level by ESA favorably results in a redistribution of its population density and overcomes the bottleneck that results from the long lower-level lifetime. Slope efficiencies obtained in this way were < 15 %. Moreover, a saturation of the output power at 2.7 µm was observed, regardless of the pump wavelength, and the highest reported output powers were in the 20 mW region [131,132]. The output power saturated, because the excitation of the metastable 4 S3/2 level (lifetime ≈ 580 µs [133]) led to inversion with respect to the 4 I13/2 level. A second laser transition at 850 nm repopulated the 4 I13/2 lower laser level of the 2.7 µm transition, see Fig. 10b, causing the 2.7 µm laser to saturate at low output powers [129].


(a)

(b) 24

4

4I15/2 4I13/2 4I11/2

-22

cm2]

20

σ ESA [10

239

F7/2 H11/2 4 S3/2 2

τ5 = 580 µs Laser Laser 1.7µm 850nm

16 12

4

F9/2

4

I9/2

4

I11/2

ESA

8 4 0 780

790

800

810

820

830

τ2 = 6.9 ms Laser 2.7µm

840

Wavelength [nm] 4

τ1 = 9.0 ms

I13/2

GSA

4

I15/2 Er3+

Fig. 10. (a) Absorption cross-sections in ZBLAN:Er3+ near 800 nm: GSA 4 I15/2 → 4 I9/2 and ESA 4 I13/2 → 2 H11/2 , 4 I11/2 → 4 F3/2 , and 4 I11/2 → 4 F5/2 . (Data taken from [130].) (b) Partial energy-level scheme of erbium indicating the processes relevant to the ZBLAN:Er3+ cascade laser: Lower loop with GSA to 4 I9/2 , multiphonon relaxation, laser transition at 2.7 µm, luminescent decay, and upper loop with ESA to 2 H11/2 , thermal relaxation, laser transition at 1.7 µm, multiphonon relaxation, laser transition at 2.7 µm. Competitive lasing at 850 nm is suppressed in the cascade regime

Significant improvement in the performance of this laser system was obtained by deliberately operating a third laser transition 4 S3/2 → 4 I9/2 at 1.7 µm, thereby suppressing the competitive laser at 850 nm and recycling the excitation energy accumulated in the 4 S3/2 level into the upper laser level, see the energy-level scheme in Fig. 10b. The slope efficiency of the 2.7 µm transition increased significantly to 23 % [134], close to the Stokes-efficiency limit of 29 % under 800-nm pumping. An output power of 150 mW was demonstrated experimentally [134]. This cascade lasing regime represents the best option for dopant concentrations of typically 0.1 mol% (≈ 1.6×1019 cm−3 ) at which ESA is important. Also a three-transition-cascade lasing regime with additional lasing at the transition 4 I13/2 → 4 I15/2 near 1.6 µm was demonstrated [135], but no further improvement was obtained. 7.2

Lifetime Quenching by Pr3+ Co-Doping

In ZBLAN fibers with higher dopant concentrations of typically 1–5 mol% (≈ 1.6–8 × 1020 cm−3 ) and with the double-clad geometry, ESA is much less

240


important, because the reduced pump intensity with low-brightness diode lasers leads to smaller excitation densities. The relatively high fabrication costs of ZBLAN double-clad fibers means that pre-calculation and optimization of the expected device performance on the basis of the available spectroscopic parameters becomes an important tool. Currently, the most successful approach towards a high-power erbium 2.7 µm fiber laser is co-doping of the fiber with Pr3+ [73]. This idea was reported already in [131,136,137] and was proposed for the double-clad fiber laser in [138]. In this approach, the Er3+ 2.7 µm transition is operated as a simple four-level laser, see the energy-level scheme in Fig. 11a: The 4 I13/2 lower laser level is depopulated by the energy-transfer process ET1 to the Pr3+ co-dopant and fast decay to the ground state by multiphonon relaxation within Pr3+ . This energy-transfer process is much more efficient than the energy-transfer process ET2 from the 4 I11/2 upper laser level to the Pr3+ co-dopant, because the oscillator strength in Pr3+ is much higher for the transition involved in the former process [133]. The relatively weak lifetime quenching of the upper laser level affects the pump threshold, but it does not influence the slope efficiency. The strong lifetime quenching of the 4 I13/2 lower laser level and the fact that the 4 I11/2 population density is clamped to laser threshold significantly reduces ground-state bleaching and excitation of the laser levels, thus making the influence of ESA negligible, but similarly preventing energy recycling by ETU [139]. Like in the cascade-lasing regime (Sect. 5.2), each pump photon can at best produce one laser photon in the Er3+ , Pr3+ -co-doped system. The theoretical limit of the slope efficiency is given by the Stokes efficiency, which is 29 % under 800-nm pumping. Experimentally, a slope efficiency of 17 % and an output power of 1.7 W were obtained [73], see Fig. 11b. Other researchers [140] reported output powers of 660 mW. Since ESA from both laser levels is negligible, the system can alternatively be pumped near 980 nm, which provides the highest possible Stokes efficiency of 35 %. In this way, the experimental slope efficiency could be increased to 25 % [141]. The fiber that was used in these experiments contained an erbium concentration of 3.5 mol.%. 7.3

Energy Recycling by Energy-Transfer Upconversion

Energy-transfer processes between erbium ions govern the population mechanisms of highly erbium-doped laser systems. In the energy-level scheme of Fig. 12a, the important ETU and cross-relaxation processes are introduced. The ETU process (4 I13/2 , 4 I13/2 ) → (4 I15/2 , 4 I9/2 ) leads to a fast depletion of the lower laser level. Half of the ions that undergo this process are upconverted to the 4 I9/2 level and, by subsequent multiphonon relaxation, are recycled to the 4 I11/2 upper laser level from where they can each emit a second laser photon, for a single pump-photon absorption. For a large number of ions participating in this process, the quantum efficiency ηq = nl /np of

Mid-Infrared Fiber Lasers (a)

241

(b)

4

2

F7/2 H11/2 4 S3/2 1

D2

4

F9/2

4

I9/2

4

I11/2

ET2 1

G4

Output Power [W]

2

1

0 0

ET1 4

3

F4 F3 3 F2 3 H6

I13/2

2

4

6

8

10

12

Launched Pump Power [W]

3

Laser GSA

3

H5

3

4

H4

I15/2 Er

3+

3+

Pr

Fig. 11. (a) Partial energy-level scheme of erbium indicating the processes relevant to the ZBLAN:Er3+ lifetime-quenching laser: GSA at 980 nm to the 4 I11/2 upper laser level (or at 790 nm to the 4 I9/2 pump level and subsequent multiphonon relaxation to 4 I11/2 ), laser transition to the 4 I13/2 lower laser level, and relaxation to the ground state via energy transfer ET1 to the Pr3+ co-dopant. The energy transfer ET2 from the 4 I11/2 upper laser level to the Pr3+ co-dopant is weak. (b) Output power at 2.7 µm under 792-nm pumping (Data taken from [73])

pump photons np converted to laser photons nl increases from 1 to 2 [122]. In a simple rate-equation system which includes the processes shown in Fig. 12a the slope efficiency is given by [142] ln (1 − T ) b2 W22 ηsl = ηSt , (2) 2 − 12 ln [(1 − T ) (1 − L)] b2 W11 with T , the transmission of the outcoupling mirror, L, the internal resonator losses, and bi and Wii , the Boltzmann factors and ETU parameters of the upper (i = 2) and lower (i = 1) laser levels, respectively. If ETU occurs only from the lower laser level, i.e., W22 = 0, we obtain a factor-of-two increase in slope efficiency from (2). The slope efficiency is reduced, however, by the resonator losses, the nonperfect mode overlap, and the ETU process from the upper laser level in the case of W22 > 0. Energy recycling by ETU is the most efficient way to operate a CW erbium laser near 3 µm. The highest slope efficiency obtained experimentally is currently 50 % in LiYF4 :15 % Er3+ [143]. Quasi-CW excitation reduces the

242


(a)

(b)

F7/2 H11/2 4 S3/2 2

ETU2 4

F9/2

4

I9/2

4

I11/2

4

I13/2

CR

ETU1

W [10 -17 cm3s -1], Ratio

4

10 9 8 7 6 5 4 3 2 1 0

W11 W22 Ratio

0

ETU1

4

5

10

15

Erbium Conc. [10 20 cm-3]

ETU2

CR

I15/2 Er3+

Fig. 12. (a) Partial energy-level scheme of erbium indicating ETU1 from 4 I13/2 , ETU2 from 4 I11/2 , and cross-relaxation (CR) from the thermally coupled 4 S3/2 and 2 H11/2 levels. (b) Macroscopic parameters of ETU1 from 4 I13/2 (W11 ) and ETU2 from 4 I11/2 (W22 ) and the ratio W11 /W22 in ZBLAN:Er3+ bulk glasses. (Data taken from [133])

slope efficiency, because the lower laser level is less populated and ETU is less efficient in this case [144]. The parameters Wii of both ETU processes in ZBLAN bulk glasses [133] versus Er3+ concentration are shown in Fig. 12b. The criterion for optimization of the slope efficiency in (2) is maximizing the ratio W11 /W22 . For Er3+ concentrations of > 2–3 mol% at which ETU processes become important, this ratio is ≈ 3, see the dashed line in Fig. 12b, a more favorable value than reported for LiYF4 :Er3+ [145]. Energy recycling by ETU at high Er3+ concentrations [146] might lead to output powers at 3 µm on the order of 10 W. So far, two research groups tried to exploit energy recycling [147,148], however the slope efficiencies in these experiments did not exceed the slope efficiencies obtained in Er3+ ,Pr3+ -co-doped fibers pumped at corresponding pump wavelengths [73,141].

8

Fiber Lasers at Wavelengths Beyond 3 µm

Generating wavelengths longer than 3 µm from fiber lasers is a task which tests the limits of current glass technology. The need for lower phonon ener-


243

gies has to be balanced with acceptable mechanical, chemical, and thermal properties. Since the highly developed ZBLAN glass is only useful for laser transitions up to 3–3.5 µm, glasses such as the chalcogenides [149] will need to fill the gap. It is because these glasses have to be drawn into low-loss fiber which has prevented long-wavelength emission to the extent that is possible in crystalline-based solid-state lasers [150]. Creating efficient and high power mid-infrared fiber lasers with output wavelengths > 3 µm is at the forefront of current fiber-laser research efforts. 8.1

ZBLAN Fiber Lasers at 3.22 µm, 3.45 µm, and 3.95 µm

The operation of lasers at the longer wavelengths of 3.22 µm [75] and 3.95 µm [151] has been obtained from Ho3+ -doped ZBLAN fiber and at 3.45 µm [76] from Er3+ -doped ZBLAN fiber. It was, however, necessary to cool the ZBLAN fiber for the 3.45 µm and 3.95 µm transitions. These two laser transitions span five or six maximum phonon energies in ZBLAN, therefore the lifetime of the upper laser level for each of these transitions is short and engenders an increase in the pump threshold compared to other ZBLAN fiber lasers operating at the shorter mid-infrared wavelengths. In addition, the lower laser levels of these transitions possess quite long lifetimes and some saturation of the output power has been observed [77]. This problem (while it can be mitigated with cascaded lasing), combined with the use of inconvenient pump sources has impeded the full utilization of these laser transitions. The 3.95 µm wavelength emitted from the cooled ZBLAN fiber laser is currently the longest laser wavelength that has been generated from a fiber laser. A laser transition that has as yet not been realized but might work well in ZBLAN fibers is the 6 H13/2 → 6 H15/2 transition at 3.0–3.4 µm in Dy3+ . 8.2

Future Mid-Infrared Fiber Lasers

As mentioned above, fiber lasers operating on laser transitions which have wavelengths > 3 µm will need to use glasses which have very low phonon energies. While rare-earth-doped heavy-metal oxides [152] have been studied for 2–3 µm mid-infrared emission, up to now, there has been no report of laser action for a fiber laser comprised of such a glass. Heavy-metal oxides do not seem to be suitable for lasers at wavelengths beyond 3 µm, because their maximum phonon energies are comparable to fluoride glasses and are too high for laser transitions beyond 3 µm. The chalcogenide glasses have, by contrast, been doped with a number of rare-earth ions including Ho3+ [153], Tm3+ [154], Tb3+ [154], Dy3+ [155], Pr3+ [156], and Er3+ [157,158] for studies into > 3-µm mid-infrared luminescence, see Table 3. Fiber-laser action has been reported, however, only for an Nd3+ -doped GLS glass operating at a wavelength of ≈ 1 µm [159]. Recent demonstrations of fabricating Bragg gratings [160], single-mode fiber [161] and holey fiber [162] with chalcogenide glass highlight the utility of this glass

244


Table 3. Examples of luminescent transitions investigated as candidates for midinfrared lasers in sulfide fibers Ion 3+

Dy Tm3+ Ho3+ Dy3+ Tb3+ Ho3+

λLaser ( µm) 3.2 3.8 3.9 4.3 4.8 4.9

Transition

Ref.

H13/2 → H15/2 3 H5 → 3 H4 5 I5 → 5 H6 6 H11/2 → 6 H13/2 7 F5 → 7 F6 5 I4 → 5 I5

[155] [154] [153] [155] [154] [153]

6

6

for fiber-based applications; however, the purity and toxicity of the starting materials and the difficulty of making ultra-low loss fiber currently impede the widespread use of chalcogenide glass for mid-infrared fiber-laser applications. Once these obstacles have been overcome, future > 3-µm fiber lasers will most likely involve the rare-earth ions Pr3+ , Nd3+ , Dy3+ , and Ho3+ doped into chalcogenide glass because most of the important mid-infrared transitions relevant to these ions can be accessed with pump-photon wavenumbers < 10 000 cm−1 . Judicious choice of the overall dopant-ion concentration and the use of particular sensitizer and quenching ions will enable the production of efficient > 3 µm output some time in the future.

9

Conclusions

On our journey through the forest of the various established and yet to be demonstrated mid-infrared fiber lasers, we have found that the shorter the wavelength, the better is the laser performance. When we approach longer wavelengths in the mid-infrared spectrum, we find that the quality and durability of the required low-phonon-energy fibers decline, Stokes and slope efficiencies decrease, whereas the thermal problems increase. The assumption that due to its large surface-to-volume ratio, the fiber geometry might avoid all thermal problems has been questioned by several recent high-power fiberlaser experiments in the near- and mid-infrared spectral region. All the above phenomena are not much different from the situation found in crystalline lasers. Nevertheless, there remain distinct differences between these two host categories. When flexibility of the resonator design, short pulses, and high peak powers are required, the fiber shows some disadvantages. On the other hand, fiber lasers are preferred when high beam quality or low pump threshold combined with medium CW output power are desired. The low pump threshold is an invaluable advantage when cascade-laser operation is required to depopulate the long-lived terminating level of one laser transition by a second laser transition. The comparatively low dopant concentrations that are useful in fiber lasers due to the long interaction lengths can minimize energy dissipation by interionic processes but, equally true, limit the exploitation of


245

these processes as a tool to optimize the population mechanisms of a certain laser system. Although still a great challenge with respect to the fabrication process, chalcogenide-glass fibers have the potential to revolutionize CW midinfrared lasers in the wavelength range between 3–5 µm.

References 1. A. E. Siegman: Lasers (University Science Books, Mill Valley, CA 1986) 220 2. O. Svelto: Principles of Lasers, 3rd ed. (Plenum Press, New York 1989) 220 3. M. J. F. Digonnet (Ed.): Rare-Earth-Doped Fiber Lasers and Amplifiers, 2nd ed. (Marcel Dekker, New York 2001) 220, 221, 230 4. R. Reisfeld, M. Eyal: Possible ways of relaxations for excited states of rare earth ions in amorphous media, J. Phys. (Paris) 46, C349 (1985) 221 5. T. Schweizer: Rare-earth-doped gallium lanthanum sulphide glasses for midinfrared fibre lasers, Dissertation, University of Hamburg, Germany (1998) 221 6. D. C. Brown, H. J. Hoffman: Thermal, stress, and thermo-optic effects in high average power double-clad silica fiber lasers, IEEE J. Quantum Electron. 37, 207 (2001) 221, 226 7. L. Wetenkamp: Charakterisierung von laseraktiv dotierten Schwermetallfluorid-Gl¨ asern und Faserlasern, Dissertation, Technical University of Braunschweig, Germany (1991) 221, 222 8. S. M. Lima, T. Catunda, R. Lebullenger, A. C. Hernandes, M. L. Baesso, A. C. Bento, L. C. M. Miranda: Temperature dependence of thermo-optical properties of fluoride glasses determined by thermal lens spectrometry, Phys. Rev. B 60, 15173 (1999) 221 9. S. M. Lima, T. Catunda, M. L. Baesso, L. D. Vila, Y. Messaddeq, E. B. Stucchi, S. J. L. Ribeiro: Thermal-optical properties of Ga:La:S glasses measured by thermal lens technique, J. Non-Cryst. Solids 247, 222 (1999) 221 10. B. J. Ainslie, S. P. Craig, S. T. Davey: The absorption and fluorescence spectra of rare earth ions in silica-based monomode fiber, J. Lightwave Technol. 6, 287 (1988) 221 11. S. D. Jackson, T. A. King: CW operation of a 1.064 µm pumped Tm-Ho-doped silica fiber laser, IEEE J. Quantum Electron. 34, 1578 (1998) 221, 236 12. O. Humbach, H. Fabian, U. Grzesik, U. Haken, W. Heitmann: Analysis of OH absorption bands in synthetic silica, J. Non-Cryst. Solids 203, 19 (1996) 222 13. D. C. Tran, G. H. Sigel, Jr., B. Bendow: Heavy metal fluoride glasses and fibers: a review, J. Lightwave Technol. 2, 566 (1984) 222 14. P. W. France, M. G. Drexhage, J. M. Parker, M. W. Moore, S. F. Carter, J. V. Wright: Fluoride Glass Optical Fibres (Blackie, Glasgow 1990) 222 15. S. T. Davey, P. W. France: Rare earth doped fluorozirconate glasses for fibre devices, BT Technol. J. 7, 58 (1989) 222 16. M. Monerie, F. Alard, G. Maze: Fabrication and characterisation of fluorideglass single-mode fibres, Electron. Lett. 21, 1179 (1985) 222 17. L. Wetenkamp, G. F. West, H. T¨ obben: Optical properties of rare earth-doped ZBLAN glasses, J. Non-Cryst. Solids 140, 35 (1992) 222 18. Y. D. Huang, M. Mortier, F. Auzel: Stark level analysis for Er3+ -doped ZBLAN glass, Opt. Mater. 17, 501 (2001) 222

246


19. P. N. Kumta, S. H. Risbud: Rare-earth chalcogenides – an emerging class of optical materials, J. Mat. Sci. 29, 1135 (1994) 222 20. J. S. Sanghera, J. Heo, J. D. Mackenzie: Chalcohalide glasses, J. Non-Cryst. Solids 103, 155 (1988) 222 21. L. D. DeLoach, R. H. Page, G. D. Wilke, S. A. Payne, W. F. Krupke: Transition metal-doped zinc chalcogenides: Spectroscopy and laser demonstration of a new class of gain media, IEEE J. Quantum Electron. 32, 885 (1996) 222 22. L. B. Shaw, B. Cole, P. A. Thielen, J. S. Sanghera, I. D. Aggarwal: Mid-wave IR and long-wave IR laser potential of rare-earth doped chalcogenide glass fiber, IEEE J. Quantum Electron. 48, 1127 (2001) 222 23. Y. D. West, T. Schweizer, D. J. Brady, D. W. Hewak: Gallium lanthanum sulphide fibers for infrared transmission, Fiber Integrated Opt. 19, 229 (2000) 222 24. J. Heo, Y. B. Shin: Absorption and mid-infrared emission spectroscopy of Dy3+ in Ge-As (or Ga)-S glasses, J. Non-Cryst. Solids 196, 162 (1996) 222 25. J. R. Lu, T. Murai, K. Takaichi, T. Uematsu, K. Misawa, M. Prabhu, J. Xu, K. Ueda, H. Yagi, T. Yanagitani, A. A. Kaminskii, A. Kudryashov: 72 W Nd:Y3 Al5 O12 ceramic laser, Appl. Phys. Lett. 78, 3586 (2001) 222 26. J. Lu, J. Lu, T. Murai, K. Takaichi, T. Uematsu, K. Ueda, H. Yagi, T. Yanagitani, T. Works, A. A. Kaminskii: Optical properties and highly efficient laser oscillation of Nd:Y2 O3 ceramics, in CLEO Europe, OSA Tech. Digest (Opt. Soc. Am., Washington, DC 2002), p. 60 223 27. G. Huber, V. Peters, K. Petermann: New solid-state lasers based on rare-earth doped sesquioxide crystals, in CLEO Europe 2000, OSA Tech. Digest (Opt. Soc. Am., Washington, DC 2000) p. 386 223 28. B. N. Samson, P. A. Tick, N. F. Borrelli: Efficient neodymium-doped glassceramic fiber laser and amplifier, Opt. Lett. 26, 145 (2001) 223 29. H. Po, E. Snitzer, R. Tumminelli, L. Zenteno, F. Hakimi, N. M. Cho, T. Haw: Doubly clad high brightness Nd fiber laser pump by GaAlAs phased array, in Proc. Optical Fibre Communication Conf. (1989) paper PD7 223 30. H. Po, J. D. Cao, B. M. Laliberte, R. A. Minns, R. F. Robinson, B. H. Rockney, R. R. Tricca, Y. H. Zhang: High power neodymium-doped single transverse mode fibre laser, Electron. Lett. 29, 1500 (1993) 223 31. I. N. Duling III, W. K. Burns, L. Goldberg: High-power superfluorescent fiber source, Opt. Lett. 15, 33 (1990) 223 32. J. D. Minelly, W. L. Barnes, R. I. Laming, P. R. Morkel, J. E. Townsend, S. G. Grubb, D. N. Payne: Diode-array pumping of Er3+ /Yb3+ co-doped fiber lasers and amplifiers, IEEE Photonics Technol. Lett. 5, 301 (1993) 223 33. H. M. Pask, J. L. Archambault, D. C. Hanna, L. Reekie, P. S. J. Russell, J. E. Townsend, A. C. Tropper: Operation of cladding-pumped Yb3+ -doped silica fibre lasers in 1 µm region, Electron. Lett. 30, 863 (1994) 223 34. A. Liu, K. Ueda: The absorption characteristics of circular, offset, and rectangular double-clad fibers, Opt. Commun. 132, 511 (1996) 224 35. H. Zellmer, U. Willamowski, A. T¨ unnermann, H. Welling, S. Unger, V. Reichel, H.-R. M¨ uller, T. Kirchhof, P. Albers: High-power CW neodymiumdoped fiber laser operating at 9.2 W with high beam quality, Opt. Lett. 20, 578 (1995) 224 36. A. Ankiewicz, W. L. R. Perera, P. L. Chu, D. A. C. C. Jayasinghe: Use of graded elliptical cladding for efficient pumping of doped fiber cores, Opt. Commun. 133, 82 (1997) 224


247

37. D. Kouznetsov, J. V. Moloney, E. M. Wright: Efficiency of pump absorption in double-clad fiber amplifiers. I. Fiber with circular symmetry, J. Opt. Soc. Am. B 18, 743 (2001) 224 38. M. H. Muendel: Optimal inner cladding shapes for double-clad fiber lasers, in CLEO 9, OSA Tech. Digest Ser. (Opt. Soc. Am., Washington, DC 1996), p. 209 224 39. A. Liu, K. Ueda: Propagation losses of pump light in rectangular double-clad fibers, Opt. Eng. 35, 3130 (1996) 224 40. V. Doya, O. Legrand, F. Mortessagne: Optimized absorption in a chaotic double-clad fiber amplifier, Opt. Lett. 26, 872 (2001) 224 41. P. Glas, M. Naumann, A. Schirrmacher, Th. Pertsch: The multicore fiber – a novel design for a diode pumped fiber laser, Opt. Commun. 151, 187 (1998) 224 42. P. Glas, M. Naumann, A. Schirrmacher: A novel design for a high brightness diode pumped fiber laser source, Opt. Commun. 122, 163 (1996) 224 43. G. P. Lees, D. Taverner, D. J. Richardson, L. Dong, T. P. Newson: Q-switched erbium doped fibre laser utilising a novel large mode area fibre, Electron. Lett. 33, 393 (1997) 224 44. K. Furusawa, A. Malinowski, J. H. V. Price, T. M. Monro, J. K. Sahu, J. Nilsson, D. J. Richardson: Cladding pumped ytterbium-doped fiber laser with holey inner and outer cladding, Opt. Express 9, 714 (2001) 224 45. T. A. Birks, J. C. Knight, P. St. J. Russell: Endlessly single-mode photonic crystal fiber, Opt. Lett. 22, 961–963 (1997) 224 46. D. J. Ripin, L. Goldberg: High efficiency side-coupling of light into optical fibres using imbedded v-grooves, Electron. Lett. 31, 2204 (1995) 225 47. T. Weber, W. L¨ uthy, H. P. Weber, V. Neuman, H. Berthou, G. Kotrotsios: A longitudinal and side-pumped single transverse mode double-clad fiber laser with a special silicone coating, Opt. Commun. 115, 99 (1995) 225 48. C. C. Cook, T. Y. Fan: Spectral beam combining of Yb-doped fiber lasers in and external cavity, in Trends Opt. Photonics Adv. Solid-State Lasers 26, 163–166 (Opt. Soc. Am., Washington, DC 1999) 225 49. W. A. Clarkson, V. Matera, T. M. J. Kendall, D. C. Hanna, J. Nilsson, P. W. Turner: High-power wavelength-combined cladding-pumped Tm-doped silica fibre lasers, in Conference on Lasers and Electro-Optics, OSA Technical Digest, Vol. 56 (Optical Society of America, Washington, DC 2001) pp. 363– 364 225 50. S. J. Augst, A. K. Goyal, R. L. Aggarwal, T. Y. Fan, A. Sanchez: Wavelength beam combining of ytterbium fiber lasers in a MOPA configuration, in CLEO, OSA Tech. Digest (Opt. Soc. Am., Washington, DC 2002) pp. 594–595 225 51. V. Dominic, S. MacCormack, R. Waarts, S. Sanders, S. Bicknese, R. Dohle, E. Wolak, P. S. Yeh, E. Zucker: 110 W fibre laser, Electron. Lett. 35, 1158 (1999) 225 52. J. Limpert, A. Liem, S. H¨ ofer, H. Zellmer, A. T¨ unnermann, S. Unger, S. Jetschke, H.-R. M¨ uller: 150 W Nd/Yb codoped fiber laser at 1.1 µm, in CLEO, OSA Tech. Digest (Opt. Soc. Am., Washington, DC 2002), pp. 590– 591 225 53. M. Pollnau, P. J. Hardman, M. A. Kern, W. A. Clarkson, D. C. Hanna: Upconversion-induced heat generation and thermal lensing in Nd:YLF and Nd:YAG, Phys. Rev. B 58, 16076 (1998) 226

248


54. M. K. Davis, M. J. F. Digonnet, R. H. Pantell: Thermal effects in doped fibers, J. Lightwave Technol. 16, 1013 (1998) 226 55. L. Zenteno: High-power double-clad fiber lasers, J. Lightwave Technol. 11, 1435 (1993) 226 56. N. A. Brilliant, K. Lagonik: Thermal effects in a dual-clad ytterbium fiber laser, Opt. Lett. 26, 1669 (2001) 226 57. B. R. Judd: Optical absorption intensities of rare-earth ions, Phys. Rev. 127, 750 (1962) 228 58. G. S. Ofelt: Intensities of crystal spectra of rare-earth ions, J. Chem. Phys. 37, 511 (1962) 228 59. L. A. Riseberg, H. W. Moos: Multiphonon orbit–lattice relaxation of excited states of rare-earth ions in crystals, Phys. Rev. 174, 429 (1968) 230 60. J. M. F. van Dijk, M. F. H. Schuurmans: On the nonradiative and radiative decay rates and a modified exponential energy gap law for 4f –4f transitions in rare-earth ions, J. Chem. Phys. 78, 5317 (1983) 230 61. D. L. Dexter: A theory of sensitized luminescence in solids, J. Chem. Phys. 21, 836 (1953) 230 62. T. Miyakawa, D. L. Dexter: Phonon sidebands, multiphonon relaxation of excited states, and phonon-assisted energy transfer between ions in solids, Phys. Rev. B 1, 2961 (1970) 230 63. V. P. Gapontsev, N. S. Platonov: Migration-accelerated quenching of luminescence in glasses activated by rare-earth ions, Mater. Sci. Forum 50, 165 (1989) 231 64. W. J. C. Grant: Role of rate equations in the theory of luminescent energy transfer, Phys. Rev. B 4, 648 (1971) 232 65. E. Delevaque, T. Georges, M. Monerie, P. Lamouler, J.-F. Bayon: Modeling of pair-induced quenching in erbium-doped silicate fibers, IEEE Photonics Technol. Lett. 5, 73 (1993) 232 66. J. L. Wagener, P. F. Wysocki, M. J. F. Digonnet, H. J. Shaw, D. J. DiGiovanni: Effects of concentration and clusters in erbium-doped fiber lasers, Opt. Lett. 18, 2014 (1993) 232 67. R. S. Quimby, W. J. Miniscalco, B. Thompson: Clustering in erbium-doped silica glass fibers analyzed using 980 nm excited-state absorption, J. Appl. Phys. 76, 4472 (1994) 232 68. E. Maurice, G. Monnom, B. Dussardier, D. B. Ostrowsky: Clustering-induced nonsaturable absorption phenomenon in heavily erbium-doped silica fibers, Opt. Lett. 20, 2487 (1995) 232 69. M. Pollnau, D. R. Gamelin, S. R. L¨ uthi, H. U. G¨ udel, M. P. Hehlen: Power dependence of upconversion luminescence in lanthanide and transition-metalion systems, Phys. Rev. B 61, 3337 (2000) 232 70. R. A. Hayward, W. A. Clarkson, P. W. Turner, J. Nilsson, A. B. Grudinin, D. C. Hanna: Efficient cladding-pumped Tm-doped silica fibre laser with high power singlemode output at 2 µm, Electron. Lett. 36, 711 (2000) 232, 234, 235 71. S. D. Jackson: 8.8 W diode-cladding-pumped Tm3+ , Ho3+ -doped fluoride fibre laser, Electron. Lett. 37, 821 (2001) 232, 235, 236 72. R. G. Smart, J. N. Carter, A. C. Tropper, D. C. Hanna: Continuous-wave oscillation of Tm3+ -doped fluorozirconate fibre lasers at around 1.47 µm, 1.9 µm and 2.3 µm when pumped at 790 nm, Opt. Commun. 82, 563 (1991) 232, 234


249

73. S. D. Jackson, T. A. King, M. Pollnau: Diode-pumped 1.7-W erbium 3 µm fiber laser, Opt. Lett. 24, 1133 (1999) 232, 240, 241, 242 74. T. Sumiyoshi, H. Sekita, T. Arai, S. Sato, M. Ishihara, M. Kikuchi: Highpower continuous-wave 3- and 2 µm cascade Ho3+ :ZBLAN fiber laser and its medical applications, IEEE J. Sel. Topics Quantum Electron. 5, 936 (1999) 232, 237 75. C. Carbonnier, H. T¨ obben, U. B. Unrau: Room temperature CW fibre laser at 3.22 µm, Electron. Lett. 34, 893 (1998) 232, 243 76. H. T¨ obben: CW-lasing at 3.45 µm in erbium-doped fluorozirconate fibers, Frequenz 45, 250 (1991) 232, 243 77. J. Schneider, C. Carbonnier, U. B. Unrau: Characterization of a Ho3+ -doped fluoride fiber laser with a 3.9 µm emission wavelength, Appl. Opt. 36, 8595 (1997) 232, 243 78. R. M. Percival, S. F. Carter, D. Szebesta, S. T. Davey, W. A. Stallard: Thuliumdoped monomode fluoride fibre laser broadly tunable from 2.25 to 2.5 µm, Electron. Lett. 27, 1912 (1991) 233, 235 79. D. C. Hanna, I. M. Jauncey, R. M. Percival, I. R. Perry, R. G. Smart, P. J. Suni, J. E. Townsend, A. C. Tropper: Continuous-wave oscillation of a monomode thulium-doped fibre laser, Electron. Lett. 24, 1222 (1988) 233 80. W. L. Barnes, J. E. Townsend: Highly tunable and efficient diode pumped operation of Tm3+ doped fibre lasers, Electron. Lett. 26, 746 (1990) 233 81. J. N. Carter, R. G. Smart, D. C. Hanna, A. C. Tropper: CW diode-pumped operation of 1.97 µm thulium-doped fluorozirconate fibre laser, Electron. Lett. 26, 599 (1990) 233 82. T. Y. Fan, G. Huber, R. L. Byer, P. Mitzscherlich: Spectroscopy and diode laser-pumped operation of Tm,Ho:YAG, IEEE J. Quantum Electron. 24, 924 (1988) 233, 235 83. R. C. Stoneman, L. Esterowitz: Efficient, broadly tunable, laser-pumped Tm:YAG and Tm:YSGG CW lasers, Opt. Lett. 15, 486 (1990) 233 84. D. C. Hanna, R. M. Percival, R. G. Smart, A. C. Tropper: Efficient and tunable operation of a Tm-doped fibre laser, Opt. Commun. 75, 283 (1990) 234 85. L. E. Nelson, E. P. Ippen, H. A. Haus: Broadly tunable sub-500 fs pulses from an additive-pulse mode-locked thulium-doped fiber ring laser, Appl. Phys. Lett. 67, 19 (1995) 234 86. R. C. Sharp, D. E. Spock, N. Pan, J. Elliot: 190-fs passively mode-locked thulium fiber laser with a low threshold, Opt. Lett. 21, 881 (1996) 234, 235 87. S. D. Jackson, T. A. King: High-power diode-cladding-pumped Tm-doped silica fiber laser, Opt. Lett. 23, 1462 (1998) 234 88. D. C. Hanna, I. R. Perry, J. R. Lincoln, J. E. Townsend: A 1-Watt thuliumdoped CW fibre laser operating at 2 µm, Opt. Commun. 80, 52 (1990) 234 89. P. S. Golding, S. D. Jackson, P.-K. Tsai, B. C. Dickinson, T. A. King: Efficient high power operation of a Tm-doped silica fiber laser pumped at 1.319 µm, Opt. Commun. 175, 179 (2000) 234 90. T. Yamamoto, Y. Miyajima, T. Komukai: 1.9 µm Tm-doped silica fibre laser pumped at 1.57 µm, Electron. Lett. 30, 220 (1994) 234 91. R. M. Percival, D. Szebesta, C. P. Seltzer, S. D. Perrin, S. T. Davey, M. Louka: 1.6 µm semiconductor diode pumped thulium doped fluoride fibre laser and amplifier of very high efficiency, Electron. Lett. 29, 2110 (1993) 234

250


92. T. Yamamoto, Y. Miyajima, T. Komukai, T. Sugawa: 1.9 µm Tm-doped fluoride fiber amplifier and laser pumped at 1.58 µm, Electron. Lett. 29, 986 (1993) 234 93. S. D. Jackson, T. A. King: Theoretical modeling of Tm-doped silica fiber lasers, J. Lightwave Technol. 17, 948 (1999) 234 94. J. Q. Xu, M. Prabhu, L. Jianren, K. Ueda, D. Xing: Efficient double-clad thulium-doped fiber laser with a ring cavity, Appl. Opt. 40, 1983 (2001) 234 95. K. S. Lai, P. B. Phua, R. F. Wu, Y. L. Lim, E. Lau, S. W. Toh, B. T. Toh, A. Chng: 120-W continuous-wave diode-pumped Tm:YAG laser, Opt. Lett. 25, 1591 (2000) 234 96. R. Allen, L. Esterowitz: CW diode pumped 2.3 µm fiber laser, Appl. Phys. Lett. 55, 721 (1989) 234 97. J. Y. Allain, M. Monerie, H. Poignant: Tunable CW lasing around 0.82, 1.48, 1.88 and 2.35 µm in thulium-doped fluorozirconate fibre, Electron. Lett. 25, 1660 (1989) 234 98. R. M. Percival, D. Szebesta, S. T. Davey: Highly efficient and tunable operation of 2 color Tm-doped fluoride fiber laser, Electron. Lett. 28, 671 (1992) 235 99. K. Tanimura, M. D. Shinn, W. A. Sibley, M. G. Drexhage, R. N. Brown: Optical transitions of Ho3+ ions in fluorozirconate glass, Phys. Rev. B 30, 2429 (1984) 235 100. J. K. Tyminski, D. M. Franich, M. Kokta: Gain dynamics of Tm:Ho:YAG pumped in near infrared, J. Appl. Phys. 65, 3181 (1989) 235 101. V. A. French, R. R. Petrin, R. C. Powell, M. Kokta: Energy-transfer processes in Y3 Al5 O12 :Tm,Ho, Phys. Rev. B 46, 8018 (1992) 235 102. L. B. Shaw, R. S. F. Chang, N. Djeu: Measurement of up-conversion energytransfer probabilities in Ho:Y3 Al5 O12 and Tm:Y3 Al5 O12 , Phys. Rev. B 50, 6609 (1994) 235 103. M. C. Brierley, P. W. France, C. A. Millar: Lasing at 2.08 µm and 1.38 µm in a holmium doped fluorozirconate fiber laser, Electron. Lett. 24, 539 (1988) 236 104. D. C. Hanna, R. M. Percival, R. G. Smart, J. E. Townsend, A. C. Tropper: Continuous-wave oscillation of holmium-doped silica fiber laser, Electron. Lett. 25, 593 (1989) 236 105. A. S. Kurkov, E. M. Dianov, O. I. Medvedkov, G. A. Ivanov, V. A. Aksenov, V. M. Paramonov, S. A. Vasiliev, E. V. Pershina: Efficient silica-based Ho3+ fibre laser for 2 µm spectral region pumped at 1.15 µm, Electron. Lett. 36, 1015 (2000) 236 106. J. Y. Allain, M. Monerie, H. Poignant: High-efficiency CW thulium-sensitized holmium-doped fluoride fiber laser operating at 2.04 µm, Electron. Lett. 27, 1513 (1991) 236 107. R. M. Percival, D. Szebesta, S. T. Davey, N. A. Swain, T. A. King: Thulium sensitized holmium-doped CW fluoride fiber laser of high-efficiency, Electron. Lett. 28, 2231 (1992) 236 108. K. Oh, T. F. Morse, A. Kilian, L. Reinhart, P. M. Weber: Continuous-wave oscillation of thulium-sensitized holmium-doped silica fiber laser, Opt. Lett. 19, 278 (1994) 236 109. Ch. Ghisler, W. L¨ uthy, H. P. Weber, J. Morel, A. Woodtli, R. D¨ andliker, V. Neuman, H. Berthou, G. Kotrotsios: A Tm3+ sensitized Ho3+ silica fibre laser at 2.04 µm pumped at 809 nm, Opt. Commun. 109, 279 (1994) 236


251

110. C. Bollig, R. A. Hayward, W. A. Clarkson, D. C. Hanna: 2-W Ho:YAG laser intracavity pumped by a diode-pumped Tm:YAG laser, Opt. Lett. 23, 1757 (1998) 237 111. P. A. Budni, M. L. Lemons, J. R. Mosto, E. P. Chicklis: High-power/highbrightness diode-pumped 1.9 µm thulium and resonantly pumped 2.1 µm holmium lasers, IEEE J. Select. Topics Quantum Electron. 6, 629 (2000) 237 112. G. Rustad, K. Stenersen: Modeling of laser-pumped Tm and Ho lasers accounting for upconversion and ground-state depletion, IEEE J. Quantum Electron. 32, 1645 (1996) 237 113. L. Wetenkamp: Efficient CW operation of a 2.9 µm Ho3+ -doped fluorozirconate fiber laser pumped at 640 nm, Electron. Lett. 26, 883 (1990) 237 114. T. Sumiyoshi, H. Sekita: Dual-wavelength continuous-wave cascade oscillation at 3 and 2 µm with a holmium-doped fluoride-glass fiber laser, Opt. Lett. 23, 1837 (1998) 237 115. T. Rothacher, W. L¨ uthy, H. P. Weber: Diode pumping and laser properties of Yb,Ho:YAG, Opt. Commun. 155, 68 (1998) 237 116. A. Diening, P. E.-A. M¨ obert, E. Heumann, G. Huber, B. H. T. Chai: Diodepumped CW lasing of Yb,Ho:KYF4 in the 3 µm spectral range in comparison to Er:KYF4 , Laser Phys. 8, 214 (1998) 237 117. B. Peng, T. Izumitani: Ho3+ doped 2.84 µm laser glass for laser knives, sensitised by Yb3+ , Rev. Laser Eng. 22, 9 (1994) 237 118. M. Robinson, P. D. Devor: Thermal switching of laser emission of Er3+ at 2.69µ and Tm3+ at 1.86µ in mixed crystals of CaF2 :ErF3 :TmF∗3 , Appl. Phys. Lett. 10, 167 (1967) 238 119. K. S. Bagdasarov, V. I. Zhekov, V. A. Lobachev, T. M. Murina, A. M. Prokhorov: Steady-state emission from a Y3 Al5 O12 :Er3+ laser (λ = 2.94 µm, T = 300 K), Kvant. Elektr. (Moscow) 10, 452 (1983) [Engl. transl. Sov. J. Quantum Electron. 13, 262 (1983)] 238 120. M. C. Brierley, P. W. France: Continuous wave lasing at 2.71 µm in an erbiumdoped fluorozirconate fibre, Electron. Lett. 24, 935 (1988) 238 121. R. S. Quimby, W. J. Miniscalco: Continuous-wave lasing on a self-terminating transition, Appl. Opt. 28, 14 (1989) 238 122. M. Pollnau, S. D. Jackson: Erbium 3 µm fiber lasers, IEEE J. Select. Topics Quantum Electron. 7, 30 (2001) 238, 241 123. V. Lupei, S. Georgescu, V. Florea: On the dynamics of population inversion for 3 µm Er3+ lasers, IEEE J. Quantum Electron. 29, 426 (1993) 238 124. J. Schneider: Kaskaden-Faserlaser im mittleren Infrarot, Dissertation, Technical University Braunschweig (Cuvillier, G¨ ottingen 1996) 238 125. B. C. Dickinson, P. S. Golding, M. Pollnau, T. A. King, S. D. Jackson: Investigation of a 791-nm pulsed-pumped 2.7 µm Er-doped ZBLAN fibre laser, Opt. Commun. 191, 315 (2001) 238 126. N. J. C. Libatique, J. Tafoja, N. K. Viswanathan, R. K. Jain, A. Cable: ‘Fieldusable’ diode-pumped ≈ 120 nm wavelength-tunable CW mid-IR fibre laser, Electron. Lett. 36, 791 (2000) 238 127. R. C. Stoneman, J. G. Lynn, L. Esterowitz: Direct upper-state pumping of the 2.8 µm Er3+ :YLF laser, IEEE J. Quantum Electron. 28, 1041 (1992) 238 128. R. S. Quimby, W. J. Miniscalco, B. Thompson: Excited state absorption at 980 nm in erbium doped glass, in Fiber Laser Sources and Amplifiers III, Proc. SPIE 1581, 72–79 (1991) 238

252


129. S. Bed¨ o, M. Pollnau, W. L¨ uthy, H. P. Weber: Saturation of the 2.71 µm laser output in erbium doped ZBLAN fibers, Opt. Commun. 116, 81 (1995) 238 130. M. Pollnau, Ch. Ghisler, W. L¨ uthy, H. P. Weber: Cross-sections of excitedstate absorption at 800 nm in erbium-doped ZBLAN fiber, Appl. Phys. B 67, 23 (1998) 238, 239 131. J. Schneider, D. Hauschild, C. Frerichs, L. Wetenkamp: Highly efficient Er3+ :Pr3+ -Co-doped CW fluorozirconate fiber laser operating at 2.7 µm, Int. J. Infrared Millimeter Waves 15, 1907 (1994) 238, 240 132. J. Schneider: Mid-infrared fluoride fiber lasers in multiple cascade operation, IEEE Photonics Technol. Lett. 7, 354 (1995) 238 133. P. S. Golding, S. D. Jackson, T. A. King, M. Pollnau: Energy-transfer processes in Er3+ -doped and Er3+ ,Pr3+ -codoped ZBLAN glasses, Phys. Rev. B 62, 856 (2000) 238, 240, 242 134. M. Pollnau, Ch. Ghisler, G. Bunea, M. Bunea, W. L¨ uthy, H. P. Weber: 150 mW unsaturated output power at 3 µm from a single-mode-fiber erbium cascade laser, Appl. Phys. Lett. 66, 3564 (1995) 239 135. M. Pollnau, Ch. Ghisler, W. L¨ uthy, H. P. Weber, J. Schneider, U. B. Unrau: Three-transition cascade erbium laser at 1.7, 2.7, and 1.6 µm, Opt. Lett. 22, 612 (1997) 239 136. J. Y. Allain, M. Monerie, H. Poignant: Energy transfer in Er3+ /Pr3+ -doped fluoride glass fibres and application to lasing at 2.7 µm, Electron. Lett. 27, 445 (1991) 240 137. L. Wetenkamp, G. F. West, H. T¨ obben: Co-doping effects in erbium3+ - and 3+ holmium -doped ZBLAN glass, J. Non-Crystal. Solids 140, 25 (1992) 240 138. M. Pollnau: The route toward a diode-pumped 1-W erbium 3 µm fiber laser, IEEE J. Quantum Electron. 33, 1982 (1997) 240 139. S. D. Jackson, T. A. King, M. Pollnau: Modelling of high-power diode-pumped erbium 3 µm fibre lasers, J. Mod. Opt. 47, 1987 (2000) 240 140. B. Srinivasan, J. Tafoya, R. K. Jain: High-power ‘watt-level’ CW operation of diode-pumped 2.7 µm fiber lasers using efficient cross-relaxation and energy transfer mechanisms, Opt. Express 4, 490 (1999) 240 141. S. D. Jackson, T. A. King, M. Pollnau: Efficient high power operation of erbium 3 µm fibre laser diode-pumped at 975 nm, Electron. Lett. 36, 223 (2000) 240, 242 142. M. Pollnau, R. Spring, Ch. Ghisler, S. Wittwer, W. L¨ uthy, H. P. Weber: Efficiency of erbium 3 µm crystal and fiber lasers, IEEE J. Quantum Electron. 32, 657 (1996) 241 143. Ch. Wyss, W. L¨ uthy, H. P. Weber, P. Rogin, J. Hulliger: Emission properties of an optimised 2.8 µm Er3+ :YLF laser, Opt. Commun. 139, 215 (1997) 241 144. M. Pollnau, R. Spring, S. Wittwer, W. L¨ uthy, H. P. Weber: Investigations on the slope efficiency of a pulsed 2.8 µm Er3+ :LiYF4 laser, J. Opt. Soc. Am. B 14, 974 (1997) 242 145. T. Jensen: Upconversion-Prozesse und Wirkungsquerschnitte in Er3+ -dotierten 3 µm Fluorid- und Granat-Lasern, gepumpt mit CW und quasi-CW Dioden-Arrays, Dissertation, University of Hamburg, Germany (1996) 242 146. M. Pollnau, S. D. Jackson: Energy recycling versus lifetime quenching in erbium-doped 3 µm fiber lasers, IEEE J. Quantum Electron. 38, 162 (2002) 242


253

147. B. Srinivasan, E. Poppe, J. Tafoya, R. K. Jain: High-power (400 mW) diodepumped 2.7 µm Er:ZBLAN fibre lasers using enhanced Er-Er cross-relaxation processes, Electron. Lett. 35, 1338 (1999) 242 148. T. Sandrock, D. Fischer, P. Glas, M. Leitner, M. Wrage, A. Diening: Diodepumped 1-W Er-doped fluoride glass M-profile fiber laser emitting at 2.8 µm, Opt. Lett. 24, 1284 (1999) 242 149. R. Reisfeld: Chalcogenide glasses doped by rare earths: structure and optical properties, Ann. Chim. Fr. 7, 147 (1982) 243 150. S. R. Bowman, L. B. Shaw, B. J. Feldman, J. Ganem: A 7 µm praseodymiumbased solid-state laser, IEEE J. Quantum Electron. 32, 646 (1996) 243 151. J. Schneider: Fluoride fibre laser operating at 3.9 µm, Electron. Lett. 31, 1250 (1995) 243 152. W. H. Dumbaugh, J. C. Lapp: Heavy-metal oxide glasses, J. Am. Ceram. Soc. 75, 2315 (1992) 243 153. T. Schweizer, B. N. Samson, J. R. Hector, W. S. Brocklesby, D. W. Hewak, D. N. Payne: Infrared emission from holmium doped gallium lanthanum sulphide glass, Infrared Phys. Technol. 40, 329 (1999) 243, 244 154. T. Schweizer, B. N. Samson, J. R. Hector, W. S. Brocklesby, D. W. Hewak, D. N. Payne: Infrared emission and ion–ion interactions in thulium- and terbium-doped gallium lanthanum sulfide glass, J. Opt. Soc. Am. B 16, 308 (1999) 243, 244 155. T. Schweizer, D. W. Hewak, B. N. Samson, D. N. Payne: Spectroscopic data of the 1.8-, 2.9-, and 4.3 µm transitions in dysprosium-doped gallium lanthanum sulfide glass, Opt. Lett. 21, 1594 (1996) 243, 244 156. D. W. Hewak, J. A. Medeiros Neto, B. N. Samson, R. S. Brown, K. P. Jedrzejewski, J. Wang, E. Taylor, R. I. Laming, G. Wylangowski, D. N. Payne: Quantum-efficiency of praseodymium doped Ga:La:S glass for 1.3 µm optical fibre amplifiers, IEEE Photonics Technol. Lett. 6, 609 (1994) 243 157. C. C. Ye, D. W. Hewak, M. Hempstead, B. N. Samson, D. N. Payne: Spectral properties of Er3+ -doped gallium lanthanum sulphide glass, J. Non-Cryst. Solids 208, 56 (1996) 243 158. T. Schweizer, D. J. Brady, D. W. Hewak: Fabrication and spectroscopy of erbium doped gallium lanthanum sulphide glass fibres for mid-infrared laser applications, Opt. Express 1, 102 (1997) 243 159. T. Schweizer, B. N. Samson, R. C. Moore, D. W. Hewak, D. N. Payne: Rareearth doped chalcogenide glass fibre laser, Electron. Lett. 33, 414 (1997) 243 160. M. Asobe, T. Ohara, I. Yokohama, T. Kaino: Fabrication of Bragg grating in chalcogenide glass fibre using the transverse holographic method, Electron. Lett. 32, 1611 (1996) 243 161. R. Mossadegh, J. S. Sanghera, D. Schaafsma, B. J. Cole, V. Q. Nguyen, R. E. Miklos, I. D. Aggarwal: Fabrication of single-mode chalcogenide optical fiber, J. Lightwave Technol. 16, 214 (1998) 243 162. T. M. Monro, Y. D. West, D. W. Hewak, N. G. R. Broderick, D. J. Richardson: Chalcogenide holey fibres, Electron. Lett. 36, 1998 (2000) 243

Index

absorption, 222–224, 228 – coefficient, 224, 228 bleaching, 231 branching ratio, 228 broadening (lifetime), 227 cascade lasing, 237, 239 ceramics, 222, 223 chalcogenide, 222, 243 – glass, 232, 243, 244 cladding – pumped, 223 co-dopant, 240 co-doped, 231, 236, 240 core pumped, 223, 226, 228 cross-relaxation, 231, 233, 235–237, 240 cross-section, 228, 238 double-clad – fiber, 224, 226, 240 – geometry, 239 – pump, 224 Dy3+ , 243, 244 electric-dipole – transition, 228 emission cross-section, 222, 223, 228, 230 emission spectra, 228 energy migration, 231, 232 energy recycling, 241, 242 energy transfer, 230–232, 235, 237, 240 – upconversion (ETU), 231, 240–242 energy-level scheme, 227, 233, 235, 237, 239, 240 Er3+ , 223, 227, 232, 243 excited-state absorption (ESA), 228, 230, 238, 239

fluoride, 230 fluoride fiber , 222 fluoride glass, 222 – fluoride, 222 four-level – CW laser, 234 – fiber laser, 237 – laser, 240 – – transition, 238 GaLaS (GLS), 221, 222, 243 ground-state, 231 – bleaching, 228, 240 – transition, 233, 236 ground-state absorption (GSA), 228, 231, 235, 238 heat, 226 Ho3+ , 232, 243, 244 lanthanide, 227 laser, 234 – self-terminating, 231 – three-level, 234, 238 – transition-metal-ion-doped, 226 lifetime, 228, 230, 231 – quenching, 226 loss, 221, 223, 224, 243, 244 multiphonon – relaxation, 220, 222, 226, 228, 230 multiplet, 227, 228 Nd3+ , 223, 225, 243, 244 nonradiative decay, 228, 230 numerical aperture, 224 oxides, 230, 243

Index parity, 228 phonon energy, 220–222, 230, 232, 243 Pr3+ , 240, 243, 244 Pr3+ -co-doped, 240 pump, 224 pump absorption, 224, 226, 228 pump cladding, 223, 224 quantum efficiency, 220, 222, 226, 230, 234, 240 quenching, 231, 233, 240, 244 radiative decay, 228, 230 radiative transition, 228 rare-earth ion, 222, 226–228, 231, 243, 244 rare-earth solubility, 222 rare-earth-doped – chalcogenides, 222 – fiber laser, 232 – heavy-metal oxide, 243 – solid-state laser, 230, 231 rate constant, 228, 230 reabsorption, 226, 230, 234 recycling, 239, 240 refractive index, 222, 224 – fluctuation, 224 resonator, 225, 230 self-terminating laser, 231

255

self-terminating transition, 237 sensitizing, 231, 235–237, 244 silica, 221 – glass, 232 silicate, 222 slope efficiency, 232, 236, 238–242 solubility, 220, 222 spectral-line broadening, 222, 227, 228 Stark, 227, 233, 238 – component, 238 – level, 234 stimulated emission, 228, 230 Stokes efficiency, 234, 238–240 sulfide fiber, 244 sulfide glass, 222 Tb3+ , 243 thermal conductivity, 221–223 three-level laser, 234, 238 threshold, 228, 230, 234–236, 240, 243 Tm3+ , 232, 235–237, 243 transition-metal ion, 223, 226 transition-metal-ion-doped laser, 226 transparency, 220–222, 232 tunability, 234, 235, 238 – of lasers, 228 Yb3+ , 223, 225, 237 ZBLAN, 221–223, 228, 230, 232, 243

Crystalline Mid-Infrared Lasers Irina T. Sorokina Institut f¨ ur Photonik, Technische Universit¨ at Wien Gusshausstr. 27/387, A-1040 Vienna, Austria [email protected] Abstract. A survey of the well-established as well as newly emerging ion-doped crystalline lasers, operating in the mid-IR spectral range between 2 and 5 µm, is presented. The review includes rare-earth- and transition-metal-based ionic crystals, as well as color-center lasers. The emphasis is made on state-of-the-art compact all-solid-state room-temperature tunable sources, belonging to the broad class of vibronic lasers. The announcement of the efficient high-power room-temperature operation and super-broad tunability, as well as the possibility of generating ultrashort pulses from the novel class of chromium doped chalcogenide lasers led to a strong revival of research interest in vibronic laser systems, involving 3dn transition-metal ions. The review outlines the underlying physics of mid-IR lasers and sorts out the essential spectroscopic and laser characteristics, allowing assessment of the suitability of laser materials to serve as active media in diode-pumped laser systems.

1

Introduction

The purpose of this chapter is to review the existing crystalline lasers, covering the wavelength range roughly between 2 and 5 µm. Although various kinds of lasers are being considered, the emphasis is made on state-ofthe-art continuous-wave, diode-pumped and tunable lasers. These lasers are based on both rare-earth ion doped crystals, which usually operate at fixed wavelengths, but allow high-power operation, and transition-metal (TM) ion doped- and color-center crystals, allowing broad tuning and in some cases also ultrashort pulse generation. The bandwidth of the gain medium is therefore one of the most important parameters when reaching the operation threshold (especially in the diodepumped regime) and tuning are to be considered. Throughout this chapter, bandwidth will be understood as the full width at half maximum relative to the gain maximum at λ0 , i.e. ∆ λ/λ0 ≈ ∆ ν/ν0 . Besides the fact that this definition is the same in the wavelength and frequency domains, it also allows one to compare gain media with different central wavelengths. Figure 1 provides a fair comparison of the existing crystalline active media, having the broadest gain in a certain wavelength range at room temperature, going on the wavelength scale through the near-infrared to mid-IR range. For ultrashort pulse generation, (∆ λ/λ0 )−1 has the physical meaning of the number of optical cycles per pulse. I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 255–351 (2003) c Springer-Verlag Berlin Heidelberg 2003

Gain spectrum

256

Irina T. Sorokina Ti:Saph Cr:LiSGaF

700

1000

Cr:YAG

1400

Cr:ZnSe

2000

Tb:KPb2Cl5

3000

4500

6000

Wavelength (nm, log scale)

Fig. 1. Broadband gain media. The abscissa is a logarithmic scale, keeping the relative bandwidth ∆ λ/λ0 constant

The mid-IR (MIR) wavelength region, which is also often called the “molecular fingerprint” region, and in particular, the range between 2 and 5 µm is characterized by the presence of strong fundamental vibrational absorption lines of atmospheric constituents, vapors and other gases (see, e.g. Fig. 1 in the chapter by Fischer and Sigrist). Those include, e.g. water vapour (H2 O), filling the whole range between 2.5 and 3 µm with a maximum around 2.7 µm; carbon monoxide (CO) with strong features around 2.3–2.4 µm; carbon dioxide (CO2 ) absorbing around 2.7–2.8 µm and 4.3 µm; nitrous oxide (N2 O), having several absorption features all through the 2–4 µm range, as well as many other species. Detection of low concentrations of these and other molecules (for details see the chapter by Richter et al.), constituting air pollutants or green-house gases for the purpose of environmental diagnostics or even human breath for the purpose of medical diagnostics, is currently done using laser systems, which are based mainly on nonlinear optical conversion techniques and include optical parametric oscillators (OPO, see the chapters by Vodopyanov and by Ebrahinzadeh) and difference-frequency generators (DFG, see the chapter by Fischer and Sigrist). Semiconductor lasers, including heterojunction lasers, lead-salt, antimonide and quantum cascade laser sources, are probably the simplest and the most cost-effective sources in the mid-IR wavelength region. However, except for a few cases they require cryogenic cooling and provide limited output power levels. Crystalline solidstate lasers, on the other hand, can provide very high power levels retaining good beam quality and narrow spectral linewidth. In combination with nearinfrared diode lasers as pump sources these lasers [1,2] can offer stability, efficiency and compactness as well as broad spectral coverage and tuning ranges, which are generally inaccessible for semiconductor lasers. Therefore, the availability of room-temperature diode-pumped tunable solid-state lasers as a simple and compact alternative in this wavelength region is a significant step forward in remote sensing and trace gas detection, as well as in other applications. Furthermore, the ultrabroad gain bandwidth of some laser crystals allows generation of ultrashort pulses of only a few optical cycles. This makes these lasers attractive for such applications as mid-IR free-space communica-

Crystalline Lasers

257

tions, optical frequency standards, as well as optical coherence tomography (OCT) in medicine. Further medical applications of mid-IR solid-state lasers include tissue cutting and welding, ophthalmology, neurosurgery, etc. (see the chapter by Jean and Bende). Summarizing, numerous applications in the mid-IR benefit from the development of the low-cost rugged broadly tunable solid-state lasers. The chapter is organized as follows. After a brief introduction and description of the historical evolution of these lasers (Sect. 2), the chapter provides the spectroscopic background necessary to understand the basics of laser operation of tunable crystalline lasers (Sect. 3) starting with the differences between rare-earth and transition-metal ion spectroscopy in octahedral and tetrahedral positions. The next sections give a systematic account of the available combinations of active ions and crystalline hosts (Sect. 3.1.1), of the commonly used pump sources, tuning and mode-locking techniques as well as power scaling approaches (Sect. 4). Although the used laser techniques are similar to those used in the visible and near-infrared regions, an attempt to simply extend those towards the mid-IR spectral region causes a number of new aspects and problems, requiring considerable modification of the experimental techniques and reconsideration of the existing theories (like e.g., in the case of Kerr-lens mode-locking). Other sections review existing continuous-wave and pulsed lasers.

2

Historical Overview

The first discussion of the feasibility of the infrared solid-state laser technique appeared in a paper by Schawlow and Townes in 1958 [3]. Crystalline lasers have played an important role in the history of lasers in general, and midIR lasers in particular. Indeed, chromium ions in a sapphire host has been a key player in laser devices since the first demonstration of the ruby laser (Cr3+-doped Al2 O3 ) by Maiman in 1960 [4,5]. Ironically, the very first brief Letter to Physical Review Letters titled “Optical maser action in ruby” was rejected by the editor, and was finally published in Nature [4]. A ruby laser is an example of a three-level system, and at that time could be operated only in the pulsed regime. In the same year Sorokin and Stevenson built the first mid-IR pulsed laser, which was based on calcium fluoride doped with trivalent uranium [6]. This laser operated at 2.6 µm with xenon flash-lamp pumping and liquid helium cooling. Later improvements in crystal quality permitted operation at room temperature. The U3+:CaF2 laser was a typical example of a fourlevel system. Shortly afterwards in 1962 Kiss and Duncan announced the successful operation of another mid-IR system, dysprosium-doped calcium fluoride at 2.36 µm [7]. This laser could be operated both pulsed and continuous-wave (CW) [7,8] and was pumped by a xenon flash-lamp. Sunlight has also been

258

Irina T. Sorokina

used to pump dysprosium-doped calcium fluoride cooled by liquid hydrogen [10]. Another interesting peculiarity of this material is the slow decrease of the active ion concentration of the divalent dysprosium, Dy2+, due to the valence change to Dy3+. Apart from this negative effect, preventing the longterm use of this material, until recently this was the only laser operating in the very interesting, for many applications, 2.3 µm water window wavelength range. The first tunable continuous-wave solid-state lasers to be demonstrated were those based on the divalent 3d transition-metal ions. In 1963 laser operation of nickel and cobalt ions in fluoride hosts was demonstrated at Bell Labs by Johnson, Dietz and Guggenheim [11]. The flash-lamp pumped Ni2+:MgF2 laser operated around 1.6 µm. A year later laser operation of Co2+:MgF2 and Co2+:ZnF2 was achieved by the same authors in the wavelength range between 1.75 and 2.16 µm [12]. Later, the operation of Ni2+ and Co2+ in several other hosts was reported [13,14]. However, all these media were characterized by low effective emission cross-sections. In the case of Ni2+-doped crystals this was due to excited state absorption. The only way to increase gain in these materials was to increase the upper laser level lifetime by cooling the crystal. Therefore, in order to be operated, all these lasers required cryogenic cooling. This, in part, hindered the commercial development of these very useful devices. The further advent of laser pumping allowed the tight focusing of the pump beam into the laser medium and therefore partially compensated for the low emission cross-section. Using laser pumping, CW operation has been achieved in Co2+ as well as in Ni2+-doped MgF2 [15,16]. Later also room-temperature operation was achieved by Welford and Moulton [17]. The discovery of color-center lasers (otherwise called F-center lasers), which may also be attributed to the class of broadly tunable vibronic lasers, was an important event in the solid-state laser community. The available tunable range of the color-centers in alkali halide crystals spans the mid-IR region between 2.3 and 4 µm [18]. When cryogenically cooled and optically pumped, these laser crystals have low threshold pump powers, relatively high output powers, and are smoothly and broadly tunable. In single-mode CW operation color-center lasers have extremely narrow spectral linewidths; in modelocked operation they can provide ultrashort pulses (down to < 50 fs [19]). In the 1970s and 1980s, F-center lasers, including commercial ones, found widespread use in various applications, especially in high-resolution spectroscopy requiring high spectral or temporal resolution, and in frequency standards. However, the main drawbacks of this type of source include the use of an insulating vacuum, and the necessity of cryogenic cooling, especially in the mid-IR region. Room-temperature color-center lasers exist nowadays only in the visible and near-infrared regions up to 1.3 µm [20]. This problem is similar to that of Co2+ and Ni2+-doped crystals, and is the reason for the decline of interest in these devices in recent years.

Crystalline Lasers

259

Another important group of ions includes rare-earth ions such as Tm3+, Ho3+ and Er3+, operating in the mid-IR spectral region at a whole number of wavelengths around 1.6, 2, and 3 µm. Limited continuous tunability is possible in lasers based on these ions, and is connected with the vibronic (i.e. phonon assisted) nature of the involved transitions. However, one should note that vibronic line broadening in lanthanide ions is much weaker than that in transition-metal ions due to the much lower interaction of ions with the lattice crystalline field. The pioneering works on Tm, Ho- and Er-doped crystalline media go back to 1962, when Johnson et al. reported the laser operation of Tm and Ho in CaWO4 at 2 µm [22,21] and Kiss and Duncan reported the operation of Er in the same crystal at 1.6 µm [24]. Those were the major milestones in the development of mid-IR crystalline (insulating) lasers during their first “discovery” or “pioneering” period. The discovered ions could be divided into four major groups: transitionmetals (e.g., Ni2+, Co2+ ), divalent and trivalent lanthanides (Dy, Er, Ho,Tm), actinides (U3+ ), and F-centers. In order to broaden the absorption pump bands of the ions, thus increasing laser efficiency, Voron’ko et al. suggested a new type of host matrix, which was a mixed disordered system (solid solution) [25]. This concept was further successfully used to develop both tunable and ultrashort-pulsed lasers. This way even femtosecond pulse generation became possible from the rather narrow-band rare-earth ions [26]. During this period, most important information on the physics of stimulated emission in crystals was accumulated. The experimental and theoretical methods known from maser pyhsics were extended towards lasers. The theoretical basis for the operation of broadly tunable lasers on vibronic transitions was established [27,28]. The maser concept of a three- and four-level scheme was successfully implemented in lasers. In the following two decades a burst of studies on the new laser materials took place, and hundreds of new laser types, operating at various wavelengths, were developed. For an extensive collection of existing crystalline laser materials and commonly used crystalline lasers the reader is referred to the books by Kaminskii [29] and Koechner [30]. The variety of energy levels of active ions provided the possibility of creating different complex excitation schemes, including step-wise excitation based on upconversion and cross-relaxation [31,32,33,34], cascaded excitation [35] and sensitization of the active ion with another ion/s, absorbing the pump energy and transferring it to the upper laser level (see, e.g. [36]). The term “crystal engineering” has become common in the laser community, and indeed active media could be very often “engineered” to possess specific spectroscopic and thermo-mechanical properties necessary to develop a new laser in the required wavelength range [37]. During the last decade the solid-state laser field was driven by a number of factors. Those include the invention of laser- and especially diodepumping [1], opening the possibilities for the creation of compact, room-

260

Irina T. Sorokina

temperature tunable sources, the success of the CW tunable and ultrashort pulsed Ti:sapphire laser [38], and the appearance of new applications for such sources. Crystals doped with Tm3+ (e.g., Tm3+:YAG, Tm3+:YLF, Tm3+:YVO4 and Tm3+:YSGG) provide tunable CW operation between 1.86 and 2.3 µm [39,40] and those doped with Er3+ can operate around the 2.7– 2.9 µm wavelength range which is very interesting for medical applications. Huber et al. [41] and Fan [42] were the first to report the room-temperature laser- and diode-pumped operation of Tm, Ho, and Er-doped garnets. Pulses as short as 40 ps could be obtained from Er-containing media [43] and 25 ps from Tm-containing mixed garnets [44]. However, when speaking about tunability and ultrashort pulse generation the priority should certainly be given to the new class of vibronic materials, Cr2+-doped II-VI compounds, that was first suggested in the mid-1990s by the group of scientists at the Lawrence Livermore Laboratory [45,46,47]. This type of lasers allows high-power room-temperature continuous-wave operation [48] tunable over more than 1000 nm between 2 and 3 µm. The principles of operation and description of the existing state-of-the-art Er, Tm, and Hodoped oxide and fluoride, as well as Cr-doped chalcogenide lasers will be given in the following sections of this chapter. All through the chapter crystalline solid-state lasers will be referred to as just “solid-state lasers”.

3

Principles of Solid-State Lasers

In this section the spectroscopic background will be provided to the extent that is necessary to understand the principles of laser operation and the characteristic features distinguishing tunable, mainly vibronic active media, from their fixed-wavelength counterparts (for a detailed recent spectroscopyrelated review see [49]). Vibronic lasers are those that operate on the transitions that involve the simultaneous absorption or emission of both photons and phonons. The emission bandwidth exceeds the energy of the maximum phonon in such media. Usually these are lasers based on transition-metal ions (exceptions are rare-earth ions such as Ce3+, Eu2+, Sm2+ or Yb3+ [50], but they are not known to lase in the mid-IR). Although it exists (see e.g. [51]), the vibronic broadening in rare-earth ion-doped crystals is much smaller than in transition-metal doped crystals. However, noticeable tunability can also be achieved in this group of crystals. The available bandwidth in this case is determined by the combined action of vibronic interaction, significant Stark splitting, and sometimes inhomogeneous broadening due to the multicenter nature of the mixed solid-solutions (see e.g. [25,52]). Nevertheless, when going into the infrared the relative contribution of the phonons to the transition (in energy units) becomes larger, and eventually, all fixed-wavelength (in the visible) lasers become broadband and potentially tunable.

Crystalline Lasers

3.1

261

Spectroscopic Background

The rare-earth ions, which operate in the mid-IR range, include: the trivalent lanthanides Tm, Ho, Er, Dy, Pr, Tb, Nd; divalent Dy and the trivalent actinide U. To the present moment the wavelength range, covered by the paramagnetic ions (with incomplete electronic shells) in crystals extends up to 7.2 µm [53]. The successfully operating tunable transition metal ions are divalent Co, Ni, Cr, and Fe in fluorides and chalcogenides. 3.1.1 Ion-Host Interaction: Octahedrally versus Tetrahedrally Coordinated Ions, Energy Levels, Configurational-Coordinate Model The lanthanide rare-earth ions have filled 4d, 5s, 5p, and 6s shells, and the valence electrons enter either 4f or occasionally 5d shells. Because of the better stability the rare-earth ions in crystals have a strong tendency to be incorporated in the trivalent state. The valence electrons form bonds with the neighboring fluorine or oxygen atoms. The remaining optically active 4f electrons are shielded from the surrounding crystalline environment by the filled 5s and 5p shells. The filled shells are spherically symmetric and do not affect the 4f electrons. Therefore, there is only limited interaction with the crystalline field. Due to the shielding by the outer shell electrons, the crystal field splitting is treated in the weak-field limit, and the spin–orbit coupling is the most significant interaction that determines the free-ion multiplets. The free-ion Hamiltonian H can be written as a sum of the unperturbed Hamiltonian H0 of the interaction of each electron with a nucleus, a Hamiltonian Hcoul of the Coulomb interaction of electrons with each other, and a Hamiltonian Hso of the spin–orbit coupling. For an ion inside the host an additional term describing the interaction with the crystalline field Hcryst is added, and the overall Hamiltonian becomes equal to [54,55] H = H0 + Hcoul + Hso + Hcryst .

(1)

As can be seen in Fig. 2 in rare-earth ions (weak crystal field limit [55]) the magnitude of the Coulomb and the spin–orbit couplings is approximately equal and is of the order of tens of thousands of wave numbers. The Stark splitting due to the crystalline field is much smaller and is of the order of hundreds of wave numbers. The energy levels of rare-earth ions generally retain the character of the corresponding levels of the triply ionized rare earth ions in a gas, and change only slightly from host to host. Therefore, it is justified to use the well-known and very useful Dieke diagram [56] to estimate the relative position of the multiplets and the widths of the Stark splittings in crystals. Figure 3 provides a simplified Dieke diagram, containing only the levels of the rare-earth ions discussed in this chapter, which are relevant to the tunable mid-IR operation. For a detailed overview the reader is referred to [57,58,56,54].

262

Irina T. Sorokina

F r e e io n

C o u lo m b fie ld

C r y s ta l fie ld

F 5

4 f1

S p in - o r b it in te r a c tio n

S 5

~ 1 0 0 0 0 c m 0

-1

I4 5 5

5

I 5

5 2 0 0 c m

-1

5

E le c tr o n c o n fig u r a tio n

T e rm s

M u ltip le ts

I5 I6 5 I7

1 7 s u b le v e ls (1 5 o b s e rv e d ) 4 7 0 c m

I8

-1

S ta rk c o m p o n e n ts

Fig. 2. The effect of the crystalline field on the energy level splitting in lanthanide ions in the 4f electron shell. The terms and numbers on the graph correspond to the Ho3+ in the YSGG crystal

In contrast to lanthanide ions, where the optically active ions have transitions in the inner 4f electron shell, the actinide rare-earth ions have partially filled 5f shells. Therefore, they are more sensitive to the crystalline environment. As a result, these ions can be found in various, up to 6+, ionization states. The radioactivity of these ions restricted their use to trivalent uranium in various fluorides: CaF2 [6,59], BaF2 [60], SrF2 [61] and YLF [62]. Similarly to transition-metal ions their spectroscopic properties depend more on the crystalline host environment, and vibronic line broadening plays a correspondingly larger role. The Stark splitting is larger and the f –f transitions are stronger for actinides than for lanthanides. For example, the ground 4 I9/2 level of U3+ in YLF is split to 1113 cm−1 , favoring the four-level operation scheme [63]. The relatively smooth emission spectrum extends between 2.2 and 2.8 µm, and laser action has been observed at various wavelengths in this range depending upon the host. There exists a large potential for U3+-doped fluorides as tunable mid-IR lasers. In 1994 room-temperature CW laser operation at 2.8 µm was demonstrated in YLF [62]. However, the interest in this laser revived at that time did not find further development, most probably due to the discovery in 1995 of a new class of tunable lasers in the same wavelength range: transition-metal doped chalcogenides [45,46,47]. In order to get a better understanding of the influence of the host on the spectral characteristics of an ion let us consider as an example a typical laser host such as a garnet. Garnets have a stoichiometric formula A3 B2 C3 O12 , comprising an eight-fold coordinated (relative to oxygen) dodecahedral A site, a six-fold octahedral B site and a four-fold tetrahedral C site [64,65,66]. The elementary cell contains eight such formula units. Most of the trivalent rare-earth ions of interest are incorporated into the dodecahedral and octahedral sites (Fig. 4) and are surrounded by a set of nearest-neighbor oxygen ions (ligands). Analogously, in the case of a fluorite host this is a set of fluo-

Crystalline Lasers

263

Fig. 3. Simplified Dieke diagram (rare-earth ions in LaCl3 ) [56]

rine ions. As can be seen in Fig. 4 the dodecahedral sites are the largest and therefore are usually occupied by the largest ions. The rare-earth ions usually enter dodecahedral and octahedral sites and transition-metal ions prefer octahedral and tetrahedral sites. The consequence of this site preference rule is illustrated in the following: with a decrease of the ionic radii from Ho3+ to Er3+ and Tm3+ (rare-earth ions with larger atomic number have smaller ionic radii, this is known as lanthanide contraction [58]) an increase in the occupation probability of the octahedral site takes place. This leads to multiple sites for these ions (see e.g. [68,69,70]), and as a result, to inhomogeneous line broadening. Although the multisite occupation is generally considered

264

Irina T. Sorokina

- o x y g e n a n - o c ta h e d ra - te tra h e d ra - d o d e c a h e

io n s l s ite s l s ite s d r a l s ite s

Fig. 4. Oxygen polyhedra in a garnet host, forming octahedral, tetrahedral, and dodecahedral sites (after [66])

to be a negative feature for a laser, this is certainly a positive feature for tunability purposes. Now let us consider the most common tunable laser media based on 3d–3d transitions of transition-metal ions of the iron group, which are most sensitive to the crystalline environment. This is so, because in contrast to lanthanides, where the active ions are characterized by the optical transitions of the electrons in the inner 4f electron shell, which is shielded from the crystalline field, the optical transitions in transition-metal ions occur in the outer, open 3d-shell. The interaction of transition-metal ions with the crystalline host creates broad absorption and emission bands, which are finally used to tune the laser. Therefore knowledge of the mechanisms of such interaction is important for understanding the principles of tunable solid-state lasers. The active ions interact with the ligands through an electrostatic Coulomb interaction. This ion–host interaction had been treated by several theories. Bethe [71] introduced crystal-field theory [71]. Later Van Vleck and coworkers proposed an alternative theory based on covalent bonding [72,73]. Nowadays it is known as ligand-field theory [74]. In spite of the drawbacks and shortcomings of this elegant and rather simple theory, the description of which is beyond the scope of this review, it provides a useful tool for analysis of spectroscopic information (for details see [75,55]). This theory allows us to qualitatively predict the crystal-field splitting of the free ion energy levels due to a certain type of crystal-field symmetry at the ion’s site. The quantitative prediction of the Stark splitting of the levels is more difficult. Nevertheless, very often it is adequate to reduce the problem to the simplified point charge model, where the active ions and the ligands are point charges, possessing specific symmetry. In this case the active ion experiences the crystal field determined by the symmetry of the ligand environment. Figure 5

Crystalline Lasers

265

Fig. 5. Ion arrangements producing electrostatic crystalline fields of octahedral, tetrahedral, and cubic symmetry (after [55])

shows the three typical ion arrangements of the specific symmetries: octahedral, tetrahedral and cubic. One uses to say that the active ions occupy then the octahedral, the tetrahedral and the cubic sites. Among the common laser crystals the most frequently occupied sites are those of octahedral and tetrahedral symmetry. However, in reality there is usually a distortion from the perfect octahedron or tetrahedron and the site has lower symmetry. Usually one considers the crystal field as consisting of two components: a strong octahedral (or tetrahedral) component and a weaker one of lower symmetry, which is considered as a perturbation in the perturbation theory analysis of the energy levels. In contrast to the rare-earth ions, where the crystalline field is relatively weak because of the screening effect of the electrons in the outer shells, in the case of transition-metal ions the crystalline field is stronger or comparable to the Coulomb interaction between the 3d electrons, and the latter is stronger than the spin–orbit coupling interaction. This holds true for the intermediate and strong crystal field limit characteristic of transition-metal ions [55]. Therefore the position of the energy levels E is calculated as a function of the crystal-field strength Dq (parameters D and q always occur as a product, and hence Dq can be regarded as a single parameter) and of the Racah parameters B and C. The latter reflect the Coulomb interaction between the ion and the ligands [55]. The Tanabe–Sugano diagrams [76,77] for the 3dn configuration for n = 2 to 8 show the dependence of the level energy E/B on the crystalline field Dq/B at the fixed value of C/B. Both the level energy, which is measured from the ground level, and the strength of the crystalline field are scaled by B as a unit. The elegance of the Tanabe–Sugano representation consists in the certain symmetry. Indeed, the energy levels of the ions in the 3dn configuration are inverted relative to those in the 3d10−n configuration. This is so because the Coulomb energy of the electrostatic interaction between the n positive charges is the same as that between n negative charges. Hence, the free ion energy levels are the same in both systems, but the interaction energy between the positive charges and the crystalline field will be opposite in sign to that for the negative charges. Moreover, all the calculations of crystal field splittings for the case of octahedral coordination can be used for the case of tetrahedral coordination with the same ion-to-ligand distance. One only needs to reverse

266

Irina T. Sorokina

the sign of the parameter Dq and reduce its value according to the following rule [55]: Dq (tetrahedral) = −4/9 Dq (octahedral) .

(2)

The conclusion that can be drawn from this general rule is that the energy levels of the ions in the tetrahedral coordination experience smaller energy splittings due to the crystalline field. This results in the correspondingly longer wavelength optical transitions. The parameter C/B mentioned above is almost constant through the transition metal series. In addition, the B parameter is of the order of 1000 cm−1 for all known ions [37]. Thus, Tanabe–Sugano diagrams also allow an approximate quantitative analysis of energy level structures of all transition-metal ions in almost any crystalline environment. Examples of Tanabe–Sugano diagrams for mid-IR active ions such as Co2+ ions in octahedral sites, occurring, e.g. in the Co2+:MgF2 laser and corresponding to the 3d7 configuration, and Cr2+ ions in tetrahedral sites, used e.g. in various chalcogenide crystals and corresponding to the 3d4 configuration, can be found in Fig. 6. The left scale shows the splitting energy both in E/B units and in cm−1 , using the empirical values of the B parameter. The lines on the diagram represent different energy levels and are classified according to their spin state and group symmetry. In addition, the corresponding electronic configurations are given in parentheses. It is easily seen that the

Fig. 6. Tanabe–Sugano diagrams for (a) octahedrally coordinated Co2+ ions (3d7 configuration) and (b) tetrahedrally coordinated Cr2+ ions (3d4 configuration). Free-ion terms are shown to the left of the diagram. Energy levels with different spin states are shown by lines with different thickness. Both ions experience low crystalline field (Dq/B < 1.1 for most hosts). The transition, marked as ESA, is spin-allowed for the Co2+ ion, but is spin-forbidden for the Cr2+ ion. Only a part of the full diagram [76] is shown, corresponding to lower energies, that are relevant to the mid-IR transitions

Crystalline Lasers

267

energy intervals between the levels belonging to the same configurations tend to be rather independent of the Dq/B (the lines go almost parallel), while energy intervals of levels, belonging to different configurations, depend strongly on crystal field strength (the tilts of the lines are different). This has a major implication on spectral parameters of the ions: as will be seen later, the absorption and emission lines corresponding to the transition between states with the same electronic configuration tend to be narrow-band, and those corresponding to transitions involving configuration change are broadband. A typical example of the d4 (Cr2+ ) configuration in tetrahedral symmetry is shown in Fig. 6b. The lowest free-ion 5 D state of Cr2+ is split by the crystal field potential by an amount defined as the crystal field splitting ∆ = 10Dq into an excited orbital doublet 5 E and a ground-state orbital triplet 5 T2 , where Dq =

Ze2 r4 . 24πa5

(3)

Here, Ze is the charge on the ligand ion, is the dielectric susceptibility, a is the distance between the ion and the ligand, and r4 is the matrix element for the 3d-orbital, where r is the distance from the ion nucleus. Both the orbital triplet ground- and the orbital doublet excited states are further split by the Jahn–Teller effect [78,79,80], which plays a large role in spectroscopic properties of transition metal ion doped crystals (e.g. in Ti:sapphire [38], as well as in Cr:ZnSe and other chalcogenides [81,82,83,84,85]). The Jahn–Teller theorem was formulated in 1937 and states that “any non-linear molecular system in a degenerate electronic state will be unstable and will undergo distortion to form a system of lower symmetry and lower energy thereby removing the degeneracy”. The simple crystal-field treatment fails to take into account the Jahn–Teller splitting. It is therefore important to treat the lattice as a dynamic system coupled to the electronic d shell of the active ion. One of the results of this coupling to the lattice is to allow Jahn–Teller deformation at the impurity ion site. These Jahn–Teller effects can be either of dynamic or static type [86,87,88]. The magnitudes of these deformations can be derived from group-theoretical considerations. It is not the purpose of this paper to go into details of this phenomenon. It is only worth noting that the splitting of the ground state and the excited states (amounting in some cases to several hundreds of wavenumbers) caused by the Jahn–Teller effect should always be kept in mind, when considering tunable transitionmetal doped lasers. As an example of the influence of such splitting on the laser characteristics, one may consider the passive losses in Cr:ZnSe at the laser wavelength of 2.5 µm [165]. The absorption around 2.5 µm arises from the overlap between the absorption band due to 5 T2 → 5 E transition and the transition within the ground state 5 T2 due to the static Jahn–Teller effect around 6.6 µm [89,81]. As a result, losses due to ground-state absorption (GSA) within the ground state are intrinsic to Cr:ZnSe crystals and provide an upper limit for the figure of merit (FOM) of this material. A characteristic

268

Irina T. Sorokina

such as FOM is a useful parameter, describing the optical quality of the lasing material. It is commonly defined as the ratio between the attenuation at the maximum of the absorption to the attenuation at the lasing wavelength. This value differs in various laser materials, achieving ∼ 100 or more in good ones. It should also be noted that transitions of the ion in a tetrahedral site are generally stronger. Strictly speaking the transitions between energy levels of the 3dn configuration are transitions of the same parity and therefore obey the Laporte rule [90], i.e. are electric-dipole forbidden. This selection rule can be weakened as a result of the influence of odd-parity crystal-field terms, which mix the d and p orbitals, having different parity. In the centrosymmetric octahedral sites, odd-parity crystal-field terms may occur due to the small site symmetry distortion, or in the presence of vibronic (vibrational and electronic) coupling between wavefunctions with opposite parities. During vibrations the ion spends part of its time away from its equilibrium position, i.e. it is no longer at a center of symmetry. In contrast to the octahedral site case, in tetrahedral sites the crystal field has no inversion symmetry, the mixing of orbitals is strong, the Laporte rule does not strictly apply and the transitions are characterized by the much higher oscillator strength. The oscillator strength f is connected to the absorption coefficient k by the following relation: mc k(ν)dν , (4) f= πN e2 where N is the number of active centres per cubic centimeter, ν is the frequency in Hz, and m and e are the mass and the charge of the electron correspondingly. Since k(ν) = σ abs N , where σ abs is the absorption crosssection, the above expression for the oscillator strength relates it to σ abs or the inverse radiative lifetime τ rad by the following relation: 8π 2 e2 ν 2 f. (5) τ rad mc3 Indeed, transition-metal ions in tetrahedral sites are characterized by transition cross-sections that are generally higher than in octahedral sites and the shorter lifetimes. Compare, e.g. the absorption cross-section in Cr3+:LiCAF [91], which is of the order of 2·10−20 cm2 , with that in the tetrahedraly coordinated Cr2+:ZnSe or Cr2+:ZnS, which is of the order of 10−18 cm2 . Such high cross-sections and short lifetimes may be good for CW and ultrashort-pulsed operation, but may in some cases be disadvantageous because of the decreased energy storage capability of the short-lived media. Finally, the tetrahedral sites provide lower crystal-field stabilization of the ions than the octahedral sites [92]. Covalent bonding plays larger role in the tetrahedral sites, causing the frequently observed multiple valence states of the transition-metal ions in these sites and the high probability of chargetransfer processes. Indeed, whereas in octahedral sites internal e ↔ t2 transitions do not significantly redistribute the charge around the active ion (i.e., 1

=

Crystalline Lasers

269

both the initial and the final one-electron wave functions are almost equally localized around the impurity, hence the term “internal” transitions), in covalent tetrahedral sites such excitations can redistribute the charge from impurity-centered orbitals (i.e. essentially purely d-like e states) to ligandcentered orbitals (i.e., p-d hybridized t2 states) [93,94]. They are, therefore, not really “internal” transitions. This is especially true for tetrahedrally coordinated Cr2+ ions in II–VI compounds, which are used as broadly tunable crystalline active media in the mid-IR wavelength range. As has been recently shown, the related charge transfer processes in some cases may affect the laser performance of tunable CW lasers with the active ions in tetrahedral sites [95,96,97,98]. For the analysis of the vibronic line bandwidth the ion in the crystalline environment can be approximated by a one-dimensional harmonic oscillator. In Fig. 7, the potential energies of the ion in different electronic states are shown as parabolae along a generalized configurational coordinate, corresponding to one of the irreducible vibrational modes of the lattice. As the electronic wavefunctions in the upper and lower states are different, the equilibrium positions of the ion along the vibrational coordinate may not be the same. As electronic transitions occur on a time scale much faster than vibrational movement of heavy ions (the Born–Oppenheimer or adiabatic approximation), absorption and emission of photons leaves the ion in an excited vibrational state. The excess energy ∆ E is transferred nonradiatively to the phonons of the lattice, causing a Stokes shift ≈ 2∆ E between absorption and emission photon energies. For laser oscillation purposes, it is advantageous to have ∆ E kT so that the thermal population of the lower laser level is low (four-level laser scheme). Large ∆ E also increases the transition bandwidth (∆ λ/λ0 ∝ ∆ E/E0 ) and the laser tuning range. At the same time, the two parabolic potential curves eventually intersect at some higher energy, allowing thermally activated nonradiative decay with activation energy Eact (Fig. 7), which inversely scales with the Stokes shift Eact ∝ E02 /∆ E. The detailed description of the nonradiative decay processes is not the pur-

Fig. 7. Energy level diagram in configuration space, indicating the four-level nature of the vibronic laser

270

Irina T. Sorokina

pose of this chapter and can be found in [55,99,100,101]. The nonradiative decay rate 1/τnr rapidly increases with temperature, having an exponential asymptotic dependence of ≈ exp(Eact /kT ) at higher temperatures. Thus, the large Stokes shift causes a high nonradiative decay rate and the condition ∆ E E0 should be fulfilled. In the mid-IR spectrum range the photon energy hν = E0 − ∆ E is small, so that the combination of the two conditions E0 ∆ E kT can hold only at low temperatures. In the case of rare-earth ions, which have a negligible shift of their electronic energy parabolae along the configurational coordinate, the competition from nonradiative decay processes may also become significant. The analytical form of the nonradiative decay rate is different from transition-metal ions and is governed by the direct multiphonon processes, having the form: 1 = Wnr (p)(n(ω, T ) + 1)p , τnr

(6)

where p = ∆ E/¯ hω is the number of phonons with frequency ω necessary to bridge the energy gap, n(ω, T ) = [exp(¯ hω/kT ) − 1]−1 is the mean thermal occupancy in this vibrational mode, and Wnr (p) exponentially decreases with p [55]. Thus, the nonradiative decay rate depends on the maximum phonon frequency of the lattice. Table 1 summarizes the typical values of the maximum phonon energies for various mid-IR laser materials. As can be seen from the table, one should expect the earliest (at the shortest wavelength) onset of the nonradiative decay in oxides, then in fluorides, and only afterwards in selenides, chlorides and bromides. As has been recently confirmed in comprehensive experimental and theoretical studies of multiphonon relaxation processes in rare-earth ions carried out by Basiev and co-workers [102,103,104,105,106,107,108] as well as by Auzel and co-workers (see e.g. [109,110,111]), an increase of the anion and cation atom mass and increase of the lattice parameter decreases the maximum phonon frequency and increases the number of phonons. According to formula (6) this leads to a decrease of the nonradiative decay rate and an increase of the fluorescence quantum yield. Thus, the heavier the ions of the lattice, the slower the multiphonon relaxation rate. Besides, there exists a strong dependence of this rate on the distance R0 between the rare-earth ion and the nearest ligands and the rare-earth ionic radius ri . The shorter R0 and the larger ri the faster the multiphonon relaxation rate is [105]. The larger the Stark splitting of rare-earth ion energy levels and the larger the matrix elements of the respective electronic transition the faster the multiphonon relaxation rate is [107]. Knowledge of these tendencies allows a directed search for new active media in the mid-IR region. Thus, nonradiative decay is one of the most important processes for midIR crystalline lasers. It sets the fundamental limit for obtaining continuouswave room-temperature laser operation from vibronic transitions at longer wavelengths. In fact, until the invention of Cr.ZnSe type materials there were no broadband lasers in the mid-IR wavelength range operating at room tem-

Crystalline Lasers

271

Table 1. Maximum phonon frequencies of different compounds Crystalline host

Oxides

Fluorides Selenides Chlorides Iodides Bromides

Maximum phonon 600–1200 350–650 frequency, cm−1

∼ 250

200–300

∼ 160

∼ 140

perature. Even narrowband fixed-wavelength lasers seldom operate at roomtemperature beyond 3 µm. 3.1.2 Intra- and Interionic Interactions: Ground-State Absorption, Excited-State Absorption and Upconversion In this final part of the section devoted to the overview of the spectroscopic background such processes as ground-state absorption (GSA), excited-state absorption (ESA) and upconversion will be defined. The relative contributions of the last two processes in the population dynamics of the upper laser level are different for broadband transition-metal ion-doped and rare-earth ion-doped crystals. ESA generally plays a large role in broadband materials, and upconversion is an important interionic process in narrow-band (longlived) media. Both can be considered as a loss mechanism (the exception is the use of upconversion in the multistage population of the upper laser level, described, e.g. in chapter by Pollnau and Jackson in this volume) in a laser. The GSA is an intraionic process, describing the absorption of a photon from the ground state. It is usually measured in a spectrophotometer and obeys the Lambert–Beer law: I(λ) = I0 (λ)e−nσGSA (λ)l ,

(7)

where n is the active ion concentration, σGSA (λ) is the GSA cross-section, and l is the active medium length. Spectroscopists use the GSA absorption spectra in order to obtain information on the position of the ion energy levels relative to the ground state and on the magnitude of the oscillator strength f of the transition. The ESA is another intraionic process, describing the absorption of a photon by the ion in the excited state, whereas the ion is excited into an even higher-lying state. As can be seen from Fig. 8, when the offset between the configuration curves of the first and second excited states is large (which is often the case in many transition-metal doped crystals), ESA appears as a broad band. It is a two-step excitation process and is sometimes referred to as reversed saturable absorption or nonlinear absorption. ESA is usually measured using the pump-probe technique, which is best described in [115]. Using this technique one measures the difference in the transmitted inten-

272

Irina T. Sorokina

Fig. 8. Configurational coordinate description of the broadband ESA

sities between the pumped and unpumped (Ip − Iu ) cases relative to the intensity in the pumped case Ip , which is equal to [115]: Ip (λ) − Iu (λ) = C [σGSA (λ) + σ(λ) − σESA (λ)] , Ip (λ)

(8)

where σGSA is the GSA cross-section, σESA is the ESA cross-section, σ is the emission cross-section, and C is a constant, which includes the sample thickness l, the population density in the excited state nexc and the lock-in amplification factor A, i.e. C = Anexc l. From (8), σESA is given by [49]: σESA (λ) = σ(λ) + σGSA (λ) −

1 Ip (λ) − Iu (λ) . C Ip (λ)

(9)

This method of determination of ESA is most widely used. However, it has some deficiencies because of the impossibility of precisely measuring the constituting parameters as well as the necessity of providing optimal pump and probe beam overlap. Another recently suggested alternative method of ESA measurement, which is partially free of the above deficiencies, is based on a relaxation oscillation study [116,117] (for the detailed treatment of relaxation oscillations in a laser see, e.g. [113]). This method is based on the fact that relaxation oscillations parameters ω, frequency, and 2γ, spectral width, both depend on pump in the excited-state absorption cross-section at the pump wavelength σESA the following way: (r − 1) , τ pump −1 σESA r n th 2γ = , 1+ 1− τ Nt σGSA

ω 2 τc =

(10)

(11)

Crystalline Lasers

273

where r is the relative pump rate, τc is the photon lifetime in the resonator, τ is the metastable state lifetime, n th is the critical inversion, Nt is the total concentration of the active ions, and σGSA is the pump absorption cross-section. The latter method has an advantage of being insensitive to misalignment or not optimal overlap between pump and probe beams, but requires knowledge of the relative threshold inversion. It can be quite sensitive to the fluctuations of the pump source if these occur close to the relaxation oscillation frequency and should better to be used in directly diode-pumped systems. Upconversion, opposite to ESA, is an interionic process and involves interaction of at least two ions, both being in the excited state. It was extensively studied in the mid-1960s, including the pioneering works by Ovsyankin et al. [118,119,120,121,122]. In the course of nonradiative energy transfer between ions in close proximity, the first ion is excited into an even higher-lying level and the second one relaxes to the ground state (Fig. 9a). In other words, the low-energy photons are converted to higher-energy photons in the course of the energy transfer. Therefore, one sometimes refers to upconversion as energy transfer upconversion (ETU). However, one should distinguish upconversion from two other ways of conversion of low-energy photons to high-energy photons. This is, first, a two-step photon absorption (Fig. 9b) that involves pumping of the upper laser level via ESA [124,125], and second, the so-called avalanche absorption upconversion process (Fig. 9c) that involves both ESA of the pump light and interionic cross-relaxation [126]. Within this chapter we limit our treatment to the first type of process, in the following referred to simply as upconversion. The opposite process is called cross-relaxation [34]. The latter is a unique process, allowing luminescence quantum yields greater than unity [123]. This process plays a major role in the population of the upper laser level in such rare-earth mid-IR lasers as Tm- and Er-doped lasers. At high active ion concentrations upconversion can be considered as a nonlinear process: its efficiency increases with the increase of the excited ion concentration n. According to the simplified and frequently used upconversion model the process rate has a dn/dt = −αn2 dependence [127], where the upconversion macroparameter α is an effective interaction rate constant between

hn2

hn1

(a)

(b)

(c)

Fig. 9. Schematic illustration of (a) upconversion, (b) twostep photon absorption, and (c) avalanche absorption upconversion

274

Irina T. Sorokina

the excited ions. However, as shown in [127,128,129] in many practically important cases the experimental upconversion kinetics cannot be described in terms of αn2 . There might be different mechanisms of upconversion: a static mechanism corresponding to direct energy transfer from the donor to acceptor ion, and migration-assisted energy transfer corresponding to the case when excitation migrates over the subsystem of donor ions before it reaches an acceptor. Varying the overall ion concentration N , one can change the mechanism of nonlinear upconversion from migration-assisted to the static one in the same type of crystal. In the latter case the initial part of the upconversion kinetics can be reasonably well fitted by the simple model accounting for the three-particle upconversion interaction described in terms of α n3 . Therefore, in order to correctly describe upconversion one should always carefully take into account the microscopic interaction constants and other parameters of the system. In [130] the analytical formula, describing upconversion beyond the αn2 approximation was obtained and the boundary ion concentrations were determined, separating different upconversion regimes. The term αn2 enters the rate equations along with the decay term n/τ , so that the upconversion is most important in media with long (millisecond) lifetimes at high levels of excitation, when αn becomes comparable to the spontaneous decay rate 1/τ . These upconversion processes take place in all kinds of laser systems based on: rare-earth doped (e.g. in Ho3+-doped) systems [131,132], mixed rareearth and transition-metal doped (e.g. in Cr3+,Er3+-doped [133,134,135] or Cr3+,Tm3+,Ho3+-doped garnets [136,137]) as well as purely transition metal doped crystals [138]. However, the probability of upconversion is higher in crystals with several metastable excited states with rather long lifetimes. These conditions are best realized in rare-earth doped (i.e. fixed-wavelength) laser materials (for details see Sect. 3.1.1). The classical objects for the study of upconversion processes are Er3+ ions, having a long millisecond lifetime of the lowest excited states [139]. This is also the reason why upconversion was better studied in the rare-earth doped systems (for a review on upconversion lasers see e.g [140]). Summarizing this section, it can be concluded that whereas ESA is a predominant loss mechanism in transition-metal ion-doped crystals, upconversion plays a major role mainly in rare-earth doped laser materials. 3.1.3

Energy Transfer Processes

In this section various types of energy transfer processes, radiative as well as nonradiative, leading to population and depopulation of excited states in laser crystals are considered. The population dynamics is explained using the examples of the energy level schemes for commonly used mid-IR lasers such as 3 µm Er and 2 µm Tm, Ho lasers. For more information on energy transfer processes in laser crystals the reader is referred to the books [141,142,143,144].

Crystalline Lasers

275

Nonradiative energy transfer processes between active ions in laser crystals (usually referred to as “donors” and “acceptors”) play an important role in both inversion population of the upper laser level as well as in its depopulation [145]. In other words they influence both gain and loss. The loss issues will be discussed in Sect. 3.2. In this section some important schemes of inversion population creation will be considered. In the mid-IR region lasers based on crystals co-doped with Cr3+, Er3+, Tm3+, or Ho3+ are of importance as sources of CW, Q-switched and modelocked radiation around 2 and 3 µm. The type of the interaction between the broadband transition-metal ions and narrow-band rare-earth ions (3d–4f interaction) differs from that between the rare-earth ions (4f –4f interaction). The principal difference is that in the former case the electrons of the outer partly filled 3d shell of transition-metal ions interact with the electrons of the inner partly filled 4f shell of rare-earth ions, which are shielded from the crystalline field by the filled s- and p-electron shells, and in the latter case the electrons of the inner 4f electron shells interact with each other. Such pair interactions between the active ions are described within various energy transfer theories. In particular, the theory of weak incoherent energy transfer in terms of the dipole–dipole interaction between ions, which was first suggested by F¨ orster [146,147], is widely applied for the description of the energy transfer between ions in laser materials. This theory is generally valid for relatively low concentrations of active ions. The validity conditions of this theory are given in [148]. Dexter [149] further developed this theory for the case of the higher-order multipole interaction. This theory is often used in the case of higher active ion concentrations. Both, dipole–dipole and multipole interactions have a Coulomb nature. Later Inokuti reported also on the exchange mechanism of energy transfer [150]. No matter what the energy transfer nature is (Coulomb or exchange), all types of interactions can be divided into resonant [151] and nonresonant (phonon assisted). Galanin, Agranovich and Agabekyan [151,142,152] developed energy transfer theories for the particular cases of strong coherent and incoherent interactions. Special attention was devoted to “hopping mechanism of energy transfer” [153], as well as to its particular case of supermigration and enhanced migration within percolation theory [154,155]. These mechanisms of energy transfer were studied, in particular, in conjunction with the problem of excitation migration over the subsystem of Tm3+ ions in various Tm3+, Ho3+ systems, where the concentration of Tm3+ ions is kept high in order to provide efficient cross-relaxation and delivery of the pump excitation to the lasing Ho3+ ions. Without going into details, these theories have laid the basis for nonradiative energy transfer studies in laser crystals. Radiative energy transfer, often referred to as radiation trapping, is another process that plays an important role in the population dynamics of the upper laser level. This process was thoroughly studied first in gases [156,157] and then in the solid state [158,159,160,161,163,164]. Radiation trapping is

276

Irina T. Sorokina

a result of reabsorption of the ions spontaneously emitted from the metastable level by the ions in the ground state. This reabsorption is followed by the subsequent emission of photons, and the whole process repeats. As a result, the effectively measured fluorescence time lengthens by up to a few times relatively to the radiative lifetime. The result is the reported discrepancy in the stimulated emission cross-sections measured by different authors. The phenomenon is characteristic of two, three and quasi-three level schemes. In the mid-IR region these are mainly Tm3+ and Er3+-doped systems. In codoped media (like e.g. Tm,Ho or Yb,Er-doped crystals) the radiation trapping increases not only the energy storage capability of the medium, but also enhances the nonradiative energy transfer between donors and acceptors and the energy transfer upconversion [164]. Recently reabsorption has been found to be significant also in Cr2+-doped lasers [165]. In particular, it was found that in four-level systems in the presence of reabsorption in order to measure passive losses it is justifiable to use only the inverse slope efficiency analysis [166] and not the commonly used Findlay–Clay analysis [167]. Before the beginning of the 1980s flash-lamp pumping was the common way to create inversion population in solid-state laser media. Sensitization of rare-earth ion luminescence by broadband chromium ions allowed increasing laser efficiency [168,169]. It is interesting to note that sensitization of luminescence in the mid-IR is more efficient than in the visible and near-infrared since with increasing wavelength the competing spontaneous emission rates within the sensitizers decrease as 1/λ3 . New excitation schemes [33] became possible with increasing use of selective laser excitation and opening the possibility of diode pumping. The obvious advantages of diode pumping are accompanied by the possibility of using nonlinear cooperative processes in exciting high-lying rare-earth levels, where direct lamp pumping is extremely ineffective. The pioneering work of Johnson, Guggenheim and Antipenko [170,171,33] laid the foundation for important diode-pumped mid-IR lasers such as 3-µm Er and 2-µm Cr,Tm,Ho lasers. The energy level diagram in Fig. 10 is an illustrative example of the multistep excitation of the upconversion-pumped 3-µm Er:BaY2 F8 laser. It incorporates nearly all the energy transfer processes discussed up to this point, e.g. cross-relaxation, upconversion, radiation trapping, etc. Although for the Er3+ ion there exist several laser transitions spanning the wavelength range between the visible and mid-IR, only two transitions are of interest in the mid-IR range, namely, 4 F9/2 →4 I11/2 around 2 µm and 4 I11/2 →4 I13/2 around 2.58–2.94 µm depending upon the host. Since the 2-µm range is well covered by Tm and Tm,Ho lasers, only the 4 I11/2 →4 I13/2 3-µm transition is referred to in connection with Er lasers. This laser transition is especially interesting because of the water absorption lines in this wavelength range (for a review of medical applications of Er lasers see the chapter by Jean and Bende). The main problem of the Er laser is the self-termination of the lower laser level 4 I13/2 , the lifetime of which is longer than that of the upper laser level

Crystalline Lasers

277

Fig. 10. Energy level diagram and main population and relaxation processes in Er3+:BaY2 F8 (after [164]). I – pumping; II – cross-relaxation; III – radiation and multiphonon relaxation; IV – radiation trapping (reabsorption); V – upconversion 4

I11/2 . At the beginning of Er laser development, this was the main obstacle, preventing continuous-wave operation in a YAG crystal. However, in highly concentrated crystals, e.g. 10% Er:BaY2 F8 , the population mechanism works as follows (Fig. 10): the pumping occurs via the 4 I15/2 →4 F7/2 transition; following the fast multiphonon relaxation 4 F7/2 →4 H11/2 →4 S3/2 there are two main nonradiative energy transfer processes [173,174,175] that play a role in the population of the upper laser level 4 I11/2 : (1) cross-relaxation according to the scheme 4

S3/2 →

4

I13/2 , 4 I15/2 →

4

I9/2 ,

(2) upconversion according to the schemes 4

I13/2 →

4

I15/2 , 4 I13/2 →

4

I9/2 and

4

I11/2 →

4

I15/2 , I11/2 →

2

H11/2 .

4

Finally, the excitation relaxes to the upper laser level 4 I11/2 . It was found that if multiphonon processes leading to population of the upper laser level are predominant, then the laser action occurs due to transitions between the Stark components of the 4 I11/2 and 4 I13/2 levels with the highest oscillator strengths. If nonradiative transitions due to ion–ion interaction are more probable, a “red shift” is observed in the lasing spectra [176]. Cross-relaxation processes contribute to the increase of the pumping efficiency. Upconversion results in depopulation of the lower laser level and additional population of the upper laser level, thus leading to a stationary inversion population. Due to reabsorption the lifetimes of both the upper and lower laser levels effectively increase. This may have two consequences: first, the effective

278

Irina T. Sorokina

threshold decreases [177], and second, enhancement of the upconversion processes [164]. Zhekov and Huber [175,41] showed that the interaction among Er-ions of the following type: 4 I13/2 →4 I15/2 , 4 I13/2 →4 I9/2 with the subsequent transition to 4 I11/2 leads to the elimination of the self termination of the laser transition 4 I11/2 →4 I13/2 . In Er:YLF this upconversion process which depopulates the lower laser level and populates the upper laser level allows this transition to operate in a CW mode [172]. The interplay of all the discussed processes together with the optimized Er-ion concentration and the proper choice of the host allowed the engineering of several 3-µm Er lasers, operating in continuous-wave, Q-switched and mode-locked regimes (see Sect. 5.3.2). These lasers are based, first of all, on (Y1−x Erx )3 Al5 O12 [173], as well as on Y3 Sc2 Ga3 O12 (YSGG) [178], Gd3 Sc2 Al3 O12 (GSAG) [179], (Lu1−x Erx )3 Al5 O12 and YAlO3 [180], CaF2 [181], SrF2 and LiYF4 [182], BaY2 F8 [183] and some other crystals [29,184]. Another illustrative example of multistep excitation of the upper laser level is the 2-µm laser operating on the 5 I7 →5 I8 transition of Ho3+ ions. Since the mid-1960s several research groups have been working towards the development of a crystalline 2-µm laser. The concept of such a laser consisted in the sensitization of the lasing Ho3+-ion by Er3+ and Tm3+ ions. One of the first and most efficient realizations of the 2-µm laser is the BaEr2 F8 laser [185] based on the combination of active ions suggested in 1965 by Johnson [186]. The excitation of the upper laser level 5 I7 of Ho ions takes place according to the scheme in Fig. 11. The relatively high efficiency of this laser (5.2%) is explained by the high quantum efficiency and the decreased thermal loading of the cross-relaxational schemes, first studied in [187]. The high concentration of Er3+ ions in the host provides good absorption of flashlamp pumping. There exist three possibilities for the energy flow towards the upper laser level 5 I7 of Ho3+ ions. The first one is depicted in Fig. 11. After two stages of cross-relaxation within the Er3+ and Tm3+ ions the excitation is accumulated in the levels 4 I13/2 of Er3+ ions and 3 H4 of Tm3+ ions and is further transferred to the upper laser level 5 I7 of Ho3+ ions. This process implies a quantum efficiency of three. Another possibility is limited to only one stage of cross-relaxation with the subsequent nonradiative transition 4 I9/2 →4 I13/2 and the final energy transfer to the upper laser level like in the previous case. This process implies a quantum efficiency of two. Finally, the third possibility includes the nonradiative cascading transfer of energy to the 4 I13/2 level of Er3+ ions with the following energy transfer to the upper laser level of Ho3+ ions. The latter process has a quantum efficiency of unity. In [188] it was shown that the thermal load in this crystal is decreased by a factor of three and that the only really working scheme is the one depicted in Fig. 11 having a quantum efficiency of three. With the advent of powerful pump diodes this scheme became less attractive for CW operation, because it is possible now to pump directly into the Tm3+ ions at 780–800 nm, achieving over 50% efficiency. However, this energy transfer scheme is very interesting

Crystalline Lasers

279

Fig. 11. Energy level diagram of Tm3+,Ho3+:BaY2 F8 . Demonstration of the crossrelaxational mechanism to populate the upper-laser level in Ho3+ ions

for another reason. Recently, Barnes and colleagues [189] made use of this energy transfer scheme and demonstrated an extremely simple and elegant wavelength selection method in Cr3+,Er3+,Tm3+,Ho3+:YAG allowing tuning to one of the wavelengths, corresponding either to 2.1 µm or 2.9 µm (for details see Sect. 5.3.1). Antipenko was the first to suggest the concept of the Cr, Tm, Ho:YAG laser (Fig. 12), where the chromium ions absorb the pump radiation and transfer it to the thulium ions with the subsequent exchange of one excitation for two [34] and the energy transfer of the electron excitation to the upper laser level 5 I7 of Ho3+ ions [190]. The crystals allowing realization of this concept include besides Y3 Al5 O12 (see e.g., [190,192,193,41,42] such crystals as Gd3 Sc2 Ga3 O12 (GSGG) [194,195], Gd3 Sc2 Al3 O12 (GSAG) [196], Y3 Sc2 Ga3 O12 (YSGG) [197], Y3 Sc2 Al3 O12 (YSAG) [198], mixed scandium garnets [199], Lu3 Al5 O12 (LuAG) [186,200] as well as LiYF4 (YLF) [201,202]. All these crystals are characterized by the high pump efficiency of the broadband visible pump sources as well as the high conversion efficiency of the absorbed light into the inversion population of the upper laser level of Ho ions. The first is explained by the use of Cr3+ as a sensitizer (donor) ion. The absorption bands of Cr3+ (together with Tm3+ and Ho3+ absorption bands)

280

Irina T. Sorokina

Fig. 12. The excitation scheme of the Ho3+ upper-laser level in Cr3+, Tm3+, Ho3+:YAG

cover most of the visible range. This also lowers the requirements to the pump source. Due to the relatively low crystalline field in these crystals and according to the corresponding Tanabe–Sugano [49] diagram the energy gap ∆ E between the levels 2 E and 4 T2 of chromium is low (≤ kT at room temperature). The latter combined with the high concentration of Tm3+ ions (nTm = 3–15 at.%) makes the energy transfer to Tm ions efficient: (1)

2

E → 4A2 (Cr3+ ), 3 H6 →

3

F3 (Tm3+ ) .

The high Tm concentration also favors efficient cross-relaxation with a quantum efficiency close to two [34]: (2)

3

F4 →

3

H4 , 3 H6 →

3

H4 (Tm3+ ) ,

and nonradiative energy transfer to the upper laser level of Ho ions [202,196]: (3)

3

H4 →

3

H6 (Tm3+ ), 5 I8 →

5

I7 (Ho3+ ) .

Like in the previous case of Er lasers the cross-relaxational mechanism of energy transfer allows considerable decrease of the thermal load in the active medium. Combined with the increased capability of the medium to absorb the flashlamp pump this leads to the increase in the slope efficiency of the triply doped crystals in comparison to the crystals which are singly doped with Ho ions. As in the case of the previous scheme, direct diode-pumping into the 3 F3 level of the Tm3+ ion is much more efficient than flashlamp pumping, and the Cr,Tm,Ho-doped laser is used only in pulsed applications. The multitude of these existing energy levels in Cr,Tm,Ho-doped systems provides many possibilities not only for the discussed energy transfer processes, but also for parasitic energy flow channels such as the energy transfer

Crystalline Lasers

281

from Tm to Ho ions via the upper energy states: (1)

3

F4 →

3

H6 (Tm3+ ), 5 I8 →

5

I4 (Ho3+ ),

as well as for two upconversion processes, including the resonant one: (2)

5

I7 →

5

I8 , 5 I7 →

5

I6 (Ho3+ ),

and the nonresonant one: (3)

3

H4 →

3

H6 (Tm3+ ), 5 I7 →

5

I5 (Ho3+ ).

These processes are not marked in Fig. 12 for the sake of simplicity. However, they have been extensively studied in the literature [203,204,205]. Their influence on the laser output characteristics is different depending upon the operation regime. It is therefore important to carefully take them into account, when choosing the optimum host or the optimum concentration of the active ions [206]. Summarizing, in this section the most important energy transfer processes and the population dynamics of the rare-earth-doped 2- and 3-µm lasers have been considered. The multistep pumping schemes of the upper laser levels have been shown to be efficient and useful for creation of inversion population in the mid-IR. 3.2

Influence of Spectroscopic Parameters on Laser Properties

In the previous paragraphs the processes influencing formation of the inversion population in mid-IR crystalline solid-state lasers were introduced. In this section the effect of these processes on laser operation, i.e. on laser threshold and efficiency, will be analyzed. The case of continuous-wave operation will be considered as the most demanding one. In this case assuming the four-level scheme the balance equation for the population of the upper laser level n can be written as [207]: n dn = F pump σGSA (Nt − n) − F las σem n − − αn2 , dt τ

(12)

where the first three terms stand for pumping, stimulated emission and spontaneous decay, respectively. The last term corresponds to upconversion with the upconversion macroparameter α. Nt is the total concentration of the active ions, n is the excited ion concentration, σGSA is the GSA cross-section at the pumping wavelength (pump absorption cross-section), σem is the emission cross-section at the laser wavelength, F pump is the pumping photon flux, F las is the intracavity emission photon flux, and τ is the temperature dependent lifetime of the upper laser level. Note that the ESA terms do not enter (12) under the assumption that the quantum yield from the upper states to the upper laser level is equal to unity, and therefore ESA, even if it is present, does not change the excited ion concentration. If the upper states have long

282

Irina T. Sorokina

lifetimes and possess different decay routes, then the laser schemes and rate equations become quite complicated (see, e.g. the chapter by Pollnau and Jackson in this volume). The rate equation for the photon flux at the laser wavelength can be written as: c dF las F las c las = F las σem nla − − F las σESA nla , dt l τc l

(13)

where the first two terms on the right-hand side stand for stimulated emission and loss, while the last term describes the ESA. The photon lifetime in the resonator is τc = l/(T + L)c, where l is the effective cavity round-trip length, and (T + L) are the logarithmic round-trip losses, consisting of the output coupling (T ) and other losses (L). From (13), the threshold population n th can be determined [207]: n th =

T +L , las ) la (σem − σESA

(14)

where la is the active medium length. The threshold absorbed pump power in longitudinally pumped configuration is then given as [207]: hν las λ las A pump σem τ λ pump −1 pump las n th σESA σESA × (1 + αn th τ ) 1 − 1+ σem σGSA (Nt − n th )

abs = (T + L) P th

(15)

which is proportional to the losses (T + L), the saturation intensity (hν/στ ), the Stokes shift λ las /λ pump , and the pumped area A pump . The last three terms represent the upconversion, ESA at the laser wavelength, and ESA at the pump wavelength, respectively. From the same equations one can obtain the expression for the slope efficiency with respect to the absorbed pump power: −1 pump las σESA n th λ pump T σESA , (16) 1 + ηslope = 1 − pump λ las T + L σem σESA (Nt − n th ) where the first two terms represent the Stokes shift and output coupling efficiency, while the last two stand for the efficiency loss due to the ESA at the laser and pump wavelengths, respectively. For completeness, we note that the quantity las λ pump σESA 1− (17) η0 = λ las σem is usually referred to as the intrinsic slope efficiency. As follows from (15) and (16), upconversion and temperature reduction of the active ion lifetime increase the laser threshold but do not affect the slope efficiency. To keep upconversion losses at minimum (αn th τ 1), one should

Crystalline Lasers

283

keep overall losses T +L low to reduce the threshold population (14). It is also seen from (15) and (16) that losses and ESA affect both threshold and slope efficiency. However, ESA at the pump wavelength is relatively unimportant in four-level lasers, where n th < Nt , and its influence can be further decreased by lowering losses (14). The expressions (12)–(16) can be readily generalized for the case of three- and quasi-three level systems. The effect of reabsorption on laser parameters is rather ambiguous. As has been discussed in Sect. 3.1.3 a straightforward consequence of reabsorption is the lengthening of the lifetime. According to (15) this leads, in the first approximation, to the effective decrease of the laser threshold. On the other hand, as discussed in Sect. 3.1.3, reabsorption increases the quantum yield of the energy transfer processes, and in particular, that of upconversion. In laser systems, where upconversion represents a loss mechanism, the effect of reabsorption is the same as described above for upconversion. However, in other laser systems that use upconversion pumping, e.g. 3-µm Er lasers, reabsorption finally leads to an improvement in efficiency. 3.3

Reaching the Threshold

It is an easily noticeable fact that laser transitions in the mid-IR tend to be broadband, even those in rare-earth ions. Indeed, the widths of the individual states (in energy units) are determined by the particular crystal-lattice interaction and inhomogeneous broadening, which are characteristic for the given ion and host. The transition bandwidth h∆ ν (in energy units) is then the sum of the widths of the upper and lower states. Since the transition energy h∆ ν is inversely proportional to the wavelength, the relative bandwidth of the transition ∆ ν/ν ≈ ∆ λ/λ increases towards the infrared. This is well illustrated in Fig. 1 on the example of room-temperature Tb3+ ion luminescence around 5 µm in KPb2 Cl5 [208]. The transition bandwidth is one of the most important characteristics, determining the ability of the laser material to lase and to be tuned. Another two characteristics of interest, which will be discussed here, are the transition lifetime and cross-section. Both are in one way or another connected with the bandwidth of the laser transition. The treatment of absorption, spontaneous and stimulated emission using an elegant thermodynamic treatment given first by Einstein [112] can be found elsewhere (see, e.g. [113,114]). Based on this treatment the spontaneous-emission cross-section σem (λ) can be obtained from the fluorescence intensity signal I(λ) using the relation: σem (λ) =

I(λ)λ5 A , 8πcn2 I(λ)λdλ

(18)

where A is the full spontaneous emission probability (radiative decay rate) from the upper laser level, c is the speed of light and n is the index of refraction. The measured fluorescence intensity signal I(λ) is usually corrected to

284

Irina T. Sorokina

the spectral sensitivity of the recording system, using, e.g. a calibration lamp. Formula (18) follows from basic principles [27,38] and does not contain assumptions other than neglecting the polarization dependence of fluorescence that can also be correctly accounted for. In practice, taking the inverse lifetime at low temperature as the best-known approximation for the radiative decay rate, one usually obtains the effective emission cross-section σem (λ). One of the direct consequences of the ion–host interaction discussed in the previous section is nonradiative decay. The usually measured luminescence lifetime τ consists of purely radiative and nonradiative components: 1 1 1 = + . τ τ rad τnr

(19)

From (18) the radiative lifetime of the broadband transition has the form 1 1 ∆λ σem (λ)dλ = A = 8cn2 ≈ 8cn2 σem (λ0 ) 3 , (20) 4 τ rad λ λ0 λ0 clearly demonstrating the general tendency that lifetime inversely scales with emission bandwidth. We can see from (15) that the minimum absorbed pump power density at threshold is proportional to the loss coefficient and the gain saturation intensity. The necessary incident pumping intensity is therefore inversely proportional to the pump absorption coefficient: I th ∝

αloss hν , α abs σ(ν)τ

(21)

where αloss may include the absorption from the ground state in the case of three- and quasi-three-level schemes. Combining (20) and (21), the following scaling rule for the required pump intensity at threshold can be obtained: I th ∝

1 ∆ λ n2 1 , F OM η QE λ0 λ40

(22)

where η QE is the quantum yield of the transition, defined as: η QE =

τ . τ rad

(23)

One important conclusion is that the threshold power density scales with the bandwidth. This explains the relatively high threshold, characteristic for broadband materials. Only the progress in pump sources, when laser pumping became available, made possible laser action of broadband materials. As a result, the first argon-laser pumped tunable Ti:sapphire appeared [191,38], followed by a number of exciting developments in femtosecond pulse generation [210,211,212,213]. Nowdays there exists only a limited number of directly diode-pumped broadband solid-state lasers. These include: Cr:LiSAF[214] and Cr:LiSGaF [207,215], Cr:YAG [96], and since recently also Cr:ZnSe [97]

Crystalline Lasers

285

and Cr:ZnS [98]. Due to the relatively high pump saturation intensity of Ti:sapphire, the diode-pumping of this material will remain a severe problem, even if high-power blue-green laser diodes become available. The analysis of the formula (22) shows also that there is a factor λ4 in the denominator, which effectively decreases the threshold for longer wavelength transitions. For mid-IR this is a very encouraging factor. Provided that the active material has a high FOM and high quantum yield (like it is in the case of e.g., Cr:ZnSe) this makes direct diode-pumping in the mid-IR feasible. To get a feeling for what difference the λ−4 factor can produce, compare, e.g. otherwise quite similar Ti:sapphire (λ0 = 780 nm) and Cr:ZnSe (λ0 = 2450 nm): the threshold intensity differs by two orders of magnitude!

4

Laser Design Considerations

In this section the typical experimental techniques will be reviewed, which are used to create a working laser: pumping arrangements and cavity schemes as well as tuning, Q-switching, and mode-locking techniques. Most of these are well-known. Therefore the present consideration will concentrate on features specific to mid-IR lasers. 4.1

Pump Sources

For optical pumping of solid-state lasers one can use a gas discharge lamp, high-power laser diodes, or another laser at shorter wavelength. As described in the Introduction, flashlamp pumping of mid-IR lasers has been used from the very first demonstrations. The high-brightness noble gas discharge lamps emit most of their energy in the visible and near-infrared region of spectrum. Thus, the flashlamps are useful only for laser media with sufficient absorption in this region and all the energy difference between the pump photons in the visible and the emission photons in the mid-IR (Stokes shift) has to be dissipated in the medium as heat. Many rare-earth ions have absorption lines, suitable for flash lamp pumping. However, these absorption lines only partially overlap with the total emission spectrum so that the absorption efficiency remains low. Some improvement in this respect can be achieved by using energy transfer schemes (see Sect. 3.1.3), with the first absorbing ion having broad bands in the visible (like, e.g. chromium). A point in case is the Cr,Tm,Ho:YAG crystal [216], which has successfully entered commercial production. However, while sensitization solves the problem of spectrum overlap, it does not reduce the Stokes shift and excessive heat production remains an important issue. High-power diode lasers have a spectrum bandwidth of ∼ 2–4 nm that can be set to match the absorption band during diode fabrication. This significantly improves the pumping efficiency. Continuous-wave diode laser arrays have output power per unit length comparable with arc lamps, but emit

286

Irina T. Sorokina

partially coherent light in a limited solid angle while the lamps emit incoherent light in all directions. With proper illumination optics, diodes allow much higher pumping intensity, thus decreasing the pumping threshold power of solid-state lasers. At the same time, efficient high-power diode lasers exist only in certain wavelength ranges and cannot always be used for direct pumping of some ions. Most of the high-power laser diodes emit in the nearinfrared. The Stokes shift and the corresponding heating are thus somewhat lower than in the case of lamp pumping. For pulsed laser operation, pump diodes perform worse than the flashlamps. The reason for this lies in the different damage behavior of diodes and flashlamps. The output power limit for the laser diode is set by the optical damage of the emitting surface and is practically the same for pulsed and CW operation. On the contrary, the maximum flashlamp pulse energy scales as the square root of the pulse duration. For a 100-µs pump pulse, this would mean a 100-fold increase in the peak power as compared to the CW case. This compensates for all the disadvantages of flashlamps as compared to laser diodes, and makes flashlamps still competitive in high-energy pulsed applications. In the case of active media with long upper level lifetime, like e.g. Ho- and Tm-doped crystals with τ ≈ 10 ms, pump diodes can be driven at much lower peak power and are preferred to flashlamps. Optical pumping by a high-quality beam of another laser, while increasing the complexity and cost of the system, has certain advantages. First, the coherent laser output can be much more easily focused to create a very high pumping intensity, enabling laser action of materials with high laser threshold. Second, laser pumping allows operation of materials with absorption bands, which cannot be directly pumped by lamps or laser diodes. And finally, multistage pumping decreases the Stokes heat production in the laser medium. This can be understood if one recalls that primary pump photons are generated in the visible (lamps) or near-IR (diodes) spectral regions. With the final laser wavelength lying in the mid-IR, most of the photon energy should dissipate as heat. As pointed out in Sect. 3.1.1 long-wavelength media are especially temperature-sensitive due to the influence of thermally activated nonradiative decay. In a multi-stage scheme a significant share of the Stokes heat can be dissipated in the pump laser, thus helping to keep the heating in the final stage of the crystal low. 4.2

Cavity Design

The cavity design guidelines for solid-state mid-IR lasers are the same as for any other spectral region [114,30,113]. An important difference however is that in many cases the atmosphere cannot be considered transparent because of the absorption by water and carbon dioxide molecules. In these cases the cavity should be kept as short as possible or purged (filled) with dry nitrogen or inert gases.

Crystalline Lasers

287

The power scaling approaches for solid-state lasers in the mid-IR are also generally the same as in other spectral regions [30]. The discussion issues include efficient pump concentration and delivery, crystal cooling, thermal lens and induced birefringence compensation, as well as crystal fracture prevention. For high-energy pulsed systems, the optical damage of the cavity elements is an important factor, even if the average power remains moderate due to the low pulse repetition rate. High-power laser diodes are typically 1–2 cm long arrays of single emitters on a heatsink, with an output power of up to 40–50 W per cm of length. The radiation divergence of a single emitter is about 12g degrees in the array plane (slow axis) and 35g degrees in the perpendicular plane (fast axis). The advanced heatsink configuration allows stacking such arrays with a pitch of 1–4 mm to form a rectangular emitting area with a surface brightness of 100–400 W/cm2 . This is about an order of magnitude less than the threshold pump power density of typical solid-state media. A number of techniques have been developed to increase the pump power density. In fiber coupled diodes, every single emitter is launched into a fiber. The fibers are then combined together to form a bundle, or a single thick multimode fiber with high numerical aperture. One can use conical lens ducts [217], which concentrate the light from the large rectangular input cross-section into the small round output cross-section using the total internal reflection from the guiding surfaces (Fig. 13). A popular approach is to collimate every array by a cylindrical microlens to reduce the large divergence in the fast axis. The output is then a collimated elliptic beam with large aspect ratio, and can be imaged into the active medium, or geometrically multiplexed with the other such beams. One of the methods is to arrange the diode arrays in a tight manner around

Fig. 13. Schematic diagram showing high-power laser diode collimation and delivery via lens duct (after [218])

288

Irina T. Sorokina

the rod, making all surfaces reflective except for the array apertures, thus assuring multiple-pass absorption in the rod. Because cooling of the crystal occurs through the surface, while heat production takes place in the volume of the crystal, the key to effective cooling of the high-power laser crystal is in the short thermal path from the heat source to the cooled surface and the high surface-to-volume ratio. These requirements can be best achieved if the active medium is a rod with a small diameter, or a thin slab. In the more traditional rod design the temperature distribution has rotational symmetry, with the highest temperature achieved at the rod axis. When the system is driven at full power, the temperature gradient can be quite high and the rod acts as a lens with high dioptric power. While the thermal lens can be partially compensated in a well-designed cavity, the ultimate limit to the absorbed pump power is set by the thermal fracture that starts at the rod faces. Modern high-power designs therefore use rods with pieces of undoped material (typically 1–2 rod diameters long) at both ends of the rod (end-caps) that also improve the thermal lensing properties [219,220]. The slab geometry comes in two important modifications – with the light propagating across or along the slab. In the former case, one can take advantage of the fact that the heat flow to the surfaces is coaxial with the light, thus reducing by orders of magnitude the thermal lens and depolarization. At the same time, the active medium length is small and the gain is correspondingly low. Another important issue is that the cooling system should be transparent to the beam, at least on one side. The solution is a so-called thin-disk laser design [221,222] (Fig. 14), that was initially invented and developed for the Yb:YAG, and later successfully applied in the infrared to the Tm:YAG [223] and Cr:ZnSe [346]. For the pump power delivery, convenient and maintenance-friendly fiber-coupling can be used. When the light is propagating along the slab (including zig-zag propagation), cooling through large faces becomes very efficient. However, for pump delivery one has to use side-pumping techniques, like e.g. for a 12-W TEM00 Tm:YLF laser [224]. In the mid-IR range, high-power systems have been pursued with Tmand Ho-doped crystals. The highest CW output power of 115 W [225] has

Fig. 14. Schematic diagram of a thin-disk laser [221]

Crystalline Lasers

289

been demonstrated in Tm:YAG using a rod design with endcaps and diode pumping via lens ducts. In pulsed mode, energies up to 17 J have been reported [226]. As a final comment it should be noted that high-power laser engineering is a very expensive and involved research field, which is application-driven. Most applications for lasers with multi-hundred-Watt output power are in material processing, where the main goal is to achieve the highest possible concentration of energy. Since the focusing spot size scales with wavelength, lasers operating at shorter wavelength have a significant natural advantage and most of the efforts in high-power laser engineering are put into 1-µm systems, based on Yb and Nd ions. There is no doubt, however, that development of mid-IR high-power lasers will proceed at a much faster pace, once the applications demanding the particular wavelengths emerge. 4.3

Tuning Methods

A number of tuning methods have been developed to control the laser wavelength. These include: prisms, gratings, birefringent filters, étalons, and acoustooptic filters. A good review can be found e.g. in [228]. Prism tuning is quite easy to implement, and the material can be easily chosen – any material, transparent in the region of interest will do. Two typical schemes are shown in Fig. 15a and b (Littrow configuration). If the prism is set to operate at the Brewster angle of incidence and if the laser itself is linearly polarized, then the insertion losses can be exceptionally low (< 0.1%). In this case no surface coating is needed, as the reflection losses stay very small in a quite broad vicinity of the Brewster angle. Yet another advantage of the prism-tuning scheme is that there is no intrinsic bandwidth limitation, like in most interference-based schemes. At the same time, the wavelength selectivity of prism tuning is rather low. The relative acceptance bandwidth (full-width at half-maximum, FWHM) is given by the formula [228] −1 dn 1 ∆λ = , (24) λ 4πw0 dλ where w0 is the beam waist radius. The bandwidth is twice as much as in the Littrow configuration. With a beam radius of 1 mm at λ = 2.5 µm this corresponds to approximately ∆ λ ≈ 50 nm using a ZnSe prism (dn/dλ = 7.9 × 10−3 µm−1 ) or ∆ λ ≈ 7.5 nm using a Ge prism (dn/dλ = 54 × 10−3 µm−1 ). Grating tuning also comes in two popular implementations (Fig. 15c,d). The relative acceptance bandwidth in the grazing incidence scheme is given by the formula [228] λ ∆λ = , λ 4πw0 tan θ

(25)

where θ is the angle of incidence. The bandwidth is twice as much in the Littrow configuration. The tuning condition is given by sin θ = λ/2d, where

290

Irina T. Sorokina

Fig. 15. Schematic representations of tuning arrangements with mechanical tuning: prism (a), Littrow prism (b), grazing incidence grating (c), Littrow grating (d), and birefringent filter (e)

d is the grating period. According to (25), the relative acceptance bandwidth thus linearly increases with wavelength. For the system in the previous example we get ∆ λ ≈ 0.44 nm with a grating period of d = 600 grooves/mm used in a grazing incidence scheme or ∆ λ ≈ 2.5 nm with a grating period of d = 300 grooves/mm used in the Littrow configuration. The grating scheme is thus about an order of magnitude more selective than a prism. The selectivity of both schemes can be further improved by expanding the beam size w0 . At the same time, losses in gratings are much higher than in prisms. Even though the reflectivity of metal-coated gratings can be quite high in the infrared, the diffraction efficiency is typically 0.9–0.95 except for the narrow region around the blaze angle. Grating tuning has been extensively used in mid-IR color-center lasers. Birefringent (Lyot) filters (Fig. 15e) make use of the fact that the phase retardation of a birefringent plate is wavelength-dependent [229,230,231]. If the cavity is polarization-sensitive, the losses will become sine-modulated with a period ∆ λ FSR ≈

λ2 , d cos θi |no − ne |

(26)

where d is the thickness of the plate, no and ne are the ordinary and extraordinary refractive indices, respectively, and θi is the incidence angle inside the plate. As with prisms, if the plate is put at the Brewster angle, then insertion losses become negligibly small, no coating is required, and the Brewster surfaces of the plate act as a polarization discriminator. The actual tuning

Crystalline Lasers

291

occurs by rotating the plate in its plane (thus changing the value of ne ). To make a filter, any transparent birefringent material will do. The wavelength acceptance ∆ λ = ∆ λ FSR /2 can be freely set by choosing the thickness d and for a Brewster-oriented 5-mm thick plate at 2.5 µm wavelength can be ∆ λ ≈ 95 nm in sapphire (|no − ne | ≈ 0.0075) and ∆ λ ≈ 5 nm in TeO2 (|no − ne | ≈ 0.14). To cover all the available bandwidth the free spectral range of a filter should be larger than the gain bandwidth of the medium. In the case of broadband media, this would imply low wavelength selectivity. One therefore uses a stack of birefringent filters (usually two or three) with different thickness to ensure the narrow-line operation in a broadband medium. When going further into the infrared, keeping the d/λ ratio constant ensures the wavelength-independent selectivity ∆ λ/λ. ´ Etalons are short Fabry–Pérot cavities placed inside the resonator. They have maxima separated by ∆ λ FSR ≈

λ2 , 2nd cos θi

(27)

where d is the etalon thickness and n is the refractive index of its body. The acceptance bandwidth is given by ∆ λ = ∆ λ FSR /F , where the finesse √ F = π R/(1 − R) can vary between 10 and 1000 depending on the reflection coefficient of the mirrors R. At normal incidence and 2.5 µm the 1-mm CaF2 étalon with R = 99% mirrors will have ∆ λ FSR = 2.1 nm and ∆ λ = 0.007 nm. To keep high finesse, the étalons operate at small incidence angles and thus allow only very limited tuning by changing the incidence angle. They provide however the strongest wavelength selectivity of all other methods and allow mode selection down to single-frequency operation. As in the case of birefringent filters, the high selectivity at any wavelength can be achieved by scaling the thickness d with λ. Tunable acousto-optical filters (AOTF), first used for dye laser tuning in 1971 [232,233], have the convenience of direct electronic control of the laser wavelength, without mechanically moving parts, required by all the tuning schemes described above. In the typical so-called non-collinear scheme (requiring linearly polarized light) the relative acceptance bandwidth of the AOTF scales as [228] λ ∆λ ≈ , λ L|no − ne |

(28)

where L is the interaction length. As with birefringent filters and étalons, the relative bandwidth is proportional to the ratio λ/L. Since, however, the interaction length is of the order of a centimeter for practical reasons, the relative bandwidth progressively increases with wavelength, thus making the filter less selective in the mid-IR. Using a typical TeO2 material (transparency range up to 5 µm) with |no − ne | ≈ 0.1 at 2.5 µm, we obtain ∆ λ ≈ 6 nm. An important advantage of the AOTF is the possibility of tuning over wide wavelength ranges, limited in practice by the electric matching of the driver

292

Irina T. Sorokina

to the transducer. The wavelength dependency of the diffraction efficiency can be written as η∝

PRF M2 , λ2

(29)

where PRF is the applied RF power and M2 is the acousto-optical figure of merit (M2 = 35 × 10−15 s3/kg for TeO2 ). The required RF power rapidly increases with wavelength, reaching ∼ 10 W levels above 2 µm, making active cooling of the device necessary. Unfortunately, such excellent acousto-optic materials as GaAs (M2 = 100 × 10−15 s3/kg) or Ge (M2 = 840 × 10−15 s3/kg) cannot be used as described, because they are not birefringent, making this type of tuning not very prospective at wavelengths above 2.5 µm. It should also be noted that the overall transmission efficiency of the AOTF is less than 100%, being typically 90–95% in a single pass. AOTF tuning is therefore more useful for pulsed lasers. Finally, tuning can be achieved by an electro-optic modulator used as a birefringent filter. In this case, applying the voltage changes the phase shift |no − ne | L, thus shifting the transmission maximum, whereas the double half-wave voltage Vλ/2 (typically in the kilovolt range) corresponds to tuning over the whole free spectral range. As the half-wave voltage scales with the wavelength, this type of wavelength tuning additionally needs progressively higher control voltages with increasing wavelength. 4.4

Q-switched and Mode-Locked Operation

Most of the modulation techniques used in the visible and near-IR ranges can be applied for Q-switching and mode-locking of mid-IR solid-state lasers. The main criterion is the correct choice of the working material, which should be transparent at the wavelength of interest and fulfill other design criteria like damage threshold, environmental stability, etc. Electro-optical modulation is probably the most widely used technique for active Q-switching, because there exist a number of technologically welldeveloped crystals, transparent in the mid-IR and possessing a high electrooptical coefficient [235]. For the electro-optic modulator, the half-wave voltage scales as Vλ/2 ∝ λ/rn3 , where rn3 is the effective electro-optic coefficient of the crystal for the chosen geometry, which does not change strongly with the wavelength. The control voltage thus scales approximately proportional to the wavelength. For example, a lithium-niobate based electro-optic modulator would require, for a Er laser wavelength of 2.9 µm, about 3.3 times higher control voltage than for the Nd laser wavelength of 1.06 µm. Electro-optical Q-switching has been demonstrated with many mid-IR solid-state materials with long upper-level lifetime: Co:MgF2 [236], as well as Ho- [237], Tm- [238], and Er-doped [239,240,241] crystals. Acousto-optic modulators also require higher control power with longer wavelength, but the dependence is quadratic on λ, as seen from (29). Using

Crystalline Lasers

293

the same example of Nd and Er lasers, the wavelength change from 1.06 to 2.9 µm corresponds to a 7.5-fold increase of required acoustic power. However, as already mentioned earlier, in the mid-IR there exist materials with an exceptionally high figure of merit that can offset the wavelength dependence. As the typical schemes for acousto-optic mode-lockers and Q-switchers do not require birefringence, cubic crystals with high figure of merit like GaAs or Ge can be implemented. Acousto-optic Q-switching has been realized at 2 µm in Tm:YAG [243], Cr,Tm:YAG [227], and Tm,Ho:YLF [244] lasers, and at 2.8–2.9 µm in Er:YAG and Er:YSGG [245,246] lasers. Acoustooptic mode-locking was demonstrated in Co:MgF2 and Ni:MgF2 lasers [262], Tm:YAG [265], Cr,Tm:YAG and Cr,Tm,Ho:YAG [263], Cr,Tm,Ho:GSAG:YSGG [264], and Cr:ZnSe lasers [266,267]. The other active Q-switching techniques in use include the rotary mirror, and frustrated total internal reflection (FTIR) [247] methods. These techniques have an advantage of being rather wavelength-independent and have been successfully implemented in Er:YAG and Er:YSGG lasers in the 2.70– 2.94 µm wavelength range [248,249,250,251]. Passive Q-switching techniques, based on saturable absorbers, require an appropriate absorbing material with low saturation intensity at the lasing wavelength. Using Cr2+:ZnSe as an absorber, Tm:YAG at 2.02 µm and Ho:YAG at 2.09 µm have been passively Q-switched [252]. In this case, the absorber acted at the edge of its absorption spectrum. However, the large difference in saturation cross-sections between the short-living Cr2+:ZnSe and longliving Ho:YAG and Tm:YAG allowed the passive Q-switching regime. Using the same absorber material in the coupled cavity, a Co:MgF2 laser has been passively Q-switched at 1.76 nm [253]. Spectacular results have been achieved with Er:YAG (at 2.94 µm) and Er:YSGG lasers, passively Q-switched using thin layers of water and ethanol [258,259], or even a soap film [261]. A Cr,Yb,Ho:YSGG laser, operating at 2.94 µm, has also been Q-switched by water absorption [260]. In these cases, one takes advantage of the strong absorption band of the OH vibration, but the saturation mechanism is still under investigation. An appropriate absorber material can also be engineered as a semiconductor bulk material with required band-gap energy or as a quantum well. This approach can be especially useful when a classical ion-doped saturable absorber does not exist at the desired wavelength. Experimental demonstration of this technique in the mid-IR has been performed in the Er:YSGG laser at 2.8 µm [256,257,249] and in Cr,Yb,Ho:YSGG at 2.94 µm [260]. The technique of optimizing the absorption wavelength by using quantum confinement effects has been demonstrated in the Ho3+-doped YAG laser[254,255] at 2.1 µm using a glass doped with PbSe quantum dots of ∼ 8.5 nm average diameter. The passive mode-locking techniques for solid-state lasers include: synchronous pumping [113,114], Kerr-lens mode-locking (KLM) [268,113], and semiconductor-based saturable absorbers [269]. The synchronous pumping

294

Irina T. Sorokina

technique is the mostly effective in media with short upper-laser level lifetime. It has therefore been used in the mid-IR only with short-living colorcenter lasers [270,271]. In the KLM technique, the crucial parameter is the ratio of the intracavity power to the critical power for self-focusing Pcrit ∝ λ/nn2 [272]. The latter increases with the wavelength. In order to realize the KLM regime in the mid-IR one has therefore to use media with significantly higher nonlinear refractive indices than the well-known oxide and fluoride hosts, working in the visible and near-IR regions. Semiconductor-based media like Cr:ZnSe and Cr:ZnS possess a very high nonlinear refractive index n2 , making the KLM-based passive mode-locking regime feasible. At the moment, a pulse duration of about 4 ps has been demonstrated in Cr:ZnSe [267]. The potential of generating few-optical cycle pulses in this material waits to be realized. Finally, mode-locking by a saturable absorber requires the existence of an absorbing medium matching the gain wavelength of the laser material and possessing very short lifetime and low saturation intensity. Semiconductorbased saturable absorbers provide a very promising way for passive modelocking, because they allow matching of the absorption wavelength by changing the composition and using quantum confinement effects. To date, this type of passive mode-locking has been demonstrated in a color-center laser at 2.8 µm using a HgCdTe quantum well [273] as well as in the Cr,Er:YSGG laser using an InAs saturable absorber [256].

5

Mid-Infrared Lasers

Broadband crystalline continuous-wave lasers have always played a key role in mid-IR solid-state photonics as primary tunable laser devices. Their advantages over other tunable laser sources include high-power room-temperature operation retaining good beam quality and spectral linewidth, as well as simplicity, stability, efficiency and compactness, when being pumped with diode lasers. Furthermore, the broad gain bandwidth of some laser crystals allows generation of ultrashort pulses. These lasers are also most suitable for shifting the wavelength from 2 to 5 µm further into the infrared by using either parametric (see the chapter by Vodopyanov), difference-frequency mixing (the chapter by Fischer and Sigrist) or Raman (the chapter by Basiev) processes in infrared-transmitting nonlinear crystals. This section reviews the tunable lasers based on the established and new transition-metal (TM) ion doped crystals as well as those used in laboratory practice color-center crystals. The emphasis is being made on novel TM2+doped crystals as sources of ultrabroadband radiation and potentially ultrashort pulses in the mid-IR. It is worth noting that among transition-metal ions only divalent ions proved to have lased in the infrared. Finally, an alternative class of tunable alkali-halide crystalline lasers employing color centers (F-centers) is considered.

Crystalline Lasers

5.1

295

TM2+-Doped Lasers

Historically crystalline TM2+-doped lasers replaced tunable near-IR dye lasers. The absence of the flowing dye jet greatly enhances the laser stability, reducing the noise by a factor of 10, and allowing linewidths down to 1 kHz in a ring cavity. The existing TM2+-ions, lasing on the 3d–3d transitions in the mid-IR, can be divided into two major groups: those ions, which occupy octahedral sites (like, e.g. Ni2+ and Co2+ in halides), and those which can be found in tetrahedral positions (like, e.g. Cr2+, Fe2+, Ni2+ and Co2+ in chalcogenides). As discussed earlier in Sect. 3.1.1 the electric-dipole transitions of the ions in octahedral sites (possessing inversion symmetry) are parity forbidden and have therefore generally low oscillator strength and long lifetime. The ions in the tetrahedral sites, lacking inversion symmetry, are characterized by the high oscillator strength and short lifetime. The other distinguishing feature of the TM2+-ions in tetrahedral sites is the relatively low crystal-field splittings, placing optical transitions further into the infrared. 5.1.1

Ni2+-, Co2+-Doped Lasers

Remarkably the first tunable solid-state laser to be demonstrated was based on the divalent TM ion in octahedral sites: Ni2+ (3d8 configuration) in MgF2 [11]. The cryogenically cooled and flash lamp pumped laser operated around 1.62 µm. This publication stimulated a lot of work on Ni2+ and Co2+doped crystals [16,274,14,275,276]. One year later laser operation from Co2+ ions in MgF2 and ZnF2 at various wavelengths between 1750 and 2160 nm was obtained [12]. In 1966 the first continuous-wave operation using a CW tungsten lamp was demonstrated in Ni2+:MgF2 along with the laser operation of Ni2+ and Co2+ in MgO, MgF2 and MnF2 [13]. The performance of the Ni2+:MgF2 laser was substantially improved with the use of the CW Nd:YAG laser as a pump source [16,277]. Cooling the crystal down to 77 K up to 2 W of the output power at 28% slope efficiency was achieved. However, this level of slope efficiency was a factor of two lower than predicted. The difference was assumed to be due to excited state absorption (16). ESA could also partially explain the limited tunability in Ni2+-doped systems, which was smaller than predicted from the fluorescence spectra. An example of a continuous tuning curve between 1.61 and 1.74 µm is given in Fig. 16. Taken collectively the Ni2+ tunability in various hosts spans the wavelength range between 1.31 and 1.94 µm, and is frequently not continuous [13] due to the inhomogeneous character of the emission spectra at high dopant concentration. The latter is required to be high due to the relatively small cross-section in the case of broadly emitting ions in octahedral sites (see Sect. 3.1.1). MgO is another nearly ideal although difficult to grow host, possessing excellent optical, mechanical and thermal properties (with thermal conductivity higher than that of sapphire). The advantage of MgO over fluoride

296

Irina T. Sorokina

Output power (mW)

100

50

0 1600

1650

1700

1750

Wavelength (nm)

Fig. 16. Tuning curve of the CW Ni2+:MgF2 laser, pumped by a 1.33 µm Nd:YAG laser [278]

hosts is that the 3 A2 →3 T2 absorption band can be pumped by the widely accessible 1.06 µm Nd lasers. Also the lifetime dependence upon temperature is nearly constant up to 400 K, indicating high quantum efficiency. In this crystal at 100 K Moulton has achieved 10 W in the CW regime at 57% slope efficiency [277,275]. However, there has been little further development in Ni2+-doped laser systems. The major disadvantage consists in the necessity of operating these lasers at low temperatures, usually at 77 K. The highest temperature at which CW operation was observed in Ni2+:MgO crystals was 240 K [279]. This restriction is caused by peculiarities of the spectroscopic properties of Ni2+-doped systems (see, e.g. [280,281]). The host crystals, MgF2 , MnF2 , MgO, ZnF2 , KMgF3 , provide the weak crystal field sites for Ni2+-ions, making pumping with 1.06 and 1.32 µm Nd:YAG lasers possible. The infrared emission occurs on the long-lived (of the order of 10 ms at 77 K in Ni2+:MgF2 ) magnetic-dipole 3 T2 →3 A2 transition. At high temperatures nonradiative decay is present in fluoride hosts, decreasing the lifetime and increasing the laser threshold. The ESA in the long-wavelength spectral wing (predicted earlier by Moulton) was reported by Moncorge [279] and Koetke [282] to overlap with the GSA at short wavelengths. Along with the nonradiative decay the latter loss mechanism is the main obstacle on the way towards room-temperature tunable operation of Ni2+-doped systems, including those based on garnets and perovskites [283,282]. Co2+ ions in octahedral sites, which are in many respects similar to Ni2+ ions and have the same 3d8 configuration, have been studied in the same rutile and perovskite fluoride hosts and the same low crystal field octahedral sites as Ni2+ ions, however, with a little bit more success with respect to roomtemperature operation [284,285,286]. The free ion term 4 F splits into 4 T1 , 4 T2 , and 4 A2 levels with the 4 T1 level being the ground state. The absorption bands resulting from the 4 T1 →4 T2 transition around 8000 cm−1 and 4 T1 →4 A2 transition around 18000 cm−1 [287] allow pumping with a 1.32 µm

Crystalline Lasers

297

Luminescence (a.u.)

Nd laser and a 514 nm Ar laser correspondingly (Fig. 6a). The large configurational coordinate offset (see Sect. 3.1.1) of the upper laser level 4 T2 results in a broadband infrared luminescence spectrum (Fig. 17) ranging from 1.5 to over 2.3 µm and corresponding to the 4 T2 →4 A2 transition. The unusually broad bandwidth is explained not only by the vibronic nature of the transition, but also by the fact that the ground state is split into six sublevels spread over 1300 cm−1 , each having an associated vibronic transition [12,289]. The corresponding zero-phonon lines lie between 1.47 and 1.85 µm. As a result, the luminescence spectrum is the sum of the six zero-phonon lines and related sidebands, which can be clearly seen in Fig. 17. ESA in the 4 T2 →4 A2 and 4 T2 →4 T1 transitions in Co2+:MgF2 and Co2+:ZnF2 , although being of less importance than in Ni2+-doped crystals, has been found to be a limiting factor in efficient laser operation of these crystals [285]. ESA also effectively decreases the gain cross-section and leads to a large saturation fluence of 100 J/cm2 [286]. The first Co2+-doped MgF2 and ZnF2 lasers using flash-lamp excitation were reported in 1964 [12], shortly after the first Ni2+-doped laser [11]. Extensive research in subsequent years [13] led to the demonstration by Moulton of Q-switched operation of MgF2 [274,288] and CW operation of MgF2 [277,289,290], KZnF3 [291,292,285] and KMgF3 [293] using laser pumping. Also active mode-locking of the Co2+:MgF2 laser was demonstrated [262]. All these lasers needed liquid-nitrogen cooling to maintain lifetimes in the millisecond range. The main problem of Co2+:MgF2 is the rapid decrease of the upper laser level lifetime with temperature (Fig. 18), which drops from 1.3 ms at 77 K down to 36 µs at 300 K [289]. Similar lifetime behavior has been reported for KMgF3 by Sturge [294] and is assumed to be caused by the increased rate of nonradiative decay due to multiphonon emission [289]. On the positive side is the large thermal conductivity of Co2+:MgF2 being 0.3 W/( mK), Mohs hardness of 6, and the negligibly small dn/dT [17].

1400

1600

1800 2000 Wavelength (nm)

2200

Fig. 17. Luminescence spectrum of Co2+ ions on the 4 T1 –4 A2 transition in MgF2 at 77 K (after [37])

298

Irina T. Sorokina

Lifeteime (ms)

1

0.1

0

50

100

150

200

250

300

Temperature (K)

Fig. 18. Temperature dependence of the fluorescence lifetime of Co2+ ions in MgF2 (after [289])

This allows pumping at high power levels without deterioration of the output characteristics. Rines et al. [286] could generate in the pulsed regime as much as 6.5 W of average output power at 2.05 µm at a repetition rate of 9 Hz and a pump energy of 2.7 J. A cryogenically cooled CW Co2+:MgF2 laser exhibited high energy conversion efficiency of 31% at 1.3 µm, and high output power of 1 W at 1.86 µm [290]. Moulton demonstrated more than 4 W of output power in the TEM00 mode at 1.92 µm in the CW regime at 15 W of pump power [289]. No indication of output power saturation due to heating effects was observed. Recently, Di Lieto realized a high power broadly tunable Co2+:MgF2 CW laser [295]. Using a 12 W Nd:YAG laser at 1.3 µm as a pump source, he could achieve up to 3 W output power in TEM00 mode at 1.67 µm, and 2 W at 1.77 µm. The slope efficiency of 32% compares with the best values reported in other works, and the broad continuous tuning range between 1.6 and 2.1 µm (Fig. 19) makes this laser a practical tool

2XWSXW SRZHU P:

:DYHOHQJWK QP

Fig. 19. Tuning curve of the CW Co2+:MgF2 laser at 77 K, pumped by a multimode Nd:YAG at 1.32 µm (after [295])

Crystalline Lasers

299

for spectroscopic [295] and solid-state laser pumping purposes [165]. In contrast to Ni2+-doped lasers room-temperature pulsed operation is possible in Co2+-doped lasers at the expense of the higher threshold and lower slope efficiency [284,285,286]. However, in this case the active medium should be pumped on a timescale short enough compared to the upper state lifetime of 36 µs. In the pulsed regime a Nd:YAlO3 pumped Co2+:MgF2 laser has been tuned over a broad wavelength range between 1.5 and 2.3 µm [289]. This laser has finally matured to the commercially available laser system [17]. 5.1.2

Cr2+-, Fe2+-Doped Lasers

As shown earlier in Sect. 3.1.1 multiphonon relaxation processes set the fundamental limit for obtaining continuous-wave room-temperature laser operation from vibronic transitions in the mid-IR region. Because of this the majority of the known vibronically broadened laser transitions in the midIR are quenched at room temperature. Those are, e.g. transitions in Ni2+ and Co2+-ions (described in the previous section) or color centers (Sect. 5.2). Unfortunately, alternative sources in this wavelength range based on semiconductor structures often also require cryogenic cooling and are not as broadband as vibronic lasers. Yet there exists a plenitude of applications for roomtemperature tunable mid-IR sources like, e.g. spectroscopic (see the chapter by Richter et al.), medical (see the chapter by Jean and Bende), and remote sensing, etc. During the last several years this has provoked increased interest for long-wavelength, i.e. above 2 µm, tunable directly diode-pumped solid-state lasers. The major breakthrough in this respect came with the invention, in the middle of the 1990s by a group of scientists at the Lawrence Livermore National Laboratory in USA, of a new class of transition-metal doped zinc chalcogenides [45,296,297,46,47]. Shortly afterwards a similar class of transition-metal doped cadmium chalcogenides based on CdSe [300,301] as well as on CdTe and compounds [298,299] was proposed simultaneously by the other two groups. Since that time TM2+-ions have been incorporated into several binary and ternary II–VI compounds, including ZnSe, ZnS, ZnTe, CdSe, CdS, Cd1−x Mnx Te, Cd1−x Znx Te, ZnMgSe and ZnMgSeTe. In all these crystals TM2+ ions occupy low crystalline field tetrahedral sites coordinated by heavy selenide, telluride or sulphide anions. As shown in Sect. 3.1.1, the low maximum phonon frequency in chalcogenides (compare 250 cm−1 in ZnSe and 850 cm−1 in YAG) leads to a decrease of the nonradiative decay rate and an increase of the fluorescence quantum yield. At room-temperature the latter is close to unity in Cr:ZnSe and is comparably high in other chalcogenide materials. This provides Cr:ZnSe with the highest gain among vibronic lasers and enables efficient room-temperature operation. It is not a surprise that in subsequent years Cr:ZnSe drew a lot of attention as a roomtemperature broadly tunable continuous-wave (CW) laser operating around

300

Irina T. Sorokina

Table 2. Material properties of Cr2+-doped laser hosts. If not indicated otherwise, the data are taken from [309] Ti3+:Al2 O3 Cd0.55 Mn0.45 Te

ZnSe

ZnS

CdSe

Crystal structure

cubic

mixed-polytype [334]

cubic, uniaxial

Hardness (Knoop)

120

160

2000

Thermal conductivity (W/m ◦ C)

18

17 (uniaxial) 4 27 (cubic)

27

dn/dT (1/ ◦ C)

70 × 10−6

46 × 10−6

12 × 10−6

Bandgap (eV)

2.8

3.8

1.7

Refractive index (at λ las )

2.45

2.27

2.47

1.76

Third-order nonlinearity n2 ( m2 /W)

170 × 10−20 90 × 10−20 (at 1.8 µm) (at 1.3 µm) [305] [306]

1300×10−20 (at 1.5 µm) [308]

3 × 10−20 (at 0.8 µm)

18 pm/V [235]

absent

Second-order nonlinearity 30 pm/V (pm/V)

uniaxial

2.3

uniaxial

8 [307]

2.5 µm [48,165,302]. Besides the Cr2+-ion, Fe2+ has demonstrated laser operation in ZnSe at cryogenic temperatures. Consideration of TM2+-doped lasers in this section will be limited to Cr2+- and Fe2+-doped materials, with the emphasis on Cr2+:ZnSe as the material providing superior performance so far. The material properties of the most important crystalline hosts for the Cr2+ ion are listed in Table 2. The remarkable characteristics of the ultrabroadband (∼ 1000 nm) Cr2+:ZnSe crystal, such as the high emission crosssection of the order of 10−18 cm2 [46,303], the negligibly low excited state absorption (ESA) [304], the fairly good chemical and mechanical stability and the thermal conductivity nearly as high as in sapphire, gives this material enormous potential as a laser medium for diode-pumped tunable MIR lasers. The only disadvantage of this material is the relatively high thermal

Crystalline Lasers

301

lensing parameter dn/dT (∼ 70×10−6 K−1 in Cr:ZnSe, 46×10−6 K−1 in ZnS compared to 12 × 10−6 K−1 in sapphire [309]). However, the latter is compensated by the generally low thermal load due to the absence of parasitic processes such as ESA or upconversion. Finally, maybe one of the most important advantages of Cr:ZnSe is the availability of technologically developed and low-cost polycrystalline material. The existing technologies of producing ceramic ZnSe, such as the chemical vapor deposition (CVD) method or the hot-press method of powders, result in high optical quality window substrates. This makes Cr:ZnSe one of the most technological and low-cost active media among known solid-state lasers. A ceramic Cr:ZnSe laser has been realized so far in CW, diode-pumped and mode-locked modes [303,310,311]. Spectroscopy of Cr2+ (d4 ) in chalcogenides has been the subject of extensive research since the early 1960s [82,312,313]. ZnSe and related compounds have been investigated as materials for light emission in the blue and green spectral regions, which can be used in light-emitting diodes, as well as lasers with lifetimes in the continuous-wave regime at room temperature in excess of 100 h [314], and in fluorescent displays [315]. At the same time ZnTe, CdTe and related compounds have been known as photorefractive materials and electro-optical limiters [316,317]. Curiously, the main interest in chromium as an impurity in ZnSe producing deep levels has been not that of an active ion, but as a quencher of luminescence that is not associated with intracentral transitions (for example, recombination luminescence). Therefore Cr is frequently referred to as a visible luminescence “killer”. Nevertheless, these fundamental studies have laid a solid basis for the latest spectroscopic investigations of Cr2+ as an active laser ion. In a tetrahedral crystal field, Cr2+ ions in the high spin configuration (S = 2) have the e2 t22 electron configuration. Two levels (5 T2 and 5 E) originate from the crystal-field splitting of the 5 D ground state of the freeion with the d4 configuration, which is the only quintet state. Since all the higher lying states are singlets or triplets, the excited state absorption (ESA) transitions from the upper state are spin-forbidden (Fig. 6b). Recently this has been experimentally proved [304]. The simplicity of the energy level scheme, which can be otherwise found only in Ti:sapphire, makes Cr:ZnSe and related compounds very attractive materials for laser applications. In Cr:ZnSe the multiplet ground state term 5 T2 is localized inside the forbidden gap at 0.46 eV [318] above the edge of the valence band as a deep donor (Fig. 20). The absorption and emission spectra in Cr:ZnSe and Cr:ZnS are depicted in Fig. 21. A broad absorption band centered around 1.8 µm [313,82,81] in Cr:ZnSe, around 1.7 µm [319] in Cr:ZnS and around 1.9 µm in Cd-compounds [320] corresponds to the transition between the ground 5 T2 and excited 5 E states, and is the only spin-allowed transition in this system. Correspondingly, the broadband emission between 2 and 3 µm reflects the parity-forbidden yet spin-allowed electronic transition between 5 E and 5 T2 states [81,321,322,323,324]. Indeed, according to the Laporte se-

302

Irina T. Sorokina

15 2+

Cr :ZnSe 2

Gain

σ (10

-19

10

Co:MgF2

5

Absorption 1400

cm )

InGaAsP diodes

Er:fiber

Tm:YALO Tm:YAG

Fig. 20. Schematic diagram of the Cr2+ ion in the bandgap of ZnSe and ZnS

1600

1800

2000

2200

2400

2600

2800

0 3000

15 2+

Co:MgF2

5

Absorption 1400

1600

1800

2 -19

10

cm )

Cr :ZnS

Gain

2000

2200

2400

2600

2800

σ (10

InGaAsP diodes

Er:fiber

Tm:YALO Tm:YAG

Wavelength, nm

0 3000

Wavelength, nm

Fig. 21. Absorption and gain cross-sections of Cr2+:ZnSe (upper graph) and Cr2+:ZnS (lower graph). The picture also shows the wavelength ranges and available power of some pump sources (not to the scale)

lection rule, electric dipole transitions between d-states are parity-forbidden. However, for Cr2+ ions substituting Zn at tetrahedral sites this restriction is lifted due to the absence of inversion symmetry at these sites. Besides,

Crystalline Lasers

303

the restriction is weakened due to phonon coupling. It is important to note that the corresponding lines are broad not solely because of the vibronic nature of the transition, but because both involved levels are additionally split due to the Jahn–Teller effect (for details see Sect. 3.1.1 and references therein). The corresponding Jahn–Teller distortion in Cr:ZnSe was subject to numerous investigations [81,85,89,325,326]. The Jahn–Teller splittings of the ground 5 T2 state and of the excited 5 E state were estimated to be 340 cm−1 and 40 cm−1 correspondingly [81]. The Jahn–Teller splitting may influence the optical properties of transition-metal doped crystals and therefore should always be kept in mind. The first thorough spectroscopic investigation of Cr2+-doped chalcogenides as laser materials and measurement of absorption and emission crosssections were carried out in [46,47]. The peak absorption and emission crosssections were measured for Cr2+:ZnSe to be 90×10−20 cm2 and 87×10−20 cm2 , respectively. The direct measurement of the emission lifetime τ was first carried out by De Loach et al. [46]. The earlier investigations [323,324,327,328] report similar measurements in other Cr2+-doped chalcogenides, where a Cr2+ lifetime of a few microseconds was found to be constant with temperatures up to ∼ 300 K. The rather low nonradiative decay rate in Cr2+:ZnSe is not characteristic for mid-IR vibronic laser materials and is due to the low phonon energy of ZnSe. The measured radiative lifetime was reported to be 8 µs in Cr2+:ZnSe [46]. However, close analysis of the temperature dependence of the lifetime in [46] reveals the somewhat curious dependence τ (T ), where τ increases from 5 µs at 4 K up to 8 µs at 300 K (Fig. 22a). In this case the radiative lifetime of 8 µs reported in [46] should be reconsidered [162]. It is not obvious whether the measured lifetime increase between 4 and 300 K is a measurement artefact (e.g. effect of reabsorption [161,163]) or a result of more complex intermultiplet interactions [329] due to Jahn–Teller splitting [78,79,89]. Recently extensive lifetime measurements were carried out for mainly polycrystalline [330] as well as for single crystalline Cr2+:ZnSe [331,303]. Both studies correlate with similar investigations in Cr2+:CdSe [332] and report

8

6

6 Lifetime (µs)

Lifetime (µs)

Cr:ZnS

8

Cr:ZnSe

4

2

4

2

(a) 0

0

(b) 100

200 Temperature (K)

300

400

0

0

100

200

300

400

Temperature (K)

Fig. 22. Temperature dependence of the Cr2+ lifetime in (a) ZnSe [46] and (b) ZnS

304

Irina T. Sorokina

concentration quenching of the Cr2+ lifetime. The authors of [331,303] suggested also the presence of a substantial reabsorption effect (see Sect. 3.1.1 and references therein), even in the rather thin, 1–2 mm, samples. Roomtemperature lifetimes of ∼ 6.5 µs [330] and ∼ 5.5 µs [331,303,333] were reported. These data lead to the correction of the initially reported absorption and emission cross-sections, being now equal to approximately 1.1×10−18 cm2 and 1.3 × 10−18 cm2 correspondingly [303]. For the same reason a similar spectroscopic investigation in Cr2+:ZnS [334] yielded a shorter (in comparison to 11 µs in [46]) radiative lifetime of 5.7 µs and larger absorption and emission cross-sections of ∼ 1 × 10−18 cm2 and 1.4 × 10−18 cm2 correspondingly. Cr:ZnS, otherwise very similar to Cr:ZnSe, is characterized by a slightly more rapid onset of nonradiative decay with temperature (Fig. 22b). Table 3 summarizes the spectroscopic parameters of the most important Cr2+ doped laser materials and compares them with those of Ti3+:sapphire as a reference material. It can be seen from the table that the lifetimes are the same in all materials and the cross-sections are noticeably larger than the corresponding values for Ti3+:sapphire [335]. However the emission saturation intensity, one of the most important parameters of the laser material characterizing its ability of reaching the threshold (Sect. 3.3), is by two orders of magnitude larger in chalcogenides than in Ti3+:sapphire. This explains, for example, why these materials are more suitable for diode pumping than Ti3+:sapphire. The semiconductor nature of chalcogenides makes the physics of ultrashort pulse formation different and the nonlinearity larger (by a factor of ∼ 60 in ZnSe) than in Ti3+:sapphire. The relative bandwidths, having the physical meaning of an inverse number of optical cycles per pulse, are comparable in all crystals. This means that achieving even two–three optical cycle pulse durations, corresponding to ∼ 15–20 fs at 2.5 µm, is feasible. The promising spectroscopic characteristics of Cr2+ based materials stimulated intensive laser related research largely focused around Cr2+:ZnSe. Since the first experimental demonstration [45,46,47] the laser performance of Cr:ZnSe and Cr:ZnS has greatly improved, including demonstration of direct diode-pumping in the pulsed [336] as well as continuous-wave regimes [98,304,337,338,339,340,341], active [337,342,343] and passive [343] modelocking, continuous-wave operation spanning several hundreds of nanometers [47,48,165,303,344] at the highest reported so far for transition-metal doped lasers and close to the quantum limit slope efficiency (> 60% [48,165]), narrow linewidth [165,303] and power levels in excess of 1.8 W in the CW regime [345] and several watts (up to 7 W) in the pulsed regime [346,347]. A crucial point in obtaining lasing in Cr:ZnSe is the choice of the pump source since the absorption band of Cr2+ is centered at 1.75 µm and only a few CW lasers are able to deliver sufficient power in this region. The pump sources implemented so far include: tunable between 1.6 and 2.1 µm Co2+:MgF2 laser [165], 1.9–2.1 µm Tm3+,Ho3+-lasers [48,304], ∼ 1.6 µm Er-

Crystalline Lasers

305

Table 3. Spectroscopic and laser characteristics of Cr2+-doped laser materials ZnSe

ZnS

CdSe (pulsed)

Cd0.55 Mn0.45 Te Ti3+:Al2 O3 (pulsed)

Peak emission cross-section σem (10−20 cm2 )

90[46] 130

75 [46] 140 [334]

200 [300]

170 [352]

39 [38] 45 [335]

λmax lum (nm)

2450

2350

2200 [300]

2480 [352]

780 [38]

Peak absorption cross-section σabs (10−20 cm2 )

87[46] 110

52 [46] 100 [334]

300 [300]

170 [352]

6.5 [38]

λmax abs (nm)

1780

1680

1900

1900

500 [38]

τem ( µs) at 300 K

8[46] 5.5

8[46] 4.3[334]

6 [300]

4.8 [352]

3 [38]

Saturation intensity 11 em Isat ( kW/cm2 )

14

8

10

210

Luminescence bandwidth (nm)

900

800

550

770 [352]

300

Relative bandwidth

0.41

0.34

0.25

0.31

0.38

Optical quantum efficiency

1[46]

0.73[46]

1 [300]

1 [352]

0.9 [191]

Slope efficiency (%)

53[349]

63 [165] 71 [304]

48 [353]

64 [352]

30

CW output power (W)

1.8[345]

0.7[98]

–

–

∼ 10 (commercial)

Output energy (average power)

0.43 mJ[346] 0.1 mJ[333] 0.8 mJ[301] 0.6 mJ[369] (11 W[347])

> 1J (commercial)

Mode-locked 80[342] output power (mW) 400 [343]

–

–

–

300 [350,351] (at τ = 5 fs)

Pulse duration

4.4 ps [342] 4 ps [343]

–

–

–

5 fs [350,351]

Diode pumping

yes

yes

no

yes

no

fiber lasers [348,98,349], ∼ 1.6 µm NaCl:OH color-center laser [302], ∼ 1.6 µm Raman-shifted Nd:YAG laser (see, e.g. the chapter by Basiev et al.), ∼ 1.6 µm Raman-fiber laser [354] as well as a variety of 1.6–1.9 µm InGaAsP/InP semiconductor lasers [98,340,341,355]. Some of these sources are also marked in Fig. 21. Direct diode pumping and pumping with Er-fiber laser [356] yield the highest wall-plug efficiency. The crystals were grown in different experiments using various crystal growth methods, including Bridgman [357], chemical vapor transport [358], and physical vapor transport [359], the latter providing the best laser results. It was also established that the introduction

306

Irina T. Sorokina

of chromium during growth yields worse results in comparison to the postgrowth diffusion-doping technique [330,360]. In the first laser experiments [47] the Bridgman grown crystals and the pulsed Co:MgF2 laser at 1.86 µm were used. In the simple two-mirror confocal cavity output energies as high as 0.6 mJ were achieved at absorbed pump energies of ∼ 3.5 mJ. Using the grating in the Littrow configuration tunability over 2150–2800 nm was achieved, however at a relatively broad linewidth of ∼ 40 nm. Later these results were improved and 45% slope efficiency was demonstrated in [361]. Scaling of output power by using a pulsed thin-disk (see Sect. 4.1 and [221]) Cr:ZnSe laser up to ∼ 4.3 W was demonstrated by McKay and Schepler in [346] (Fig. 23). A ∼ 1 mm-thick Cr:ZnSe was used in an active mirror configuration. The laser was pumped by a Q-switched Tm,Ho:YLF laser and operated at ∼ 50% optical to optical efficiency. More recently the output power in the pulsed Cr:ZnSe laser was further scaled up to 11 W, and the first Q-switched operation was demonstrated [347]. The pump source was a diode-pumped Tm:YAlO3 laser. The first CW operation was reported by Wagner in 1999 [48,362]. A threemirror folded cavity and a Brewster oriented Cr:ZnSe plate were used, pumped by the Tm:YAlO3 laser at 1.9 µm. Up to 380 mW output power at up to 63% slope efficiency was demonstrated. The laser was broadly tunable over the 2138–2760 nm range, and the free-running laser linewidth was 35–50 nm. The output power was afterwards scaled to 0.5–1.8 W by several groups [165,304,345,363]. The authors of [165] implemented an astigmatically compensated X-folded four-mirror cavity which is known to yield best results in terms of the mode quality and linewidth (Fig. 24). Indeed, narrow-linewidth operation was achieved for the first time in this laser, retaining polarization and TEM00 mode. The paper also reports observation of dual Q-switching and lasing at both pump and laser wavelengths due to cavity coupling in the case of multipass pumping or occasional reflections, as well as the line broadening associated with this phenomenon. The noticeable GSA was measured and 5

3

2

2+

Cr :ZnSe output (W)

4

1

0 0

2

4

6

Incident pump power (W)

8

10

Fig. 23. High-power operation of a Cr:ZnSe laser at 2.5 µm in a thindisk pumping arrangement [346] (data courtesy J. McKay)

Crystalline Lasers

307

Fig. 24. The astigmatically compensated X-folded cavity for Cr2+:ZnSe. All intracavity elements have Brewster surfaces. Lyot: birefringent filter for wavelength tuning. As an alternative, prism tuning was realized (shown by dashed line)

found to be responsible for discrepancies in internal loss measurements by Findlay-Clay [364] and inverse slope efficiency methods [166]. More recently these results were further improved and tunability over 1100 nm (between 2000 and 3100 nm) was achieved, preserving the narrow linewidth, which was measured to be less than 600 MHz [344,303] (Fig. 25). As already mentioned, because of the extremely broad gain and high nonlinearity, being ∼ 60 times higher than in Ti3+:sapphire (Table 3), Cr:ZnSe is especially attractive for ultrashort pulse generation. Two groups demonstrated transform-limited Gaussian-shaped pulses as short as 4 ps using an acousto-optic modulator [342,343]. The output power of 400 mW [343] was limited only by the pump source. Scaling to higher powers in excess of 1 W is therefore feasible. The mechanisms of pulse formation in the passive modelocking regime are currently under investigation [343,365]. The first diode-pumped operation was obtained by the authors of [336] in pulsed mode by using transversal pumping by a 75 W peak power Mirror set 2450 nm Mirror set 2750 nm

Mirror set 1950 nm

100

Air 90

400 300

Ambient air N2 purged

200

80

70

100

Transmission, %

Output power, mW

500

0 2000

2200

2400

2600

2800

3000

3200

Wavelength, nm

Fig. 25. Tuning range, demonstrated in Cr2+:ZnSe using three different mirror sets

308

Irina T. Sorokina

InGaAsP/InP laser diode array at 1.65 µm [366]. Peak output energies of ∼ 18 µJ and average output powers of ∼ 0.3 W were obtained. More recently the first CW end-pumped diode-pumped laser has been developed [337,338,339,304]. Nowadays ∼ 100 mW level output powers in CW regime have been demonstrated from Cr:ZnSe [97,341,355] and somewhat lower power from Cr:ZnS [98], using various diode lasers emitting between 1.6 and 1.9 µm. With the progress in laser diodes at this wavelength scaling to higher powers should be straightforward. The alternative to Cr:ZnSe materials, which could be used for lasing, include: Cr2+:ZnS, Cr2+:CdSe as well as Cr2+:Cd1−x Mnx Te. Until recently all these materials operated only in the gain-switched, pulsed mode. The first CW operation, tuning and spectroscopic study of Cr2+:ZnS was published in [334]. It was noticed that the formally cubic crystal of Cr2+:ZnS exhibited birefringence. The X-ray analysis revealed the hexagonal symmetry in the crystal. However, the structure of the crystals was not that of wurtzite, but a modification of the cubic one. Cr:ZnS is one of the most structurally rich polytypical compounds and can exist in several structure types, sphalerite and wurtzite structures being the most common. However, even so-called cubic Cr:ZnS as a rule reveals a certain degree of “hexagonality”, usually amounting to 10–20%. Such crystals demonstrate natural birefringence. The cause of the hexagonal symmetry may be twinning and fault-stacking enhanced by the Cr-doping in the case of melt grown Cr:ZnS. This problem has not yet been solved by crystal growers in the case of ZnS single crystals (note that opposite to ZnS the fault stacking-free purely cubic crystals of ZnSe are known to have been grown by different methods). However, this did not prevent demonstrating up to 700 mW output power at 2.35 µm, when pumped by the Er-fiber laser at 1.6 µm, as well as diode-pumped operation [98]. It is interesting to note that the diode-pumped version of this laser allowed even broader tunability than Cr:ZnSe, i.e. 400 nm over 350 nm in Cr:ZnSe [340] (Fig. 26). The Er-fiber pumped microchip laser, based on both Cr:ZnS as well as Cr:ZnSe, has been simultaneously realized [349]. A slope efficiency of 53% has been realized from a Cr:ZnS microchip laser.

Output power, rel.u.

Cr:ZnS Cr:ZnSe

98

96

94 2200

2400

Wavelength, nm

2600

Air transmission, %

100

Fig. 26. Tuning curves of diodepumped Cr:ZnSe and Cr:ZnS (after [98])

Crystalline Lasers

309

Chromium doped cadmium chalcogenide compounds are very attractive, because their emission is shifted to longer wavelengths. Cr2+:Cd1−x Mnx Te compounds are an alternative to Cr2+:ZnSe and Cr2+:ZnS materials [352,367,368,369]. As can be seen from Tables 2 and 3 they all have similar spectroscopic properties. However, one of the incentives to work with them is that it is easier to grow them in comparison to selenides [370]. One can also tailor the desired spectroscopic properties by variation of chemical composition [371]. Free-running operation of pure Cr2+:CdTe, has been reported in [371]. The laser delivered 132 µJ at 2 Hz repetition rate, when pumped by 1.4 mJ from a Cr,Tm,Ho:YAG laser at 2.09 µm. In Cd0.85 Mn0.15 Te tuning was accomplished by using a quartz birefringent filter and extended from 2.3 to 2.6 µm [367], as well as by a grating in a Littrow configuration and covered the 2.17–3.01 µm range [369]. Output energies as high as 0.6 mJ could be achieved at an absorbed pump energy of ≈ 1.6 mJ. More recently the first CW and direct diode-pumped operation was demonstrated in this material [341], however, at the expense of the output power, which in the CW regime did not exceed 6 mW (15 mW in the pulsed regime). This can be explained by the poor thermal conductivity and strong thermal lensing in this material. Therefore in the future the diode-pumped Cr2+:Cd1−x Mnx Te laser is expected to operate only in the pulsed regime. Another interesting cadmium compound is Cr2+:CdSe. Pulsed radiation at 2.6 µm has been generated from this crystal [300] at an average power of 500 mW and pulse repetition rate of 1 kHz and 48% conversion efficiency (815 mW at 27 % conversion efficiency) [301,372]. Remarkably broad tuning was reported in this laser, allowing operation between 2.45 and 3.4 µm in pulsed operation (Fig. 27) [373], pumped by a 2.05 µm Tm, Ho:YLF laser. CW operation could not be obtained so far mainly due to the rather poor thermal properties of Cr2+:CdSe in comparison to Cr2+:ZnSe and Cr2+:ZnS. There are only a few room-temperature lasers beyond 3.4 µm. The need for yet longer wavelengths has stimulated research in Fe2+:ZnSe, the only TM2+doped chalcogenide to be operated beyond 3.5 µm. Fe2+ ions in ZnSe have an 0.5

Absorbed energy (mJ)

0.4

0.3

0.2

0.1

0.0 2400

2600

2800

3000

Wavelength (nm)

3200

3400

Fig. 27. Oscillation threshold versus wavelength for a pulsed Cr2+:CdSe laser, demonstrating long-wavelength operation [373] (data courtesy J. McKay)

310

Irina T. Sorokina

energy level structure similar to Cr2+ ions, and therefore ESA is not expected to be a problem here. In fact, the first laser, based on Fe2+-doped n-InP, was demonstrated by Klein et al. [374] by using above-bandgap excitation at 2 K. On the other hand in Fe2+:ZnSe, pulsed laser operation could be obtained at a much higher temperature of 130 K and yielded a maximum output energy of 12 µJ/pulse [375,376]. Tuning with temperature was realized in the wavelength range between 3.98 and 4.54 µm. Concluding this section, I would like to mention two important directions of further development of Cr2+ lasers. The first direction is power scaling. In this connection technical approaches such as thin-disk laser, waveguide laser as well as edge-pumped zig-zag slab laser approaches (see Sect. 4.1) should be explored. The second direction is wavelength scaling towards the infrared using OPOs (see the chapters by Vodopyanov and by Ebrahimzadeh). For example, the operation wavelength of the first Cr:ZnSe-based pulsed ZGP OPO [377,378] was successfully shifted from 2.5 to 4–6 µm. 5.2

Color Center Lasers

This section reviews the existing color-center lasers in the mid-IR (between ∼ 2 and 5 µm) (Table 4). Color center lasers (otherwise called F-center lasers) are a broad class of mainly CW tunable and mode-locked light sources based on the allowed vibronic laser transitions between electronic levels of the active color centers. They possess two major features, making them ideal efficient CW broadly tunable light sources, operating between 0.8 and 5 µm. The first is the large oscillator strength due to the allowed nature of the transitions. Just for comparison, typical emission cross-sections (∼ 10−16 cm2 ) are by three–four orders of magnitude larger than in octahedrally coordinated Cr3+or Ti3+-doped crystals, and by two orders of magnitude larger than in tetrahedrally coordinated Cr2+-doped crystals (Sect. 5.1.2). Naturally, the lifetime is typically small (tens of nanoseconds), which does not allow storing of energy in the Q-switched regime. The second characteristic feature of color-center lasers is that they posess homogeneously broadened, Stokes shifted and wellseparated absorption and emission spectra (as an example see, e.g. Fig. 28). As a result they do not suffer from self-absorption effects (Sect. 3.1.3). Generally, vibronically broadened transitions in color-center lasers are in many respects similar to those in transition-metal doped crystals and can be described in the same way, using the configurational-coordinate model, introduced in Sect. 3.1.1. Color centers are essentially electrons trapped by various defects in dielectric crystals. The simplest color center is the F-center (F comes from the german Farbe, which stands for color), which is an electron trapped at an anion vacancy. It does not lase. However, there is a large variety of more complex color centers and their aggregates, which do lase. Figure 29 illustrates a number of such centers along with the classic F-center. For example, the F+ 2 center consists of an electron trapped at two adjacent anion vacancies and an

Crystalline Lasers

311

Table 4. Color-center lasers in the mid-IR Host

Active color center

Pump source

Mode of operation

KCl:Li+ KCl:Li+ KCl:Li+

FA FA FA

pulsed 77 pulsed tunable 77 CW tunable 77–200

2.72 2.4–2.9 2.6–2.8

[388] [397] [389]

KCl:Li+

FA

CW tunable

77

2.32–3.12

[292]

KCl:Li+, KCl:Na+

CW tunable

77

0.9–3.3

[390]

RbCl:Li+ KCl:Na+

FA (II), FB (II), F+ 2 FA (II) FB (II)

Flash-lamp Flash-lamp Kr laser, 647 nm Kr laser, 647 nm Kr laser, 647 nm

pulsed CW tunable

77 77

2.55–3.28 2.25–2.9

[398] [292,399]

RbCl:Li+: Na+

FA (II)FB (II)

CW tunable

77

2.5–3.15

[400]

RbCl:Li+

FA (II)

CW tunable

77

2.6–3.33

[400]

RbCl: Na+

FB (II)

CW tunable

77

2.5 –2.9

[400]

KCl:Li+

(F+ 2 )A

CW tunable

77

2.0–2.5

[401]

+

KCl:Li , KCl:Na+ KBr:O2+

(F+ 2 )A

CW tunable

77

1.67–2.46

[403]

CW tunable

77

1.8–2.4

[404]

KI

(F+ 2 )A

pulsed tunable 77

3.28–3.99

[405]

KI

+ F+ 2 -(F2 )A

pulsed tunable 77

1.98–3.85

[405]

KBr

FH (CN− )

pulsed

1.7

4.86

[406]

KBr

FH (CN− )

Dye laser Kr laser, 647 nm Kr laser, 647 nm, 752 nm Kr laser, 647 nm, 752 nm Kr laser, 647 nm, 752 nm Nd:YAG, 1.34 µm Nd:YAG, 1.32 µm NaCl(F+ 2 )H laser Er:YLF, 1.72 µm Er:YLF, 1.72 µm 2ω CO2 , 4.81 µm (F+ 2 )A KCl, 2.42 µm (F+ 2 )A KCl, 2.42 µm (F+ 2 )A KCl, 2.42 µm

CW

1.7

4.86

[407]

CW

30

5

[409]

CW

77

4.89; 4.94: 5

[410]

(F+ 2 )AH

−

CsBr

FH (CN )

CsCl

FH (CN− )

Temp. (K)

Wavelength range ( µm)

Ref.

+ + (F+ 2 )A center is an F2 center trapped next to an impurity alkali ion (e.g. Li + or Na in KCl lattice). For more information on the physics and chemistry of color centers the interested reader is referred to the excellent books and reviews by Fowler [379], Mollenauer [380,381], and Pollock [383,384]. Although the first tunable mid-IR lasers, used mainly in scientific instrumentation, were those based on color centers, there are still no roomtemperature mid-IR color-center lasers up to now. Room-temperature colorcenter lasers exist in the near-IR range up to ∼ 1.3 µm [385,386,387]. Because of the necessity of cryogenic cooling color-center lasers, operating between 1.5 and 3 µm, are nowadays less attractive for real-world applications and are more rarely used than paramagnetic ion-doped crystalline lasers, at least in the wavelength range up to 3 µm. However, the wavelength range between 3

312

Irina T. Sorokina Wavelength (µm) 1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

1.0

Absorption 0.5

Emission

Laser

0.0

1.0

0.9

0.8

0.7

0.6

0.5

Energy (eV)

Fig. 28. Normalized absorption and emission spectra of the lithium (F+ 2 )A center in KCl at 77 K, and the tuning curve (solid circles) of the corresponding laser (after [401])

Fig. 29. The classical F-center and various laseractive color centers (after [382])

and 5 µm, where there are no room-temperature competitors among tunable transition-metal doped CW laser sources, is still largely covered by cryogenically cooled color-center lasers. The first flash-lamp pumped pulsed operation around 2.72 µm of a FA color-center laser based on KCl:Li+ was demonstrated by Fritz and Menke as early as 1965 [388]. A decade later, in 1974, Mollenauer and Olson reported the first Kr laser pumped (around 647 nm) CW tunable between 2.6 and 2.8 µm operation in the same type of laser [389]. The laser worked over the temperature range 77–200 K. Later the authors extended the tunability range up to 3.3 µm [390]. They also reviewed the appropriate laser configurations, the latter being very similar to those used in dye lasers. This is natural since the first color-center laser was conceived as a direct solid-state analog of a dye laser. Considering threshold values (down to a few mW), output powers (hundreds of mW), tuning ranges and efficiencies, color-center lasers are comparable with dye lasers. The obvious advantage of the solidstate active medium over the flowing dye jet is its static character, providing the amplitude stability. The high amplitude stability leads to good frequency

Crystalline Lasers

313

stability, yielding a much narrower linewidth than is possible with dye lasers. Usually, the free-running color-center laser shows a frequency jitter of about 1 MHz [391]. Considering the measured dependence of the linewidth on the disturbing effects of temperature and pump power fluctuations, a linewidth of 25 kHz for the free running single mode color-center laser was shown to be achievable [392]. It is therefore not surprising that active stabilization techniques lead to phenomenal improvement in laser linewidth, allowing it to be reduced down to ∼ 1 kHz [393]. Because of this, color-center lasers are attractive for applications in laser spectroscopy and chemistry requiring high spectral or temporal resolution in the near- and mid-IR [394]. In fact, these sources are frequently used in the frequency standard measurement systems [395]. Beigang and co-authors studied the achievable linewidth and demonstrated the utility of using color-center lasers for atomic spectroscopy [396]. They described [397] a pulsed narrow-band laser using FA (II) centers in KCl:Li+, which provides peak powers up to 20 kW with a bandwidth below 0.3 cm−1 in the wavelength range from 2.4 to 2.9 µm. The extension of the tuning range towards infrared (from 2.55 to 3.28 µm) was demonstrated for FA (II) center lasers in RbCl:Li+ [398]. Similar tuning ranges (2.25–2.9 µm) could be obtained from FB (II) centers in KCl:Na+ lasers [399,292]. Combined FA (II)FB (II) type RbCl:Li+:Na+ crystals exhibited an overall tunability between 2.5 and 3.33 µm [400]. The tuning range below 2.5 µm (between 2 and 2.5 µm) is equally interesting, because it corresponds to the water-free window. The authors of [401] obtained continuous tunability in this wavelength range maintaining high absorbed pump power efficiency of 39% and a low absorbed power threshold of only 5 mW. Based on this laser, the authors later built an oscillator-amplifier system, tunable between 2 and 2.4 µm and yielding output energies of 2.3 mJ and peak powers exceeding 0.4 MW [402]. Continuous tunability (between 1.67 and 2.46 µm) of KCl crystal containing Li+ and Na+ (F+ 2 )A centers and pumped by a 1.3 µm line of Nd:YAG laser was reported in [403]. Output powers as high as 150 mW could be obtained from the (F+ 2 )H centers in + 2− KBr:O2− and (F+ ) centers in KBr:Na :O crystals pumped by another 2 AH ) color-center laser. The lasers were continuously tunable between NaCl (F+ H 2 1.8 and 2.4 µm [404]. The wavelength range around and above 3 µm is especially interesting for spectroscopic applications, because of several gas absorption lines in this region (see, e.g. the chapters by Richter et al. and Fischer et al.). Schneider was the first to extend tunability of a pulsed Er:YLF-pumped (at 1.72 µm) + KI laser based on (F+ 2 )A up to ∼ 4 µm (3.28–3.99 µm) [405]. For the F2 and + (F2 )A centers coexisting in KI, the combined lasing range was demonstrated to be 1.98–3.85 µm. At this point it is worth noting that except for the very first publication on CW operation of color centers, where the latter operated in the temperature range 77–200 K, all the other laser systems described above operated at

314

Irina T. Sorokina

liquid nitrogen temperature. Going further into the infrared, it gets more and more difficult to obtain laser action even at this temperature. As discussed in Sect. 3.1.1, to overcome the barrier set by multiphonon relaxation it is necessary to use hosts with a minimal value of the maximum phonon energy. According to Table 1 bromides represent the ideal hosts in this respect. The first pulsed operation of a KBr crystal with a CN− molecule around 4.86 µm was realized only at 1.7 K [406]. A year later Gosnell et al. [407] achieved CW operation in this laser, pumped by another tunable (F+ 2 )A KCl color-center laser at 2.42 µm with only 13 mW of the incident pump power. The laser linewidth was measured to be less than 1 GHz, corresponding to only two or three longitudinal modes of the cavity. The authors discussed the physical mechanisms preventing lasing at the elevated temperatures. They give promising prospects for lasing at 80 K and even room temperature [408]. Later CW operation at around 5 µm of the F center/CN− defect pairs, “FH (CN− ) centers”, was obtained in CsBr crystals at up to 30 K temperature [409]. In the host CsCl laser oscillation was obtained on the same kind of centers at temperature up to 77 K at a few mW threshold on the three strongest vibrational emission transitions at 4.89, 4.94 and 5.0 µm [410]. These results are promising enough. However, more material science work has to be done to develop a practical mid-IR source around 5 µm. An alternative approach consists in developing a rare-earth doped laser in halide hosts (Fig. 1). 5.3 5.3.1

Rare-Earth Doped Lasers Tm3+- and Ho3+-Doped Lasers

The first experiments with Tm- and Ho-doped crystals were performed by Johnson et al. using Tm:CaWO4 [22], Ho:CaWO4 [21], and Tm,Ho:CaWO4 [23] crystals at 77 K. It took a while to realize that both ions allow pulsed operation at room temperature. This was first achieved in Ho3+:YLF in 1971 [411] and later in Tm, Cr -doped YAG and YAlO3 in 1975 [412]. Laser operation in the 3 µm range in Ho was demonstrated in different hosts in 1976 [413]. It took another decade to demonstrate CW operation of a Cr,Ho,Tm-doped YSAG and YSGG [414] under a Kr+ion laser pumping at 647 nm using a co-doping scheme suggested by Antipenko et al. [34]. Soon after that, the availability of laser diodes in the wavelength region 780–790 nm made possible CW diode-pumped operation in Tm,Ho:YAG [415,416,417] at room temperature. Since the laser ion, Ho3+, has a high peak emission cross-section, the pumping threshold was lower than 5 mW. Progress in laser diodes allowed demonstration of diodepumped Tm:YAG [418], single-frequency monolithic Tm:YAG [421] and tunable Tm,Ho:YLF [419,420]. At the moment, lasers based on Tm-, Ho-, and Tm,Ho-doped crystals are well-developed and available commercially. The dopant concentrations and laser schemes vary, and are chosen depending upon the desired wavelength,

Crystalline Lasers

Intensity (a.u.)

1

315

Cr,Ho,Tm:YSGG:GSAG fluorescence spectrum

0 1600

1800

2000

2200

Output power (mW)

Wavelength (nm) 30

Diode-pumped tuning curve 20

10

Cr,Ho,Tm:YSGG:GSAG 0 2040

2060

2080

2100

2120

Wavelength (nm)

Fig. 30. Fluorescence spectrum and tuning curve of the Cr,Tm,Ho:YSGG:GSAG laser (after [264])

mode of operation, output beam parameters, and output power. Using Ho3+ as the laser ion, one can obtain oscillation on a 5 I7 –5 I8 transition in the 1.95– 2.15 µm range on a number of distinct lines (depends upon the crystal), as well as on the 5 I6 –5 I7 transition in the 2.85–3.05 µm region [260,413,422,423,424]. There exist plenty of literature reports [53], as well as commercial systems, for Ho-based lasers in the 2-µm range. Recent developments concentrate on improving the efficiency and tuning range. To obtain broad continuous tuning, one can make use of inhomogeneous broadening, which should be sufficient to fill the gaps between the individual spectral lines. One such example is the Cr,Tm,Ho:YSGG:GSAG crystal, where the host matrix is a solid solution of GSAG and YSGG garnets [52]. Such a solid solution retains the garnet structure, but substitution involves both dodecahedral sites (yttrium and gadolinium) and tetrahedral sites (gallium and aluminium), creating a high degree of disorder. The resulting inhomogeneous broadening allows as much as 80-nm continuous tuning (Fig. 30) [264,425]. Recently, a continuous tuning range of 85 nm has also been achieved with Tm,Ho:BaY2 F8 [426]. While Tm-doped lasers demonstrate even broader tuning ranges and higher powers in the same spectral region, Ho-based lasers are attractive for compact low-threshold devices [427]. Finally, Ho-doped crystals can be diode-pumped at 1.9 µm directly into the upper 5 I7 manifold, achieving up to 0.7 W of CW output [428]. In this scheme the low Stokes shift improves the power- and temperature-handling capabilities of the system, allowing operation at the heatsink temperature up to 60 ◦ C.

316

Irina T. Sorokina

Ho-based lasers in the 3 µm range are less common. In this case, the metastable 5 I7 level serves as a lower laser level (Fig. 12), and the laser operates only in pulsed mode, because the lower laser level accumulates population during oscillation (self-terminated transition). Recently, using a Yb3+ ion as a sensitizer, a diode-pumped Ho,Yb:YSGG laser at 2.84 µm has been realized [424], achieving 10 mJ output energy at 2.5% efficiency. It is also possible to obtain oscillation on both transitions of the Ho3+ ion in the same resonator using the cascade oscillation scheme in Cr,Yb,Ho:YSGG [429]. Recently it was shown that variation of the flashlamp pump pulse length provides a new method to discriminate between lasing at 2.1 and 2.9 µm in YAG doped with Cr3+, Er3+, Tm3+ and Ho3+ [189]. According to the authors this effect results from the Er3+ to Tm3+ to Ho3+ energy transfer, the short lifetime of the upper lasing manifold in Er3+, the 4 I11/2 manifold, and the long upper laser level lifetime in Ho3+, the 5 I7 manifold. This simple and elegant tuning method, allowing two widely separated wavelengths without the use of any optical tuning elements has a large potential for applications such as remote sensing and medical lasers. Finally, it is possible to obtain pulsed operation of a 5 I5 –5 I6 transition. In this case, the upper-level lifetime is again shorter than that of the lower level, the transition is self-terminated, and can only be operated in pulsed mode. Pumping should occur directly into the upper laser level at ≈ 900 nm. Recently, a room-temperature 3.9 µm Ho:BaY2 F8 laser was demonstrated using pulsed Cr:LiSAF laser pumping [430]. Tm-doped crystalline lasers can provide tunable operation in two spectral regions: 1.9–2 µm using the 3 H4 –3 H6 transition and 2.2–2.4 µm on the 3 F4 –3 H5 transition (Fig. 12). The 3 H4 –3 H6 transition is especially attractive for high-power applications. In a diode-pumed Tm:YAG as much as 115 W of CW output power has been obtained at 2.01 µm, using the end-capped rod geometry and the lens duct pump delivery [242]. This output power has been obtained at the expense of the beam quality (M 2 ≈ 20) and polarization properties. Almost diffraction-limited beam quality (M 2 = 1.3) could be maintained up to 14 W of the output power from the Tm:YAG and Tm:LuAG crystals (unpolarized) [432], using a special beam-shaping diode-laser collimation technique with conventional rod geometry. Using fiber-coupled pump delivery, over 20 W has been obtained in the end-capped Tm:YLF and Tm:YAlO3 at 1.91 and 1.94 µm, respectively. Due to the natural birefringence, the output was linearly polarized in both materials and of good beam quality (M 2 = 3 and 2.5, respectively). Finally, as much as 28 W of polarized output power with M 2 < 1.3 beam quality has recently been obtained in Tm:YLF (21 W from Ho:YLF) using a novel side-pumped slab geometry [224]. Another important feature of Tm-based lasers is the possibility of continuous tuning on both mentioned transitions. In the 1.9–2 µm range tuning has been demonstrated in Tm:YAG and TmYSGG (1.87–2.16 µm and 1.84– 2.14 µm, respectively) [433], Tm:YAlO3 (1.93–2.00 µm) [434] in Tm:Y2 O3

Crystalline Lasers

317

3+

-1

Absorption coefficient (cm )

40

30

6.8 at. % Tm 20 2 (8.3x10 cm )

π-polarized

20

σ-polarized

pump diodes @ 808 nm

10

0

780

790

800

810

820

Wavelength (nm)

Fig. 31. Absorption spectrum of Tm:GdVO4

and Tm:Sc2 O3 (1.93–2.09 µm and 1.93–2.16 µm, respectively) [435], Tm:YLF (1.91–2.07 µm) [224], and in Tm:GdVO4 (1.86–1.99 µm) [436]. The latter crystal [437,438,439,440] is especially attractive because of its high absorption at 808 nm (Fig. 31). All the previously mentioned Tm and Tm,Ho doped crystals possess usable pumping bands in 780–795 nm. However, the pumping diodes in this range are more expensive and possess lower brightness than in the 805–810 nm range. The extremely high absorption coefficient of the Tm:GdVO4 crystal (and of its analog Tm:YVO4 [441,442]) at typical working concentrations allows pumping in the 805–810 nm range. The high absorption coefficient allows operation in the microchip arrangement [439]. And finally, the luminescence band of this crystal (Fig. 32) and tuning range (Fig. 33) are slightly shifted towards shorter wavelengths, making this crystal attractive as a pump source for Cr:ZnSe (Sect. 5.1.2).

3+

σ

Luminescence (rel. units) 1600

Tm :GdVO4 lifetime 390 µs (room temperature)

π

1700

1800

Observed tuning range

1900

2000

2100

Wavelength (nm)

Fig. 32. Fluorescence spectrum of Tm:GdVO4

318

Irina T. Sorokina

Intensity (rel. units)

σ - polarization π - polarization

free-running spectrum

1860 1880 1900 1920 1940 1960 1980 2000 Wavelength (nm)

Fig. 33. Tuning range and free-running oscillation of the Tm:GdVO4 laser at room temperature

In the 2.3 µm range (3 F4 –3 H5 transition), tuning has been demonstrated in Tm:YLF (2.2–2.46 µm) [443] and in Tm:GSGG (2.2–2.37 µm) [444]. It is not possible to achieve 2 µm and 2.3 µm operation in the same crystal, as has been done for Ho-doped crystals, because the two transitions require different Tm3+ ion concentrations. As described in Sect. 3, pumping the 3 H4 –3 H6 transition relies on the cross-relaxation process that requires high concentration of Tm3+ ions, typically > 5 %. This process, however, would be detrimental to operation on the 3 F4 –3 H5 transition, because cross-relaxation will depopulate the upper laser level 3 F4 (Fig. 12). Therefore, operation in the 2.3 µm range requires a low concentration of Tm3+ ions, typically < 2%. Because the pump absorption becomes a problem at low Tm3+ concentrations, the Tm:GdVO4 crystal should be particularly suitable for this application. Mode-locking experiments have been performed with both Tm- and Ho-based lasers. Pulses as short as 35 ps at an output power of 70 mW have been demonstrated in Tm:YAG [265] and 41-ps pulses have been realized in Cr,Tm:YAG at 190 mW of output power [263]. In Ho-based systems pulses have been by an order of magnitude longer, 800 ps in Cr,Tm,Ho:YAG [445] and 370 ps in Tm,Ho:YLF [446]. The significantly longer pulse duration in Ho-doped media can be explained by the spectrum structure of Ho3+, consisting of narrow peaks. However, using the disordered garnet solid solutions Cr,Tm,Ho:YSGG:GSAG mentioned above it was possible to obtain pulse duration down to 25 ps at 100 mW of output power in Cr,Tm,Ho:YSGG:GSAG[264,425]. These remain the shortest pulses from a mode-locked laser at 2 µm.

Crystalline Lasers

5.3.2

319

Er3+-Doped Lasers

Erbium lasers operating in the 3 µm wavelength range are extremely attractive for applications such as laser surgery and ophthalmology (for details see the chapter by Jean and Bende). Their emission coincides with O-H vibrations in water [447] and is therefore strongly absorbed within less than a micrometer [448,449]. These lasers exhibit a rather complicated multistage population dynamics, involving several cross-relaxation and upconversion processes. Since the 3 µm laser transition is self-terminated, these processes play a key role in the proper population of the upper laser level. The spectroscopic basis for understanding the operation of Er3+-ion doped crystalline lasers has been provided in Sect. 3.1.3 as well as in the chapter by Pollnau. During the last two and a half decades of Er laser technology development, hundreds of new active media containing Er3+ ions have been demonstrated using various excitation schemes and pump sources. The present discussion will be focused on the description of the most important experimental results with the aim of finding the best performance room-temperature and preferably diode-pumped systems which have been developed so far, and determine their main advantages and limitations. YAG was the first crystalline host, incorporating Er3+ ions for the matter of obtaining stimulated radiation around 3 µm. Stimulated emission from Er3+ ions in YAG crystals at λ = 2.94 µm was reported in 1975 by Zharikov et al. [450]. Later, an efficient cross-relaxation laser making use of the above-mentioned energy transfer processes was reported by Zhekov et al. [451]. As discussed in Sect. 3.1.3, the main problem of Er lasers is the selftermination of the lower laser level 4 I13/2 . This was the main obstacle preventing continuous-wave operation in a Er:YAG crystal. This problem could be solved by using increased Er3+ concentration. Several other potential hosts for erbium have been studied and laser action has been obtained in many of them using flash-lamp pumping as well as laser pumping. For more information the interested reader is referred to the book by Kaminskii [29]. Various garnets with large lattice constants have been studied with the aim of finding an optimum host, which would incorporate a large erbium concentration and would provide conditions favorable to efficient sensitization with chromium ions. One such host was YSGG (yttrium scandium gallium garnet) [452]. Huber et al. achieved Kr laser pumping at room temperature and CW operation on various transitions in YAG, YGG, YSAG and YSGG doped with Cr and Er-ions [41]. In [453] it was demonstrated that the flashlamp pumped operation in YSGG with an average output power of 2.7 W could be achieved at 2.796 µm at much lower threshold than in other crystals (i.e. at a pump energy of 5 J, and repetition rate up to 10 Hz). Also, many crystals have been studied which have 100% of Er3+ ions in substitutional sites (for example, ErAG can be considered as 100% substituted YAG) [454]. Later, interesting results on diode-pumped operation in

320

Irina T. Sorokina

Er3+:LiYF4 were reported by Kintz et al. [455]. The first true-CW 3-µm operation was observed and compared with that in the pulsed regime. It was found that the threshold in the pulsed regime is much lower due to the lifetime of the terminal laser level being longer than the upper laser level. Depopulation of the lower laser level, which permits CW operation, was due to a cooperative upconversion process. Pulsed outputs of 21 µJ and CW outputs of 180 µW have been obtained. These low outputs and correspondingly low efficiencies were due to operating in the self-terminating mode of the laser system. It was concluded that higher concentrations and higher pumping rates should have significantly improved the efficiency. The authors of [181,456] reported on the pulsed and CW laser operation in CaF2 –ErF3 using a stepwise upconversion pumping scheme, described in Sect. 3.1.3. The new experimental data, which are important for understanding of the physics of stimulated emission on the self-saturated 4 I11/2 – 4 I13/2 channel of Er3+ ions, were obtained in [457]. In this work the authors studied CW operation of the diode-pumped Er3+:LiYF4 . A threshold as low as 5 mW of absorbed power at 0.795 µm (diode-laser) was obtained for the 2.81 µm wavelength. It follows from this paper that lasing on this channel proceeds according to the three-level scheme. The results confirm the explanation of the “red shifted” 3-µm-emission lines of Er3+ ions in crystals such as (Y1−x Erx)3 Al5 O12 and (Lu1−x Erx )3 Al5 O12 , which was first given in [458] (see also the chapter by Pollnau). All these works contributed to understanding the complexity of Er3+ energy levels and to finding ways to optimize the population of the upper laser level by choosing the right concentration and the right pumping conditions. This, together with the progress in diode lasers, led to the development of a number of room-temperature diode-pumped Er lasers in the mid-1990s. For example, the authors of [459] report on the diode-pumped CW operation of Er3+:LiYF4 . The authors of [460] report on the first CW operation of monolithic Er:YAG, Er:GGG and Er:YSGG lasers at wavelengths near 3 µm. Pumping with an InGaAs diode laser at 970 nm, they have observed thresholds as low as 5 mW and power outputs as high as 0.5 W, with absolute efficiencies approaching 30%. In addition tunable single-frequency operation was demonstrated from Er:YAG. The authors of [461] claim the first Q-switched operation in the diodepumped regime of Er:YLF and Er:YSGG, producing 5 mJ/19 ns pulses. Microlasers began appearing, now being diode pumped. The first one was made on the basis of Er:YSGG [462]. In BaY2 F8 with 10% Er3+ ions the authors of [463] obtained up to 160 mW output power at 32% slope efficiency (compare with 35% quantum defect), pumping by 550 mW from laser diodes at 970 nm. More recently two groups reported high-power operation from diodepumped Er:YAG. First, the group from the Lawrence Livermore Laboratory [464,220] announced a composite-slab Er:YAG laser side diode-pumped

Crystalline Lasers

321

Fig. 34. Side-pumping with the laser diode arrays for a novel composite slab concept, where the Er:YAG is diffusion bonded to a sapphire plate, allowing both optical access and cooling through the substrate. After [220]

and producing output power levels in excess of 1 W with peak slope efficiencies ∼ 20% (Fig. 34). They could achieve this by employing the pump light coupling through a sapphire plate diffusion-bonded to the laser slab, removing heat directly at the pump face of the slab instead of requiring conduction through to its far side (Fig. 35). This reduced the temperature in the gain region and the thermal lensing, thus, providing exceptionally good beam quality (M 2 ≈ 1.3). Somewhat later the authors of [465] also demonstrated more than 1 W of Gaussian-like CW output at 2.94 µm from a diode laser end-pumped monolithic laser crystal composed of Er-doped YAG bonded to undoped YAG. The output at 2.94 µm was generated with 34% slope efficiency and greater than unity quantum efficiency. Recently, as much as 4 W of polarized CW output power was obtained from Er:YLF laser [469] at the end of the first paragraph. It is important to note that all the described systems operated on the 4 I11/2 –4 I13/2 transition of Er3+. However, some low phonon energy hosts allow operation on the longer wavelength transition, 4 I9/2 –4 I11/2 . The flashlamp pumped operation on this transition at 110 K in (Y,Er)AlO3 was demonstrated at 4.75 µm [29]. However, CW diode-pumped laser operation at 120 K was achieved in Er3+:LiYF4 at 3.41 µm [466], as well as pulsed diode-pumped operation at 4.6 µm [467] at room temperature. These are the longest wavelengths accessible to diodepumped lasers so far.

6

Conclusion and Outlook

Summarizing, state-of-the-art crystalline solid-state lasers, operating in the broad range of wavelengths between roughly 2 and 5 µm as well as several operational modes, including the free-running, continuous-wave, gain-switched, Q-switched and mode-locked regimes have been reviewed. The chapter began with a brief excursion into the history of solid-state lasers. The historical insight gained into the development of solid-state laser technology helped to

322

Irina T. Sorokina

Fig. 35. Output characteristics for the Er:YAG composite slab. Repetition rate is 120 Hz; 500-µs pulse duration; 18% slope efficiency. After [220]

correctly understand the place of the considered laser sources in the historical context as well as to get a vision of the avenues for the future work. The physics of the ion-doped crystalline active media has been considered. In particular, an attempt has been made to explain the difference between the rare-earth ions and transition-metal ions as well as between the ions occupying different substitutional sites, i.e. the difference between the tetrahedral and octahedral positions, the weak and the strong crystalline field, etc. The spectroscopic background was provided, which was necessary to understand the principles of laser operation and the characteristic features, distinguishing the tunable, mainly vibronic active media from their fixed-wavelength counterparts. In the future research efforts will certainly continue to extend the operational wavelength of solid-state lasers further into the infrared at room temperature. As can be seen from the present overview conventional crystalline lasers are restricted to operation at wavelengths shorter than 3 µm because of the rapid luminescence quenching due to multiphonon relaxation. In order to extend the wavelength range of solid-state-lasers it is necessary to explore ions such as Er3+, Dy3+, Tb3+, etc. in low-phonon hosts like chlorides, bromides and iodides. Work in this direction is underway and several promising results have been reported recently [468]. Another extension of the operational wavelength would be the wavelength conversion of light from diode-pumped mid-IR crystalline solid-state lasers by means of state-of-the-art OPOs and other nonlinear conversion techniques.

Crystalline Lasers

323

Acknowledgements It is my pleasure to thank a number of colleagues who provided me with information on their recent results, reprints, and figures used in this review: Jason McKay and Ken Schepler (Air Force Research Laboratory), Mike Bass (CREOL), Greg Wagner and Tim Carrig (Coherent Technologies, Inc.), Michael Mond and Stefan K¨ uck (Hamburg University), Uwe Hömmerich (Hampton University), Ray Beach and Steve Payne (Lawrence Livermore National Laboratory), Mikhail Noginov (Norfolk State Unversity), Alberto Di Lieto and Mauro Tonelli (Pisa University), and Sergey Mirov (University of Alabama-Birmingham). Last, but not least, I would like to thank my close colleagues Vladimir Kalashnikov, Sergei Naumov, and Evgeni Sorokin from the Technical University of Vienna for their constant support. Finally, I appreciate the financial support from the Austrian National Science Foundation (FWF), Austrian Ministry of Education, Science and ¨ Culture (BMBWK), Austrian Academy of Science (OAD), and International Science and Technology Center (ISTC).

References 1. R. L. Byer: Diode laser-pumped solid-state lasers, Science 239, 742 (1988) 256, 259 2. T. Y. Fan, R. L. Byer: Diode laser-pumped solid-state lasers, IEEE J. Quantum Electron. 24, 895 (1988) 256 3. A. L. Schawlow, C. H. Townes: Infrared and optical masers, Phys. Rev. 112, 1940–1949 (1958) 257 4. T. H. Maiman: Stimulated optical radiation in ruby, Nature 187, 493–494 (1960) 257 5. T. H. Maiman, R. H. Hoskins, I. H. D’Haenens, C. K. Asawa, V. Evtukhov: Spectroscopy and stimulated emission in ruby, Phys. Rev. 123, 1151 (1960) 257 6. P. P. Sorokin, M. J. Stevenson: Stimulated infrared emission from trivalent uranium, Phys. Rev. Lett. 5, 557–559 (1960) 257, 262 7. Z. J. Kiss, R. C. Duncan: Pulsed and continuous optical maser action in CaF2 :Dy2+, Proc. IRE (Corresp.) 50, 1531–1532 (1962) 257 8. A. Yariv: Continuous operation of a CaF2 :Dy2+ optical maser, Proc. IRE (Corresp.) 50, 1699 (1962) 257 9. Z. J. Kiss, H. R. Lewis, R. C. Duncan: Sun-pumped continuous optical maser, Appl. Phys. Lett. 2, 93–94 (1963) 10. A. A. Kaminskii, L. S. Kornienko, A. M. Prokhorov: Continuous solar laser using Dy2+ in CaF2 , Dokl. Akad. Nauk SSSR 161, 1063–1064 (1965) [English transl. Sov. Phys. Dokl. 10, 334–335 (1965)] 258 11. L. F. Johnson, R. E. Dietz, H. J. Guggenheim: Optical maser oscillation from Ni2+ in MgF2 involving simultaneous emission of phonons, Phys. Rev. Lett. 11, 318–320 (1963) 258, 295, 297 12. L. F. Johnson, R. E. Dietz, H. J. Guggenheim: Spontaneous and stimulated emission from Co2+ ions in MgF2 and ZnF2 , Appl. Phys. Lett. 5, 21–22 (1964) 258, 295, 297

324

Irina T. Sorokina

13. L. F. Johnson, H. J. Guggenheim, R. A. Thomas: Phonon terminated optical masers, Phys. Rev. 149, 179–185 (1966) 258, 295, 297 14. L. F. Johnson, H. J. Guggenheim, D. H. Bahnck: Phonon terminated laser emission from Ni2+ ions in KMgF3 , Opt. Lett. 8, 371–373 (1983) 258, 295 15. P. F. Moulton, A. Mooradian: Broadly tunable CW operation of Ni:MgF2 and Co:MgF2 lasers, Appl. Phys. Lett. 35, 838–840 (1971) 258 16. P. F. Moulton, A. Mooradian, T. B. Reed: Efficient CW optically pumped Ni:MgF2 laser, Opt. Lett. 3, 164–166 (1978) 258, 295 17. D. Welford, P. F. Moulton: Room-temperature operation of a Co:MgF2 laser, Opt. Lett. 13, 975–977 (1988) 258, 297, 299 18. W. Gellermann: Color center lasers, J. Phys. Chem. Solids 52, 249–97 (1991) 258 19. L. F. Mollenauer, R. H. Stolen: The soliton laser, Opt. Lett. 9, 14 (1984) 258 20. S. B. Mirov, T. Basiev: Progress in color-center lasers, IEEE J. Sel. Top. Quantum Electron. 1, 22–30 (1995) 258 21. L. F. Johnson, G. D. Boyd, K. Nassau: Optical maser characteristics of Ho3+ in CaWO4 , Proc. IRE 50, 87 (1962) 259, 314 22. L. F. Johnson, G. D. Boyd, K. Nassau: Optical maser characteristics of Tm3+ in CaWO4 , Proc. IRE 50, 86 (1962) 259, 314 23. L. F. Johnson: Optical maser characteristics of rare-earth ions in crystals, J. Appl. Phys. 34, 897–909 (1963) 314 24. Z. J. Kiss, R. C. Duncan: Optical maser action in CaWO4 :Er3+, Proc. IRE 50, 1531 (1962) 259 25. Yu. K. Voron’ko, A. A. Kaminskii, V. V. Osiko, A. M. Prokhorov: A new type of crystal for optically pumped lasers, Izv. Akad. Nauk SSSR, Neorg. Mater. 2, 1161–1170 (1966) [English transl. Inorg. Mater. (USSR) 991–998 (1966)] 259, 260 26. E. Sorokin, M. H. Ober, I. Sorokina, E. Wintner, A. J. Schmidt, A. I. Zagumennyi, G. B. Loutts, E. W. Zharikov, A. I. Shcherbakov: Femtosecond solid-state lasers using Nd3+-doped mixed scandium garnets, J. Opt. Soc. Am. B 10, 1436–1442 (1993) 259 27. D. E. McCumber: Theory of phonon-terminated optical masers, Phys. Rev. 134, A299–A306 (1964) 259, 284 28. D. E. McCumber: Einstein relations connecting broadband emission and absorption spectra, Phys. Rev. 136, A954–A957 (1964) 259 29. A. A. Kaminskii: Laser Crystals (Springer, 1990) p. 456 259, 278, 319, 321 30. W. Koechner: Solid-State Laser Engineering (Springer, Berlin, Heidelberg 1999) p. 746 259, 286, 287 31. L. F. Johnson, H. J. Guggenheim: Infrared-pumped visible laser, Appl. Phys. Lett. 19, 44–47 (1971) 259 32. A. A. Mak, B. M. Antipenko: Rare-earth converters of neodymium laser radiation Zh. Prikl. Spektrosk. 37, 1029 (1982) [English transl. J. Appl. Spectr. 37, 1458–71 (1982)] 259 33. B. M. Antipenko, A. A. Mak, B. V. Sinizyn, O. B. Raba, T. V. Uvarova: New excitation schemes for laser transitions, Zh. Tekh. Phys. 52, 1982; p. 521–522 (1982) [English transl. Sov. Phys.-Tech. Phys. 27, 333–334 (1982)] 259, 276 34. B. M. Antipenko: Cross-relaxation schemes for pumping laser transitions, Zh. Tekh. Phys. 54, 385–388 (1984) [English transl. Sov. Phys.-Tech. Phys. 29, 228–230 (1984)] 259, 273, 279, 280, 314

Crystalline Lasers

325

35. A. A. Kaminskii: Spectroscopic study of stimulated radiation of Er3+ activated CaF2 -YF3 crystals, Optika-i-Spektroskopiya 31, 938–43 (1971) [English transl. Opt. Spectrosc. 31, 507–510 (1971)] 259 36. L. F. Johnson, L. G. Van Uitert, J. J. Rubin, R. A. Thomas: Energy transfer from Er3+ to Tm3+ and Ho3+ ions in crystals, Phys. Rev. 133, A494–A498 (1964) 259 37. B. Henderson, R. H. Bartram: Crystal-Field Engineering of Solid-State Laser Materials (Cambridge Univ. Press, Cambridge 2000) p. 398 259, 266, 297 38. P. F. Moulton: Spectroscopic and laser characteristics of Ti:Al2 O3 , J. Opt. Soc. Am. B 3, 125–133 (1986) 260, 267, 284, 305 39. J. R. C. Stoneman, L. Esterowitz: Efficient, broadly tunable, laser-pumped Tm:YAG and Tm:YSGG CW lasers, Opt. Lett. 15, 486–488 (1990) 260 40. J. F. Pinto, L. Esterowitz, G. H. Rosenblatt: Tm3+:YLF laser continuously tunable between 2.20 and 2.46 µm, Opt. Lett. 19, 883–885 (1994) 260 41. G. Huber, E. W. Duczynski, K. Petermann: Laser pumping of Ho-, Tm-, Erdoped garnet lasers at room temperature, IEEE J. Quantum Electron. 24, 920–923 (1988) 260, 278, 279, 319 42. T. Y. Fan, G. Huber, R. L. Byer, P. Mitzscherlich: Spectroscopy and diode laser-pumped operation of Tm,Ho:YAG, IEEE J. Quantum Electron. 24, 924– 933 (1988) 260, 279 43. K. L. Vodop’yanov, L. A. Kulevskii, P. P. Pashinin, A. F. Umyskov, I. A. Shcherbakov: Bandwidth-limited picosecond pulses from a YSGG:Cr3+:Er3+ laser (λ = 2.79 µm) with active mode locking, Kvant.Elektron. 14, 1219–24 (1987) [English transl. Sov. J. Quantum Electron. 17, 776–779 (1987)] 260 44. E. Sorokin, I. T. Sorokina, A. Unterhuber, E. Wintner, A. I. Zagumenny, I. A. Shcherbakov, V. Carozza, A. Toncelli, M. Tonelli: A novel CW tunable and mode-locked 2 µm Cr,Tm,Ho:YSGG:GSAG laser, in W. R. Bosenberg, M. M. Fejer (Eds.), OSA Trends Opt. Photonics 19, 197–200 (Opt. Soc. Am., Washington, DC 1998) 260 45. R. H. Page, L. D. DeLoach, G. D. Wilke, S. A. Payne, R. J. Beach, W. F. Krupke: Cr2+-doped II–VI crystals: new widely tunable, room-temperature mid-IR lasers, Lasers and Electro-Optics Society Annual Meeting, 1995, 8th Annual Meeting Conference Proceedings, Vol. 1, IEEE 449–450 (1995) 260, 262, 299, 304 46. L. D. DeLoach, R. H. Page, G. D. Wilke, S. A. Payne, W. P. Krupke: Transition metal-doped zinc chalcogenides: spectroscopy and laser demonstration of a new class of gain media, IEEE J. Quantum Electron. 32, 885–895 (1996) 260, 262, 299, 300, 303, 304, 305, 331 47. R. H. Page, K. I. Schaffers, L. D. DeLoach, G. D. Wilke, F. D. Patel, J. B. Tassano, S. A. Payne, W. F. Krupke, K.-T. Chen, A. Burger: Cr2+-doped zinc chalcogenides as efficient, widely tunable mid-infrared lasers, IEEE J. Quantum Electron. 33, 609–619 (1997) 260, 262, 299, 303, 304, 306 48. G. J. Wagner, T. J. Carrig, R. H. Page, K. I. Schaffers, J. Ndap, X. Ma, A. Burger: Continuous-wave broadly tunable Cr2+:ZnSe laser, Opt. Lett. 24, 19–21 (1999) 260, 300, 304, 306 49. S. K¨ uck: Laser-related spectroscopy of ion-doped crystals for tunable solidstate lasers, Appl. Phys. B. 72, 515–562 (2001) 260, 272, 280

326

Irina T. Sorokina

50. B. F. Aull, H. P. Jenssen: Impact of ion–host interactions on the 5d-to-4f spectra of lanthanide rare-earth-metal ions. I. A phenomenological crystal-field model, Phys. Rev. B 34, 6640–6 (1986) 260 51. B. F. Aull, H. P. Jenssen: Vibronic interactions in Nd:YAG resulting in nonreciprocity of absorption and stimulated emission cross sections, IEEE J. Quantum Electron. 18, 925–930 (1982) 260 52. E. V. Zharikov, A. I. Zagumennyi, G. B. Lutts, V. A. Smirnov, I. T. Sorokina, I. A. Shcherbakov: New laser crystals – solid solutions based on scandiumcontaining aluminum-gallium garnets, doped with chromium and neodymium, Laser Phys. 1, 216–219 (1991) 260, 315 53. M. J. Weber: Handbook of Lasers (CRC, Boca Raton, FL 2000) p. 1198 261, 315 54. S. Hufner: Optical Spectra of Transparent Rare-Earth Compounds (Academic Press, New York 1978) 261 55. B. Henderson, G. F. Imbusch: Optical Spectroscopy of Inorganic Solids (Clarendon Press, Oxford 1989) 261, 264, 265, 266, 270 56. G. H. Dieke: Spectra and Energy Levels of Rare-Earth Ions in Crystals (WileyInterscience, New York 1968) 261, 263 57. B. R. Judd: Optical absorption intensities of rare-earth ions, Phys. Rev. 127, 750–761 (1962) 261 58. B. G. Wybourne: Spectroscopic Properties of Rare-Earths (Wiley-Interscience, New York 1965) 261, 263 59. G. D. Boyd, R. J. Collins, S. P. S. Porto, A. Yariv, W. A. Hargreaves: Excitation, relaxation and continuous maser action in the 2.613-micron transition of CaF2 :U3+, Phys. Rev. Lett. 8, 269–272 (1962) 262 60. S. P. S. Porto, A. Yariv: Optical maser characteristics of BaF2 :U3+, Proc. IRE 50, 1542–1543 (1962) 262 61. S. P. S. Porto, A. Yariv: Excitation, relaxation and optical maser action at 2.407 micron in SrF2 :U3+, Proc. IRE 50, 1543–1544 (1962) 262 62. D. Meichenin, F. Auzel, S. Hubert, E. Simoni, M. Louis, J. Y. Gesland: New room-temperature CW laser at 2.82 µm U3+/LiYF4 , Electron. Lett. 30, 1309 (1994) 262 63. H. P. Jenssen, M. A. Noginov, A. Cassanho: U:YLF, a prospective 2.8 µm laser crystal, OSA Proc. Adv. Solid-State Lasers 15, 463–467 (Opt. Soc. Am., Washington, DC 1993) 262 64. G. Menzer: Die Kristallstruktur von Granat, Centralbl. Min. A, 344–345 (1925); Z. Kristallogr. 63, 157–158 (1926) (in German) 262 65. G. Menzer: Die Kristallstruktur der Granate, Z. Kristallogr. 69, 300–396 (1928) (in German) 262 66. S. Geller: Crystal chemistry of the garnets, Z. Kristallogr. 125, 1–47 (1967) 262, 264 67. S. C. Abrahams, S. Geller: Refinement of the structure of a grossularite garnet, Acta Crystallogr. 11, 437–441 (1958) 68. W. Nie, Y. Kalisky, C. Pedrini, A. Monteil, G. Boulon: Energy transfer from Cr3+ multisites to Tm3+ multisites in yttrium aluminum garnet, Opt. Quantum Electron. 22, 123–131 263 69. Y. Kalisky, S. R. Rotman, A. Brenier, C. Pedrini, G. Boulon, F. X. Hartman: Spectroscopic properties, multisite dynamics, and operation of pulsed Ho laser, OSA Proc. Adv. Solid-State Lasers 13, 144–147 (Opt. Soc. Am., Washington, DC 1992) 263

Crystalline Lasers

327

70. V. Lupei: RE3+ emission in garnets: multisites, energy transfer and quantum efficiency, Opt. Mater. 19, 95–107 (2002) 263 71. H. Bethe: Thermaufspaltung in Kristallen, Ann. Phys. (Leipzig) 3, 133–208 (1929) 264 72. J. H. Van Vleck: The puzzle of rare-earth spectra in solids, J. Phys. Chem. 41, 67–80 (1937) 264 73. J. H. Van Vleck, A. Sherman: The quantum theory of valence, Rev. Mod. Phys. 7, 167–228 (1935) 264 74. C. J. Ballhausen: Introduction to Ligand Field Theory (McGraw-Hill, New York 1962) 264 75. P. Schuster: Ligandenfeldtheorie (Verlag Chemie, Weinheim 1973) 264 76. Y. Tanabe, S. Sugano: On the absorption spectra of complex ions, Part I, J. Phys. Soc. Jap. 9, 753–766 (1954); Part II, J. Phys. Soc. Jap. 9, 766–779 (1954) 265, 266 77. S. Sugano, Y. Tanabe, H. Kamimura: Multiplets of Transition-Metal Ions in Crystals (Academic Press, New York 1970) 265 78. H. A. Jahn, E. Teller: Stability of polyatomic molecules in degenerate electronic states. I. Orbital degeneracy, Proc. R. Soc. London A 161, 220–235 (1937) 267, 303 79. R. Engleman: The Jahn–Teller Effect in Molecules and Crystals (WileyInterscience, New York 1972) 267, 303 80. I. B. Bersuker: The Jahn–Teller Effect and Vibronic Interactions in Quantum Chemistry (Plenum, New York 1984) 267 81. M. Kaminska, J. M. Baranovski, S. M. Uba, J. T. Vallin: Absorption and luminescence of Cr2+ (d4 ) in II–VI compounds, J. Phys. C 12, 2197–2214 (1979) 267, 301, 303 82. J. T. Vallin, G. A. Slack, S. Roberts, A. E. Hughes: Infrared absorption in some II-VI compounds doped with Cr, Phys. Rev. B 2, 4313–4333 (1970) 267, 301 83. S. W. Biernacki, Jahn–Teller coupling of Cr2+ ion in ZnS, ZnSe and ZnTe, Phys. Stat. Sol. (b) 87, 607–612 (1978) 267 84. A. L. Natadze, A. I. Ryskin, Jahn–Teller coupling of Cr2+ ion with degenerate modes in ZnS, ZnSe, and ZnTe crystals: microscopic treatment, Phys. Stat. Sol. (b) 97, 175–185 (1980) 267 85. G. Goetz, H. Zimmermann, H. J. Schulz, Jahn–Teller interaction at Cr2+ (d4 ) centres in tetrahedrally coordinateqd II-VI lattices studied by optical spectroscopy, Z. Physik B 91, 429–436 (1993) 267, 303 86. F. S. Ham: Dynamical Jahn–Teller effect in paramagnetic resonance spectra: Orbital reduction factors and partial quenching of spin–orbit interaction, Phys. Rev. 138, A1727 (1965) 267 87. F. S. Ham: Effect of linear Jahn–Teller coupling on paramagnetic resonance in a 2 E state, Phys. Rev. 166, 307 (1968) 267 88. M. D. Sturge: in Advances in research and applications, Vol. 20, Solid State Physics, F. Seitz, D. Turnbull, H. Ehrenreich (Eds.): (Academic Press, New York 1967) p. 91 267 89. B. Nygren, J. T. Vallin, G. A. Slack: Direct observation of the Jahn–Teller splitting in ZnSe:Cr2+, Solid State Commun. 11, 35 (1972) 267, 303 90. O. Laporte: Die Struktur des Eisenspektrums, Z. Phys. 23, 135 (1924) 268 91. S. A. Payne, L. L. Chase, H. W. Newkirk, L. K. Smith, W. F. Krupke: LiCaAlF6 :Cr3+: A promising new solid-state laser material, IEEE J. Quantum Electron. 24, 2243–2252 (1988) 268

328

Irina T. Sorokina

92. L. E. Orgel: An Introduction to Transition-Metal Chemistry Ligand-Field Theory (Methuen, London 1966) 268 93. A. Zunger, U. Lindefelt: Electronic structure of transition-atom impurities in semiconductors: Substitutional 3d impurities in silicon, Phys. Rev. B 27, 1191 (1983) 269 94. V. Singh, A. Zunger: Electronic structure of transition-atom impurities in GaP, Phys. Rev. B 31, 3729 (1985) 269 95. I. T. Sorokina, S. Naumov, E. Sorokin, A. G. Okhrimchuk: The mechanisms of slow bleaching in YAG:Cr4+ under CW pumping, in Laser Optics 2000, V. I. Ustyugov (Ed.), Proc. SPIE 4350, 99–105 (2001) 269 96. I. T. Sorokina, S. Naumov, E. Sorokin, E. Wintner, A. V. Shestakov: Directly diode-pumped tunable continuous-wave room-temperature Cr4+:YAG laser, Opt. Lett. 24, 1578–1580 (1999) 269, 284 97. E. Sorokin, I. T. Sorokina: Tunable diode-pumped continuous-wave Cr2+:ZnSe laser, Appl. Phys. Lett. 80, 3289–3291 (2002) 269, 284, 308 98. I. T. Sorokina, E. Sorokin, S. Mirov, V. Fedorov, V. Badikov, V. Panyutin, K. Schaffers: Broadly tunable compact continuous-wave Cr2+:ZnS laser, Opt. Lett. 27, 140–142 (2002) 269, 285, 304, 305, 308 99. H. V. Lauer, F. K. Fong: Coupling strength in the theory of radiationless transitions: f to f and d to f relaxation of rare earth ions in YAlO3 and Y3 Al5 O12 , J. Chem. Phys. 60, 274–280 (1974) 270 100. W. H. Fonger, C. W. Struck: Unified model of energy transfer for arbitrary Franck–Condon offset and temperature, J. Lumin. 17, 241–261 (1978) 270 101. M. Yamaga, Y. Gao, F. Rasheed, K. P. O’Donnel, B. Henderson, B. Cockayne: Radiative and non-radiative decays from the excited state of Ti3+ ions in oxide crystals, Appl. Phys. B 51, 329 (1990) 270 102. T. T. Basiev, A. Yu. Dergachev, E. O. Kirpichenkova, Yu. V. Orlovskii, V. V. Osiko: Direct measurement of the rate of nonradiative relaxation and luminescence spectra from the 4 G7/2 , 4 G 5/2 + 2 G7/2 , and 4 F9/2 levels of Nd 3+ ions in LaF3 , SrF2 , and YAlO3 laser crystals, Sov. J. Quantum Electron. 17, 1289–1291 (1987) 270 103. Yu. V. Orlovskii, R. J. Reeves, R. C. Powell, T. T. Basiev, K. K. Pukhov: Multiple-phonon nonradiative relaxation: Experimental rates in fluoride crystals doped with Er3+ and Nd3+ ions and a theoretical model, Phys. Rev. B 49, 3821–3830 (1994) 270 104. Yu. V. Orlovskii, K. K. Pukhov, T. T. Basiev, T. Tsuboi: Nonlinear mechanism of multiphonon relaxation of the energy of electronic excitation in optical crystals doped with rare-earth ions, Opt. Mat. 4, 583–595 (1995) 270 105. T. T. Basiev, Yu. V. Orlovskii, K. K. Pukhov, V. B. Sigachev, M. E. Doroshenko, I. N. Vorob’ev: Multiphonon relaxation rate measurements and theoretical calculations in the frame of non-linear and non-Coulomb model of a rare-earth ion-ligand interaction, J. Lumin. 68, 241–253 (1996) 270 106. T. T. Basiev, Yu. V. Orlovskii, K. K. Pukhov, F. Auzel: Multiphonon relaxation of the energy of electronic excitation in optical crystals doped with rare-earth ions, Laser Phys. 7, 1139–1152 (1997) 270 107. Yu. V. Orlovskii, T. T. Basiev, I. N. Vorob’ev, E. O. Orlovskaya, N. P. Barnes, S. B. Mirov: Temperature dependencies of excited state lifetimes and relaxation rates of 3–5 phonon (4–6 µm) transitions in the YAG, LuAG and YLF

Crystalline Lasers

108.

109. 110.

111. 112. 113. 114. 115.

116.

117.

118. 119.

120.

121.

122.

123. 124. 125.

329

crystals doped with trivalent holmium, thulium, and erbium, Opt. Mater. 18, 355–365 (2002) 270 T. T. Basiev, Yu. V. Orlovskii, B. I. Galagan, M. E. Doroshenko, I. N. Vorob’ev, L. N. Dmitruk, A. G. Papashvili, V. N. Skvortsov, V. A. Konyushkin, K. K. Pukhov, G. A. Ermakov, V. V. Osiko, A. M. Prokhorov, S. Smith: Evaluation of rare-earth doped crystals and glasses for 4–5-µm lasing, Laser Phys. 12, 859–877 (2002) 270 F. Pelle, F. Auzel: Phonon bottleneck effect in multiphonon non-radiative transitions of rare-earth ions, J. Lumin. 76–77, 623–627 (1998) 270 F. Pelle, N. Gardant, F. Auzel: Effect of excited-state population density on nonradiative multiphonon relaxation rates of rare-earth ions, J. Opt. Soc. Am. B 15, 667–679 (1998) 270 F. Pelle, N. Gardant, F. Auzel: Saturation effect on multiphonon relaxation rates, J. Alloys Compounds 275/277, 430 (1998) 270 A. Einstein: Zur Quantentheorie der Strahlung, Phys. Z. 18, 121–128 (1917) 283 O. Svelto: Principles of Lasers (Plenum, New York 1998) 272, 283, 286, 293 A. E. Siegman: Lasers (Cambridge Univ. Press, Cambridge 1986) 283, 286, 293 J. Koetke, G. Huber: Infrared excited-state absorption and stimulatedemission cross sections of Er3+-doped crystals, Appl. Phys. B 61, 151 (1995) 271, 272 I. T. Sorokina, E. Sorokin, E. Wintner, A. Cassanho, H. P. Jenssen: In situ measurement of ESA, upconversion, and thermal quenching in Cr:LiSAF and Cr:LiSGaF lasers, in OSA Trends Opt. Photonics 10, 411–414, C. R. Pollock, W. R. Bosenberg (Eds.) (Opt. Soc. Am., Washington, DC 1997) 272 S. Naoumov, I. T. Sorokina, E. Sorokin: Measurement of the excited state absorption cross-section in Cr4+:YAG using relaxation oscillation study, paper CtuK38, Conference on Lasers and Electro-Optics Europe, Nice, France; in CLEO Europe’2000 Tech. Digest (2000) p. 105 272 V. V. Ovsyankin, P. P. Feofilov: On the mechanism of adding of electronic excitations in doped crystals, Sov. Phys. JETP Lett. 3, 322–323 (1966) 273 V. V. Ovsyankin, P. P. Feofilov: Cooperative sensitization of luminescence in crystals activated with rare earth ions, Sov. Phys. JETP Lett. 4, 317–318 (1966) 273 F. Auzel: Compteur quantique par transfert d’energie entre deux ions de terres reres dans un tungstate mixte et dans un verre, C. R. Acad. Sci. Paris B 262, 1016–1019 (1966) 273 F. Auzel: Compteur quantique par transfert d’energie entre de Yb3+ a Tm3+ dans un tungstate mixte et dans verre germanate, C. R. Acad. Sci. Paris B 263, 819–821 (1966) 273 V. V. Ovsyankin, P. P. Feofilov: Cooperative processes in luminescent systems (in Russian), Izv. Akad. Nauk, Ser. Fiz. 37, 262–272 (1973) [English transl. Bulletin Acad. Sci. USSR, Phys. Ser.] 273 D. L. Dexter: Possibility of luminescence quantum yields greater than unity, Phys. Rev. 108, 630–633 (1957) 273 N. Bloembergen: Solid-state infrared quantum counter, Phys. Rev. Lett. 2, 84–85 (1959) 273 John F. Porter, Jr.: Fluorescence excitation by the absorption of two consecutive photons, Phys. Rev. Lett. 7, 414–415 (1961) 273

330

Irina T. Sorokina

126. J. S. Chivian, W. E. Case, D. D. Eden: The photon avalanche: a new phenomenon in Pr3+ -based infrared quantum counters, Appl. Phys. Lett. 35, 124–125 (1979) 273 127. V. A. Benderskiy, V. Kh. Brikensteyn, M. A. Kozhushner, I. A. Kuznetsova, P. G. Filippov: Nonlinear fluorescence quenching of high density localized electron excitations in molecular crystals, Sov. Phys. JETP 43, 268 (1976) 273, 274 128. M. A. Noginov, H. P. Jenssen, A. Cassanho: Upconversion in Cr:LiSGaF and Cr:LiSAF, OSA Proc. Adv. Solid-State Lasers 15, 376–380, A. A. Pinto: Tso Ye Fan (Eds.) (Opt. Soc. Am., Washington, DC 1993) 274 129. V. Ostroumov, T. Jensen, J.-P. Meyn, G. Huber, M. A. Noginov: in Concentration Quenching and Upconversion of Neodymium Ions in LaSc3 (BO3 )4 and GdVO4 Crystals, OSA Proc. Adv. Solid-State Lasers 24, 509–513, B. H. T. Chai, S. A. Payne (Eds.) (Opt. Soc. Am., Washington, DC 1995) 274 130. D. A. Zubenko, M. A. Noginov, V. A. Smirnov, I. A. Scherbakov: Different mechanisms of nonlinear quenching of luminescence, Phys. Rev. B 55, 8881– 8886 (1997) 274 131. B. M. Antipenko, B. A. Buchenkov, A. S. Glebov, T. I. Kiseleva, A. A. Nikitichev, V. A. Pismennyi: Spectroscopy of YAG:CrTmHo laser crystals, Opt. Spectrosc. 64, 772–774 (1988) 274 132. B. M. Antipenko, A. S. Glebov, T. I. Kiseleva, V. A. Pismennyi: Interpretation of the temperature dependence of the YAG:Cr+Tm+Ho lasing threshold, Opt. Spectrosc. 63, 230–232 (1988) 274 133. Kh. S. Bagdasarov, V. I. Zhekov, V. A. Lobachev, T. M. Murina, A. M. Prokhorov: Steady-state emission from a Y3 Al5 O12 :Er3+ laser (λ = 2.94µ, T = 300 K), Sov. J. Quantum Electron. 13, 262–263 (1983) 274 134. M. A. Noginov, V. A. Smirnov, I. A. Shcherbakov: Non-linear population processes of Er3+ laser levels in chromium-doped garnet crystals, Opt. Quantum Electron. 22, S61–S74 (1990) 274 135. S. A. Pollack, D. B. Chang, N. L. Moise: Continuous wave and Q-switched infrared erbium laser, Appl. Phys. Lett. 49, 1578–80 (1986) 274 136. D. A. Zubenko, M. A. Noginov, V. A. Smirnov, I. A. Shcherbakov: Interaction of excited holmium and thulium ions in yttrium scandium garnet crystals, J. Appl. Spectrosc. 52, 391–394 (1990) 274 137. D. A. Zubenko, M. A. Noginov, S. G. Semenkov, V. A. Smirnov, I. A. Shcherbakov: Interionic interactions in YSGG:Cr:Tm and YSGG:Cr:Tm:Ho laser crystals, Sov. J. Quantum Electron. 22, 133–138 (1992) 274 138. D. A. Zubenko, M. A. Noginov, V. G. Ostroumov, V. A. Smirnov, I. A. Shcherbakov: Interaction of excited Cr3+ ions in laser crystals, J. Appl. Spectrosc. 56, 56–59 (1992) 274 139. P. P. Feofilov: in Physics of Impurity Centers in Crystals T. S. Zavt (Ed.) (Academy of Sciences of the Estonian SSR, Tallin 1972), p. 539 274 140. W. Lenth, R. M. Macfarlane: Upconversion lasers, Opt. Photonics News, 8–15 (March 1992) 274 141. B. DiBartolo (Ed.): Optical Properties of Ions in Solids (Plenum, New York 1975) 274 142. M. D. Agranovich, M. D. Galanin: Electron Excitation Energy Transfer in Condensed Media, (Nauka, Moscow 1978) p. 383 274, 275

Crystalline Lasers

331

143. B. DiBartolo (Ed.): Energy Transfer Processes in Condensed Matter (Plenum, New York 1983) 274 144. A. A. Kaplyanski, R. M. MacFarlane (Eds.): Spectroscopy of Rare-Earth Ions in Crystals (North-Holland, Amsterdam 1987) 274 145. A. I. Burstein: Concentration quenching of noncoherent excitations in solids, Sov. Phys. Usp. 27, 579–606 (1984) 275 146. T. F¨ orster: Zwischenmolekulare Energiewanderung und Fluoreszenz, An. Phys. 2, 55–77 (1948) 275 147. Th. F¨ orster: Experimentelle und theoretische Untersuchung des zwischen¨ molekularen Ubergangs von Elektronenanregungsenergie, Z. Naturforsch. 4, 321–327 (1949) 275 ¨ 148. H. Kallman, F. London: Uber quantenmechanische Energie¨ ubertragung zwischen atomaren Systemen, Z. Phys. Chem. 2, 207–243 (1929) 275 149. D. L. Dexter: A theory of sensitized luminescence in solids, J. Chem. Phys. 21, 835–850 (1953) 275 150. M. Inokuti, F. Hirayama: Influence of energy transfer by the exchange mechanism on donor luminescence, J. Chem. Phys. 43, 1978–1989 (1965) 275 151. M. D. Galanin: Resonant energy transfer of excitation in solutions, Trudy FIAN 12, 3–35 (1960) (in Russian) 275 152. A. S. Agabekyan: Energy transfer in strong incoherent interactions, J. App. Spectrosc. 34, 465–468 (1981) 275 153. A. I. Burstein: Hopping mechanism of energy transfer, Soviet Phys. JETP 35, 882–885 (1972) 275 154. A. Brenier, J. Rubin, R. Moncorge, C. Pedrini: Excited-state dynamics of the Tm3+ ions and Tm3+ to Ho3+ energy transfers in LiYF4 , J. Phys. (Paris) 50, 1463–1482 (1989) 275 155. V. K. S. Shante, S. Kirkpatrick: An introduction to percolation theory, Adv. Phys. 20, 325–357 (1971) 275 156. E. A. Milne: The diffusion of imprisoned radiation through a gas, J. London Math. Soc. 1, 40 (1926) 275 157. A. C. G. Mitchell, M. W. Zemansky: Resonance Radiation and Excited Atoms (Cambridge Univ. Press, Cambridge 1961) 275 158. F. Varsanyi, D. l. Wood, A. L. Schawlow: Self-absorption and trapping of sharp-line resonance radiation in ruby, Phys. Rev. Lett. 3, 544–545 (1959) 275 159. T. T. Basiev, Yu. K. Voron’ko, V. V. Osiko, A. M. Prokhorov, I. A. Shcherbakov: Experimental observation of ‘excitation trapping’ in a system of strongly interacting particles, Zhur. Eksp. Teoretich. Fiz. 39, 1042 (1974) [English transl. Sov. Phys. JETP] 275 160. P. Lacovara, H. K. Choi, C. A. Wang, R. L. Aggarwal, T. Y. Fan: Roomtemperature diode-pumped Yb:YAG laser, Opt. Lett. 16, 1089 (1991) 275 161. D. S. Sumida, T. Y. Fan: Effect of radiation trapping on fluorescence lifetime and emission cross section measurements in solid-state laser media, Opt. Lett. 19, 1343–1345 (1994) 275, 303 162. S. A. Payne, one of the authors of the original publication [46], shared this view 303 163. M. P. Hehlen: Reabsorption artifacts in measured excited-state lifetimes of solids, J. Opt. Soc. Am. B 14, 1312–1318 (1997) 275, 303

332

Irina T. Sorokina

164. M. A. Noginov: Reabsorption trapping of luminescence in laser crystals, enhancement of energy storage and upconversion, Appl. Opt. 36, 4153–4158 (1997) 275, 276, 277, 278 165. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. H. Page, K. I. Schaffers: Efficient broadly tunable continuous-wave Cr2+:ZnSe laser, J. Opt. Soc. Am. B 18, pp. 926–930 (2001) 267, 276, 299, 300, 304, 305, 306 166. J. A. Caird, S. A. Payne, P. R. Staver, A. J. Ramponi, L. L. Chase, W. F. Krupke: Quantum electronic properties of the Na3 Ga2 Li3 F12 :Cr3+ laser, IEEE J. Quantum Electron. 24, 1077–1099 (1988) 276, 307 167. D. Findlay, R. A. Clay: The measurement of internal losses in 4-level lasers, Phys. Lett. 20, 277 (1966) 276 168. V. A. Smirnov, I. A. Shcherbakov: Rare-earth scandium chromium garnets as active media for solid-state lasers, IEEE J. Quantum Electron. 24, 949–959 (1988) 276 169. I. A. Shcherbakov: Optically dense active media for solid-state lasers, IEEE J. Quantum Electron. 24, 979–984 (1988) 276 170. L. F. Johnson, H. J. Guggenheim: New laser lines in the visible from Er3+ ions in BaY2 F8 , Appl. Phys. Lett. 20, 474–477 (1972) 276 171. B. M. Antipenko: A cooperative mechanism of excitation of Ho3+ lasing ions in BaYb2 F8 , Soviet Tech. Phys. Lett. 6, 417–418 (1980) [transl. from: Pis’ma v Zhurnal Tekhnicheskoi Fiz. 6, 968–72 (1980)] 276 172. L. Esterowitz: Diode-pumped holmium, thulium, and erbium lasers between 2 and 3 µm operating CW at room temperature, Opt. Engin. 29, 676–680 (1990) 278 173. V. I. Zhekov, B. V. Zubov, V. A. Lobachev, T. M. Murina, A. M. Prokhorov, A. F. Shevel: Mechanism of a population inversion between the 4 I11/2 and 4 I13/2 levels of the Er3+ ion in Y3 Al5 O12 crystals, Soviet J. Quantum Electron. 10 (4), 428–30 (1980) [transl. from: Kvant. Elektron. 7(4), 749–53 (1980)] 277, 278 174. V. I. Zhekov, V. A. Lobachev, T. M. Murina, A. M. Prokhorov: Cooperative phenomena in yttrium erbium aluminium garnet crystals, Soviet J. Quantum Electron. 11(1), 128–30 (1984) [transl. from: Kvant. Elektron. 14(1), 189–92 (1984)] 277 175. V. I. Zhekov, T. M. Murina, A. M. Prokhorov, M. I. Studenikin, S. Georgescu, V. Lupei, I. Ursu: Cooperative process in Y3 Al5 O12 :Er3+ crystals, Soviet J. Quantum Electron. 16(2), 274–276 (1986) [transl. from: Kvant. Elektron 13(2), 419–422 (1986)] 277, 278 176. V. I. Zhekov, V. A. Lobachev, T. M. Murina, A. M. Prokhorov: Spectrum of stimulated emission due to self-saturating transitions in high-concentration media, Soviet J. Quantum Electron. 11(2), 279–81 (1981); [transl. from: Kvant. Elektron. 8(2), 451–454 (1981)] 277 177. M. A. Noginov, H. P. Jenssen: Reabsorption trapping: enhancement of energy storage and upconversion, in Digest IEEE Lasers Electro-Optics Society 1992 Annual Meeting (Institute Electrical Electronics Engineers, New York 1992), postdeadline paper PD9 278 178. E. V. Zharikov, N. N. Il’ichev, S. P. Kalitin, V. V. Laptev, A. A. Malyutin, ˙ Osiko, P. P. Pashinin, A. M. Prokhorov, Z. S. Saidov, V. A. Smirnov, V.V. A. F. Umyskov, I. A. Shcherbakov: Spectral, luminescence, and lasing properties of a yttrium scandium gallium garnet crystal activated with chromium

Crystalline Lasers

179.

180.

181. 182.

183.

184. 185.

186.

187. 188.

189.

190.

191.

192.

333

and erbium, Sov. J. Quantum Electron. 16(5), 635–639 (1986) [transl. from: Kvant. Elektron. 13(5), 973–979 (1986)] 278 Yu. K. Voron’ko, S. B. Gessen, I. V. Gribkov, A. A. Kiryukhin, S. V. Lavrishev, D. I. Melikhov, V. V. Osiko, A. A. Sobol’, V. M. Tatarintsev, S. N. Ushakov, L. I. Tsymbal: Laser utilizing erbium-activated gadolinium aluminum scandium garnet crystals coactivated with Cr3+ and emitting in the three-micron wavelength region, Sov. J. Quantum Electron. 19(9), 1147–1148 (1989) [transl. from: Kvant. Elektron. 16(9), 1785–1786 (1989)] 278 A. A. Kaminskii, T. I. Butaeva, A. O. Ivanov, I. V. Mochalov, A. G. Petrosyan, G. I. Rogov, V. A. Fedorov: New data for stimulated radiation of crystals with Er3+ and Ho3+ ions, Pis’ma Zhur. Techn. Fiz. 2, 787–793 (1976) [English transl. in: Sov. Tech. Phys. Lett. 2, 308 (1976)] 278 S. A. Pollack, D. B. Chang, N. L. Moise: Upconversion-pumped infrared erbium laser, J. Appl. Phys. 60, 4077–4086 (1986) 278, 320 S. A. Pollack, D. B. Chang: Ion-pair upconversion pumped laser emission in Er3+ ions in YAG, YLF, SrF2 , and CaF2 crystals, J. Appl. Phys. 64, 2885– 2893 (1988) 278 A. A. Kaminskii, B. P. Sobolev, S. E. Sarkisov, V. A. Fedorov, V. V. Ryabchenkov, T. V. Uvarova: A new self-activated crystal for producing three-micron stimulated emission, Inorg. Mater. (USSR) 17, 829–830 (1981) [Transl. from: Izv. Akad. Nauk SSSR, Neorganich. Mater. 17, 1121–1122 (1981)] 278 M. J. Weber: Handbook of Lasers (CRC Press, Boca Raton, FL 2000) 278 B. M. Antipenko, A. A. Mak, L. K. Sukhareva: Cross-relaxation BaEr2 F8 :Tm,Ho laser, Soviet Technical Physics Letters 10 (5), 217–218 (1984) [transl. from: Pis’ma Zhur. Tekhnich. Fiz. 10, 513–517 (1984)] 278 L. F. Johnson, J. E. Geusic, L. G. Van Uitert: Coherent oscillations from Tm3+ , Ho3+ , Yb3+ , Er3+ ions in YAG, Appl. Phys. Lett. 7, 127–128 (1965) 278, 279 L. G. VanUitert, E. F. Dearborn, J. J. Rubin: J. Chem. Phys. 46, 3551 (1967) 278 B. M. Antipenko, A. S. Glebov, R. V. Dumbravyanu, B. P. Sobolev, T. V. Uvarova: Spectroscopy and lasing characteristics of BaEr2 F8 :Tm,Ho crystals, Sov. J. Quantum Electron. 17 (4), 424–427 (1987) [transl. from: Kvant. Elektron. 14(4), 677–681 (1987)] 278 B. M. Walsh, K. E. Muray, N. P. Barnes: Cr:Er:Tm:Ho:yttrium aluminum garnet laser exhibiting dual wavelength lasing at 2.1 and 2.9 µm: Spectroscopy and laser performance, J. Appl. Phys. 91, 11–17 (2002) 279, 316 B. M. Antipenko, A. S. Glebov, T. I. Kiseleva, V. A. Pismennyi: 2.12 µm HoYAG laser, Soviet Tech. Phys. Lett. 11 (6), 284–285 (1985) [transl. from: Pis’ma Zhur. Tekhnich. Fiz. 11, 682–5 (1985)] 279 P. Albers, E. Stark, G. Huber: Continuous-wave laser operation and quantum efficiency of titanium-doped sapphire, J. Opt. Soc. Am. B 3, 134–139 (1986) 284, 305 B. M. Antipenko, A. S. Glebov, T. I. Kiseleva, V. A. Pismennyi: A new spectroscopic scheme of an active medium for the 2-µm band, Opt. Spectrosc. 60(1), 95–97 (1986) [transl. from: Optika I Spektroskopiya 60(1), 153–157 (1986)] 279

334

Irina T. Sorokina

193. E. W. Duczynski, G. Huber, V. G. Ostroumov, I. A. Shcherbakov: CW double cross pumping of the 5 I7 →5 I8 laser transition in Ho3+-doped garnets, Appl. Phys. Lett. 48, 1562–1563 (1986) 279 194. B. M. Antipenko, A. S. Glebov, L. I. Krutova, V. M. Solntsev, L. K. Sukhareva: Active medium of lasers operating in the 2-µm spectral range and utilizing gadolinium scandium gallium garnet crystals, Sov. J. Quantum Electron. 16, 1986 [transl. from: Kvant. Elektron. 13, 1521–3 (1986)] 279 195. B. M. Antipenko, L. I. Krutova, L. K. Sukhareva: Cascade lasing of CSGGCr3+Tm3+Ho3+ crystals, Opt. Spectrosc. 61, 414–416 (1986) [transl. from: Optika I Spektroskopiya 61(3), 659–661 (1986)] 279 196. I. T. Sorokina, A. F. Umyskov, V. A. Smirnov, I. A. Shcherbakov: The Tm– Tm and Tm–Ho energy transfer studies in YSGG and GSAG crystals at the resonant excitation of Tm and Ho ions, OSA Proc. Adv. Solid State Lasers 17, 159–161, T. Y. Fan, B. E. Chai (Eds.) (Opt. Soc. Am., Washington, DC 1994) 279, 280 197. A. N. Alpat’ev, E. V. Zharikov, S. P. Kalitin, V. V. Laptev, V. G. Ostroumov, A. M. Prokhorov, Z. C. Saidov, V. A. Smirnov, I. T. Sorokina, A. F. Umyskov, I. A. Shcherbakov: The room-temperature laser action on 5 I7 →5 I8 transition of holmium ions in yttrium-scandium-gallium garnet with chromium, thulium, and holmium ions (YSGG:Cr3+, Ho3+, Tm3+ ), Sov. J. Quant. Electron. 16, 1404 (1986) 279 198. A. I. Feodorov, G. B. Loutts, M. A. Noginov, I. A. Shcherbakov, V. A. Smirnov, I. T. Sorokina, V. B. Tsvetkov, A. I. Zagumennyi, E. V. Zharikov, D. A. Zubenko: New promising low Sc3+ content garnet YSAG:Cr3+,Ho3+,Tm3+ for 2-µm lasers, OSA Proc. Adv. Solid State Lasers 13, 148–151, L. L. Chase, A. A. Pinto (Eds.) (Opt. Soc. Am., Washington, DC 1992) 279 199. E. Sorokin, I. T. Sorokina, A. Unterhuber, E. Wintner, A. I. Zagumenny, I. A. Shcherbakov, V. Carozza, A. Toncelli, M. Tonelli: A novel CW tunable and mode-locked 2 µm Cr,Tm,Ho:YSGG:GSAG laser, OSA Trends Opt. Photonics 19, 197–200, W. R. Bosenberg, M. M. Fejer (Eds.) (Opt. Soc. Am., Washington, DC 1998) 279 200. N. P. Barnes, E. D. Filer, F. L. Naranjo, W. J. Rodriguez, M. R. Kokta: Spectroscopy and lasing properties of Ho:Tm:LuAG, Opt. Lett. 18, 708–10 (1993) 279 201. H. Hemmati: 2.07 µm CW diode-laser-pumped Tm,Ho:YLiF4 roomtemperature laser, Opt. Lett. 14, 435 (1989) 279 202. S. A. Payne, L. K. Smith, W. L. Kway, J. B. Tassano, W. F. Krupke: The mechanism of Tm to Ho energy transfer in LiYF4 , J. Phys. 4, 8525–8542 (1992) 279, 280 203. M. A. Noginov, S. G. Semenkov, I. A. Shcherbakov, V. A. Smirnov: Energy transfer (Tm to Ho) and upconversion process in YSGG:Cr3+:Tm3+:Ho3+ laser crystals, OSA Proc. Adv. Solid State Lasers 10, 178–82 (Opt. Soc. Am., Washington, DC 1991) 281 204. E. Wintner, F. Krausz, M. A. Noginov, V. A. Smirnov, I. T. Sorokina, C. Spielmann, I. A. Shcherbakov: Interaction of Ho3+ and Tm3+ ions in YSGG:Cr3+,Ho3+,Tm3+ at the strong selective excitation, Laser Phys. 2, 138 (1992) 281 205. R. R. Petrin, M. G. Jani, R. C. Powell, M. Kokta: Spectral dynamics of laserpumped Y3 Al5 O12 :Tm,Ho lasers, Opt. Mater. 1, 111–24 (1992) 281

Crystalline Lasers

335

206. B. M. Antipenko, A. S. Glebov, T. I. Kiseleva, V. A. Pis’mennyi: Conversion of absorbed energy in YAG:Cr3+ ,Tm3+ ,Ho3+ crystals, Opt. Spectrosc. 64, 221–224 (1988) [transl. from: Opt. Spektrosk. 64, 373–377 (1988)] 281 207. I. T. Sorokina, E. Sorokin, E. Wintner, A. Cassanho, H. P. Jenssen, M. Noginov: Efficient continuous-wave TEM00 and femtosecond Kerr-lens mode-locked Cr:LiSrGaF laser, Opt. Lett. 21, 204–206 (1996) 281, 282, 284 208. A. G. Okhrimchuk, L. N. Butvina, E. M. Dianov, N. V. Lichkova, V. N. Zavgorodnev, E. Sorokin, I. T. Sorokina: Population dynamics of the 7 F5 Level of Tb3+ ions doped in the KPb2 Cl5 , Int. Quantum Electron. Conf. (IQEC 2002) Moscow, (2002) paper QSuR24 283 209. S. I. Pekar: On the influence of the lattice deformations by electrons on the optical and electrical properties of crystals, Sov. Phys. Usp. 50, 197 (1953) 210. D. Spence, P. Kean, W. Sibbett: 60-fsec pulse generation from a self-modelocked Ti:sapphire laser, Opt. Lett. 16, 41 (1991) 284 211. M. Piche: Beam reshaping and self-mode-locking in nonlinear laser resonators, Opt. Commun. 86, 156–160 (1991) 284 212. Ch. Spielmann, P. F. Curley, T. Brabec, F. Krausz: Ultrabroadband femtosecond lasers, IEEE J. Quantum Electron. 30, 1100–1114 (1994) 284 213. Y. Chen, F. X. K¨ artner, U. Morgner, S. H. Cho, H. A. Haus, E. P. Ippen, J. G. Fujimoto: Dispersion-managed mode locking, J. Opt. Soc. Am. B 16, 1999–2004 (1999) 284 214. R. Mellish, N. P. Barry, S. C. W. Hyde, R. Jones, P. M. W. French, C. J. Van der Poel, A. Valster: Diode-pumped Cr:LiSAF all-solid-state femtosecond oscillator and regenerative amplifier, Opt. Lett. 20, 2312–2314 (1995) 284 215. V. P. Yanovsky, F. W. Wise, A. Cassanho, H. P. Jenssen: Femtosecond diodepumped Cr:LiSGAF lasers, IEEE J. Quantum Electron. 2, 465 (1996) 284 216. B. M. Antipenko, A. S. Glebov, T. I. Kiseleva, V. A. Pis’mennyi: 2.12 µm Ho:YAG laser, Pisma Zh. Tech. Fiz. 11, 682 (1985) [English transl. in: Sov. Tech. Phys. Lett. 11, 284 (1985)] 285 217. R. J. Beach: Theory and optimization of lens ducts, Appl. Opt. 35, 2005–2015 (1996) 287 218. R. J. Beach, M. A. Emanuel, B. L. Freitas, J. A. Skidmore, N. W. Carlson, J. Benett, R. W. Solarz: Applications of microlens-conditioned laser diode arrays, Proc. SPIE 2383, 283–297 (1995) 287 219. M. Tsunekane, N. Taguchi, H. Inaba: Improvement of thermal effects in a diode-end-pumped, composite Tm:YAG rod with undoped ends, Appl. Opt. 38 (1999) 288 220. S. A. Payne, R. J. Beach, C. Bibeau, C. A. Ebbers, M. A. Emanuel, E. C. Honea, C. D. Marshall, R. H. Page, K. I. Schaffers, J. A. Skidmore, S. B. Sutton, W. F. Krupke: Diode arrays, crystals, and thermal management for solid-state lasers, IEEE J. Sel. Top. Quantum Electron. 3, 71–81 (1997) 288, 320, 321, 322 221. A. Giesen, H. H¨ ugel, A. Voss, K. Wittig, U. Brauch, H. Opower: Scalable concept for diode-pumped high-power solid-state-lasers, Appl. Phys. B 58, 365 (1994) 288, 306 222. C. Stewen, K. Contag, M. Larionov, A. Giesen, H. H¨ ugel: A 1-kW CW thin disc laser, IEEE J. Sel. Topics Quantum Electron. 6, 650–657 (2000) 288 223. N. Berner, A. Diening, E. Heumann, G. Huber: Tm:YAG: a comparison between endpumped laser-rods and the “thin-disk”-setup, OSA Trends Opt. Photonics 26, 463–467 (Opt. Soc. Am., Washington, DC 1999) 288

336

Irina T. Sorokina

224. A. Dergachev, K. Wall, P. F. Moulton: A CW side-pumped Tm:YLF laser, OSA Trends Opt. Photonics 68, 343–350 (Opt. Soc. Am., Washington, DC 2002) 288, 316, 317 225. E. C. Honea, R. J. Beach, S. B. Sutton, J. A. Speth, S. C. Mitchell, J. A. Skidmore, M. A. Emanuel, S. A. Payne: 115-W Tm:YAG diode-pumped solid-state laser, IEEE J. Quantum Electron. 33, 1592–1600 (1997) 288 226. B. Struve, G. Huber: Properties and medical applications of near-IR solidstate lasers, J. Phys. IV Colloque. 1, 3–6 (1991) 289 227. C. Li, J. Song, D. Shen, Y. Cao, N. S. Kim, K. Ueda: Flash-lamp-pumped acousto-optic Q-switched Cr-Tm:YAG Laser, Opt. Rev. 7, 58–61 (2000) 293 228. F. J. Duarte: Tunable Lasers Handbook (Academic Press, San Diego 1995) 289, 291 229. B. Lyot: Un monochromateur a grand champ utilizant les interferences en lumiére polarisée, Compt. Rend. 197, 1593 (1933) 290 230. J. W. Evans: The birefringent filter, J. Opt. Soc. Am. 39, 229–242 (1949) 290 231. A. L. Bloom: Modes of a laser resonator containing tilted birefringent plates, J. Opt. Soc. Am. 64, 447–452 (1974) 290 232. D. J. Taylor, S. E. Harris, S. T. K. Nieh, T. W. H¨ ansch: Electronic tuning of a dye laser using the acousto-optic filter, Appl. Phys. Lett. 19, 269–271 (1971) 291 233. T. Yano, A. Watanabe: Acoustooptic TeO2 tunable filter using far-off-axis anisotropic Bragg diffraction, Appl. Opt. 15, 2250–2258 (1976) 291 234. C. Chappuis, J. P. Goedgebuer: Broad band electro-optic tuning of a CW dye laser, Opt. Commun. 47, 12–17 (August, 1983) 235. V. G. Dmitriev, G. G. Gurzadyan, D. N. Nikogosyan: Handbook on Nonlinear Optical Crystals, Springer Ser. Opt. Sci. 64 (Springer, Berlin, Heidelberg 1999) 292, 300 236. S. Lovold, P. F. Moulton, D. K. Killinger, N. Menyuk: Frequency tuning characteristics of a Q-switched Co:MgF2 laser, IEEE J. Quantum Electron. 21, 202–208 (1985) 292 237. M. Frenz, H. Pratisto, F. Konz, E. D. Jansen, A. J. Welch, H. P. Weber: Comparison of the effects of absorption coefficient and pulse duration of 2.12-µm and 2.79-µm radiation on laser ablation of tissue, IEEE J. Quantum Electron. 32, 2025–2036 (1996) 292 238. R. Brinkmann, C. Hansen: Beam-profile modulation of thulium laser radiation applied with multimode fibers and its effect on the threshold fluence to vaporize water laser surgery application, Appl. Opt. 39, 3361–3371 (2000) 292 239. N. M. Wannop, M. R. Dickinson, A. Charlton, T. A. King: Q-switching the erbium-YAG laser, J. Mod. Opt. 41, 2043–2053 (1994) 292 240. B. Pelz, M. K. Schott, M. H. Niemz: Electro-optic mode locking of an Erbium:YAG laser with a RF resonance transformer, Appl. Opt. 33, 364–367 (1994) 292 241. H. Voss, F. Massmann: Diode pumped, Q-switched erbium lasers with short pulse duration, OSA Trends Opt. Photonics Ser. 10, 217–221 (Opt. Soc. Am., Washington, DC 1997) 292 242. E. C. Honea, R. J. Beach, S. B. Sutton, J. A. Speth, S. C. Mitchell, J. A. Skidmore, M. A. Emanuel, S. A. Payne: 115-W Tm:YAG diode-pumped solid-state laser, IEEE J. Quantum Electron. 33, 1592–1600 (1997) 316

Crystalline Lasers

337

243. C. Li, J. Song, D. Shen, N. S. Kim, K. Ueda, Y. Huo, S. He, Y. Cao: Diodepumped high-efficiency Tm:YAG lasers, Opt. Express 4, 12–18 (1999), 293 244. A. Finch, J. H. Flint: Diode-pumped 6- mJ repetitively-Q-switched Tm,Ho,YLF laser, CLEO’95, Tech. Digest Series 15, Paper CWH2 232 (1995) 293 245. S. Schnell, V. G. Ostroumov, J. Brequet, W. A. R. L¨ uthy, H. P. Weber, I. A. Scherbakov: Acoustooptic Q-switching of erbium lasers, IEEE J. Quantum Electron. 26, 1111–1114 (1990) 293 246. P. Maak, L. Jakab, P. Richter, H. J. Eichler, B. Liu: Efficient acousto-optic Q switching of Er:YSGG lasers at 2.79-µm wavelength, Appl. Opt. 39, 3053– 3059 (2000) 293 247. I. N. Court, K. K. von Willisen: Frustrated total internal reflection and application of its principle to laser cavity design, J. Appl. Opt. 3, 719–726 (1964) 293 248. Kh. S. Bagdasarov, V. P. Danilov, V. I. Zhekov, T. M. Murina, A. A. Manenkov, M. I. Timoshechkin, A. M. Prokhorov: Pulse-periodic Y3 Al5 O12 :Er3+ laser with high activator concentration, Sov. J. Quantum Electron. 8, 83 (1978) 293 249. F. K¨ onz, M. Frenz, V. Romano, M. Forrer, H. P. Weber, A. V. Kharkovskiy, S. I. Khomenko: Active and passive Q-switching of a 2.79 µm Er:Cr:YSGG laser, Opt. Commun. 103, 398–404 (1993) 293 250. A. H¨ ogele, G. H¨ orbe, H. Lubatschowski, H. Welling, W. Ertmer: 2.70 µm CrEr:YSGG with high output energy and FTIR-Q-Switch, Opt. Commun. 125, 90–94 (1996) 293 251. H. J. Eichler, B. Liu, M. Kayser, S. I. Khomenko: Er:YAG-laser at 2.94 µm Q-switched by a FTIR-shutter with silicon output coupler and polarizer, Opt. Mater. 5, 259–265 (1996) 293 252. Y. T. Tzong, M. Birnbaum: Q-switched 2-µm lasers by use of a Cr2+:ZnSe saturable absorber, Appl. Opt. 40, 6633–6637 (2001) 293 253. E. Sorokin, I. T. Sorokina: Coupled-cavity passive Q-switching, paper CThO6, Conference on Lasers and Electro-Optics Europe, Nice, France 2000; in CLEO Europe’2000 Tech. Digest (2000) p. 359 293 254. A. M. Malyarevich, I. A. Denisov, N. N. Posnov, P. V. Prokoshin, K. V. Yumashev, A. A. Lipovskii: PbS(Se)-doped glasses as passive Q-switches for 1.5- µm Er:glass and 2.1-µm Ho:YAG lasers, OSA Trends Opt. Photonics Adv. SolidState Lasers 34, 225 (Opt. Soc. Am., Washington, DC 2000) 293 255. A. M. Malyarevich, P. V. Prokoshin, M. I. Demchuk, K. V. Yumashev, A. A. Lipovskii: Passively Q-switched Ho3+:Y3 Al5 O12 laser using a PbSedoped glass, Appl. Phys. Lett 78, 572–573 (2001) 293 256. K. L. Vodopyanov, A. V. Lukashev, C. C. Phillips, I. T. Ferguson: Passive mode locking and Q-switching of an erbium 3 µm laser using thin InAs epilayers grown by molecular beam epitaxy, Appl. Phys. Lett. 59, 1658–1660 (1991) 293, 294 257. K. L. Vodopyanov, A. V. Lukashev, C. C. Phillips: Nano- and picosecond 3 mm Er:YSGG lasers using InAs as passive Q-switchers and mode-lockers, Opt. Commun. 95, 87–91 (1993) 293 258. K. L. Vodopyanov, L. A. Kulevskii, P. P. Pashinin, A. M. Prokhorov: Water and ethanol as bleachable absorbers of radiation in an yttrium-erbium-aluminum garnet laser (λ = 2.94 µm), Sov. Phys. JETP 55, 1049–1051 (1982) 293

338

Irina T. Sorokina

259. K. L. Vodopyanov, R. Shori, O. M. Stafsudd: Generation of Q-switched Er:YAG laser pulses using evanescent wave absorption in ethanol, Appl. Phys. Lett. 72, 2211–2213 (1998) 293 260. Yu. D. Zavartsev, A. I. Zagumennyi, L. A. Kulevskii, A. V. Lukashev, P. A. Studenikin, I. A. Shcherbakov, A. F. Umyskov: Q-switching in a Cr3+:Yb3+:Ho3+:YSGG crystal laser based on the 5 I6 –5 I7 (λ = 2.92 µm) transition, J. Quantum Electron. 29, 295–297 (1999) [transl. from: Kvant. Elektron. 27, 13–15 (1999)] 293, 315 261. J. Breguet, W. L¨ uthy, H. P. Weber: Q-switching of YAG:Er laser with a soap film, Opt. Commun. 82, 488–490 (1991) 293 262. B. C. Johnson, P. F. Moulton, A. Mooradian: Mode-locked operation of Co:MgF2 and Ni:MgF2 lasers, Opt. Lett. 10, 116–118 (1984) 293, 297 263. F. Heine, E. Heumann, G. Huber, K. L. Schlepper: Mode locking of roomtemperature CW thulium and holmium lasers, Appl. Phys. Lett. 60, 1161– 1162 (1992) 293, 318 264. E. Sorokin, I. T. Sorokina, A. Unterhuber, E. Wintner, A. I. Zagumenny, I. A. Shcherbakov, V. Carozza, A. Toncelli, M. Tonelli: A novel CW tunable and mode-locked 2 µm Cr,Tm,Ho:YSGG:GSAG laser, OSA Trends Opt. Photonics 19, 197–200, W. R. Bosenberg, M. M. Fejer (Eds.) (Opt. Soc. Am., Washington, DC 1998) 293, 315, 318 265. J. F. Pinto, L. Esterowitz, G. H. Rosenblatt: Continuous-wave mode-locked 2-µm Tm:YAG laser, Opt. Lett. 17, 731–732 (1992) 293, 318 266. T. J. Carrig, G. J. Wagner, A. Sennaroglu, J. Y. Jeong, C. R. Pollock: Modelocked Cr2+:ZnSe laser, Opt. Lett. 25, 168–170 (2000) 293 267. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. H. Page, K. I. Schaffers: Active and passive mode-locking of the Cr2+:ZnSe laser, in Tech. Dig. Adv. Solid-State Lasers (Opt. Soc. Am., 2001), pp. 87–89 293, 294 268. D. E. Spence, P. N. Kean, W. Sibbett: 60-fs pulse generation from a self-modelocke Ti:sapphire laser, Opt. Lett. 16, 42–44 (1991) 293 269. U. Keller, K. Weingarten, F. X. K¨ artner, D. Kopf, B. Braun, I. D. Jung, R. Fluck, C. H¨ onninger, N. Matuschek, J. Aus der Au: Semiconductor saturable absorber mirrors (SESAM’s) for femtosecond to nanosecond pulse generation in solid-state lasers, IEEE J. Sel. Topics Quantum Electron. 2, 435–453 (1996) 293 270. G. Steinmeyer, U. Morgner, M. Ostermeyer, F. Mitschke, H. Welling: Subpicosecond pulses near 1.9 µm from a synchronously pumped color-center laser, Opt. Lett. 18, 1544–1546 (1993) 294 271. K. M¨ ollmann, M. Schrempel, B. K. Yu, W. Gellermann: Subpicosecond and + continuous-wave laser operation of (F+ 2 )H and (F2 )AH color-center lasers in the 2-µm range, Opt. Lett. 19, 960–962 (1994) 294 272. T. Brabec. Ch. Spielmann, P. F. Curley, F. Krausz: Kerr lens mode locking, Opt. Lett. 17, 1292–1294 (1992). 273. C. L. Cesar, M. N. Islam, C. E. Soccolich, R. D. Feldman, R. F. Austin, K. German: Femtosecond KCl:Li and RbCl:Li color-center lasers near 2.8 µm with a HgCdTe multiple-quantum-well saturable absorber, Opt. Lett. 15, 1147– 1149 (1990) 294 294 274. P. F. Moulton: Pulsed-pumped operation of divalent transition-metal lasers, IEEE J. Quantum Electron. 18, 1185 (1982) 295, 297

Crystalline Lasers

339

275. P. F. Moulton: Tunable paramagnetic-ion lasers, in Laser Handbook, Vol. 5, M. Bass, M. L. Stich (Eds.) (North-Holland, Amsterdam, 1985) p. 203 295, 296 276. J. M. Breteau, D. Mekhenin, F. Auzel: Study of the Ni2+:MgF2 tunable laser, Rev. Phys. Appl. 22, 1419 (1987) (in French) 295 277. P. F. Moulton, A. Mooradian: Broadly tunable CW operation of Ni:MgF2 and Co:MgF2 lasers, Appl. Phys. Lett. 35, 838 (1979) 295, 296, 297 278. P. F. Moulton, A. Mooradian: Tunable transition-metal-doped solid-state lasers, in Laser Spectroscopy IV, H. Walther, K. W. Rothe (Eds.) (Springer, Heidelberg, Berlin 1979) 296 279. R. Moncorge, T. Benyattou: Excited-state absorption of Ni2+ in MgF2 and MgO, Phys. Rev. B 37, 9186 (1988) 296 280. R. G. Pappalardo, D. L. Wood, R. C. Linares: Optical absorption study of Nidoped oxide systems I, J. Chem. Phys. 35, 1460 (1961) 296 281. M. V. Iverson, W. A. Sibley: Temperature dependence of Ni2+ luminescence in KZnF3 , MgF2 and MgO, J. Lumin. 20, 311 (1979) 296 282. J. Koetke, K. Petermann, G. Huber: Infrared excited state absorption of Ni2+doped crystals, J. Lumin. 60&61, 197–200 (1994) 296 283. J. Koetke, K. Petermann, G. Huber: Spectroscopy of Ni2+-doped garnets and perovskites for solid state lasers, J. Lumin. 48&49, 564–568 (1991) 296 284. D. Welford, P. F. Moulton: Room-temperature operation of a Co:MgF2 laser, Opt. Lett. 13, 975 (1988) 296, 299 285. H. Manaa, Y. Guyot, R. Moncorge: Spectroscopic and laser properties of Co2+doped single crystals, Phys. Rev. B 48, 3633 (1993) 296, 297, 299 286. D. M. Rines, P. F. Moulton, D. Welford, G. A. Rines: High-energy operation of a Co:MgF2 laser, Opt. Lett. 19, 628 (1994) 296, 297, 298, 299 287. R. G. Pappalardo, D. L. Wood, R. C. Linares: Optical absorption study of Codoped oxide systems II, J. Chem. Phys. 35, 2041 (1961) 296 288. S. Lovold, P. F. Moulton, D. K. Killinger, N. Menyuk: Frequency tuning characteristics of a Q-switched Co:MgF2 laser 21, 202–208 (1985) 297 289. P. F. Moulton: An investigation of the Co:MgF2 laser system, IEEE J. Quantum Electron. 21, 1582–1595 (1985) 297, 298, 299 290. A. M. Fox, A. C. Maciel, J. F. Ryan: Efficient CW performance of a Co:MgF2 laser operating at 1.5–2.0 µm Opt. Commun. 59, 142–144 (1986) 297, 298 291. W. K¨ unzel, W. Knierim, U. D¨ urr: CW infrared laser action of optically pumped Co2+:KZnF3 , Opt. Commun. 36, 383 (1981) 297 292. K. R. German, U. D¨ urr, V. K¨ unzel: Tunable single-frequency continuous-wave laser action in Co2+:KZnF3 , Opt. Lett. 11, 12–14 (1986) 297, 311, 313 293. U. D¨ urr: Vibronic solid state lasers: the transition metal ion laser, Laser und Optoelektronik 15, 31–39 (1983) 297 294. M. D. Sturge: Temperature dependence of multiphonon nonradiative decay at an isolated impurity center, Phys. Rev. B. 8, 6–14 (1973) 297 295. A. Di Lieto: Development of a CW Co:MgF2 laser, Opt. Laser Eng. 39, 305– 308 (2003) 298, 299 296. R. H. Page, L. D. DeLoach, K. I. Schaffers, F. D. Patel, R. L. Beach, S. A. Payne, W. F. Krupke: Recent developments in Cr2+-doped II-VI compound lasers, in OSA Trends Opt. Photonics 1, 130–136, S. A. Payne, C. R. Pollock (Eds.) (Opt. Soc. Am., Washington, DC 1996) 299

340

Irina T. Sorokina

297. W. F. Krupke, R. H. Page, L. D. DeLoach, S. A. Payne: Transition-metal doped sulphide, selenide, and telluride laser crystal and lasers, US Patent 5,541,948 from July 30 (1996) 299 298. U. H¨ ommerich, X. Wu, V. R. Davis, S. B. Trivedi, K. Grasze, R. J. Chen, S. W. Kutcher: Demonstration of room-temperature laser action at 2.5 µm from Cr2+:Cd0.85 Mn015 Te, Opt. Lett. 22, 1180–1182 (1997) 299 299. S. B. Trivedi, K. Grasze, R. J. Chen, S. W. Kutcher, U. H¨ ommerich, X. Wu: Jae Tae Seo, Chromium doped cadmium manganese telluride: a novel solidgain media for mid-infrared lasers, Phys. Semicond. Dev. 2, 762–767 (1998) 299 300. K. L. Schepler, S. Kueck, L. Shiozawa: Cr2+ emission spectroscopy in CdSe, J. Lumin. 72–74, 116–117 (1997) 299, 305, 309 301. J. McKay, K. L. Schepler, G. C. Catella: Efficient grating-tuned mid-infrared Cr2+:CdSe laser, Opt. Lett. 24, 1575–1577 (1999) 299, 305, 309 302. A. Sennaroglu, A. Ozgun Konza, C. R. Pollock: Continuous-wave power performance of a 2.47-µm Cr2+: ZnSe laser: experiment and modelling, IEEE J. Quantum Electron. 36, 1199–1205 (2000) 300, 305 303. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, V. G. Shcherbitsky, V. E. Kisel, N. V. Kuleshov, V. I. Levchenko: Ultrabroadband tunable continuous-wave mid-infrared Cr:ZnSe lasers, to be published 300, 301, 303, 304, 307 304. A. V. Podlipensky, V. G. Shcherbitsky, N. V. Kuleshov, V. I. Levchenko, V. N. Yakimovich, M. Mond, E. Heumann, G. Huber, H. Kretschmann, S. K¨ uck: Efficient laser operation and continuous-wave diode pumping of Cr2+: ZnSe single crystals, Appl. Phys. B. 72, 253–255 (2001) 300, 301, 304, 305, 306, 308 305. G. Tocci, M. Vannini, E. Giorgetti, E. Sorokin, I. Sorokina, to be published 300 306. T. D. Krauss, F. W. Wise: Femtosecond measurement of nonlinear absorption and refraction in CdS, ZnSe and ZnS, Appl. Phys. Lett. 65, 1739–1741 (1994) 300 307. M. E. Innocenzi, R. T. Swimm, M. Bass, R. H. French, M. R. Kokta: Optical absorption in undoped yttrium aluminum garnet, J. Appl. Phys. 68, 3, 1200–4, (1990) 300 308. G.-M. Schucan, R. G. Ispasoiu, A. M. Fox, J. F. Ryan: Ultrafast two-photon nonlinearities in CdSe near 1.5 µm studied by interferometric autocorrelation, IEEE J. Quantum Electronics 34, 1374 (1998) 300 309. Landolt B¨ ornstein, Numerical data and functional relationships in Science and Technology, in O. Madelung (Ed.): Non-Tetrahedrally Bonded Elements and Binary Compounds I, New Series, Group III, Vol. 41c, Semiconductors (Springer, Berlin, Heidelberg 1998) 300, 301 310. E. Sorokin, I. T. Sorokina, A. DiLieto, M. Tonelli, R. H. Page: Tunable diodepumped continuous-wave single crystal and ceramic Cr2+:ZnSe lasers, in CLEO/Pacific Rim 2001, Tokyo Tech. Digest (2001) paper WJPD1–1 301 311. E. Sorokin, I. T. Sorokina, A. DiLieto, M. Tonelli, P. Minguzzi: Mode-locked ceramic Cr2+:ZnSe laser, Advanced Solid-State Photonics’2003, St. Antonio, USA (2003) paper TUB-17 301 312. A. Zunger: Solid State Physics 89 (Academic Press, New York 1986) p. 275 301

Crystalline Lasers

341

313. J. T. Vallin, G. A. Slack, S. Roberts, A. E. Hughes: Near and far infrared absorption in Cr doped ZnSe, Solid State Commun. 7, 1211 (1969) 301 314. S. Taniguchi, T. Hino, S. Itoh, K. Nakano, N. Nakayama, A. Ishibashi, M. Ikeda: 100 h II–VI blue-green laser diode, Electron. Lett. 32, 552 (1996) 301 315. H. Ulchiike: Review of flat-panel displays, liquid crystal displays, plasma displays, etc., in Proc. Fourth Int. Workshop of Electroluminescence, S. Shionoya, H. Kobayashi (Eds.) (Springer, Berlin, Heidelberg 1989) p. 238–245 301 316. K. Grasza, S. B. Trivedi, Yu-Zengchen, S. W. Kutcher, W. Palosz, G. A. Brost: Low supersaturation nucleation and contactless growth of photorefractive ZnTe crystals, J. Crystal Growth 174, 719–725 (1997) 301 317. A. Partovi, J. Millerd, E. M. Garmire, M. Ziari, W. H. Steier, S. B. Trivedi, M. B. Klein: Photorefractivity at 1.5 µm in CdTe:V, Appl. Phys. Lett. 57, 846–848 (1990) 301 318. A. Fazzio, M. J. Caldas, A. Zunger: Many-electron multiplet effects in the spectra of 3d-impurities in heteropolar semiconductors, Phys. Rev. B 30, 3430 (1984) 301 319. C. S. Kelley, F. Williams: Optical absorption spectra of chromium-doped zinc sulphide crystals, Phys. Rev. B 2, 3 (1970) 301 320. R. Pappalardo, R. E. Dietz: Absorption spectra of transition ions in CdS crystals, Phys. Rev. 123, 1188 (1961) 301 321. H. Nelkowski, G. Grebe: IR-luminescence of ZnS:Cr, J. Lumin. 1/2, 88–93 (1970) 301 322. G. Grebe, H. J. Schulz: Interpretation of excitation spectra of ZnS:Cr2+ by fitting the eigenvalues of the Tanabe–Sugano matrices, Phys. Stat. Sol. (b) 54, 69–72 (1972) 301 323. G. Grebe, H. J. Schulz: Luminescence of Cr2+ centers and related optical interactions involving crystal field levels of chromium ions in ZnS, Z. Naturf. 29a, 1803 (1974) 301, 303 324. G. Grebe, G. Roussos, H. J. Schulz: Infrared luminescence of ZnSe:Cr crystals, J. Lumin. 12/13, 701 (1976) 301, 303 325. B. D. Bhattacharya: Jahn–Teller effect in ESR study of Cr2+ in II–VI semiconductors, Phys. Stat. Sol. (b) 71, K181–K185 (1975) 303 326. A. L. Natadze, A. I. Ryskin: Jahn–Teller coupling of Cr2+-ion with degenerate modes in ZnS, ZnSe, and ZnTe crystals: microscopic treatment, Phys. Stat. Sol. (b) 97, 175–185 (1980) 303 327. R. Renz, H. J. Schulz: Temperature dependence of the lifetime of excited states for 3d transition element centers in II–VI crystals, J. Lumin. 24/25, 221–224 (1981) 303 328. G. Goetz, H. J. Schulz: Decay of internal luminescence transitions of 3dimpurities in II–VI compounds – recent experiments and refined interpretations, J. Lumin. 40/41, 415–416 (1988) 303 329. P. Dahan, V. Fleurov: Possibility of a metastable state for transition-metal impurity in semiconductor, Phys. Rev. B 53, 12845 (1996) 303 330. A. Burger, K. Chattopadhyay, J.-O. Ndap, X. Ma, S. H. Morgan, C. I. Rablau, C. H. Su, S. Feth, R. H. Page, K. I. Schaffers, S. A. Payne: Preparation conditions of chromium doped ZnSe and their infrared luminescence properties, J. Crystal Growth 225, 249–256 (2001) 303, 304, 306

342

Irina T. Sorokina

331. V. E. Kisel, V. G. Shcherbitsky, N. V. Kuleshov, V. I. Konstantinov, L. I. Postnova, V. I. Levchenko, E. Sorokin, I. T. Sorokina: Luminescence lifetime measurements in diffusion doped Cr:ZnSe, ECLEO’2003, Munich (June 2003) 303, 304 332. J. McKay, K. L. Schepler: Solid-state and diode-laser technology review, May 21–24, Albuquerque, NM, SS-6 (2001) 303 333. S. B. Mirov, V. V. Fedorov, K. Graham, I. S. Moskalev, E. Sorokin, I. T. Sorokina, V. Gapontsev, D. Gapontsev, V. V. Badikov, V. V. Panyutin: Diode, fiber, and potentially electrically pumped Cr2+:ZnS mid-IR external cavity and microchip lasers, Tech. Digest 5th Int. Conf. Mid-Infrared Optoelectronics Materials and Devices, Sept. 2002, Annapolis, MD (2002) 304, 305 334. I. T. Sorokina, E. Sorokin, S. Mirov, V. Fedorov, V. Badikov, V. Panyutin, A. Di Lieto, M. Tonelli: Continuous-wave tunable Cr2+:ZnS laser, Appl. Phys B 74, 607–611 (2002) 300, 304, 305, 308 335. P. Albers, E. Stark, G. Huber: Continuous-wave laser operation and quantum efficiency of titanium-doped sapphire, J. Opt. Soc. Am. B 3, 134 (1986) 304, 305 336. R. H. Page, J. A. Skidmore, K. I. Schaffers, R. J. Beach, S. A. Payne, W. F. Krupke: Demonstrations of diode-pumped and grating tuned ZnSe:Cr2+ lasers, Opt. Soc. Am. Trends Opt. Photonics, C. R. Pollock, W. R. Bosenberg (Eds.) (Opt. Soc. Am., Washington, DC 1997), pp. 208–210 304, 307 337. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. Page, K. Schaffers: Efficient broadly tunable Cr2+:ZnSe laser, Conf. Lasers and Electro-Optics Europe, Nice, France, in CLEO Europe’2000 Tech. Digest (2000) p. 261, paper CWH4 304, 308 338. E. Sorokin, I. T. Sorokina, R. H. Page: Room-temperature CW diode-pumped Cr2+:ZnSe laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 46, 101– 105, S. Payne, C. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2001) 304, 308 339. M. Mond, E. Heumann, G. Huber, H. Kretschmann, S. K¨ uck, A. V. Podlipensky, V. G. Shcherbitsky, N. V. Kuleshov, V. I. Levchenko, V. N. Yakimovich: Continuous-wave diode pumped Cr2+:ZnSe and high power laser operation, OSA Trends Opt. Photonics Adv. Solid-State Lasers 46, 162–165, S. Payne, C. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2001) 304, 308 340. E. Sorokin, I. T. Sorokina: Tunable diode-pumped continuous-wave Cr2+:ZnSe laser, Appl. Phys. Lett. 80, 3289–3291 (2002) 304, 305, 308 341. M. Mond, D. Albrecht, E. Heumann, G. Huber, S. K¨ uck, V. I. Levchenko, V. N. Yakimovich, V. G. Shcherbitsky, V. E. Kisel, N. V. Kuleshov, M. Rattunde, J. Schmitz, R. Kiefer, J. Wagner: 1.9-µm and 2.0-µm laser diode pumping of Cr2+:ZnSe and Cr2+:CdMnTe, Opt. Lett. 27, 1034–6 (2002) 304, 305, 308, 309 342. T. J. Carrig, G. J. Wagner, A. Sennaroglu, J. Y. Jeong, C. R. Pollock: Modelocked Cr2+:ZnSe laser, Opt. Lett. 25, 168–170 (2000) 304, 305, 307 343. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. H. Page, K. I. Schaffers: Active and passive mode-locking of the Cr2+:ZnSe laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 46, 157–161, S. Payne, C. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2001) 304, 305, 307

Crystalline Lasers

343

344. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. H. Page, K. I. Schaffers: Tunable diode-pumped continuous-wave operation and passive mode-locking of Cr2+:ZnSe laser, Tech. Digest CLEO/Europe Focus Meeting 2001, Munich (2001) 304, 307 345. G. J. Wagner, T. J. Carrig: Power scaling of Cr2+:ZnSe lasers, OSA Trends Opt. Photonics Adv. Solid-State Lasers 50, 506–510, C. Marshall (Ed.) (Opt. Soc. Am., Washington, DC 2001) 304, 305, 306 346. J. B. McKay, W. B. Roh, K. L. Schepler: 4.2 W Cr2+:ZnSe face cooled disk laser, paper CMY3, in Opt. Soc. Am. Trends in Optics and Photonics, Vol. 73, Conference on Lasers and Electro-Optics, OSA Tech. Digest, Postconference Edition (Opt. Soc. Am., Washington, DC 2002), pp. 119–120 288, 304, 305, 306 347. W. Alford, G. Wagner, J. Keene, T. Carrig: High power and Q-switched Cr:ZnSe laser, paper MA-8, Advanced Solid-State Photonics 2003, St. Antonio, USA (2003) 304, 305, 306 348. M. Mond, D. Albrecht, H. M. Kretschmann, E. Heumann, G. Huber, S. K¨ uck, V. I. Levchenko, V. N. Yakimovich, V. G. Shcherbitsky, V. E. Kisel, N. V. Kuleshov: Er doped fiber amplifier pumped Cr2+:ZnSe laser, Phys. Stat. Sol. (a) 188 (4), R3–R5 (2001) 305 349. S. B. Mirov, V. V. Fedorov, K. Graham, I. Moskalev, V. Badikov, V. Panyutin: Er-fiber laser pumped continuous-wave microchip Cr2+ :ZnS and Cr2+ :ZnSe lasers, Opt. Lett. 27, 909–911 (2002) 305, 308 350. D. H. Sutter, G. Steinmeyer, L. Gallmann, N. Matuschek, F. Morier-Genoud, U. Keller, V. Scheuer, G. Angelow, T. Tschudi: Semiconductor saturable absorber mirror assisted Kerr-Lens mode-locked Ti:sapphire laser, producing cycles in the two-cycle regime, Opt. Lett. 24, 631–633 (1999) 305 351. U. Morgner, F. X. K¨ artner, S. H. Cho, Y. Chen, H. A. Haus, J. G. Fujimoto, E. P. Ippen, V. Scheuer, G. Angelow, T. Tschudi: Sub-two-cycle pulses from a Kerr-Lens mode-locked Ti:sapphire laser, Opt. Lett. 24, 920–922 (1999) 305 352. J. T. Seo, U. H¨ ommerich, H. Zong, S. B. Trivedi, S. W. Kutcher, C. C. Wang, R. J. Chen: Mid-infrared lasing from a novel optical material: Chromiumdoped Cd0.55 Mn0.45 Te, Phys. Stat. Sol. (a) 175, R 3 (1999) 305, 309 353. J. McKay, D. Krause, K. L. Schepler: Optimization of Cr2+:CdSe for efficient laser operation, OSA Trends Opt. Photonics 34, 218–224, H. Injeyan, U. Keller, Ch. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2000) 305 354. P. A. Champert, S. V. Popov, J. R. Taylor: Efficient lasing of Cr2+:ZnSe at 2.2. µm pumped by all-fiber-format seeded Raman source, Electron. Lett. 38, 448 (2002) 305 355. T. J. Carrig, G. J. Wagner, I. T. McKinnie, W. J. Alford: Recent progress in the development of practical Cr:ZnSe lasers, paper presented at the Solid State & Diode Laser Technology Review, Albuquerque, NM (2002) 305, 308 356. T. J. Carrig: Transition-metal-doped chalcogenide lasers, J. Electron. Mater. 31, 759–769 (July 2002) 305 357. J. F. Wang, A. Omino, M. Isshiki: Bridgman growth of twin-free ZnSe single crystals, Mater. Sci. Eng. B 83, 185–191 (2001) 305 358. C. S. Fang, Q. T. Gu, J. Q. Wei, Q. W. Pan, W. Shi, J. Y. Wang: Growth of ZnSe single crystals, J. Crystal Growth 209, 542–546 (2000) 305 359. V. I. Levchenko, V. N. Yakimovich, L. I. Postnova, V. I. Konstantinov, V. P. Mikhailov, N. V. Kuleshov: Preparation and properties of bulk ZnSe:Cr single crystals, J. Crystal Growth 198–199, 980–983 (1999) 305

344

Irina T. Sorokina

360. A. Burger, J.-O. Ndap, K. Chattopadhyay, O. O. Adetunji, D. E. Zelmon, A. Burger: Thermal diffusion of Cr2+ in bulk ZnSe, J. Crystal Growth 240, 176–184 (2002) 306 361. A. V. Podlipensky, V. G. Shcherbitsky, N. V. Kuleshov, V. P. Mikhailov, V. I. Levchenko, V. N. Yakimovich, L. I. Postnova, V. I. Konstantinov: Pulsed laser operation of diffusion-doped Cr2+:ZnSe, Opt. Commun. 167, 129 (1999) 306 362. G. J. Wagner, T. J. Carrig, R. H. Jarman, R. H. Page, K. I. Schaffers, J.-O. Ndap, X. Ma, A. Burger: High-efficiency, broadly tunable continuouswave Cr2+:ZnSe laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 26, 427–434, S. Payne, C. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2001) 306 363. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. Page, K. Schaffers: 0.5 W efficient broadly tunable continous-wave Cr2+:ZnSe laser, OSA Trends Opt. Photonics 34, 188–193, H. Injeyan, U. Keller, Ch. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2000) 306 364. D. Findlay, R. A. Clay: Measurement of internal losses in 4-level lasers, Phys. Lett. 20, 277 (1966) 307 365. V. L. Kalashnikov, E. Sorokin, I. T. Sorokina: Multipulse operation and limits of the Kerr-lens mode locking stability, IEEE J. Quantum Electron. 39, 323– 336 (2003) 307 366. J. A. Skidmore, B. L. Freitas, C. E. Reinhardt, E. J. Utterback, R. H. Page, M. A. Emanuel: High-power operation of InGaAsP/InP laser diode array at 1.73 µm, IEEE Photonics Technol. Lett. 9, 1334 (1997) 308 367. J. T. Seo, U. H¨ ommerich, S. B. Trivedi, R. J. Chen, S. Kutcher: Slope efficiency and tunability of a Cr2+-doped Cd0.85 Mn0.15 Te mid-infrared laser, Opt. Commun. 153, 267–270 (1998) 309 368. J. T. Seo, U. H¨ ommerich, S. B. Trivedi, R. J. Chen, S. Kutcher, C. C. Wang, H. Zong, P. R. Boyd, W. Tardiff: Spectroscopy and tunable mid-infrared lasing of a novel gain-medium: Cr2+-doped Cd0.85 Mn0.15 Te, Korean Phys. Society 34, 221–226 (1999) 309 369. U. H¨ ommerich, J. T. Seo, M. Turner, A. Bluiett, S. B. Trivedi, H. Zong, S. Kutcher, C. C. Wang, R. J. Chen: Mid-infrared laser development based on transition metal doped cadmium manganese telluride, J. Lumin. 87–89, 1143–1145 (2000) 305, 309 370. S. B. Trivedi, S. W. Kutcher, C. C. Wang, G. V. Jagannathan, U. H¨ ommerich, A. G. Bluiett, M. Turner, J. T. Seo, K. L. Schepler, B. Schumm, P. R. Boyd, G. Green: Transition metal doped cadmium manganese telluride: A new material for tunable mid-infrared lasing, J. Electron. Mater. 30, 728 (2001) 309 371. A. G. Bluiett, U. H¨ ommerich, J. T. Seo, R. Shah, S. B. Trivedi, S. W. Kutcher, C. C. Wang, P. R. Boyd: Observation of lasing from Cr2+:CdTe and compositional effects in Cr2+ doped II–VI semiconductors, J. Electron. Mater. 31, 806–810 (July 2002) 309 372. J. McKay, K. L. Schepler, G. Catella: Kilohertz, 2.6 µm Cr2+:CdSe laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 26, 420, M. Fejer, H. Injeyan, U. Keller (Eds.) (Opt. Soc. Am., Washington, DC 1999) 309 373. J. McKay, W. B. Roh, K. L. Schepler: Extended Mid-IR tuning of a Cr2+:CdSe laser, Opt. Soc. Am. Trends Opt. Photonics Adv. Solid-State Lasers 68, 371– 373 (Opt. Soc. Am., Washington, DC 2002) 309

Crystalline Lasers

345

374. P. B. Klein, J. E. Furneaux, R. L. Henry: Laser oscillation at 3.53 µm from Fe2+ in n-InP:FeAppl. Phys. Lett. 42, 638 (1983) 310 375. J. J. Adams, C. Bibeau, R. H. Page, S. A. Payne: Tunable laser action at 4 µm from Fe:ZnSe, OSA Trends Opt. Photonics Adv. Solid-State Lasers 26, 435– 440, M. Fejer, H. Injeyan, U. Keller (Eds.) (Opt. Soc. Am., Washington, DC 1999) 310 376. J. J. Adams, C. Bibeau, R. H. Page, D. M. Krol, L. H. Furu, S. A. Payne: 4.0– 4.5-µm lasing of Fe:ZnSe below 180 K, a new mid-infrared laser material, Opt. Lett. 24, 1720 (1999) 310 377. W. S. Peloach, G. J. Wagner, T. J. Carrig, W. J. Scharpf: Mid-wave ZGP OPOs pumped by a Cr:ZnSe laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 46, 670–674, S. Payne, C. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 2001) 310 378. I. T. McKinnie, G. J. Wagner, S. Christensen, T. Carrig, C. B. Rawle: Dual band mid-wave/long-wave ZGP OPO pumped tuned by a Cr:ZnSe laser, in Opt. Soc. Am. Trends in Optics and Photonics 73, 170–171, Conf. Lasers Electro-Optics, OSA Technical Digest, Postconf. Edition (Opt. Soc. Am., Washington, DC 2002) 310 379. W. Beall Fowler: Physics of Color Centers, W. Beall Fowler (Ed.) (Academic Press, New York 1968) Chap. 2 311 380. L. F. Mollenauer: Color center lasers, in Quantum Electronics, part B, vol. 15, C. L. Tang (Ed.) (Academic Press, New York 1979) Chap. 6 311 381. L. F. Mollenauer: in Laser Handbook Vol. 4, M. Bass (Ed.) (North-Holland, Amsterdam 1985) 311 382. L. F. Mollenauer: Colorcenter lasers, in Tunable Lasers, 2nd. edition, L. F. Mollenauer, J. C. White, C. R. Pollock (Eds.) Springer-Verlag, Berlin (1992), p. 225–275 312 383. C. R. Pollock: Optical properties of laser-active color centers, J. Lumin. 35, 65–78 (1986) 311 384. C. R. Pollock: Color center lasers, in Encyclopedia of Lasers and Optical Technology, R. A. Meyers (Ed.) (Academic, San Diego 1991) 311 385. T. T. Basiev, S. B. Mirov, V. V. Osiko: Room-temperature color-center lasers, IEEE J. Quantum Electron. 24, 1052 (1988) 311 386. T. T. Basiev, S. B. Mirov: Room-Temperature Tunable Color-Center Lasers, Laser Sci. Technol. Ser. 16, p. 1 (Gordon, Breach, New York 1994) 311 387. S. B. Mirov, T. Basiev: Progress in color center lasers, IEEE J. Sel. Top. Quantum Electron. 1, 22–30 (1995) 311 388. B. Fritz, E. Menke: Laser effect in KCl with Fa(Li) centers, Solid-State Commun. 3, 61 (1965) 311, 312 389. L. F. Mollenauer, D. H. Olson: A broadly tunable CW laser using color centers, Appl. Phys. Lett. 24, 386 (1974) 311, 312 390. L. F. Mollenauer, D. H. Olson: Broadly tunable lasers using color centers, J. Appl. Phys. 46, 3109–3118 (1975) 311, 312 391. K. R. German: Optimization of FA (II) and FB (II) color-center lasers, J. Opt. Soc. Am. B 3, 149–157 (1986) 313 392. R. Beigang, G. Liftin, H. Welling: Frequency behavior and linewidth of CW single mode color-center lasers, Opt. Commun. 22, 269–271 (1977) 313 393. C. H. Breant, T. Baer, D. Nesbitt, J. L. Hall: State-dependent hyperfine coupling of HF studied with a frequency controlled color-center laser spectrome-

346

394.

395.

396. 397.

398. 399. 400.

401. 402. 403. 404.

405. 406. 407.

408.

409.

410.

411.

Irina T. Sorokina ter, in Laser Spectroscopy VI, H. P. Weber, W. L¨ uthy (Eds.) (Springer, Berlin, Heidelberg 1983) 313 H. Welling, G. Liftin, R. Beigang: Tunable infrared lasers using color centers, in Laser Spectroscopy III, J. L. Hall, J. L. Carlsten (Eds.) (Springer, Berlin, Heidelberg 1977) p. 370–375 313 C. R. Pollock, J. A. Jennings, F. R. Petersen, J. S. Wells, R. E. Drullinger, E. C. Beaty, K. M. Evenson: Direct frequency measurements of transitions at 520 THz (576 nm) in iodine and 260 THz (1.15 µm) in neon, Opt. Lett. 8, 133 (1983) 313 R. Beigang, J. J. Wynne: Atomic Rydberg-state spectroscopy of Ca I using pulsed color-center lasers, Opt. Lett. 6, 295–7 (1981) 313 A. S. Subdo, M. M. T. Loy, P. A. Roland, R. Beigang: High peak power FA (II) color-center laser for time-resolved infrared spectroscopy, Opt. Commun. 37, 417–20 (1981) 311, 313 R. Beigang: Color center laser pumped by a flashlamp-pumped dye laser, Opt. Commun. 34, 249–251 (1980) 311, 313 G. Liftin, R. Beigang, H. Welling: Tunable CW laser operation in FB (II) type color-center crystals, Appl. Phys. Lett. 31, 381 (1977) 311, 313 K. P. Koch, G. Liftin, H. Welling: Continuous-wave laser oscillation with extended tuning range in FA (II)-FB (II) color-center lasers, Opt. Lett. 4, 387 (1979) 311, 313 I. Schneider, C. L. Marquardt: Tunable, CW laser action using (F+ 2 ) centers in Li-doped KCl, Opt. Lett. 5, 214 (1980) 311, 312, 313 I. Schneider, C. L. Marquardt: Broadly tunable oscillator-amplifier system using lithium (F+ 2 )A centers in KCl, Opt. Lett. 10, 13–15 (1985) 313 I. Schneider, S. C. Moss: Color-center laser continuously tunable from 1.67 to 2.46 µm, Opt. Lett. 8, 7–8 (1983) 311, 313 K. M¨ ollmann, M. Schrempel: Bing Kun Yu, W. Gellermann: Subpicosecond + and continuous-wave laser operation of (F+ 2 )H and (F2 )AH color-center lasers in the 2-µm range, Opt. Lett. 19, 960–62 (1994) 311, 313 I. Schneider: Continuous tuning of a color-center laser between 2 and 4 µm, Opt. Lett. 7, 271 (1982) 311, 313 R. W. Tkach, T. R. Gosnell, A. J. Sievers: Solid-state vibrational laser KBr:CN−, Opt. Lett. 10, 122 (1984). 311, 314 T. R. Gosnell, A. J. Sievers, C. R. Pollock: Continuous-wave operation of the KBr:CN− solid-state vibration laser in the 5-µm region, Opt. Lett. 10, 125 (1985) 311, 314 T. R. Gosnell, R. W. Tkach, A. J. Sievers: High temperature vibrational fluorescence of CN− ions in alkali halides, Solid-State Commun. 53, 419–421 (1985) 314 W. Gellermann, F. L¨ uthy: Stable multi-line CW vibration laser near 5 µm based on FH (CN− ) centers in CsBr, Opt. Commun. 72, 214–218 (1989) 311, 314 W. Gellermann, Y. Yang, F. L¨ uthy: Laser operation near 5 µm of vibrationally excited F-center/CN− molecule defect pairs in CsCl crystals, pumped in the visible, Opt. Commun. 57, 196–200 (1986) 311, 314 E. P. Chicklis, C. S. Naiman, R. C. Folweiler, D. R. Gabbe, H. P. Jenssen, A. Linz: High-efficiency room-temperature 2.06-µm laser using sensitized Ho3+:YLF, Appl. Phys. Lett. 19, 119–120 (1971) 314

Crystalline Lasers

347

412. J. A. Caird, L. G. DeShazer, J. Nella: Characteristics of room-temperature 2.3-µm laser emission from Tm3+ in YAG and YAlO3 , IEEE J. Quantum Electron. 11, 874–881 (1975) 314 413. A. A. Kaminskii, T. I. Butaeva, A. O. Ivanov, I. V. Mochalov, A. G. Petrosyan, G. I. Rogov, V. A. Fedorov: New data on stimulated emission of crystals containing Er3+ and Ho3+ ions, Sov. Tech. Phys. Lett. 2, 308–310 (1976) 314, 315 414. E. W. Duczynski, G. Huber, V. G. Ostroumov, I. A. Shcherbakov: CW double cross pumping of the 5 I7 –5 I8 laser transition in Ho3+-doped garnets, Appl. Phys. Lett. 48, 1562–1563 (1986) 314 415. L. Esterowitz, R. Allen, L. Goldberg, J. F. Weller, M. Storm, I. Abella: Diodepumped 2 µm holmium laser, in Tunable Solid-State Lasers II, Springer Ser. Opt. Sci. 52 (Springer, Berlin, Heidelberg 1986) pp. 291–292 314 416. G. J. Kintz, L. Esterowitz, R. Allen: CW diode-pumped Tm3+, Ho3+:YAG 2.1 µm room-temperature laser, Electron. Lett. 23, 616 (1987) 314 417. T. Y. Fan, G. Huber, R. L. Byer, P. Mitzscherlich: Continuous-wave operation at 2.1 µm of a diode-laser-pumped, Tm-sensitized Ho:Y3 Al3 O12 laser at 300 K, Opt. Lett. 12, 678–680 (1987) 314 418. G. J. Kintz, R. Allen, L. Esterowitz: CW laser emission at 2.02 µm from diodepumped Tm3+:YAG at room temperature, paper FB2, Proc. Conf. Lasers and Electro-Optics, CLEO’88 (Opt. Soc. Am., Washington, DC 1988) 314 419. B. T. McGuckin, R. T. Menzies: Efficient CW diode-pumped Tm, Ho:YLF laser with tunability near 2.067 µm, IEEE J. Quantum Electron. 28, 1025– 1028 (1992) 314 420. A. Di Lieto, P. Minguzzi, A. Toncelli, M. Tonelli, H. P. Jenssen: A diodelaser-pumped tunable Ho:YLF laser in the 2 µm region, Appl. Phys. B 57, 3172–3175 (1993) 314 421. T. J. Kane, T. S. Kubo: Diode-pumped single-frequency lasers and Q-switched laser using Tm:YAG and Tm,Ho:YAG, OSA Proc. Adv. Solid-State Lasers 6, 136–139 (Opt. Soc. Am., Washington, DC 1990) 314 422. W. S. Rabinovich, S. R. Bowman, B. J. Feldman, M. J. Winings: Tunable laser pumped 3 µm Ho:YAlO3 laser, IEEE J. Quantum Electron. 27, 895–897 (1990) 315 423. A. F. Umyskov, Yu. D. Zavartsev, A. I. Zagumennyi, V. V. Osiko, P. A. Studenikin: Cr3+:Yb3+:Ho3+:YSGG crystal laser with a continuously tunable emission wavelength in the range 2.84–3.05 µm, Sov. J. Quantum Electron. 26, 563–564 (1996) [transl. from: Kvant. Elektron. 23, 579–580 (1996)] 315 424. A. Diening, S. K¨ uck: Spectroscopy and diode-pumped laser oscillation of Yb3+,Ho3+-doped yttrium scandium gallium garnet, J. Appl. Phys. 87, 4063 (2000) 315, 316 425. E. Sorokin, I. T. Sorokina, A. Unterhuber, E. Wintner, A. I. Zagumenny, I. A. Shcherbakov, V. Carozzo, M. Tonelli: A novel CW mode-locked 2 µm Cr,Tm,Ho:YSGG:GSAG laser, paper CWM1, Conf. Lasers Electro-Optics, Techn. Digest (1998) p. 296 315, 318 426. F. Cornacchia, E. Sani, A. Toncelli, M. Tonelli, M. Marano, S. Taccheo, G. Galzerano, P. Laporta: Optical spectroscopy and diode-pumped laser characteristics of codoped Tm-Ho:YLF and Tm-Ho:BaYF: a comparative analysis, Appl. Phys. B 75, 817–822 (2002) 315

348

Irina T. Sorokina

427. H. Nakajima, T. Yokozawa, T. Yamamoto, H. Hara, N. Djeu: Power optimization of 2 µm Tm,Ho:YAG laser in monolithic crystal, Jpn. J. Appl. Phys. 33, 1010–1011 (1994) 315 428. C. D. Nabors, J. Ochoa, T. Y. Fan, A. Sanchez, H. K. Choi, G. W. Turner, Ho:YAG laser pumped by 1.9-µm diode lasers, IEEE J. Quantum Electron. 31, 1603–1605 (1995) 315 429. Yu. D. Zavartsev, V. V. Osiko, S. G. Semenkov, P. A. Studenikin, A. F. Umyskov: Cascade laser oscillation due to Ho3+ ions in a (Cr,Yb,Ho):YSGG yttrium-scandium-gallium garnet crystal, Sov. J. Quantum Electron. 23, 312–316 (1993) [transl. from: Kvant. Elektron. 20, 366–370 (1993)] 316 430. A. M. Tabirian, H. P. Jenssen, A. Cassanho: Efficient, room temperature midinfrared laser at 3.9 µm in Ho:BaY2 F8 , OSA Trends Opt. Photonics Adv. Solid-State Lasers 50, 170–176 (Opt. Soc. Am., Washington, DC 2001) 316 431. L. A. Pomeranz, P. A. Budni, M. L. Lemons, C. A. Miller, J. R. Mosto, T. M. Pollak, E. P. Chicklis: Power Scaling Performance of Tm:YLF and Tm:YALO Lasers, OSA Trends Opt. Photonics Adv. Solid-State Lasers 26, 458–462 (Opt. Soc. Am., Washington, DC 1999) 432. R. A. Hayward, W. A. Clarkson, D. C. Hanna: High-power diode-pumped room-temperature Tm:YAG and intracavity-pumped Ho:YAG lasers, OSA Trends Opt. Photonics Adv. Solid-State Lasers 34, 90–94 (Opt. Soc. Am., Washington, DC 2000) 316 433. R. C. Stonemann, L. Esterowitz: Efficient, broadly tunable, laser-pumped Tm:YAG and Tm:YSGG CW lasers, Opt. Lett. 15, 486–488 (1990) 316 434. R. C. Stonemann, L. Esterowitz: Efficient 1.94-µm Tm:YALO laser, IEEE J. Sel. Topics Quantum Electron. 1, 78–80 (1995) 316 435. L. Fornasiero, N. Berner, B.-M. Dicks, E. Mix, V. Peters, K. Petermann, G. Huber: Broadly tunable laser emission from Tm:Y2 O3 and Tm:Sc2 O3 at 2 µm, OSA Trends Opt. Photonics Adv. Solid-State Lasers 26, 450–453 (Opt. Soc. Am., Washington, DC 1999) 317 436. E. Sorokin, A. N. Alpatiev, I. T. Sorokina, A. I. Zagumennyi, I. A. Shcherbakov: Tunable efficient continuous-wave room-temperature Tm3+:GdVO4 laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 68, 347–350 (Opt. Soc. Am., Washington, DC 2002) 317 437. P. A. Studenikin, A. I. Zagumennyi, Yu. D. Zavartsev, P. A. Popov, I. A. Shcherbakov: GdVO4 as a new medium for solid-state lasers: some optical and thermal properties of crystals doped with Nd3+, Tm3+ and Er3+ ions, Quantum Electron. 25, 1162–1165 (1995) 317 438. V. A. Mikhailov, Yu. D. Zavartzev, A. I. Zagumennyi, V. G. Ostroumov, P. A. Studenikin, E. Heumann, G. Huber, I. A. Shcherbakov: Tm3+:GdVO4 – a new efficient medium for diode-pumped 2 µm lasers, Quantum Electron. 27, 13–14 (1997) 317 439. Ch. P. Wyss, W. L¨ uthy, H. P. Weber, V. I. Vlasov, Yu. D. Zavartsev, P. A. Studenikin, A. I. Zagumennyi, I. A. Shcherbakov: Emission properties of a Tm3+:GdVO4 microchip laser at 1.9 µm, J. Appl. Phys. B 67, 1–4 (1998) 317 440. A. I. Zagumennyi, V. A. Mikhailov, Yu. D. Zavartsev, S. A. Koutovoi, F. Zerrouk, I. A. Shcherbakov: A diode-pumped 0.6 W CW Tm:GdVO4 microchip laser, CLEO/Europe-EQEC Focus Meeting 2001, Prog. Solid-State Lasers, Techn. Digest, p. 99, paper C-PSL 129 317

Crystalline Lasers

349

441. H. Saito, S. Chaddha, R. S. F. Chang, N. Djeu: Efficient 1.94-µm Tm3+ laser in YVO4 host, Opt. Lett. 17, 189–191 (1992) 317 442. K. Ohta, H. Saito, M. Obara, N. Djeu: Characterization of a longitudinally pumped CW, room-temperature operation of Tm3+:YVO4 laser, Jpn. J. Appl. Phys. 32, 1651–1657 (1993) 317 443. J. F. Pinto, L. Esterowitz, G. H. Rosenblatt: Tm3+:YLF laser continuously tunable between 2.20 and 2.46 µm Opt. Lett 19, 883–885 318 444. G. H. Rosenblatt, J. F. Pinto, R. C. Stonemann, L. Esterowitz: Continuously tunable 2.3 µm Tm:GSGG laser, LEOS ’93 Conf. Proc. IEEE, 689–690 (1993) 318 445. F. Heine, E. Heumann, G. Huber, K. L. Schepler: Mode locking of roomtemperature CW thulium and holmium lasers, Appl. Phys. Lett. 60, 1161– 1162 (1992) 318 446. K. L. Schepler, B. D. Smith, F. Heine, P. A. Budni: Mode-locking of a diodepumped Tm,Ho:YLF, OSA Proc. Adv. Solid-State Lasers 20, 257–259 (Opt. Soc. Am., Washington, DC 1994) 318 447. K. L. Vodop’yanov: Bleaching of water by intense light at the maximum of the λ ∼ 3 µm absorption band, Sov. Phys. JETP 70, 114–121 (1990) 319 448. V. M. Zolotarev, B. A. Mikhailov, L. I. Alperovich, S. L. Popov: Dispersion and absorption of liquid water in the infrared and radio regions, Opt. Spectrosc. 27, 430 (1969) 319 449. F. Frauchiger, W. L¨ uthy: Interaction of 3 µm radiation with matter, Opt. Quantum Electron. 19, 231 (1987) 319 450. E. V. Zharikov, V. J. Zhekov, L. A. Kulevskii, T. M. Murina, V. V. Osiko, A. M. Prokhorov, V. D. Savel’ev, V. V. Smirnov, B. P. Starikov, M. I. Timoshechkin: Stimulated emission from Er3+-ions in yttrium aluminum garnet crystals at λ = 2.94 µm, Sov. J. Quantum Electron. 4(8), 1039–1040 (1975) 319 451. V. I. Zhekov, V. A. Lobachev, T. M. Murina, A. M. Prokhorov: Efficient crossrelaxation laser emitting at λ = 2.94 µm, Sov. J. Quantum Electron. 13, 1235–1237 (1983) 319 452. E. V. Zharikov, V. V. Osiko, A. M. Prokhorov, I. A. Shcherbakov: Crystals of rare-earth gallium garnets with chromium as active media for solid-state lasers, Inorg. Mat. 48, 81–94 (1984) 319 453. P. F. Moulton, J. G. Manni, G. A. Rines: Spectroscopic and laser characteristics of Er,Cr:YSGG, IEEE J. Quantum Electron. 24, 960–973 (1988) 319 454. A. M. Prokhorov, A. A. Kaminskii, V. V. Osiko, M. I. Timoshechkin, E. V. Zharikov, T. I. Butaeva, S. E. Sarkisov, A. G. Petrosyan, V. A. Fedorov: Investigations of the 3 µm stimulated emission from Er3+ ions in aluminium garnets at room temperature, Phys. Stat. Sol. (a) 40, K69–K77 (1977) 319 455. G. J. Kintz, R. E. Allen, L. Esterowitz: Diode-pumped 2.8-µm laser emission from Er3+:YLF at room temperature, Appl. Phys. Lett. 50, 1553 (1987) 320 456. S. A. Pollack, D. Chang, N. L. Moise: Continuous wave and Q-switched infrared erbium laser, Appl. Phys. Lett. 49, 1578–1580 (1986) 320 457. F. Auzel, S. Hubert, D. Meichenin: Multifrequency room-temperature continuous diode and Ar-laser-pumped Er3+ laser emission between 2.66 and 2.85 µm, Appl. Phys. Lett. 54, 681 (1989) 320 458. A. A. Kaminskii, A. G. Petrosyan, G. A. Denisenko, T. I. Butaeva, V. A. Fedorov, S. E. Sarkisov: Spectroscopic properties and 3 µm stimulated emission

350

459.

460. 461.

462.

463.

464.

465. 466. 467.

468.

469.

Irina T. Sorokina of Er3+ ions in the (Y1−x Erx )3 Al5 O12 and (Lu1−x Erx )3 Al5 O12 garnet crystal systems, Phys. Stat. Sol. (a) 71, 291 (1982) 320 S. Hubert, D. Meichenin, B. W. Zhou, F. Auzel: Emission properties, oscillator strength and laser parameters of Er3+ in LiYF4 at 2.7 µm, J. Lumin. 50, 7 (1991) 320 B. J. Dinerman, P. F. Moulton: 3-µm CW laser operations in erbium-doped YSGG, GGG, and YAG, Opt. Lett. 19, 1143 (1994) 320 H. Voss, F. Massmann: Diode-pumped Q-switched erbium lasers with short pulse duration, OSA Trends Opt. Photonics Adv. Solid-State Lasers 10, 217– 221, C. R. Pollock, W. R. Bosenberg (Eds.) (Opt. Soc. Am., Washington, DC 1997) 320 R. Waarts, D. Nam, S. H. Sanders, J. Harrison, B. J. Dinerman: Two dimensional Er:YSGG microlaser array pumped with a monolithic two-dimensional laser diode array, Opt. Lett. 19, 1738 (1994) 320 H. J. Eichler, J. Findeisen, B. Liu, A. A. Kaminskii, A. V. Butachin, P. Peuser: Highly efficient diode-pumped 3-µm Er3+:BaY2 F8 laser, IEEE J. Sel. Topics Quantum Electron. 3, 90 (1997) 320 R. H. Page, R. A. Bartels, R. J. Beach, S. B. Sutton, L. H. Furu, J. E. LaSala: 1 W composite-slab Er:YAG laser, OSA Trends Opt. Photonics Adv. SolidState Lasers 10, 214–216, C. R. Pollock, W. R. Bosenberg (Eds.) (Opt. Soc. Am., Washington, DC 1997) 320 D. Chen, C. L. Fincher, T. S. Rose, F. L. Vernon, R. A. Fields: Diode pumped 1 W continuous-wave Er:YAG 3 µm laser, Opt. Lett. 24, N6 (1999) 321 J. F. Pinto, G. F. Rosenblatt, L. Esterowitz: Continuous-wave laser action in Er3+:YLF at 3.41 µm, Electron. Lett. 30, 1596 (1994) 321 S. R. Bowman, S. K. Searles, N. W. Jenkins, S. B. Qadri, E. F. Skelton, J. Ganem: Diode pumped room temperature mid-infrared erbium laser, OSA Trends Opt. Photonics Adv. Solid-State Lasers 50, 154–156 (Opt. Soc. Am., Washington, DC 2001) 321 M. C. Nostrand, R. H. Page, S. A. Payne, W. F. Krupke, P. G. Schunemann, L. I. Isaenko: Room-temperature CaGa2 S4 :Dy3+ laser action at 2.43 and 4.31 µm and KPb2 Cl5 :Dy3+ laser action at 2.43 µm, OSA Trends Opt. Photonics Adv. Solid-State Lasers 26, 441–449, S. Payne, C. Marshall (Eds.) (Opt. Soc. Am., Washington, DC 1999) 322 A. Dergachev, P. Moulton: Tunable CW Er:YLF diode-pumped laser, in Advanced Solid-State Photonics, OSA Technical Digest, 5–7 (2003) 321

Index

bandwidth – acceptance, 289 – relative, 255 Cr2+:Cd1−x Mnx Te, 308 Cr2+:CdSe, 309 Cr2+:ZnS, 308 Cr2+:ZnSe, 304 cross-relaxation, 273 Dieke diagram, 261

– Fe2+ , 309 – Ho3+ , 314 – Ni2+ , 295–296 – power scaling, 287 – pumping, 285 – thin-disk, 288 – Tm3+ , 316 – transition-metal, 295–310 – tuning, 289–292 – vibronic, 260 lens ducts, 287

end-caps, 288 Er:Cr:YSGG laser, 293 Er:YAG laser, 319, 320 excited-state absorption (ESA), 271

nonlinear figure of merit (FOM), 267 nonradiative decay, 269

FWHM, 289

quantum yield, 284

garnet, 262 ground-state absorption (GSA), 271

radiation trapping, 275

ion–host interaction, 261 Jahn–Teller effect, 267 lanthanide – contraction, 263 laser – Co2+ , 296 – color-center, 310–314 – Cr2+ , 300 – Er3+ , 319 – Er:Cr:YSGG, 293

oscillator strength, 268

Tanabe–Sugano diagram, 265 threshold, 282 tuning – étalon, 291 – acousto-optical, 291 – birefringent filter, 290 – electro-optical, 292 – grating, 289 – prism, 289 upconversion, 273 vibronic laser, 260

Crystalline and Fiber Raman Lasers T. T. Basiev1 , V. V. Osiko1, A. M. Prokhorov1, and E. M. Dianov2 1

2

General Physics Institute Russian Academy of Sciences Vavilova ul., 38, Moscow 119991, Russia Tel/Fax: (7 095) 135-02-67 [email protected] Fiber Optics Research Center at the General Physics Institute, Russian Academy of Sciences Vavilova ul., 38, Moscow 119991, Russia Phone/Fax:(7 095) 135-05-66 [email protected]

Abstract. This chapter describes the state of the art of crystalline and fiber Raman lasers based on the simulated Raman scattering (SRS) effect in crystals and silica-based fibers. It includes historical and theoretical background, analysis of properties of known and newly developed high-efficient SRS crystals, such as LiIO3 , Ba(NO3 )2 , NaNO3 , PbNO3 , CaCO3 , KGW, BaWO4 , SrWO4 ,BaMoO4 ,SrMoO4 , PbWO4 , and germanosilicate and phosphosilicate fibers. A large set of data on IR Raman shifters and lasers operating in the CW, nanosecond, and picosecond regimes with low and high repetition rates is given. Some applications of Raman lasers in medicine, ecology, fiber optics, and communications are discussed.

1 Introduction (Historical and Theoretical Background) Stimulated Raman scattering (SRS) was first observed 40 years ago by Woodbury and Ng [1] and was tightly related to the development of lasers, one of the most important discovery of the 20th century. Previously, spontaneous Raman scattering (RS) was discovered simultaneously and independently by Raman and Krishnan in liquids [2] and by Landsberg and Mandel’shtam in crystals [3]. The stimulated processes of Raman scattering were accounted for in the general Placzek theory [4] developed soon after the discovery of the RS effect. However, only high-peak-power (megawatt) laser sources of light allowed one observe stimulated Raman scattering. SRS scattering was observed [1] when investigating Q-switching of a ruby laser with a nitrobenzene Kerr cell inside the laser cavity. In the laser spectrum, the authors observed an intense IR component with a frequency shifted from the laser line by 1345 cm−1 , which was later explained by stimulated Raman scattering [5]. Since then, this interesting nonlinear phenomenon and its applications in laser spectroscopy and laser engineering have attracted great attention of physicists (see, for instance, the reviews [6,7,8,9,10,11,12,13,14,15,16,17,18] [19,20]). SRS can be used not only when studying the macroscopic behavior of I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 351–400 (2003) c Springer-Verlag Berlin Heidelberg 2003

352

T. T. Basiev et al.

materials under high-power laser radiation, which is important for developing high-power sources of coherent radiation at new wavelengths, but also in the microphysics of individual atoms, ions, and molecules. It is convenient to present the RS process as a two-photon resonance, when the difference of the absorption (νL ) and emission (νS ) optical frequencies is equal to the frequency of an atomic or molecular vibration in Raman media: νL − νS = νR . Due to the two-photon nature, RS cross-sections are very small, σ ≈ 10−30 cm2 . That is why even today this effect is hard to observe in bulk materials without low-divergence high-power monochromatic laser pumping. According to quantum theory [4,5,8,9,10], the intensity of stimulated Raman scattering is proportional to the occupation densities of the states of three interacting fields: laser pump photons (NL ), photons of the generated Stokes wave (Ns ), and crystal phonons (molecular vibrations) in the ground (Ng ) or excited (Ne ) state (N = Ng − Ne ) dNS ∼ NL NS (Ng − Ne ) dz

(1)

In the steady-state regime, when the pump duration tp is much longer than the Raman mode dephasing time TR (tp TR ), the intensity IS of the SRS Stokes beam passing through the Raman medium with length l is given by IS (l) = IS (0) exp(gIL l)

(2)

where IL is the intensity of the incident pump laser beam and g is the gain coefficient, dσ λL λ2S N g= . (3) πcn2S ∆ νR dΩ Here λL and λS are the wavelengths of the pump laser beam and the Stokes Raman beam, respectively; c is the speed of light, nS is the refractive index at the Stokes wavelength; ∆ νR is the full width at half maximum of the Raman spectral line in s−1 units; and Ω is the solid scattering angle. The gain coefficient is the most important parameter for solid-state laser applications, and its magnitude is usually given in units of cm/GW. These expressions show that the intensity of the Raman beam increases with the intensity of the pump laser and with the interaction length. In addition, the gain is greatest for materials with a high density of scatters N , large Raman scattering integral cross-section dσ/dΩ, and small Raman linewidth ∆ νR . In the transient regime when the pump pulse duration is of the same order or shorter than the dephasing time TR and the pump spectral width is broader than that of the Raman line (tp < TR ; ∆ νp > ∆ νR ), the analytical

Crystalline and Fiber Raman Lasers

353

expression for the Stokes intensity in the limit of large gain can be written as [12] 12 tp λL λ2S N dσ . (4) IS (l) = IS (0) exp − exp 2 IL tp l TR n2S dΩ Comparing expression (4) with (2) and (3) one can see that the transient Raman gain has a slower (square root) dependence on the crystal length and the total integral Raman scattering cross-section dσ/dΩ and does not depend on the Raman line broadening ∆ νR at all. The linear gain dependence on the pump intensity for steady-state regimes changes to the square root dependence on the pump pulse energy (E = IL tp ) for transient regimes. Evaluations show that in the steady-state regime it is necessary to have a gain-pump-length product (increment) gIL l = 25 to reach the SRS threshold, when IS ≈ IL . To meet this condition for gaseous and liquid SRS-active media, having g = 1–3 cm GW−1 (for hydrogen g = 1.5 cm GW−1 , for nitrobenzene g = 3 cm GW−1 ), one should use either a very long medium (l = 30–100 cm) or a pump intensity IL > 1 GW cm−2 . The advantages of SRS-active molecular gases (H2 , N2 , O2 , CH4 ) are optical homogeneity, simple formation of the active medium, high frequencies of SRS-active vibrations (2 000–4 000 cm−1 ), and, as a result, large Raman frequency shift νS = νL − νR . Of great importance is the small homogeneous spectral width of an SRS gas vibration (∆ νR = 10−3 –10−1 cm−1 ) and, hence, the relatively large peak RS cross-section for an individual molecule [12,13]. These advantages of molecular gases are balanced by several drawbacks. The first one is the low particle density even for pressures much higher than atmospheric (N < 1021 cm−3 for 10 bar), which must be compensated by a long interaction length (up to several meters). To ensure the high pump intensity (about 109 W cm−2 ) in such a long length is a quite difficult problem. The second limitation is the low thermal conductivity of gases, which makes one use bulky and expensive constant flow gas exchange systems. This considerably complicates the whole device, especially for high pressures. The merits of liquid SRS media [12,13] are the simplicity of preparing an optical SRS cell and the high density of active particles (1022 cm−3 ). But the latter leads to the line broadening of vibrational transitions, particularly for high temperatures (∆ νR ≈ 0.1 cm−1 for T = 77 K and ∆ νR ≈ 1–4 cm−1 for T = 300 K), and reduces the RS peak cross-section, even for an individual molecule. However, these negative features do not eliminate the positive effect of high density, and, thus, molecular liquids have higher Raman gain (g ≈ 3 cm GW−1 ) and allow one to build much more compact devices in comparison with Raman gases. The main drawbacks of liquid SRS-active media are the low thermal conductivity and a high value of dn/dT , resulting in significant thermal optical distortions accompanied by the problems of homogeneity of laminar flows. The high nonlinear refractive index n2 causes self-focusing of the pump beam,

354

T. T. Basiev et al.

which reduces the interaction length, impairs the divergence, and may result in optical breakdown of the medium. Cryogenic molecular liquids (N2 , O2 ) possess high density, small line broadening, and large scattering cross-section and gain coefficient (17 cm GW−1 [13]). When working with these liquids, the main problems are to keep the cryogenic temperatures of the SRS-active medium and provide thermal insulation of optical windows. In addition, most liquids exhibit high scattering losses due to suspended particles and bubbles, which distort the wave fronts of the pump and output SRS radiation. Many of the above-mentioned problems can be overcome by using solidstate SRS media, in particular, single crystals and glass optical fibers. The high density of active particles in crystals (N ≈ 1023 cm−3 ) defines the high SRS gain and, due to the short light–crystal interaction length, makes possible the development of the most compact laser systems [15,16,17,18,19,20,21,22]. The high symmetry of the atomic and molecular positions in crystals prevents variations in the frequencies and linewidths of Raman-active vibrations. Due to the small inhomogeneous broadening of the SRS-active line in the best crystals, the joined response of nuclear vibrations to the pump radiation is very strong, which manifests itself in a lower interaction threshold and in an increase of the SRS gain and conversion efficiency. SRS crystals and Raman lasers are especially important for the mid and far IR regions, where the development of new active laser materials and solidstate lasers based on population inversion is limited by fast nonradiative multiphonon relaxation and low quantum yield. The great number of natural and especially artificially synthesized crystals offers a great variety of properties, such as types of chemical bonds (ionic, ionic-molecular, covalent), transparency range (from 200 nm to 100 µm), vibrational frequencies active in the RS process (10–2000 cm−1 ), and integral cross-section of Raman scattering. Search and development of new effective SRS crystals strongly depend on the insight into the intermolecular and molecule-to-phonon interaction, which defines the line broadening and the phase and population relaxation of vibrational excitations. Let us summarize the advantages of SRS crystals and mention their shortcomings or limitations. Advantages • • • • • • • •

High concentrations of Raman scattering centers, N = 1022 –1023 cm−3 High Raman gain, g = 1–50 cm/GW Small size, 1–5 cm3 Great variety of Raman frequency shifts, from 10 to 2000 cm−1 Wide optical transparency region, from 200 nm to 100 µm High thermal conductivity, k = 1–20 W/cm Low refractive index–temperature derivatives, dn/dT = 10−6–10−5 deg−1 Mechanical hardness


355

• Moisture resistance • Broad range of temperature stability Limitations • Complicated individual technology for every single crystal synthesis, growth, orientation of crystal axes, optical treatment, and coating • Limited pulsed laser damage threshold (0.2–1 GW/cm2 ) and damage threshold due to average power • High price of SRS crystals Glass optical fibers are an excellent Raman medium. Although Raman gain in glass is typically three orders of magnitude lower than in crystals and liquids, low thresholds have been achieved by maintaining high power densities over long lengths of low-loss fibers.

2

Nanosecond SRS Based on Ba(NO3 )2 Crystals

In the 1980s two new synthetic crystals were developed, which defined progress in SRS lasers for several years. These are barium nitrate Ba(NO3 )2 crystal (see [18,19,20,23,24,25,26,27]) and potassium-gadolinium tungstate KGd(WO4 )2 crystals (KGW) [21,22]. The former has a record-high gain (47 cm/GW) in the nanosecond steady-state regime and a high laser damage threshold of 10–20 J/cm2 (see Table 1), but is very soft, plastic, and hygroscopic and has low thermal conductivity. The latter has much better mechanical characteristics and higher thermal conductivity and can be activated by rare earth laser ions, such as neodymium, erbium, etc. [22]. Unfortunately, this crystal has much lower laser damage threshold for nanosecond pulses and smaller steady-state RS peak cross-section. In spite of these drawbacks, KGW crystals found wide application in picosecond lasers, where the integral cross-section is much more important than the peak cross-section, and the laser damage threshold is much higher. One of the first studies of visible SRS in synthetic nitrate crystals Ba(NO3 )2 , Pb(NO3 )2 , and NaNO3 in comparison with natural CaCO3 crystals was reported in [23]. The SRS crystals 15 mm long were studied in a single-pass scheme under 530-nm excitation. The pump laser, consisting of a Nd-glass master oscillator with passive Q-switching, a double-stage amplifier, and a second harmonic generator, produced pulses of green radiation with an energy of 0.8 J and duration of 15 ns. The pump energy density was increased from 1 J/cm2 to the point of laser damage of the SRS crystal, which was 6–15 J/cm2 . For the Ba(NO3 )2 crystal, the minimum SRS threshold energies were about 1 J/cm2 for the first Stokes and 3 J/cm2 for the second Stokes radiation, and the highest total conversion efficiency was equal to 26 %. The first Stokes maximum output energy was as high as 0.15 J at the pump energy density of about 8 J/cm2 , which was close to the surface

356

T. T. Basiev et al.

Table 1. Comparative SRS and thermo-mechanical characteristics of Ba(NO3 )2 , KGd(WO4 )2 , and BaWO4 crystals Raman crystal Characteristic

Ba(NO3 )2 KGd(WO4 )2

BaWO4

Raman frequency shift νR , cm−1

1047

901; 767

924

Raman linewidth ∆ νR , cm−1

0.4

5.4; 6.4

1.6

Dephasing time TR , ps

28

2.0

6.6

Gain coefficients, cm/GW: g(0.53 µm) steady-state (ns) g(0.53 µm) transient (20 ps) g(1.06 µm) steady-state (ns) g(1.06 µm) transient (30–50 ps) g(1.3 µm) steady-state (ns)

47 4.7 11 1.1

11 11 4 3

36–40 14.4 8.5 3.8 5.8

Transparency range, µm

0.33–1.8

0.3–5

0.255–5

Moisture resistance

Low

High

High

Thermal conductivity at 25 ◦ C, W/K·m

1.17

2.5–3.4

3.0

Thermal expansion coefficient α/◦ C 13 × 10−6

(1.6–8.5) × 10−6 6 × 10−6

Hardness

4–4.5 (Moos)

19.2 (NaNO3 Knoop)

4 (Moos) 400 (Knoop)

damage threshold. For the Pb(NO3 )2 crystal, the SRS thresholds were approximately two times higher, while the maximum conversion efficiency was lower by about 20 %, and the first Stokes output of about 70 mJ was reached at the pump density of 6–7 J/cm2 , starting the bulk damage of the sample. The NaNO3 and CaCO3 crystals showed even higher SRS thresholds equal to about 4–5 J/cm2 , but also had higher laser damage threshold, of 14 J/cm2 . The output energies were measured to be 0.15 J with a conversion efficiency of 17 % for NaNO3 and 0.12 J with 14 % efficiency for CaCO3 . Thus, by the first and second Stokes frequency shifts of 1045–1085 cm−1 , all these crystals can provide highly efficient yellow (560 nm) and orange (598 nm) visible radiation, with the best performance shown by the Ba(NO3 )2 crystal. Raman lasers based on Ba(NO3 )2 , NaNO3 , and CaCO3 crystals under external pumping by nanosecond (tp = 10 ns) 532-nm radiation were studied in [24,25]. In a flat–flat resonator 26 cm long with the reflectivity of mirrors R1 ≈ 100 % and R2 = 8 %, the threshold pump densities were equal to 5– 7 MW/cm2 for nitrate crystals and 20 MW/cm2 for calcites (the crystals were 5–8 cm long) [24]. These data well correlate with the directly measured SRS gain coefficients, gBa(NO3 )2 = 47 ± 5 cm/GW, gCaCO3 = 13 ± 3 cm/GW. At a pump density five times higher than the threshold one, the coefficient of the pump energy conversion to all the Stokes components reached


357

the maximum value of 50–65 %. Since the variations in R2 from 4 to 20 % did not change the conversion efficiency, the exit face of the Raman crystal can be used instead of the output mirror. The maximum conversion coefficient in the Ba(NO3 )2 crystal was 40 % for the first Stokes component and 25 % for the second Stokes for single-pass pumping. It was shown that the increase in the number of pump beam passes through the SRS crystal resulted in a better selection of the first Stokes radiation and an increase in the conversion efficiency to 50 % for the first Stokes component for two-pass pumping. The second Stokes component in a flat and unstable resonator can be efficiently suppressed by the application of the forward scattering geometry. Transient stimulated Raman scattering in Ba(NO3 )2 crystals was first studied in [28] under pumping by the Nd:YAG second harmonic. The pump pulse duration was about 22 ± 2 ps, which is less then the dephasing (transverse) relaxation time TR = 28 ps of internal symmetrical vibrations of quasimolecular (NO3 )− complexes of Ba(NO3 )2 crystals. In the single-pass scheme, the SRS threshold intensities of Ba(NO3 )2 crystals 50 and 40 mm long were estimated to be 1.1 and 1.45 GW/cm2 , respectively, which is much higher than for the steady-state nanosecond oscillation. This corresponds to an approximately 10-fold decrease of the picosecond Raman gain in comparison with steady nanosecond pumping. It should be noted that the overall conversion efficiency of a Ba(NO3 )2 crystal at the pump density of 2 GW/cm2 was 25 % for the first Stokes Raman shift (563 nm) and 5 % for the second Stokes (599 nm). The above data, as well as a fast decrease in the SRS gain and fourfold increase in the SRS threshold with the pump wavelength changing from visible 532 to infrared 1064 nm, strongly restrict application of Ba(NO3 )2 crystals for picosecond SRS in the infrared spectral region. Solid-state Raman lasers, discretely or continuously tunable, are of special importance for the IR spectral region, where no dye lasers exist with high quantum yield and good photo and thermal stability, and only a few IR bands are covered by available tunable solid-state lasers based on activated crystals. At the same time, SRS becomes more a complicated problem when moving to the IR spectral region, because the SRS gain coefficient can decrease fourfold with a twofold increase in the pump wavelength. For example, Ba(NO3 )2 crystals have g = 47 cm GW−1 for λ = 532 nm and g = 11 cm GW−1 for λ = 1064 nm [26,27]. By the early 1980s, unique tunable lasers for the near IR range were de+ veloped employing LiF crystals with F− 2 and F2 color centers [29,30,31,32]. These lasers were tunable in the IR region from 0.82 to 1.28 µm. To extend the tuning range further in the IR region, new SRS-active crystals, Ba(NO3 )2 and KGd(WO4 )2 , were offered [33]. The investigations of [33] showed that Ba(NO3 )2 crystals advantageously operate in the near IR range up to 1.7 µm with efficiencies as high as 60 % under nanosecond pumping with power densities well below the laser damage threshold. Further extension into the IR

358

T. T. Basiev et al.

is restricted by the Ba(NO3 )2 fundamental absorption beginning at approximately 1.8 µm. The KGd(WO4 )2 crystals have a much wider transparency region, up to 5 µm, and can also convert the IR radiation to the Stokes region, but the four times smaller steady-state gain coefficient and low radiation resistance restrain their application in nanosecond lasers. The development of SRS shifters based on Ba(NO3 )2 crystals enabled one to extent the spectral range of tunable solid-state lasers to the 1.3–1.7-µm IR region and, as a result, to develop a solid-state laser spectrometer for a very wide near-infrared region, from 0.82 to 1.66 µm. This spectrometer also covered the visible (0.42–0.82 µm) and UV (0.25–0.45 µm) regions by using frequency mixing and second and higher harmonics of IR radiation (Fig. 1) [32,33,34]. It was the first all-solid-state-laser spectrometer operating in all three spectral regions, UV, visible, and IR. Interest in SRS lasers was rekindled in the early 1990s, when solid-state lasers based on impurity doped crystals found so wide application that this aroused the problem of how to prevent the human eye from being hurt by a laser beam. The laser wavelength safe for human eyes was standardized to be λ = 1.54 µm, and the radiation of the majority of lasers developed found themselves in the prohibited wavelength region with λ < 1.5 µm (as an example, Nd:YAG lasers operate at 0.53, 1.06, and 1.32 µm; ruby at 0.69 µm; alexandrite at around 0.78 µm; and Ti:sapphire at 0.7–1.1 µm). In this respect, the development of SRS frequency shifters for available laser radiation to the eye-safe wavelength of 1.54 µm is of great importance. Small atmospheric absorption and high transmittance of fibers at this wavelength allow wide applications of such devices for lidars, free space communication, and fiber links. One of the simplest and most effective ways to get 1.54-µm radiation was proposed in 1993–1994 in [35,36,37,38,39,40,41,42], where a Ba(NO3 )2 Ra-

+

30

Efficiency, %

4ω ω

LiF:F2

3ω ω

-

LiF:F2

1 - St.

2ω ω

10

Ba(NO3)2

2 - St.

3

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

Wavelength, µm − Fig. 1. Tuning curve of an all-solid-state laser based on LiF:F+ 2 ; F2 color center crystals, Ba(NO3 )2 Raman shifters, and frequency doubling and mixing (for UV, visible, and near IR spectral regions)


359

man laser was pumped by the 1.064- and 1.320-µm radiation of a Q-switched neodymium laser. Under 1064-nm nanosecond pumping, a Ba(NO3 )2 crystal 30–50 mm long in a compact cavity with dichroic mirrors, specially designed for optimum lasing of the first, second, and third Stokes waves, produced a single-mode radiation at new, including eye-safe, wavelengths (1.2– 1.4 µm and 1.5–1.6 µm), with light-to-light conversion efficiency of 20–50 % (Fig. 2) [35,36,37,38]. The 1320-nm pumping provides a very good opportunity to build highly efficient, compact eye-safe Raman lasers (λ = 1553 µm), using only the first Stokes radiation conversion. A very high real conversion efficiency of 50 % was obtained in [38] (Fig. 3) at only a few millijoule (4–5 mJ) pulsed pumping of a 40 mm long Ba(NO3 )2 crystal in a compact Raman laser cavity. In this case, a slope efficiency as high as 65 % and a quantum efficiency of 58 % were obtained in the TEM00 mode operation regime. The 1.320 µm nanosecond intracavity pumping of a Ba(NO3 )2 crystal, when the pump and Raman lasers had one common cavity or two optically coupled cavities, enabled increasing the conversion efficiency to the first Stokes at 1.5 µm up to 90 % [39,40,41,42]. In this case, the effect of nonlinear cavity dumping makes it possible to shorten the pump pulse a few times and increase the peak power. An important additional advantage of SRS lasers is that their output beam can have much better spatial quality and smaller beam divergence (“beam cleanup”), and, hence, higher brightness than the pump beam [36,37,38,43].

Output Energy, mJ

6 1st Stokes ηslope=65% 4 2nd Stokes ηslope=50% 2 3d Stokes ηslope=40% 0 0

5

10

15

Pump Energy, mJ

Fig. 2. The first (1197 nm), second (1369 nm), and third (1590 nm) Stokes output energy of Ba(NO3 )2 Raman lasers with optimized coupling reflectivity (18, 26, and 25 %) for 1064-nm pumping and a 51 mm long SRS crystal

360

T. T. Basiev et al.

Output Energy, mJ

3

2

ηslope=65% 1

0 0

2

4

6

Pump Energy (mJ)

Fig. 3. Output energy of the first Stokes (1556 nm) radiation of an eye-safe Raman laser based on a 40 mm long Ba(NO3 )2 crystal versus 1338-nm pump energy of a Nd:YAG laser

A diffraction-limited high-energy (0.25 J) eye-safe Raman laser with wavelengths of 1535 and 1560 nm was developed in 1995 [44,45] when an intracavity Ba(NO3 )2 crystal was pumped by a high-power nanosecond Nd:YAG laser operating at 1.318–1.338 µm. The two standing-wave resonators, the pump laser cavity 69 cm long and the Raman cavity 20 cm long, were coupled by a dichroic beam splitter. Using a 50-mm uncoated Ba(NO3 )2 crystal, the authors had 12 mJ of Raman laser output with 48 % conversion efficiency at the pump power of 25 mJ. The SRS pulse was about two times shorter than the 10-ns pump pulse. Nonlinear Raman beam cleanup was observed in the near and far fields of the 1.5-µm beam profile. The second Stokes radiation at 1.820–1.860 µm was generated through the cascaded SRS process. In an enhanced (more powerful) system with a Nd:YAG rod diameter of 7×110 mm and two flash lamp pump cavities, the output energy of diffraction-limited eye-safe radiation was increased to 0.25 J at the repetition rate of 1 Hz. The repetition rate was limited by the thermal effects due to low thermal conductivity of Ba(NO3 )2 crystals. In order to develop a marine-based lidar laser transmitter, operated around 580 nm, for very turbid water of a surf-zone, the authors of [46] have studied and optimized a Ba(NO3 )2 Raman laser with intracavity pumping by a Nd:YLF laser and with second harmonic generation by a LBO crystal. With a Nd:YAG rod diameter of 7 × 76 mm and a T-shaped high-Q pump laser cavity coupled with a low-Q Raman cavity (400 and 85 mm in length, respectively), the authors realized nonlinear cavity-dumping regimes. They used a 20 mm long MgF2 AR coated Ba(NO3 )2 SRS crystal and the Raman


361

cavity output coupler with R = 25 %. At 50 J of the flash lamp pump energy and 10 Hz, the maximum output for the Raman wavelength of 1176 nm was 150 mJ, which was transferred into 90 mJ of yellow (588 nm) radiation by the LBO (type I) crystal 12 mm long. To stop pulse modulation and frequency cascading, the authors included a prism in the Raman cavity. In this case, they had 1.5-ns pulsed radiation at 588 nm wavelength, which is well suited for coastal water lidar applications.

3

LiIO3 Raman Lasers

One of the first nonlinear crystals with good SRS properties that was used for developing Raman laser systems was lithium iodate, LiIO3 . This crystal is well known by its high nonlinear coefficient for frequency doubling, mixing, and parametric oscillation. In addition, its strong birefringence allows wide angular phase matching for various pumping wavelengths. This crystal also demonstrates a rather wide transparency spectral range (0.3–6.0 µm) and strong Raman scattering lines around 760 and 820 cm−1 . A nanosecond LiIO3 crystalline Raman laser operating in the IR region at a high repetition rate was first described in [47,48]. The pumping Nd:YAG laser operated in a quasi-CW mode at 1080 nm with repetition rate of 1–10 kHz and had a pulse duration of 100 ns. In a shared laser and Raman two-mirror cavity with a LiIO3 SRS crystal 25 mm long, cut at θ ≈ 20.5◦ , the first Stokes radiation (the Raman shift νR = 818 cm−1 ) had an average output power of 1.26 W and pulse duration of 50 ns. In a separate experiment, the second Stokes radiation at 1310 nm was obtained with a power of 0.55 W and pulse duration of 15 ns. Using three-mirror coupled cavities, the authors of [48] had strong pulse shortening, tp(1St) = 4–8 ns with an average output power of 0.9 W, peak power of 7 kW, and conversion efficiency of 77 %. After optimization, the second Stokes pulse duration was even shorter, tp(2St) = 2 ns with average power of 0.2 W and peak power of 100 kW. A picosecond Raman laser based on a LiIO3 SRS crystal was studied in [49]. The authors used a Y-cut lithium iodate crystal 10 mm long and, as the source of synchronous pumping, a self-mode-locked neodymium phosphate glass laser operating at λ = 1054 nm. The pump pulses had a duration of 6–10 ps at the axial interval of 6.7 ns and the duration of the train envelop of about 200 ns. The Raman laser cavity 1 m long was formed by an entrance edge mirror with R1.05µ = 25 % and R1.14c = 98 %, an output mirror with R(1.05µ,1.14µ) = 50 %, and a two-lens intracavity telescope with the LiIO3 crystal inside. The first Stokes component was observed at the wavelength of 1.143 µm with a frequency shift of 760 cm−1 . In the vicinity of the SRS threshold (1 mJ), the output train consisted of 25 pulses and had a total duration of 170 ns. Varying the Raman cavity length, the authors found that the maximum shortening

362

T. T. Basiev et al.

(by a factor of 7–8) of the SRS pulses in comparison with the pump pulses occurs when the Raman cavity length is equal to the length of the pump cavity. The minimum SRS pulse duration was 1 ps at the energy conversion efficiency of 30 %. The maximum conversion efficiency reached 40 %. Recently, a series of studies [50,51,52,53,54] was devoted to practical development of Raman lasers based on a LiIO3 crystal. With flash-lamp quasi-CW (3 kHz) Nd:YAG intracavity pumping, the authors [50] got a 770 cm−1 Raman frequency shift and the first Stokes radiation at the wavelength of 1179 nm with average power of 1.8 W, pulse duration of 40 ns, and optical-to-optical conversion efficiency of 50 %. By using a transversely diode-pumped Nd:YAG laser, the authors of [51,52,53] reached 2.7 W power of the first Stokes output radiation at 1155 nm with electrical-to-optical efficiency of 10 %. The longitudinal diode pumping of the Nd:YAG-LiIO3 Raman laser and optimization of thermal lensing resulted in an increase of the output average power of yellow radiation to 1.42 W at 20 ns pulse duration and 8-kHz repetition rate with electrical-to-optical efficiency of 7.5 % [54]. Unfortunately, till now no data have been published on the steadystate or transient SRS threshold and gain measurements in the LiIO3 nonlinear crystal. Spontaneous Raman scattering characteristics can be found in [15,16,17,18] (see also Table 2 and the discussion below). Among the limitations and shortcomings of LiIO3 SRS crystals, we should mention low values of its thermal conductivity, moisture resistance, and laser damage threshold.

4

Picosecond Raman Lasers Based on KGW Crystals

A specific feature of picosecond Raman converters is that SRS has a nonstationary character when the pump pulse duration is of the same order or shorter than the time of phase relaxation of Raman vibrations. At the same time, the laser damage threshold of nonlinear crystals is much higher when the crystal is irradiated by picosecond laser pulses than by nanosecond ones, i.e. the picosecond operation mode allows pumping with a higher power density without crystal damage. As was shown earlier, for nonstationary pumping, the Raman gain coefficient sharply decreases in comparison with the steadystate regime, loses its dependence on the scattering peak cross-section and pump power, and keeps only the slow square root dependence on the integral Raman scattering cross-section and picosecond pump pulse energy density. The first results on SRS in Nd:KY(WO4 )2 (Nd:KYW) and Nd:KGd(WO4 )2 (Nd:KGW) crystals were obtained in 1984 [55] under picosecond pumping. At the same time, it was also the first study where the Raman material worked simultaneously as an active and nonlinear SRS material.


363

Table 2. Parameters of spontaneous Raman scattering in crystals Material

Lattice Molecular space group group

Diamond

O7h

-

Ba(NO3 )2 NaNO3 CaCO3

T6h D63d D63d

[NO3 ] -”[CO3 ]

1048.6 0.4 1069.2 1.0 1086.4 1.2 Tungstates

21 23 6.0

63 44 10.6

//C4 //C4 //C3 ⊥C3 //C3 ⊥C3

CaWO4 SrWO4 BaWO4 NaGd(WO4 )∗ 2 NaY(WO4 )∗ 2 Sc2 (WO4 )∗ 3 In2 (WO4 )∗ 3 LiIn(WO4 )∗ 2 NaSr(WO4 )∗ 2 ∗ Na2 WO4 KGd(WO4 )2 -”-”-”KY(WO4 )2 -”-”-”KYb(WO4 )2 -”-”-”-

C64h -”-”-”-”-

[WO4 ] -”-”-”-”-”-”-”-

52 50 52 -

18.6 41 64 -

⊥C4 //C4 -

O7h C62h -”-”-”-”-”-”-”-”-”-”-”-

[WO6 ] -”-”-”-”-”-”-”-”-”-”-”-

910.7 6.95 921.5 3 926.5 1.63 919 14∗ 918 15∗ 1024 15 1023 13 920 8 924 2.5 929.2 1.8 901 5.4 901 5.4 768 6.4 768 6.4 905.6 7 905.6 7 767.4 8.4 767.4 8.4 908 7.4 908 7.4 757 15∗∗ 757 15∗∗ Molibdates

50 43 19 59 46 41 17 58 48 48 18∗∗ 58∗∗

35 30 9.2 37 35 30 10 35 34 34 13.8∗∗ 30∗∗

⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2 ⊥C2

CaMoO4 SrMoO4 BaMoO4

C64h -”-”-

[MoO4 ] -”-”-

879.3 887.7 892.4 Iodate and

64 55 52

34 51 64

⊥C4 //C4 ⊥C4 //C4 ⊥C4 //C4

LiIO3 LiNbO3 -”-”LaNbO4

C66 C63v -”-”C32h

[IO3 ] [NbO6 ] -”-”[NbO4 ]

821.6 5.0 872 21.4 632 27 250 28 805 9 Phosphates

25 5 18 22 7.1

//C2 //C3 ⊥C3 ⊥C3 ⊥C2

Ca5 (PO4 )3 F Sr5 (PO4 )3 F LiPO∗ 4

C26h -”-

[PO4 ] -”-”-

964.7 950.3 951

NaClO∗∗∗ 3 NaBrO∗∗∗ 3 ∗∗∗ NH4 Cl -”NH4 SO∗∗∗ 4 Ba3 (B3 O6 )2 SiO2

O1h -”C63v D63

[ClO3 ] [BrO3 ] [NH4 ] -”[SO4 ] [B3 O6 ] [SiO4 ]

937 799.5 1712 3052 976.5 636 464

∗

Raman frequency νV (cm−1 )

Integral crosssection (rel. u.)

1332.9 2.7 100 Nitrates and Calcites

policrystalline sample line with inhomogeneous splitting ∗∗∗ crystal without orientation ∗∗

Raman linewidth ∆ν (cm−1 )

5.0 2.8 2.1 Niobates

2.8 2.8 7.7 Other 4.9∗∗ 2.5∗∗ 6 85∗∗ 3.5 4.5 7

54 44 166 22

Peak Geometry of intensity excitation (rel. u.) scattering K E 100

//C3 ⊥C3

⊥C2 //C2 ⊥C2 //C2 ⊥C2 //C2 ⊥C2 //C2 ⊥C2 //C2 ⊥C2 //C2

⊥C2 ⊥C3 //C3 //C3 //C2

3.4 3.4 -

3.8 3.8 -

⊥C6 //C6 ⊥C6 //C6 -

1 2.2

0.6 1.2

//C3 ⊥C3 ⊥C3 //C3

364

T. T. Basiev et al.

At about 20-J flash lamp pump energy, active elements 5 mm in diameter and 50 mm long oscillated at the fundamental Nd3+ wavelength of 1067 nm in a single-mode regime, producing a train of about 10 ultrashort pulses, each having a pulse width of 11–12 ps at the spectral width of 1.8 cm−1 and energy of 0.5 mJ. The efficiency of stimulated Raman self-conversion to the first Stokes component (λ = 1182.7 nm) for a single picosecond pulse from the beginning of the train was 6.14 % of the total energy and reached 35 % in the middle of the train. In the case when two Nd:KYW crystals were used (one as the lasing and self-converting Raman master oscillator, and the second as the synchronously laser-pumped Raman amplifier (without flash lamp pumping), the authors had a maximum conversion efficiency η1St = 37 % to the first Stokes, η2St = 16 % to the second Stokes (λ2St = 1323.7 nm), and ηaS < 5 % to the anti-Stokes (λaS = 947.9 nm). When the Nd:KYW amplifier was additionally pumped by 30 J of flash lamp radiation, the total output energy increased to 11 mJ. For two Nd:KGW crystals 50 mm long used in one laser cavity, the conversion efficiency of the whole train to the Raman scattering radiation was not so high, η1St = 3.84 % (λ1St = 1180.5 nm), probably due to low quality of the crystals, and was improved with a SRS amplifier to η1St = 16.58 %, η2St = 0.96 %, and ηaS = 0.03 %. The authors also briefly mentioned an oscillation observed in Nd:KLa(MoO4 )2 and Nd:NaLa(MoO4 )2 crystals. The pulse duration at the fundamental wavelength was 5–6 ps and at the first Stokes component, 1.5– 2.5 ps, with the conversion efficiency η1St = 40 %. Almost simultaneously with [55], the intracavity Raman oscillation on Nd:KGW crystals was reported in [56]. The authors studied the frequency doubling of the picosecond radiation of a Nd:KGW laser with passive mode locking. New radiation wavelengths, of 1180 and 1320 nm, were found corresponding respectively to the first and second Stokes components of the SRS conversion of the Nd3+ fundamental wavelength (1067 nm) in a Nd:KGW crystal with a frequency shift of 900 ± 10 cm−1 . This was proved by the SRS radiation of the Nd:KGW crystal that was placed in the picosecond Nd:YAG laser cavity and was not flash lamp pumped. Estimation of the SRS conversion efficiency showed that only 10 % of the total energy belongs to the 1180-nm radiation and the rest (90 %) belongs to the 1067-nm radiation. Therewith, the SRS pulse duration was 3.8 ± 0.4 ps, which was considerably shorter than the pump pulse, by a factor of 2.5. The SRS of undoped KGW crystals with different orientation in an external pumping scheme under 30-ps Nd:YAG laser pumping was studied later in [57]. In the spectral region from 1.06 to 1.32 µm, the SRS radiation of 12 new spectral lines with different combinations of Stokes shifts of 84, 767, and 901.5 cm−1 was recorded. The normalized thresholds and gain coefficients for different polarization orientations were measured. The maximum gain coefficients were measured to be 6.4 cm/GW for the 901.5 cm−1 Stokes frequency


365

shift and 4.3 cm/GW for 767 cm−1 . With increasing pump density, the SRS radiation of higher Stokes components, up to the fourth component with wavelength λ4St = 1730 nm, was recorded. The SRS pulse compression in Nd:KGW and Nd:KYW picosecond lasers was specially investigated in [58,59]. In a cavity 1.3 m long formed by a back dichroic mirror with R1064 nm = 99.94 % and R1180 nm = 72.6 % and by a glass plate as the output coupler, the first Stokes radiation comprised 22–40 % of pump pulse energy at the middle of the pulse train, the first Stokes pulse duration was as short as t1St = 1.3 ps, and the pump pulse duration was also shortened to 3.9 ps. Application of the SRS amplifier on a Nd:KGW crystal with synchronous laser pumping in a passive regime (without flashlamp pumping of Nd3+ ions) allowed the authors to shorten the Stokes pulse to the minimum value of 0.9 ps and increase the conversion efficiency above 40 %. For active amplifying, the Stokes pulse duration grew continuously from 1.1 to 6 ps as the energy Ep of Nd:KGW amplifier pumping increased from 25 to 80 J. The duration of the Nd laser fundamental radiation pulse decreased from 10 to 4.4 ps. A detailed review of the KGW crystal structure, lattice cell parameters, atomic positions and coordinates and interatomic distances can be found in [60,61]. Spectroscopic study of spontaneous and stimulated Raman scattering in KGW crystals with different orientation of the pump beam direction and polarization with respect to the crystallographic axes was also presented. The Stokes components of stimulated Raman scattering, from the first (1162.2 nm) to the fourth (1734.8 nm) ones, were observed under pumping by the 1064-nm radiation of a Nd:YAG laser with pulse duration of 30 ps. The threshold power was 4.8–24.3 GW/cm2 . The threshold pump density for the KGW first Stokes with shifts of 767 and 901.5 cm−1 was measured for different orientations and compared with those measured for CaWO4 and Ba(NO3 )2 crystals [60]. The normalized SRS threshold for KGW was measured to be from 3.9 to 14.8 GW/cm2 depending on the crystal orientation, which was much less than 23–25 GW/cm2 for CaWO4 crystals. The SRS threshold pump density measured for the Ba(NO3 )2 crystal to be 6.2 GW/cm2 , is much less than that for CaWO4 , but higher than the minimum value for KGW at optimum orientation. The nonstationary gain coefficient calculated from the measured thresholds turned out to be minimal for CaWO4 crystals (g = 1–1.1 cm/GW), four times higher for Ba(NO3 )2 crystals (g = 4 cm/GW), and highest for KGW crystals for optimum orientation p[mm]p (g = 6.4 cm/GW for the 901.5 cm−1 Stokes shift). For the KGW Stokes shift of 767 cm−1 and the orientation q[pp]q, this value was g = 5.2 cm−1 . The above data on KGW and KYW Raman lasing properties demonstrate very efficient operation of these crystals in the pico- and subpicosecond regimes of oscillation, where the high density pumping (1–10 GW/cm2 ) can be used without radiation damage of the crystal. These crystals demonstrate the multiwavelength frequency shift of

366

T. T. Basiev et al.

coherent radiation in the most hard-to-reach IR region from 1 to 1.7 µm with a high conversion efficiency (about 40 %) and very effective radiation pulse shortening (more than 10 times) down to subpicosecond pulse durations.

5

Nanosecond Raman Lasers Based on KGW Crystals

The main problem that arises when one tries to use KGW crystals as Raman shifters for nanosecond IR laser radiation is a rather high SRS threshold (due to the low Raman gain) in combination with the low laser damage threshold at nanosecond pumping. The narrow gap between the SRS and laser damage thresholds essentially restricts the SRS conversion efficiency and imposes strong requirements upon the pump radiation profile and the optical quality of the crystal and all the elements of the Raman cavity. The first paper where the authors succeeded in realizing nanosecond intracavity SRS generation on a Nd:KGW crystal was [62]. The pulse duration of the fundamental radiation was tp = 7.5 ns, and the duration of the first Stokes pulse at 1180 nm was measured to be threefold shorter, tSRS = 2.5 ns. The intracavity SRS threshold was estimated to be 150–300 MW/cm2 . However, no data on the conversion efficiency and output energy were reported. Nanosecond SRS in a KGW crystal in a single-pass scheme with an external pumping was also studied in [63]. The steady-state SRS gain coefficients for both Stokes shifts, of 901.5 cm−1 and 767.3 cm−1 , were found to be 6 cm/GW when pumped by a Nd:YAG laser with the radiation wavelength of 1064 nm. For nanosecond (tp = 10 ns) pumping by a Nd:KGW laser in a single-pass scheme, the total efficiency of SRS conversion to both (first and second) Stokes was 70 % with the output energy of 10 mJ and did not depend on whether the Stokes shift frequency was 767.3 or 901.5 cm−1 . Application of KGW crystals for SRS frequency shifting of nanosecond radiation with longer wavelength, 1.1–1.4 µm, where the gain coefficient is much weaker, was first demonstrated in [33]. A KGW crystal 40 mm long was used for nanosecond SRS frequency conversion of a tunable solid-state LiF:F− 2 color center laser MALSAN-201. The LiF:F− 2 color center laser radiation was tuned from 1.1 to 1.23 µm and provided the pump density of about 1 GW/cm2 inside the SRS crystal. Using the single-pass Raman shifting scheme with the pump pulse duration of 8 ns, energy of about 20 mJ, and spectral width of 0.1 cm−1 , the authors observed tunable radiation in new spectral regions, 1.23–1.37 µm (first Stokes) and 1.43–1.6 µm (second Stokes). The maximum conversion efficiency reached 30 % for the first Stokes and 20 % for the second. These results were among the first experiments on solid-state Raman lasers operated in the eye-safe spectral region (≈ 1.54 µm). It should be noted that the laser damage threshold was only 2–3 times higher than the SRS threshold, which strongly limited the reported application of KGW crystals. Ten years later, a compact flash lamp pumped solid-state laser for a range finder operating in the eye-safe spectral region at 1538 nm was developed


367

in [64,65] with application of SRS in a Nd:KGW active crystal self-pumped by 1350-nm (fundamental wavelength) Nd3+ laser radiation. The authors used Nd:KGW active-nonlinear elements with dimensions of ∅3 × 50 and ∅4 × 65 mm. The cavity mirrors had reflectivity R1 = R2 = 99.99 % at the fundamental wavelength of 1350 nm, and the output mirror had 35–40 % reflectivity at the first Stokes wavelength of 1538 nm. Passive polymer and electrooptical LiNbO3 Q-switches were used. In the case of the Nd:KGW crystal with small dimensions, 3 mm in diameter and a length of 50 mm, the threshold pump energy of single-pulse operation changed from 2.5 to 5.5 J, and the energy of the 40-ns SRS pulse increased from 1 to 6 mJ, as the passive Q-switch transmission changed from 45 to 33 %. The KGW crystal with dimensions of 4 mm in diameter and a length of 65 mm allowed increase of the output energy to 9 mJ at the pump energy of 7.5 J. With electrooptical Q-switching by a Brewster cut LiNbO3 crystal, the output energy of 12-ns pulses reached 13.5–15 mJ at the pump energy of 10 J. The overall efficiency was quite high, 0.15 %. Using diode laser pumping of Nd-doped laser crystals, one can build compact and efficient solid-state lasers for the IR and green spectral regions. A transversely diode pumped Nd:KGd(WO4 )2 Raman laser with selffrequency conversion was described in [66]. An AlGaAs/GaAs quasi-CW diode laser operating at 808 nm with a power of 300 W and pulse duration of 300 µs was used for side pumping of a Nd:KGd(WO4 )2 crystal rod 4 mm in diameter and 22 mm in length. The plane-concave Raman laser cavity with an acousto-optical Q-switch inside had a high reflectivity for the fundamental wavelength (1067 nm) and the output coupler transmission of 5 % for the first Stokes wavelength (1162 nm). As a result, the first Stokes pulsed radiation had an energy of 0.1 mJ at a pulse duration of 50 ns, while the pump pulse duration was 140 ns at the repetition rate of 47 Hz. The second harmonic radiation (581 nm) was generated with a 30 % efficiency by a LiB3 O5 crystal with non-critical phase matching at a temperature of 54 ◦ C. Recently, the intracavity self-frequency conversion in microchip Yb:KGW and Yb:KYW crystals was reported by two groups of authors [67,68]. They used 1.7 and 1.1-mm thick plates of Yb-doped laser crystals, pumped by a 980-nm laser diode, and a Cr4+ :YAG plate as the saturable absorber. For output coupler reflectivity of 7 and 5 %, the authors obtained nanosecond pulses of the first Stokes radiation at 1139-nm wavelength with an average power of 7 and 2 mW and repetition rate of 17 and 49 kHz, respectively. The above historical review of the best SRS materials for nano- and picosecond Raman lasers shows that till the mid 1990s, there was no clear understanding of why one SRS material is better for nanosecond operation and another material operates better in the picosecond regime. There was no complete comparative analysis of the main fundamental properties of SRS materials, which would allow one to select the most suitable materials among

368

T. T. Basiev et al.

the known synthetic crystals or to develop SRS materials with required properties.

6 Search for New SRS Materials and Comparative Spectroscopy Study New SRS converters and lasers could not be purposefully developed without a deep insight into the physics of Raman-active vibrational excitations, which determine the nature of light scattering and its properties. In the early 1980s, the complex inhomogeneous nature of RS resonance in fused quartz at a frequency of 460 cm−1 was revealed in [69]. Using LiF:F− 2 color center tunable lasers and the techniques of biharmonic pumping (λ1 = 1.064 µm and λ2 = 1.119 µm), the authors eliminated the inhomogeneous broadening and, using RS amplification spectroscopy, for the first time measured the homogeneous width (16 cm−1 ) of the inhomogeneously broadened (≈ 200 cm−1 ) vibrational spectrum of fused quartz. At the same time, a very efficient pumpto-Stokes radiation conversion (38 %) was observed in 2–3-m long silica fibers. The physics of SRS processes in the unique Ba(NO3 )2 crystal at various temperatures, down to liquid-helium, was studied in [70] by the methods of nanosecond SRS amplification spectroscopy with high spectral resolution and in [71] by transient picosecond spectroscopy of Coherent AntiStokes Raman Scattering (CARS). The unique properties of Ba(NO3 )2 crystal, namely, extremely low SRS threshold and high gain and efficiency, were not understood before. It was found that the high-frequency (1047 cm−1 ) totally symmetric Raman active vibration A1g of NO3 quasi-molecular groups exhibits anomalously low homogeneous and inhomogeneous line broadening (∆ ν = 0.4 cm−1 at room temperature), which reflects the anomalously weak coupling with other quasi-molecular and lattice vibrations [70]. The directly measured dephasing time TR of the A1g vibration proved to be surprisingly long (TR(300 K) = 28 ps) at room temperature and almost ten times longer at liquid helium temperatures (TR(11 K) = 220 ps) [71]. These values are several tens of times higher than those for diamond, calcite, NaNO3 , and other SRS crystals. As the temperature increases to 400–600 K, the homogeneous line broadening ∆ νR (T ) and dephasing rate TR−1 of Ba(NO3 )2 rapidly grow, but still keep much lower values than for other matrices. Experimental data on the dephasing rate TR−1 versus temperature, obtained by RS amplification and picosecond CARS at T = 11–600 K, were described by the theory of multiphonon decay, absorption, and dephasing of vibrational excitations and lattice phonons in a crystal. The analysis showed that the anomalously long lifetime of the A1g vibration in Ba(NO3 )2 crystals mainly resulted from the absence of three-phonon decay processes (when one high-frequency RS-active vibration decays into two vibrations with lower frequencies) [70,71]. At temperatures from 11 to 300 K, these three-phonon relaxation and dephasing


369

processes dominate over all the others in most SRS active materials (except A1g Raman vibrations in Ba(NO3 )2 ) and cause high dephasing rates and significant homogeneous line broadening [20]. It was found that the dephasing and broadening in Ba(NO3 )2 crystals occurs not only due to the allowed four-phonon decay processes ν1047 → ν817 + ω144 + ω81 dominating at low temperatures (4–77 K) but is also due to the six-phonon decay processes ν1047 → ν731 + 4ω81 . The latter, despite a substantially lower probability, exhibit a rapid growth with increasing temperature and make a significant contribution to the line broadening at above-room temperatures [70,71]. Nevertheless, the sum of these two high-order processes demonstrates much less broadening and rate of dephasing than the low-ordered three-phonon processes in other SRS materials. Due to the extending application of SRS crystals, the great variety of their properties, and the wide scatter in published data, in recent years a demand arose for comparative quantitative analysis of SRS-active crystals, which is very important for the search and development of new efficient SRS crystals. Such a detailed analysis both of known SRS crystals and new crystals with promising properties for SRS lasers was performed in [72,73,74,75,76]. Carbonate, nitrate, phosphate, tungstate, niobate, iodate, bromide, borate, and silicate crystals were analyzed. The frequencies and linewidths of the most intense and narrow quasi-molecular Raman vibrations of [CO3 ], [NO3 ], [WO4 ], [MoO4 ], [NbO6 ], [NbO4 ], [IO3 ], [PO4 ], [ClO3 ], [B3 O6 ], [BrO3 ], [SiO2 ], [SO4 ], and [NH4 ] anion groups were compared. The relative intensities (integral and peak) of spontaneous RS lines were measured for more than 30 synthetic crystals in comparison with the reference diamond crystal (100 %). The RS line intensities were studied for various orientations of the crystals with respect to the exciting and scattered beams (see Table 2) [72,73,74,75,76]. It was shown that among the above-listed crystals only the iodate, niobate, tungstate, and molybdate crystals possess significant integral scattering cross-sections (40–60 %) for high-frequency RS lines (600–900 cm−1 ). With the exception of lithium niobate crystals (166 %), these values are somewhat lower than for diamond (100 %) but several times larger than for the known SRS-active crystals of calcite (6 %) and nitrates (20 %). As one can see from (3) and (4), the integral RS cross-section is directly responsible for the SRS threshold and gain both in the steady-state and transient regimes and at the same time demonstrates slow variation for crystals with the same anion group. A different situation was observed for the peak cross-sections of Raman scattering, which determine the SRS threshold and gain coefficient in the steady-state regime under nano- and subnanosecond pumping tp > 10 TR . For example, the uniquely small broadening of RS lines in nitrate crystals results in record high peak values (44–63 %), though they have quite moderate integral RS cross-sections (21–23 %).

370

T. T. Basiev et al.

The analysis of the data given in Table 2 shows that the best crystals for low-threshold SRS frequency converters should have both a large integral RS cross-section and small RS transition broadening. Two promising classes for this purpose are tungstates and molybdates, which possess large integral RS cross-sections (40–60 %) and a great variety of line broadenings (from 1.6 to 15 cm−1 ). Since the line broadening (both due to upconversion and decay dephasing) of high-frequency SRS-active vibrational modes relates to their interaction with lattice phonons, the authors of [73,74] selected the crystals with the heaviest cations to reduce the frequencies of lattice phonon modes. In [73,74,77] simple tungstates of alkaline-earth metals (Ca, Sr, and Ba) with a scheelite structure were synthesized and investigated in detail. As can be seen from Table 2, the CaWO4 SRS crystal frequently encountered in the scientific literature [7,60] has a moderate peak scattering cross-section, equal to 18 %, which is 3.5 times smaller than in the Ba(NO3 )2 crystal and almost two times smaller than in KGd(WO4 )2 . This results from the significant broadening of its RS line, ∆ ν = 6.95 cm−1 at room temperature. Replacement of light calcium cations for heavier strontium and barium cations resulted in a much lower relaxation rate and several times narrower spectral line (1.63 cm−1 ) of BaWO4 Raman vibrations at room temperature. As the cation mass increases in the sequence Ca, Sr, and Ba, the maximum frequency ωlat of lattice phonons in fact decreases, from 274 to 194 cm−1 . In addition, the increase of the lattice constant and cation radius results in the higher frequency of the totally symmetric A1g Raman mode of the [WO4 ]2− quasi-molecular complex (see Table 2). These two factors weaken the vibronic–lattice interactions and considerably narrow the spectrum of Raman vibration, which should manifest itself in a growth of the peak scattering cross-section and steady-state SRS gain. The detailed study of RS fine structure of the A1g oscillation mode in scheelites [78] shows that due to the largest values of lattice constants and W–W distances for the BaWO4 crystal (as well as for Ba(NO3 )2 ), the Davydov splitting (DS) for these crystals is very small, ∆ EDS ≈ 0.5–1 cm−1 . The very low acoustic phonon density of states at this frequency, hωph = ∆ EDs = 0.5–1 cm−1 , cannot provide fast dephasing processes and strong broadening of these Davydov-split components by means of direct single phonon bridge processes. Conversely, in CaWO4 , PbWO4 , and partly in SrWO4 crystals with smaller lattice constants, the Davydov splitting was found to be much higher: 50, 32, and 28 cm−1 , respectively. This can explain the much faster dephasing and stronger broadening (6.9, 4.7, and 3 cm−1 ) in these crystals due to direct single-phonon bridge processes from one DS component to another with absorption or emission of acoustic phonons of frequencies 28–50 cm−1 , 2 whose density of states can be rather high (ρ ≈ ωph ). Similar regularities were observed in the series of alkaline-earth molybdates with a scheelite structure, CaMoO4 , PbMoO4 , SrMoO4 , and BaMoO4 .


371

Table 2 demonstrates the relative values of integral and peak scattering coefficients. As is seen, even with the broader RS lines, the new crystals BaWO4 and BaMoO4 can surpass the unique Ba(NO3 )2 crystal in the peak crosssection, while the SrWO4 and SrMoO4 crystals are only slightly below. The data on the peak cross-sections of spontaneous RS found by the comparative study were subsequently confirmed by direct measurements of SRS gain and SRS threshold [75,76,77,78,79,80]. This gave high potentiality of the newly developed crystals for application in SRS amplifiers and lasers, which will be discussed below.

7

New BaWO4 SRS Crystals

The newly developed BaWO4 crystals for stimulated Raman scattering exhibit many advantages in comparison with the two previously discussed best synthetic crystals, Ba(NO3 )2 and KGd(WO4 )2 (the characteristics of these crystals are compared in Table 1). As one can see, the BaWO4 crystal has much higher hardness, thermal conductivity, and moisture resistance than Ba(NO3 )2 . The transparency range in IR is also much wider, up to 5 µm, for BaWO4 than for Ba(NO3 )2 crystals, which is promising for mid-IR Raman laser development. The Raman scattering integral cross-section of BaWO4 is at least two times higher, which is important for picosecond application, where BaWO4 can provide much higher gain than Ba(NO3 )2 . The extremely small Raman line broadening (0.4 cm−1 ) in Ba(NO3 )2 results in the record-high gain in the nanosecond (steady-state) regime of SRS, but at the same time restricts the steady-state regime on the picosecond time scale. As a result, the Raman gain of Ba(NO3 )2 falls down steeply with shortening the picosecond pulse duration due to the transient character of SRS. The BaWO4 crystal exhibits only a slightly smaller SRS peak cross-section and gain coefficients as compared to those of Ba(NO3 )2 for nanosecond pulses, but, due to the four times larger linewidth (1.6 cm−1 ) and correspondingly four times shorter dephasing time TR , its steady-state regime can be extended to four times shorter laser pulses on the picosecond time scale. Comparing the BaWO4 crystal with KGd(WO4 )2 , one can see almost equally good hardness, thermal conductivity, moister resistance, and transparency range. The integral cross-sections of these two crystals are also quite similar and high. A difference can be found only in the line broadening and the dephasing time. Due to the three times broader Raman line, the KGd(WO4 )2 crystals have much lower nanosecond (steady-state) SRS crosssection and gain coefficient than Ba(NO3 )2 and BaWO4 crystals. Even for the picosecond region 10−11 –10−9 s, the new BaWO4 crystal can show much better SRS operation than KGd(WO4 )2 . Only at few-picosecond or subpicosecond pumping, can KGW crystals overtake the BaWO4 crystal in gain

372

T. T. Basiev et al.

and efficiency due to the deeper transient behavior and faster gain decrease of BaWO4 in comparison with KGW. Similar properties are demonstrated by the SrWO4 and SrMoO4 crystals, whose line broadening and dephasing times lie between those of BaWO4 and KGW crystals and define the intermediate values of the SrWO4 and SrMoO4 SRS cross-section and Raman gain, which are better than for KGW and poorer than for BaWO4 in the nano- and picosecond steady-state regime.

8

BaWO4 and SrWO4 Nanosecond Raman Lasers

The IR SRS laser oscillation using newly developed BaWO4 crystals and two well-known nonlinear crystals, was comparatively studied in [79], using a 9 cm long Fabry–Pérot cavity formed by two plane dielectric mirrors, an input mirror with R1.064 < 5 % and R1.1–1.25 > 98 % and an output mirror with R1.0–1.2 ≈ 55 %. A pulsed transverse single-mode Nd:YAG laser passively Q-switched by a LiF:F2 − crystal was used as the pump source. The pump laser produced 12-ns pulses at 10 Hz repetition rate. The pump beam was focused on the center of the SRS crystal by a lens with a focal length of 50 cm. The beam focal diameter was measured with a CCD camera to be 90 µm at half maximum. The beam shape was close to Gaussian. The Raman laser energy efficiency reaches 26 % for Ba(NO3 )2 and 20.5 for BaWO4 . In both cases, the slope efficiency exceeds 75 %, and the laser thresholds are very close. At the same time, no efficiency saturation is observed for BaWO4 with increasing pump power, hence, one can expect further improvement of the laser characteristics. One can get the best results at the BaWO4 orientation E//C4, and slightly poorer at E⊥C4. An even longer KGW crystal demonstrates a much higher threshold and considerably lower slope efficiency (about 45 %) in the same experimental scheme. IR Raman shifting of the Nd:GGG solid-state laser fundamental wavelength λL = 1062.1 nm in the BaWO4 crystalline Raman laser to the first Stokes wavelength λ1St = 1177.9 nm was studied in [81,82]. Results of the pump-to-output conversion in the BaWO4 Raman laser are shown in Fig. 4, which demonstrates the more than 30-% slope efficiency and the 15 mJ output energy with a megawatt peak power level for a 10 Hz repetition rate. As we discussed above, application of the intracavity IR pumping scheme of a Raman laser can give an opportunity to increase the IR Raman conversion efficiency up to 100 % (by use of the nonlinear intracavity dumping regime), to improve the spatial beam quality (beam clean up), and to decrease the pulse duration with increasing peak power. The Raman laser efficiency of the SrWO4 SRS crystal was studied in [80] under similar conditions. The laser scheme differed from that considered above by a shorter cavity (70 mm), lower reflectivity of the output cou-


373

Output Energy, mJ

15

10

5

0 0

10

20

30

40

50

60

70

Pump Energy, mJ

Fig. 4. The first Stokes (1179, 9 nm) output energy of a BaWO4 Raman laser versus 1062.1 nm pump energy of a Nd:GGG laser

pler (R1.05–1.32 = 30 ± 5 %), double the focal length of the focusing lens (f = 100 cm), and a twice larger focal spot (∅ = 200 µm). To prevent the coupling between the SRS crystal (without antireflection coatings), the Raman laser cavity, and the pump laser cavity, they were realigned to about 1◦ from each other. Figure 5 shows the Raman laser output energy versus the pump energy for the new SrWO4 and BaWO4 crystals in comparison with KGW and Ba(NO3 )2 . One can see that the SrWO4 crystal has a lower threshold and about twice the efficiency (≈ 12 %) than KGW. In spite of the poorer threshold, gain, and output characteristics of SrWO4 in comparison with those

Ba(NO3)2 BaWO4 SrWO4

1st Stokes, mJ

1

KGd(WO4)2

0.1 0

2

4

6

8

10

Pump, mJ

Fig. 5. Dependences of the first Stokes output energy on the pump energy for Ba(NO3 )2 49 mm long (rhombs), BaWO4 43 mm long (triangles), SrWO4 47 mm long (squares), and KGd(WO4 )2 50 mm long (circles) crystals

374

T. T. Basiev et al.

of BaWO4 and Ba(NO3 )2 crystals, the SrWO4 crystal allows a rather high level of rare earth doping, which makes it promising for many applications as a multifunctional nonlinear (SRS) and active material. The SRS self-conversion of laser radiation inside a Nd3+ :SrWO4 laser crystal was first observed under flash lamp pumping in a Q-switched regime [80]. With cavity mirrors of high reflectivity at the fundamental wavelength of 1.06 µm, the first Stokes oscillation at 1.18-µm wavelength was recorded with an output energy of 3 mJ. The first Stokes oscillation of this Raman laser combined two oscillation mechanisms: the broadband stimulated emission of F− 2 color centers and the SRS inside the laser medium. This combination allowed SRS at lower intracavity energy density, i.e. reduced the SRS threshold. 8.1

Mid-IR BaWO4 Raman Laser

An important advantage of barium tungstate crystals is their transparency in the much deeper IR region than the transparency of a Ba(NO3 )2 highgain Raman crystal. This could allow one to use these materials to develop a Raman laser for the mid-IR spectral region (from 2 to 5 µm), which is now one of the most interesting but hard-to-reach regions. The main problem here (besides the transparency) is that the Raman frequency shift of crystals is usually about 1000 cm−1 , hence to obtain mid-IR radiation one should start from as long pump wavelength as possible and use the second or even third Stokes of stimulated scattering. In addition, the Raman gain drops rapidly with longer wavelength of pumping. An attempt to get Raman laser radiation with a longer wavelength was done in [83] using 1.34-µm pump radiation of an accousto-optically Q-switched Nd3+ :YAG laser. A 42-mm long BaWO4 SRS crystal was placed in the cavity formed by a rear plain HR mirror with 90–100 % reflectivity for 1.5–2.0 µm spectral range and about 80 % transmittance at 1.34 µm pumping wavelength. The pump radiation was focused within the Raman laser cavity by a lens with a focal length of 120 mm. The measured oscillation spectra of the BaWO4 Raman laser for an output mirror with R1.54 = 50 % and R1.79 = 14 % reflectivity is shown in Fig. 6a. As can be seen, even for 50 % reflectivity at the first Stokes wavelength (1.53 µm), two additional peaks in the Raman laser output spectrum appear at the wavelengths of 1.79 µm (second Stokes) and 2.15 µm (third Stokes). Two low-intensity maxima near the oscillation lines of the first and second Stokes are associated with another weaker vibronic mode observed in the BaWO4 spontaneous Raman and SRS spectrum with the frequency shift of 332 cm−1 [85]. Examples of input–output characteristics of the mid-IR Raman laser are shown in Fig. 6b. The total conversion slope efficiency from a 1.34-µm pump to the first 1.53 and second 1.79 µm Stokes wavelengths was about 22 %. In this regime, the conversion efficiency for the first and second Stokes radiation was about 10 % each. Latest improvement of this Raman laser provide up to


375

Fig. 6. Oscillation spectra of a BaWO4 Raman laser pumped with a YAG:Nd Q-switched laser (λP = 1.34 µm) (a) and input–output characteristics of a BaWO4 Raman laser in a plane–plane cavity (b)

376

T. T. Basiev et al.

2.2 mJ of 1.7 µm radiation and 1.7 mJ of 2.1 µm radiation with slope efficiency till 14 % each. 8.2 Mid-IR Lasing in a Cr2+ :ZnSe Crystal with BaWO4 Raman Laser Pumping As was shown in the chapter on crystalline mid-IR lasers by Sorokina, Cr2+ doped crystalline hosts such as ZnSe are very promising sources for tunable 2.1–2.7 µm mid-IR lasers. Cr2+ :ZnSe is very attractive due to its near unity fluorescence quantum efficiency at room temperature, high gain cross-section and wide vibronic absorption and emission bands. This crystal has been investigated under different pump sources in CW and pulsed operation. A diffusion Cr2+ -doped ZnSe crystal, 2.5 mm thick, was placed perpendicular relative to the pump and oscillation beams. The crystal demonstrated approximately 75 % absorption at the 1.53-µm pump wavelength. The Cr2+ :ZnSe laser cavity was about 50 mm long and was formed by an HR curved mirror with curvature of 100 mm and a plane output mirror with variable reflectivity. The 1.53-µm pump radiation was focused into the crystal by a lens with a focal length of 120 mm placed near the HR mirror of the Cr2+ :ZnSe laser cavity. The input–output characteristics of the Cr2+ :ZnSe laser with different output mirror reflectivities are shown in Fig. 7a. As can be seen, the lasing threshold is about 200 µJ and the maximum slope efficiency is 11 %. The maximum total efficiency for the output mirror with R = 34 % was determined to be about 8 %. The measured oscillation spectra of the Cr2+ :ZnSe laser are shown in Fig. 7b. The maxima of all the oscillation spectra are in the range 2.45–2.65 µm. The Cr2+ :ZnSe laser oscillation pulse width was measured to be about 7 ns, which is demonstrated in Fig. 7c and reflects the detector response time. Strong pulse shortening of Cr2+ :ZnSe oscillation is a characteristic feature of this regime and could be explained by gain switching in the active medium. As can be seen from Fig. 7c, the delay between the fronts of the pump and oscillation pulses is rather short (50 ns) and goes down to 25 ns with increasing output mirror reflectivity and pump energy.

9

BaWO4 Picosecond Raman Frequency Shifters

A comparative study of picosecond stimulated Raman scattering in the new BaWO4 crystal and in KGd(WO4 )2 and KY(WO4 )2 crystals was fulfilled in [86,87,88,89,90,91,92] for 28 and 40-ps pulses with a wavelength of 532 and 1064 nm, respectively. At visible, green light pumping, the picosecond Raman gain coefficient calculated for the BaWO4 crystals from the pump threshold measurements by the formula gexp = 25/Ip l, was equal to 14.4 cm/GW, which is 30 % higher than that for KGW (11.5 cm/GW) but 30 % lower then for KYW (18.7 cm/GW).


377

0.20

a)

η slope =11% E th =200 µJ

0.15

Eout, mJ

R=34%

0.10

0.05 R=75% R=55% 0.00 0.0

0.5

1.0

1.5

2.0

2.5

E in , mJ

2.5 1.0

2.0

0.8

2 1.5

0.6 R=75% 0.4

Intensity

Intensity, a.u.

R=34%

0.2

c)

Delay

b)

Cr2+:ZnSe oscillation pulse

τ=7 ns

1.0

1

R=55%

BaWO4 SRS pump pulse

0.5

0.0 2400 2450 2500 2550 2600 2650 2700

0.0 -50

Wavelength, nm

0

50

100

150

Time, ns

Fig. 7. Input–output characteristics of a Cr :ZnSe laser (a), oscillation spectra of a Cr2+ :ZnSe laser with different output mirror reflectivities (b) and temporal shapes of BaWO4 SRS pump (1) and Cr2+ :ZnSe laser oscillation (2) pulses for 55 % reflectivity of the output mirror at maximum pump energy (c) 2+

378

T. T. Basiev et al.

In comparison with the steady-state regimes, where the BaWO4 crystals show approximately 3 times higher gain than KGW, this ratio at picosecond pumping is moderate, though the BaWO4 still has a higher gain than the KGW crystal. This is caused by the deeper transient behavior of BaWO4 due to its smaller linewidth and longer dephasing time TR = 6.6 ps of BaWO4 . A much stronger difference can be found when comparing the BaWO4 crystal with Ba(NO3 )2 . Under nanosecond pumping (steady-state regime), BaWO4 has only a 20–30 % lower Raman gain (40 cm/GW against 52 cm/GW) [79,82]), but at the picosecond mode its gain is about three times higher (see Table 1). For 1064-nm picosecond pumping, the BaWO4 Raman gain reaches 3.8 cm/GW, which is slightly lower that for KYW (4.7 cm/GW), but higher than for KGW (3 cm/GW). The maximum conversion efficiency for a 30-mm long BaWO4 crystal under 532-nm picosecond pumping was 30 % for the first Stokes. This is much higher than the 18 % efficiency of an even longer (40 mm) KGW crystal. For the second Stokes scattering, both crystals show a similar efficiency of about 15 %. The detailed study of BaWO4 SRS under 1.06-µm, 50-ps pulsed pumping [91] allowed the authors to obtain the first Stokes radiation at 1180 nm with a conversion efficiency of 25 % in the single-pass and 35 % in the doublepass regimes. Placing a BaWO4 crystal into the short (3.8 cm) cavity with an optimized output coupler provide an almost 100 % slope efficiency and 55 % pump-to-first-Stokes conversion efficiency with the output energy as high as 3 mJ (Fig. 8).

Fig. 8. The first Stokes (1180 nm) output energy of a picosecond BaWO4 Raman laser with output coupler reflectivity R = 33 % and resonator length of 3.8 cm versus 1064-nm pump energy


379

Thus, we can conclude that BaWO4 is the first universal crystal, which shows the almost record-high gain both under nanosecond and picosecond modes of operation.

10

PbWO4 SRS Shifters and Raman Lasers

Another interesting SRS crystal was recently introduced to Raman laser development. This is a well known scintillating material, PbWO4 , with a scheelite structure similar to BaWO4 . Analogously to other tungstates, one can expect a high integral cross-section for the PbWO4 crystal, but the measured linewidth of the totally symmetrical A1g 904 cm−1 Raman vibration shows a rather high value of 5.6 cm−1 for a polycrystalline sample [75] close to KGW. Detailed study of the A1q Raman line broadening at different temperatures up to the melting point shows that, similar to CaWO4 and SrWO4 , lead tungstate PbWO4 has a rather high Davydov splitting ∆ EDS = 32 cm−1 caused by a strong [WO4 ]2− -[WO4 ]2− intermolecular interaction [78]. This value is one or two orders of magnitude higher than ∆ EDS in BaWO4 and Ba(NO3 )2 crystals (with the record high peak value of gain) ∆ EDS = 0–1 cm−1 . The high density of acoustic phonon states leads to fast dephasing and large Raman line broadening ∆ νR (300 K) = 4.7 cm−1 (for single crystal) due to the direct relaxation processes between DS components with single-phonon absorption or emission. The study of the PbWO4 crystal by the picosecond SRS spectroscopy technique and under nanosecond Raman laser pumping was first reported in [93,94,95]. The authors presented the measured data on the gain coefficient (8.4 cm/GW at 532 nm and 3.1 cm/GW at 1064 nm) and some other characteristics of the PbWO4 single crystal. They observed 18 different spectral lines of SRS Stokes and anti-Stokes radiation in the spectral region from 485.5 to 1117 nm under pumping by 1064-nm Nd:YAG laser pulses with a duration of 100 ps. Raman laser oscillation of PbWO4 crystals under external pumping by 100-ps pulses of a Nd:YAG laser at 1064 nm was studied in [93,94]. In the 200 mm long plane-concave laser cavity with spherical (r = 0.5 m) output coupling mirror with a reflectivity of 78 %, the first Stokes radiation energy of 1 mJ with slope efficiency as high as 40 % was obtained at the pump energy of 5 mJ. In those experiments, the pump beam was focused inside the 30 mm long PbWO4 crystal by a lens with a focal length of 12 cm. There are also presented data on the steady-state Raman gain measured in visible light under pumping by 30-ps pulses of the frequency-doubled radiation of a Nd:YAG laser. The authors concluded that the gain coefficient of PbWO4 , measured to be 8.4 cm/GW, is slightly higher than 8 cm/GW of a KGW crystal, measured by the same technique. IR Raman lasing by one or two 45-mm long PbWO4 crystals was studied in a ring Raman laser cavity 285 mm long with two highly reflecting mirrors

380

T. T. Basiev et al.

and one output coupler with a reflectivity of 27 % for the first Stokes wavelength [95]. The Raman crystals were pumped by a Nd:YAG laser (1064 nm) beam 2.1 mm in diameter with a pulse duration of 7 ns. At the pumping energy of 36 mJ (close to the Raman crystal laser damage threshold), the output energy of the first Stokes radiation was as high as 5.3 mJ for one PbWO4 crystal. The highest conversion efficiency reached 20 % for two SRS crystals in the cavity. The authors noted that the laser damage threshold for IR radiation was surprisingly low, at about 2 J/cm2 , which is less than half that for visible radiation. The PbWO4 crystal can also be easily activated by Nd3+ ions [93]. The first attempt to develop a diode-pumped passively Q-switched Nd3+ :PbWO4 Raman laser with self-frequency conversion was described in [96]. An active and SRS Nd(0.5 at.%):PbWO4 crystal with dimensions of 3 × 3 × 15 mm3 was transversely pumped by a 100-W GaAlAs quasi-CW diode laser array (λ = 808 nm) with a pulse duration of 160 µs and a repetition rate of 20 Hz. In a high-reflective cavity 24 mm long, passively Q-switched by a Cr:YAG saturable absorber with T = 95 %, under a pump energy of 4.5 mJ, the Raman output energy at 1170 nm was rather low, at 2.5 µJ, with an efficiency of 0.6 % and threefold output pulse shortening to 8 ns.

11

Diode-Pumped CW Raman Fiber Lasers

Another class of highly efficient medium-power Raman lasers are CW singlemode Raman fiber lasers for the near and mid-infrared. The lasers can operate practically at any wavelength in the region from 1.1 to 1.6 µm with highquality output power of several watts and are widely used for pumping optical fiber amplifiers. Stimulated Raman scattering in a glass fiber was observed for the first time by Stolen et al. in 1971 [97]. The authors observed SRS in a single-mode glass fiber both as single-pass superradiant emission and as Raman oscillator output radiation. The great interest aroused after this pioneering work in the development of Raman fiber oscillators is explained by two of their advantages, namely, long interaction lengths in low-loss optical fibers and broad Raman gain bandwidth in glasses (extended to 500 cm−1 ). In the late 1970s, low-threshold tunable Raman oscillators were created both for visible [98,99,100,101] and near-IR [102,103] regions. The development of highquality low-loss fibers allowed one both to reduce the SRS threshold (below 1 W) and avoid the problems of fiber deterioration [99]. Laser radiation tunable within the spectral region from 5200 to 5600 A was obtained by using four orders of Stokes oscillation [101]. These results showed that laser diodes could be used for pumping Raman fiber oscillators. This promised the possibility of developing compact, efficient, and inexpensive sources of near infrared. But the problem was to achieve an efficient coupling of laser diode radiation to a single-mode fiber


381

core. This problem was solved by the development of double-clad fiber lasers (DCFL) pumped by laser diodes [104]. The authors suggested surrounding an Nd (Yb)-doped single-mode core by a layer with lower refractive index and thus form the core of a multimode waveguide. This structure was in turn surrounded by an outer cladding of still lower refractive index. This multimode structure guided the pump radiation around and through the central rare-earth-doped single-mode core. The shape and dimensions of the inner cladding were specifically chosen to provide efficient end-coupling for the output of high-power laser diode arrays. Such a scheme turned out to be very efficient for pumping Raman fiber lasers and later allowed the development of high-power (up to 35 W in CW mode) diffraction-limited fiber Nd (Yb)-doped lasers operating at around 1 µm (see for example [105,106,107]). The principal elements of modern Raman fiber lasers are fiber Bragg gratings, serving as mirrors for different Stokes cavities. It was found that as many as six Raman Stokes can be obtained by using highly reflective infiber Bragg gratings [108]. In such a scheme, each intermediate Raman Stokes can be resonated and thereby efficiently converted to the next, higher Stokes until the cascade is terminated by a suitable output coupling for the desired Stokes order. Thus, in-fiber Bragg gratings allowed one to transform the multiresonant Raman fiber oscillator [101] into an elegant all-fiber structure. The listed achievements served as a base for the development of Raman fiber lasers, which were required for pumping 1.3 µm Raman fiber amplifiers in the 1990s [109,110]. Further investigations showed that these lasers are also promising as pumping sources for other optical fiber amplifiers, including Er-doped fiber amplifiers. Wide applications of optical amplifiers in optical fiber communication systems, especially in WDM ones, stimulated active investigations of Raman fiber lasers. There were developed various types of efficient Raman fiber lasers, which were based on different fibers and had various Stokes frequency shifts, Stokes cavity designs, and pumping sources. 11.1

Raman Fiber Lasers Based on Germanosilicate Fibers

A widely used type of Raman fiber laser is based on well-developed, commercially available germanosilicate fibers. These fibers have high photosensitivity and thus allow writing Bragg gratings directly in them. The Raman scattering cross-section in germanosilicate fibers is higher than in silica fibers and increases (as well as the photosensitivity) with increasing Ge content. For the first time, the third Stokes (1239 nm 300 mW CW) of a high Gedoped fiber pumped by a 1060-nm Nd:YAG laser was used for pumping a 1.3µm Raman fiber amplifier in [109]. The CW radiation of a diode-pumped germanosilicate Raman fiber laser for the same purpose was first obtained in [110] using a fiber 800 m long, which was H2 -sensitized to increase the photosensitivity. Three pairs of Bragg gratings forming the Stokes cavities were written directly in the germanosilicate fiber. The source of 1060-nm pumping was a diode-pumped Nd DCFL. With a 20 % output coupler, the

382

T. T. Basiev et al.

third Stokes radiation (1240 nm) had a threshold of 175 mW and a slope efficiency of 53 %. A fifth-order cascaded Raman laser based on highly doped germanosilicate fiber was constructed in [108]. Five pairs of Bragg gratings were written in the H2 -sensitized fiber. The 1117-nm pump source was a diode-pumped Yb DCFL. With a 20 % output coupler, the authors obtained the fifth Stokes radiation (1484 nm) with a threshold of 660 mW, slope efficiency of 46 %, and output power of 1.5 W. The laser can be used to pump high-power and remotely pumped Er-doped fiber amplifiers. Even more impressive results on fifth-order cascaded Raman fiber lasers were reported by Innis et al. [107]. The authors managed to obtain 1472-nm output radiation with a power of 8.5 W using pumping by a 20.5-W Yb double-clad fiber laser at 1101 nm. Another way to construct cascaded Raman fiber lasers is to use fused WDM couplers instead of Bragg gratings to form Raman cavities. Such a cascaded Raman fiber laser operating at 1240 nm under pumping by a 1060-nm Yb fiber laser was developed on the base high Ge-doped fiber in [111]. In [112], also using high Ge-doped fiber as the active medium, a 1480-nm Raman fiber laser based on both a fused WDM coupler and Bragg gratings was developed. The efficiency of cascaded Raman lasers with fused WDM couplers turned out to be lower than that of lasers with Bragg gratings [107,108,110]. It should be noted that an essential drawback of Raman lasers based on germanosilicate fibers is the relatively small Raman frequency shift equal to about 440 cm−1 . The problem is that the best pumping sources for Raman fiber lasers operate in the vicinity of 1 µm, while the most desired output wavelength is 1450–1480 nm, which can be used for pumping optical amplifiers (Er-doped and Raman) operating in the 1.55-µm spectral window. Such a conversion can be obtained with six Stokes, i.e. it is necessary to write six pairs of Bragg gratings in a fiber, which makes the laser design too complicated and reduces the conversion efficiency. To have efficient conversion, one should use Raman Stokes of lower orders, i.e. choose fibers with large Stokes shifts. In this connection, it should be noted that in the early 1980s researchers from AT&T Bell Laboratories constructed a Raman fiber laser for the 1.55-µm spectral region pumped by a 1060-nm Nd:YAG laser (see [113] and references therein). The active medium was molecular D2 diffused into a l00-m long single-mode germanosilicate fiber. The Raman shift of D2 is 2972 cm−1 , which allowed the authors to obtain laser radiation at 1560 nm. However, this fiber cannot be used in practice because it must be stored at liquid nitrogen temperature to prevent D2 outdiffusion. 11.2

Phosphosilicate Fiber-Based Raman Lasers

A large Stokes shift can be obtained in phosphosilicate fibers, which show a strong Raman line shifted by 1330 cm−1 [114]. In addition, phosphosili-


383

cate fibers have rather high photosensitivity to permit writing Bragg gratings directly in them by 193-nm radiation [115]. However, phosphosilicate fibers have a grave drawback, specifically, phosphosilicate fibers containing more than 10 mol% of P2 O5 have a high level of optical losses. Low-loss high P-doped fibers were suggested as active media Raman fiber lasers in [116]. In the first experiment [117], the radiation of phosphosilicate fiber pumped by a 1060-nm Nd:YAG laser was obtained at 1.24 and 1.48 µm for the first and second Stokes, respectively. These results demonstrated the high potential of Raman phosphosilicate fiber lasers. Further improvements in the technology of phosphosilicate fibers resulted in reducing optical losses to a level below 1 dB/km for the spectral region from 1.2 to 1.6 µm [118]. As shown in [119], the phosphosilicate fiber Raman scattering spectrum (Fig. 9a) has a narrow line shifted by 1330 cm− 1, which relates to P=O bonds, and a broad band with the maximum at 490 cm−1 , which consists of overlapping SiO2 - and P2 O5 -related components. Hence, it is possible to construct Raman fiber lasers using the two frequency shifts, at 490 cm−1 and 1330 cm−1 , to get new wavelengths of laser radiation. Efficient phosphosilicate fiber-based Raman lasers operating at wavelengths of 1240 nm were developed in [120,121]. The active phosphosilicate fiber of the Raman fiber laser pumped by a Nd fiber laser (1060 nm) contained 13 mol% of P2 O5 (∆ n = 0.011) and was 200 m long. To improve the fiber photosensitivity, the phosphosilicate fiber was loaded with hydrogen for four days under a pressure of 150 bars at a temperature of 50 ◦ C. The gratings were written by an ArF excimer laser (193 nm) using a phase mask technique. The output power of the 1240 nm Raman fiber laser (Fig. 9b) reached a maximum of 2.3 W at a pump power of 3.5 W. The quantum efficiency was 12

a

2.5

2.0

8

6

P1.24, W

Light intensity, a.u.

10

4

b

1.5

1.0

0.5

2

0 0

200

400

600

800 1000 1200 1400 -1

Raman shift, cm

0.0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

P1.06, W

Fig. 9. (a) Raman spectrum of phosphosilicate fiber. (b) Output power of the 1.24-µm Raman laser versus 1.06-µm pump power: (circles) experimental points and (solid line) numerical simulation

384

T. T. Basiev et al.

77 %, which considerably exceeded the published data for the lasers based on germanosilicate fibers [110,111]. A 1.48-µm CW Raman fiber laser was developed in [122]. It consisted of a pigtailed LD array pump module, a Yb-doped double-clad fiber laser, and a 1.48-µm cascaded Raman converter (Fig. 10). The pump light is launched into the first cladding of the Yb fiber through a short piece of standard fiber (Flexcor 1060 with low-index polymer coating, length 1 cm) with a highly reflecting 1.06-µm Bragg grating written in the core. The piece of Flexcor 1060 fiber served as a multimode waveguide for the pump radiation. The output coupler of the Yb fiber laser was formed by a 5 % Bragg grating. The estimated mode field diameters (MFDs) of the Yb fiber and the standard fiber at 1.06 µm were 6.9 and 7.1 µm, respectively, and permitted splicing of these fibers (splicing points S2 and S3) with optical loss of 0.1 dB. The length of the Yb-doped double-clad fiber was 13 m, long enough to absorb the pump radiation at 976 nm. The mean diameter of the first cladding was 125 µm. The multimode pump was transformed into high-brightness 1.06-µm radiation with a slope efficiency of 80 %. The laser power was 3.3 W at the maximum LD array power of 4.5 W, which corresponds to a total light-tolight efficiency of 73 %. The cascaded resonant Raman laser cavity was formed by two pairs of Bragg gratings with a phosphosilicate fiber between them. All the gratings were written in Flexcor fiber after H2 preloading. The Raman laser gratings have reflectivity > 99 %, except the 1.48-µm output coupler, whose reflectivity was 15 %. A small nonresonant excess loss of approximately 0.1–0.15 dB was found in each of the two chains consisting of three Bragg gratings. The phosphosilicate fiber was 1 km long. The fiber core contained 13 mol% of phosphorous, yielding a refractive-index difference of 0.011. The optical losses of the fiber length were 1.7, 1.0, and 0.8 dB at 1.06, 1.24, and 1.48 µm, respectively. The P2 O5 -doped fiber had MFDs of 6.3 and 10.4 µm at 1.06 and 1.48 µm, whereas the corresponding MFDs in the Flexcor fiber were 7.1 and 12.7 µm, respectively. In spite of the mismatched MFDs, the authors managed to achieve optical losses as low as 0.05 dB by optimizing the splicing conditions (splices S4 and S5).

Fig. 10. Experimental setup of 1.48 µm two cascaded Raman laser pumped with a Yb-doped double-clad fiber laser


385

An important feature of the 1.48-µm Raman converter emission spectrum (Fig. 11a) is the absence of silica Stokes (440 cm−1 ) peaks at 1.12 and 1.31 µm, hence rejection optical filters (such as long-period fiber gratings) for suppressing the silica Stokes peaks are not necessary. The suppression of the 1.24-µm radiation corresponding to the first phosphorous Stokes order was 20 dB. The output power of the second-order phosphosilicate Stokes is shown in Fig. 11b. The first Stokes radiation had a threshold of about 0.7 W and increased until the pump power reached the second Stokes threshold (1.5 W); then, the first Stokes power was fixed at a level of 10 mW. The slope efficiency of the 1.48-µm second Stokes radiation with respect to the LD array power was 34 %. The maximum output power was 1 W (at a pump power of 4.5 W), and the spectral width was 0.75 nm (FWHM). Real efficiency could be further improved by reducing the Raman cavity loss.

0

a

Res. 1nm

Intensity, dB

-20

-40

-60

-80 1000

1100

1200

1300

1400

1500

1600

Wavelength, nm

1000

Output power, mW

b 800

600

400

200

0 0

1

2

3

4

5

Pump power, W @980nm

Fig. 11. (a) Emission spectrum of the 1.48-µm phosphosilicate fiber Raman converter. (b) Second-order Stokes 1.48-µm output power versus pump power

386

T. T. Basiev et al.

Raman lasers based on phosphosilicate fibers using 1330 and 490-cm−1 Stokes frequency shifts were developed in [123,124]. The two Raman bands with considerably different frequencies provide more possibilities of generating new laser wavelengths. With a tunable Yb double-clad fiber laser as a pump source, it is possible to obtain any wavelength in the spectral region from 1.1 to 1.6 µm with only three cascades of Raman conversion. The authors of [123,124] suggested Raman fiber lasers operating at 1407 and 1430 nm, which can be used, respectively, for pumping I500-nm Raman fiber amplifiers and in medical devices. The authors used a 25-m Nd-doped fiber as the pump source and a specially developed P-doped silica fiber 975 m long as the Raman active fiber of the 1407-nm laser. It is important that the active fiber had relatively low optical losses at the wavelength of 1.4 µm associated with OH groups (1.1 dB/km). The Raman gain properties of a fiber can be characterized by the fiber Raman gain coefficient g0 = GR /Aeff (in dB/km W), where GR is the mean Raman gain coefficient of the core material (usually expressed in m/W) and Aeff is the effective area of the fiber core. Unlike GR and Aeff , g0 can be measured directly for each fiber ([101,121]). The fiber Raman gain coefficient for the 1407-nm laser were measured to be g0 (1.3 µm/1.24 µm) = 5.4 dB/(km · W) (SiO2 -related Stokes) and g0 (1.24 µm/1.06 µm) = 6.8 dB/(km · W) (P2 O5 -related Stokes). The close values of these Raman gain coefficients enabled the authors to use the generation of both SiO2 - and P2 O5 -related Stokes components. The Raman fiber laser had three optical cavities for the three successive Stokes wavelengths λS1 = 1236 nm, λS2 = 1316 nm, and λS3 = 1407 nm. Here λS1 corresponds to the P-associated Stokes shift of 1330 cm−1 , λS2 and λS3 to the SiO2 -associated Stokes shift of 490 cm−1 . Hence, the successive generation of Stokes components of both fiber core constituents (P2 O5 and SiO2 ) was used. The fiber laser cavities were formed by the pairs of fiber Bragg gratings, written in the germanosilicate fiber and spliced with the Nddoped and P-doped fibers. Figure 12a shows the output spectrum of this Raman laser. The dependence of the third Stokes output power (λS3 ) on the pump power (λP ) is shown in Fig. 12b. The 1047-nm radiation had a linewidth of 0.5 nm and laser slope efficiency of 35 %. This efficiency can be considered high enough taking into account that the generation wavelength λ = 1407 nm coincides with the maximum of OH absorption in the fiber, which leads to high losses in the Raman cavity. A similar laser scheme with a 1.089-nm Yb DCFL pumping laser was used for a laser operating at the wavelength of 1430 nm. Such lasers are very promising for medical applications because their wavelength coincides with a water absorption band. The RFL output power at 1430 nm was 1.4 W. Long wavelength 1.65-µm operation of two stage fiber Raman lasers on phosphosilicate and germanosilicate fibers with output power as high as


387

-20 λ s3

-30

Power, dBm

λp

λ s2

λ s1

-40 -50 -60 -70 1100

1200

a)

1300

1400

W avelength, nm 1.1 1.0 0.9

1407 nm

Output power, W

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0

b)

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Pump power (1062nm), W

Fig. 12. (a) Output spectrum and (b) output power of the 1407-nm phosphosilicate fiber Raman laser under pumping by a Nd fiber laser

1.2 W and conversion efficiency 15 % was realized in [125] under LD pumped Yb fiber laser pumping. Thus, CW Raman fiber lasers pumped by Yb (Nd) double-clad fiber lasers can operate in the spectral region from 1.1 to 1.6 µm with single-mode output power of 1–10 W, spectral bandwidth of radiation of about 1 nm, and conversion efficiency close to 50 %. At present the main application of Raman fiber lasers is the pumping of Raman and Er-doped amplifiers. Using higher power Yb (Nd) DCFL for pumping it is possible to obtain an output power of several tens of watts. Due to the high brightness and quality of output beams and the possibility of choosing generation wavelengths, Raman fiber lasers can also find wide applications in material processing, printing, marking, medicine, and free space optical communication.

388

12

T. T. Basiev et al.

Conclusion

We have demonstrated the wide potentialities of synthetic optical SRS crystals and glass optical fibers and methods of research and development. The large size and high optical quality of synthesized single crystals and fibers with various nonlinear parameters, transparency ranges, and Raman frequencies and linewidths give the opportunity to develop a great variety of solid-state Raman lasers and shifters. Application of stimulated Raman scattering in crystals and fibers extends the spectral range, improves the beam quality, and decreases the pulse width of pump laser radiation. SRS solid-state laser technology is especially important for the mid-IR spectral region, where the development of directly oscillating crystalline lasers with population inversion runs into the problems caused by fast multiphonon relaxation and decreasing quantum yield. A number of crystals and glasses transparent in the mid-IR and with high cubic optical nonlinearity can provide a step-by-step SRS shift of radiation frequency of existing tunable lasers (e.g. Cr:ZnSe 2–3 µm lasers, see the chapter by Sorokina in this volume) to the longer mid-IR region. Acknowledgements This work was partially supported by the International Science and Technology Center and European Office of Aerospace Research and Development (Partner project #2022p).

References 1. E. J. Woodbury, W. K. Ng: Excitation of a ruby oscillator in a near infrared region, Proc. IRE 50, 2367 (1962) 351 2. C. V. Raman, K. S. Krishnan: A new type of secondary emission, Nature (London) 121, 501 (1928) 351 3. G. S. Landsberg, L. I. Mandel’shtam: A new phenomenon of light scattering, Zh. Rus. Fiz. Khim. Obshch. 60, p. 335 (1928) 351 4. G. Placzek: Rayleigh-Streuung und Raman Effect, in Handbuch der Radiologie 6, 2, 2. Auflage (Leipzig 1934) 351, 352 5. R. W. Hellwarth: Theory of stimulated Raman scattering, Phys. Rev. 130, 1850 (1963) 351, 352 6. V. A. Zubov, M. M. Sushchinskii, I. K. Shuvalov: Stimulated Raman scattering of light, Usp. Fiz. Nauk 83, 197 (1964) (in Russian) [Engl. transl. Sov. Phys. Usp., 7419 (1964)] 351 7. G. Eckhart: Selection of Raman laser materials, IEEE J. Quantum Electron. 2, (1966) 351, 370 8. N. Bloembergen: The stimulated Raman effect, Am. J. Phys. 35, 989 (1967) 351, 352 9. R. Loudon: The Raman effect in crystals, Adv. Phys. 13, 423 (1967) 351, 352


389

10. V. N. Lugovoi: Vvedenie v Teoriyu Vynuzhdennogo Kombinalsionnogo Rasseyaniya (Moscow, Nauka 1968) (in Russian) [Introduction to the Theory of Stimulated Raman Scattering] 351, 352 11. A. Glass: Design consideration for Raman lasers, IEEE J. 3, 11, 516–520 (1967) 351 12. W. Kaiser, M. Maier: Stimulated Rayleigh, Brillouin and Raman scattering, in Laser Handbook, F. T. Arecchi, E. O. Shultz-Dubois (Eds.) (North-Holland, Amsterdam 1972) p. 1077 351, 353 13. A. Z. Grasyuk: Raman lasers, Kvantovaya Elektron. 3, 485 (1974) (in Russian) [Sov. J. Quantum Electron.] 351, 353, 354 14. Yu. E. D’yakov, S. Yu. Nikitin: Stimulated Raman scattering, Kvantovaya Elektron. 14, (10) 1925 (1987) (in Russian) [Sov. J. Quantum Electron.] 351 15. T. T. Basiev, R. C. Powell: Introduction to the special issue on solid-state Raman lasers, Opt. Mater. 11, 301 (1999) 351, 354, 362 16. T. T. Basiev: Spectroscopy of new Raman-active crystals and solid-state Raman lasers, Physics Uspekhi 42, 1051–1056 (1999) 351, 354, 362 17. P. G. Zverev, T. T. Basiev, A. M. Prokhorov: Stimulated Raman scattering of laser radiation in Raman crystals, Opt. Mater. 11, 335 (1999) 351, 354, 362 18. R. C. Powell, T. T. Basiev: Seventy years of Raman scattering, in Advances in Laser Physics, V. S. Letokhov, P. Meystre (Eds.) (Harwood Academic, Amsterdam 2000) pp. 55–66 351, 354, 355, 362 19. L. I. Belevtseva, V. N. Voitsekhovskii, N. A. Nazarova, G. I. Romanova, M. V. Shvedova, V. E. Yakobson: Basic properties of barium nitrate optical crystals, Opt. Mekh. Promst. 56, 38 (1989) (in Russian) V. N. Voitsekhovskii, S. M. Karpukhin, V. E. Yakobson: Single crystal barium nitrate and sodium nitrate as efficient materials for laser radiation frequency conversion based on stimulated Raman scattering, J. Opt. Tech. 62, 770–776 (1995) 351, 354, 355 20. P. G. Zverev, T. T. Basiev, V. V. Osiko, A. M. Kulkov, V. N. Voitsekhovskii, V. E. Yakobson: Physical, chemical and optical properties of barium nitrate Raman crystal, Opt. Mater. 11, 315–334 (1999) 351, 354, 355, 369 21. V. A. Berenberg, S. N. Karpukhin, I. V. Mochalov: Nonlinear optics of laser crystal of potassium gadolinium tungstate activated with neodymium Nd3+ :KGd(WO4 )2 , Kvantovaya Elektron. 14, 1849 (1987) (in Russian) [Engl. transl. Sov. J. Quantum Electron. 17, 1178 (1987)] 354, 355 22. I. V. Mochalov: Nonlinear optics of laser crystal of potassium gadolinium wolframate activated with neodymium KGd(WO4 )2 :Nd3+ , Optich. Zhurn. 11, 4 (1995) 354, 355 23. A. S. Eremenko, S. N. Karpukhin, A. I. Stepanov: Stimulated Raman scattering of the neodymium laser second harmonic in nitrate crystals, Kvantovaya Elektron. 7, 196–197 (1980) (in Russian) [Engl. transl. Sov. J. Quantum Electron. 10, 113 (1980)] 355 24. S. N. Karpukhin, A. I. Stepanov: Stimulated emission from the cavity under SRS in Ba(NO3 )2 , NaNO3 , and CaCO3 crystals, Kvant. Elektr. 13, 1572–1577 (1986) (in Russian) [Sov. J. Quantum Electron.] 355, 356 25. S. A. Vitsinskii, V. K. Isakov, S. M. Karpukhin, L. L. Lovchil: Stimulated Raman scattering of copper-vapor laser radiation in barium nitrate crystal, Kvant. Elektr. 20, 1555–1558 (1993) (in Russian) [Sov. J. Quantum Electron.] 355, 356

390

T. T. Basiev et al.

26. S. N. Karpukhin, V. E. Yashin: Generation and amplification of Raman emission in crystals, Kvantovaya Elektron. 10, 1992 (1984) (in Russian) [Sov. J. Quantum Electron. 14, 1337 (1984)] 355, 357 27. V. N. Voitsekhovskii, S. N. Karpukhin, V. E. Yakobson: Barium nitrate and sodium nitrate single crystals as efficient materials for Raman conversion, Opt. Zhurn. Ll, 30–37 (1995) 355, 357 28. P. G. Zverev, J. T. Murray, R. C. Powell, R. J. Reeves, T. T. Basiev: Nonlinear optics of laser crystal of potassium gadolinium wolframate activated with neodymium KGd(WO4 )2 :Nd3+ , Opt. Commun. 97, 59–64 (1993) 357 29. T. T. Basiev, Yu.K. Voron’ko, S. B. Mirov, V.V, Osiko, A. M. Prokhorov: Tunable solid-state lasers based on color centers in ionic crystals, Izv. Akad. Nauk SSSR, Ser. Fiz. 46, 8, 1600–1610 (1982) (in Russian) 357 30. T. T. Basiev: Tunable color-center lasers and their application in selective spectroscopy of disordered media, Izv. Akad. Nauk SSSR, Ser. Fiz. 49, 68–77 (1985) (in Russian) [Bull. Sov. Acad. Sci. USSR Phys. Ser.] 357 31. T. T. Basiev, V. V. Osiko, S. B. Mirov: Room-temperature color center lasers, IEEE J. Quantum Electron. 24, 1052 (1988) 357 32. T. T. Basiev, S. B. Mirov: Room Temperature Tunable Color Center Lasers (Gordon, Breach, New York 1994) 357, 358 33. T. T. Basiev, V. N. Voitsekhovskii, P. G. Zverev, F. V. Karpushko, A. V. Lyubimov, S. B. Mirov, V. P. Morozov, I. V. Mochalov, A. A. Pavlyuk, G. V. Sinitsyn, V. E. Yakobson: Conversion of tunable radiation from a laser utilizing an LiF crystal containing F− 2 color centers by stimulated Raman scattering in Ba(NO3 )2 and KGd(WO4 )2 crystals, Kvantovaya Elektron. 14, 2452–2453 (1987) (in Russian) [Engl. transl. Sov. J. Quantum Electron. 17, 1560–1561 (1987)] 357, 358, 366 34. T. T. Basiev, P. G. Zverev, F. V. Karpushko, V. A. Konyushkin, S. M. Kulashchik, S. B. Mirov, V. P. Morozov, V. S. Motkin, A. G. Papashvili, N. A. Saskevich, G. V. Sinitsyn, V. V. Fedorov: Lasing characteristics of tunable color center lasers, MALSAN 2000, Izv. Akad. Nauk SSSR, Ser. Fiz. 54, 1450–1455 (1990) (in Russian) 358 35. P. G. Zverev, T. T. Basiev: Compact SRS laser on barium nitrate crystal, Proc. All-Union Conf. Opt. Las. 2, Leningrad (1993) p. 363 (in Russian) 358, 359 36. P. G. Zverev, T. T. Basiev: Barium nitrate Raman laser, J. Phys. IV (Paris) 4, C4–599 (1994) 358, 359 37. P. G. Zverev, T. T. Basiev, A. M. Prokhorov: Barium nitrate lasers operating in the near IR spectral region, CLEO/Europe’94, Tech. Digest (1994) paper 94TH0614-8, p. 154 358, 359 38. P. G. Zverev, T. T. Basiev, I. V. Ermakov, A. M. Prokhorov: Stimulated Raman scattering in barium nitrate crystal in an external optical cavity, in Laser Methods of Surface Treatment and Modification, Proc. SPIE 2498, 164 (1994) 358, 359 39. T. T. Basiev, V. B. Sigachev, M. E. Doroshenko, P. G. Zverev, V. V. Osiko, A. M. Prokhorov: Crystalline passive Q-switching of 1.3 µm Nd-lasers and Raman shifting to eye-safe region, CLEO/Europe’94, Tech. Digest (1994) paper 94TH0614-8, p. 125 358, 359 40. T. T. Basiev, V. B. Sigachev, M. E. Doroshenko, P. G. Zverev, V. V. Osiko, A. M. Prokhorov: Passive Q-switching of 1.3 µm Nd-lasers with Nd2+ :SrF2 and V3+ :YAG crystalline saturable absorbers and application to Raman shifting


41.

42.

43. 44. 45.

46.

47.

48. 49.

50.

51. 52.

53. 54.

55.

391

to eye-safe region, in Laser Methods of Surface Treatment and Modification, Proc. SPIE 2498, 171 (1994) 358, 359 P. G. Zverev, T. T. Basiev: Barium nitrate Raman laser for near IR spectral region, OSA Trends Opt. Photonics Adv. Solid-State Lasers 24, 288, B. H. T. Chai, S. A. Payne (Eds.) (Opt. Soc. Am., Washington, DC 1995) 358, 359 P. G. Zverev, T. T. Basiev, V. Jia, H. Liu: Raman spectroscopic and nonlinear optical properties of barium nitrate crystal, OSA Trends Opt. Photonics Adv. Solid-State Lasers 1, 554, S. A. Payne, C. Pollok (Eds.) (Opt. Soc. Am., Washington, DC 1996) 358, 359 J. T. Murray, W. L. Austin, R. C. Powell, N. Peyghambarian: Theory of Raman gain spectrum transformation, Opt. Mater. 11, 373 (1999) 359 J. T. Murray, R. C. Powell, E. L. Austin: High brightness, eye-safe lasers, Opt. Photon. News 6, 32 (1995) 360 J. T. Murray, R. C. Powell, N. Peyghambarian, D. Smith, W. Austin, R. A. Stolzenberger: Generation of 1.5 µm radiation through intracavity solidstate Raman shifting in Ba(NO3 )2 nonlinear crystals, Opt. Lett. 20, 1017– 1019 (1995) 360 J. T. Murray, W. L. Austin, L. K. Calmes, R. C. Powell, G. J. Quarles: Nonlinear cavity-damped intracavity solid state Raman laser transmitters, in OSA Trends Opt. Photonics Adv. Solid-State Lasers 20, 72–76, S. A. Payne, C. K. Pollock (Eds.) (Opt. Soc. Am., Washington, DC 1996) 360 E. O. Ammann, C. D. Decker: Raman spectroscopic and nonlinear optical properties of barium nitrate crystal, J. Appl. Phys. 48, 1973–1975 (1977) 361 E. O. Ammann: High average power Raman oscillator employing a shared resonator configuration, Appl. Phys. Lett. 32, 52–54 (1978) 361 G. G. Grigoryan, S. B. Sogomonyan: A synchronously pumped picosecond SRS-laser on a LiIO3 crystal, Kvantovaya Electronika 16, 2180–2183 (1989) [Sov. J. Quantum Electron.] 361 H. M. Pask, J. A. Piper: Practical 580-nm source based on frequency doubling of an intracavity-Raman-shifted Nd:YAG laser, Opt. Commun. 148, 285–288 (1998) 362 H. M. Pask, J. A. Piper: Efficient all-solid-state yellow laser source producing 1.2 W average power, Opt. Lett. 24, 1490–1492 (1999) 362 M. Revermann, H. M. Pask, Justin L. Blows, T. Omatsu: Thermal lensing measurements in an intracavity LiIO3 Raman laser, in OSA Trends Opt. Photonics Adv. Solid-State Lasers 34, 506–509, H. Injeyan, U. Keller, Ch. Marshal (Eds.) (Opt. Soc. Am., Washington, DC 2000) 362 H. M. Pask, J. A. Piper: Diode-pumped LiIO3 intracavity Raman laser, IEEE J. Quantum Electron. 36, 949–855 (2000) 362 H. M. Pask, J. A. Piper: Design and operation of an efficient 1.4 W diodepumped Raman laser at 578 nm, in Tech. Digest, Int. Conf. Adv. Solid-State Lasers, Quebec, Canada, Feb. 2002 (Opt. Soc. Am., Washington, DC 2002) WD1 362 K. Andryunas, Yu. Vishakas, V. Kabelka, I. V. Mochalov, A. A. Pavlyuk, G. T. Petrovskii, V. Syrus: Raman self-conversion of Nd3+ laser radiation in crystals of double tungstates, Pis’ma v Zhurnal Eksperi Teor. Fiz. 42, 333 (1985) (in Russian) [Engl. transl. JETP Lett. 42, 410–412 (1986)] K. Andryunas, Yu. Vishakas, V. Kabelka, I. V. Mochalov, A. A. Pavlyuk, N. V.

392

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

T. T. Basiev et al. Ionina, V. Syrus: A method of generation of picosecond pulses at a Raman frequency, USSR patent no. 1227074 (1984) (in Russian) 362, 364 A. M. Ivanyuk, M. A. Ter-Pogosyan, P. A. Shakhverdov, V. D. Belyaev, V. L. Ermolaev, H. P. Tikhonova: Picosecond radiation pulses at intracavity stimulated Raman scattering inside the active element of neodymium laser, Optika i Spectrosk. 59, 950–952 (1985) (in Russian) [Sov. Opt. Spectrosc.] 364 A. V. Mikhailov, I. V. Mochalov, A. V. Lyubimov: Raman scattering in KGW crystals, in Lasers and Optical Nonlinearity, E. K. Maldutis (Ed.) (Vilnyus, 1987) pp. 265–269 (in Russian) 364 K. Andryunas, A. Barila, Yu. Vishchakas, V. Syrus: Investigation of pulse duration dynamics in a laser based on KGd(WO4 )2 :Nd3+ and KY(WO4 )2 :Nd3+ in Lasers and Optical Nonlinearity, E. K. Maldutis (Ed.) (Vilnyus, 1987) pp. 43–50 (in Russian) 365 K. Andryunas, A. Barila, Yu. Vishchakas, I. V. Mochalov, G. T. Petrovskii, V. Syrus: Temporal characteristics of picosecond pulses of SRS self-conversion, Optika i Spektroskopiya 64, 397–401 (1988) (in Russian) [Sov. Opt. Spectrosc.] 365 Yu. K. Vishchakas, I. V. Mochalov, A. V. Mikhailov, R. F. Klevtsova, A. V. Lyubimov: Crystal structure and Raman scattering in KGd(WO4 )2 crystals, Lietuvos fizikos rinkinys 28, 224–234 (1988) 365, 370 I. V. Mochalov: Laser and nonlinear properties of the potassium gadolinium tungstate laser crystal KGd(WO4 )2 :Nd3+ (KGW:Nd), Opt. Eng. 36, 1660– 1669 (1997) 365 A. M. Ivanyuk, V. A. Sandulenko, M. A. Ter-Pogosyan, P. A. Shakhverdov, V. G. Chervinskii, A. V. Lukin, V. L. Ermolaev: Intracavity stimulated Raman scattering in a nanosecond neodymium laser based on potassium-gadolinium tungstate, Opt. Spektrosc. 62, 961–962 (1987) (in Russian) []Sov. Opt. Spectrosc.] 366 V. A. Berenberg, S. N. Karpukhin, I. V. Mochalov: SRS of nanosecond pulses in KGd(WO4 )2 crystals, Kvantovaya Electronika 14, 1849–1850 (1987) (in Russian) [Sov. Quant. Electron.] 366 A. A. Kaminskii, N. S. Ustimenko, A. V. Gulin, S. N. Bagaev, A. A. Pavlyuk: Raman-parametric interactions in monoclinic KGd(WO4 )2 and KY(WO4 )2 :Nd3+ crystals: picosecond multicomponent Stokes and anti-Stokes generation and nanosecond self-SRS conversion in human eye-safe 1.5 µm range, Dokl. Akad. Nauk 359, 179–183 (1998) (in Russian) [Sov. Phys. Docl.] 367 N. S. Ustimenko, A. V. Gulin: 1.538 µm wavelength radiation of KGd(WO4 )2 :Nd3+ lasers based on stimulated Raman scattering with self-conversion, Pribory i Tekhnika Eksperimenta. 3, 99–101 (1998) (in Russian) [Sov. Instr. Exp. Techn.] 367 J. Fideisen, N. J. Eichler, P. Peuser: Self-stimulating, transversely diode pumped Nd:KGd(WO4 )2 Raman laser, Opt. Commun. 181, 129–133 (2000) 367 A. A. Lagatsky, A. Abdolvand, N. V. Kuleshov: Passive Q-switching and selffrequency Raman conversion in a diode-pumped Yb: KGd(WO4 )2 laser, Opt. Lett. 25, 616–618 (2000) 367 A. S. Grabtchikov, A. N. Kuzmin, V. A. Lisinetskii, V. A. Orlovich, A. A. Demidovich, N. V. Eichler, A. Bednarkiewics, W. Strek, A. N. Titov: Yb:KGW


69.

70.

71. 72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

393

microchip laser performance: fundamental frequency generation and Raman self-frequency conversion, in Tech. Digest, Int. Conf. Adv. Solid State Lasers, Quebec, Canada, Feb. 2002 (Opt. Soc. Am., Washington, DC 2002) WB 6 367 T. T. Basiev, E. M. Dianov, E. A. Zakhidov, A.Ya. Karasik, S. B. Mirov, A. M. Prokhorov: Selective nonlinear spectroscopy of inhomogeneously-broadened phonon resonances in disordered medium, Pis’ma Zh. Eksp. Tear. Fiz. 37, 192 (1983) (in Russian) [Engl. transl. JETP Lett. 37229 (1983)] 368 P. G. Zverev, T. T. Basiev: Study of line broadening of Raman-active vibration in barium nitrate crystal by the method of two-photon Raman gain spectroscopy, Kvantovaya Elektron. 22, 12, 1241 (1995) (in Russian) [Sov. J. Quantum Electron.] 368, 369 P. G. Zverev, W. Jia, H. Liu, T. T. Basiev: Vibrational dynamics of the Ramanactive mode in barium nitrate crystal, Opt. Lett. 20, 2378 (1995) 368, 369 T. T. Basiev, A. A. Sobol, P. G. Zverev, Yu. K. Voron’ko, V. V. Osiko, R. C. Powell: Comparative Raman spectroscopy study of crystals for Raman lasers, in OSA Trends Opt. Photonics Adv. Solid-State Lasers 19, 546, W. R. Rosenberg, M. M. Fejer (Eds.) (Opt. Soc. Am., Washington, DC 1998) 369 T. T. Basiev, A. A. Sobol, P. G. Zverev, V. V. Osiko, R. C. Powell: Comparative spontaneous Raman spectroscopy of crystals for Raman lasers, Appl. Opt. 38, 594 (1999) 369, 370 T. T. Basiev, A. A. Sobol, P. G. Zverev, L. I. Ivleva, V. V. Osiko, R. C. Powell: Raman spectroscopy of crystals for stimulated Raman scattering, Opt. Mater. 11, 307 (1999) 369, 370 T. T. Basiev, A. A. Sobol, P. G. Zverev, V. V. Fedorov, M. E. Doroshenko, L. I. Ivleva, N. M. Polozkov, V. V. Osiko, A. M. Prokhorov, G. Hager: Raman spectroscopy of tungstate crystals for Raman lasers, in Laser ’98, V. J. Corcoran, T. A. Goldman (Eds.), Proc. Int. Conf. Tucson, Dec. 1998 (SRS, McLean, VA 1999) p. 712 369, 371, 379 T. T. Basiev, P. G. Zverev, A. A. Sobol, V. V. Fedorov, M. E. Doroshenko, V. V. Skornyakov, L. I. Ivleva, V. V. Osiko: Perspectives of tungstate crystals for Raman lasers, in Novel Lasers and Devices – Basic Aspects, Tech. Digest CLEO/Europe Focus Meeting 1999, Munich (2001) p. 160 369, 371 T. T. Basiev, A. A. Sobol, P. G. Zverev, L. I. Ivleva, V. V. Osiko: Laser material for stimulated Raman scattering, Russian patent no. 2178938 (27.01.2002, C1 Bull N3) 370, 371 T. T. Basiev, A. A. Sobol, Yu.K. Voron’ko, P. G. Zverev: Spontaneous Raman spectroscopy of tungstate and molybdate crystals for Raman lasers, Opt. Mater. 15, 205–216 (2000) 370, 371, 379 P. G. Zverev, T. T. Basiev, A. A. Sobol, V. V. Skornyakov, L. I. Ivleva, N. M. Polozkov, V. V. Osiko: Stimulated Raman scattering in alkaline-earth tungstate crystal, J. Quantum Electron. 30, 55–59 (2000) 371, 372, 378 T. T. Basiev, P. G. Zverev, A. A. Sobol, V. V. Osiko: Search and charcterization of new crystals for Raman lasers, in Tech. Digest Adv. Solid State Lasers (Opt. Soc. Am., Washington, DC 2002) MB 10 371, 372, 374 T. T. Basiev, N. E. Doroshenko, P. G. Zverev, A. M. Prokhorov: Solid-state laser of yellow spectral region, Russian patent no. 2178939 (27.01.2002, C1 Bull N3) 372

394

T. T. Basiev et al.

82. P. G. Zverev, T. T. Basiev, M. E. Doroshenko, V. V. Osiko: Barium tungstate Raman laser – A new coherent source for sodium star experiments, OSA Trends Opt. Photonics Ser. 34, 348–354 (Opt. Soc. Am., Washington, DC 2000) 372, 378 83. T. T. Basiev, M. E. Doroshenko: Eye-safe mid IR BaWO4 Raman laser oscillation in 1.5–2.15 µm spectral region, Optics of Lasers, Int. Conf., St. Petersburg, 2003 374 84. T. T. Basiev, T. T. Doroshenko, et al.: Short pulse Cr2+ :ZnSe laser oscillation under BaWO4 Raman laser pumping, J. Quant. Electron. (to be published) 85. P. G. Zverev, T. T. Basiev, A. A. Sobol, I. V. Ermakov, W. Gellerman: BaWO4 crystal for quasi-CW yellow Raman laser, in Tech. Digest Adv. Solid-State Lasers (Opt. Soc. Am., Washington, DC 2001), pp. 124–125 374 ˇ 86. P. Cern´ y, H. Jel´ınkov´ a, T. T. Basiev, P. G. Zverev: Properties of transient and steady-state stimulated Raman scattering in KGd(WO4 )2 and BaWO4 tungstate crystals, in Growth, Fabrication, Devices and Applications of Laser and Nonlinear Devices, Proc. SPIE 4268, 101–108 (2001) 376 ˇ ˇ 87. P. Cern´ y, H. Jel´ınkov´ a, J. Sulc, P. G. Zverev, T. T. Basiev: Evaluation of BaWO4 steady-state Raman gain in single- and double-pass arrangement, in OSA Trends Opt. Photonics Adv. Solid-State Lasers 50, (Opt. Soc. Am., Seattle, WA 2001) 376 ˇ 88. P. Cern´ y, P. G. Zverev, H. Jel´ınkov´ a, T. T. Basiev: Efficient Raman shifting of picosecond pulses using BaWO4 crystal, Opt. Commun. 177, 397–404 (2000) 376 ˇ 89. P. Cern´ y, P. G. Zverev, H. Jel´ınkov´ a, T. T. Basiev, L. I. Ivleva, V. V. Osiko: Comparison of stimulated Raman scattering of picosecond pulses in tungstate crystals, in Nonlinear Materials, Devices, and Applications, J. W. Pierce (Ed.), Proc. SPIE 3928, 124–131 (2000) 376 ˇ 90. P. Cern´ y, H. Jel´ınkov´ a, P. Zverev, T. Basiev, V. Kubeˇeek: Picosecond Raman gain of CaCO3 , KGd(WO4 )2 and BaWO4 crystals, in Book of Abstracts, paper CthE14, Conf. Lasers and Electro-Optics Europe, Nice, France; in CLEO Europe’2000 Tech. Digest (2000) p. 94 376 ˇ 91. P. Cern´ y, H. Jel´ınkov´ a, M. Miyagi, T. T. Basiev, P. G. Zverev: Efficient picosecond Raman lasers based on BaWO4 and KGd(WO4 )2 tungstate crystals emitting in 1.15 to 1.18-µm spectral region, accepted for presentation at LASE 2002, paper no. 4630-17 376, 378 92. P. Cherny, H. Jelinkova: Picosecond stimulated Raman scattering in BaWO4 crystal with close to quantum limit efficiency, Tech. Digest, Int. Conf. Adv. Solid-State Lasers, Quebec, Canada, Feb. 2002 (Opt. Soc. Am., Washington, DC 2002) paper MB 11 376 93. A. A. Kaminskii, H. J. Eichler, K. Ueda, N. V. Klassen, B. S. Redkin, L. E. Li, J. Findeisen, D. Jaque, J. Garcia-Sole, J. Fernandez, R. Balda: Properties of Nd3+ -doped and undoped tetragonal PbWO4 , NaY(WO4 )2 , CaWO4 , and undoped monoclinic ZnWO4 and CdWO4 as laser-active and stimulated Raman scattering-active crystals. Appl. Opt. 38, 4533–4546 (1999) 379, 380 94. J. Findeisen, H. J. Eichler, A. A. Kaminskii: Efficient picosecond PbWO4 and two-wavelength KGd(WO4 )2 Raman lasers in the IR and visible, IEEE J. Quant. Electron. 35, 173–178 (1999) 379 95. A. A. Kaminskii, C. L. McCray, H. R. Lee, S. W. Lee, D. A. Temple, T. H. Chyba, W. D. Marsh, J. C. Barnes, A. N. Annanenkov, V. D. Legun, H. J. Eich-


96.

97. 98. 99. 100. 101. 102.

103. 104.

105.

106.

107.

108.

109.

395

ler, G. M. A. Gad, K. Ueda: High efficiency nanosecond Raman lasers based on tetragonal PbWO4 crystals, Opt. Comm. 183, 277–287 (2000) 379, 380 W. Chen, Y. Inagawa, T Omatsu, M. Tateda, N. Takeuchi, Y. Usiki: Selfstimulating, passively Q-switched, diode-pumped Nd3+ :PbWO4 Raman laser, Techn. Digest, Int. Conf. Adv. Solid-State Lasers, Seattle, WA, Jan. 2001 (Opt. Soc. Am., Washington, DC 2001) 380 R. H. Stolen, E. P. Ippen, A. R. Tynes: Raman oscillation in glass optical waveguide, Appl. Phys. Lett. 20, 62–65 (1972) 380 K. O. Hill, B. S. Kawasaki, D. C. Johnson: Low-threshold CW Raman laser, Appl. Phys. Lett. 29, 181–183 (1976) 380 R. K. Jain, Ch. Lin, R. H. Stolen, W. Pleibel, P. Kaiser: A high-efficiency tunable CW Raman oscillator, Appl. Phys. Lett. 30, 162–164 (1977) 380 D. C. Johnson, K. O. Hill, B. S. Kawasaki, D. Kato: Tunable Raman fiber-optic laser, Electron. Lett. 13, 53–54 (1977) 380 R. K. Jain, Ch. Lin, R. H. Stolen, A. Ashkin: A tunable multiple Stokes CW fiber Raman oscillator, Appl. Phys. Lett. 31, 89–90 (1977) 380, 381, 386 Ch. Lin, L. G. Cohen, R. H. Stolen, G. W. Tasker, W. G. French: Near- infrared sources in the 1–1.3 µm region by efficient stimulated Raman emission in glass fibers, Opt. Commun. 20, 426–428 (1977) 380 Ch. Lin, R. H. Stolen, L. G. Cohen: A tunable 1.1-µm fiber Raman oscillator, Appl. Phys. Lett. 31, 97–99 (1977) 380 H. Po, E. Snitzer, L. Tumminelli, F. Hakimi, N. M. Chu, T. Haw: Doubly clad high brightness Nd fiber laser pumped by GaAlAs phased array, Opt. Fiber Commun. Conf., Houston (1989) PD 7 381 H. Po, J. D. Cao, B. M. Laliberte, R. A. Minns, R. F. Robinson, B. H. Rockney, R. R. Tricca, Y. H. Zhang: High power neodymium-doped single transverse mode fiber laser, Electron. Lett. 29, 1500–1501 (1993) 381 M. Muendel, B. Engstrom, D. Kea, B. Laliberte, R. Minns, R. Robinson, B. Rockney, Y. Zhang, R. Collins, P. Gavrilovich, A. Rowley: 35-Watts CW single-mode Ytterbium Fiber laser at 1.1 µm, CLEO’97, Tech. Digest (1997) paper CPD-30 381 D. Innis, D. J. DiGiovanni, T. A. Stasser, A. Hale, C. Headley, A. J. Stentz, R. Pedrazzani, D. Tipton, S. G. Kosinski, D. L. Brownlow, K. W. Quoi, K. S. Kranz, R. G.Huff, R.Espindola, J. D. LeGrange, G. Jacobovitz-Veselka: Ultrahigh-power single-mode fiber lasers from 1.065 to 1.472 µm using Ybdoped cladding-pumped and cascaded Raman lasers, CLEO’97, Tech. Digest (1997) paper CPD-31 381, 382 S. G. Grubb, T. Strasser, W. Y. Cheung, W. A. Reed, V. Mizrahi, T. Erdogan, P. J. Lemaire, A. M. Vengsarkar, D. J. DiGiovanni, D. W. Peckham, B. H. Rockney: High-power 1.48 µm cascaded Raman laser in germanosilicate fibers, Proc. Top. Meet. Opt. Amplifiers Appl. Davos (1995) paper SaA4, pp. 197–199 381, 382 Annual Report General Phys. Inst. Russ. Acad. Sci. 1993, sect. Optical fiber communication, integrated optics, p. 7 January 1994 E. M. Dianov, D. G. Fursa, A. A. Abramov, M. I. Belovolov, M. M. Bubnov, A. V. Shipulin, A. M. Prokhorov, G. G. Devyatykh, A. N. Gur’yanov, V. F. Khopin: Low-loss high germania-doped fiber: A promising gain medium for 1.3 µm Raman amplifier, Proc. 20th Europ. Conf. Opt. Commun., Firenze (1994) p. 427 381

396

T. T. Basiev et al.

110. S. Grubb, T. Erdogan, V. Mizrahi, T. Strasser, W. Y. Cheung, W. A. Reed, P. J.Lemaire, A. E. Miller, S. G. Kosinski, G. Nykolak, P. C. Becker: 1.3 µm cascaded Raman amplifier in germanosilicate fibers, Proc. Top. Meet. Opt. Amplifiers Appl., Breckenridge (1994) p. 187, paper PD 3-1 381, 382, 384 111. S. V. Chernikov, N. S. Platonov, D. V. Gapontsev, Do III Chang, M. J. Guy, J. R. Taylor: Raman fiber laser operating at 1.24 µm, Electron. Lett. 34, 680– 681 (1998) 382, 384 112. Do III Chang, Hak Kyu Lee, Kyong Hon Kirn: Cascaded Raman fiber laser operating at 1.48 nm, Electron. Lett. 35, 1951–1952 (1999) 382 113. A. R. Chraplyvy, J. Stone: Synchronously pumped D2 gas-in-glass fiber Raman laser operating at 1.56 µm, Opt. Lett. 9, 241–242 (1984) 382 114. V. V. Grigoryants, B. L. Davydov, M. E. Zhabotinski, V. F. Zolin, G. A. lvanov, V. I.Smirnov, Yu. K. Chamorovski: Spectra of stimulated Raman scattering in silica fiber waveguides, Opt. Quantum Electron. 9, 351–352 (1977) 382 115. J. Canning, R. Pasman, M. G. Seats: Photosensitisation of phosphosilicate fiber Bragg gratings, Topical Meeting: Photosesitivity and quadratic nonlinearity in glass waveguides, Fundamentals and applications, Portland (1995), Tech. Dig. Ser. 29, 86–87 (1995) 383 116. E. M. Dianov, A. M. Prokhorov: Raman fiber laser (variants) and fiber Bragg grating, RU patent no. 2095902 (Nov. 10, 1997), US patent no. 5,838,700, (Nov. 17, 1998) 383 117. E. M. Dianov, M. V. Grekov, I. A. Bufetov, S. A. Vasiliev, O. I. Medvedkov, V. G. Plotnichenko, V. V. Koltashev, A. V. Belov, M. M. Bubnov, S. L. Semjonov, A. M. Prokhorov: CW high power 1.24 µm and 1.48 µm Raman lasers based on low loss phosphosilicate fiber, Electron. Lett. 33, 1542–1543 (1997) 383 118. M. M. Bubnov, E. M. Dianov, O. N. Egorova, S. L. Semjonov, A. N. Guryanov, V. F. Khopin, E. M. DeLiso: Fabrication and investigation of single-mode highly phosphorus-doped fibers for Raman lasers, in Advances in Fiber Optics, Proc. SPIE 4083, 12–22 (2000) 383 119. F. L. Galeener, J. C. Mikkelsen, R. H. Geils, W. J. Mosby: The relative Raman cross sections of vitreous SiO2 , GeO2 , B2 O5 and P2 O5 , Appl. Phys. Lett. 32, 34–36 (1978) 383 120. E. M. Dianov, I. A. Bufetov, M. M. Bubnov, A. V. Shubin, S. A. Vasiliev, O. I. Medvedkov, S. L. Semjonov, M. V. Grekov, V. M. Paramonov, A. N. Guryanov, V. F. Khopin, D. Varelas, A. Iocco, D. Costantini, H. G. Limberger, R.-P. Salathe: CW highly efficient 1.24 µm Raman laser based on low-loss phosphosilicate fiber, Tech. Digest Opt. Fibre Conf.’99 (1999) paper PD 25 383 121. E. M. Dianov, I. A. Bufetov, M. M. Bubnov, M. V. Grekov, A. V. Shubin, S. A. Vasiliev, O. I. Medvedkov, S. L. Semjonov, A. N. Gur’yanov, V. F. Khopin: Investigation of CW highly efficient 1.24 µm Raman laser based on low-loss phosphosilicate fiber, Proc. ECOC’99 1, 72–73 (1999) 383, 386 122. V. I. Karpov, E. M. Dianov, V. M. Paramonov, O. I. Medvedkov, M. M. Bubnov S. L. Semyonov, S. A. Vasiliev, V. N. Protopopov, O. N. Egorova: Laser-diode pumped phosphosilicate-fiber Raman laser with an output power of 1 W at 1.48 µm, Opt. Lett. 24, 887–889 (1999) 384 123. E. M. Dianov, I. A. Bufetov, M. M. Bubnov, M. V. Grekov, S. A. Vasiliev, O. I. Medvedkov: Three-cascaded 1407 nm Raman laser based on phosphorusdoped silica fiber, Opt. Lett. 25, 402–404 (2000) 386


397

124. A. S. Kurkov, E. M. Dianov, V. M. Paramonov, O. I. Medvedkov, ˙ Laptev, A. Umnikov, S. A. Vasiliev, N. N. Vechkanov, A. N. Guryanov, A.Yu. E. V. Pershina: 1.43 µm fiber laser for medical applications, in Advances in Fiber Optics, Proc. SPIE 4083, 127–130 (2000) 386 125. E. M. Dianov, A. S. Kurkov, O. I. Medvedkov, V. M. Paramonov, O. N. Egorova, N. Kurukitkoson, S. K. Turitsyn: Raman fiber source for a range of 1.6–1.75 mm, Proc. Optical Fiber Conf. 2003, Atlanta, USA, March 25–27, 2003, paper MF26 387

Index

A1g , 368–370, 379 D2 , 382 F+ 2 , 357 F− 2 , 357 H2 -sensitized fiber, 382 P2 O5 , 383, 384, 386 (NO3 )− , 357 acoustic phonon, 370, 379 acousto-optical Q-switch, 367 alexandrite, 358 AlGaAs/GaAs quasi-CW diode laser, 367 anti-Stokes, 364, 379 ArF excimer laser, 383 Ba(NO3 )2 , 355–357, 359, 360, 368, 370–374, 378, 379 BaMoO4 , 370, 371 BaWO4 , 370–374, 376, 379 beam cleanup, 359, 360 biharmonic pumping, 368 Bragg grating, 381–384, 386 CaCO3 , 355, 356 CaMoO4 , 370 cascaded Raman fiber laser, 382 cavity dumping, 359 CaWO4 , 365, 370, 379 cladding, 384 Coherent Anti-Stokes Raman Scattering (CARS), 368 conversion efficiency, 355 coupled cavities, 359, 361 Cr2+:ZnSe, 376 Cr:YAG saturable absorber, 380 Davydov splitting (DS), 370, 379

dephasing, 368–370, 379 – rate, 368 – time, 352, 368, 371, 372 diamond, 368, 369 diode pumping, 362 divergence, 354 double-clad – fiber lasers (DCFL), 381, 387 double-pass regime, 378 dye lasers, 357 efficiency saturation, 372 electrooptical Q-switching, 367 end-coupling, 381 external pumping, 356, 364, 366, 379 eye-safe – radiation, 360 – Raman laser, 359, 360 – spectral region, 366 – wavelength, 358 fiber deterioration, 380 first Stokes – radiation, 355, 357, 359, 361, 362, 364, 365, 367, 378, 380, 385 – wave, 359 flat–flat resonator, 356 frequency doubling, 364 frequency mixing, 358 fused quartz, 368 GaAlAs quasi-CW diode laser, 380 gain coefficient, 352, 369 germanosilicate fiber, 381, 382, 384 glass fiber, 380 higher Stokes components, 365 homogeneous line broadening, 368, 369 homogeneous width, 368

Index inhomogeneous broadening, 354, 368 integral cross-section, 354, 369, 371, 379 interaction length, 352, 354, 380 intracavity – pumping, 359, 360, 362 – Raman oscillation, 364 – SRS generation, 366 Kerr cell, 351 KGd(WO4 )2 , 355, 357, 371, 376 KGW, 355, 364–366 KY(WO4 )2 , 376 laser – cavity, 351 – damage threshold, 355–357, 362, 366, 380 – DCFL, 387 – line, 351 – Nd:YAG, 358 – Raman – – fiber, 380–383, 386, 387 – – nanosecond, 379 – – picosecond, 361, 367 – ruby, 351, 358 – source, 351 – spectrum, 351 – tunable, 357 lattice phonon, 368, 370 LiF, 357 LiF:F2 , 372 LiF:F− 2 , 366 – color center laser MALSAN-201, 366 – color center tunable lasers, 368 light detection and ranging (LIDAR), 358, 360, 361 LiIO3 , 361, 362 LiNbO3 , 367 liquid SRS media, 353 mode field diameter (MFD), 384 multimode waveguide, 381, 384 multiphonon – decay, 368 – relaxation, 354 NaNO3 , 355, 356, 368 nanosecond pumping, 366

399

nanosecond Raman laser, 379 nanosecond SRS, 366 – amplification spectroscopy, 368 Nd DCFL, 381 Nd3+ :PbWO4 , 380 Nd:GGG, 372 Nd:KGd(WO4 )2 (Nd:KGW), 362, 367 Nd:KGW, 364–367 Nd:KLa(MoO4 )2 , 364 Nd:KY(WO4 )2 (Nd:KYW), 362 Nd:KYW, 364, 365 Nd:NaLa(MoO4 )2 , 364 Nd:YAG laser, 358 nonlinear intracavity dumping, 372 nonlinear refractive index, 353 nonstationary pumping, 362 optical breakdown, 354 optical loss, 383, 384, 386 optimum orientation, 365 P=O bonds, 383 passive Q-switching, 355, 367 passive mode locking, 364 Pb(NO3 )2 , 355, 356 PbMoO4 , 370 PbWO4 , 370, 379, 380 peak cross-section, 369, 370 phase matching, 361, 367 phosphosilicate fibers, 382, 383, 386 picosecond pumping, 362, 378 picosecond Raman gain, 357, 376 picosecond Raman laser, 361, 367 pigtailed LD array, 384 plane-concave (Raman) laser cavity, 367, 379 polarization, 364, 365 population inversion, 354 pulse compression, 365 Q-switching, 351 quantum yield, 354, 357 quasi-CW, 362 Raman amplifier, 364 Raman fiber amplifier, 381, 386 Raman fiber laser, 380–383, 386, 387 Raman frequency shift, 353, 354, 362, 374, 382 Raman gain, 354

400

Index

– coefficient, 362, 386 – picosecond, 357, 376 – transient, 353 Raman laser – eye-safe, 359, 360 – fiber, 380–383, 386, 387 – nanosecond, 379 – picosecond, 361, 367 Raman line broadening, 353, 371, 379 Raman linewidth, 352 Raman scattering (RS), 351 – cross-section, 352 – integral cross-section, 352, 371 – peak cross-section, 353 Raman spectral line, 352 refractive index, 352, 381 relaxation rate, 370 ruby laser, 351, 358 saturable absorber, 367 scheelite, 370 – structure, 370, 379 second harmonic, 357, 367 second harmonic generation (SHG), 355 second Stokes – radiation, 355, 360, 361, 374 – wave, 359 self-conversion, 364, 374 self-focusing, 353 self-frequency conversion, 367 side pumping, 367 silica – fiber, 381, 386 single phonon bridge processes, 370 single-mode fiber core, 381 single-pass pumping, 357 single-pass Raman shifting scheme, 366 single-pass scheme, 355, 357, 366 SiO2 , 383, 386 solid-state laser spectrometer, 358 SrMoO4 , 370, 371 SrWO4 , 370–374, 379 standing-wave resonators, 360

steady-state gain coefficient, 358 steady-state nanosecond oscillation, 357 steady-state regime, 352, 353, 355, 362, 369, 371, 378 steady-state RS peak cross-section, 355 stimulated Raman scattering (SRS), 351 – -active molecular gases, 353 – amplifier, 364, 365, 371 – converters, 368 – gain, 354, 357, 370, 371 – gain coefficient, 356, 357, 366 – lasers, 358 – peak cross-section, 371 – shifters, 358 – threshold, 353, 355–357, 361, 362, 365, 366, 368, 369, 371, 374 Stokes shift, 364–366, 382, 386 Stokes wave, 352, 359 subpicosecond pumping, 371 subpicosecond regime, 365 superradiant emission, 380 symmetrical vibrations, 357 TEM00 , 359 third Stokes, 374, 381, 382, 386 Ti:sapphire, 358 totally symmetric Raman active vibration, 368 transient Raman gain, 353 transient regime, 352, 353, 369 tunable laser, 357 two-photon resonance, 352 upconversion, 370 vibrational excitations, 354, 368 vibronic mode, 374 W–W distance, 370 wavelength division multiplexer (WDM), 381, 382 Yb DCFL, 382, 386

Narrow-Linewidth Tunable Terahertz-Wave Sources Using Nonlinear Optics Kodo Kawase1 , Jun-ichi Shikata2 , and Hiromasa Ito2,3 1

2

3

Kawase Initiative Research Unit, RIKEN 2-1 Hirosawa, Wako 3510198, Japan [email protected] Research Institute of Electrical Communication, Tohoku University 2-1-1 Katahira, Sendai 9808577, Japan Photo Dynamics Research Center, RIKEN 519-1399 Aramaki-Aoba, Sendai 9800845, Japan

Abstract. We review methods of generation of widely tunable terahertz (THz) waves using nonlinear optical effects. Using parametric generation of MgO-doped LiNbO3 crystals pumped by a nanosecond Q-switched Nd:YAG laser, we have built widely tunable coherent THz wave sources with a simple configuration. Fourier transform limited THz wave spectrum narrowing was achieved by introducing the injection seeding method. At the same time, the THz wave output was increased hundreds of times higher than that of a conventional generator which has no injection seeder. Also, tunable THz wave generation was demonstrated by the difference frequency generation between dual signal-wave quasi-phase-matched optical parametric oscillations, using periodically poled LiNbO3 (PPLN) with a series of gratings. An organic ionic salt, 4-dimethylamino-N-methyl-4-stilbazolium-tosylate (DAST) was used as a nonlinear crystal. These room-temperature operated, tabletop systems promise to be new widely tunable THz wave sources suited to a variety of applications.

1

Introduction

In recent years, the generation of terahertz (THz) radiation by optical rectification or photoconductive switching has been extensively studied using femtosecond laser pulses [1,2]. Applied research, such as time domain spectroscopy, makes use of the high time resolution of THz waves and ultrabroad bandwidth up to the THz region. In contrast, our research focuses on the development of tunable THz wave sources with high temporal and spatial coherence using nonlinear optical effects. Specifically, widely tunable coherent sources have a wide range of applications, such as in material science, solid state physics, molecular analysis, atmospheric research, bioscience, chemistry, gas tracing, material testing, food inspection, differential imaging, etc. Novel tunable sources already exist in the sub-THz (several hundreds

Corresponding author

I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 397–424 (2003) c Springer-Verlag Berlin Heidelberg 2003

398

Kodo Kawase et al.

GHz) frequency region, such as the backward-wave-oscillator (BWO). However, a widely tunable THz wave source has long been desired in the frequency region above 1 THz, where the tuning capability of a BWO rapidly decreases. Widely tunable sources covering the region between 1 and 3 THz (with wavelengths from 100 to 300 µm) are limited to free electron lasers or photomixers. The University of California at Santa Barbara plays a central role in exploring the THz region with free electron lasers [3]. However, the number of researchers who can access such apparatus is limited, due to its large scale. In photomixers using low-temperature-grown GaAs, two optical lasers are mixed to generate tunable continuous wave THz waves [4,5]. This should become a reliable source if the output power can be much improved. A p-type germanium (p-Ge) laser is another widely tunable source in this region, but it is primarily for use in pure science since liquid He is required [6,7]. Although the aforementioned widely tunable sources can be used at the laboratory level, they do not satisfy all of the needs of researchers interested in practical applications. Therefore, there is significant potential for an explosion of applied research on the THz band if simple, convenient tunable sources are made available. Much research effort was carried out a couple of decades ago concerning the generation of tunable coherent far-infrared radiation based on optical technology [8,9]. Among them, THz oscillation and amplification were expected by Nishizawa of our institute to be made available by utilizing the resonant frequency of the crystal lattice or of the molecule itself in 1963 [10,11]. Efficient and widely tunable THz generation had been reported in the pioneering works of Pantell et al. in the late 1960s to early 1970s [12,13,14]. This is based upon tunable light scattering from the long-wavelength side of the A1 -symmetry softest mode in LiNbO3 . The input (pump) photon at near infrared stimulates a near infrared Stokes photon (idler) at the difference frequency between the pump photon and the vibrational mode. At the same time, the THz wave (signal) is generated by the parametric process due to the nonlinearity arising from both electronic and vibrational contributions of the material. The tuning is accomplished by controlling the propagation direction. Although the interaction is highly efficient, it should be noted that most of the generated THz waves are absorbed or totally reflected inside the crystal due to the large absorption coefficient, as well as the large refractive index in the THz range. To allow THz radiation, a cut-exit had been made in the corner of the crystal. It is to our surprise that no research has been reported on this novel method since 1976 due to the alternative and successful use of submillimeter moleculer gas lasers. In the last seven years, we have developed an efficient and widely tunable source of coherent THz waves based on the principle of the previous works, but far better characteristics by introducing a monolithic grating coupler, an arrayed Si-prism coupler, doped crystals, and injection seeding.

Narrow-Linewidth Tunable THz Source

399

2 Theory of THz-Wave Parametric Generation Using Polaritons The generation of coherent tunable THz waves results from the efficient parametric scattering of laser light via a polariton. A polariton is a quantum of the coupled phonon–photon transverse wave field, and stimulated polariton scattering occurs when the pump excitation is sufficiently strong in polar crystals, such as LiNbO3 , LiTaO3 , and GaP, which are both infrared-active and Raman-active. The scattering process involves both second- and thirdorder nonlinear processes. Thus, strong interaction occurs among the pump, the idler, and the polariton (THz) waves. LiNbO3 is one of the most suitable materials for generating THz waves efficiently because of its large nonlinear coefficient (d33 = 25.2 pm/ V, λ = 1.064 µm [13]) and its transparency over a wide wavelength range (0.4–5.5 µm). LiNbO3 has infrared- and Ramanactive transverse optical (TO) phonon modes, called A1 , and the lowest mode (ω0 ≈ 250 cm−1 ) is useful for efficient tunable far-infrared generation because it has the largest parametric gain, as well as the smallest absorption coefficient [14,15]. The principle of tunable THz wave generation is as follows. Polaritons exhibit phonon-like behavior in the resonant frequency region (near the TOphonon frequency ωTO ). However, they behave like photons in the nonresonant low-frequency region (Fig. 1), where a signal photon at THz frequency (ωT ) and a near-infrared idler photon (ωi ) are created parametrically from a near-infrared pump photon (ωp ), according to the energy conservation law ωp = ωT + ωi (p: pump, T: THz, i: idler).1 In the stimulated scattering process, the momentum conservation law kp = ki + kT (noncollinear phase matching condition; see the insets of Fig. 1) also holds. This leads to the angle-dispersive characteristics of the idler and THz waves. Thus, a coherent THz wave is generated efficiently by using an optical resonator for the idler wave, and continuous and wide tunability is accomplished simply by changing the angle between the incident pump beam and the resonator axis. In stimulated polariton scattering, four fields mutually interact: the pump Ep , idler Ei , THz wave ET , and ionic vibration Q0 (lowest A1 mode). The parametric gain coefficients for the idler and THz waves are obtained by solving the classical coupled-wave equations that describe this phenomenon. Assuming a steady state and no pump depletion, the coupled-wave equations 1

Unlike the usual convention for OPOs in the optical domain (see chapters by Vodopyanov and Ebrahimzadeh in this volume), where the “idler” has longest wavelength, in our case the longest wavelength corresponds to the terahertz wave (ωT ) and the “idler” wave (ωi ) corresponds to a near-infrared wave

400

Kodo Kawase et al. Fig. 1. Dispersion relation of the polariton, an elementary excitation generated by the combination of a photon and a transverse optical phonon (ωTO ). The polariton in the low-energy region behaves like a photon at THz frequency. Due to the phase-matching condition as well as the energy conservation law which hold in the stimulated parametric process, the tunable THz wave is obtained by control of the wavevector k T . The inset shows the noncollinear phase-matching condition

in CGS units are written as [14,16,17], ω2 ω2 2 , ET = − 2 ∇ + 2 c εT c χP Ep Ei∗ ω2 ω2 2 ∗ ∇2 + 2i εi + χR |Ep | , Ei = − 2i χP Ep ET c c

(1)

where ωi (= ωp − ω) and ω denote the frequencies of the idler and THz wave, respectively, εβ (β = T, i) is the permittivity in the material (LiNbO3 ), and c is the speed of light in vacuum. The nonlinear susceptibilities χP and χR denote parametric and Raman processes, respectively, and are expressed as S0 ω02 dQ , ω02 − ω 2 S0 ω02 χR = 2 d2 , ω0 − ω 2 − iωΓ0 Q χP = dE +

(2) (3)

where ω0 , S0 , and Γ0 are the eigenfrequency, oscillator strength, and damping coefficient (or linewidth) of the lowest A1 -symmetry phonon mode, respectively. The coefficients dE (= 16πd33 ) and dQ denote the second- and thirdorder nonlinear processes, which originate from electronic and ionic polarization, respectively. According to the rate equation analysis, the expression for dQ in CGS units is given by dQ =

8πc4 np (S33 /L∆ Ω)0 S0 ω0 ωi4 ni (¯ n0 + 1)

1/2 ,

(4)

where nβ (β = p, i) is the refractive index and n ¯ 0 = [exp(ω0 /kT ) − 1]−1 (: Planck constant, k: Boltzman constant, T : temperature) is the Bose distribution function. The quantity (S33 /L∆ Ω)0 denotes the spontaneous Raman (Stokes) scattering efficiency of the lowest A1 -symmetry phonon mode, where S33 is the fraction of incident power that is scattered into a solid angle


401

∆ Ω near a normal to the optical path length L, and is proportional to the scattering cross-section. The coupled wave equations (1) can be solved using the plane wave approach, and analytical expressions of the exponential gain for the power of the THz wave and idler are  

2 1/2  g0 αT  gT = gi cos φ = (5) 1 + 16 cos φ −1 ,  2  αT where φ is the phase-matching angle between the pump and the THz wave; g0 is the low-loss limit; and αT is the absorption coefficient in the THz region. In CGS units, they are written as: 1/2 πωωi Ip g0 = χP , (6) 2c3 nT ni np 1/2 2ω S0 ω02 αT = 2|ImkT | = Im ε∞ + 2 . (7) c ω0 − ω 2 − iωΓ0 The low-loss parametric gain g0 has the same form as the parametric gain in the optical region [18], but the nonlinear susceptibility χP , which involves both second- and third-order processes (2), is almost entirely determined by the third-order (ionic) dQ -term (more than 80 % contribution). Figure 2 shows the calculated parametric gain gT for LiNbO3 at typical pump intensities. A gain in the order of several cm−1 is feasible in the frequency domain up to 3 THz at room temperature. The gain is remarkably enhanced by cooling the crystal to liquid nitrogen temperature (T = 77 K). Actually, THz power enhancement by two orders of magnitude was observed [19]. In the low-frequency region far from resonance, the absorption coefficient αT (7) is nearly in proportion to the linewidth (damping coefficient) Γ0 .2 The linewidth Γ0 is reduced at low temperature (Γ0T =77 K ∼ = Γ0T =300 K /3[20]), and therefore, reduction by an order of magnitude in the absorption loss of THz waves can be achieved for the propagation distance of 1 mm (αTT=77 K ∼ = 20 cm−1 at 1.5 THz [19]). Also, the parametric gain gT is enhanced, because it is a monotonically decreasing function of the absorption coefficient αT (5). Physically, when the polariton damping caused by random thermal activation is reduced, the polariton has a longer lifetime, and therefore, coherent parametric interaction efficiently occurs. 2

∼ ω02 and the numerical values of ε∞ = 4.6, S0 = 16.0, Γ0 = Considering ω02 − ω = 29.0 ([14] and references therein), the first term (ε∞ ) in the square bracket of (7) can be neglected. Thus the Taylor expansion of (7) based on ω02 ωΓ0 is proportional to Γ0

402

Kodo Kawase et al.

Fig. 2. Calculated gain coefficient for parametric THz wave generation using LiNbO3 crystal pumped at 1.064 µm. The gain is enhanced by cooling the crystal due to the reduced absorption loss at THz frequency

3 Injection-Seeded THz-Wave Parametric Generator (is-TPG) Tunable THz-wave generation, using nonlinear optical methods, had been widely reported where difference-frequency mixing between two laser sources was utilized [8,9], though the conversion efficiency observed was poor. In contrast, higher conversion efficiency was obtained by simultaneous Raman and parametric oscillation, utilizing the polariton mode scattering of LiNbO3 based on the 248 cm−1 A1 -symmetry soft mode, as described in Sect. 2. This method used a single fixed optical source, and was performed at room temperature. The idler (Stokes) and signal (THz) waves were generated from the pump (near-infrared) wave in the direction consistent with the noncollinear phase matching condition inside the LiNbO3 crystal. The idler wavelength was longer by a few nanometers than the pump wavelength. We have researched a THz wave parametric oscillator (TPO, [23,24]) and THz wave parametric generator (TPG, [16]) using LiNbO3 or MgO:LiNbO3 crystals. The TPO (Fig. 3) is continuously tunable in the 1–3 THz range in one operation by slightly changing the incident angle of the pump beam as shown in Fig. 4, and can emit peak powers of up to several tens of a milliwatt. The difference between a TPO and a TPG is that the former has an idler cavity as shown in Fig. 3 while the latter does not. The THz wave linewidth of a conventional TPG exceeds 500 GHz and the THz wave output is much smaller than that from a TPO. Therefore, we previously concentrated our efforts on the development of a TPO system, although its linewidth was tens of GHz. This section is primarily concerned with new work on TPGs which appear to perform better than TPOs. The TPG spectrum was narrowed to the


403

Fig. 3. Experimental cavity arrangement of the THz wave parametric oscillator (TPO) using a Si-prism coupler

Fig. 4. The tuning characteristic between the incident angle of the pump to the x-surface of the crystal normal, generated THz wavelength. Solid curve indicates the calculated tuning curve

Fourier transform limit of the pulse width by introducing injection seeding to the idler [25,26]. The purity of the THz wave frequency was dramatically improved to ∆ ν/ν < 10−4 . Simultaneously, the output obtained was several hundred times higher than that of a conventional TPG. In addition, wide tunability and fine resolution were demonstrated using a tunable seeder. Even in the optical region, injection seeding to a nanosecond (ns) optical parametric generator (OPG) has not been reported until recently [27], due to the limit of parametric gain. 3.1

Experimental Setup

Figure 5 shows the setup of our experimental injection-seeded TPG (is-TPG). Arrangements were tested using one, two, and three nonlinear crystals (65 mm in length, 5 mol% MgO doped LiNbO3 ). The maximum THz wave output was obtained when two crystals were used in series. The TPG efficiency of

404

Kodo Kawase et al.

Fig. 5. The setup used for our experimental injection-seeded THz wave parametric generator (is-TPG). The pump was a single longitudinal mode Q-switched Nd:YAG laser (1.064 µm), and the seed for the idler was a continuous-wave Yb-fiber laser (1.070 µm) or tunable laser diode (1.066–1074 µm)

a MgO:LiNbO3 is several times higher [16] than that of an undoped LiNbO3 . Both crystals were cut into 65×6×5-mm (x×y×z-axis) pieces. The x-surfaces were polished, and antireflection coated. An array of seven Si-prism couplers was placed on the y-surface of the secondary MgO:LiNbO3 crystal for efficient coupling of the THz wave as shown in Fig. 5. The pump used was a single longitudinal mode (SLM) Q-switched Nd:YAG laser (Spectron SL404T, wavelength: 1.064 µm, energy:< 50 mJ/ pulse, pulse width: 15 ns, beam profile: TEM00 ). The pump beam diameter was decreased to 0.8 mm using a telescope in order to increase the power density. The pump power density was < 530 MW/ cm2 at the crystal surface and could be varied with an attenuator. The pump beam was almost normal to the crystal surfaces as it entered the crystals and passed through the crystal close to the y-surface. A continuous wave SLM Yb-fiber laser (wavelength: 1.070 µm fixed, power: < 300 mW) or tunable diode laser (wavelength: 1.066–1.074 µm, power: 50 mW) was used as an injection seeder for the idler. Observation of the intense idler beam easily confirmed injection-seeded THz wave generation. The polarizations of the pump, seed, idler, and THz waves were all parallel to the z-axis of the crystals. The THz wave output and temporal waveform were measured with a 4 K Si-bolometer and a Schottky barrier diode (SBD) detector [28], respectively. 3.2

Power Enhancement

Energy enhancement of the THz and idler waves by injection seeding is shown in Fig. 7a,b, respectively. The THz and idler outputs are roughly proportional to each other. Comparison of the output from 0 and 200-mW seeding enabled us to determine that the THz wave and idler energy increased by factors of nearly 300 and 500, respectively. The maximum conversion efficiency was achieved when the pump and seed beams almost fully overlapped at the incident surface of the first MgO:LiNbO3 crystal, as shown in Fig. 5. This was confirmed by the fact that initial excitation is an essential feature of injection seeding. The maximum THz wave output of 900 pJ/ pulse (peak > 100 mW) was obtained with a pump of 45 mJ/ pulse and a seed of 250 mW. In our previous studies, the maximum THz-wave output from


405

the conventional TPGs and TPOs was 3 and 190 pJ/ pulse, respectively [29]. The Si-bolometer became saturated at about 5 pJ/ pulse, so we used several thick calibrated papers as an attenuator. As the minimum sensitivity of the Si-bolometer was almost 1 fJ/ pulse, the dynamic-range of the is-TPG system was 900 pJ to 1 fJ ≈ 60 dB, which is sufficient for most applications. The dynamic range can be significantly increased using a lock-in amplifier. The outputs began to saturate with a seed power of almost 100 mW. A relatively high seed power was required in this experiment because the seed energy did not fully contribute to the idler generation. The idler beam runs parallel to the seed beam for 3 mm distance. This is because the pump and seed beams were spatially separated inside the secondary MgO:LiNbO3 crystal (Fig. 5), and because most of the idler energy was generated inside the secondary MgO:LiNbO3 crystal. On the other hand, when one crystal was used, 10 mW of seed power was enough to obtain idler saturation because the pump and seed beams were not separated, though the THz-output was very low. Therefore, it is important to somehow confine the pump and seed beams in a long interaction volume, in order to decrease the required seed power and to increase the efficiency. The idler beam pattern was expanded in the z-axis direction probably due to the photorefractive effect inside the crystal. The angle between the idler beam and the crystal x-surface normal was almost 1.5◦ , proving that the cavity effect of the crystal surfaces has no relation to this parametric generation. Figure 6 shows examples of temporal waveforms of the pump, idler, and THz wave using a pump energy of 45 mJ and a seed power of 250 mW. The pulse widths of the pump and idler are 15 and 4 ns, respectively. The observed pump depletion (28 %) was the largest depletion encountered during our TPG/TPO research. The THz waveform was also found to be depleted, probably due to back conversion of the pump. The second peak of the pumpwaveform is due to back conversion, and the product of Ep times Ei resulted in the second peak of the THz waveform. Depletion of the THz waveform was not observed with pump energies below 35 mJ/ pulse. The THz waveform began to deplete as the pump energy increased, although the THz wave energy continued to increase, as shown in Fig. 7, due to the pulse width expansion. Figure 8 shows the THz wave beam pattern in the horizontal (upper) and vertical (lower) directions, respectively, at a distance of ≈ 40 cm from the Si-prism array. The beam pattern was nearly Gaussian and had a diameter of 7 mm, which is suitable for many applications. The original vertical divergence was about 6◦ , as determined from the pump beam diameter and the wavelength according to diffraction theory. A cylindrical lens (f = 30 mm) made of polymethylpentene (PMP or ‘TPX’) was used, as shown in Fig. 5, to collimate the THz wave divergence in the vertical (z-axis) direction. As for the horizontal direction, the beam diameter decreased as it propagated, due to the phased array like effect of the Si-prism array [29]. Furthermore, the

406

Kodo Kawase et al.

Fig. 6. The temporal waveforms of the pump (1.064 µm), idler (1.07 µm), and THz wave (190 µm). The pump energy was 45 mJ and the seed power was 250 mW. This is the largest pump depletion (28.4 %) observed during our research on TPGs and TPOs

THz beam can be tightly focused into a ≈ 0.5 mm spot using a short focus TPX or Si lens. 3.3

Spectrum Narrowing

Figure 9 shows the effect of injection seeding on idler spectrum narrowing. The dotted line indicates the idler spectrum of a conventional TPG without injection seeding, and the solid line indicates the idler spectrum of an is-TPG. The resolution limit of the spectrum analyzer used was 0.2 nm, so the real idler spectrum was much narrower than shown in this figure. Using a solid etalon, the idler spectrum was assured to be less than 1 GHz. The THz wavelength and linewidth were measured using a scanning Fabry–Pérot etalon consisting of two Ni metal meshes with a 65 µm grid. Figure 10 shows the transmitted THz wave power as a function of etalon spacings of (a) ≈ 80 mm and (b) ≈ 210 mm. Figure 10a demonstrates the stability of the spectrum and output during the 20-min scan. The displacement between the two periods (190 µm) directly corresponds to the wavelength. The merit of an is-TPG lies in its output stability due to the mode-hop-free characteristic, since it has no cavity. On the other hand, as with an injectionseeded TPO [30], the cavity-length must be actively controlled to match the seed wavelength in order to stabilize the output. In Fig. 10b, the free spectral range (FSR) of the etalon is 750 MHz and the THz wave linewidth was measured to be less than 200 MHz (0.0067 cm−1 ), which is our measurement resolution limit. Since the etalon spacing was up to 210 mm, the nanosecond THz wave pulse made less than three round trips in the etalon cavity; thus the resolution is inevitably limited. The Fourier transform limit of the spectral width was calculated from the pulse shape of the THz wave as measured by SBD. The typical pulse width


407

Fig. 7. The input–output characteristics of a is-TPG, showing energy enhancement of the (a) THz wave (190 µm) and (b) idler (1.070 µm) by a factor of several hundreds with injection seeding

Fig. 8. The beam pattern of the THz wave in the horizontal (upper ) and vertical (lower ) directions, at a distance of ≈ 40 cm from the Siprism array

408

Kodo Kawase et al.

Fig. 9. Narrowing of the idler (1.07 µm) spectrum by injection seeding. The dotted and solid lines indicate the idler spectrum of a conventional TPG and an is-TPG, respectively. The resolution limit of the spectrum analyzer used was 0.2 nm, so the real idler linewidth was much narrower than in this figure

of the THz wave was 3.4 ns, as shown in Fig. 11a, and was almost identical to that of the idler, which was measured with a high-speed photodetector. Figure 11b shows the power spectrum of the THz wave calculated from the upper graph, and indicates that the linewidth was 136 MHz. In this calculation, we ignored any fluctuations in the background noise near the zero level in Fig. 11a. Figures 10 and 11 confirmed that the linewidth of the THz wave was narrowed to near the Fourier transform limit. Even with an is-TPG the THz wave linewidth was still more than 10 GHz when a multifrequency Nd:YAG laser was used as the pump. Thus, both the pump and seed must be SLM lasers to obtain the transform-limited THz wave with a TPG. 3.4

Wide Tunability

It was possible to tune the THz wavelength using an external cavity laser diode as a tunable seeder. A wide tunability from 125 to 430 µm (frequency: 0.7 to 2.4 THz, wave number: 23 to 80 cm−1 ) was observed as shown in Fig. 12 by changing both the seed wavelength and the seed incident angle. Squares and circles indicate the tunability of the THz and idler waves, respectively. Both crystals were MgO:LiNbO3 in this experiment. The wavelength of 430 µm (0.7 THz) was the longest ever observed during our study of TPGs and TPOs. In the longer wavelength region, the angle between the pump and idler becomes less than 1◦ as shown in Fig. 4, thus it is difficult for the TPO to oscillate only the idler inside the cavity without scattering the pump. In the shorter wavelength region, the THz output is comparatively smaller than the idler output, due to the larger absorption loss inside the crystal.


409

Fig. 10. The THz-linewidth and wavelength measured with a scanning Fabry–Pérot etalon consisting of two metal-mesh plates. (a) The stability of the spectrum is demonstrated and the displacement between the two periods (190 µm) corresponds directly to the wavelength. (b) The FSR of the etalon is 750 MHz and the linewidth of the THz wave is measured to be less than 200 MHz (0.0067 cm−1 ), which is our measurement resolution limit

The absorption spectrum of low-pressure (< 1 Torr) water vapor was measured to demonstrate the continuous tunability and the high resolution of the is-TPG. The absorption gas cell used was an 87-cm-long stainless light pipe with TPX windows at both ends. Figure 13 shows an example of measurements at around 1.92 THz, where two neighboring lines exist. Resolution of less than 100 MHz (0.003 cm−1 ) was clearly shown. In fact, it is not easy for FTIR spectrometers in the THz wave region to demonstrate a resolution better than 0.003 cm−1 because of the instability of the scanning mirror for more than a meter. The system is capable of continuous tuning at high spectral resolution in 4 GHz segments anywhere in the 0.7 to 2.4-THz region. The range of continuous tuning is currently restricted by the mode hop of the tunable laser diode. Since there is no cavity to be slaved, continuous tun-

410

Kodo Kawase et al.

Fig. 11. (a) Temporal THz wave output measured by the SBD, and (b) the calculated Fourier transform limit of the spectral width from the measured temporal THz waveform. The typical pulse width of the THz wave was 3.4 ns, as shown in the upper figure, and the calculated linewidth was 136 MHz, as shown in the lower figure

ing is extendible, in principle, to the full tunability of the is-TPG by using a mode-hop-free seeder, such as a Littman-type external cavity diode laser. Figure 14 shows the change in THz wave output as a function of the seed incident angle. In this experiment, the seed wavelength (1.07 µm) and THz wavelength (190 µm) were fixed, and the calculated noncollinear phase matched angle was 1.43◦ . Here, it is important that injection seeding was not overly sensitive to the seed incident angle. In addition, the linewidth was assured to be less than 200 MHz at any deviated incident angle. From this, we see that continuous tuning is possible, to some extent, by simply varying the seed wavelength without having to adjust the incident angle. In practice, the tuning in Fig. 13 was produced without changing the seed incident angle. Tuning without mechanical movement will lead to a stable and compact spectroscopic system. Even when the incident angle must be varied for wide tuning such as in Fig. 12, there is no requirement to precisely control the angle due to this tolerance. On the other hand, as with the injection seeded TPO [30], the incident angle must be precisely controlled so that it is always perpendicular to the cavity mirror.


411

Fig. 12. The wide tunability of an injection-seeded TPG. Squares and circles indicate the tunability of the THz and idler waves, respectively

Fig. 13. An example of the absorption spectrum measurement of low-pressure (< 1 Torr) water vapor at around 1.919 THz. Resolution of less than 100 MHz (0.003 cm−1 ) was clearly shown

4 Arrayed Silicon Prism Coupler for a THz-Wave Parametric Oscillator In generating a THz wave by parametric process in a LiNbO3 crystal, some way of extracting the THz wave from the crystal is necessary, since the refractive index of the LiNbO3 crystal for the THz wave is large enough to cause total internal reflection. We introduced a Si-prism coupler (n ≈ 3.4) to extract the THz wave generated inside a nonlinear crystal, thereby substantially improving the exit characteristics [31]. This section describes the characteristics of the oscillation, and a novel coupling method for THz waves

412

Kodo Kawase et al.

Fig. 14. Variation in THz-wave output as a function of the seed incident angle. The seed wavelength (1.070 µm) and generated THz wavelength (190 µm) were confirmed constant. The angle of incidence shows significant tolerance

using an arrayed Si prism [29]. Using the arrayed prism coupler, there is a 6-fold increase in coupling efficiency and a 40 % decrease in the far-field beam diameter, compared with using a single-prism coupler. We also discuss the negative effect of the free carriers at the Si-prism surface excited by the scattered pump beam, and the positive effect of cavity rotation on the unidirectivity of THz-wave radiation from the Si prism. 4.1

Experimental Setup

The basic configuration of the source consisted of a Q-switched Nd:YAG laser (LOTIS LS-2136, 1.064 µm) and a TPO, as shown in Fig. 15. The LiNbO3 crystal used in the experiment was cut from a wafer 65 × 6 × 5 mm (x × y × z). The x-surfaces were mirror polished and antireflection coated. The y-surface was also mirror polished, in order to minimize the coupling gap between the prism base and the crystal surface, and to prevent scattering of the pump beam, which excites a free carrier at the Si-prism base. The pump wave passed through the crystal close to the y-surface, to minimize the travel distance of the THz wave inside the crystal. The idler wave was amplified in an oscillator consisting of flat mirrors with a half-area HR coating. In the experiment, the beam diameter, pulse width, and repetition of the pump wave were 1.5 mm, 25 ns, and 50 Hz, respectively, and the typical pump energy was 30 mJ/pulse. The mirrors and crystal were installed on a precise, computer-controlled rotating stage for precise tuning. An array of seven Si prism couplers was placed on the y-surface of the LiNbO3 , as shown in Fig. 15. The right angle prisms were fabricated from high resistivity Si (ρ > 1 k Ω · cm, α ∼ = 0.6 cm−1 ); each was cut from a bulk Si crystal, using a precise diamond cutter, to dimensions of 8.0 (base) × 6.1 (face) × 5.1 (side) × 5.0 (thickness) mm, and the angles were 50 – 40 – 90◦ . The total base length was 8 mm × 7 = 56 mm. A prism opening angle of ξ = 40◦ was chosen so that the THz wave would emerge almost normal to the prism face (6.1 × 5.0 mm2 ). The base of the prisms was


413

Fig. 15. A THz wave parametric oscillator (TPO) with a Si-prism array. The Siprism array was introduced to increase the output, and to reduce the diffraction angle of the THz wave by increasing the coupling area. An array of seven rightangle Si-prism couplers was placed on the y-surface of the LiNbO3 . The total base length was 8 mm × 7 = 56 mm

Fig. 16. Typical input–output characteristics of a THz wave parametric oscillator (TPO) with a Si prism array. By increasing the prism base area, THz wave output of more than 6 times that of a single-prism coupling was obtained

pressed by a specially designed holder against the y-surface of the LiNbO3 crystal to maximize coupling efficiency. 4.2

Experimental Results

Typical input–output characteristics of a TPO with the Si-prism array are shown in Fig. 16, in which the oscillation threshold was 18 mJ/ pulse. With a pump power of 34 mJ/ pulse, the THz wave output from the prism array was 192 pJ/ pulse (∼ = 19.2 mW at the peak). In the case of single-prism coupling, the typical output was about 30 pJ/ pulse (3 mW at the peak) at best under the same conditions. In comparison, the prism array was capable of emitting more than 6 times as much THz wave energy as the single-prism coupling due to the 7 times wider base area.

414

Kodo Kawase et al.

A small portion of the pump beam was reflected from the end mirror or scattered at the crystal edge and shone on the emitting face of the Si prism, generating free carriers that strongly absorbed the THz wave. A HR mirror was therefore installed in front of the last Si prism (rightmost in Fig. 15) to intercept the reflected pump beam, as illustrated in Fig. 15. The HR mirror needed to be close to the y-surface of the LiNbO3 crystal to intercept the pump beam, as only 10 µJ/ cm2 of pump is enough to generate the free carrier in Si [32]. Without the HR mirror, the THz wave output from the last Si prism decreased to 10−2 –10−3 of the output with the HR mirror in place. Because of the difficulty in perfectly shielding the scattered pump, it had been difficult to obtain maximum THz wave output (≈ 3 mW) using a single-prism coupling, even with the HR mirror. On the other hand, it is much easier to extract maximum THz wave output (≈ 20 mW) from the arrayed prism, because the last prism acts as a perfect auxiliary shield for the scattered pump. The spatial intensity distribution of the THz wave radiation was measured by transversely shifting a Si-bolometer with a 1.4 mm wide incident slit. The beam pattern and diffraction of the THz wave in the z-plane are shown in Figs. 17 and 18 for the single- and arrayed-prism couplers respectively. In both figures, d indicates the distance between the prism coupler and the slit. In both measurements, the THz wavelength was 170 µm and the angle between the THz beam and the y-surface of the crystal was 50◦ . With singleprism coupling, both the near- and far-field patterns were Gaussian-like, as shown in Fig. 17, and the diffraction angle in the far field was measured to be 1.4◦ . With arrayed-prism coupling, the far-field pattern was almost Gaussianlike, whereas the near-field pattern was asymmetric, as shown in Fig. 18. In the near field, higher output was observed from the prisms closest to the pump exit surface (right in Figs. 15 and 18) because as the distance between the pump and the y-surface of the crystal shortens, the absorption loss of the THz wave decreases. The output of each prism was distinguishable at d < 10 cm, whereas the beam pattern became continuous at d > 20 cm. The diffraction angle of the THz wave emitted from the prism array coupler was apparently smaller than that of single-prism coupling, comparing the FWHM of the beam pattern at distance d = 100 cm, where the FWHM is 58 mm for the single-prism coupling and 34 mm for the arrayed-prism coupling. The far-field diffraction angle is defined by the emitting aperture width and the wavelength. The smaller diffraction angle was obtained by the prism array due to the 7 times wider emitting aperture than the single prism. In previous experiments [14], a cut-exit was used to avoid total internal reflection, as illustrated in the inset of Fig. 19. A cut exit was made at the corner of the LiNbO3 crystal, so that the THz wave emerged approximately normal to the exit surface. In this case, the refractive index dispersion of LiNbO3 and the change of the phase-matching angle, δ, directly influenced the THz wave direction change, ∆ θc . On the other hand, when a Si-prism


415

Fig. 17. The intensity cross-section of the THz wave in the horizontal direction at distance d from a single Si-prism coupler. The measured diffraction angle of the THz wave in the far field was 1.4◦

Fig. 18. The intensity cross-section of the THz wave in the horizontal direction at distance d from the Siprism array. Comparing the FWHM at d = 100 cm, the diameter of the THz wave was 40 % smaller than that of a single-prism coupler

coupler is used, radiation is almost in one direction and variation in the phase matching angle is substantially reduced. Using the phase-matching conditions, the output angle of the THz wave θp yields [33]

ni ni λT c + c sin θp = nT sin arcsin c 2nT 2nT (λT − λp ) 2 np − n2i λ2T n2T λp − c + −ξ , (8) 2nT ni (λT − λp ) 2ncT ni λp (λT − λp ) where λp and λT are the pump and THz wavelength, np , ni , and nT are the refractive indices of the pump, idler, and THz waves in LiNbO3 , ncT is the

416

Kodo Kawase et al.

Fig. 19. Calculated radiation angle changes for two different THz coupling methods: cut exit and Si-prism coupler. The dotted and broken lines indicate the radiation angle changes for the cut exit, ∆ θc , and Si prism coupler, ∆ θp , respectively. The solid line shows the actual change in THz-beam direction when observed from outside the TPO. Since the TPO cavity can be angle-tuned by rotating the stage in the direction counter to ∆ θp , the actual angle change becomes much smaller than ∆ θp . The changes in the radiation angle are set at zero for λTHz = 200 µm, for comparison

refractive index of the THz wave in silicon, and ξ denotes the opening angle of the Si-prism, respectively. Figure 19 shows the calculated changes in radiation angle for these two methods of coupling. The changes in the radiation angle are set at zero for λTHz = 200 µm, for comparison. The dotted and broken lines indicate the changes in the radiation angle for the cut exit, ∆ θc , and Si-prism coupler, ∆ θp , respectively. The solid line shows the change in THzbeam direction when observed from outside the TPO. It is important to note that since the TPO cavity can be angle-tuned by rotating the stage in the direction counter to ∆ θp , the actual angle change becomes much less than ∆ θp . For the tuning range of 100–420 µm, ∆ θc = 16.5◦ , ∆ θp = 4.0◦ and the actual change, ∆ θp - (cavity rotation), is 1.5◦ .

5 Tunable THz-Wave Generation from DAST Crystal Using Dual Signal-Wave Parametric Oscillation of PPLN In this section, tunable THz wave generation by difference frequency generation (DFG) of dual signal-wave quasi-phase-matched optical parametric oscillation (DSW-QPM-OPO) was performed, using periodically poled LiNbO3


417

(PPLN) with a series of gratings [34]. An organic ionic salt, 4-dimethylaminoN-methyl-4-stilbazolium-tosylate (DAST), was used as a nonlinear crystal. A compact THz wave source resulted, and the wavelength was varied between 120 and 160 µm by changing the PPLN temperature. The wavelength can be tuned from 100 to 700 µm by proper selection of periodically poled grating combinations. 5.1

Experimental Setup

A DSW-QPM-OPO was applied to THz wave DFG using PPLN with a series of gratings. This type of PPLN has two periods on the pump-path, and therefore has two energy conservation and phase matching conditions such that the two signal waves are collinearly generated at one pump wave [35]. This makes DSW-QPM-OPO promising as a compact DFG THz wave source. For the DFG mixer, the organic ionic salt DAST was used. We have developed a stable, reproducible method of growing large DAST crystals, using the fixed seed crystal method [36]. DAST, which was invented and patented by Nakanishi [37], has quite large nonlinear and electro-optic coefficients. In addition, DAST is suitable for high-speed modulation and detection, as well as THz wave radiation, because of its low dielectric constant [36,38,39,40]. Efficient THz short pulse radiation from DAST was reported using optical rectification with a subpicosecond laser pulse as a source [41,42]. Also, tunable THz wave generation was demonstrated using an electronically tuned Ti:Al2 O3 laser as the optical source [43]. We have also conducted studies on the use of DAST as a millimeter-wave receiver [36]. The experimental setup is illustrated in Fig. 20. The PPLN was fabricated by means of the electric-field poling process. It consisted of two grating periods (29.3 and 29.5 µm), each having an interaction length of 20 mm, giving a total length of 40 mm. The crystal was placed in an oven, both to avoid the photorefractive effects and to tune the oscillating wavelengths. The OPO cavity consisted of 50-mm radius concave mirrors, separated by 50 mm. The reflectivities of the input and output mirrors at the signal wavelengths were 99 % and 80 %, respectively. The two idler waves were absorbed in the mirror substrate. The pump source was a Q-switched Nd:YAG laser (wavelength: 1.064 µm, pulse width: 120 ns, repetition: 1–5 kHz, energy: 1 mJ/ pulse at 1 kHz). The pump beam was focused to 150 µm (1/e2 radius) with a lens (f = 200 mm) at the center of the PPLN secondary grating. The output for each signal was typically 100 µJ/ pulse, and the polarizations were parallel to each other. The two signal beams were focused onto a ≈ 0.5-mm spot on the DAST surface (total crystal size: 12 × 11 × 1 mm3 ) using a lens with f = 100 mm. The THz wave generated from the DAST was focused into a 4 K Si-bolometer using two parabolic metal mirrors (f = 120 mm). The transmitted signal beams were eliminated by a small mirror.

418

Kodo Kawase et al.

Fig. 20. Experimental setup for tunable THz wave generation from a nonlinear crystal of 4-dimethylamino-N-methyl-4-stilbazolium-tosylate (DAST) by DFG between near-IR inputs. The input waves were generated by dual signal-wave quasiphase-matched optical parametric oscillation (DSW-QPM-OPO) using periodically poled LiNbO3 (PPLN) with a series of gratings

5.2

Experimental Results

Figure 21 shows the examples of wavelength measurements, (a) oscillated signal-waves at a temperature 150◦ , and (b) generated THz wave. The THz wavelength was measured using a scanning Fabry–Pérot etalon consisting of two metal-mesh plates (65-µm grid). The displacement between the two periods was 139 µm, which corresponds directly to the difference frequency between the two incident signal wavelengths of 1.529 and 1.546 µm. The signal wavelength interval was tuned from 15 nm to 20 nm by varying the temperature from 100 to 200◦ , which resulted in tuning the THz wavelength from 120 to 160 µm. Through proper selection of grating periods, it was possible to tune the signal wave interval from 2 nm to 22 nm corresponding to THz wavelength tuning from 100 to 700 µm. Further, we have recently succeeded in the very widely tunable generation from about 50 µm to 700 µm using DAST-DFG [44]. The purity of the THz wave is determined by the signals’ linewidth. A narrower THz linewidth can be obtained by introducing an injection seeding technique to DSW-QPM-OPO. The maximum and minimum THz wave outputs were observed when the signal polarizations were parallel to the a-axis and b-axis of the DAST crystal, respectively. The polarization direction of the THz wave was measured by a wire grid polarizer, and was confirmed to be parallel to the a-axis of the DAST crystal. The THz wave polarization was independent of the signal polarization. These results indicate that the d111 property of DAST crystals is utilized in THz wave generation. The obtained maximum THz wave output was 52 fJ/ pulse at a repetition rate of 1 kHz (52 pW on average) with an in-


419

Fig. 21. Examples of wavelength measurements: (a) output of dual-signal waves QPM-OPO, (b) generated THz wave. The THz wavelength was measured by a scanning metal-mesh etalon

cident sum signal energy of 270 µJ/ pulse. The repetition rate was increased up to 5 kHz, which was the upper limit of the response of the Si-bolometer. It should be added that we performed DFG experiments using other crystals, such as LiNbO3 , LiTaO3 , KTP, and GaAs under the same experimental conditions, however, only DAST generated a detectable THz wave. The conversion efficiency could be increased if we use the less absorptive frequency range of DAST in the sub-THz region [45] and the DFG interaction with longer coherence length [40].

6

Conclusion

We reviewed the widely tunable THz wave generation based upon nonlinear optical effects. In Sect. 3, we demonstrated a high spectral resolution, injection-seeded THz wave parametric generator (is-TPG). We measured the power enhancement, spectrum narrowing, and tunability of this is-TPG. In comparison with a conventional TPG without injection seeding, the output was increased from 3 to 900 pJ/ pulse, and the linewidth was decreased from > 500 GHz to ≈ 100 MHz. Wide tunability from 125 to 430 µm (frequency: 0.7

420

Kodo Kawase et al.

to 2.4 THz) was assured using a tunable seeder, and fine-tuning was demonstrated by THz spectroscopy of low-pressure water vapor. These measurements prove this method to be suitable for many application fields. These include spectroscopy, communication, medical and biological applications, THz imaging, and so forth. Further improvement of our system is possible. As OPGs and OPOs have improved tremendously in the last decade, the use of TPGs and TPOs shows great potential to move towards a lower threshold, higher efficiency, and wider tunability. A lower threshold and narrower linewidth can be expected using a nonlinear optical waveguide and longer pump pulse width, respectively. Operation in other wavelength regions, through proper crystal selection, should also be possible. Success in this will prove the practicality of a new widely tunable THz wave source, the is-TPG, that will compete with free-electron lasers and p-Ge lasers. In Sect. 4, efficient radiation of the THz wave was demonstrated by introducing an arrayed Si-prism coupler. Output was more than 6 times greater than with a single-prism coupling. The diffraction angle observed using the prism array was smaller than with single-prism coupling, because of the phased-array-like effect. We also explained the unidirectional THz wave radiation from the Siprism coupler. In Sect. 5, a compact and widely tunable THz wave source was realized by the combination of DAST crystal and DSW-QPM-OPO using PPLN with a series of gratings. The obtained THz wave power was 52 pW (rep. rate 1 kHz). For tunable THz wave applications, the simplicity of the wave source is an essential requirement since cumbersome systems do not encourage new experimental thoughts and ideas. Compared with the available sources, the present nonlinear optical methods have significant advantages in compactness, tunability, and ease of handling. Acknowledgements The authors would like to acknowledge the continuing guidance and encouragement of Prof. J. Nishizawa. The authors thank Prof. K. Mizuno of the Research Institute of Electrical Communication, Tohoku University, for providing the Schottky barrier diodes; C. Takyu for his excellent work coating the crystal surface, and T. Shoji for polishing the crystals superbly.

References 1. P. R. Smith, D. H. Auston, M. C. Nuss: Subpicosecond photoconducting dipole antennas, IEEE J. Quantum Electron. 24, 255–256 (1988) 397 2. X.-C. Zhang, B. B. Hu, J. T. Darrow, D. H. Auston: Generation of femtosecond electromagnetic pulses from semiconductor surfaces, Appl. Phys. Lett. 56, 1011–1013 (1990) 397 3. Far infrared region 1000 to 10 µm, Committee on Free Electron Lasers and Other Advanced Coherent Light Sources, National Research Council (National Academy Press, Washington, DC 1994) pp. 24–31 398


421

4. E. R. Brown, K. A. McIntosh, K. B. Nichols, C. L. Dennis: Photomixing up to 3.8 THz in low temperature-grown GaAs, Appl. Phys. Lett. 66, 285–287 (1995) 398 5. S. Matsuura, M. Tani, H. Abe, K. Sakai, H. Ozeki, S. Saito: High-resolution terahertz spectroscopy by a compact radiation source based on photomixing with diode lasers in a photoconductive antenna, J. Molec. Spectrosc. 187, 97– 101 (1998) 398 6. S. Komiyama: Far-infrared emission from population-inverted hot-carrier system in p-Ge, Phys. Rev. Lett. 48, 271–273 (1982) 398 7. E. Br¨ undermann, A. M. Linhart, H. P. R¨ oser, O. D. Dubon, W. L. Hansen, E. E. Haller: Miniaturization of p-Ge lasers: Progress toward continuous wave operation, Appl. Phys. Lett. 68, 1359–1362 (1996) 398 8. X. Zernike Jr., P. R. Berman: Generation of far infrared as a difference frequency, Phys. Rev. Lett. 15, 999–1001 (1965) 398, 402 9. R. Morris, Y. R. Shen: Theory of far-infrared generation by optical mixing, Phys. Rev. A 15, 1143–1156 (1977) 398, 402 10. J. Nishizawa: History and characteristics of semiconductor laser, Denshikagaku 14, 17–20 (1963) (in Japanese) 398 11. J. Nishizawa, K. Suto: Semiconductor Raman laser, J. Appl. Phys. 51, 2429– 2431 (1980) 398 12. J. M. Yarborough, S. S. Sussman, H. E. Puthoff, R. H. Pantell, B. C. Johnson: Efficient tunable optical emission from LiNbO3 without a resonator, Appl. Phys. Lett. 15, 102–105 (1969) 398 13. B. C. Johnson, H. E. Puthoff, J. SooHoo, S. S. Sussman: Power and linewidth of tunable stimulated far-infrared emission in LiNbO3 , Appl. Phys. Lett. 18, 181–183 (1971) 398, 399 14. M. A. Piestrup, R. N. Fleming, R. H. Pantell: Continuously tunable submillimeter wave source, Appl. Phys. Lett. 26, 418–421 (1975) 398, 399, 400, 401, 414 15. I. Shoji, T. Kondo, A. Kitamoto, M. Shirane, R. Ito: Absolute scale of secondorder nonlinear-optical coefficients, J. Opt. Soc. Am. B 14, 2268–2294 (1997) 399 16. J. Shikata, K. Kawase, K. Karino, T. Taniuchi, H. Ito: Tunable terahertz-wave parametric oscillators using LiNbO3 and MgO:LiNbO3 crystals, IEEE Trans. Microwave Theory Tech. 48, 653–661 (2000) 400, 402, 404 17. E. D. Palik (Ed.): Handbook of Optical Constants of Solids (Academic, New York 1991) 400 18. Y. R. Shen: The Principles of Nonlinear Optics (John Wiley, New York 1984) pp. 169–172 401 19. S. S. Sussman: Tunable light scattering from transverse optical modes in lithium niobate, Microwave Lab. Report, No. 1851 (Ph.D. Thesis), Stanford University (1970) 401 20. A. Yariv: Quantum Electronics (John Wiley, New York 1989) pp. 407–452 21. J. Shikata, K. Kawase, M. Sato, T. Taniuchi, H. Ito: Enhancement of THz wave output from LiNbO3 optical parametric oscillators by cryogenic cooling, Opt. Lett. 24, 202–204 (1999) 22. W. D. Johnston, Jr. I. P. Kaminow: Temperature dependence of Raman and Rayleigh scattering in LiNbO3 and LiTaO3 , Phys. Rev. 158, 1045–1054 (1968) 23. K. Kawase, M. Sato, T. Taniuchi, H. Ito: Coherent tunable THz wave generation from LiNbO3 with monolithic grating coupler, Appl. Phys. Lett. 68, 2483–2485 (1996) 402

422

Kodo Kawase et al.

24. K. Kawase, M. Sato, T. Taniuchi, H. Ito: Characteristics of THz wave radiation using a monolithic grating coupler on a LiNbO3 crystal, Int. J. Infrared Millim. Waves 17, 1839–1850 (1996) 402 25. K. Kawase, J. Shikata, K. Imai, H. Ito: A transform-limited, narrow-linewidth, THz wave parametric generator, Appl. Phys. Lett. 78, 2819–2821 (2001) 403 26. K. Kawase, H. Minamide, K. Imai, J. Shikata, H. Ito: Injection-seeded terahertz-wave parametric generator with wide tunability, Appl. Phys. Lett. 80, 195–197 (2002) 403 27. S. Wu, V. A. Kapinus, G. A. Blake: A nonosecond optical parametric generator/ amplifier seeded by an external cavity diode laser, Opt. Commun. 159, 74–79 (1999) 403 28. T. Nozokido, J. J. Chang, C. M. Mann, T. Suzuki, K. Mizuno: Optimization of a Schottky barrier mixer diode in the submillimeter wave, Int. J. Infrared Millim. Waves 15, 1851–1865 (1994) 404 29. K. Kawase, J. Shikata, H. Minamide, K. Imai, H. Ito: Arrayed silicon prism coupler for a THz wave parametric oscillator, Appl. Opt. 40, 1423–1426 (2001) 405, 412 30. K. Imai, K. Kawase, J. Shikata, H. Minamide, H. Ito: An injection-seeded terahertz-wave parametric oscillator, Appl. Phys. Lett. 78, 1026–1028 (2001) 406, 410 31. K. Kawase, M. Sato, K. Nakamura, T. Taniuchi, H. Ito: Uni-directional radiation of widely tunable THz wave using a prism coupler under non-collinear phase matching condition, Appl. Phys. Lett. 71, 753–755 (1997) 411 32. T. Nozokido, H. Minamide, K. Mizuno: Dulation of submillimeter wave radiation by laser-produced free carriers in semiconductors, Electron. Commun. Jpn. 80, Part 2, 1–9 (1997) 414 33. T. Walther, K. R. Chapin, J. W. Bevan: Comment on “Unidirectional radiation of widely tunable THz wave using a prism coupler under noncollinear phase matching condition”, Appl. Phys. Lett. 73, 3610–3611 (1998) 415 34. K. Kawase, T. Hatanaka, H. Takahashi, K. Nakamura, T. Taniuchi, H. Ito: Tunable THz wave generation from DAST crystal using dual signal-wave parametric oscillation of PPLN, Opt. Lett. 25, 1714–1716 (2000) 417 35. T. Hatanaka, K. Nakamura, T. Taniuchi, H. Ito, Y. Furukawa, K. Kitamura: Quasi-phase-matched optical parametric oscillation with periodically poled stoichiometric LiTaO3 , Opt. Lett. 25, 651–653 (2000) 417 36. S. Sohma, H. Takahashi, T. Taniuchi, H. Ito: Organic nonlinear optical crystal DAST growth and its device applications, J. Chem. Phys. 245, 359–364 (1999) 417 37. H. Nakanishi, Japan patent no. 1716929 (1986) 417 38. F. Pan, G. Knopfle, Ch. Bosshard, S. Follonier, R. Spreiter, M. S. Wong, P. Gunter: Electro-optic properties of the organic salt 4-N,N-dimethylamino-4N-methyl-stilbazolium tosylate, Appl. Phys. Lett. 69, 13–15 (1996) 417 39. R. Spreiter, Ch. Bosshard, F. Pan, P. Gunter: High-frequency response and acoustic phonon contribution of the linear electro-optic effect in DAST, Opt. Lett. 22, 564–566 (1997) 417 40. U. Meier, M. Bosch, Ch. Bosshard, F. Pan, P. Gunter: Parametric interactions in the organic salt 4-N,N-dimethylamino-4-N-methyl-stilbazolium tosylate at telecommunication wavelengths, J. Appl. Phys. 83, 3486–3489 (1998) 417, 419


423

41. T. J. Carrig, G. Rodriguez, T. S. Clement, A. J. Taylor: Generation of terahertz radiation using electro-optic crystal mosaics, Appl. Phys. Lett. 66, 10–12 (1995) 417 42. T. J. Carrig, G. Rodriguez, T. S. Clement, A. J. Taylor: Scaling of terahertz radiation via optical rectification in electro-optic crystals, Appl. Phys. Lett. 66, 121–123 (1995) 417 43. K. Kawase, M. Mizuno, S. Sohma, H. Takahashi, T. Taniuchi, Y. Urata, S. Wada, H. Tashiro, H. Ito: Difference-frequency terahertz-wave generation from 4-dimethylamino-N-methyl-4-stilbazolium-tosylate by use of an electronically tuned Ti:sapphire laser, Opt. Lett. 24, 1065–1067 (1999) 417 44. T. Taniuchi, N. Osaki, J. Shikata, H. Ito: Widely tunable THz wave generation in DAST crystal by nonlinear difference frequency mixing, CLEO 2002, Techn. Digest, (Opt. Soc. Am., Washington, DC 2002) paper CTuC5 418 45. M. Walther, K. Jensby, S. R. Keiding, H. Takahashi, H. Ito: Far-infrared properties of DAST, Opt. Lett. 25, 911–913 (2000) 419

Index

4-dimethylamino-N-methyl-4stilbazolium-tosylate (DAST), 417 DAST, 417 DAST-DFG, 418 difference frequency generation (DFG), 416, 417, 419 dual signal-wave quasi-phase-matched optical parametric oscillation (DSW-QPM-OPO), 416, 417 idler, 399, 402 injection seeding, 403–410 injection-seeded TPG (is-TPG), 403–410 laser – Nd:YAG, 404, 408, 412, 417 LiNbO3 , 398, 399, 401, 402 – periodically poled (PPLN), 417 LiTaO3 , 399 MgO:LiNb3 , 402

Nd:YAG laser, 404, 408, 412, 417 noncollinear phase matching, 399, 402, 410 nonlinear processes, 399, 400 optical parametric generator (OPG), 403 phase matching, 401 – noncollinear, 399, 402, 410 photon (idler), 398 photon (pump), 398 polariton, 399 PPLN, 417 pump, 399, 402, 404 SBD, 406 Schottky barrier diode (SBD), 404 Si-prism coupler, 411, 415 THz wave, 397, 399, 402, 411, 416 – parametric generator (TPG), 402 – parametric oscillator (TPO), 402, 412, 413, 416

Mid-Infrared and THz Coherent Sources Using Semiconductor-Based Materials Hiroshi Takahashi , Hidetoshi Murakami, Hideyuki Ohtake, and Nobuhiko Sarukura Institute for Molecular Science Nishigonaka, Myodaiji, Okazaki 4448585, Japan [email protected] Abstract. After the appearance of ultrafast lasers, various emitters were developed to generate mid-infrared and far-infrared coherent light sources. These were called “THz radiation” emitters, and a great deal of attention has been devoted to improve their performance. In this chapter, we summarize mid-infrared and far-infrared light sources that use semiconductor-based materials.

1

Introduction

“THz radiation” is a rather new technical term generally used to categorize far-infrared radiation. The frequency region of THz radiation is between the well-developed microwave region at GHz frequency and the optical electromagnetic wave region at PHz frequency. Since coherent light sources in this new frequency region have been difficult to access until recently, many frontier works are still considered to be part of this region. For example, spectroscopic properties of many materials, such as semiconductors, superconductors, and proteins, are poorly characterized in the THz region. In contrast to the visible and near-infrared regions, nonlinear parametric frequency conversion in the mid-infrared and far-infrared regions is difficult to achieve because of phonon absorption features present in solid materials. To overcome such a fundamental limitation, various semiconductor-based devices have been proposed, including photoconductive switches, quantum confined structures, semiconductor surfaces, and nonlinear optical processes. These are categorized into two groups depending on the waveform of emitted THz radiation: pulse or continuous-wave (CW). THz pulses are primarily generated by utilizing ultrafast laser pulses as excitation sources. In this case, a THz pulse consists of a single cycle of electromagnetic radiation and provides an extremely broad bandwidth. Due to the invention of THz time-domain spectroscopy, such as EO sampling or photoconductive sampling, the time-domain waveform of single-cycle THz radiation can be successfully observed [1,2,3,4]. Since this time-domain waveform can provide the refractive index of materials, THz time-domain spectroscopy is also used for various applications such

Corresponding author.

I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 425–445 (2003) c Springer-Verlag Berlin Heidelberg 2003

426

Hiroshi Takahashi et al.

as imaging or sensing [5,6,7,8]. Recent progress in this technique allows the construction of full three-dimensional images of common objects, and THz pulse is considered to be a promising light source to replace conventional X-rays [9,10]. In contrast, the CW source exhibits a narrow bandwidth with broad tunability in the THz region. Therefore, the CW source is considered to be a powerful tool for spectroscopic works, such as molecular analysis, atmospheric research, and bioscience, since it provides a high-resolution spectrum in the THz region. In this chapter, we primarily focus on semiconductorbased materials as coherent THz radiation sources and briefly review their basic physics.

2

THz Radiation from Semiconductor-Based Materials

2.1

Photoconductive Switch

Photoconductive switches irradiated with ultrafast optical pulses are widely used to generate THz radiation. The basic idea of THz radiation from a photoconductive switch was suggested by Auston et al. [11], and the generation and detection scheme of THz radiation was demonstrated by using fast photoconducting material as time-varying Hertzian dipoles. The experimental configuration demonstrated by Auston et al. is shown in Fig. 1. Two identical photoconductors are located symmetrically on opposite sides of a thin slab of insulating material. To obtain a very high response speed, a 1 µm epitaxial silicon film implanted with 3 × 1015 Ar+ ion/ cm2 at an energy of 2 MeV was used as a photoconducting film. The ion-implantation technique is generally used to produce the large density of defects that work as the carrier trap and effectively reduce the carrier lifetime. The active region of the photoconductive switch was a 10 µm gap between two aluminum electrodes. Subpicosecond optical pulses were focused on the active region of the transmitting photoconductive switch (left side) and produced a photocurrent. The photocurrent had a rapid rise time due to excitation of free carriers by optical pulses, and decayed rapidly due to their capture by the traps. For a point source (dimension wavelength), the electric field of THz radiation is given by the second derivative of the induced polarization: ∂2P . (1) ∂t2 Therefore, this photocurrent creates a transient polarization and emits THz radiation. The optical pulses were used to detect THz radiation.They illuminated the receiving photoconductive switch (right side), and were delayed relative to the optical pulses illuminating the transmitting photoconductive switch. Scanning the optical delay enabled measuring the time-domain waveform of THz radiation as the average current in the receiving photoconductive switch. The measured average current is shown in Fig. 2. The observed signal was extremely fast, and its full width at half maximum (FWHM) was 2.3 ps. E (t) ∝

Mid-Infrared and THz Coherent Sources

427

Fig. 1. Experimental configuration used to generate and detect fast electromagnetic pulses. Both the generator and detector are time-varying dipoles. The transmitting photoconductive switch (left side) has a dc bias and the receiving photoconductive switch (right side) which is connected directly to a low-frequency amplifier, receives its bias from the electromagnetic pulse [11]

Fig. 2. Measured response of the experiment illustrated in Fig. 1. The average current in the receiving dipole is plotted as a function of the relative delay between the optical pulses used to excite the two photoconductive switches. The horizontal time scale is 2 ps/div [11]

Fattinger et al. improved this system and generated freely propagating, diffraction-limited beams of single-cycle THz radiation [12]. The schematic diagram of the sample structure is shown in Fig. 3. The transmission line consisted of two 5 µm wide aluminum electrodes separated by 25 µm. The substrate was heavily implanted silicon-on-sapphire to ensure a short carrier lifetime, as mentioned above. A spherical sapphire lens on the back of the substrate was used to collimate the THz radiation, which provides a relatively large beam diameter with a diffraction-limited divergence. Grischkowsky et al.

428


Fig. 3. (a) Schematic diagram of the charged coplanar transmission line. The laser excitation beam spot defines the location of the transient electric dipole. The monitor beam measures the electrical pulse coupled to the line. (b) Schematic diagram of the THz detector. The laser detection beam spot is shown centered on the gap in the focal spot of the THz radiation. (c) Schematic diagram of the collimating and focusing optics, which consists of crystalline, sapphire lenses in contact with the backside of the SOS chips [12]

fabricated a photoconductive switch on high-resistivity GaAs as a substrate, and reported that the amplitude of THz radiation was five times larger than that of silicon-on-substrate and the spectrum became broader [13]. To widen the bandwidth of THz radiation, several techniques have been proposed and demonstrated to reduce the carrier lifetime, since a shorter lifetime reduces the decay time of transient polarization, and leads to radiation in the higherfrequency region. A photoconductive switch fabricated on a low-temperaturegrown GaAs (LT-GaAs) substrate has recently been exploited by many re-


429

searchers, since it provides both shorter lifetime and higher mobility of carriers, depending on the growth condition of molecular beam epitaxy. The photoconductive switch is also used to generate THz radiation by difference-frequency mixing [14,15]. Figure 4 shows an experimental setup for tunable narrow-band THz radiation, as suggested by Weling et al. [14]. The linearly chirped optical pulses were split into two pulses by a beam splitter and one pulse was delayed with respect to the other. The two pulses were then recombined, and sent to the large-aperture photoconductive switch. The photoconductive switch was fabricated by depositing 2-mm-wide aluminum electrodes with a 4-mm spacing on a semi-insulating GaAs substrate. Since the frequency of linearly chirped pulses varied linearly as a function of time, overlap of two relatively delayed optical pulses generated interference at a constant beat frequency, which was proportional to the time delay and the amount of frequency chirp. The photoconductive switch, which was excited by the interferometer output, served as an optical mixer and converted this optical modulation into THz radiation with a beat frequency. Figure 5 shows the interferograms of THz radiation observed for various time delays between two linearly chirped optical pulses. As expected, the oscillation frequency exhibited a linear dependence on the relative delay. A different method is used to generate THz radiation from an antenna structure. In contrast to the semiconductor-based dipole antenna, in which optical pulses acted as gate pulses to excite carriers for the transient photocurrent, Hangyo et al. proposed a completely different scheme [16]. They

Fig. 4. Experimental setup to generate tunable THz radiation from the overlap of two chirped pulses [14]

430


Fig. 5. THz radiation interferograms from beating of 22.5-ps chirped pulses at various time delays [14]

Fig. 6. (a) Schematic structure of the sample. Thin films of YBa2 Cu3 O7−δ (YBCO) have been deposited on a 0.5-mm-thick MgO (100) substrate by rf magnetron sputtering. A coplanar transmission line with a bridge structure was patterned into the films by using photolithographic techniques and a wet chemical etch. (b) THz radiation spectrum measured for the +100 mA bias current [16]


431

used a dipole antenna fabricated on superconducting material as a THz radiation source. The schematic diagram of the sample structure is shown in Fig. 6a. A coplanar transmission line with a bridge structure was patterned on the film by using a photolithographic technique and wet chemical etching. The THz radiation spectrum measured at 11 K is shown in Fig. 6b. The THz radiation is ascribed to the ultrafast supercurrent modulation produced by optical excitation. In this case, optical pulses decreased the supercurrent through the break up of Cooper pairs. The schematic difference of THz radiation from semiconductor-based and superconducting-based antenna structures is depicted in Fig. 7.

Semiconductor-based antenna structure Bias voltage

Superconducting-based antenna structure Bias current

Gap photoconducting semiconductor Superconducting Insulator Electrode thin film thin film substrate Excitation laser pulse Excitation laser pulse t J

t J

E ~ dJ / dt

E ~ dJ / dt

(a)

(b)

Fig. 7. Schematic difference of THz radiation from (a) semiconductor-based and (b) superconducting-based antenna structures

2.2

Quantum Confined Structure

THz radiation from a semiconductor-based quantum confined structure has attracted much attention as a compact, coherent, and intense solid-state source. There are two types of excitation schemes to generate THz radiation: optical excitation and electrical excitation. In the optical excitation scheme, the coherent charge oscillation in the quantum confined structure is utilized as a radiation source. Planken et al.

432


observed THz radiation from a single quantum well after coherent optical excitation of light hole (lh) and heavy hole (hh) excitons [17]. The lh and hh envelope states of the quantum well in an electric field are shown in Fig. 8. The charge oscillations result from time evolution of the superposition of the lh and hh exciton states, which induces lh–hh valence band mixing of a transition dipole moment between the lh and hh subbands. A density matrix formulation in the three-band model provides clear insight into the physical origin of THz radiation: P (t) =

|e| [(z11 − z33 ) ρ11 + (z22 − z33 ) ρ22 + 2Re (z12 ρ12 )] , V

(2)

where the self-dipole moments are defined as z11 = lh|z|lh, z22 = hh|z|hh, and z33 = e|z|e, and the intersubband transition dipole moment as z12 = lh|z|hh. The generated lh density is ρ11 , the generated hh density is ρ22 , and the coherence between the hh and lh states is ρ12 . The last term is proportional to ej(ω2 −ω1 ) and is responsible for the charge oscillation. The THz radiation is E(t) ∝ d2 P/dt2 , and contains both the initial transient caused by the excitation of polarized electron–hole pairs and the oscillatory tail caused by the charge oscillation. Figure 9 shows the time-domain waveform of THz radiation as a function of wavelength at an electrical field strength of 4 kV/ cm. The oscillatory tail was only observed for 1.527 eV excitation, which excited both the lh and hh excitons. The oscillation has a frequency of 1.4 THz, which corresponds to the lh–hh splitting. Christian et al. used the Bloch oscillation in the superlattice as a THz radiation source [18]. The Bloch oscillation is caused by Bragg reflections of ballistic electrons at the Brillouin zone boundary, and its oscillation frequency is given as νB = eF d/h, where d is the lattice constant, and F is the electric field. In bulk materials, scattering easily destroyed the coherence of the Bloch oscillation before an oscillation cycle was completed. However, for a superlattice, the Bloch oscillation becomes observable since the lattice constant in the above equation is replaced by the superlattice period, and high-oscillation frequency is achieved even with a modest electric field. Figure 10 shows the Wannier–Stark (WS) states in the superlattice. THz radiation results from the Bloch oscillation of the carriers after coherent excitation of several WS

Fig. 8. Wavefunction envelopes of the electron, lh and hh in a quantum well biased with an electric field [17]


433

Fig. 9. Time-domain waveform of THz radiation for several excitation wavelengths. The electric field in the MQW is 4 kV/ cm. The inset shows the measured photocurrent spectrum [17]

Fig. 10. Schematic representation of the transitions from the valence band to the conduction band of a superlattice in the Wannier–Stark bias regime [18]

ladder transitions. Figure 11 depicts the observed coherent THz radiation as a function of time for different bias voltages. At higher bias, the peak frequency shifts toward the higher-frequency regime, which corresponds to the energy separation between adjacent WS transitions. A quantum cascade structure based on an intersubband transition is used for the electrical excitation scheme. Benjamin et al. used an interwell– intersubband transition as a THz radiation source [19]. Figure 12 shows the schematic diagram of a multiple-quantum-well structure, that illustrates the formation of a three-level system. The energy separation between E2 and E1 (≈ 40 meV) is designed to be greater than the GaAs LO-phonon energy (36 meV), so that electrons on the E1 level can be effectively depopulated by LO-phonon scattering. The scattering from E3 to E1 is also dominated by LO-

434


Fig. 11. Left side: Measured coherent electromagnetic transients emitted from the superlattice for different reverse-bias voltages. Right side: Fourier transforms of the time-domain date [18]

phonon scattering, but the small overlap of wavefunctions leads to a longer scattering time. The scattering process from E3 to E2 is due to acousticphonon scattering, electron–electron scattering, and LO-phonon scattering. When τ32 exceeds τ21 , population inversion occurs between the E3 and E2 levels. THz radiation is generated by the optical transition between E3 and E2 . Figure 13 presents the THz radiation measured at 5 K and 80 K. The FWHM of the THz radiation spectrum is only 0.47 THz. The large part of this bandwidth is induced by the energy splitting of injection levels E1 and E3 . Therefore, the bandwidth can be further reduced by increasing the thickness of the injection barrier B3 , which decreases the coupling between E1 and E3 . This, however, leads to a relatively slow extraction time for the lower laser level. Recently, a THz laser based on the interminiband transition in the conduction band of the heterostructure was reported by K¨ ohler et al. [20]. They adopted a chirped superlattice as an active region to avoid the above draw-


435

Fig. 12. Calculated band diagram, subband levels, and squared magnitude of wave functions of two cascade-connected triple-quantum-well modules under a bias of 51 mV/module. At this bias, the E1 and E3 subbands anticross, E3 − E2 = 11.3 meV, and E2 − E1 ≈ 40 meV [19]

Fig. 13. THz intersubband emission taken at (a) 5 K and (b) 80 K under a bias of 1.6 V. The inset shows the spectrum under a bias of 4.0 V [19]

back. Figure 14 presents the schematic diagram of wave functions and energies in the conduction band. The optical transition takes place across the 18 meV minigap between the second and first miniband. The lower state 1 is strongly coupled to a wide injection miniband. This provides a large phase space for electrons scattered from the injector and leads to the first depletion of state 1. Additionally, the wide miniband allows efficient electrical transport, even at high current densities. Figure 15 shows the THz radiation spectrum for different drive currents. The characteristic narrowing of the emission line and the nonlinear dependence of the intensity are clearly observed. An output power

436


Fig. 14. Self-consistent calculation of the wavefunctions and energies in the conduction band of a THz laser under a field of 3.5 kV/ cm. Injectors and active superlattice (SL) are alternating [20]

Fig. 15. THz radiation spectra from a 1.24-mm-long and 180-µm-wide laser device recorded at 8 K for different drive currents. Subthreshold spectra were recorded with a resolution of 2 cm−1 (60 GHz) using the FTIR with a liquid-helium-cooled Si bolometer in step-scan mode. Laser spectra were colleted in rapid-scan mode with the maximum possible resolution of 0.125 cm−1 (3.75 GHz) using a DTGS detector. The inset shows the laser line on a logarithmic scale [20]


437

exceeding 2 mW with low-threshold current densities of about a few hundred A/cm2 was demonstrated. These results are very attractive, and will provide the most promising candidate for a continuous-wave THz radiation source in real-world applications if it can be improved to operate at room temperature. 2.3

Semiconductor Surface

Semiconductor surfaces irradiated with ultrafast optical pulses are widely used as a THz radiation source since they provide intense THz radiation and do not require chemical processes and microfabrication techniques for sample preparation. The THz radiation from a semiconductor surface is categorized into two major processes: difference-frequency mixing and transient photocurrent induced by photoexcitation. The former process creates an instantaneous polarization via optical rectification, which is described by the second-order nonlinear susceptibility, and becomes dominant under high-excitation conditions [21,22]. Using the (111) surface significantly enhances THz radiation originating from this process since it has an asymmetric structure and the nonlinear optical process is effectively enhanced. In contrast, under lowexcitation conditions, the contribution from the transient photocurrent becomes the dominant source for THz radiation. Depending on surface properties, the transient photocurrent originates from two sources: the acceleration of photoexcited carriers by the surface-depletion field and the surge current induced by different diffusion velocities between photoexcited electrons and holes (photo-Dember field) [23,24]. In semiconductors with a wide bandgap, such as GaAs or InP, the former process is dominant in generating THz radiation. Figure 16 illustrates a band diagram of InP, which has negative band bending near the air–semiconductor interface [25]. For InP, the photocurrents are produced through carrier separation by the depletion field, the electrons moving to the surface and the holes to the wafer. The rise time of the photocurrent is on the order of the laser pulse duration, and the decay time is the transit time of carriers crossing the depletion field. This photocurrent creates a transient polarization and emits THz radiation. In contrast, for narrow-bandgap semiconductors such as InAs and InSb, the latter origin is thought to be the main source for THz radiation, as both the narrow absorption depth and the high kinetic energy of the photoexcited carriers largely enhance the diffusion process. In this case, ultrafast build-up and relaxation of the photo-Dember field create a transient polarization. Since Zhang et al. reported the quadratic magnetic-field dependence of THz radiation power from GaAs, [26] many studies have shown that applying a magnetic field enhances THz radiation by an order of magnitude. This enhancement is explained by the change in the direction of carrier acceleration, which is induced by the Lorentz force in a magnetic field [27]. Owing to the high contrast of the refractive index at the semiconductor surface, the change of acceleration direction enhances THz radiation transmitted through the air–semiconductor interface.

438


Fig. 16. Band diagram of semi-insulating InP. The surface state near the conduction-band edge causes Fermi level pinning at the interface. Photocarriers are swept across the depletion width ld by the built-in field [25]

Fig. 17. Experimental setup for the THz radiation emitter in a magnetic field. A liquid-helium-could InSb bolometer with calibrated sensitivity was provided for detection. The sample was placed at 45◦ to the magnetic field, and the excitation laser was parallel to the magnetic field. The fourier spectrum of the THz radiation was measured by the polarizing Michelson interferometer [28]

Although GaAs is used as a THz radiation source, its electron-effective mass at the Γ -valley is approximately three times larger than that of InAs. The smaller effective mass of photoexcited carriers is considered to be advantageous for enhancing THz radiation power. Therefore, the THz radiation from InAs is expected to be more intense than that of GaAs. Sarukura et al. demonstrated an intense THz radiation source from a femtosecond laser irradiating InAs in a magnetic field [28]. The schematic diagram of their experimental setup is depicted in Fig. 17. They reported that the THz radiation from InAs exhibited a quadratic magnetic-field dependence, and its power was one order of magnitude higher than that from GaAs as shown in Fig. 18. Ohtake et al. applied a magnetic field up to 5 T to InAs by using a superconducting magnet [29]. They found clear saturation of the THz radiation power at 3 T, as shown in Fig. 19. The average power from this system will be approximately 100 µW. The saturation mechanism is still being discussed among many researchers [30,31,32]. Figure 20 illustrates twodimensional plots for THz radiation spectra as a function of the magnetic field and geometrical layouts. The maximum bandwidth is approximately 3 THz,


439

Fig. 18. Excitation-power dependence of THz radiation power. THz radiation from InAs increases rapidly as the excitation power increases. THz radiation from GaAs is also shown for comparison [28]

Fig. 19a,b. Magnetic-field dependence of THz radiation power. The insets indicate the experimental setup for the geometrical layouts of the excitation laser and magnetic field [29]

440


Fig. 20a–c. Two-dimensional plots for THz radiation spectra on different magnetic fields. Open circles and bars show the center frequency and spectral bandwidth. Hpol and V-pol indicate horizontal and vertical polarization [29]

and the center frequency of the THz radiation spectra is found to be easily controlled by the external magnetic field. 2.4

Nonlinear Optical Process

After the appearance of commercially available 10-fsec mode-locked Ti:sapphire lasers, it became much easier to generate ultrabroadband THz radiation or mid-infrared light pulses by nonlinear optical processes. In contrast to other semiconductor-based schemes, in which the radiated frequencies are limited to below a few THz, the nonlinear process does not have such limitations. This chapter introduces non-phase-matched and phase-matched optical rectification schemes. Bonvalet et al. generated a mid-infrared emission from a 0.1-mm-thick GaAs plate grown along the 110 direction and irradiated with 15 fs optical pulses at a 100 MHz repetition rate [33]. Figure 21 presents the midinfrared emission spectrum observed from the GaAs at normal incidence. By using optical pulses in the 10-fsec regime, coherent infrared light pulses with a spectrum extending up to 7 µm are successfully generated via optical rectification. Although this scheme provides ultrabroadband THz radiation, it is a non-phase-matched process and the average power is extremely low. Kaindl et al. demonstrated phase-matched optical rectification to enhance the average power of a mid-infrared light pulse [34]. They used a GaSe crystal


441

Fig. 21. Spectrum of mid-infrared emission at normal incidence through the GaAs sample, as measured by the HgCdTe detector. The dashed line shows how the theoretical spectrum corresponds to the expected emission when using 15-fs Gaussian pulses and a pure nonresonant rectification effect [33]

Fig. 22. Normalized spectra of femtosecond mid-infrared pulses continuously tunable from 9 to 18 µm. The spectra are derived from electric field correlations. The inset shows the tunning curve of GaSe. The phasematching angle is plotted versus the center wavelength of the pulses [34]

as a nonlinear optical crystal, since it has an exceptionally high second-order nonlinear susceptibility. Broadband 20 fs optical pulses were injected into a GaSe crystal, in which different components of the input spectrum contribute to the phase-matched optical rectification. Figure 22 depicts the normalized spectra of femtosecond mid-infrared light pulses obtained with different phase-matching angles. The spectra exhibited a bandwidth of 10–15 meV and demonstrated broad tunability from 9 to 18 µm. The average power of the pulses generated by the phase-matched optical rectification is ≈ 1 µW at a wavelength of 15 µm. This scheme has many advantages, including its inherent simplicity, broadband tunability, and high average power, and may

442


be a promising candidate as a coherent mid-infrared light source in practical use.

3

Concluding Remarks

This chapter briefly introduced historical and currently developing works on generating THz radiation. Since the photon energy in the mid-infrared and far-infrared regions is quite close to the thermal energy of room temperature, the realistic design of a gain medium to realize population inversion has been extremely difficult, except for the gas molecule rotational transition. Unlike the frequency conversion in the shorter wavelength region, nonlinear crystals cannot be widely used at longer wavelengths because of their strong phonon absorption. To overcome these limitations, various nontraditional radiationgeneration schemes have been successfully developed, as described in this chapter. Owing to these inventions, many applications of THz radiation have been reported recently not only in the field of spectroscopy but also for imaging applications that will potentially substitute for some X-ray imaging. However, these applications still have much room for improvement because of the limited power available from the current light sources. In this sense, the key issue should still be developing an intense and compact light source as well as an image sensor to expand THz radiation applications.

References 1. D. H. Auston, P. R. Smith: Generation and detection of millimeter waves by picosecond photoconductivity, Appl. Phys. Lett. 43, 631–633 (1983) 425 2. Q. Wu, M. Litz, X. C. Zhag: Broadband detection capability of ZnTe electrooptic field detectors, Appl. Phys. Lett. 68, 2924–2926 (1996) 425 3. F. G. Sun, G. A. Wagoner, X. C. Zhang: Measurement of free-space terahertz pulses via long-lifetime photoconductors, Appl. Phys. Lett. 67, 1656–1658 (1999) 425 4. S. Kono, M. Tani, K. Sakai: Ultrabroadband photoconductive detection: Comparision with free-space electro-optic sampling, Appl. Phys. Lett. 79, 898–890 (2001) 425 5. B. B. Hu, M. C. Nuss: Imaging with terahertz waves, Opt. Lett. 20, 1716–1718 (1995) 426 6. D. M. Mittleman, S. Hunsche, L. Boivin, M. C. Nuss: T-ray tomography, Opt. Lett. 22, 904–906 (1999) 426 7. H. Harde, R. A. Cheville, D. Grischkowsky: Terahertz studies of collisionbroadened rotational lines, J. Phys. Chem. A. 101, 3646–3660 (1997) 426 8. R. A. Cheville, D. Grischkowsky: Far-infrared terahertz time-domain spectroscopy of frames, Opt. Lett. 20, 1646–1648 (1995) 426 9. P. Y. Han, G. C. Cho, X. C. Zhang: Time-domain transillumination of biological tissues with terahertz pulses, Opt. Lett. 25, 242–244 (2000) 426 10. B. Ferguson, S. Wang, D. Gray, D. Abbot, X. C. Zhang: T-ray computed tomography, Opt. Lett. 27, 1312–1314 (2002) 426


443

11. D. H. Auston, K. P. Cheung, P. R. Smith: Picosecond photoconducting Herzian dipoles, Appl. Phys. Lett. 45, 284–286 (1984) 426, 427 12. C. Fattinger, D. Grischkowsky: Terahertz beams, Appl. Phys. Lett. 54, 490–492 (1989) 427, 428 13. D. Grischkowsky, S. Keiding, M.Exter, C. Fattinger: Far-infrared time-domain spectroscopy with terahertz beams of dielectrics and semiconductors, J. Opt. Soc. Am. B 7, 2006–2015 (1990) 428 14. A. S. Weling, B. B. Hu, N. M. Froberg, D. H. Auston: Generation of tunable narrow-band THz radiation from large aperture photoconducting antennas, Appl. Phys. Lett. 64, 137–139 (1994) 429, 430 15. J. Y. Sohn, Y. H. Ahn, D. J. Park, E. Oh, D. S. Kim: Tunable terahertz radiation using femtosecond pulse shaping, Appl. Phys. Lett. 81, 13–15 (2002) 429 16. M. Hangyo, S. Tomozawa, Y. Murakami, M. Tonouchi, M. Tani, Z. Wang, K. Sakai, S. Nakamura: Terahertz radiation from superconducting YBCO thin films excited by femtosecond optical pulses, Appl. Phys. Lett. 69, 2122–2124 (1996) 429, 430 17. P. C. M. Planken, M. C. Nuss, I. Brener, K. W. Goossen, M. S. C. Luo, S. L. Chuang, L. Pfeiffer: Terahertz emission in single quantum wells after coherent optical excitation of light hole and heavy hole excitons, Phys. Rev. Lett. 69, 3800–3803 (1992) 432, 433 18. C. Waschke, H. G. Roskos, R. Schwedler, K. Leo, H. Kruz, K. Köhler: Coherent submillimeter-wave emission from Bloch oscillations in a semiconductor superlattice, Phys. Rev. Lett. 70, 3319–3322 (1993) 432, 433, 434 19. B. S. Williams, B. Xu, Q. Hu, M. R. Melloch: Narrow-linewidth terahertz intersubband emission from three-level systems, Appl. Phys. Lett. 75, 2927–2929 (1999) 433, 435 20. R. K¨ ohler, A. Tredicucci, F. Beltram, H. E. Beere, E. H. Linfield, A. G. Davies, D. A. Ritchie, R. C. Iatti, F. Rossi: Terahertz semiconductor-heterostructure laser, Nature 417, 156–159 (2002) 434, 436 21. S. L. Chuang, S. Schmitt-Rink, B. I. Greene, P. N. Saeta, A. F. J. Levi: Optical rectification at semiconductor surfaces, Phys. Rev. Lett. 68, 102–105 (1992) 437 22. M. Migita, M. Hangyo: Pump-power dependence of THz radiation from InAs surfaces under magnetic fields excited by ultrashort laser pulses, Appl. Phys. Lett. 79, 3437–3439 (2001) 437 23. S. Kono, P. Gu, M. Tani, K. Sakai: Temperature dependence of terahertz radiation from n-type InSb and n-type InAs surfaces, Appl. Phys. B 71, 901–904 (2000) 437 24. P. Gu, M. Tani, S. Kono, K. Sakai, X. C. Zhang: Study of terahertz radiation from InAs and InSb, J. Appl. Phys. 91, 5533–5537 (2002) 437 25. X. C. Zhang, D. H. Auston: Optoelectronic measurement of semiconductor surfaces and interfaces with femtosecond optics, J. Appl. Phys. 71, 326–338 (1992) 437, 438 26. X. C. Zhang, Y. Liu, T. D. Hewitt, T. Sangsiri, L. E. Kingsley, M. Weiner: Magnetic switching of THz beams, Appl. Phys. Lett. 62, 2003–2005 (1993) 437 27. M. B. Johnston, D. M. Whittaker, A. Corchia, A. G. Davies, E. H. Linfield: Simulation of terahertz generation at semiconductor surfaces, Phys. Rev. B 65, 165301 (2002) 437

444


28. N. Sarukura, H. Ohtake, S. Izumida, Z. Liu: High average-power THz radiation from femtosecond laser-irradiated InAs in a magnetic field and its elliptical polarization characteristics, J. Appl. Phys. 84, 654 (1998) 438, 439 29. H. Ohtake, S. Ono, M. Sakai, Z. Liu, T. Tsukamoto, N. Sarukura: Saturation of THz radiation power from femtosecond-laser-irradiated InAs in a high magnetic field, Appl. Phys. Lett. 76, 1398–1400 (2000) 438, 439, 440 30. J. Shan, C. Weiss, R. Beigang, T. F. Heinz: Origin of magnetic field enhancement in the generation of terahertz radiation from semiconductor surfaces, Opt. Lett. 26, 849–851 (2001) 438 31. J. Heyman, P. Neocleous, D. Hebert, P. A. Crowell, T. M¨ uller, K. Unterrainer: Terahertz emission from GaAs and InAs in a magnetic field, Phys. Rev. B 64, 085202 (2001) 438 32. G. Meinert, L. Banyai, P. Gartner, H. Haug: Theory of THz emission from optically excited semiconductors in crossed electric and magnetic fields, Phys. Rev. B 62, 5003 (2000) 438 33. A. Bonvalet, M Joffre, J. L. Martin, A. Migus: Generation of ultrabroadband femtosecond pulses in the mid-infrared by optical rectification of 15 fs light pulses at 100 MHz repetition rate, Appl. Phys. Lett. 67, 2907–2909 (1995) 440, 441 34. R. A. Kaindl, D. C. Smith, M. Joschko, M. P. Hosselbeck, M. Woerner, T. Elsaesser: Femtosecond infrared pulses tunable from 9 to 18 µm at an 88-MHz repetition rate, Opt. Lett. 23, 861–863 (1998) 440, 441

Index

ballistic electron, 432 Bloch oscillation, 432

Lorentz force, 437 LT-GaAs, 428

carrier lifetime, 426–428 charge oscillation, 431, 432 Cooper pairs, 431

multiple quantum-well, 433 non-phase-matched, 440 non-phase-matched process, 440

difference-frequency mixing, 429, 437 optical rectification (OR), 437, 440 effective mass, 438 exciton, 432 GaAs, 428, 429, 433, 437, 438, 440 GaSe, 440, 441 heavy hole, 432 InAs, 437, 438 InP, 437 InSb, 437 interminiband transition, 434 interwell–intersubband transition, 433 light hole, 432 LO-phonon scattering, 433, 434

phase-matched optical rectification, 440, 441 phonon absorption, 425, 442 photo-Dember field, 437 photoconductive switch, 425, 426, 428, 429 population inversion, 434, 442 quantum well, 432 semiconductor surface, 425, 437 superlattice, 432, 434 surface-depletion field, 437 Wannier–Stark (WS) state, 432

Mid-Infrared Laser Applications in Spectroscopy Frank K. Tittel1 , Dirk Richter2 , and Alan Fried2 1

2

Rice Quantum Institute, Rice University Houston, TX 77251–1892, USA [email protected] The National Center for Atmospheric Research 1850 Table Mesa Dr., Boulder, CO 80305, USA {dr,fried}@ucar.edu

Abstract. The vast majority of gaseous chemical substances exhibit fundamental vibrational absorption bands in the mid-infrared spectral region (≈ 2–25 µm), and the absorption of light by these fundamental bands provides a nearly universal means for their detection. A main feature of optical techniques is the non-intrusive in situ detection capability for trace gases. The focus time period of this chapter is the years 1996–2002 and we will discuss primarily CW mid-infrared laser spectroscopy. We shall not attempt to review the large number of diverse mid-infrared spectroscopic laser applications published to date. The scope of this chapter is rather to discuss recent developments of mid-infrared laser sources, with emphasis on established and new spectroscopic techniques and their applications for sensitive, selective, and quantitative trace gas detection. For example, laboratory based spectroscopic studies and chemical kinetics, which will also benefit from new laser source and technique developments, will not be considered.

1

Solid-State Mid-IR Spectroscopic Laser Sources

Mid-infrared laser sources come in many different varieties. For their application to spectroscopic measurement, and specifically for quantitative measurements, the ideal source would have the following properties: (1) sufficient optical power to overcome inherent electronic detection noise and ensure high laser signal-to-noise ratios, (2) narrow linewidth to obtain high selectivity and sensitivity, (3) single longitudinal mode operation with low amplified spontaneous emission output for high selectivity and elimination of intermode competition noise; (4) ease of tailoring the inherent laser operating wavelength (design of gain material and/or cavity structure) to access the desired absorption region; (5) low source noise and low amplitude modulation; (6) high beam quality, i.e., small beam divergence, small astigmatism, and stable, predictable beam output direction, for optimum coupling into and through a gas sampling cell; (7) low temperature and current tuning rates to minimize wavelength jitter induced by controller noise; (8) rapid wavelength tunability for fast response and high data acquisition rates; (9) minimal susceptibility to changing environmental conditions of temperature, pressure, humidity, I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 445–516 (2003) c Springer-Verlag Berlin Heidelberg 2003

446

Frank K. Tittel et al.

and vibrations; (10) no long term changes in laser wavelength and/or spatial output characteristics; (11) highly reliable performance for many years; and (12) compact and robust overall sensor package size. Of course it is challenging to realize all these idealized attributes in any one real world mid-infrared laser source, and they are listed here to serve as a general guideline. However, some of the attributes are more important than others for a given application to obtain the best possible measurement performance. In the sections that follow, we will elucidate wherever possible those important attributes necessary for achieving optimal performance employing the technique being discussed. It is common practice in infrared spectroscopy to express transition frequencies in inverse centimeters (cm−1 ), or wavenumbers, defined simply as the inverse of the transition wavelength in vacuum, ν = λ−1 . Multiplying this quantity by c gives the frequency in hertz; thus 1 cm−1 is roughly 30 GHz. We shall use both units throughout this chapter, where appropriate. The spectral coverage of the most frequently employed tunable CW midinfrared sources is shown in Fig. 1. In general, one can distinguish between two classes of mid-IR laser sources. Class “A” includes sources which generate tunable mid-IR laser radiation directly from gain in gas discharge, semiconductors, rare-earth and transition-metal doped solid-state bulk materials or optical fibers. Class “B” laser sources are based on nonlinear optical parametric frequency conversion of near-infrared (≈ 0.9 to 2 µm) laser sources. Rotation-Vibrational

Class ‘A’

Overtone Region Gas (Line Tunable)

CO Laser

I

QC-Laser Lead-Salt

Semiconductor

Antimonide III-V Cr2+ II-VI

Direct Solid-State

Class ‘B’

II

CO2 Laser

QPM GaAs PPLN / PPKTP / PPRTA 3+ 3+ Tm / Ho 3+ Solid State and Er Fiber Laser / Pr3+ Amplifier Yb3+ DFG / OPO / OPA 3+ Nd DFB Diode Laser OPSL

Frequency Conversion Based Sources

0.1

1

2

3

Wavelength ("Pm) m)

4

5 6 7 8 9 10

Fig. 1. Laser sources and typical wavelength coverage. Shown are also the wavelengths of the two atmospheric windows I (2.9 –5.3 microns) and II (7.6–16 microns), typically accessed for trace gas sensing. OPSL, optically pumped semiconductor laser

Mid-Infrared Laser Applications in Spectroscopy

447

Many hundreds of non-linear optical crystals have been developed for this purpose, however only very few of these are practical and in use. In this chapter, examples of parametric frequency conversion sources will be limited to quasi-phase-matched materials (QPM). While class “A” mid-IR laser sources are inherently more compact, class “B” laser sources can be made to operate over a much wider wavelength spectrum, often in a single optical arrangement, which is a key advantage for certain applications. Figure 1 shows the spectral coverage of Class “A” and “B” laser sources in the midIR wavelength region. The merits of the various spectroscopically important laser sources shown will be discussed in this section. Although line-tunable laser sources such as CO and CO2 gas lasers have and are being successfully used in trace gas detection [1,2], they are not within the scope of this book, which focuses specifically on solid-state laser sources. 1.1

Class “A” Laser Sources

Class “A” lasers as defined above, are direct laser sources. In the following subsections these lasers are discussed with respect to their typical operating characteristics and optical set-up conditions. The first group of class “A” lasers is further subdivided into semiconductor based lasers, namely lead-salt, antimonide and quantum cascade laser sources, which share some common features such as cryogenic cooling (in some cases thermoelectric) and the set-up of collection optics. This section is followed by a discussion of tunable solid-state laser sources. 1.1.1

Lead-Salt Diode Lasers

Lead-salt diode lasers have been developed for operation at wavelengths from 3 to 30 µm and have been available since the mid-1960s. These lasers are comprised of PbTe, PbSe, and PbS and various alloys of these compounds with the same materials above and with SnSe, SnTe, CdS and other materials. The lead-salt diode laser consists of a single crystal of these semiconductor materials to form a p–n junction. The crystal is shaped into an optical cavity employing parallel end faces that are approximately 250 µm square at both the front and rear faces. Typical cavity lengths range from ≈ 300 to 500 µm. Lasing action is achieved by applying a forward bias current which injects charge carriers (electrons or holes) across the p–n junction, and this in turn populates the nearly empty conduction band. Stimulated emission across the band gap between the conduction and nearly full valance band provides the gain mechanism for lasing action. Since the emission photon energy approximately equals the band gap energy, which in the case of Pb-salt materials is small, these lasers require cryogenic cooling to achieve population inversion. These lasers therefore are subject to temperature extremes (≈ 10–300 K) and this places stringent demands on the entire laser package.

448


The energy band gap, which is dependent upon semiconductor composition and crystal temperature, determines the output lasing wavelength. One can tailor the lasing wavelength regime at the time of manufacture by either varying the stoichiometry between Pb and the other constituents or by employing different alloys of Pb. Any given device can be actively tuned in wavelength over ≈ 100 cm−1 by changing the device temperature, or over tens of cm−1 by changing the injection current. Both tuning mechanisms, however, produce semicontinuous wavelength coverage, since the laser structure is a Fabry–Pérot device. Varying the injection current generally allows continuous tunability over a ≈ 1–2 cm−1 spectral region before the output jumps to a new longitudinal mode. In some cases the gain is broad enough to support multiple longitudinal modes simultaneously, resulting in wavelength regions where the lasing output gradually shifts from one mode to another. Tuning by temperature takes advantage of the change in band gap, and hence the wavelength with temperature. Typical rates range between 2 and 5 cm−1 / K. Since this mechanism involves changing the temperature of the entire laser package, including the stage on which the laser is mounted, temperature tuning is very slow on the order of seconds. Since typical absorption linewidths are in the 0.001 cm−1 range, at reduced pressures of several tens of Torr typically employed in sample cells, stable operation requires temperature control to better than 1 mK over long time periods. Tuning by current, which involves ohmic heating of the active region only, causes a change in the refractive index of this region, which in turn produces a change in wavelength. This tuning mechanism is very rapid, thus making it possible to employ high frequency modulation techniques in the kHz to MHz regime. Typical current tuning rates range between 0.02 and 0.07 cm−1 / mA (≈ 606– 2121 MHz/ mA). This rapid tuning mechanism, however, also requires low noise laser current controllers in order to avoid linewidth broadening. Given the typical absorption linewidths above, low noise controller operation of 10 µA or better is required. Employing such a low noise temperature and current controller, Reid et al. [3] determined that the measured linewidths from Pb-salt diode lasers varies dramatically from laser to laser, and for any given laser, depends strongly upon the junction temperature and injection current. Typical linewidths (FWHM), in the absence of refrigerator shocks caused by the closed cycle cooling system employed, were 0.6–25 MHz. Vibrations from the closed cycle cooler degraded this linewidth to ≈ 60 MHz. However, linewidths of ≈ 100 MHz or more were sometimes observed by Reid et al. and by Lundqvist et al. [4]. Sams and Fried [5] discussed the effect of such mechanical vibrations on quantitative spectroscopic determinations. Since most diode lasers now operate at or above liquid nitrogen temperatures, where liquid nitrogen dewars have replaced closed cycle refrigerator systems, one should expect linewidths ≤ 25 MHz. Older Pb-salt devices were grown by a diffusion process, which unfortunately often resulted in a poorly defined lasing region. In many instances


449

lasing occurred over nearly the entire junction width in multiple filaments, which in some cases produced a laser output in different directions with different wavelengths [6]. Even when such multiple beams were not present, the main emission lobe was sometimes emitted at an angle relative to the optic axis of the laser. The resulting poor spatial beam quality, which made it extremely difficult to optimally collect the output beam, together with other problems just discussed, has no doubt contributed to the poor performance reported by many early users of Pb-salt devices. This has led to the development of newer Pb-salt lasers based on mesa structure and ultimately buried heterostructure or double heterostructure devices in which the active lasing area is highly confined to a region less than 1 µm thick and less than 10 µm wide. These devices, which are discussed by Preier et al. [7], are prepared by molecular beam epitaxy using PbEuSeTe or PbSnTe active layers. In addition to improved output beam quality, which as we will show is very important for trace gas detection employing multi-pass absorption cells, these newer devices exhibit threshold currents as low as 1 mA and higher operating temperatures. Most of these devices typically operate at temperatures between 77 and 120 K, which is accessible using liquid nitrogen cryogenic dewars instead of bulky and noisy closed cycle refrigerator systems required by earlier devices. Despite this significant effort, Pb-salt diode lasers have not been manufactured to routinely operate CW at temperatures much higher than the above range and certainly less than the highest reported temperature of 215 K for CW operation published by Wall [8]. Typical CW single mode output powers for Pb-salt lasers are in the range 100–500 µW. For comprehensive reviews of Pb-salt diode lasers, their electrical and optical properties, performance characteristics and device materials along with the corresponding manufacturing techniques, the reader is referred to [7,8,9], and the many references found in Grisar [10]. Brassington also presents an excellent review of these topics as well as applications of Pb-salt lasers [11]. Unfortunately due to the small market for Pb-salt lasers, significant advances in device structure based on buried quantum well, distributed feedback and distributed Bragg reflectors, which appeared so promising in the 1990s, are not being aggressively pursued at present. Since lead-salt diode lasers can be modulated at very high frequencies (tens of kHz up to several hundred MHz), similar to those of near-infrared diode lasers, harmonic detection and two-tone modulation techniques can be employed as an efficient means of reducing noise (Sect. 1.2.2). As discussed previously, all CW Pb-salt diode lasers require some form of cryogenic cooling for operation. Although cryogenic operation is feasible and indeed commonplace in field environments, even in rugged airborne laser systems [12], such operation may impose limitations that can ultimately affect system performance. For example, all cryogenically cooled lasers are temperature cycled many times over their lifetime and may result in unrecoverable changes of the lasing frequency [13]. Other effects include long-term changes of tuning characteristics, slight changes of the

450


spatial mode quality, and reduction in laser power. Although all present Pbsalt devices are temperature-cycled by the manufacturer to minimize these problems, the finite possibility of such changes is disconcerting in certain demanding applications where it is imperative to access routinely a specific absorption feature. Airborne measurements of formaldehyde, where there is a limited choice of strong and interference-free absorption lines, is one such example [14,15]. In addition, since liquid nitrogen dewars frequently contain as many as four lasers, one runs the risk of mechanically disturbing all lasers in changing out any given laser. Small manipulations to the laser lead wires, which are unavoidable in this procedure, may result in irreversible changes to laser performance, similar to that from temperature cycling. Lead-salt diode lasers also exhibit large beam divergence and astigmatism, which places critical and stringent alignment requirements on the collection optics, particularly the first optical element, which is placed in front of the dewar. Subtle changes in the alignment between the laser and the first collection optic, due to small mechanical changes in the position of either the laser or the optical element, necessitates periodic adjustments to the first collection element, and this may add some mechanical instability to the alignment. Such instability, even when relatively small, in turn often leads to optical noise in IR absorption systems. This noise source, which is produced by unwanted scattering off various optical elements, results in a periodically undulating background structure of varying amplitude, spacing, and temporal frequency. Despite these drawbacks of lead-salt diode lasers, spectrometers employing these sources still yield excellent sensitivity, even in rugged field environments. As we will show in a later section, a liquid-nitrogen cooled Pbsalt diode laser spectrometer can routinely measure ambient formaldehyde (CH2 O) levels as low as 20–50 parts-per-trillion by volume (pptv) on an airborne platform employing 1-minute integration [16]. This corresponds to a minimum detectable absorbance, Amin , of 0.7 to 1.7 × 10−6 (S/N = 1) using a 100-meter absorption pathlength. Table 1. Summary of lead-salt diode laser characteristics Wavelength range ( µm)

Tuning (coarse/fine)

Power (mW)

Linewidth

3–30

100 cm−1 / 1–2 cm−1

0.1–0.5

1–1000 MHz – Elliptical – Highly astigmatic – Highly divergent

1.1.2

Beam profile characteristics

Operating requirements Cryogenic cooling

Antimonide Diode Lasers

Continuous wave lasing at room temperature at wavelengths above 2 µm with output optical powers up to 20 mW/ facet has been achieved using structures grown by molecular beam epitaxy (MBE) on GaSb substrates


451

and employing compressively strained GaInSbAs quantum wells (QWs) between Ga(Al)Sb(As) barriers in the active region. Narrow ridge Fabry–Pérot GaInSbAs/GaSb type II electrically pumped QW lasers emitting at 2.35 µm have been reported [17,18]. For further details on mid-infrared heterojunction lasers see the chapter by Joullié et al. These lasers emit in a fundamental spatial mode and exhibit single frequency operation over a range of currents and temperatures. They emit in a spectral region where overtone and combination absorption lines of such gases as CO, CH4 , NH3 and NO2 can be accessed conveniently. In efforts to extend coverage to the fundamental spectral region, several groups have reported the development of antimonide diode lasers in the 2–3 µm spectral range [19] and InAsSb/InAs lasers between 3 and 5 µm [20,21,22,23,24]. However, this technology is not yet robust and single frequency operation by means of DFB (Distributed Feedback) laser designs has not been realized. Hence these devices must be operated in an external cavity configuration to achieve stable and narrow-linewidth operation. Although room-temperature operation has been demonstrated, reliable single-frequency operation still requires cryogenic cooling. Table 2. Summary of antimonide diode laser characteristics Wavelength range ( µm)


Power (mW)

Linewidth Beam profile characteristics

Operating requirements

2–3; 3–5

1–2 cm−1

0.1 (singlemode)

50 MHz

– Thermoelectric (2–3 µm) – Cryogenic cooling (> 3 µm)

1.1.3

– Elliptical – Astigmatic – Highly divergent

Quantum Cascade Lasers

Quantum cascade (QC) lasers are unipolar semiconductor injection lasers based on intersubband transitions in a multiple quantum-well heterostructure. They are designed by means of band-structure engineering and grown by molecular beam epitaxy. The emission wavelength of a QC laser depends on the thickness of the quantum well and barrier layers of the active region rather than the band gap of diode lasers. These lasers operate either as CW or pulsed devices. The chapter by Hofstetter and Faist as well as several papers in [25,26,27,28] provide details of their design and operating characteristics. In the following, an overview will be given on emission wavelengths, output powers and design approaches. Quantum cascade lasers grown in a InGaAs/AlInAs lattice matched to the InP material system have been fabricated for emission wavelengths from 3.5 to 24 µm. Quantum cascade lasers have excellent spectroscopic properties in terms of optical power, but their tuning range is limited and their beam divergence is large and astigmatic. Multi-mode devices with 100 stages (quantum

452


well gain regions in series) have demonstrated peak powers of 0.6 W at room temperature. Until recently room temperature operation was only feasible for pulsed operation, but CW multi-spectral mode operation up to temperatures of 312 K was reported for a Fabry–Pérot style QC laser in 2002 [29]. Reliable single frequency operation has been achieved through the integration of a Bragg grating into the laser waveguide, resulting in a distributed feedback (DFB) laser, operating at cryogenic temperatures. The latest generation of QC-DFB lasers is based on a “top-grating” approach that takes advantage of the characteristics of a mid-infrared waveguide. For mid-infrared wavelengths below 15 µm, dielectric waveguides built from low-doped semiconductor layers that have appropriate refractive index modulation are used [30]. At longer wavelengths, the waveguide is overlaid with metal. In this case the radiation is guided not only by the dielectric but also by a surface plasmon mode [31]. Continuous wavelength tunability without mode hops is achieved through the temperature dependence of the waveguide parameters. The temperature can either be varied by a temperature change of the heat sink on which the device is mounted or more rapidly by dissipative heating through changing the direct QC laser excitation current. Characteristic total tuning ranges per current sweep are typically around 0.4 % of the emission wavelength. For many spectroscopic purposes, the spectral linewidth of the laser emission is as important as continuous tunability. The linewidth of selected CW DFB QC lasers ranges from a few MHz [32] through current stabilization to a few kHz with frequency stabilization [28], but exceeds 150 MHz (HWHM) in pulsed operation. Device reliability and long-term wavelength tuning characteristics are excellent as a result of using robust materials, such as InP and GaAs based heterostructures. To achieve mode-hop free tuning for Fabry– Pérot-type QC lasers in the mid-IR, a grating-coupled external cavity has been used to obtain a wavelength tuning of hundreds of nanometers, or up to 8 % of the central wavelength in the 3–4 µm region for InAs/InAsSb or GaSb/InAsSb heterostructure lasers with a few hundred milliwatts (mW) peak power [33]. Fabry–Pérot QC lasers [34] at 4.5 and 5.1 µm have been tuned with an external cavity. The principal technical issue is the need to deposit a low loss broadband antireflection coating or an angled surface on one of the laser output facets. A practical consideration of QC lasers is their operating current requirements, often drawing multiple amperes of current in CW operation. In addition, a QC laser typically requires compliance voltages of 5–10 V. The resulting thermal load to the laser is significant and good thermal management is important to achieve room-temperature operation. Low noise drivers have been developed based on the modified Libbrecht design [35]. The use of batteries also permits low noise operation of QC lasers and linewidths below 1 MHz are obtained without frequency locking. To date, QC laser-based chemical sensors primarily use InGaAs/InAlAs type-I QC-DFB devices. There are two limitations inherent to this kind of


453

Table 3. Summary of quantum cascade laser characteristics Wavelength range ( µm)


Power (mW)

Linewidth


4.3–24 (Selected regions where singlefrequency operation has been demonstrated)

35 cm−1 with external grating cavity/ 3 cm−1

1–100 (CW, single frequency) 50 (avg., pulsed)

0.001– – Elliptical 10 MHz, – Astigmatic CW – Highly (locked– divergent unlocked) > 150 MHz (HWHM), pulsed

Operating requirements – Cryogenic cooling (CW) – Peltier cooling (pulsed) – High voltage, high current low noise driving electronics

laser for chemical sensing. First, they cannot access the spectral region of C–H and O–H stretch vibrations near 3000 cm−1 . This shortcoming can be overcome by developing QC lasers based on alternative materials and structures. For example, the 3000 cm−1 region is accessible by type-II lasers [36], but no single frequency devices of this kind have yet been demonstrated. Another issue is the limited tunability of each QC-DFB laser, which restricts the feasibility of multi-component chemical sensing. This requirement can be addressed by separating the gain medium from the wavelength-selective element [34,37]. In [34], a QC laser tunability of ≈ 35 cm−1 at a fixed temperature was demonstrated in an external cavity configuration with a diffraction grating. This is about a ten times wider range than that typically achieved for QC-DFB lasers by means of current tuning. On-line concentration measurements of ≈ 20 gaseous compounds and several isotopomers in ambient air have been realized in the first three years of various QC laser based chemical sensors. Mid-infrared semiconductor laser sources as discussed above share several common characteristics. Low temperature operation is yet the most reliable means of obtaining CW tunable single frequency emission. For the purpose of applying the generated radiation to a specific spectroscopic technique and application, the laser radiation has to be collected and mode matched to the spectroscopic sampling path or cell. Semiconductor lasers exhibit a large divergence and astigmatism and four representative approaches for beam collection, shaping and delivery to the spectroscopic sampling path are illustrated in Fig. 2. Illustrated in Fig. 2a is the approach used by Fried et al. [16]. Here, all reflective elements are used to collect, collimate and image the laser radiation to a long-path absorption cell. Fig. 2b shows the approach used by Nelson [38], in which an all-reflective objective is used for the same purpose. However, collection efficiency is typically lower than the technique depicted in Fig. 2a due to center mirror obscuration by as much as 20 %. Care must be taken to avoid residual feedback of the center mirror, which may increase the laser noise and at times induce mode-hopping. In approach 2c (Laser Components, Inc.), a toroidal mirror is used as a first collection element to correct the astigmatism/emission aspect ratio and lower the divergence of

454


Fig. 2. Representative approaches for optical beam collection, shaping and imaging of highly divergent, astigmatic laser radiation. See text for explanation of figures. OAP, off-axis parabolic mirror; OAE, off-axis ellipsoidal mirror; AP, aperture; RO, reflecting objective; M, mirror; OAT, off-axis toroidal mirror; L, lens

the beam. An OAE re-images the beam to the sampling path or cell. This design provides a well-defined beam shape and quality, however at the cost of complexity and the use of four optical elements, all subject to drift, which can lead to beam pointing instabilities. The advantage of this configuration is that the beam can be optimally collected while maintaining a fixed output axis. The approach shown in Fig. 2d uses two lenses to collimate and reimage the laser [39], avoiding the spherical aberrations introduced by OAP and OAEs. Similar to the approach shown in Fig. 2b, feedback may be an issue to the laser wavelength stability and is subject to creating stronger optical interference fringes between refractive optical elements. 1.1.4

Tunable Solid-State Lasers

A large and important class of tunable lasers is based on the vibronically broadened transitions that can occur in certain gain media, such as color centers and certain transition metal or rare-earth ions in crystalline hosts [40,41,42,43]. When such a medium is placed in a tunable cavity and pumped above laser threshold, stimulated emission can be made to occur at any desired frequency within the emission band. Tunable laser media based on 3d–3d transitions of transition-metal ions and 4f–5d transitions of rare-earth ions cover the mid-infrared spectral range between 1.0 µm and 4.7 µm. The tuning range of such lasers can be widely varied by the choice of impurities and by selecting different hosts. Recent spectroscopic studies demonstrated that chromium-doped zinc selenide chalcogenides, such as Cr2+ :ZnSe and Cr2+ :ZnS, have favorable characteristics as tunable mid-infrared solid-state materials near 2.5 µm [44,45]. These include room-temperature operation,


455

broad tunability, the possibility of direct diode-pumping, erbium fiber amplifier and CW operation. Table 4. Summary of tunable solid-state laser characteristics Wavelength range ( µm)


Power (mW)

Linewidth

2–5

600 nm/ 1 cm−1

200–1000 → 800 MHz, 1 s 20 → 0.1 MHz, 1 s



– TEM00 Low technical – Cavity subject noise to pointing environment instability

For selective spectroscopic detection at reduced pressures, the output power of the above mentioned solid-state laser sources is significantly reduced if one or more frequency selective optical elements are employed to obtain laser linewidths of less than 10 MHz. For further details see the chapter by Sorokina on mid-IR crystalline lasers and references therein. 1.2

Class “B” Laser Sources

Class “B” laser sources, as defined earlier, are based on parametric frequency conversion of near-IR laser source(s). These can be configured either in resonant (cavity) arrangements with a single pump laser or non-resonant (singlepass) arrangements with two pump lasers and are referred to as optical parametric oscillator (OPO) [46] and difference frequency generation (DFG) based sources, respectively. Figure 3 illustrates these two concepts.

Fig. 3. Shown are two representative examples of DFG and OPO. DFG based mixing uses single pass parametric interaction and requires spatial overlap of the input sources via discrete or optical fiber components. For wavelength tuning, the OPO ring-cavity is wavelength tuned with a piezo driven cavity mirror M. A frequency selector, e.g. étalon, is synchronously tuned to maintain single-mode operation [47]. M, mirror; ET, étalon; L, lens; PPLN, periodically poled lithium niobate; SM, semi-transparent mirror; BS, beamsplitter/combiner

456


Before discussing the two approaches of parametric frequency conversion, we will discuss a variety of commonly used room-temperature near-IR laser sources that can serve as effective pump sources, followed by a tabulated overview of these sources and their performance characteristics. These laser sources are not emitting mid-IR radiation but can operate at wavelengths ranging from the visible to near-infrared region. However, used in combination with frequency conversion, their spectroscopic characteristics (linewidth, tuning range, etc.) are partially (OPO) or completely transferred to the mid-IR. 1.2.1 Near-IR Pump Laser Sources for Parametric Frequency Conversion Many of the problems in tuning and wavelength stability of Fabry–Pérot type diode lasers can be enormously reduced by incorporating an external or internal grating structure to provide more defined feedback [48]. In an ECDL (External Cavity Diode Laser) , one or both faces of the laser chip are antireflection-coated to eliminate optical feedback. Instead, the feedback required for laser action is provided by an external cavity. The cavity acts as a narrow wavelength selector, which determines a specific operating wavelength out of the usually broad gain spectrum of the semiconductor laser material. Several cavity configurations have been developed that differ in the method of tuning, number of components, output beam characteristics, and output coupling efficiency. Mode-hop-free single frequency tuning ranges of over 1000 GHz have been demonstrated for an ECDL [49,50,51]. Gratingtuned external cavity diode lasers with large tuning ranges have been commercialized in the near-infrared from 0.9 to 1.6 µm. All ECDL designs have been based on tuning a frequency selector by mechanical means. In environments with technical noise such as vibrations this can lead to additional frequency jitter. More recent designs have incorporated Micro Electro Mechanical Systems (MEMS) technologies, which dramatically reduce the size (5 mm × 5 mm) of the external cavity [52]. Implementation of integrated wavelength mode filters in the form of distributed feedback (DFB) structures or distributed Bragg reflectors (DBR) also enable single frequency operation and have been developed for the optical telecommunication industry at wavelengths ranging from 1.3 to 2 µm. Other novel developments include broadly tunable monolithic integrated multisection diode laser chips employing gain, filter and Bragg tuning elements [53]. Unlike external grating controlled diode lasers, these lasers [54,55,56] offer fast and versatile electronic coarse wavelength tunability (≈ 70 nm at 1.56 µm) of potential pump sources for nonlinear optical frequency conversion devices. Other forms of single frequency near-IR diode laser include VCSEL (Vertical Cavity Surface Emitting Laser), which are simpler and more cost effective to produce, but have not yet reached comparable output powers and IR wavelength coverage. VCSELs can tune over ≈ 30 cm−1 by changing their operating current (threshold-maximum). Such tuning rates are attractive for


457

certain applications such as rapid combustion diagnostic studies, but not desirable for precision gas sensing. Grating stabilized near-IR laser sources have also been developed in the form of rare-earth doped single mode DFB fibers in the 1 µm (Yb) and 1.5 µm (Er) wavelength region. The operating wavelength within the dopant gain region can be precisely engineered to ≈ 0.1 nm accuracy by writing a DFB grating via photolithography into a photosensitive doped fiber region. Fiber lasers of this type can be wavelength tuned by means of temperature and straining the DFB fiber length with a piezoelectric element [57]. Ultranarrow linewidths on the order of a few kHz have been achieved without the use of frequency locking techniques and hence offer convenient highresolution wavelength tuning capability. In addition DFB fiber lasers offer low relative intensity noise and high side-mode suppression ratios of > 60 dB [http://www.koheras.com] [58]. Another important solid-state laser class demonstrated for operation in the near-IR wavelength region is the monolithic non-planar ring oscillator based on Nd, Yb or Tm:YAG materials. This type of laser offers single-frequency CW or pulsed output with diffraction-limited beam quality. Narrow-linewidth CW powers exceeding several watts have been demonstrated [59]. This laser source is specifically suited and used for pumping optical parametric oscillators. For further details see the chapters by Vodopyanov and Ebrahimzadeh. 1.2.2

Sources Based on Difference Frequency Generation (DFG)

Numerous DFG based mid-IR sources have been designed and used for spectroscopy [60,61,62,63,64,65,66,67,68,69,70,71,72]. For further details see also the chapter by Fischer and Sigrist. In the case of DFG, two laser beams (“pump” and “signal”) at different frequencies combined in a nonlinear material with suitable dispersion characteristics generate a beam at the differencefrequency (“idler”). The narrow emission spectra of the “pump” (highest frequency) and “signal” (middle frequency) are convolved during the frequency conversion and hence translate into a similarly narrow spectrum of the idler wave. Idler wavelength tuning is accomplished by tuning of the pump laser, or signal laser, or both. In order that the idler wave continue to build up as the beams pass collinearly through the nonlinear material, the three waves must stay in phase (the “phase matching condition”). This imposes a condition on the refractive indices of the three waves. This condition can often be satisfied with a birefringent nonlinear crystal by having some of the three waves polarized along an ordinary axis and some polarized along a direction that includes the extraordinary axis. If the polarization direction that includes the extraordinary axis is not parallel to it (angle tuning), the three waves will not propagate in the same direction (double refraction) and the beams will separate as they pass through the crystal (“walk-off”) limiting the overlap region and the DFG power. In order

458


Table 5. Summary of tunable near-IR diode, fiber laser and solid-state source characteristics Wavelength range ( µm)

Tuning Power (coarse/fine) (mW)



ECDL

0.650–1.655

10–100 nm/ 1 cm−1 (by Piezo)

1–50

1–5 MHz, 1s

– Elliptical – Astigmatic – Beam pointing instability with tuning (Littrow design)

Low technical noise environment for best performance. Vibrations lead to increased linewidth of > 100 MHz

SGDBRDL

1.55

200 cm−1 / 1 cm−1

2–10

25 MHz, 1s

– Single mode – Telecom fiber environment diffraction – Requires limited complex Gaussian electronics beam quality for tuning – Inherent beam pointing stability

DFBDL

1.3–2

3 nm/ 1 cm−1

2–30

0.1– 1 MHz, 1s

– Single mode – Telecom fiber environment diffraction limited Gaussian beam quality – Inherent beam pointing stability

DFB Fiber Laser

1.03–1.2 (Yb) 1.528–1.61 (Er)

5 cm−1 (Temp.)/ 10 cm−1 (PZT)

2–10

0.1 MHz, 1s

– Single mode – Telecom fiber environment diffraction limited Gaussian beam quality – Inherent beam pointing stability

SolidState

0.946, 1.064, 1.319, 1.444, (Nd:YAG) 2.02 (Tm:YAG)

30 GHz (Temp.)/ 100 MHz (piezo)

150– 2000 (CW)

0.02 MHz, 1s

– TEM00 (M2 < 1.1)

– Room temperature

ECDL, External Cavity Diode Laser ; SG-DBR-DL, Sampled Grating Distributed Bragg Reflector Diode Laser; DFB, Distributed Feedback Diode Laser

to satisfy phase matching keeping all waves exactly parallel or perpendicular to the optic axis (“90◦ phase matching”), the refractive indices must be tuned as the difference frequency is tuned by varying the temperature of the crystal or by tuning pump and signal simultaneously. In the early demonstration of the DFG method by Pine [73], single mode argon-ion and dye laser outputs were combined in bulk lithium niobate crystal to produce narrow-band (15 MHz) radiation tunable from 2.2 to 4.2 µm by temperature tuning the crystal. Simultaneous tuning of both signal and pump


459

has been used in AgGaS2 [64] to provide tunable single frequency radiation from 3.5 to 9 µm. For DFG radiation longer than 5 µm it is also possible to use birefringent bulk nonlinear optical materials, such as AgGaSe2 [74], ZnGeP2 [75], or GaSe [76,77,78]. Another approach to phase matching is the introduction of periodic short (about 10–30 µm wide) regions in which the sign of the second order susceptibility alternates, thus bringing the three waves back into the right phase relationship. This is called quasi-phase-matching (QPM). It is most easily achieved in ferro-electric materials, where the direction of the extraordinary axis can be permanently reversed locally by the application of an external electric field (≈ 20 kV) at elevated temperatures [79]. Other advantages of QPM are no walk-off effects, access to the crystal’s largest diagonal nonlinear coefficient (not accessible by birefringent phase-matched crystals), wide acceptance bandwidth, and relative ease of alignment. The implementation of diode-pumped mid-infrared frequency conversion sources received a significant boost from the development of novel periodically poled nonlinear materials, such as lithium niobate (PPLN) [80], lithium tantalite (LiTaO3 ), and ferroelectric crystals of the potassium titanyl phosphate (KTiOPO4, or KTP) family at wavelengths in the 2.5–5.2 µm spectral region [81,82,83]. The quasi-phase-matching properties of each of these crystals can be engineered for interaction of any pump and signal wavelengths within the transparency range of the crystal, allowing significant flexibility in the choice of laser sources for frequency mixing [84,85,86]. In the future, quasi-phase matched GaAs [87] should become available greatly extending the long wavelength region covered by DFG. In the work reported by Eyres et al. orientation-patterned GaAs films of 200 µm thickness have been grown by hydride vapor phase epitaxy (HVPE) on an orientation-patterned template fabricated by molecular beam epitaxy (MBE). The availability of PPLN permits near-infrared diode or fiber lasers to be used as pump lasers [68,69,70,71,88,89,90,91,92] instead of much larger dye or Ti:sapphire lasers making it feasible to construct compact mid-infrared spectrometers that operate at room temperature and can generate CW output powers up to 1 mW [93]. Thus the practicality of near-infrared diode laser and optical fiber technology are combined to achieve the analytical power of midinfrared spectroscopy in a single instrument. Such an instrument inherits the single-frequency operation and high modulation speed capabilities of diode lasers, and takes advantage of their relatively wide tuning range. For example, a typical 780-nm diode laser can be grating tuned over 20 nm, or 2.6 % in wavelength without appreciable changes in output power. When the output of such a laser is down-converted by mixing with a 980 nm diode laser, the tuning range in frequency units remains the same, in this case a significant tuning range: 3.6–4.1 µm, or 13 % in wavelength. A detailed quantitative theory of this nonlinear optical process is beyond the scope of this review chapter. Instead, the reader is referred to a paper by

460


Zondy [94] and the references contained therein. Therefore, we will simply state that the maximum idler power generated in a given crystal is proportional to the product of crystal length, pump power, signal power, and the square of the second-order nonlinear coefficient of the crystal. Optimum DFG output power is achieved by means of precise spatial overlap and focusing of the pump and signal beams. There is an optimal focus because a point is reached at which any further increase in beam intensity through tighter focusing is offset by a decrease in interaction length due to diffraction, resulting in loss of output power and in some cases clipping at the crystal aperture. Although this type of source is routinely used for spectroscopy and gas detection, DFG in bulk nonlinear (either birefringent or quasi-phase matched) crystals is characterized by low conversion efficiency, typically [95] in the range 0.002–0.05 % W−1 cm−1 . The tradeoff between beam size and interaction length can be eliminated in guided-wave DFG. Optical confinement of pump and signal radiation near the waveguide core creates a region of high intensity and good modal overlap that can be maintained throughout the length of the waveguide. Thus the interaction length for tight focusing is now limited by the length of the waveguide, not by diffraction. Guided-wave parametric processes, such as OPO, SHG, and DFG, have been demonstrated [96] in periodically poled LiNbO3 , LiTaO3 , and KTP. In LiNbO3 , for example, a waveguide can be formed by titanium in-diffusion, or by a Li+ – H+ ion exchange typically followed by several hours of annealing at elevated temperature to create a graded index distribution. A DFG waveguide designed to carry a single spatial mode at the idler wavelength is necessarily multimode at the shorter, pump and signal wavelengths. The presence of multiple spatial modes complicates waveguide phase matching characteristics. For example, a TEM00 (fundamental) mode at the signal wavelength will interact with TEM02 and TEM10 modes at the pump wavelength, but not with TEM01 or TEM11 modes. Efficient and reproducible fundamental-mode excitation of a DFG waveguide was first achieved by Chou et al. [96,97] using a combination of a mode filter and an adiabatic taper. An improved device featuring separate inputs for the pump and signal beams followed by a directional coupler has also been demonstrated. DFG waveguides have been used to build sources of mid-infrared radiation for spectroscopic purposes [98,99,100]. Surprisingly, though, reported idler power from waveguide-DFG have not exceeded powers of more than 0.1 mW, whereas with bulk QPM DFG crystals, powers exceeding 1 mW have been obtained through use of high input power sources. Difficulties in efficiently coupling the pump laser power into the waveguide and maintaining the waveguide properties in the presence of higher power fields have so far prevented the generation of higher mid-IR power levels that exceed those obtained with bulk QPM DFG crystals. In addition, waveguide-QPM structures have relatively narrow conversion bandwidths, whereas bulk-QPM can be poled to


461

have either multiple grating periods or a fan-out structure for continuous phasematching over 750 cm−1 [69]. Table 6. Summary of difference frequency generation based sources Wavelength range ( µm)


Power (mW)

2.3–4.6 (based on available near-IR telecom laser wavelengths)

15 cm−1 per diode laser pair;

0.1–1 mW 1 MHz

1.2.3

wavelength multiplexing possible/ 2 cm−1



– Near Gaussian Telecom (fiber pump environment beam delivery only) – Beam pointing stability – large f/# ≈ 100

Tunable Optical Parametric Oscillators

Optical parametric oscillators (OPOs) are progressing as useful spectroscopic tools for the generation of coherent radiation that is continuously tunable over large spectral ranges [101,102,103,104,105,106,107]. For further details see the chapters by Vodopyanov and Ebrahimzadeh. Unlike DFG, the nonlinear crystal is placed in a cavity and is used to generate output beams at two new frequencies (signal and idler) ν1 and ν2 from a single pump beam at ν3 (Fig. 3). Energy conservation requires ν1 +ν2 = ν3 . How the frequency is divided between the new waves, signal and idler is determined by the phasematching condition. The development of pulsed OPOs is mature, and these devices are available commercially. On the other hand, CW OPOs still have some practical problems in terms of requiring high power pump sources, efficient and high quality nonlinear crystals, low loss broadband optics, and mode-hop-free operation with good frequency stability in order to realize their potential usefulness in mid-IR spectroscopic power. OPO devices are arranged so that the signal frequency resonates inside the cavity (singly resonant OPO), or in addition the idler can also be made resonant (doubly resonant OPO). In some special cases the pump frequency can be made resonant as well (triply resonant OPO). In principle, both doubly resonant and singly resonant OPO configurations can be used although as a practical matter doubly resonant OPOs are difficult to construct and cumbersome to tune. As in DFG devices quasi-phase-matching in periodically poled ferroelectric crystals offers several distinct advantages for their use in CW OPOs, such as non-critical phase matching and a high effective nonlinear coefficient, deff . A particularly significant development was the demonstration of the use of PPLN as the parametric gain medium, in which case the oscillation threshold of externally pumped CW singly resonant OPOs can be reduced to the few watt level hence making it feasible to use diode pumped solid-state pump

462


lasers [101,102,103,104,105,106,108,109,110,111,112,113,114]. Pump powers as low as 800 mW were used to pump a PPLN singly resonant OPO in which both the pump and signal were resonated [115]. The parametric process can also be used in optical parametric amplifiers (OPAs) to boost infrared output powers. Continuous wave OPOs amplified by pulsed OPAs offer a competitive alternative to other tunable lasers in the 1–5 µm spectral region in terms of linewidths, wavelength tunability and output powers. Table 7. Summary of OPO (PPLN based) characteristics [116] Wavelength range ( µm)


Power (mW)

Linewidth



1.45–2 2.3–4

1900 cm−1 / 0.05 cm−1

10– 100 mW

0.15 MHz, instantaneous

TEM00

Low technical noise environment

2 Fundamentals of Absorption Spectroscopy for Trace Gas Detection Spectroscopic trace gas detection is a method allowing one to determine the concentration of a known gas, or gases, from a measured optical absorption spectrum of the gas mixture (in practice, a small fragment of the spectrum is measured). The procedure requires a good quantitative knowledge of the gas absorption characteristics. This knowledge is the realm of molecular spectroscopy, a complex and highly developed subject. A few fundamental spectroscopic concepts and formulae that are directly applicable to gas detection are, however, summarized in this section. Each atom or molecule, small or large, is uniquely characterized by a set of energy levels. Transitions between levels by absorption or emission of electromagnetic radiation result in highly specific spectroscopic features. These features allow both the identification and quantification of the molecular species, such as atmospheric trace gases. Molecules may undergo transitions between electronic, vibrational, and rotational states when exposed to electromagnetic radiation, resulting in absorption spectra. These spectra consist of a number of discrete absorption lines. Each line will have a certain linewidth and shape that depends on temperature and what surrounds the molecule. The lines may in some cases be resolved and in other cases the line density may be too high to be spectrally resolved. Transitions between molecular rotationalvibrational (“ro-vibrational”) states occur in the infrared “fingerprint” region of the electromagnetic spectrum, approximately between the wavelengths of 2.5 and 25 µm. Also, overtone and combination-overtone ro-vibrational lines are possible with significantly lower intensities as compared to those for fundamental vibrational bands and the corresponding wavelengths are in the

Mid-Infrared Laser Applications in Spectroscopy CH2O

463

HCN

N2O

CO2

CO N2O

HCl

CH4

L=0.5 m; p=40 Torr; Resolution=2.4 cm-1 All species plotted at 250 ppm relative concentration

Fig. 4. HITRAN simulation of absorption bands of various molecules in the 3–5 µm spectral region. All species are plotted with identical relative concentration. Spectral overlap limits the choices of interference free absorption lines

0.8–2.5 µm spectral region. Transitions between electronic states of atoms and molecules occur in the ultraviolet and visible spectral region. All polyatomic molecules, with the exception of homonuclear diatomic molecules (e.g. N2 ), absorb infrared radiation. The absorption changes the state of molecular rotation and vibration. An absorption spectrum therefore depends on the physical properties of the molecule such as size and shape and hence each molecule is characterized by a unique spectral “signature”. Spectra of linear and some nonlinear polyatomic molecules consist of an array of individual or small groups of lines. In the case of large polyatomic molecules (e.g. benzene, C6 H6 ) at atmospheric pressure, there are many lines overlapping each other, resulting in broad spectral features with some occasional peaks. There are numerous atmospheric trace gases and their concentrations are normally in the parts-per-trillion (pptv, 10−12 ), parts-per-billion (ppbv, 10−9 ) to parts-per-million (ppmv, 10−6 ) range. However, species such as water vapor may have concentrations of up to a few percent (%, 10−2 ). Because of this, even weak water features, where absorption cross-sections are as much as a factor of 10−10 weaker than the molecule of interest, can be a problem. There is much spectroscopic data available in the literature and in electronic form which are important tools in the identification and development of specific detection strategies, especially in the presence of interfering species [117,118]. Direct gas phase laser absorption spectroscopy based on the Beer– Lambert absorption law is often used for quantitative measurements. In the absence of optical saturation and particulate-related scattering, the intensity of light I(x) propagating in a homogeneous gas of sample length L

464


Fig. 5. Illustration of the Beer–Lambert absorption law

follows the Beer–Lambert law: I(x) = I0 exp [−σ(ν)N L] .

(1)

Here N represents the molecular concentration and σ(ν) the absorption crosssection. The molecular absorption cross-section depends on frequency and has units of cm2 cm−1 per molecule when integrated over the absorption line, and units of cm2 per molecule at the line center. For simplicity we assume only one absorbing species. The peak absorption cross-section at line center (ν0 ) is related to the integrated line strength through a lineshape function Γ (ν). This function Γ (ν) has the same analytical form for all transitions, and in mid-infrared spectroscopy, the broadening of an individual transition due to finite upper-level lifetime is insignificant compared to broadening by the other two important mechanisms – thermal motion and molecular collisions. Their individual and combined effects on a molecular transition at a frequency νn are expressed as follows: Thermal motion (Gaussian): 2 0 ln 2 − ν−ν 1 ln 2 γD e ; ΓD (ν) = γD π

γD = 3.58 × 10

−7

ν0

T , M

(2)

Molecular collisions (Lorentzian): ΓL (ν) =

γL 1 ; π (ν0 − ν)2 + γL2

γL = γL0 p

T , T0

(3)

Combined broadening (Voigt): y Γ (ν) = k(ν0 )D π with, x=

ν − ν0 γD

∞ −∞

2

e−t dt ; y 2 + (x − t)2

√ γL √ ln 2 ; y = ln 2 ; t = γD

k(ν0 )D =

δ γD

√

σ(ν)dν γD

ln 2 , π

(4)

ln 2 .

Here T is the gas temperature (K), M the molecular weight, P the gas pressure (atm), and γL0 the coefficient of pressure broadening (cm−1 atm−1 ), k(ν0 )D is the peak Doppler cross-section, δ is the parameter of integration


465

and is used to express the Doppler and Lorentz frequency differences during the convolution process in terms of a single variable. The quantities γD and γL are referred to as the Doppler- and pressure-broadened half width at half maximum (HWHM) linewidths. The line shape that results from the combined effect of Doppler- and pressure-broadening is a convolution of the two respective line shapes, and it is known as the Voigt profile. The physical significance of the convolution is that the Voigt profile has different asymptotic shapes for very low and very high gas pressure. At low pressure, molecular collisions are less frequent, leaving thermal motion the dominant broadening mechanism – the corresponding line shape is near-Gaussian. As the gas pressure increases the collisions take over, and the resulting line shape is near-Lorentzian (Fig. 6). The previous expressions do not include the effect of pressure shift, which is typically in the range of several megahertz per atmosphere. The shift is very small compared to the width of an atmospheric-pressurebroadened line, typically several gigahertz. It can be verified by integrating the absorption cross-section of an individual transition over frequency, that the shift is independent of the broadening mechanism and is equal to the line intensity S, in units of cm2 cm−1 per molecule. The line intensity is proportional to the lower-state population density of a transition and thus depends on temperature. These parameters have been measured and calculated for many lightweight gas molecules in the mid-infrared 100.00

Transmission (%)

99.95

99.90

A B C CH2O

H2O

99.85

99.80

CH2O A: p=1 Torr, 50 ppm CH2O, 40 % H2O B: p=40 Torr, 5 ppm CH2O, 4 % H2O C: p=200 Torr, 5 ppm CH2O, 4 % H2O

99.75

99.70 2831.5

2831.6

2831.7

2831.8

2831.9

2832.0

-1

Frequency (cm )

Fig. 6. Computed mid-IR absorption spectra of CH2 O and H2 O at 1, 40 and 200 Torr. Lineshapes correspond to near-Gaussian (A), Voigt (B) and nearLorentzian (C). Also note the higher relative concentration at lower pressures to obtain a comparable absorption strength. Optimum sampling pressure with good signal strength and selectivity ranges typically between 30 and 60 Torr

466


spectrum, and compiled into extensive databases such as HITRAN [117], GEISA [118], NIST (http://ois.nist.gov/srmcatalog/datafiles), and PNNL (http://nwir.pnl.gov). Numerically accurate absorption spectra can be computed based on these data, not only for single gas species but for gas mixtures as well. The analytical formulae apply also to multi-component gas mixtures. The total absorption cross-section σ(ν) is then a weighted average of absorption cross-sections of individual species, with the mole fraction Cm of each species used as the weight coefficient: Cm σm (ν) , C : m = 1. (5) σ(ν) = m

m

For each of the m species, the pressure broadening coefficients γL0 generally depend on the transition. They also depend on the type of molecule with which the collisions occur. In general, partial pressures in conjunction with the appropriate pressure-broadening coefficients should be used to compute the overall pressure broadening from all gases present in the background (this includes self-broadening). Air-broadening coefficients are useful in calculations, and are listed in spectroscopic databases [117,118]. In trace gas sensing applications, however, the species of interest are often present in very low concentrations, so that self-broadening and broadening against other trace gases can be neglected in calculations, and air-broadening alone will suffice. For the conditions of atmospheric pressure broadening, γL0 P γD , the Doppler contribution to the overall linewidth can often be neglected, and the line shape be treated as pure Lorentzian. Likewise, at pressures low enough to ensure γL0 P γD , the line shape can be treated as pure Gaussian. In either case, calculation of the line profile is simplified considerably. At intermediate total pressures, γL0 P ≈ γD , which for most lightweight gases range from 5 to 100 Torr, calculation of the Voigt profile is necessary to obtain numerically accurate absorption spectra. Methods for approximate calculation of the Voigt profile, and the related plasma dispersion function, are now a well-developed subject. The approximations published in [119] are particularly useful.

3 Spectroscopic Techniques: Signal Enhancement and Noise Reduction In direct absorption approaches, quantitative information can be obtained using the expressions discussed in the previous section. This section only discusses in situ techniques, where the source, sampling region and the detector are in close proximity. Active remote sensing such as differential optical absorption spectroscopy (DOAS) and light detection and ranging (LIDAR) are well developed, but not covered here and the interested


467

reader is referred to the following recent book chapters by Platt [120] and by Svanberg [121]. For in situ measurements, various sensitivity enhancement and noise reduction spectroscopic detection techniques have been developed in order to achieve quantification of trace gas species at concentrations ranging from ppmv to pptv. Each detection technique has its distinct advantages and should be chosen depending on the specific application targeted, in particular in terms of sensitivity and selectivity requirements. Although laser source noise is an important aspect for sensitive detection, in practice various other noise sources affect the measurement of a small change of signal and can severely limit the detection sensitivity. Various noise reduction techniques can be employed. These include modulation, balanced beam and zero-background subtraction detection techniques. In addition, one can improve the sensitivity by signal enhancement methods based on long pathlength and cavity enhanced spectroscopy (Sects. 3.3 and 3.4), which are capable of increasing significantly the effective sample optical pathlength L to tens of meters and to kilometers, respectively, in absorption cells of typical physical lengths of 0.3 to 1 m. It is difficult to compare the expressions of sensitivity in terms of trace gas detection systems that employ different sources and detection techniques and their significance for a specific application. In this context, different expressions of sensitivities reported in the literature are a relative statement and can be at times misleading if applied or compared to a different signal enhancement or noise reduction technique. Perhaps the most appropriate method to compare the performance of trace gas detection systems (laser and even non-laser based) is to determine the minimum detectable concentration for the target molecule of interest for a given sampling and acquisition time. For comparisons with other techniques one can relate this to a minimum detectable absorbance per pathlength for the pathlength conditions employed in the concentration measurement. We make this distinction since, for example, in photoacoustic spectroscopy one obtains very high sensitivity for unit pathlength, however it does not directly scale with increasing pathlength. In many cases, such as isotopic ratio measurements, the replication precision is most important. These and other attributes such as wavelength dependent absorption strengths should be given detailed consideration before selecting a laser source and technique for a specific trace gas sensor. Table 8 gives an overview of the most commonly used expressions and depicts examples of values achieved and reported in the literature. 3.1 Balanced Beam and Balanced Ratiometric Detection (Noise Reduction) These techniques have been developed in order to eliminate technical noise including laser intensity noise to approach the fundamental limit of shotnoise. By measuring the laser signal with and without the absorption signal

468


Table 8. Expressions of detection limits and sensitivity Parameter

Expression

Common values/units

Minimum detectable fractional absorption

∆ Pmin P0

10−4 –10−7

Minimum detectable fractional absorption scaled to path length

∆ Pmin 1 P0 L

10−8 –10−12 cm−1

Minimum relative Detectable Concentration per unit volume (MDC)

∆ Pmin 1 1 P0 L σNtot

ppm (1 part in 1 million) 10−6 ppb (1 part in 1 billion) 10−9 ppt (1 part in 1 trillion) 10−12

Minimum detectable fractional absorption scaled to path length, relative concentration, and per shot measurement time

∆ Pmin 1 1 √1 P0 L σNtot BW

Absolute Laser Instrument Response Factor (LIRF)

with Tmeas,total

Measurement precision (measured with a stable input concentration)

Std.Dev. (LIRF) or Std.Dev. (MDC)

∆ Pmin 1 P0 L

10−8 –10−12 cm−1 Hz−1/2 #pts BW = Tsample or ENBW n n Bandwidth(BW) without or with a frequency selective filter (e.g. lock-in amplifier) 10−8 cm−1 (1 s)–10−11 cm−1 (60 s)

Legend in order of appearance P , optical power; L, length of effective light–matter interaction; σ, molecular absorption cross-section [cm2 ]; Ntot , total number of molecules per unit volume; #pts, number of acquired points per scan; Tsample , time for single scan; n, number of acquired scans; ENBW, equivalent noise bandwidth of a frequency selective filter; Tmeas,total , measurement time to generate one concentration data point

simultaneously, common mode noise can be subtracted and small absorption signals can be recovered. Several approaches using dual-beam detection have appeared in the literature and are briefly discussed here. Conventional dual-beam detection systems use optical balancing schemes [122]. The detected noise of an equal-intensity replica of a probe beam, such as that created by a variable-ratio beamsplitter, is subtracted from noise detected in the probe beam and thus leaving only the uncompensated weak absorption signals of interest. For example, such a beamsplitter can be realized by placing a polarization rotator (a half-wave plate) in series with a polarizing beamsplitter cube. With the input polarization ro-


469

tated about 45◦ , the beams emerging from the beamsplitter cube carry equal amounts of power P , and power noise ∆P . In the absence of absorption, the photocurrents generated by identical signal and reference detectors can be subtracted to cancel each other exactly. If one of the beams is attenuated due to small absorption a, by a gas, the balance of photocurrents is disturbed, and a signal is seen at the output of the amplifier. Care must be exercised to ensure that the signal and reference detectors have equal amplitude and frequency responses. An implementation of this method that avoids the need for exact balancing the signal and reference light was proposed by Hobbs [123]. It is known as balanced ratiometric detection (BRD). It employs electronic circuitry to produce a log ratio of photocurrents, rather than their difference, and to cancel noise currents at the same time. This analog divider uses logarithmic conformance and tight symmetry of base-emitter curves of a matched transistor pair. This scheme provides nearly perfect cancellation of noise currents even when the reference beam carries twice the power of the signal beam. Since the signal versus reference current balancing is performed by means of electronic feedback, no physical adjustment of the beam splitting ratio is necessary. The BRD differential response to absorption signals depends on the ratio of the signal and reference currents, which changes when the signal beam is partially absorbed. It also depends on temperature because the transistor base-emitter voltage does, and additional compensation circuitry is needed to produce a useful output voltage Vout that is linearly proportional to the absorbance. Noise-equivalent absorbances in the near-infrared spectral region as low as 2 × 10−7 Hz−1/2 have been demonstrated by Allen and co-workers [124], close to the limit imposed by shot noise. Application of this concept to the mid-IR spectral region using a quasi-CW QC laser was recently demonstrated by Sonnenfroh et al. [125] However, the sensitivity was reduced due to inherent differences of the employed mid-IR detectors with large background currents derived from each detector and low average photocurrents. Improvements to the design may well translate the results obtained in the near-IR and yield sensitivities of ≈ 1 × 10−5 . Ratiometric detectors have been shown to greatly reduce technical noise and enhance the sensitivity with short path distances. Technical noise originating from optical elements that are only present in either sample or reference beam are not cancelled. This limits the use of a BRD to noise reduction with common optical paths. For example, technical noise from multi-pass optical cells cannot be eliminated by this technique and hence extrapolation of short-path sensitivities does not scale with longer pathlengths. BRD is therefore well suited for low-noise measurement application using compact short-path absorption cells.

470


3.2 Wavelength and Frequency-Modulation Spectroscopy (Noise Reduction) As discussed in Sect. 1.1, the output wavelength of diode lasers can be changed by either changing the entire device temperature or by changing the injection current. The latter tuning mechanism is very rapid, thus making it possible to employ high frequency modulation techniques in the kHz to MHz regime. Traditionally, these two frequencies are used to describe two different modulation approaches. Modulation employing frequencies in the kHz regime is denoted as wavelength modulation spectroscopy (WMS). This approach, which is also known as harmonic detection or derivative spectroscopy, uses a modulation frequency (≈ 1–100 kHz) that is much less than the half width of the laser source employed, which typically ranges between several MHz to hundreds of MHz. By contrast, frequency modulation spectroscopy (FMS) is characterized by modulation frequencies greater than the laser line half width and typically involves frequencies up to several hundred MHz. Frequency modulation is always accompanied by amplitude modulation, as the injection current also controls the laser output power: E(t) = A[1 + m cos(Ωt)] sin[ωt + β cos(Ωt + Φ)] .

(6)

Here E(t) is the laser electric field, ω = 2πc/λ the laser frequency, and Ω the modulation frequency. The quantities m and β are the amplitude and frequency modulation indices, respectively, and Φ is the generally nonzero phase shift. Sine-wave modulation of the diode laser has the effect of creating multiple side-bands in its otherwise nearly monochromatic emission spectrum. Each side-band is separated from the carrier by an integer multiple of the modulation frequency Ω, and its relative intensity depends on β. In frequency-modulation spectroscopy, Ω significantly exceeds the laser linewidth that is typically several tens of megahertz, and m, β are both small, so that only the two first-order side-bands, ω + Ω and ω − Ω, have appreciable magnitude. After uniform attenuation, such as that encountered in nonresonant optical systems or media, the side-bands add up coherently with the carrier and balance each other to produce a beam of nearly constant intensity, A2 . If the attenuation strongly depends on frequency, however, as is the case with most gases, one of the side-bands may become unbalanced and lead to the appearance of multiple harmonics of Ω in the detected laser intensity. The strength of absorption determines the magnitude of these harmonics, which may be measured separately and with high noise immunity, by using a lock-in amplifier for example. This is usually done while the laser carrier frequency ω is scanned in the vicinity of the absorption line of interest. This detection technique was first applied by Bjorklund to a CW dye laser [126]. It has proved very effective and is used in diode laser spectroscopy today, sometimes in modified form such as two-tone frequency-modulation (TTFM) [127], or amplitude-modulated phase-modulation (AMPM) spectroscopy [128].


471

In WMS, which is really another form of FM spectroscopy, the modulation frequency Ω is smaller than the laser linewidth, and the modulation indices m and β are both large [129,130,131]. The side-bands are then present to a very high order and due to their small separation from each other, merge into a continuous spectrum. The detection is again performed at the first, second, or higher harmonics of Ω as the laser carrier frequency ω is scanned in the vicinity of a gas absorption line. Lock-in amplifiers or mixers are employed in this approach in selecting the harmonic of choice, which in most cases is the second harmonic. WMS is used in applications that rely on relatively low-speed detectors, and its inherent sensitivity is typically limited by the laser amplitude 1/f noise [132]. However, optical noise often limits the sensitivity achievable using both WMS and FMS techniques in real world systems. In addition, as discussed by Werle [133], an FM spectrometer can be interpreted as an optimized single beam interferometer. In this model, as the modulation frequency is increased the measurements become more sensitive to small changes in the optical path, thus ultimately resulting in less stability relative to WMS techniques. Both of these issues may explain the fact that the expected improvements with FMS over WMS have not been realized on a routine basis. As an alternative to modulation spectroscopy, some research groups have been very successful in achieving noise reduction by sweeping the laser injection current at kHz rates and detecting the resulting direct absorption spectrum using a signal averager (see for example, Zahniser et al. [134]). When coupled with background subtraction (to be discussed in Sect. 4.1.1) absorption sensitivities in the 10−6 range have been achieved. This approach has the advantage over modulation techniques in providing a direct measure of the sample concentration using the Beer–Lambert absorption law without the need for calibration standards and lock-in amplifiers or mixers, required for modulation approaches. In addition, as all modulation techniques effectively smear the peak absorbance at the line center with absorbance in the wings, the effective line center absorbance is reduced. Fried et al. [135], Iguchi [136] and references therein indicate that the effective line center absorbance using modulation approaches is only 30–50 % of that achieved employing direct absorption approaches. The exact reduction depends upon the particular modulation function employed. On the other hand, modulation spectroscopy presents some advantages over rapid sweep integration direct absorption spectroscopy. In modulation spectroscopy one has some flexibility to choose the modulation amplitude and frequency to minimize dominant optical noise features that may be present in direct absorption techniques. Furthermore, since modulation techniques rely on a “fast” change in the absorption coefficient with wavelength, these approaches discriminate against broad featureless absorptions, such as those from the wings of atmospheric pressure water lines and those from big unresolved organic molecules. This aspect, which is often overlooked, becomes important as an added degree of selectivity when

472


measuring trace gas concentrations in the atmosphere at levels of 100 pptv or less. Even though one may select an isolated absorption line to quantify a particular gas of interest using the HITRAN database [117], there exists the possibility that numerous broad featureless organic molecules, that can be present in the atmosphere, may spectrally overlap the absorption line of interest. In such cases, direct absorption measurements may yield systematic errors. 3.3 Long Optical Path Length Spectroscopy (Signal Enhancement) One of the most obvious ways to enhance the absorption signal is inherent to the Beer–Lambert absorption law, where the linear signal improves with longer optical pathlengths. Traditionally, this has been implemented by the use of optical multi-pass cells. Four types of multi-pass cells are most commonly applied: White cells [137,138], Herriott cells [139], Chernin [140] and astigmatic mirror multi-pass cells [141]. For all four types of cells, the focusing mirror curvature, applied to the beam at each reflection, keeps the beam from diverging. The White cell [137,138] is the oldest arrangement. It consisits of two semicircular mirrors, called the “D” mirrors, closely spaced along a common diameter facing a third notched mirror in a nearly confocal arrangement. The probe beam enters through one notch and emerges through the other. The number of passes is varied by changing the “D” mirror angle. The Herriott cell [139] has two identical spherical mirrors separated by nearly their diameter of curvature (nearly concentric) facing each other. A probe beam launched through a hole in one of the mirrors at an angle to the optical axis, completes a certain number of passes between the mirrors, and exits through the same hole (or a hole in the other mirror). The beam bounce pattern and pathlength are controlled by adjusting the mirror separation. For both the White and the Herriott configurations, the number of passes, if not limited by attenuation of light due to the finite mirror reflectivity, is limited by the overlapping of spots on the mirrors. Spot overlapping creates optical interferences causing base line oscillations superimposed on the absorption feature. Astigmatic mirror cells [141] are variations of the Herriott cell that spread the light spots over the entire mirror surfaces. This greatly increases the number of spots achievable without overlapping spots and therefore the number of passes. This cell type is also more compact and possesses the smallest cell volume to effective path length ratio. In such a multi-pass cell, the number of passes is typically configured for 90–238, which translates to effective optical pathlengths from 18 m to 210 m for mirror separations of 0.3 m–0.9 m, thus providing a commensurate improvement in signal strength. The cell volume of multi-pass cells scales with the number of passes and mirror separation. Respective volumes for the aforementioned astigmatic cells range from 0.3–5.2 l. Hence, longer optical pathlengths also increase the surface area and flushing


473

time and ultimately determine the speed of measurement. The four multipass arrangements are commonly enclosed in a vacuum housing, and used for the measurement of static gas samples or controlled gas flows, but can also be operated without enclosure for open-path ambient air trace gas monitoring applications. Multi-pass-cell mirrors can be configured for broadband laser wavelength operation (from the near IR to the long-mid-IR region by use of gold or silver coated mirrors) or can be optimized for specific wavelength regions by using dielectrically coated mirrors similar to mirrors used in cavity ring down spectroscopy. Metallically coated mirrors have a typical reflectivity of > 99 %, resulting in a cell transmission of 16 % for 182 passes. Cell transmissions can be greatly enhanced by the use of dielectrically coated mirrors, but at the expense of a narrower wavelength operation. The beam entrance and output coupling holes are a few mm in diameter. To avoid extensive beam aperture clipping resulting in forward and backscattering, the entering beam must be matched to the f/# of the multi-pass cell. Furthermore, the beam spots on the mirror are separated by a finite width. Thus, the higher the f/# and the smaller the beam size, the less optical interference may occur. The beam pointing stability is another important factor as it can have a multiplicative effect on the effective pathlength and hence produce jitter in the absorption signal. Slow minute changes of the beam direction may also superimpose baseline fluctuations. By flushing the multi-pass cell with “zero-air” and acquiring the background, this effect can be captured and removed. To obtain a high measurement duty cycle, the beam pointing instability must be minimized and the multi-pass cell be mounted on a rugged platform. 3.4 Cavity-Enhanced Spectroscopy Methods (Signal Enhancement) Another technique, which takes advantage of long optical path length absorption in high finesse optical cavities, is called cavity-enhanced spectroscopy. Various methods have been developed. Cavity ring-down spectroscopy (CRDS), first demonstrated by O’Keefe and Deacon [142], is based on the observation of the decay rate of an injected laser beam stored in a cavity comprised of ultra-high reflective spherical mirrors. The rate of decay (inverse of the ring-down time constant) is determined by a) mirror absorption and scattering, and b) wavelength dependent absorption loss by the inserted sample gas. If the decay rate of a) is determined in the presence of a non-absorbing gas or at a non-gas absorbing wavelength, then the gas concentration is exactly proportional to the difference of the observed inverse sample decay rates. Figure 7 illustrates the concept of CRDS. Also shown is the inherent difference to direct absorption spectroscopy. The decay rate is independent of laser amplitude, which relaxes the requirements of the laser source, in particular for pulsed sources with high pulse-to-pulse variations.

474


Fig. 7. Concept of CRDS. Shown are the laser signal as a function of time and wavelength before an absorption cell, and the time evolving signal at the detector, after the laser radiation has stopped or is blocked from entering an absorption cell, with and without a sample present. The difference of the decay rate of the sample and ‘empty’ cell is recorded. The integrated area or difference of the inverse decay rates is directly proportional to the gas concentration

Due to the large effective pathlengths (1 to 10 km is typical, 100 km has been demonstrated), this technique offers significantly higher signal enhancement than is obtainable in conventional absorption spectroscopy. A typical CRDS cell has a base pathlength ranging from 0.5 m to 1 m. The mirrors in such a cell can be designed for specific wavelength regions from the ultraviolet (≈ 300 nm) to the mid-IR spectral region (10 µm). High reflectivity coatings of 99.95–99.99 % in the mid-infrared are commercially available. Although these are very small differences in the mirror reflectivity, the resulting effective pathlength can be significant. For example, a 0.5 m cavity equipped with identical 99.98 % reflectivity mirrors yields a pathlength of ≈ 2500 m, whereas with 99.99 % reflectivity, an effective pathlength of 5000 m may be obtained. These pathlengths give ring-down times exceeding 10 µs. If the minimum detectable fractional absorption determined by the decay rates is constant, this results in a 50 % change of minimum detectable concentration. The high reflectivity is usually retained over ≈ 10 % of the center wavelength, thus over a typical absorption scan no changes of pathlength need to be considered and even multiple species detection can be implemented with a single set of mirrors. Tests with a variety of gases indicate stability of the mirror reflectivity without noticeable degradation [143]. Condensation on mirrors can be avoided by heating the mirrors to 70 ◦ C or higher without damage [144]. Nevertheless, one should avoid deposits of any kind on the mirrors, and one method successfully used is to employ a clean air purge flow over the mirrors. A second practical consideration relating to the very high mirror reflectivity deals with the amount of light that can be injected into the cavity and the amount that can be extracted and detected. For CRDS, the output intensity can be on the order of 10 % of the input intensity. For non-resonant cavity enhanced spectroscopy methods, the transmitted intensity is on the order of the transmission through a single mirror. In the following two sections, we provide an overview of the inherent merits of this absorption measurement technique in the mid-IR wavelength region. We describe technical challenges and solutions of several approaches. We will also give examples on how this technique compares to established spectro-


475

scopic techniques. Recent articles will provide a more detailed discussion on this subject than is possible within the scope of this chapter [145,146,147,148]. 3.4.1

Cavity Ring-Down Spectroscopy

Cavity ring-down spectroscopy (CRDS) is a direct absorption technique, which can be performed with pulsed or continuous light sources. This technique was derived from the characterization process of high reflectors, in which the ring down rate indicated directly the reflectivity of the mirrors. Since the original demonstration for use in spectroscopic measurements [142], many papers have since reported the improvement and applicability of this technique and extended it to longer near-IR and mid-IR wavelengths [142,145,149,150,151,152,153,154,155]. Today, CRDS is used extensively in the visible and near infrared. In one example, a pulsed laser is employed for measurements of hydrogen-bonded clusters formed in molecular beams [156]. The progress in the mid-IR has been stimulated by the availability of ultralow loss cavity mirrors and convenient mid-IR solid-state sources, such as pulsed quantum cascade laser and parametric conversion sources. With pulsed lasers, CRDS requires a short laser pulse to be injected into a high finesse optical cavity to produce a sequence of pulses leaking out through the end mirror from consecutive traversals of the cavity by the pulse. Typically, the laser pulse is short and has a small coherence length compared to a relatively large physical cavity length. Under these conditions interference effects are avoided and the intensity of the cavity pulses decays exponentially with a time constant τ=

1 l c αl − ln R

(7)

where α is the absorption coefficient of the intracavity medium, l is the cavity length and R is the mirror reflectivity (both mirrors are assumed to have the same reflectivity and the refractive index of the medium is assumed to be 1). By measuring the ring-down time, τ , without and with the absorbing gas present, the value of α can be determined. This technique is simple and immune to laser power fluctuations. In the previous discussion, pulsed laser sources with low coherence length were considered and as a result of high-mirror reflectivities, the light levels reaching the detector are small. An alternative approach utilizes the resonance of the cavity by employing CW laser sources with a long coherence length. This is illustrated in the “wavenumber time domain” depicted in Fig. 8a. The laser line represented by the dashed curve is broader than the cavity mode spacing, or free spectral range FSR = c/2l. The cavity throughput can be made much higher if the laser linewidth ∆ νL FSR. This condition can be satisfied if a narrow-line CW laser is used. When the laser frequency coincides with one of the cavity modes as shown in Fig. 8b, the cavity throughput is approximately T = ∆ νC /∆ νL , where ∆ νC is the spectral

476


(a)

(b)

Fig. 8. Two situations of laser radiation filtering by an optical cavity (idealized one-dimensional consideration). (a) The laser linewidth is much broader than the cavity mode spacing (i.e. pulsed). In this case, the cavity throughput does not depend on the laser linewidth but is solely defined by the cavity finesse. (b) The laser linewidth is less than the cavity mode spacing. The cavity throughput is determined by the ratio of the cavity mode width to the laser linewidth

width of the cavity mode (1–10 kHz). The laser emission can be interrupted and the ring-down decay measured in the same manner as it is done with a pulsed laser. The use of CW laser sources for ring down was first proposed by Lehmann in 1996 [157], followed by Romanini et al. [158,159] using CW near-infrared DFB diode lasers. In order to maintain a good overlap of the laser with the cavity mode, the two must be locked to each other. The first work on CRDS measurements with a QC-DFB laser was reported by Paldus et al. [160]. The authors used a CW laser generating 16 mW at λ = 8.5 µm. The measured ring-down time of the empty three-mirror cavity was 0.93 µs. An acousto-optic modulator was used to interrupt the cavity injection for ring-down time measurements. The system was tested on diluted ammonia mixtures, and a noise-equivalent sensitivity of 0.25 ppbv achieved. An estimated 1.0 × 10−9 cm−1 detectable absorbance limit was reported. A spectroscopic gas sensor for nitric oxide (NO) detection at 5.2 µm based on CDRS was reported by Kosterev et al. [161]. Measurements of parts per billion (ppb) NO concentrations in N2 with a 0.7 ppb standard error for an 8 s data acquisition time were performed. Interesting work [162,156] has been done combining a novel OPO with cavity ring-down spectroscopy. A practical advantage of pulsed CRDS over CW CRDS is its applicability to study transient species formed for example by laser ablation sources [163]. 3.4.2

Cavity-Enhanced Absorption Spectroscopy

A simpler (as compared to CDRS) method to exploit a high finesse optical cavity for increasing the sensitivity to absorption has been developed [164,165,166,167,168,169] and is called “integrated cavity output spectroscopy” (ICOS) or “cavity-enhanced absorption spectroscopy” (CEAS). Here, laser light is coupled into the high-finesse cavity via accidental coincidences of the light with the cavity eigenmodes by dithering the cavity length. The time-integrated intensity radiation leaking out of such an optical cavity, averaged over many cavity modes, can be used to determine the absorption of the intracavity medium. Effectively this is equivalent to a time integration of the ring-down curve. Just as in cavity ring-down spectroscopy, an effective optical pathlength of several kilometers can be obtained in a very small volume.


477

However, in its most simple configuration, the noise levels are relatively high since the cavity transmission varies significantly, depending on whether the laser is on or off resonance with the cavity. The ability to effectively average over these frequency response functions is the limiting noise factor to date. A novel approach by Paul et al. [167] uses an off-axis cavity alignment similar to the mirror configuration in astigmatic Herriott cells. This effectively lowers the FSR of the cavity and generates a very dense cavity mode spectrum. In fact, as the reentrant condition becomes longer (many multiple paths until the input beam again overlaps with the cavity beam) the mode spacing will become so dense that the cavity transmission will be almost independent of the laser wavelength. This effectively eliminates the requirement of any laser wavelength (as in CEAS) or cavity dither by a piezoelectric transducer (as in ICOS). However, this in turn also collapses the Fabry–Pérot condition and the transmission is significantly lowered and thus higher laser powers or more sensitive detectors may be needed. This technique has been demonstrated first in the visible wavelength region [167] and has been recently extended to the near-IR region for measurements of a variety of species (CH4 , C2 H2 , CO, CO2 , NH3 ) and demonstrated sensitivities of up to 3 × 10−11 cm−1 Hz−1/2 at 1.51 µm [170]. A minimum detectable concentration of 0.3 ppb (S/N = 3) of C2 H2 at 50 Torr was obtained (1 s averaging time). This latest development demonstrates the ability of cavity enhanced spectroscopic techniques to achieve similar detection sensitivities as obtained with traditional multipass cell absorption spectroscopy. However, the true figure of merit for a real world measurement is the ability to replicate the same result when one samples a constant input concentration. Since low frequency noise sources from a variety of causes may also be important in the measurement process, the ability to replicate the same answer in many successive measurements can be far worse than the inherent sensitivity, and this will be further discussed in Sect. 4.1. One of the most advanced methods, which utilize cavity enhancement is called “Noise-Immune Cavity-Enhanced Optical Heterodyne Spectroscopy” (NICE-OHMS) technique [171,172]. This technique combines the power of signal enhancement of cavity enhanced spectroscopy with the noise reduction of FM spectroscopy. In this technique, the laser frequency is locked to the frequency of a cavity mode. This method has the potential to provide shot-noise limited sensitivity with an effective pathlength determined by the cavity ringdown time. Ma et al. [171] reported a sensitivity of 10−14 cm−1 . This spectacular sensitivity is superior to that achieved with CRDS. To achieve such sensitivity, the laser requires active frequency stabilization below the kHz linewidth level, which is comparable or less than the spectral width of the cavity mode. In addition NICE-OHMS typically requires milli-Torr sample pressure, which effectively reduces the gas density by two orders of magnitude, and hence the effective minimum detectable concentration. High optical saturation (≈ 300 W of power inside the cavity) makes the Beer–Lambert ab-

478


sorption law no longer valid and hence the extraction of quantitative data difficult. The first implementation of this technique using a QC-DFB laser has been reported [173]. The NICE-OHMS approach is technically highly sophisticated and is only suitable for fundamental laboratory applications. The application and suitability of this technique for quantitative trace gas measurements has not yet been demonstrated. 3.5 Photoacoustic and Photothermal Spectroscopy (Signal Enhancement) Photoacoustic spectroscopy (PAS) has found its principal use in sensitive trace gas detection. It is based on the photoacoustic effect, in which acoustic waves result from the absorption of radiation. In its application to laser spectroscopy, the laser beam impinges on a selected target gas in a specially designed cell [174,175,176,177,178,179,180,181,182,183]. In contrast with other mid-IR absorption techniques, PAS is an indirect technique in which the effect on the absorbing medium and not direct light absorption is detected. Light, from either pulsed or chopped CW laser sources produces a transient temperature rise in an absorbing medium via non-radiative relaxation processes, which then translates into a pressure change or sound wave as illustrated in Fig. 9. This is detected with a sensitive microphone(s). The acoustic signal is directly related to the concentration of the absorbing molecules in the cell. For CW laser sources, there are two modes of operation for PAS, either the exciting light can be modulated at a frequency away from any cell resonance or it can be adjusted to coincide with an acoustic resonant frequency. The in-resonance mode is usually employed with the low-power pump lasers to provide larger signals. However, precautions may be necessary to minimize changes of the instrument response due to the change of the speed of sound caused by temperature and gas compositional changes. PAS is ideally a background-free technique: the absorbing gas generates the signal, and in the absence of an absorbing gas there is no acoustic signal. In real PAS experiments, background signals

Fig. 9. Principle of photoacoustic spectroscopy. The incoming photons excite the target molecule at a resonant wavelength. Collisional de-excitation converts the absorbed energy into local heating and pressure waves, which can be detected by a microphone. (Illustration by courtesy of M. Webber, Pranalytica, Inc.)


479

can originate from nonselective absorption of the gas cell windows (coherent noise) and from outside acoustic (incoherent) noise, and from scattering of the laser radiation by aerosols onto the microphone. PAS signals are proportional to the pump laser intensity and therefore PAS is mostly used with high-power laser sources, in particular CO2 and CO lasers [2,184]. In addition, diode (and in combination with high power optical fiber amplifiers) [185] and QC lasers, solid-state lasers, DFG and OPOs in the infrared have been applied to photoacoustic trace gas detection. Recently, a new approach called quartz enhanced PAS has been developed by Kosterev et al. [186]. Instead of using a gas-filled resonant acoustic cavity, the sound energy is accumulated in a high-Q crystal element. Feasibility experiments utilizing √ a quartz watch tuning fork demonstrate a sensitivity of 1.2 × 10−7 cm−1 W/ Hz. However, in trace gas monitoring applications, PAS is limited to extractive point monitoring due to the requirement of an absorption cell. In addition, PAS requires sufficient sampling pressures (≈ 100 Torr − 1 atm) for efficient collisional transfer and generation of the acoustic waves, thus limiting the selectivity in some cases. Furthermore, the effective collisional transfer can depend on the relative composition of the gas sample. For example, the deexcitation rates and hence the strength of the generated acoustic wave can differ by a factor of ≈ 2, as shown by Fried and Berg for the mid-IR detection of HCl under condition of dry samples and samples with 44 % relative humidity [187]. Another form of the photoacoustic effect is called photothermal absorption spectroscopy. Here, the photoacoustic signal is detected by means of recording the phase change in, for example, an unfolded Jamin interferometer. In this case, a non-resonant probe laser beam (e.g. He–Ne) is spatially overlapped with the resonant excitation beam (e.g. CO2 -laser). The probe beam is split and directed through a parallel path without the absorber. The modulated phase difference of these probe beams is measured [188]. Such a system has been configured for the detection of NH3 and demonstrated a 2σ precision of 250 ppt and 31 ppt in a 1 s and 100 s integration time, respectively. The key features of the photoacoustic technique include (1) excellent detection sensitivities down to sub-ppbv concentrations with powers in the watt range, (2) a large dynamic range, (3) PAS detector responsivity is almost independent of the pump wavelength, and (4) a PAS signal that is directly proportional to the absorbed radiation intensity, but does not scale with pathlength as with the previously discussed signal enhancement techniques. Indeed, the PAS signal will increase if a laser beam passes through the same volume/detection area of a microphone. However, it must pass through the same limited small interacting volume, otherwise more microphones are needed to be co-located along the laser path. Each added microphone will add to the noise floor. Therefore, only a moderate increase of the signal to

480


noise per unit Hz−1/2 measurement time can be attained. On the other hand, the photothermal approach does scale linearly with pathlength. Implementation of a QC-DFB laser to target fundamental absorptions has the potential of considerably improved flexibility and allows one to access many absorption features, whereas line-tunable CO2 laser sources depend upon accidental overlap, and are restricted to the 9–11 µm region. Ammonia and water vapor photoacoustic spectra were obtained using a CW cryogenically cooled QC-DFB laser with a 16 mW power output at 8.5 µm as reported by Paldus et al. [189]. A PAS cell resonant at 1.66 kHz was used. The QC-DFB was used for frequency scans using temperature tuning and for real-time concentration measurements with a fixed laser temperature. Measured concentrations ranged from 2, 200 ppmv to 100 ppbv. A detection limit of 100 ppbv ammonia (≈ 10−5 noise-equivalent absorbance) at standard temperature and pressure was obtained for a 1 Hz bandwidth and a measurement interval of 10 min. Recently, Hofstetter et al. [190] reported PAS measurements of ammonia, methanol and carbon dioxide using a pulsed 10.4 µm QC-DFB laser operated at 3–4 % duty cycle with 25 ns long current pulses (2 mW average power) and close to room temperature with Peltier cooling. Temperature tuning resulted in a wavelength range of 3 cm−1 with a linewidth of 0.2 cm−1 . This sensor used a 42 cm long PAS cell with a radial 16-microphone array for increased detection sensitivity. In addition the cell was placed between two concave reflectors resulting in 36 passes through the cell (with an effective pathlength of 15 m). The laser beam was mechanically chopped at a resonant cell frequency of 1.25 kHz, with a PAS signal enhancement by a Q factor of 70. A pyroelectric detector recorded the QC laser power to normalize the PAS signal. Detection of ammonia concentrations at the 300 ppbv level with a SNR of 3 was achieved at a pressure of 400 mbar.

4

Mid-Infrared Spectroscopic Applications

Tunable mid-infrared spectroscopic sources and spectroscopic techniques provide four important performance characteristics: sensitivity, selectivity, fast response time, and compactness. These are also based on similar optical component design and hence offer the unique ability to mix and match laser sources and techniques to be most useful for a given application. Recent progress in this field and growing optical industrial resources (e.g. optical fiber telecommunication) has led to the evolution and utilization of mid-IR spectroscopic techniques to a wide range of gas sensor applications. These include such diverse fields as: 1) environmental monitoring (CO, CO2 , CH4 and CH2 O are important gas species in various aspects of atmospheric chemistry studies); 2) industrial emission measurements (e.g. fence line perimeter monitoring in the petrochemical industry, combustion sites, waste incinerators, down gas well monitoring, gas–pipe and compressor station safety);


481

3) urban (e.g. automobile traffic, power generation) and rural emissions (e.g. horticultural greenhouses, fruit storage and rice agro-ecosystems); 4) chemical analysis and control for manufacturing processes (e.g. in the semiconductor, pharmaceutical, and food industries); 5) detection of medically important molecules (e.g. NO, CO, CO2 , NH3 , C2 H6 and CS2 ), toxic gases, drugs, and explosives relevant to law enforcement and public safety; and 6) spacecraft habitat air-quality and planetary atmospheric science (e.g. such planetary gases as H2 O, CH4 , CO, CO2 and C2 H2 ). 4.1 Detailed Examples Using Selected Spectroscopic Sources and Techniques There are literally hundreds of examples that can be cited where various midIR sources and spectroscopic techniques as discussed in the previous sections have been employed, including studies in the laboratory and industrial settings, and on ground-based, aircraft, balloon-borne, and rocket-borne platforms. Rather than cite examples of each, we present in this section three representative examples of spectroscopic sources utilizing lead-salt diode lasers, quantum cascade lasers, and difference-frequency generation coupled with either long-path and dual-beam absorption, wavelength modulation, or cavityenhanced spectroscopy. 4.1.1 Lead-Salt Diode Laser Based Spectrometer for Airborne Atmospheric Chemistry Studies In the following, we describe a lead-salt diode laser based trace gas sensor, which incorporates advances collectively developed by many research groups over the years, and thus illustrate the high performance that can be achieved routinely in a rugged airborne field setting. Aircraft measurements present unique and demanding challenges for one has to contend with: (1) changes in cabin temperature by as much as 20 ◦ C over relatively short time periods of 1 h or less; (2) changes in cabin pressure by as much as 100 to 300 mbar over time periods of several minutes; (3) changes in system attitude; and (4) changes in aircraft vibrations that can couple beneficially or detrimentally into the system. As will be shown, by careful attention to numerous details, one can routinely measure absorbances as small as 0.7 to 1.7 × 10−6 for 1-min integration times employing pathlengths of 100 m on aircraft platforms. Figure 10a illustrates the optical layout for a dual channel airborne spectrometer, which contains two Pb-salt diode lasers mounted in a liquid nitrogen dewar. Figure 10b shows a three-dimensional diagram of this system mounted in a temperature stabilized enclosure. This system has successfully acquired ambient measurements of the important atmospheric gas, formaldehyde (CH2 O), on numerous airborne campaigns. As shown, this system was also configured for simultaneous measurements of hydrogen peroxide (H2 O2 ), another important atmospheric trace gas. The performance of this second

482


Fig. 10. (a) Optical set-up of a dual channel absorption laser spectrometer (DCALS); (b) three-dimensional digitally rendered model of DCALS configured for airborne operation

channel, however, was significantly inferior to that of the CH2 O channel, and will not be discussed here. Comprehensive details regarding the airborne system, the associated airborne measurements, as well as detailed background information on calibration, sampling, measurement accuracy, inlet tests, and ground-based and airborne comparisons studies can be found in Fried et al. [16] and references therein. The IR radiation from the lead diode laser operating at 3.5 µm is directed through a multi-pass astigmatic Herriott cell (Aerodyne Research, Inc.) using a series of off-axis mirrors (two parabolic and one elliptical mirror), as shown in Fig. 10a. The IR beam, which traces out a Lissajous pattern in the cell, achieves a total optical pathlength of 100 m in a 3-l sampling volume. Upon exiting the cell, the IR beam is directed onto sample and reference indiumantimonide photovoltaic detectors. The optical system employs a minimum number of components, and of those, only two are readily adjustable. Each component, furthermore, is mounted on a rigid mount with as low a center height as possible. Both measures attempt to avoid subtle mechanical alignment changes. The entire optical enclosure, including the optical bench, is temperature stabilized to around 30 ◦ C, typically to better than ±1 ◦ C over time periods of many hours and significantly better than this over shorter time periods. A series of heaters mounted in the lid (not shown) are used for this purpose. Sheets comprised of an aluminum-balsa wood sandwich (partially removed in drawing) are mounted to the frame structure shown in Fig. 10b. These sheets provide both good thermal insulation and structural support. All the above precautions are critical for high performance aircraft measurements and are essential in extending the system stability period (to be discussed) out to 1-min and longer. Without temperature control, for example, one encounters rather significant changes in optical alignment,


483

background optical structure, and detector dark count signals, as the aircraft repeatedly ascends and descends throughout a typical flight pattern. In addition, owing to repeated changes in cabin pressure, it is equally critical to control the liquid-nitrogen pressure using some type of absolute pressure control valve. Without such control, the liquid-nitrogen boiling point would change, resulting in rather significant shifts in laser wavelength as the cryogen base temperature changes. Such temperature changes could also result in changes in dewar dimensions, and this may produce a consequent change in diode laser position which is integrally mounted to the dewar casing. Absorption data are acquired using second harmonic detection coupled with sweep integration, as discussed by Fried et al. [12]. In this procedure, the diode laser wavelength is repetitively scanned across an isolated absorption feature of CH2 O (2831.6417 cm−1 ) using a 200-point sawtooth ramp applied to the laser tuning current at a frequency of 50 Hz. A 50-kHz quasi-square wave modulation waveform is simultaneously applied to the laser tuning current, and the 2f signal at 100 kHz is detected using a digital lock-in amplifier. Second harmonic signals from the reference arm, which contains a high concentration cell of pure CH2 O, are treated identical to the sample arm. The lock-in amplifier outputs are then digitized (16-bit A/D converters) and coaveraged by a computer. The line center of the reference arm, which has very high signal-to-noise and serves as a wavelength reference, is determined on every scan. Each scan is then appropriately shifted in memory to align the peak centers before co-averaging. This fast spectral shifting is extremely important for high instrument performance [191]. Upon completion of each ambient measurement block, typically 60 s (3000 scans), the wavelength reference line center is determined using a polynomial fit, and an appropriate correction voltage is applied to the laser current controller to keep the absorption feature centered in the scan window. In addition, the mean value for the amplitude of each scan is forced through zero before co-averaging. This procedure effectively removes small scan-to-scan dc variability and thus improves the co-averaging effectiveness. The co-averaged spectrum at the end of each scan cycle is then transformed into the frequency domain employing an FFT algorithm, band pass filtered, and transformed back into the time domain. This approach helps to reduce both high and low frequency noise without significantly affecting the retrieved ambient signals. By far the most dramatic improvement in instrument performance is achieved using rapid background subtraction. Fried et al. [15,12], Zahniser et al. [134], and Werle et al. [191] among others have presented the merits of this approach for tunable diode laser absorption spectroscopy. If carried out correctly, rapid background subtraction effectively captures and removes optical noise, which ultimately limits the performance of most if not all tunable diode laser instruments. As discussed previously, such noise is caused by light scattering from various optical elements, and this generates a somewhat random undulating background structure. Often such structure contains mul-

484


tiple frequencies, amplitudes, and time constants originating from multiple scattering sources. As addressed by Werle et al. [191], acquisition of sample and background spectra along with the associated cell flush times need to be accomplished within a characteristic system stability period, topt , in order for background subtraction to be effective. One can characterize this time period using the Allan variance, as first presented by Werle et al. [191] and subsequently by Fried et al. [12]. During time periods of constant optical enclosure temperature and constant pressure (not presently controlled in our airborne system), system stability periods typically range between 40 and 100 s before drifts become prevalent. During airborne operation ambient air is drawn through a heated Teflon inlet at typical flow rates between 8 and 10 standard liters min−1 (slm, where standard conditions are defined as 273 K and 1 atmosphere pressure) and through the Herriott cell at sampling pressures around 40 Torr. Background spectra are acquired by passing ambient air from a second inlet through a heated Pd/Al2 O3 scrubber, which removes CH2 O without significantly affecting the ambient water concentration, and this airflow is re-routed to the inlet tip at flow rates exceeding the sample flow. During a typical sampling sequence[16], six successive 10 s ambient spectra are acquired, and each 1 min ambient sampling block is preceded and followed by a 10 s background acquisition. An appropriate delay period of 9 s (≈ 5 inlet/cell e-folding residence times) is employed after each switch before data are acquired. The backgrounds surrounding each 1 min block are averaged (time weighted) and subtracted point by point from each of the 10 s ambient spectra. The average laser power is determined for each 1 min sampling block using a third data acquisition channel to continuously record the sample detector dc voltage. The detector dark voltage is obtained every few 1-minute ambient cycles by blocking the laser beam with a shutter for a few seconds. The laser power is determined from the difference of the two measurements, and the ratio of the laser power obtained during calibration to that during a sample measurement is applied. Calibration spectra are typically acquired every 30–60 min by adding CH2 O standards, from a permeation calibration system, to the zero airflow near the inlet entrance. At the cell flows employed, typical CH2 O standard concentrations of 12–14 parts-per-billion by volume (ppbv) are generated at the Herriott cell entrance. Periodically, 7 ppbv standards are also added on top of ambient to check for inlet loss and to check the veracity of the data retrieval algorithm. Each 10 s background-subtracted ambient spectrum is fitted in real time to a background-subtracted calibration spectrum (acquired for 20 s) employing a multiple linear regression approach [192,193]. Each complete ambient acquisition cycle, which includes the acquisition of a 10 s background, 1 minute of ambient averaging, two 9 s delay periods and computer-processing overhead, typically takes 90 s, and this typically falls within the system stability period.


485

1000 900

-6

A = 1 x 10

Number of Measurements

800 700 600 500 400 300 200 100 0 0

10

20

30

40

50

60

70

80

90 100 110 120 130 140 150 160 170 180 190 200

1-Minute Measurement Precision [pptv]

Fig. 11. Formaldehyde (CH2 O) measurement precision diagram at the 1σ level for airborne measurements throughout the TOPSE campaign [16]

Figure 11 shows the instrument performance obtained during the TOPSE (Tropospheric Ozone Production about the Spring Equinox) airborne campaign [16]. Here the 1-minute (1σ) measurement precision, obtained from the standard deviation of replicate measurements upon sampling relatively constant and low ambient CH2 O levels, is given in terms of a histogram. The histogram reflects the fact that the measurement precision during airborne operation is somewhat variable, depending upon the exact alignment stability and the degree to which the acquired backgrounds truly represent the actual backgrounds underlying the ambient spectra. As the real precision for any technique will vary, even in a laboratory setting, the histogram approach gives a more realistic assessment of overall instrument performance than a single figure of merit often reported for many techniques. In addition, the results of Fig. 11 are obtained from replicate measurements, and not the precision of any given measurement, which in the case of tunable diode laser measurements, can be expressed in terms of an individual fit precision. As discussed by Fried et al. [12] this fit precision is proportional to the square root of the fit deviations, and in most instances was found to be a factor of 3 to 4 too optimistic. In this case, like that for many other techniques, there are additional sources of variance, which produce ambient results that have larger run-to-run variability than the precision of any individual measurement. Thus, from these two standpoints the results of Fig. 11 truly represent the meaningful performance that is obtain-

486


able with an airborne tunable diode laser system. As can be seen, most of the 1 min measurements (which require 90 s to acquire) yield 1σ replication precisions in the 20–50 pptv concentration range (median value = 40 pptv) for CH2 O during airborne operation. Employing a sampling pressure of 40 Torr, a temperature of 303 K, a 100 m pathlength, an integrated absorption crosssection of 5.44 × 10−20 cm2 cm−1 molecule−1 and an air-broadening coefficient of 0.107 cm−1 atm−1 , one calculates using a Voigt function that a line center absorbance of 1× 10−6 equates to a CH2 O concentration of 30 pptv. Thus the precisions above correspond to minimum detectable line center absorbances of 0.7 to 1.7 × 10−6 . This in turn computes to pathlength-normalized values of 7 × 10−11 to 1.7 × 10−10 cm−1 employing a 90 s sampling sequence. 4.1.2 Quantum-Cascade Laser Based Trace Gas Sensors in the Life Sciences Laser spectroscopy is finding increasing applications in medicine and the life sciences [194,195,196,197,198,199]. A particular role for spectroscopy is in the monitoring of small molecules that have been shown to be important. The role of simple molecules such as nitric oxide (NO) in physiological processes has received considerable attention in recent years and was a subject of the 1998 Nobel Prize in Medicine. One application is the measurement of NO in human breath samples, since exhaled air is an indicator of several processes taking place in the human body, in particular in assessing the severity of airway inflammation [200]. To observe NO in breath, a cavity-enhanced absorption spectroscopy (CEAS) sensor [169] with a CW QC-DFB laser operating at 5.2 µm with an output power of 80 mW was used. A direct performance comparison was carried out between a sensor configuration where the CES optical cavity was replaced with a 100 m pathlength multi-pass cell. It was found that in spite of having an effective pathlength of 670 m, CES had a lower absorption sensitivity because of baseline noise of ≈ 1 % (averaging 104 QC laser scans). These baseline fluctuations are intrinsic to CES and result from the mode structure of the cavity transmission spectrum. Some improvement in the CES baseline noise can be achieved with the recently developed off-axis ICOS technique [167]. In [161] a spectroscopic gas sensor for NO detection based on a cavity ringdown technique is described. NO is the major oxide of nitrogen formed during high-temperature combustion as well as an important nitrogen-containing species in the atmosphere (NO is a precursor of smog and acid rain). NO is also involved in a number of vital physiological processes, and its detection in human breath has potential applications (e.g. as a marker for diseases like asthma or inflammatory processes) in noninvasive medical diagnostics. A CW QC-DFB laser operating at 5.2 µm was used as a tunable single-frequency light source. The technique used consists of the following features:


487

1) The laser frequency is slowly scanned across the absorption line of interest. 2) One of the cavity mirrors is dithered back and forth to ensure periodic, random coincidences of the laser frequency with a cavity mode. 3) Once such a resonance occurs and the cavity is filled, the laser beam entering the cavity is abruptly interrupted or set off-resonance, and the decay rate of the exiting light is measured. From (7) in Sect. 3.4, the absorption coefficient can be determined as 1 1 1 − α= (8) c τ τ empty where τ empty is the decay constant of the cavity without absorber. The sensor schematic is shown in Fig. 12. It depicts a simpler design than the first CRDS experiment with a QC-DFB laser described in [160]. In that work a variable temperature cryostat for laser frequency tuning was used with the inherent complexity and additional cost of an acousto-optic modulator (AOM) to interrupt the QC laser beam. In the work reported in [161] both frequency tuning and the laser emission interruption were realized by manipulation of the QC laser pump current. No active temperature control was applied to the QC-DFB laser located in a liquid nitrogen optical cryostat: The laser current was supplied by a low-noise current source and monitored using the 0.5 Ω resistor denoted by r1 in Fig. 12. The laser frequency was tunable from 1922.9 to 1920.8 cm−1 when the pump current changed from 300 mA (lasing threshold) to 660 mA. At higher current levels the laser emission was multimode. The tuning range permitted NO detection by accessing absorption lines at 1921.599 cm−1 and 1921.601 cm−1 (R(13.5) components of Function Generator

Removable mirror

Current Source

Function Generator

QCL

r1

A/D Converter

Ringdown Cell

r2

Reference Cell

PZT Driver

Gas flow

Trigger Circuit IR Detectors

Fig. 12. Schematic of a QC laser based cavity ring-down spectrometer [161]

488


the fundamental absorption band). Absorption lines of water vapor and CO2 were also observed. The l = 37 cm long, linear high-finesse optical cavity was formed by two concave mirrors with a 6 m radius of curvature. The measured ring-down time of the cavity without absorber was τ ≈ 3.5 µs, corresponding to ∆ νC ≈ 45 kHz. The laser linewidth was estimated to be ∆ νL = 3 MHz. When a certain level of the detector signal was reached signaling that a TEM00 cavity mode is coupled to the laser, a triggering circuit (TC) opened a metal-oxide semiconductor field-effect transistor (MOSFET) to shunt the laser current, thereby reducing it to a subthreshold value. At the same time, the TC triggered an analog–digital converter, and the detector signal showing the cavity ring-down was digitized for ≈ 30 µs and stored in computer memory. The MOSFET was kept open for 35 µs, so that the laser radiation would not interfere with measurements of the cavity decay constant. This triggering-and-acquisition process was repeated and consecutive ring-down transients were stacked in the A/D memory until a desired number of transients was acquired. The results were post-processed to fit each transient with an exponential decay function yielding a ring-down time τ . The inverse ring-down time plotted as a function of the laser current provides the absorption spectrum. An example of the NO absorption in a mixture with pure N2 is presented in Fig. 13. The noise-equivalent sensitivity was estimated to be 0.7 ppbv for an 8 s data acquisition time. It was not possible to use this sensor directly for measurements of NO concentration in exhaled air (≈ 10 ppbv) because of strong CO2 interference, which can be avoided if the appropriate NO absorption line is chosen (e.g. like R(7.5) components at 1903.123 and 1903.134 cm−1 ) with a QC laser that accesses this wavelength. 0.40

(a) 0.38

1/W, Ps

-1

0.36 0.34 0.32 0.30 0.28

-0.04

-0.02

0.00

0.02

0.04

-1

Relative frequency, cm

Fig. 13. NO absorption in a mixture with pure N2 obtained with the sensor depicted in Fig. 12


489

Recent work indicates that other gases, such as carbon monoxide (CO), can also play a very significant physiological role. CO is produced from heme catabolism by the enzyme heme oxygenase. Previous work has shown that CO promotes blood flow by inhibiting vascular tone and platelet aggregation and that neuronal CO production may modulate the NO-cGMP (guanosine 3 , 5 -cyclic monophosphate) signaling system, demonstrating important biochemical interactions between the two diatomic gases. The extremely low levels of gas production in living cells and the relatively short in vivo lifetime of cell cultures have complicated detailed understanding of the kinetic, or time-dependent processes responsible for their generation. A typical production rate of CO, for example from vascular smooth muscle cells (VSMCs), is 1 to 10 pmol/min/107 cells. Instrumentation for in vivo measurement of gas production should have sensitivities on the parts per billion (ppb) level in order that the dynamics of gas production can be followed with laboratoryscale cell sample populations [201]. Because of low CO production rates from biological tissues, measurements of CO concentrations have been limited to gas chromatography and radioisotope counting techniques. Although these methods are highly sensitive, they cannot measure CO directly, requiring several time-consuming intermediate steps requiring ≈ 15 min, and may be affected by interference from water, oxygen, and carbon dioxide. Infrared laser absorption spectroscopy is an attractive alternative approach for the detection of biological CO at the parts-per-billion (ppb) level in real time [202,203]. A compact gas sensor measured endogenenous CO production from vascular cells using a mid-IR spectroscopic laser source based on differencefrequency generation (DFG) of two near-IR lasers. The CO absorption was detected in the fundamental vibration band near 4.6 µm. In this work, an extractive technique was used with gas samples taken from the flask containing the cell culture to an 18 m pathlength optical multi-pass cell so that the measurements could be performed at a reduced pressure of 100 Torr. Kosterev et al. [203] reported an improved design and performance of an optical mid-IR CO sensor intended for continuous monitoring of cell culture activity at ambient atmospheric pressure. The same fundamental absorption band region was used for CO detection, but a pulsed quantum cascade laser with a distributed feedback structure (QC-DFB) [25] was employed instead of the DFG source. The high output power of the QC-DFB laser and an advanced data analysis approach made it possible to detect biological CO and CO production rates with ≈ 1 m optical pathlength folded above a standard culture flask of VSMCs. A further improvement of the pulsed QC-DFB based sensor was reported by Kosterev et al. [204]. The laser beam was split into two channels, one being used to probe the gas absorption and the other as a reference to measure the laser pulse energy. The subsequent normalization eliminated pulse-topulse energy fluctuations as an error source, which was the predominant cause of error previously [203]. This automated sensor was used for contin-

490


uous monitoring of CO in ambient air detected by its R(3) absorption line at 2158.300 cm−1 (λ ≈ 4.6 µm). A noise-equivalent detection limit of 12 ppbv was experimentally demonstrated with a 1 m optical pathlength. This sensitivity corresponds to a standard error in fractional absorbance of 3×10−5. All the measurements were carried out at atmospheric pressure, and hence it was not possible to periodically acquire a baseline with an evacuated sample container. In order to keep the baseline (which included weak unwanted interference fringes from optical elements) stable during multi-hour measurements, the slow drifts of the laser frequency were actively compensated by computer-controlled corrections to the subthreshold current. A constant CO production rate of 44 ppbv/hour was observed, taking into account the 0.5 liter volume of the cell culture container. This corresponds to a net CO production rate of 0.9 nmol/107 cells/hour, which is in agreement with previous measurements [202] obtained with similar cells and treatment regimes. A compact mobile ammonia sensor based on a thermoelectrically cooled pulsed QC-DFB laser operating at ≈ 10 µm was described in [205]. High sensitivity detection of NH3 is also of interest in the control of deNOx chemistry, industrial safety and medical diagnostics of kidney related diseases. The optical configuration of this sensor was similar to that described in [206], but the multi-pass cell was replaced with a simple 50 cm long double pass gas cell, and no zero air subtraction employed. The laser housing was improved by replacing the previous beam-shaping optical system consisting of two off-axis parabolic mirrors and a lens with a single aspheric lens. The laser was scanned over two absorption lines of the NH3 fundamental ν2 band. The sensor was completely automated and only required the LN2 dewar of the detector to be refilled every 12 h. This sensor was applied to the continuous monitoring of NH3 concentration levels at the ppm level present in bioreactor vent gases in a water reprocessing system located at NASA’s Johnson Space Center in Houston, TX. A sensitivity of better than 0.3 ppmv was estimated which was sufficient to quantify expected ammonia levels of 1 to 10 ppmv. 4.1.3 Design and Applications of Fiber Based Difference Frequency Based Mid-IR Gas Sensors In the following, the optical architecture and performance of several field portable gas sensors based on difference frequency sources are discussed. The gas sensors described here utilize fiber based near-IR lasers, high-power rare earth doped fiber amplification and single-pass difference-frequency generation in QPM-PPLN crystals (Sect. 1.2). Difference frequency generation utilizing optical fiber coupled and fiber based pump sources allows great flexibility in designing a robust mid-infrared source with power levels typically ranging from 0.1 mW to 1 mW. Figure 14a depicts four representative fiber based DFG-source configurations, which have been developed and applied to trace gas detection. In order to evaluate and


491

Fig. 14. (a) Fiber based difference-frequency generation sources using narrowlinewidth diode laser sources and high-power optical fiber amplifiers. Shown are respective wavelengths, input and generated power levels [207]. (b) Optical fiber pumped DFG beam profile: measured 37.5 cm from PPLN exit facet, beam diameter @ 1/e2 : x = 0.80 mm, y = 0.80 mm Gaussian fit > 95 % [208]

confirm robust performance of these sources itself and their applicability, direct long-pathlength absorption as the simplest form of sensitive detection was used. Advanced signal enhancing techniques and noise reduction techniques have been applied and laboratory results indicate unique advantages of these sources. In the examples shown, low-power seed diode laser sources are used in combination with high-power, wide bandwidth fiber amplifiers. Several inexpensive low-power diode laser sources whose wavelengths overlap with the gain region of the fiber optic amplifier can be used, one at a time, or can also be multiplexed to be amplified simultaneously or sequentially in time. This permits easy modification of a fixed optical fiber platform by choosing any desirable seed wavelength within the fiber amplifier(s) gain bandwidth in the difference frequency mixing process and generate the desired mid-IR wavelength(s) (Fig. 1). Fiber amplifiers retain the spectroscopic properties of the seed laser sources, and thus decouple the high-power requirement for efficient DFG from low-noise, narrow-linewidth operation. This approach is more cost effective and technically easier than constructing a pump laser source which meets all of the three requirements at the same time, namely

492


narrow-linewidth, high-power (W), and multiple-wavelength operation. In addition, near-IR diode lasers have similar tuning rates with current and temperature. This in turn provides predictable ease of exchange of sources and mid-IR frequency stabilization. For example, an ambient temperature change will introduce the same wavelength shift to two near IR diode laser sources with similar tuning rates and use of similar types of current/temperature controller. If those laser sources are difference-frequency mixed, the drift is subtractive and affords high inherent mid-IR wavelength stability, provided the diode lasers and controllers are carefully selected and operate at the same ambient conditions. The use of fiber optics for DFG pump beam delivery also provides a stable and inherent spatial overlap to produce a circular, homogeneous beam with a near-Gaussian intensity distribution as shown in Fig. 14b. This offers small beam sizes (≈ 2 mm) with f/# ≈ 100, and are obtained by the use of only one relay optical element to collect and image the DFG beam to the spectroscopic cell (e.g. multi-pass cell). Quasi-phase matched periodically poled LiNbO3 (QPM-PPLN) has been shown to be efficient for parametric frequency down conversion (see Sect. 1.2). It also offers flexibility in the conversion bandwidth by either using crystals incorporating multiple QPM periods or a fan-out design for continuous QPM [69]. The DFG sources depicted in Fig. 14a have been used in several field applications to demonstrate their feasibility in real-world environments. These include remote volcanic gas measurements in Nicaragua, operation in an industrial setting and analysis of a space rated trace contaminant control system (using the source shown in Fig. 14a A) [209,210]. The DFG source depicted in Fig. 14a B), shown in more detail in Fig. 15, is configured for high sensitivity dual-beam long-path absorption spectroscopy of urban pollutants. Using this source, Rehle et al. [211] conducted extensive urban gas detection of CH2 O over extended time periods and achieved a detection limit of 0.32 ppbv of atmospheric formaldehyde at 3.53 µm (2832 cm−1 ). This corresponds to a sensitivity of 1 × 10−9 cm−1 in combination with a 100 m pathlength low-volume (3.3 l) astigmatic mirror Herriott gas cell. A dual-beam absorption configuration that employs two dc-coupled Peltier-cooled HgCdTe (MCT) detectors was used to eliminate the optical interference fringes originating from the refractive optical elements of the DFG conversion stage and optical fiber components. A typical absorption spectrum obtained is shown in Fig. 16. To further enhance the signal-to-noise, mainly limited by electronic noise and technical noise from the multi-pass cell, wavelength modulation and zero-background subtraction could be used as described in the lead-salt diode laser based detection system above. However, these techniques were not used with this device, because of relatively high urban formaldehyde concentrations ranging from 1 to 50 ppbv. The sensor as shown in Fig. 15 has been operated autonomously for a continuous nine and five-day period at two separate field sites in the Greater Houston area, administered by


493

Fig. 15. High-power continuous-wave DFG source employing dual-beam spectroscopy used for urban pollution monitoring of CH2 O. DL, diode laser; DFB-DL, distributed feedback-DL; DBR-DL, distributed Bragg reflector-DL; OI, optical isolator; WDM, wavelength division multiplexer

the Texas Natural Resource Conservation Commission (TNRCC) and the Houston Regional Monitoring Corporation (HRM). The acquired spectroscopic data were compared with results obtained by a well-established wetchemical o-(2,3,4,5,6-pentafluorobenzyl) hydroxylamine (PFBHA) technique with good agreement. While the accuracy of the determined concentrations is comparable with results from conventional wet-chemical techniques, the described DFG sensor offers excellent time resolution on the order of seconds and permits unattended continuous operation for long periods of time. The maintenance-free design of a tunable infrared DFG based diode laser spectrometer and the capability of remotely controllable computerized operation makes such instrumentation a convenient, robust tool for mobile trace gas detection. Thus, formaldehyde concentration measurements using direct absorption laser spectroscopy have proved to be a sensitive and effective method for online trace gas monitoring in an urban setting. Further testing of a similar DFG system by Richter et al. [208] utilizing additional wavelength modulation spectroscopy and zero air background subtraction has generated laboratory based replicate detection precisions (1σ, 1 min average) of better than 2.5 × 10−10 cm−1 , which corresponds to 74 pptv minimal detectable CH2 O concentration.

494


1.00005

Transmission

1.00000 0.99995 0.99990

concentration: (8.49 ± 0.57) ppbv

0.99985

goodness of fit: 2 F = 3.4272 E-10 V = ± 1.852 E-05

0.99980 0.99975 0.99970 2831.61

2831.64

2831.66

2831.69

2831.72

Wavenumbers [cm-1]

Fig. 16. Absorption spectra of CH2 O using the DFG based gas sensor depicted in Fig. 15

The recent rapid advances in the optical fiber telecommunication industry has produced a number of new laser sources and components to continuously expand the DFG wavelength coverage and generate higher midIR power at the mW level. Many of the requirements for telecom applications overlap and directly enhance the spectroscopic performance of DFG based mid-IR sources. These include narrower linewidth (< 1 MHz, 1 s), stable higher power using highly saturated two-stage fiber amplifiers, and wavelength stability. Relatively simple, yet powerful, mid-IR sources can be designed. Fig. 14a D) shows the optical design of an all telecom wavelength based mid-IR source [212]. Here, an optically pumped VCSEL operating at 980 nm provides over 500 mW coupled into a single-mode optical fiber with linewidths of ≈ 10 MHz. The other DFG pump source is a standard low noise and narrow-linewidth (< 1 MHz) DFB diode laser amplified by a 500 mW Er/Yb fiber amplifier. Mid-IR powers in excess of 0.25 mW at 2.64 µm are easily obtained with this DFG mixing configuration in combination with a 2 cm long QPM-PPLN crystal. Using this source, spectroscopic detection of fundamental absorption lines of hydrogen fluoride (HF) has been performed. Employing direct absorption spectroscopy, fractional absorption sensitivities of 1 × 10−4 was obtained. The DFG frequency stability of this source was measured by tracking a high resolution HF absorption line over several hours, which indicated a maximum peak-to-peak wavelength drift of 20 MHz during a 2 h time interval. Given the strong absorption cross-section of HF at this wavelength range (S = 1 × 10−18 cm molecule−1 ), sub-ppb detection sensitivity can be obtained over relatively short pathlengths of several meters. The spectroscopic performance characteristics of DFG sources and their inherent flexibility to access specific fundamental mid-IR absorption lines can be further facilitated for precision ratio measurements of, for example, isotopic 12 CO2 /13 CO2 . The ability of measuring isotopic ratios with a high


495

precision and selectivity makes this technique particularly attractive for a variety of applications, such as carbon cycle research, volcanic gas emission studies, and the use of isotopic tracers in medical diagnostics. In order to be useful, measurement precisions of 0.180/00 to 10/00 are required and are defined as [(Rsample − Rstd )/Rstd ] × 1000, where R = [13 CO2 ]/[12 CO2 ]. Isotopic ratios can be measured by detection of closely paired absorption lines of 12 CO2 /13 CO2 found for example in the 4.3 micron wavelength range and comparing this spectroscopic signatures with a known reference gas. To obtain high precision, several key requirements of the source and spectroscopic technique have to be addressed. Most important are the inherent stability and measurement precision of absorption spectra and the temperature dependence of the probed absorption lines. Closely paired absorption lines of 12 CO2 /13 CO2 with similar absorption strengths usually originate from different ground energy levels and hence have different temperature dependences. Figure 17 shows three suitable line pairs for the detection of 12 CO2 /13 CO2 isotopic ratios. Also indicated are the temperature coefficients based on the Boltzman distribution [213]. Using the DFG source depicted in Fig. 14a A) in combination with a single-pass 20 cm long dual chamber absorption cell, Erdelyi et al. [214] have obtained a 1σ measurement precision of ≈ 0.80/00 . The dual chamber absorption cell was built from a solid piece of brass and incorporated two parallel small bore extrusions and was end-fitted with common Brewster angle windows. This allowed rapid comparison measurements of sample gases with 12

13

CO2

CO2

100

c

a

b

Transmission (%)

99

98

97

a: 22.9 b: 18.8 c: 18.9

0

-1

/00 K

0

-1

0

-1

/00 K

100

/00 K

99

96

L=0.1 m p=40 Torr cCO2= 350 ppm Voigt lineshape

98

2295.8

2295.9

2296.0

2296.1

95 2286

2289

2292

2295

2298 -1

Frequency (cm )

Fig. 17. HITRAN simulation of isotopic CO2 lines

2301

2304

496


a known reference gas mixture. The cell design also ensured good thermal mixing of the sample and reference gas standard and because of the small volume, a minimal amount of expensive reference gases. The measurement precision of this system was mainly limited by electronic noise due to low pump power from an ECDL operating at the edge of its tuning range, and in turn only provided a mid-IR power of ≈ 0.2 µW. However, the achieved precision is sufficient for many applications including volcanic gas emission studies. Other approaches with similar detection levels ranging from 0.20/00 to 20/00 have been reported in the literature, including lead-salt diode, color-center-, CO2 -, and near-IR diode laser sources in combination with various multi-pass cell designs and cavity enhanced spectroscopy [213,215,216,217,218].

5

Summary and Outlook

This chapter has attempted to survey the current status of various tunable CW solid-state laser based sources and techniques suitable for mid-infrared laser applications in spectroscopy. The emphasis of our discussion has been to acquaint the reader with the key fundamentals and options in realizing optimized performance, different available sensitivity enhancement schemes and minimum detectable absorbances (10−4 to 10−6 ). Wherever possible, we tried to elucidate the factors that are important for real-world sensors and applications. In this regard we presented three representative examples of laser sources, techniques and their specific applications. Each of these critical areas will constantly undergo incremental improvements of the underlying enabling technologies and lead to further advances in in situ and remote gas sensing techniques. Reliable mid-IR sources in combination with cavity-enhanced spectroscopy and optical fiber technologies will improve detection limits. These new technologies and approaches will also lead to new effective sensor configurations. For example, one could imagine the use of low-loss single mode fibers with high reflective coatings employed as a ring down cavity and combined with evanescent field absorption of a partially stripped fiber. These potential and other new inventions will improve the simplicity, the cost and robustness of spectroscopic laser based gas sensors and broaden their range of applications [219,220,221]. As these new laser sources and spectroscopic techniques evolve in maturity, an emphasis on the instrument replicate sensitivity and precision of quantitative trace gas detection must be given, because it is the unequivocal merit of usefulness in many applications. Such issues may not solely depend on the laser or measurement principle employed, but also on numerous other factors which may be application dependent, such as ambient temperature and pressure conditions. It will thus be important to address these important key factors without increasing the system complexity. Such spectroscopic


497

laser based systems will have a large impact on the means and quality with which we can sense the world around us. Acknowledgements The authors would like to thank Dr. Douglas S. Baer (Los Gatos Research, Inc.), Dr. James F. Kelly (Pacific Northwest National Laboratories), Dr. Anatoliy A. Kosterev and Dr. Robert F. Curl (Rice University), and Dr. Michael E. Webber (Pranalytica, Inc.) for their helpful ideas, comments and invaluable scientific discussions during the preparation of this manuscript.

References 1. L. R. Narasimhan, W. Goodman, C. K. N. Patel: Correlation of breath ammonia with blood urea nitrogen and creatinine during hemodialysis, Proc. National Acad. Sci. 98, 4617–4621 (2001) 447 2. M. E. Webber, M. Pushkarsky, C. K. N. Patel: Fiber-amplifier enhanced photoacoustic spectroscopy using near-infrared tunable diode lasers, Appl. Opt. 42, 2119–2126 (2003) 447, 479 3. J. Reid, D. T. Casidy, R. T. Menzies: Linewidth measurements of tunable diode lasers using heterodyne and étalon techniques, Appl. Opt. 21, 3961– 3965 (1982) 448 4. S. Lundqvist, J. Margolis, J. Reid: Measurements of pressure-broadening coefficients of NO and O3 using a computerized tunable diode laser spectrometer, Appl. Opt. 21, 3109–3113 (1982) 448 5. R. L. Sams, A. Fried: Microphone triggering circuit for elimination of mechanically induced frequency-jitter in diode laser spectrometers: implications for quantitative analysis, Appl. Opt. 26, 3552–3558 (1987) 448 6. R. Sams, A. Fried: Potential sources of systematic errors in tunable-diode-laser absorption measurements, Appl. Spectrosc. 40, 24–29 (1986) 449 7. H. Preier, Z. Feit, J. Fuchs, D. Kostyk, W. Jalenak, J. Sproul: Status of leadsalt diode laser development at spectra-physics, in Monitoring of Gaseous Pollutants by Tunable Diode Lasers, Proc. Int. Symposium, Freiburg, Germany (1988), G. S. R. Grisar, M. Tacke, G. Restelli (Eds.) (Kluwer Academic, Dordrecht 1989) pp. 85–102 449 8. D. L. Wall, J. C. Sproul, Z. Feit, G. W. Sachse: Development of IR tunable diode lasers and source assemblies for atmospheric monitoring and related applications, in Tunable Diode Laser Spectroscopy, Lidar, and DIAL Techniques for Environmental and Industrial Measurements, A. Fried, D. K. Killinger, H. I. Schiff (Eds.) (SPIE, Atlanta, GA 1993) pp. 2–11 449 9. M. Tacke: Recent results in lead-salt laser development at the IPM, in Monitoring of Gaseous Pollutants by Tunable Diode Lasers, Proc. Int. Symposium, Freiburg, Germany (1988), G. S. R. Grisar, M. Tacke, and G. Restelli (Eds.) (Kluwer Academic, Dordrecht 1989) pp. 103–118 449 10. R. Grisar: Monitoring of Gaseous Pollutants by Tunable Diode Lasers, Proc. Int. Symposium, Freiburg, Germany (1991) (Kluwer Academic, Dordrecht 1992) 449

498


11. D. J. Brassington: Tunable diode laser absorption spectroscopy for the measurement of atmospheric species, in Advances in Spectroscopy, R. E. Hester (Ed.) (John Wiley, New York 1995) pp. 83–148 449 12. A. Fried, B. Henry, B. Wert, S. Sewell, J. R. Drummond: Laboratory, groundbased, and airborne tunable diode laser systems: performance characteristics and applications in atmospheric studies, Appl. Phys. B 67, 317–330 (1998) 449, 483, 484, 485 13. P. Werle: A review of recent advances in semiconductor laser-based gas monitors, Spectrochim. Acta 54, 197–236 (1998) 449 14. A. Fried, S. McKeen, S. Sewell, J. Harder, B. Henry, P. Goldan, W. Kuster, E. Williams, K. Baumann, R. Shetter, C. Cantrell: Photochemistry of formaldehyde during the 1993 tropospheric OH photochemistry experiment, J. Geophys. Res. Atmospheres 102, 6283–6296 (1997) 450 15. A. Fried, S. Sewell, B. Henry, B. P. Wert, T. Gilpin, J. R. Drummond: Tunable diode laser absorption spectrometer for ground-based measurements of formaldehyde, J. Geophys. Res. Atmospheres 102, 6253–6266 (1997) 450, 483 16. A. Fried, Y. Wang, C. Cantrell, B. Wert, J. Walega, B. Ridley, E. Atlas, R. Shetter, B. Lefer, M. T. Coffey, J. Hannigan, D. Blake, N. Blake, S. Meinardi, B. Talbot, J. Dibb, E. Scheuer, O. Wingenter, J. Snow, B. Heikes, D. Ehhalt: Tunable diode laser measurements of formaldehdye during the TOPSE 2000: Distributions, trends, and model comparisons, J. Geophys. Res. 108, D4, 8365 (2003) 450, 453, 482, 484, 485 17. A. N. Baranov, Y. Cuminal, G. Boissier, C. Alibert, A. Joullié: Low-threshold laser diodes based on type-II GaInAsSb/GaSb quantum-wells operating at 2.36 µm at room temperature, Electron. Lett. 32, 2279–2280 (1996) 451 18. Y. Rouillard, F. Genty, A. Perona, A. Vicet, D. A. Yarekha, G. Boissier, P. Grech, A. N. Baranov, C. Alibert: Edge and vertical surface emitting lasers around 2.0–2.5 µm and their applications, Philos. Trans. R. Soc. London A 359, 581–597 (2001) 451 19. J. C. Nicolas, A. N. Baranov, Y. Cuminal, Y. Rouillard, C. Alibert: Tunable diode laser absorption spectroscopy of carbon monoxide around 2.35 µm, Appl. Opt. 37, 7906–7911 (1998) 451 20. B. Lane, D. Wu, A. Rybaltowski, H. Yi, J. Diaz, M. Razeghi: Compressively strained multiple quantum well InAsSb lasers emitting at 3.6 µm grown by metal-organic chemical vapor deposition, Appl. Phys. Lett. 70, 443–445 (1997) 451 21. P. Werle, A. Popov: Application of antimonide lasers for gas sensing in the 3–4-µm range, Appl. Opt. 38, 1494–1501 (1999) 451 22. P. Werle, R. Mucke, F. D’Amato, T. Lancia: Near-infrared trace-gas sensors based on room-temperature diode lasers, Appl. Phys. B 67, 307–315 (1998) 451 23. P. Werle: Spectroscopic trace gas analysis using semiconductor diode lasers, Spectrochim. Acta A 52, 805–822 (1996) 451 24. T. C. Hasenberg, R. H. Miles, A. R. Kost, L. West: Recent advances in Sbbased midwave-infrared lasers, IEEE J. Quant. Electron. 33, 1403–1406 (1997) 451 25. F. Capasso, C. Gmachl, R. Paiella, A. Tredicucci, A. L. Hutchinson, D. L. Sivco, J. N. Baillargeon, A. Y. Cho, H. C. Liu: New frontiers in quan-


26.

27. 28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

38.

39. 40.

499

tum cascade lasers and applications, IEEE J. Sel. Topics Quant. Electron. 6, 931–947 (2000) 451, 489 C. Gmachl, A. Straub, R. Colombelli, F. Capasso, D. L. Sivco, A. M. Sergent, A. Y. Cho: Single-mode, tunable distributed-feedback and multiplewavelength quantum cascade lasers, IEEE Quantum Electron. 38, 569–581 (2002) 451 C. Gmachl, F. Capasso, D. L. Sivco, A. Y. Cho: Recent progress in quantum cascade lasers and applications, Rep. Prog. Phys. 64, 1533–1601 (2001) 451 R. M. Williams, J. F. Kelly, J. S. Hartman, S. W. Sharpe, M. S. Taubman, J. L. Hall, F. Capasso, C. Gmachl, D. L. Sivco, J. N. Baillargeon, A. Y. Cho: Kilohertz linewidth from frequency-stabilized mid-infrared quantum cascade lasers, Opt. Lett. 24, 1844–1846 (1999) 451, 452 M. Beck, D. Hofstetter, T. Aellen, J. Faist, U. Oesterle, M. Ilegems, E. Gini, H. Melchior: Continuous wave operation of a mid-infrared semiconductor laser at room temperature, Science 295, 301–305 (2002) 452 R. Kohler, C. Gmachl, A. Tredicucci, F. Capasso, D. L. Sivco, S. N. G. Chu, A. Y. Cho: Single-mode tunable, pulsed, and continuous wave quantumcascade distributed feedback lasers at lambda congruent to 4.6–4.7 µm, Appl. Phys. Lett. 76, 1092–1094 (2000) 452 J. Faist, D. Hofstaetter, M. Beck, T. Aellen, M. Rochat, S. Laser: Bound-tocontinuum and two-phonon resonance quantum cascade lasers for high duty cycle, high temperature operation, IEEE J. Quant. Electron. 38, 533–546 (2002) 452 M. S. Taubman, T. L. Myers, B. D. Cannon, R. M. Williams, F. Capasso, C. Gmachl, D. L. Sivco, A. Y. Cho: Frequency stabilization of quantumcascade lasers by use of optical cavities, Opt. Lett. 27, 2164–2166 (2002) 452 H. Q. Le, G. W. Turner, J. R. Ochoa, M. J. Manfra, C. C. Cook, Y. H. Zhang: Broad wavelength tunability of grating-coupled external cavity mid-infrared semiconductor lasers, Appl. Phys. Lett. 69, 2804–2806 (1996) 452 G. P. Luo, C. Peng, H. Q. Le, S. S. Pei, W. Y. Hwang, B. Ishaug, J. Um, J. N. Baillargeon, C. H. Lin: Grating-tuned external-cavity quantum-cascade semiconductor lasers, Appl. Phys. Lett. 78, 2834–2836 (2001) 452, 453 K. G. Libbrecht, J. L. Hall: A low noise high-speed diode laser controller, Rev. Sci. Instrum. 64, 2133 (1993) 452 R. Q. Yang, J. L. Bradshaw, J. D. Bruno, J. T. Pham, D. E. Wortman: Midinfrared type-II Interband cascade lasers, IEEE Quantum Electron. 38, 559– 568 (2002) 453 G. P. Luo, C. Peng, H. Q. Le, S. S. Pei, H. Lee, W. Y. Hwang, B. Ishaug, J. Zheng: Broadly wavelength-tunable external cavity mid-infrared quantum cascade lasers, IEEE J. Quant. Electron. 38, 486–494 (2002) 453 D. D. Nelson, M. S. Zahniser, J. B. McManus, C. E. Kolb, J. L. Jimenez: A tunable diode laser system for the remote sensing of on-road vehicle emissions, Appl. Phys. B 67, 433–441 (1998) 453 D. E. Cooper, C. B. Carlisle: High-sensitivity FM spectroscopy with a lead-salt diode laser, Opt. Lett. 13, 719–721 (1988) 454 L. F. Mollenauer, White, J. C.: General principles and some common features, in Tunable Lasers, L. F. Mollenauer, White, J. C. (Eds.) (Springer, Berlin Heidelberg 1987) pp. 1–18 454

500


41. P. F. Moulton: Tunable solid-state lasers, Proc. IEEE 80, 348–364 (1992) 454 42. N. P. Barnes, F. Allario: Tunable solid-state lasers: an emerging technology for remote sensing of planetary atmospheres, Proc. SPIE 868, 107–116 (1988) 454 43. S. K¨ uck: Laser-related spectroscopy of ion-doped crystals for solid-state lasers, Applied Physics B-Lasers & Optics B 72, 515–562 (2001) 454 44. A. Sennaroglu, A. O. Konca, C. R. Pollock: Continuous-wave power performance of a 2.47-µm Cr2+ : ZnSe laser: Experiment and modeling, IEEE J. Quant. Electron. 36, 1199–1205 (2000) 454 45. I. T. Sorokina, E. Sorokin, A. Di Lieto, M. Tonelli, R. H. Page, K. I. Schaffers: Efficient broadly tunable continuous-wave Cr2+ : ZnSe laser, J. Opt. Soc. Am. B 18, 926–930 (2001) 454 46. S. Schiller, J. Mlynek: Special Issue, Continuous-wave optical parametric oscillators, materials, devices, applications, Appl. Phys. B 66, 664–764 (Springer, Berlin, Heidelberg 1998) 455 47. P. E. Powers, T. J. Kulp, S. E. Bisson: Continuous tuning of a continuouswave periodically poled lithium niobate optical parametric oscillator by use of a fan-out grating design, Opt. Lett. 23, 59–61 (1998) 455 48. R. J. Lang, D. G. Mehyus, D. F. Welch: External cavity, continuously tunable wavelength source, US patent no. 5,771,252 (1998) 456 49. A. T. Schremer, C. L. Tang: External-cavity semiconductor laser with 1000 GHz continuous piezoelectric tuning range, IEEE Photon. Technol. Lett. 2, 3–5 (1990) 456 50. D. Wandt, M. Laschek, K. Przyklenk, A. T¨ unnermann, H. Welling: External cavity laser diode with 40 nm continuous tuning range around 825 nm, Opt. Commun. 130, 81–84 (1996) 456 51. T. Kiguchi, A. Uematsu, M. Kitano, H. Ogura: Grating external cavity diode lasers with broad tunable range and narrow spectral linewidth for highresolution spectroscopy, Jap. J. Appl. Phys. 35, 5890–5895 (1996) 456 52. B. Pezeshki: New approaches to laser tuning, Opt. Photon. News 12, 34–38 (2001) 456 53. L. A. Coldren: Monolithic tunable diode lasers, IEEE J. Sel. Topics Quant. Electron 6, 988–999 (2000) 456 54. B. Schmidt, S. Illek, R. Gessner, M. C. Amann: Design and realization of a buried-heterostructure tunable-twin-guide laser diode with electrical blocking regions, IEEE J. Quant. Electron. 35, 794–802 (1999) 456 55. G. Morthier, B. Moeyersoon, R. Baets: A lambda/4-shifted sampled or superstructure grating widely tunable twin-guide laser, IEEE Photon. Technol. Lett. 13, 1052–1054 (2001) 456 56. H. Ishii, H. Tanobe, F. Kano, Y. Tohmori, Y. Kondo, Y. Yoshikuni: Broadrange wavelength coverage (62.4 nm) with superstructure-grating DBR laser, Electron. Lett. 32, 454–455 (1996) 456 57. E. T. Wetjen, D. M. Sonnenfroh, M. G. Allen, T. F. Morse: Demonstration of a rapidly tuned Er3+ -doped fiber laser for sensitive gas detection, Appl. Opt. 38, 3370–3375 (1998) 457 58. H. Simonsen, J. Henningsen, S. Sogaard, J. E. Pedersen: CO2 stabilized Er3+ doped fibre laser at 1578 nm, Eur. Frequency Time Forum (EFTF) (2000) 457


501

59. I. Freitag, A. T¨ unnermann, H. Welling: Power scaling of monolithic miniature Nd:YAG ring lasers to output powers of several watts, Opt. Commun. 115, 511 (1995) 457 60. D. Boucher, W. Chen, J. Burie, P. Peze: A novel CW optical laser-based difference-frequency infrared spectrometer, Ann. Phys. (Paris) 23, 247–248 (1998) 457 61. W. D. Chen, J. Burie, D. Boucher: Investigation on infrared laser absorption spectroscopy measurement of acetylene trace quantities, Infrared Phys. Technol. 41, 339–348 (2000) 457 62. W. Schade, T. Blanke, U. Willer, C. Rempel: Compact tunable mid-infrared laser source by difference frequency generation of two diode-lasers, Appl. Phys. B 63, 99–102 (1996) 457 63. M. Seiter, M. W. Sigrist: Trace-gas sensor based on mid-IR differencefrequency generation in PPLN with saturated output power, Infrared Phys. Technol. 41, 259–269 (2000) 457 64. D. Lee, T. Kaing, J. J. Zondy: An all-diode-laser-based, dual-cavity AgGaS2 CW difference-frequency source for the 9–11 µm range, Appl. Phys. B 67, 363–367 (1998) 457, 459 65. H. D. Kronfeldt, G. Basar, B. Sumpf: Application of a CW tunable infrared spectrometer based on difference-frequency generation in AgGaS2 for selfbroadening investigations of NO at 5 µm, J. Opt. Soc. Am. B 13, 1859–1863 (1996) 457 66. M. Seiter, D. Keller, M. W. Sigrist: Broadly tunable difference-frequency spectrometer for trace gas detection with noncollinear critical phase-matching in LiNbO3 , Appl. Phys. B 67, 351–356 (1998) 457 67. M. Seiter, M. W. Sigrist: Compact gas sensor using a pulsed differencefrequency laser spectrometer, Opt. Lett. 24, 110–112 (1999) 457 68. D. G. Lancaster, A. Fried, B. Wert, B. Henry, F. K. Tittel: Differencefrequency-based tunable absorption spectrometer for detection of atmospheric formaldehyde, Appl. Opt. 39, 4436–4443 (2000) 457, 459 69. D. Richter, D. G. Lancaster, F. K. Tittel: Development of an automated diodelaser-based multicomponent gas sensor, Appl. Opt. 39, 4444–4450 (2000) 457, 459, 461, 492 70. W. D. Chen, F. Cazier, F. Tittel, D. Boucher: Measurements of benzene concentration by difference-frequency laser absorption spectroscopy, Appl. Opt. 39, 6238–6242 (2000) 457, 459 71. D. G. Lancaster, D. Richter, F. K. Tittel: Portable fiber-coupled diode-laserbased sensor for multiple trace gas detection, Appl. Phys. B 69, 459–465 (1999) 457, 459 72. U. Simon, F. K. Tittel: Nonlinear optical frequency conversion techniques, in Atomic, Molecular, and Optical Physics: Electromagnetic Radiation, R. Celotta, R. G. Hulet, F. B. Dunning (Eds.), Experimental Methods in the Physical Sciences 29, 231–278 (Academic Press, New York 1997) 457 73. A. S. Pine: Doppler-limited molecular spectroscopy by difference-frequency mixing, J. Opt. Soc. Am. 64, 1683–1690 (1974) 458 74. B. Sumpf, D. Rehle, T. Kelz, H. D. Kronfeldt: A tunable diode-laser spectrometer for the MIR region near 7.2 µm applying difference-frequency generation in AgGaSe2 , Applied Physics B-Lasers & Optics 67, 369–373 (1998) 459

502


75. P. A. Budni, L. A. Pomeranz, M. L. Lemons, C. A. Miller, J. R. Mosto, E. P. Chicklis: Efficient mid-infrared laser using 1.9-µm-pumped Ho:YAG and ZnGeP2 optical parametric oscillators, J. Opt. Soc. Am. B 17, 723–728 (2000) 459 76. W. C. Eckhoff, R. S. Putnam, S. X. Wang, R. F. Curl, F. K. Tittel: A continuously tunable long-wavelength CW IR source for high-resolution spectroscopy and trace-gas detection, Appl. Phys. B 63, 437–441 (1996) 459 77. W. Chen, G. Mouret, D. Boucher: Difference-frequency laser spectroscopy detection of acetylene trace constituent, Appl. Phys. B 67, 375–378 (1998) 459 78. R. S. Putnam, D. G. Lancaster: Continuous-wave laser spectrometer automatically aligned and continuously tuned from 11.8 to 16.1 µm by use of diodelaser-pumped difference-frequency generation in GaSe, Appl. Opt. 38, 1513– 1522 (1999) 459 79. W. K. Burns, W. McElhanon, L. Goldberg: Second harmonic generation in field poled, quasi-phase matched, bulk LiNbO3 , IEEE Photon. Technol. Lett. 6, 252–254 (1994) 459 80. L. E. Myers: Periodically poled materials for nonlinear optics, in Advances in Lasers and Applications, D. M. Finlayson, B. D. Sinclair (Eds.) (Scottish Univ. Science Press, 1999) pp. 141–180 459 81. K. Fradkin-Kashi, A. Arie, P. Urenski, G. Rosenman: Mid-infrared differencefrequency generation in periodically poled KTiOAsO4 and application to gas sensing, Opt. Lett. 25, 743–745 (2000) 459 82. K. Fradkin, A. Arie, A. Skliar, G. Rosenman: Tunable midinfrared source by difference frequency generation in bulk periodically poled KTiOPO4 , Appl. Phys. Lett. 74, 914–916 (1999) 459 83. K. Fradkin-Kashi, A. Arie: Multiple-wavelength quasi-phase-matched nonlinear interactions, IEEE J. Quant. Electron. 35, 1649–1656 (1999) 459 84. S. Sanders, R. J. Lang, L. E. Myers, M. M. Fejer, R. L. Byer: Broadly tunable mid-ir radiation source based on difference frequency mixing of high power wavelength-tunable laser diodes in bulk periodically poled LiNbO3 , Electron. Lett. 32, 218–219 (1996) 459 85. W. Chen, G. Mouret, D. Boucher, F. K. Tittel: Mid-infrared trace gas detection using continuous-wave difference frequency generation in periodically poled RbTiOAsO4 , Appl. Phys. B 72, 873–876 (2001) 459 86. D. Mazzotti, P. De Natale, G. Giusfredi, C. Fort, J. A. Mitchell, L. W. Hollberg: Difference-frequency generation in PPLN at 4.25 µm: an analysis of sensitivity limits for DFG spectrometers, Appl. Phys. B 70, 747–750 (2000) 459 87. O. Levi, T. J. Pinguet, T. Skauli, L. A. Eyeres, K. R. Parameswaran, J. S. Harris, Jr., M. M. Fejer, T. J. Kulp, S. E. Bisson, B. Gerard, E. Lallier, L. Becouarn: Difference frequency generation of 8-micron radiation in orientationpatterned GaAs, Opt. Lett. 27, 2091–2093 (2002) 459 88. D. G. Lancaster, D. Richter, R. F. Curl, F. K. Tittel: Real-time measurements of trace gases using a compact difference-frequency-based sensor operating at 3.5 µm, Appl. Phys. B 67, 339–345 (1998) 459 89. D. G. Lancaster, L. Goldberg, J. Koplow, R. F. Curl, F. K. Tittel: Fibre coupled difference frequency generation utilising ytterbium-doped fibre amplifier and periodically poled LiNbO3 , Electron. Lett. 34, 1345–1347 (1998) 459


503

90. D. G. Lancaster, D. Richter, R. F. Curl, F. K. Tittel, I. Goldberg, J. Koplow: High-power continuous-wave mid-infrared radiation generated by difference frequency mixing of diode-laser-seeded fiber amplifiers and its application to dual-beam spectroscopy, Opt. Lett. 24, 1744–1746 (1999) 459 91. D. G. Lancaster, R. Weidner, D. Richter, F. K. Tittel, J. Limpert: Compact CH4 sensor based on difference frequency mixing of diode lasers in quasiphasematched LiNbO3 , Opt. Commun. 175, 461–468 (2000) 459 92. D. Richter, D. G. Lancaster, R. F. Curl, W. Neu, F. K. Tittel: Compact midinfrared trace gas sensor based on difference-frequency generation of two diode lasers in periodically poled LiNbO3 , Appl. Phys. B 67, 347–350 (1998) 459 93. D. Richter, M. Erdelyi, F. K. Tittel: Ultra-compact mid-IR spectroscopic source based on frequency converted Yb-Er/Yb fiber amplified CW diode lasers, in C. Marshall (Ed.): Adv. Solid-State Lasers, OSA Trends Opt. Photon. 50 (Opt. Soc. Am., Washington, DC 2001) pp. 96–100 459 94. J. J. Zondy: The effects of focusing in type-I and type-II difference-frequency generations, Opt. Commun. 149, 181–206 (1998) 460 95. T. Topfer, K. P. Petrov, Y. Mine, D. Jundt, R. F. Curl, F. K. Tittel: Roomtemperature mid-infrared laser sensor for trace gas detection, Appl. Opt. 36, 8042–8049 (1997) 460 96. M. H. Chou, M. A. Arbore, M. M. Fejer: Adiabatically tapered periodic segmentation of channel waveguides for mode-size transformation and fundamental mode excitation, Opt. Lett. 21, 794–796 (1996) 460 97. M. H. Chou, K. R. Parameswaran, M. M. Fejer, I. Brener: Multiple-channel wavelength conversion by use of engineered quasi-phase-matching structures in LiNbO3 waveguides, Opt. Lett. 24, 1157–1159 (1999) 460 98. K. P. Petrov, A. T. Ryan, T. L. Patterson, L. Huang, S. J. Field, D. J. Bamford: Mid-infrared spectroscopic detection of trace gases using guided-wave difference-frequency generation, Appl. Phys. B 67, 357–361 (1998) 460 99. K. P. Petrov, A. P. Roth, T. L. Patterson, T. P. S. Thoms, L. Huang, A. T. Ryan, D. J. Bamford: Efficient difference-frequency mixing of diode lasers in lithium niobate channel waveguides, Appl. Phys. B 70, 777–782 (2000) 460 100. D. Hofmann, G. Schreiber, C. Haase, H. Herrmann, W. Grundkotter, R. Ricken, W. Sohler: Quasi-phase-matched difference-frequency generation in periodically poled Ti:LiNbO3 channel waveguides, Opt. Lett. 24, 896–898 (1999) 460 101. P. E. Powers, T. J. Kulp, S. E. Bisson: Continuous tuning of a continuouswave periodically poled lithium niobate optical parametric oscillator by use of a fan-out grating design, Opt. Lett. 23, 159–161 (1998) 461, 462 102. G. M. Gibson, M. H. Dunn, M. J. Padgett: Application of a continuously tunable, CW optical parametric oscillator for high-resolution spectroscopy, Opt. Lett. 23, 40–42 (1998) 461, 462 103. T. J. Edwards, G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh, H. Karlsson, G. Arvidsson, F. Laurell: Continuous-wave singly resonant optical parametric oscillator based on periodically poled RbTiOAsO4 , Opt. Lett. 23, 837–839 (1998) 461, 462 104. E. V. Kovalchuk, D. Dekorsy, A. I. Lvovsky, C. Braxmaier, J. Mlynek, A. Peters, S. Schiller: High-resolution Doppler-free molecular spectroscopy with a continuous-wave optical parametric oscillator, Opt. Lett. 26, 1430–1432 (2001) 461, 462

504


105. R. Altahtamouni, K. Bencheikh, R. Storz, K. Schneider, M. Lang, J. Mlynek, S. Schiller: Long-term stable operation and absolute frequency stabilization of a doubly resonant parametric oscillator, Appl. Phys. B 66, 733–739 (1998) 461, 462 106. M. E. Klein, D. H. Lee, J. P. Meyn, B. Beier, K. J. Boller, R. Wallenstein: Diode-pumped continuous-wave widely tunable optical parametric oscillator based on periodically poled lithium tantalate, Opt. Lett. 23, 831–833 (1998) 461, 462 107. G. A. Turnbull, M. H. Dunn, M. Ebrahimzadeh: Continuous-wave, intracavity optical parametric oscillators – an analysis of power characteristics, Appl. Phys. B 66, 701–710 (1998) 461 108. W. R. Bosenberg, A. Drobshoff, J. I. Alexander, L. E. Myers, R. L. Byer: Continuous-wave singly resonant optical parametric oscillator based on periodically poled LiNbO3 , Opt. Lett. 21, 713–715 (1996) 462 109. L. E. Myers, R. C. Eckardt, M. M. Fejer, R. L. Byer, W. R. Bosenberg: Multigrating quasi-phase-matched optical parametric oscillator in periodically poled LiNbO3 , Opt. Lett. 21, 591–593 (1996) 462 110. L. E. Myers, W. R. Bosenberg: Periodically poled lithium niobate and quasiphase-matched optical parametric oscillators, IEEE J. Quant. Electron. 33, 1663–1672 (1997) 462 111. F. Hanson, P. Poirer, M. A. Arbore: Single-frequency mid-infrared optical parametric oscillator source for coherent laser radar, Opt. Lett. 26, 1794–1796 (2001) 462 112. M. A. Arbore, M. M. Fejer: Singly resonant optical parametric oscillation in periodically poled lithium niobate waveguides, Opt. Lett. 22, 151–153 (1997) 462 113. A. J. Henderson, P. M. Roper, L. A. Borschowa, R. D. Mead: Stable, continuously tunable operation of a diode-pumped doubly resonant optical parametric oscillator, Opt. Lett. 25, 1264–1266 (2000) 462 114. M. van Herpen, S. te Lintel Hekkert, S. E. Bisson, F. J. M. Harren: Wide singlemode tuning of a 3.0–3.8 µm, 700 mW continuous-wave Nd:YAG-pumped optical parametric oscillator based on periodically poled lithium niobate, Opt. Lett. 27, 640–642 (2002) 462 115. S. Schiller, K. Schneider, J. Mlynek: Theory of an optical parametric oscillator with resonant pump and signal, J. Opt. Soc. Am. B 16, 1512–1524 (1999) 462 116. U. Strossner, J. P. Meyn, R. Wallenstein, P. Urenski, A. Arie, G. Rosenman, J. Mlynek: Single-frequency continuous-wave optical parametric oscillator system with an ultrawide tuning range of 550 to 2830 nm, J. Opt. Soc. Am. B 19, 1419–1424 (2002) 462 117. L. S. Rothman, C. P. Rinsland, A. Goldman, S. T. Massie, D. P. Edwards, J. M. Flaud, A. Perrin, C. Camypeyret, V. Dana, J. Y. Mandin, J. Schroeder, A. McCann, R. R. Gamache, R. B. Wattson, K. Yoshino, K. V. Chance, K. W. Jucks, L. R. Brown, V. Nemtchinov, P. Varanasi: The HITRAN molecular spectroscopic database and hawks (HITRAN atmospheric workstation) 1996 edn., J. Quant. Spectrosc. Radiat. Transfer 60, 665–710 (1998) 463, 466, 472 118. N. Jacquinet-Husson, E. Arie, J. Ballard, A. Barbe, G. Bjoraker, B. Bonnet, L. R. Brown, C. Camy-Peyret, J. P. Champion, A. Chedin, A. Chursin, C. Clerbaux, G. Duxbury, J. M. Flaud, N. Fourrie, A. Fayt, G. Graner, R. Gamache,


119.

120.

121. 122. 123. 124.

125.

126. 127.

128.

129.

130.

131. 132.

133. 134.

505

A. Goldman, V. Golovko, G. Guelachvili, J. M. Hartmann, J. C. Hilico, J. Hillman, G. Lefevre et al.: The 1997 spectroscopic GEISA databank, J. Quant. Spectrosc. Radiat. Transfer 62, 205–254 (1999) 463, 466 J. Humlicek: An efficient method of evaluation of the complex probability function: the Voigt function and its derivatives, J. Quant. Spectrosc. Radiat. Transfer 21, 309 (1979) 466 U. Platt: Air monitoring by differential optical absorption spectroscopy, in Encyclopedia of Analytical Chemistry, L. E. Myers (Ed.) (Wiley, Chichester 2000) pp. 1936–1959 467 S. Svanberg: Atomic and Molecular Spectroscopy: Basic Aspects and Practical Applications (Springer, Berlin, Heidelberg 2001) 467 G. D. Houser, E. Garmire: Balanced detection technique to measure small changes in transmission, Appl. Opt. 33, 1059–1062 (1994) 468 P. C. D. Hobbs: Ultrasensitive laser measurements without tears, Appl. Opt. 36, 903–920 (1997) 469 M. G. Allen, K. L. Carleton, S. J. Davis, W. J. Kessler, C. E. Otis, D. A. Palombo, D. M. Sonnenfroh: Ultra-sensitive dual beam absorption and gain spectroscopy: Applications for near-IR and visible diode laser sensors, Appl. Opt. 34, 3240 (1995) 469 D. M. Sonnenfroh, W. T. Rawlins, M. G. Allen, C. Gmachl, F. Capasso, A. L. Hutchinson, D. L. Sivco, J. N. Baillargeon, A. Y. Cho: Application of balanced detection to absorption measurements of trace gases with roomtemperature, quasi-CW quantum-cascade lasers, Appl. Opt. 40, 812–820 (2001) 469 G. C. Bjorklund: Frequency-modulation spectroscopy: a new method of measuring weak absorption and dispersions, Opt. Lett. 5, 15 (1980) 470 D. E. Cooper, T. F. Gallagher: Double frequency modulation spectroscopy: high modulation frequency with low-bandwidth detectors, Appl. Opt. 24, 1327 (1985) 470 R. Grosskloss, P. Kersten, W. Demtroeder: Sensitive amplitude and phasemodulated absorption spectroscopy with a continuously tunable diode laser, Appl. Phys. B 58, 137 (1994) 470 D. S. Bomse, A. C. Stanton, J. A. Silver: Frequency modulation and wavelength modulation spectroscopies: Comparison of experimental methods using a lead-salt diode laser, Appl. Opt. 31, 718–731 (1992) 471 F. S. Pavone, M. Inguscio: Frequency and wavelength modulation spectroscopties: comparison and experimental methods of using an AlGaAs diode laser, Appl. Phys. B 56, 118 (1993) 471 D. B. Oh, D. C. Hovde: Wavelength modulation detection of acetylene with near-IR external cavity diode laser, Appl. Opt. 34, 7002–7005 (1995) 471 K. Uehara: Dependence of harmonic signals on sample-gas parameters in wavelength-modulation spectroscopy for precise absorption measurements, Appl. Phys. B 67, 517–523 (1998) 471 P. Werle: Laser excess noise and interferometric effects in frequency-modulated diode laser spectrometers, Appl. Phys. B 60, 499–506 (1995) 471 M. S. Zahniser, D. D. Nelson, J. B. McManus, P. Kebabian: Measurement of trace gas fluxes using tunable diode laser spectroscopy, Phil. Trans. R. Soc. London A 351, 371–382 (1995) 471, 483

506


135. A. Fried, J. R. Drummond, B. Henry, J. Fox: Versatile integrated tunable diode laser system for high precision: application for ambient measurements of OCS, Appl. Opt. 30, 1916–1932 (1991) 471 136. T. Iguchi: Modulation waveforms for second-harmonic detection with tunable diode lasers, J. Opt. Soc. Am. B 3, 419–423 (1986) 471 137. R. G. Pilston, J. U. White: A long path gas absorption cell, J. Opt. Soc. Am. 44, 572–573 (1954) 472 138. J. U. White: Long optical paths of large aperture, J. Opt. Soc. Am. 32, 285– 288 (1942) 472 139. D. Herriot, H. Kogelnik, R. Kompfner: Off-axis paths in spherical mirror interferometers, Appl. Opt. 3, 523 (1964) 472 140. S. M. Chernin, E. G. Barskaya: Optical multi-pass matrix systems, Appl. Opt. 30, 51–58 (1991) 472 141. J. B. McManus, P. L. Kebabian, M. S. Zahniser: Astigmatic mirror multi-pass absorption cells for long-path length spectroscopy, Appl. Opt. 34, 3336 (1995) 472 142. A. O’Keefe, D. A. G. Deacon: Cavity ring-down optical spectrometer for absorption measurements using pulsed laser sources, Rev. Sci. Instrum. 59, 2544 (1988) 473, 475 143. J. Lauterbach, D. Kleine, K. Kleinermanns, P. Hering: Cavity-ring-down spectroscopic studies of NO2 in the region around 613 nm, Appl. Phys. B 71, 873–876 (2000) 474 144. D. Kleine, S. Stry, J. Lauterbach, K. Kleinermanns, P. Hering: Measurement of the absolute intensity of the fifth CH stretching overtone of benzene using cavity ring-down spectroscopy, Chem. Phys. Lett. 312, 185–190 (1999) 474 145. G. Berden, R. Peeters, G. Meijer: Cavity ring-down spectroscopy: Experimental schemes and applications, Int. Rev. Phys. Chem. 19, 565–607 (2000) 475 146. R. Peeters, G. Berden, A. Apituley, G. Meijer: Open-path trace gas detection of ammonia based on cavity-enhanced absorption spectroscopy, Appl. Phys. B 71, 231–236 (2000) 475 147. R. Peeters, G. Berden, G. Meijer: Sensitive absorption techniques for spectroscopy, Am. Lab. 33, 60 (2001) 475 148. Y. B. He, B. J. Orr: Rapidly swept, continuous-wave cavity ringdown spectroscopy with optical heterodyne detection, Single- and multi-wavelength sensing of gases, Appl. Phys. B 75, 267–280 (2002) 475 149. K. K. Lehmann, D. Romanini: The superposition principle and cavity ringdown spectroscopy, J. Chem. Phys. 105, 10263–10277 (1996) 475 150. B. A. Paldus, J. S. Harris, J. Martin, J. Xie, R. N. Zare: Laser diode cavity ring-down spectroscopy using acousto-optic modulator stabilization, J. Appl. Phys. 82, 3199–3204 (1997) 475 151. B. A. Paldus, C. C. Harb, T. G. Spence, B. Wilke, J. Xie, J. S. Harris, R. N. Zare: Cavity-locked ring-down spectroscopy, J. Appl. Phys. 83, 3991– 3997 (1998) 475 152. J. J. Scherer, J. B. Paul, A. O’Keefe, R. J. Saykally: Cavity ringdown laser absorption spectroscopy, Chem. Rev. 97, 25–51 (1997) 475 153. R. A. Provencal, J. B. Paul, E. Michael, R. J. Saykally: Absorption spectroscopy technique provides extremely high sensitivity, Photon. Spectra 32, 159 ff. (1998) 475 154. J. B. Paul, R. J. Saykally: Cavity ring-down laser absorption spectroscopy, Anal. Chem. 69, A287–A292 (1997) 475


507

155. K. W. Busch, M. A. Busch: Cavity Ring-Down Spectroscopy, An UltratraceAbsorption Measurement Technique (American Chemical Society, Washington, DC 1999) 475 156. J. J. Scherer, D. Voelkel, D. J. Rakestraw, J. B. Paul, C. P. Collier, A. O’Keefe, R. J. Saykally: Infrared Cavity Ringdown Laser Absorption Spectroscopy (IRCRLAS), Chem. Phys. Lett. 245, 273–280 (1995) 475, 476 157. K. K. Lehmann: Ring-down cavity spectroscopy cell using continuous wave excitation for trace species detection, US patent no. 5,528,040 (1996) 476 158. D. Romanini, A. A. Kachanov, N. Sadeghi, F. Stoeckel: CW cavity ring down spectroscopy, Chem. Phys. Lett. 264, 316–322 (1997) 476 159. D. Romanini, A. A. Kachanov, F. Stoeckel: Diode laser cavity ring down spectroscopy, Chem. Phys. Lett. 270, 538–545 (1997) 476 160. B. A. Paldus, C. C. Harb, T. G. Spence, R. N. Zare, C. Gmachl, F. Capasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho: Cavity ring-down spectroscopy using mid-infrared quantum-cascade lasers, Opt. Lett. 25, 666– 668 (2000) 476, 487 161. A. A. Kosterev, A. L. Malinovsky, F. K. Tittel, C. Gmachl, F. Capasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho: Cavity ringdown spectroscopic detection of nitric oxide with a continuous-wave quantumcascade laser, Appl. Opt. 40, 5522–5529 (2001) 476, 486, 487 162. K. W. Aniolek, P. E. Powers, T. J. Kulp, B. A. Richman, S. E. Bisson: Cavity ring-down laser absorption spectroscopy with a 1 kHz mid-infrared periodically poled lithium niobate optical parametric generator optical parametric amplifier, Chem. Phys. Lett. 302, 555–562 (1999) 476 163. R. J. Saykally, R. Casaes: Cavity ring-down technique measures absorption, Laser Focus World 37, 159–162 (2001) 476 164. A. O’Keefe: Integrated cavity output analysis of ultra-weak absorption, Chem. Phys. Lett. 293, 331–336 (1998) 476 165. R. Engeln, G. Vonhelden, G. Berden, G. Meijer: Phase shift cavity ring down absorption spectroscopy, Chem. Phys. Lett. 262, 105–109 (1996) 476 166. R. Engeln, G. Berden, R. Peeters, G. Meijer: Cavity enhanced absorption and cavity enhanced magnetic rotation spectroscopy, Rev. Sci. Instrum. 69, 3763–3769 (1998) 476 167. J. B. Paul, L. Lapson, J. G. Anderson: Ultrasensitive absorption spectroscopy with a high-finesse optical cavity and off-axis alignment, Appl. Opt. 40, 4904– 4910 (2001) 476, 477, 486 168. M. Murtz, B. Frech, W. Urban: High-resolution cavity leak-out absorption spectroscopy in the 10-µm region, Appl. Phys. B 68, 243–249 (1999) 476 169. L. Menzel, A. A. Kosterev, R. F. Curl, F. K. Tittel, C. Gmachl, F. Capasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho, W. Urban: Spectroscopic detection of biological NO with a quantum cascade laser, Appl. Phys. B 72, 859–863 (2001) 476, 486 170. D. S. Baer, J. B. Paul, M. Gupta, A. O’Keefe: Sensitive absorption measurements in the near-infrared region using off-axis integrated cavity output spectroscopy, Appl. Phys. B 75, 261–265 (2002) 477 171. L. S. Ma, J. Ye, P. Dube, J. L. Hall: Ultrasensitive frequency-modulation spectroscopy enhanced by a high-finesse optical cavity: Theory and application to overtone transitions of C2 H2 and C2 HD, J. Opt. Soc. Am. B 16, 2255–2268 (1999) 477

508


172. J. Ye, L. S. Ma, J. L. Hall: Sub-doppler optical frequency reference at 1.064-µm by means of ultrasensitive cavity-enhanced frequency modulation spectroscopy of a C2 HD overtone transition, Opt. Lett. 21, 1000–1002 (1996) 477 173. M. S. Taubman, R. M. Williams, T. L. Myers, F. Capasso, C. Gmachl, D. L. Sivco, A. L. Hutchinson, A. Y. Cho: Sub-Doppler NICE-OHMS spectroscopy at 8.5 microns using a quantum cascade laser, in OSA Trends in Optics and Photonics (TOPS) Vol 73, Conference on Lasers and Electro-Optics, OSA Technical Digest (Opt. Soc. of Am., Washington, DC., 2002) pp.625–626 478 174. M. A. Moeckli, C. Hilbes, M. W. Sigrist: Photoacoustic multicomponent gas analysis using a Levenberg–Marquardt fitting algorithm, Appl. Phys. B 67, 449–458 (1998) 478 175. J. Oomens, H. Zuckermann, S. Persijn, D. H. Parker, F. J. M. Harren: COlaser-based photoacoustic trace-gas detection – applications in postharvest physiology, Appl. Phys. B 67, 459–466 (1998) 478 176. M. A. Moeckli, M. Fierz, M. W. Sigrist: Emission factors for ethene and ammonia from a tunnel study with a photoacoustic trace gas detection system, Environ. Sci. Technol. 30, 2864–2867 (1996) 478 177. M. Nagele, M. W. Sigrist: Mobile laser spectrometer with novel resonant multipass photoacoustic cell for trace-gas sensing, Appl. Phys. B 70, 895–901 (2000) 478 178. B. G. Ageev, Y. N. Ponomarev, V. A. Sapozhnikova: Laser photoacoustic spectroscopy of biosystems gas exchange with the atmosphere, Appl. Phys. B 67, 467–473 (1998) 478 179. F. J. M. Harren, G. Cotti, J. Oomens, S. L. Hekkert: Photoacoustic spectroscopy in trace gas monitoring, in Encyclopedia of Analytical Chemistry, R. A. Meyers (Ed.) (Wiley, Chichester 2000) pp. 2203–2226 478 180. A. Bohren, M. W. Sigrist: Optical parametric oscillator based difference frequency laser source for photoacoustic trace gas spectroscopy in the 3-µm midIR range, Infrared Phys. Technol. 38, 423–435 (1997) 478 181. Z. Bozoki, J. Sneider, Z. Gingl, A. Mohacsi, M. Szakall, Z. Bor, G. Szabo: A high-sensitivity, near-infrared tunable-diode-laser-based photoacoustic watervapour-detection system for automated operation, Measurement Sci. Technol. 10, 999–1003 (1999) 478 182. F. G. C. Bijnen, F. J. M. Harren, J. H. P. Hackstein, J. Reuss: Intracavity CO laser photoacoustic trace gas detection – cyclic CH4 , H2 O and CO2 emission by cockroaches and scarab beetles, Appl. Opt. 35, 5357–5368 (1996) 478 183. V. A. Kapitanov, Y. U. N. Ponomarev, K. Song, H. K. Cha, J. Lee: Resonance photoacoustic spectroscopy and gas analysis of gaseous flow at reduced pressure, Appl. Phys. B 73, 745–750 (2001) 478 184. M. B. Pushkarsky, M. E. Webber, O. Baghdassarian, L. R. Narasimhan, C. K. N. Patel: Laser-based photoacoustic ammonia sensors for industrial applications, Appl. Phys. B 75, 391–396 (2002) 479 185. A. Schmohl, A. Miklos, P. Hess: Detection of ammonia by photoacoustic spectroscopy with semiconductor lasers, Appl. Opt. 41, 1815–1823 (2002) 479 186. A. A. Kosterev, Y. A. Bakhirkin, R. F.Curl, F. K. Tittel: Quartz-enhanced photoacoustic spectroscopy, Opt. Lett. 27, 1902–1904 (2002) 479 187. A. Fried, W. W. Berg: Photoacoustic detection of HCl, Opt. Lett. 8, 160–162 (1983) 479


509

188. M. Avery-Owens, C. C. Davis, R. R. Dickerson: A photothermal interferometer for gas-phase ammonia detection, Anal. Chem. 71, 1391–1399 (1999) 479 189. B. A. Paldus, T. G. Spence, R. N. Zare, J. Oomens, F. J. M. Harren, D. H. Parker, C. Gmachl, F. Cappasso, D. L. Sivco, J. N. Baillargeon, A. L. Hutchinson, A. Y. Cho: Photoacoustic spectroscopy using quantumcascade lasers, Opt. Lett. 24, 178–180 (1999) 480 190. D. Hofstetter, M. Beck, J. Faist, M. Nagele, M. W. Sigrist: Photoacoustic spectroscopy with quantum cascade distributed-feedback lasers, Opt. Lett. 26, 887–889 (2001) 480 191. P. Werle, R. Mucke, F. Slemr: The limits of signal averaging in atmospheric trace-gas monitoring by tunable diode-laser absorption spectroscopy (TDLAS), Appl. Phys. B 57, 131–139 (1993) 483, 484 192. S. Sewell, A. Fried, B. Henry, J. R. Drummond: A field diode laser spectrometer employing an astigmatic Herriott cell, in Tunable Diode Laser Spectroscopy, Lidar, and DIAL Techniques for Environmental and Industrial Measurements, A. Fried, D. K. Killinger, H. I. Schiff (Eds.) (SPIE, Atlanta, GA 1993) pp. 72–80 484 193. A. Fried, B. P. Wert, J. G. Walega, D. Richter, W. T. Potter: Airborne measurements of formaldehyde employing a high-performance tunable diode laser absorptoion system, in Diode Lasers and Applications in Atmospheric Sensing, A. Fried (Ed.), Proc. SPIE 4817, 177–183 (2002) 484 194. R. W. Waynant, I. K. Ilev, I. Gannot: Mid-infrared laser applications in medicine and biology, Phil. Trans. R. Soc. London A 635–644 (2001) 486 195. P. Murtz, L. Menzel, W. Bloch, A. Hess, O. Michel, W. Urban: LMR spectroscopy: a new sensitive method for on-line recording of nitric oxide in breath, J. Appl. Physiol. 86, 1075–1080 (1999) 486 196. H. Dahnke, D. Kleine, P. Hering, M. Murtz: Real-time monitoring of ethane in human breath using mid-infrared cavity leak-out spectroscopy, Appl. Phys. B 72, 971–975 (2001) 486 197. D. E. Cooper, R. U. Martinelli, C. B. Carlisle, H. Riris, D. B. Bour, R. J. Menna: Measurements of 12 CO2 :13 CO2 ratios for medical diagnostics with 1.6-micron distributed-feedback semiconductor diode lasers, Appl. Opt. 32, 6727–6731 (1993) 486 198. L. R. Narasimhan, W. Goodman, C. K. N. Patel: Correlation of breath ammonia with blood urea nitrogen and creatinine during hemodialysis, Proc. National Acad. Sci. United States Am. 98, 4617–4621 (2001) 486 199. C. Roller, K. Namjou, J. D. Jeffers, M. Camp. A. Mock, P. J. McCann, J. Grego: Nitric oxide breath testing by tunable-diode laser absorption spectroscopy, Application in monitoring respiratory inflammation, Appl. Opt. 41, 6018–6029 (2002) 486 200. I. Tsujino, M. Nishimura, A. Kamachi, H. Makita, M. Munakata, K. Miyamoto, Y. Kawakami: Exhaled nitric oxide – Is it really a good marker of airway inflammation in bronchial asthma?, Respiration 67, 645–651 (2000) 486 201. T. Ingi, J. Cheng, G. V. Ronnett: Carbon monoxide – an endogenous modulator of the nitric oxide-cyclic GMP signaling system, Neuron 16, 835–842 (1996) 489 202. Y. Morimoto, W. Durante, D. G. Lancaster, J. Klattenhoff, F. K. Tittel: Realtime measurements of endogenous CO production from vascular cells using an ultrasensitive laser sensor, Am. J. Physiol. 280, H483–H488 (2001) 489, 490

510


203. A. A. Kosterev, F. K. Tittel, W. Durante, M. Allen, R. Köhler, C. Gmachl, F. Capasso, D. L. Sivco, A. Y. Cho: Detection of biogenic CO production above vascular cell cultures using a near-room temperature QC-DFB laser, Appl. Phys. B 74, 95–99 (2002) 489 204. A. A. Kosterev, F. K. Tittel, R. K¨ ohler, C. Gmachl, F. Capasso, D. L. Sivco, A. Y. Cho, S. Wehe, M. G. Allen: Thermoelectrically cooled quantum-cascadelaser-based sensor for the continuous monitoring of ambient atmospheric carbon monoxide, Appl. Opt. 41, 1169–1173 (2002) 489 205. A. A. Kosterev, R. F. Curl, F. K. Tittel, R. K¨ ohler, C. Gmachl, F. Capasso, D. L. Sivco, A. Y. Cho: Transported automated ammonia sensor based on a pulsed thermoelectrically cooled QC-DFB laser, Appl. Opt. 41, 573–578 (2002) 490 206. A. A.Kosterev, F. K. Tittel: Chemical sensorsbased on quantum cascade lasers, IEEE Quant. Electron. 38, 582–591 (2002) 490 207. D. Richter: Portable mid-infrared gas sensors: development and applications, Dissertation, Rice University, Houston (2001) 491 208. D. Richter, A. Fried, B. P. Wert, J. G. Walega, F. K. Tittel: Development of a tunable mid-IR difference-frequency laser source for highly sensitive airborne trace gas detection, Appl. Phys. B 75, 281–288 (2002) 491, 493 209. D. Richter, M. Erdelyi, R. F. Curl, F. K. Tittel, C. Oppenheimer, H. J. Duffell, M. Burton: Field measurements of volcanic gases using tunable diode laser based mid-infrared and Fourier transform infrared spectrometers, Opt. Lasers Engin. 37, 171–186 (2002) 492 210. D. Richter, F. K. Tittel, J. A. Bahr, D. T. Wickham, J. D. Wright, J. C. Graf: Diode laser based formaldehyde measurements in a catalytic trace contaminant control system, paper 2302, Proc. 30th Int. Conf. Environmental Systems (ICES), Toulouse, France (2000) 492 211. D. Rehle, D. Leleux, M. Erdelyi, F. Tittel, M. Fraser, S. Friedfeld: Ambient formaldehyde detection with a laser spectrometer based on differencefrequency generation in PPLN, Appl. Phys. B 72, 947–952 (2001) 492 212. D. Richter, A. Fried, G. S. Tyndall, E. Oteiza, M. Erdelyi, F. K. Tittel: Tunable telecom diode laser based mid-IR source at 2.64 microns, OSA Trends Optics Photonics, Advanced Solid State Lasers 68, 351–353 (2002) 494 213. P. Bergamaschi, M. Schupp, G. W. Harris: High-precision direct measurements of 13 CH4 /12 CH4 and 12 CH3 D/12 CH4 ratios in atmospheric methane sources by means of a long-path tunable diode laser absorption spectrometer, Appl. Opt. 33, 7704–7716 (1994) 495, 496 214. M. Erdelyi, D. Richter, F. K. Tittel: Measurement of 13 CO2 / 12 CO2 isotopic ratio using a difference-frequency-based sensor operating at 4.35 µm, Appl. Phys. B 75, 289–295 (2002) 495 215. J. B. McManus, M. S. Zahniser, D. D. Nelson, L. R. Williams, C. E. Kolb: Infrared laser spectrometer with balanced absorption for measurement of isotopic ratios of carbon isotopes, Spectrochim. Acta A 58, 2465–2479 (2002) 496 216. E. R. T. Kerstel, R. v. Trigt, N. Dam, J. Reuss, H. A. J. Meijer: Simultaneous determination of the H-2/H-1, O-17/O-16, and O-18/O-16 isotope abundance ratios in water by means of laser spectrometry, Anal. Chem. 71, 5297–5303 (1999) 496


511

217. K. Uehara, K. Yamamoto, T. Kikugawa, N. Yoshida: Isotope analysis of environmental substances by a new laser-spectroscopic method utilizing different pathlengths, Sensors Actuators B 74, 173–178 (2001) 496 218. E. R. Crosson, K. N. Ricci, B. A. Richman, F. C. Chilese, T. G. Owano, R. A. Provencal, M. W. Todd, J. Glasser, Alex A. Kachanov, B. A. Paldus, T. G. Spence, R. N. Zare: Stable isotope ratios using cavity ring-down spectroscopy: determination of 13 C / 12 C for carbon dioxide in human breath, Anal. Chem. 74, 2003–2007 (2002) 496 219. L. Hvozdara, N. Pennington, M. Kraft, M. Karlowatz, B. Mizaikoff: Quantum cascade lasers for mid-infrared spectroscopy, Vibr. Spectrosc. 30, 53–58 (2002) 496 220. D. Richter, A. Fried, F. K. Tittel: Special Issue: Trends in laser sources, spectroscopic techniques and their applications to trace gas detection, Appl. Phys. B 75, 143–403 (2002) 496 221. R. F. Curl, F. K. Tittel: Tunable Infrared Laser Spectroscopy, Ann. Rep. Progr. Chem. C 98, 217–270 (2002) 496

Index

H2 O2 , 481 12 CO2 , 494 13 CO2 , 494 in situ measurement, 467 absorption – combination, 451 – line, 450, 451, 462, 463, 487, 488, 490, 494, 495 – linewidth, 448 – overtone, 451 – spectroscopy, 462, 463, 471, 473, 474, 477, 483, 489, 494 – spectrum, 471, 488 acousto-optic modulator (AOM), 476, 487 air-broadening coefficient, 466 airborne tunable diode laser system, 486 ambient air, 453, 473, 484, 490 ambient temperature, 492, 496 ammonia – concentration, 480 – mixture, 476 – sensor, 490 amplitude modulation, 445, 470 amplitude-modulated phasemodulation spectroscopy (AMPM), 470 analog–digital converter, 488 angle tuning, 457 antimonide diode laser, 450, 451 astigmatic beam profile, 451, 453, 458 astigmatic mirror Herriott cell, 492 astigmatic mirror multi-pass cell, 472 balanced beam, 467

balanced ratiometric detection (BRD), 467, 469 bandwidth, 459, 460, 468, 480, 491, 492 base-emitter curve, 469 beam bounce pattern, 472 beam divergence, 445, 450, 451 beam pointing stability, 458, 461, 473 beam profile, 451, 453, 455, 458, 461, 462 beamsplitter/combiner (BS), 455 Beer–Lambert absorption law, 463, 464, 471, 472, 478 birefringent nonlinear crystal, 457 Boltzman distribution, 495 Bragg grating, 452 Bragg reflector, 449, 493 Brewster angle window, 495 broadly tunable monolithic integrated multi-section diode laser chips, 456 C2 H2 , 477, 481 C2 H6 , 481 cavity eigenmode, 476 cavity finesse, 476 cavity ring-down spectroscopy (CRDS), 475, 476 cavity transmission spectrum, 486 cavity-enhanced absorption spectroscopy (CEAS), 476, 477, 486 CH2 O, 450, 465, 480, 481, 483–486, 492–494 CH4 , 451, 477, 480 Chernin multi-pass cell, 472 chromium-doped zinc selenide chalcogenide, 454 CO (carbon monoxide), 447, 451, 477, 479–481, 489, 490 CO2 , 447, 477, 479–481, 488, 495

Index CO2 laser, 496 collisional de-excitation, 478 combination-overtone ro-vibrational line, 462 combined broadening, 464 conduction band, 447 continuous wave, 448, 450, 462 continuous wavelength tunability, 452 cryogenic cooling, 447, 449, 451, 453 CS2 , 481 CW dye laser, 470 dcc-coupled Peltier-cooled HgCdTe (MCT), 492 dielectric plasmon mode, 452 difference frequency generation (DFG), 455, 457, 461, 490, 492 – beam profile, 491 differential optical absorption spectroscopy (DOAS), 466 direct diode-pumping, 455 distributed Bragg reflector (DBR), 449, 456 distributed feedback diode laser (DFB), 458 divergent beam profile, 451, 453 dopant gain region, 457 doped solid-state bulk material, 446 Doppler cross-section, 464 double refraction, 457 dual chamber absorption cell, 495 dual channel absorption laser, 482 dual-beam detection, 468 dual-beam long-path absorption spectroscopy, 492 ellipsoidal off-axis mirror (OAE), 454 elliptical beam profile, 451, 453, 458 equivalent noise bandwidth (ENBW), 468 Er/Yb fiber amplifier, 494 erbium fiber amplifier, 455 external cavity diode laser (ECDL), 456, 458, 496 Fabry–Pérot laser, 448, 451, 452, 456, 477 ferroelectric crystal, 459, 461 fiber amplifier, 479, 491, 494

513

fiber based pump source, 490 fiber optic amplifier, 491 FM spectrometer, 471 FM spectroscopy, 471 formaldehyde, 450 free spectral range (FSR), 475, 477 frequency locking technique, 457 frequency modulation spectroscopy (FMS), 470, 471 FWHM, 448 GaInSbAs quantum well, 451 GaInSbAs/GaSb, 451 Gaussian, 464–466 Gaussian beam quality, 458 Gaussian fit, 491 GEISA databank, 466 guided-wave parametric process, 460 heme oxygenase, 489 Herriott multi-pass cell, 472, 477, 482, 484 high inherent mid-IR wavelength stability, 492 high sensitivity, 490 high-power continuous-wave DFG source, 493 HITRAN database, 463, 466, 472, 495 homonuclear diatomic molecule, 463 HWHM, 452, 453, 465 hydride vapor phase epitaxy (HVPE), 459 hydrogen fluoride (HF), 494 hydrogen-bonded cluster, 475 idler, 457, 460, 461 indium-antimonide photovoltaic detector, 482 integrated cavity output spectroscopy (ICOS), 476, 477, 486 intersubband transition, 451 Jamin interferometer, 479 laser – beam, 457, 473, 478–480, 487, 489 – CO2 , 496 – ECDL, 456, 458, 496 – Fabry–Pérot, 448, 451, 452, 456, 477

514

Index

– quantum cascade (QCL), 447, 451, 453, 475, 481, 489 – radiation filtering, 476 – source, 445–447, 453, 455–457, 459, 475, 476, 478–480, 491, 492, 494, 496 lasing threshold, 487 lead-salt diode laser, 447, 449, 450, 481, 492, 496 Libbrecht design, 452 light detection and ranging (LIDAR), 466 LiNbO3 , 460, 492 – periodically poled (PPLN), 455, 459, 461, 462 linear high-finesse optical cavity, 488 lineshape, 464, 465 liquid nitrogen – dewar, 448, 450, 481 – optical cryostat, 487 LIRF, 468 Lissajous pattern, 482 LiTaO3 , 459, 460 lock-in amplifier, 468, 470, 471, 483 logarithmic conformance, 469 long optical path length spectroscopy, 472 Lorentzian, 466 low-loss single mode fiber, 496 MEMS, 456 methanol, 480 mid-IR detector, 469 mid-IR spectroscopic application, 480 minimal detectable CH2 O concentration, 493 minimum detectable concentration, 467 minimum detectable fractional absorption, 468 minimum relative detectable concentration, 468 molecular beam, 475 molecular beam epitaxy (MBE), 450, 459 molecular collision, 464, 465 molecular rotational-vibrational (ro-vibrational) state, 462 molecule, 462, 463, 465, 468, 471, 472, 478, 481, 486 monochromatic emission spectrum, 470

monolithic non-planar ring oscillator (NPRO), 457 MOSFET, 488 multi-pass cell, 472, 473, 477, 486, 489, 492, 496 multiple harmonics, 470 multiple linear regression approach, 484 multiple quantum-well heterostructure, 451 multiple spatial mode, 460 N2 , 463, 476, 488 narrow-linewidth diode laser source, 491 near-Gaussian intensity distribution, 492 NH3 , 451, 477, 479, 481, 490 NO, 476, 481, 486–488 NO-cGMP, 489 noise reduction, 466, 467, 469–471, 477, 491 noise-immune cavity-enhanced optical heterodyne spectroscopy (NICE-OHMS), 477, 478 non-planar ring oscillator (NPRO), 457 non-resonant probe laser beam, 479 off-axis cavity alignment, 477 off-axis mirror, 482 optical absorption spectrum, 462 optical fiber, 446 optical isolator, 493 optical parametric oscillation (OPO), 462 optical parametric oscillator (OPO), 455, 457, 462, 476, 479 optical pathlength, 467, 472, 476, 482, 489, 490 overtone ro-vibrational line, 462 p–n junction, 447 parabolic off-axis mirror (OAP), 454 parametric conversion source, 475 parametric frequency conversion, 446, 447, 455, 456 Pb-salt diode laser source, 448–450, 481 PbEuSeTe, 449 PbSnTe, 449 Pd/Al2 O3 , 484

Index Peltier cooling, 453, 480 PFBHA, 493 phase matching – condition, 457 photoacoustic spectroscopy (PAS), 467, 478–480 photothermal absorption spectroscopy, 479 photothermal spectroscopy, 478 platelet aggregation, 489 polarization – rotator, 468 polyatomic molecule, 463 portable gas sensor, 490 potassium titanyl phosphate (KTiOPO4 , or KTP), 459 ppbv, 463, 476, 479, 480, 484, 490, 492 PPLN, 455, 459, 461, 462, 491 ppmv, 463, 467, 480, 490 pptv, 450, 463, 467, 472, 486, 493 probe beam, 468, 472, 479 QPM-PPLN, 490, 492, 494 quantum cascade laser (QCL), 447, 451, 453, 475, 481, 489 quantum well, 449, 451, 452 quasi-phase-matched material (QPM), 447 quasi-phase-matching (QPM), 459, 461 – property, 459 radioisotope counting technique, 489 rapid background subtraction, 483 rapid sweep integration, 471 reference detector, 469 reflecting objective (RO), 454 refractive index – modulation, 452 resonant excitation beam, 479 ring-down decay, 476 ro-vibrational line, 462 sampled grating distributed bragg reflector diode laser (SG-DBR-DL), 458 selectivity, 445, 465, 467, 471, 479, 480, 495 sensitivity, 445, 467–469, 471, 476, 477, 479, 480, 486, 490, 492, 496

515

short-path absorption cell, 469 signal enhancement, 466, 467, 472–474, 477–480 single beam interferometer, 471 single-pass difference-frequency generation, 490 spectroscopic gas sensor, 476, 486 spectroscopy – cavity ring-down (CRDS), 475, 476 – FM, 471 – photoacoustic (PAS), 467, 478–480 stoichiometry, 448 sub-ppb detection sensitivity, 494 surface-plasmon mode, 452 TEM00 , 455, 458, 460, 462, 488 thermal motion, 464, 465 toroidal off-axis mirror (OAT), 454 trace gas detection, 445, 447, 449, 462, 467, 478, 479, 490, 496 triggering circuit (TC), 488 tunable optical parametric oscillator, 461 tunable solid-state laser characteristics, 455 tunable solid-state laser source, 447, 486 tuning – étalon, 455 two-tone frequency-modulation (TTFM), 470 ultra-high reflective spherical mirror, 473 unipolar semiconductor injection laser, 451 variable-ratio beamsplitter, 468 vascular smooth muscle cells (VSMCs), 489 vertical cavity surface emitting laser (VCSEL), 456, 494 vibronically broadened transition, 454 Voigt profile, 465, 466 waveguide phase matching, 460 wavelength division multiplexer (WDM), 493 wavelength modulation, 481

516

Index

– spectroscopy (WMS), 470, 471, 493 wavelength tunability, 445, 456, 462 wavenumber time domain, 475 White multi-pass cell, 472

YAG material, 457 zero-background subtraction detection, 467

Mid-IR Laser Applications in Medicine Benedikt Jean and Thomas Bende Experimental Ophthalmic Surgery, University Eye Hospital T¨ ubingen Derendingerstr. 41, 72072 T¨ ubingen, Germany [email protected] Abstract. This chapter reviews medical applications of a variety of mid-infrared lasers. These applications are based on strong absorption of laser light in human tissue due to the presence of naturally occurring chromophores, specific and unspecific absorbers. Medically relevant laser-tissue interactions are described. Experimental data, obtained with free electron lasers describe photoablation quantitatively in the mid-IR as well as collateral adverse effects. Feedback technologies for online therapy control are presented; they enhance the selectivity of the laser–tissue interaction. Typical medical and surgical applications in gynecology, otorhinolaryngology, neurosurgery, dermatology, urology, dental surgery, ophthalmology and cardiovascular surgery are briefly summarized.

1

Introduction

Attempts to use lasers in medicine are so numerous and, historically, so diverse, that it proves difficult to describe them all in this chapter. Many laser ideas were introduced experimentally, evaluated clinically and later abandoned, giving way to other concepts in accordance with the availability of laser sources or delivery technologies. This process continues today. The present collection of laser applications thus describes the laser innovations at a given point in time only. A fundamental trend stands out with regard to laser sources: laser equipment is reduced in complexity, maintenance requirements, and costs, thus giving way to more widespread use. This may lead to new emerging fields – such as laser welding or tissue molding – exemplifying this trend. IR solid-state lasers and their future development are predicted to play a central role as the prevailing laser sources of the future, possibly followed by semiconductor lasers. Also notable is another demand-driven trend towards more selective laser applications: photosensitizer techniques (photodynamic therapy) and laser tunability. The plethora of different laser applications described in this chapter fall into two categories: • applications replacing conventional tools: e.g. cutting, replacing the scalpel with the advantage of collateral hemostasis, dry surgical fields, reduced edema due to less tissue manipulation; and • applications with unique capabilities unavailable in other technologies: such as no-touch tissue removal, layer selective area ablation (dermatology and refractive surgery), submicron accuracy (refractive and corneal reshaping), non-contact membranes cutting in the eye (capsulotomy, photodisruption). I. T. Sorokina, K. L. Vodopyanov (Eds.): Solid-State Mid-Infrared Laser Sources, Topics Appl. Phys. 89, 511–546 (2003) c Springer-Verlag Berlin Heidelberg 2003

512

Benedikt Jean and Thomas Bende

The surgical requirements of “selectivity,” “efficacy,” and “safety” form a magic triangle, imposing a rigid rule for allowable compromises in various medical and surgical applications. Important improvements are imaginable for the future should tunable IR lasers become available, leading to increased selectivity and thus improved safety. With water as an omnipresent and prevailing absorber, however, other technologies, like on-line control and feed back elements, may become indispensable.

1.1

Absorbers and Naturally Occurring Chromophores

Efficacy and selectivity of the laser–tissue interaction is related to the presence of target specific absorbers, typically naturally occurring chromophores. The absorbencies of oxyhemoglobin, desoxyhemoglobin, melanin and xantophyll all reside in the visible, thus off the emission range of mid-IR lasers (Fig. 1). The following graphs show the three most important target materials for mid-IR laser surgery.

Fig. 1. Relative absorption as a function of wavelength for naturally occurring chromophores

Mid-IR Laser Applications in Medicine

513

Figure 2 shows the absorbance of water, measured in corneal tissue, typical also for brain tissue, cartilage, and disc material, with comparably high water content. Typical absorbers for hard substances are enamel and dentin, the typical absorbers for dental surgery. Their absorption spectrum is shown in Fig. 3. A complex pattern of absorbers characterizes artherosclerotic plaques of different composition; their rather complex pattern of absorption is shown

Fig. 2. Absorption coefficient as a function of wavelength between 2.5 and 7 µm for human corneal stroma

Fig. 3. Relative absorption of dentin and enamel in teeth between 2.5 and 20 µm

514


Fig. 4. Relative absorption of apatit and cholesterol compared to water as a function of wavelength between 5.5 and 12 µm, measured in human artherosclerotic plaque

in Fig. 4. Various UV and IR lasers have been used in minimally invasive cardiac surgery and angioplasty to match the above absorbencies with regard to maximal selectivity and safety. Typical laser sources, their uses, and the underlying laser–tissue interaction are described in the following sections. 1.2

Laser Sources

Since the commercial introduction of solid-state mid-IR lasers, with typical pulse duration of 50–350 µs at wavelengths between 2 and 3 µm, numerous practical applications in surgery have become available. Due to the high absorption of water, the laser emission wavelength shows relatively shallow penetration and small collateral thermal damage zones. In many respects these lasers have replaced excimer lasers, due to lower investment and maintenance costs, smaller and more compact designs, and fiber optic delivery options. The preferential switch from excimer lasers to solid-state mid-IR lasers stems also from the absence of mutagenic side effects [1], inherent to the primary radiation of excimers (248, 308 nm) as well as to the secondary radiation, triggered by 193 nm excimers. 1.2.1

Erbium YAG Laser

Depending upon the crystalline host material, erbium lasers can emit radiation between 2.64 and 2.79 µm (Er:YSGG) laser, to 2.90–2.94 µm (Er:YAG). These comparably small differences in the emission wavelengths play a major role in future preferred laser delivery technologies. The CTE:YAG (chrome, thulium erbium), 2.69 µm [2], the THC:YAG (thulium holmium) 2.15 µm and


515

the HF laser (3 µm) have also been suggested and investigated experimentally [3,4]. The Er:YAG in the pulsed free running mode has a typical pulse characteristic, influencing its medical use. Pulses in the spiking mode are shown in Fig. 5. The spikes create pressure waves and also damage the fiber tips. On the other hand, they can enhance the ablation of a hard dental substance. Recent technical developments (Fig. 5a) led to a substantial reduction of the number and amplitude of these spikes, allowing for easier fiber delivery, more homogenous energy distribution, and better applicability for area ablation as well as reduced pressure wave generation. High power Er:YAG lasers used in medicine are predominantly operated in the pulsed mode. Its 2.94 µm wavelength shows a higer absorption in water than the CO2 laser. Pulse energies vary between 1 and 2 J, allowing for large area ablation. Current technologies allow a maximum power output of more than 5 J and pulse repetition rates of up to 30 Hz. Corneal reshaping has long been considered a particular challenge for IR photoablation, principally due to its complex requirements: optimized surface quality, minimized thermal damage, small ablation rate for high accuracy corrections, etc. Certain experimental applications, including wound healing studies [6], have been performed successfully with a free running Er:YAG [7,8] as well as Q-switched system [9].

Fig. 5. Dynamical behavior of 3 µm Erbium lasers: (a) Er:YAG, optimized spiking mode; (b) Er:YAG normal spiking mode

516


1.2.2

Holmium YAG Laser

The Ho:YAG laser has an emission wavelength of 2.1 µm and utilizes a water absorption peak of about 2.0 µm. Compared to the 1064 nm absorption of the Nd:YAG laser, the Ho:YAG laser’s water absorption is higher, however, it is substantially lower than 2.94 µm for the Er:YAG laser. Typical emission energies lie in the order of 1–2 J/pulse. Fiber optic delivery systems are available for this wavelength and allow for a maximum flexibility and versatility for different clinical applications, including arthroscopy procedures, neurosurgery (discectomy), and prostatectomy, among many others.

2 2.1

Laser–Tissue Interaction Photothermal Laser–Tissue Interaction

The temperature achieved during a procedure can characterize the thermal effects that occur during laser irradiation of tissue. Depending on varying optical and thermal properties, different combinations of energy and exposure time can be used to obtain the same thermal effect. In the visible and mid-IR spectral range, laser light is absorbed by the chromophores; the IR radiation triggers molecular vibrations and rotations. The atomic or molecular absorption and subsequent relaxation of the excited particles transforms the optical energy into heat energy. Due to the heat capacity of the material, the heat energy from the laser then determines the resultant tissue temperature. With increasing depth, the radiation absorbed by the tissue decreases according to the Lambert–Beer law (I = I0 exp(−ad): I = intensity, I0 = intensity on the surface, a = absorption coefficient of the material, d = depth). This indicates that the heat energy, and thus the temperature, decreases with increasing depth. As part of the absorbed energy scatters off the path of the incident radiation, thermal conductivity and convection in the blood stream dissipate heat so that two gradients are formed: one for increasing depth and one for increasing lateral distance. The resultant tissue temperature, achieved with a defined laser irradiation regime, is determined by the optical and thermal properties of the tissue. The thermal tissue effects are listed in Table 1. The achieved thermal effect depends upon the tissue temperature and the time of exposure. Any defined degree of thermal tissue alteration, including irreversible tissue damage, e.g. required for tumor therapy, thermal inactivation of tissue (cyclophotocoagulation) etc. can be achieved by different combinations of energy and exposure times (Fig. 6). A typical complex clinical application in which compromise strategies must be applied is the controlled heating of corneal tissue for Ho:YAG thermo keratoplasty, laser thermo keratoplasty (LTK) [10,11]. This is a refractive procedure of the central cornea to make it steeper; peripheral coagulation spots in deep corneal stroma create tension lines that make the central corneal radius


517

Table 1. Thermal, mechanical and optical effects as a function of temperature Effects on tissue Temp [ ◦ C]

Thermal

Optical

Mechanical

< 37 40–45

Reversible damage, only enzyme induction edema, membrane alteration (cell death)

– –

– –

60–65

Protein denaturing, coagulation, necrosis

Whitening, light scattering

Incipiant weakening

70–85

Collagen denaturing, membrane damage, incipient necrosis

Opacification

Waterless tissue desiccation

90–100

Desiccation

> 150

Carbonization

Blackening increased absorption

Strong mechanical damage

300 >

Vaporization

fumes

Ablation

Shrinkage, drying

Fig. 6. Irreversible tissue coagulation as a function of temperature and exposure time

518


steeper. In order to achieve a defined temperature increase in corneal tissue of about 30 K, a combination of pulse repetition rate (e.g. 25 Hz) and the energy per pulse must be applied: the energy/pulse must be limited to 25 mJ, in order to avoid local ablation (ablation threshold: 30mJ) with a spot measuring 800 µm. In this manner, the subthreshold energies of each pulse avoid unwanted local corneal ablation in the contact application mode (800 µm fiber tip diameter). A saw tooth shaped temperature increase pattern with cooling intervals between the pulses leads to the therapeutically needed temperature plateau; it serves as a balance between energy uptake, accumulated thermal load, and dissipation. 2.1.1

Pulsed Versus CW Irradiation

Using subthreshold fluencies (less than 1 J/cm2 ), mid-IR lasers trigger thermal effects in tissue only. Tissue heating, desiccation, and shrinkage, however, can be better controlled if CW irradiation is used instead of pulsed radiation. For the LTK procedure described above in refractive surgery, optimally controlled tissue heating is essential: the temperature increase must reside in a corridor between 65 ◦ C and 80 ◦ C. These are the temperatures at which collagen uncoils but collagen denaturation, leading to mechanical weakening, is not yet observed (Fig. 7). Brinkmann and co-workers [12] showed that this delicate balance can be better maintained using CW irradiation of the cornea. Pulsed radiation, typically performed with a Ho:YAG laser, triggers fast on–off effects and intermittent cooling requires peak temperatures well above 150 ◦ C. These temperature spikes cannot be neglected; they trigger a relaxation-like effect, clearly shown by mechanical spectroscopy measurements as a function of corneal tissue temperature [13]. At higher temperatures, the incipient denaturing of collagen leads to the aforementioned mechanical weakening, counteracting the intended corneal steepening effect. Several groups have studied this thermo-mechanical behavior [14]; it is a good example of the high demands for laser sources in order to comply with the specific biophysical behavior of tissues. For refractive surgery procedures, the clinical relevance is obvious. Thermo-mechanical reactions of laser-irradiated tissues (cornea, skin, disk, cartilage) may not be the only or even the most prevalent reaction to tissue heating. Meticulous control of laser heating may thus become more and more important. A novel approach to monitor tissue heating is dielectric spectroscopy [15]. It describes the dipole density as a function of tissue temperature; maximal dipole density per volume means maximal tissue “shrinkage”, a criterion describing coagulation of tissue. Applied on human corneal tissue in vitro (Fig. 8), the tissue’s reactions as a function of temperature is measured non-invasively; these measurements [16] confirm the results of mechanical spectroscopy [13] – dipole density, relaxation time. Pulsed and CW irradiation create different thermal gradients. The difference is shown in Fig. 9, using Monte Carlo calculations. The temperature


519

Fig. 7. Mechanical spectroscopy: Shear compliance as a function of tissue temperature of human corneal tissue in vitro [13]

Fig. 8. Dielectric spectrosopy of human corneal tissue in vitro; dipole density and relaxation time as a function of tissue temperature [15,16]

distribution and the geometry of the gradients are substantially different (Fig. 9a,b). The triangular shaped coagulation cone is the result of absorption and focusing (Fig. 10). The introduction of diode laser thermal keratoplasty (DTK) [10,11], replacing pulsed Ho:YAG LTK [17], is an example of optimized laser tissue heating. Using CW irradiation, a better controlled temperature increase has become clinically available, avoiding counterproductive temperature peaks, reaching up to 160 ◦ C [11] and triggering a relaxation-like effect. Better clinical results seem to be achievable. The same mechanism should be principally applicable for optimized and lasting collagen contraction in asthetic

520


Fig. 9. Monte Carlo calculation, temperature distribution pattern for (a) CW, and (b) pulsed irradiation

Fig. 10. Corneal cross-section (6.0 µm FEL): birefringence of unaltered tissue; H&E eosiniphilic zone underlying the superficial carbonization Porcine cornea


521

surgery: the delicate balance between collagen shrinkage and collagen denaturing likewise plays an important role for skin resurfacing and removal of deep rhytids/wrinkles. Unintentional overheating of the collagen layer may only lead to a temporary collagen contraction, later overcompensated by relaxation, due to denatured and thus weakened collagen fibers. The modeling and calculation of temperature distributions and gradients as a function of the incident radiation can be particularly complex if the laser energy is coupled into a multi-layer tissue. Cyclophotocoagulation, for the controlled thermal destruction of the ciliary body in the anterior segment of the human eye via a transscleral route, is one such example. Selective targeting and the destruction of a distinct chromophore bearing layer (melanin) is achievable if energy and wavelength are combined with a defined pulse width. Pustovalov and Jean [18] developed thermal models in order to predict and quantify the extent and localization of thermal lesions. In this manner, fairly precise 3-D positioning of a thermal spot becomes possible for therapeutic purposes. 2.1.2

Quasi-CW Irradiation

Although optimized laser tissue heating (e.g. interstitial thermotherapy) requires a CW laser source, pulsed lasers have been used at high repetition rates. They generate a quasi-CW thermal irradiation that also leads to thermal pooling; individual laser pulses are confluent to a CW like irradiation pattern, not unlike CW irradiation. The pulse repetition rate has to be higher than the low specific heat diffusion of water (in tissues with high water content). At those pulse repetition rates, individual laser pulses are confluent to a CW-like irradiation pattern. 2.2

IR Photothermal Ablation

UV and IR photoablation are based upon different laser–tissue interactions that substantially affect the clinical applicability. UV photoablation is based upon photodissociation. It is a direct bond breaking interaction using the excimer laser’s high photon energy of 6.43 eV to dissociate C=C bonds (4.34 eV) or single C–C bonds at even lower energies (2.54 eV) [19]. In the IR photoablation, however, the basic mechanism is “photothermal ablation”, for either soft tissues (skin, cornea, etc.) or hard tissues (teeth, bone). IR lasers provide photon energies at around 1 eV and lower; this lower energy level does not trigger bond breaking but stimulates molecular vibrational and rotation modes, resulting in rapid tissue temperature increase. Overheated intercellular and intracellualar water forms steam vacuoles. The relief of pressure leads to a fast explosion-like event and subsequent tissue decomposition and vaporization. The incident radiation is absorbed in various steps depending upon the local tissue absorbencies. Temperature levels of 180 ◦ C and higher lead to tissue fragmentation, decomposition and particle ejection. The temperature

522


and pressure increase during irradiation until the critical vaporization temperature of the absorber is reached. The process has therefore been named “photothermal ablation”. The ablation can be understood as a continuous process, commencing when the energy density is high enough to vaporize the water in the tissue, and continuing until the end of the laser pulse [20,21]. For large area ablation with high surface quality, photothermal ablation may create substantial problems if there is different local hydration. Minimal heterogeneities of the laser beam profile create spots of local desiccation, reduced heat dissipation, and consecutive local coagulation and carbonization. A subsequent pulse then “sees” a different surface and enhances the heterogeneous ablation. Early attempts to overcome these effects using “wet” ablation techniques with a water film did not convince clinically [22,23]. The influence of the absorption coefficient of the target material on the ablation efficiency has been studied, using a tunable free electron laser (FEL) . In the case of a rectangular beam profile, the ablation efficiency can be expressed by the rate or depth per pulse, if the ablation diameter is known. For primarily Gaussian-shaped beam profiles, or more complex patterns of interaction, between the incident radiation and tissue, the ablation volume per pulse is better described by the term “ablation efficiency”. It is a function of the fluency applied. A linear slope of this function depends on the absorption coefficient of the tissue and the emission wavelength of the laser. Different tissue absorbencies or different emission wavelengths in the same material likewise show different slopes [24,26]. Another indicator of ablation efficiency – in cases where the energy output of laser systems is limited (such as FELs) – is the determination of the ablation threshold. Unlike UV photoablation, where the ablation rate as a function of fluency ends in a plateau, IR photoablation efficiency continues to increase with increasing fluency. The influence of pulse length on the ablation efficiency has been of lesser interest in a clinical context, because increasing the fluency applied can compensate for the lower ablation rate of shorter pulses. However, even when increasing the pulse length by a factor of 1000, Bende and coworkers [27] found only minimal reaction to the ablation rate, almost irrelevant in a clinical context. 2.3

Photospallation

For very short pulse lengths a different ablation process, called photospallation, has been described [28,29,30]. While Q-switched Er:YAG lasers achieve 70 ns of pulse length and free running Er:YAGs at around 70 µs, Tellfair et al. have achieved 7 ns using an OPO at 2.94 µm at fluencies between 100 and 500 mJ/ cm2 . Earlier attempts with short pulses in the mid-IR have been made by Seiler et al. [3], using a HF laser at 50 ns pulse length. Stern et al. [31] used a Raman shifted Nd:YAG at 2.92 µm as early as in 1988, achieving short pulses of 8 ns only. Tellfair et al. [28] were able to prove that an


523

OPO at 2.94 µm created the smallest collateral thermal damage of only 0.1– 0.2 µm, a minimal debris layer of 0–0.4 µm and small reactions (0.3–2.0 µm) of eosinophilic staining in the H&E micrometry. Short laser pulses with no relevant thermal diffusion and super-threshold fluencies seem to trigger a photothermal vaporization effect. With water as the primary absorber, the short pulses lead to minimized thermal collateral effects, highly efficient ablation (or vaporization), and almost athermal cutting edges. According to histology and histomicrometry, photospallation in the IR comes closest to the primarily athermal photoablation of the UV laser [28]. Photospallation with ps and fs pulses should be possible and clinically efficient. It also may mark the transition from linear to nonlinear ablation characteristics. The Cr:ZnSe laser is a potential candidate for this exciting perspective (desribed by Sorokina in this volume). 2.4

Collateral Thermal Damage

The medical use of IR lasers for photoablation is characterized by the ablation efficiency and the collateral adverse effects (thermal damage, pressure wave and recoil momentum). The correlation between tissue absorbencies and collateral thermal damage has been investigated systematically, using tunable free electron laser [32,33]. With increasing absorption, the collateral damage decreases, with a minimum at the absorption maximum around 3 µm and 6.1 µm; inversely, it is maximal at 5.5 µm (Fig. 11). The primary thermal lesion shows the typical pattern of a superficial layer of carbonization, followed by a rim of coagulated tissue and an underlying layer of eosinophilic tissue in H&E (hematoxilin-eosin) staining. Experimental investigations with an FEL as a tunable laser source showed that both the irreversible damage zone as well as the H&E positive zone, interpreted as reversibly damaged, react as a function of wavelength. The eosinophilic zone can be understood as an accumulation of low temperature damage with primarily reversible cell injury. For corneal tissue, the cornea’s collagen birefringence indicates the border between unaltered tissue and reversible damage. The exposure time and thus the total thermal load in relation to the tissue’s heat dissipation determines the extent of a tissue’s specific behavior of reversible versus irreversible damage. 2.4.1

Histology and Histo-micrometry

For the interpretation of histology (e.g. Fig. 10) after IR laser irradiation, one has to take into account that a typical ablation area (e.g. an ablation crater) shows substantial differences between the bottom and flanks of the crater: the optically deposited energy at the bottom creates the said pattern of surface carbonization, coagulation and reversible H&E positive staining. Each new pulse in a sequence ablates the carbonized and coagulated tissue

524


Fig. 11. Tissue absorbance and Collateral thermal damage as a function of wavelength. zone 1: H&E staining (reversibles thermal damage) zone 2: surface carbonization

of the former pulses to a certain degree. The collateral thermal damage thus appears smaller at the bottom and maximal at the flanks. Only the flanks accumulate the optically deposited (scattered light) and thermally deposited energy in a sequence of pulses; the flanks thus show the maximum collateral damage. Biophysical models of the understanding and prediction of the photothermal ablation process as a function of wavelengths have recently been presented [34,35].

3 Experimental Photothermal Ablation Using Free Electron Lasers (FEL) Medical and surgical demands are complex and often require compromise strategies. Certain procedures require small ablation rates, for instance for highly accurate reshaping of the cornea for refractive procedures. High ablation rates are required for volume ablation for instance needed in dental surgery. Vascularized tissues need sufficient surface coagulation during cutting in order to provide dry surgical fields and achieve hemostasis. Cutting edges in the avascular and transparent cornea have to be athermal, etc. Proper material targeting for complex or unpredictable local absorbencies (e.g. atherosclerotic plaque, laser angioplasty, Fig. 4) require optimized wavelength selection, but also pressure wave minimizing in order to avoid accidental pressure related damage.


525

Many of these elements have been clinically investigated over time on a trial-and-error basis using the available technology at a given period without being able to custom select the biophysically optimal parameters for a procedure [36]. Therefore, experimental investigations in vitro with a tunable laser source, the FEL, have been performed for a variety of materials [20,37,38]. 3.1

Collateral Thermal Effects as a Function of Wavelength

The first systematic investigation of IR photothermal ablation with FELs has been performed with CLIO in Paris (LURE), with FELIX at Nieuwegein (Institute of Plasma Physics) and initially the Vanderbilt Free Electron Laser [33]. Wavelengths between 1.8 and 70 µm have been used. The correlation between tissue absorbencies and the ablation rate is shown in Fig. 12a. Another criterion is the determination of the ablation threshold, if the fluence is limited (Fig. 12b). The relevance of this model, however, may be limited by the different pulse structure of FELs, compared to conventional laser sources. 3.2

Photomechanical Interaction

Little attention has been paid clinically to photomechanical interactions during photoablation. The effects can be macroscopic (spallation of tissue) or microscopic (cellular disruption) [39]. They can be transient or quasi-steadystate stress. Photomechanical and photoacoustic effects during UV photoablation trigger pressure waves in the order of several hundred bars [40]. IR photothermal ablation likewise triggers a pressure wave. These pressure waves effects can be measured in several ways: the recoil momentum assessed by a pendulum in vitro (Fig. 13), the initial velocity of the ejected particles, and the extent of the ablation plume. Auerhammer has conducted such studies, using FELIX [20]. They all increase with increasing absorption and should increase with increased fluency. According to an advanced ablation model, the pressure wave represents about 1% of the total incident laser energy [34]. In the absence of tunability, the pressure wave has to be “accepted” as an invariable collateral effect. Several clinical applications such as laser angioplasty have been limited in the past by the clinically intolerable pressure generation. Laser tunability thus appears to be an essential element to influence or control the collateral pressure exposure of tissue, for instance by pressure sensing or photoacoustic on line monitoring [41,42]. Tunability, e.g. along the 3 µm water absorption flank, would allow one to fine tune laser tissue interactions, minimizing the collateral effects in a tissue specific manner. 3.3

The Influence of Specific Absorbers in the Target Material

Water is the quantitatively prevailing absorber in soft tissues. However, proteins can act as additional or even prevailing absorbers in the mid-IR. Edwards et al. [43] first described the potential of protein absorption in corneal

526


Fig. 12. (a) Ablation rate and absorption coefficient between 2.5 and 7 µm in porcine corneal tissue; (b) ablation threshold as a function of wavelength between 4 and 20 µm.

tissue; at around the wavelengths of 6.45 µm, the amide II absorption peak, (Fig. 14). Amide II can be targeted quasi-selectively as an absorber, although this peak still represents a superimposition of the water and the protein absorption. Auerhammer et al. [20] investigated the effect with the FELIX free electron laser , tuning it to the 6.45 µm peak. The additional absorption of proteins altered the ablation process with regards to the photomechanical effects, the initial velocities of the ejected particles, the extent of the ablation plume, and the recoil momentum. It also altered the ablation threshold to a lesser extent. The amide II absorption at 6.45 µm however had a spectacular effect on fragmentation of tissue and the resultant size of the ejected particles. They


527

Fig. 13. Absorption coefficient (dotted line) and recoil momentum for water as a function of wavelength between 4 and 20 µm

Fig. 14. Relative absorption; specific absorbers in corneal tissue between 5 and 12 µm

were found to be substantially smaller up to almost one order of magnitude, possibly creating smoother ablation surfaces (which had not been measured). 3.4

Feedback Laser Control

The magic triangle of selectivity, safety and efficiency for laser surgery is certainly based primarily in the wavelength selection wherever differences of the

528


absorbencies are found. Tunability thus optimizes ablation and potentially minimizes the adverse thermal effects as well as the pressure wave exposure which may have to be more considered in the future. However, selectivity must be limited if the predominant absorber is water, forming 85% of our body mass. Feedback control using optical spectroscopy have been suggested. Lately, the exciting potential of Non-Contact Photoacoustic Spectroscopy, NCPAS [44,45] has been explored. This procedure uses the pressure wave, triggered by the photobalation effect, captures the sound wave and analyses it using different technical approaches. In this manner, the ablation coefficient of the ablated tissue can be assessed, as shown in Fig. 15. Furthermore, the NCPAS signal correlates with the fluence applied and the ablation zone diameter. The frequency shift can be used online to discriminate materials, e.g. scar tissue in the cornea can be discriminated from stromal tissue, epithelium and even Bowman’s membrane as shown in Fig. 16. Online analysis can be achieved up to 100 Hz with relatively inexpensive technology [45]. Online therapy control or feedback elements like NCPAS certainly offer an important option to ajust dynamically a tunable laser source in order to optimize the pattern of ablation parameters, eventually minimizing the adverse effects, beyond the presently available technology.

Fig. 15. Photoacustic signals of water and absorption coefficient as a function of wavelength between 4 and 17 µm, measured with FELIX


529

Fig. 16. Cluster analysis of photoacustic signal (NCPAS), characterized by three frequency peaks (f(1)–f(3)): online material discrimination during photoablation in human cornea in vivo [44]

4 4.1

Clinical Laser Applications Gynecology

Gynecology has a large potential volume for the use of lasers in minimally invasive and conventional surgery. The CO2 laser presently dominates this field with its potential for vaporization [46,47,48,49]. With regard to fiber optic delivery systems, solid state IR lasers like the KTP laser (frequency doubled Nd:YAG), the Nd:YAG and the Ho:YAG laser, a more widespread use in the future appears likely. Among the procedures currently carried out with lasers, the removal of superficial leukoplakia or neoplasia rank first: Cervical intraepithelial neoplasia (CIN) [50,51,52], vaginal intraepithelial neoplasia (VAIN), vulva intraepithelial neoplasia (VIN), removal of condylomata accuminata [53] and endometrial ablation for menorrhagia. Only the Ho:YAG and the Er:YAG have so far been used for laprascopic procedures. Laser colposcopy in conjunction with external laser radiation can replace knife conisation or kryoptherapy, offering the advantage of an outpatient or office procedure. Essential for the minimally invasive use of lasers is the combination of tissue removal with surface hemostasis, not easily achievable with CO2 lasers which create surface carbonization but often only unstable surface hemostasis. Nd:YAG and KTP lasers have been proven beneficial for various types of operative hysteroscopy and endometrial ablation [54] and for treatments of the uterine cervix [55]. The use of photosensitizing drugs for photodynamic therapy (PDT) is only at its beginning. It can be anticipated that a tunable IR laser and a wide range of pulse length modifications could favor the future use of lasers in gynecology. The predominantly used CO2 laser [56] could be replaced potentially by a mid-IR CW laser source as the absorption of the

530


10.6 µm wavelength is considerably lower than the one in the 3 µm water absorption band. 4.2

Otorhinolaryngology (ENT)

Microsurgery in this field often uses no-touch techniques and long reach of the laser beam and needs the advantage of the absence of postoperative swelling or stenosis, a dry operative field, hemostasis and also benefits from reduced postoperative pain, due to localized application, minimizing collateral edema. Microlaryngoscopy in conjunction with a laser allows for highly precise removal or excision of tissue in procedures or entities like vocal cord nodules, polyps, keratosis, granulomas, arytenoidectomy, Reinickes edema, webs and laryngeal stenosis. The KTP laser has been used successfully for various procedures [57,58,59,60]. Recurrent respiratory papillomatosis, occurring within the anterior nasal cavities, subglottis and main stem bronchi also favors the use of lasers; presently it is the CO2 laser. Laser removal of visible papilloma and vaporization is ideal for locations of relative inaccessibility, where complete removal, hemostasis or partial tissue removal can be accomplished under constant visual control. Lasers are most frequently used in laser stapedectomy [61], laryngeal subglottis stenosis, tonsillectomy, snoring laser assisted uvula palatoplasty (LAUP). LAUP has become a popular approach for antisnoring treatments [62,63], often performed as an outpatient procedure [64]. Several laser surgical options have been tried in otosclerosis [65]. Among the mid-IR lasers, only the Ho:YAG and the Er:YAG have been used in ENT so far [66]. ENT has many examples where tissue- or layer-selective coagulation has to be achieved. Fibre optic delivery systems are often essential and provide access to bronchi beyond the carina; rigid bronchoscopes are often preferred for the ability to better debride tissue; they offer better succion and manipulation at the scope tip. 4.3

Neurosurgery

For more than 25 years, lasers have been used in this specialty. However many of the more recent technologies have not yet been incorporated into neurosurgery procedures and much of the literature still favors the CO2 laser. With its high water content, brain tissue can be removed with much higher precision with a pulsed Er:YAG laser, compared to a CO2 laser. With their typical ablation rate in the order of several microns per pulse, the Er:YAG laser could potentially replace many of the present CO2 laser uses. Again, the ablation efficiency of the Er:YAG laser is higher than that of the CO2 laser due to the higher absorption in water. Increasing use of minimally invasive procedures in conjunction with fiber optic delivery favors solid state lasers. The Nd:YAG laser is frequently use for shriveling very vascular tumors. The treatment of aneurysm or arterovenous malformation is an example. No-touch techniques


531

are particulartly benificial for spinal or intramedulary tumors, leading to less damage to both cord and nerve roots. Lasers may be used for fenestration of syringomyelia to offer a permanent fluid pathway. Among the mid-IR lasers, only the Ho:YAG has found more widespread use for percutaneous discectomy [67]. Decompression is achieved by vaporization or thermal shrinkage of dislocated and fragmented disc. Transcutaneous volume reduction by thermal shrinkage and tissue removal by vaporization form the basis of these procedures. With regard to the epidemiology and morbidity profile, minimally invasive procedures like percutaneous disectomy could certainly be further improved if the thermal shrinkage with consecutive volume reduction could be optimized, for instance using a CW tunable IR laser. The high water content of the disc would also favor tissue ablation. A tunable source would offer the additional benefit that the ablation efficiency, the collateral thermal effect and the pressure wave exposure (!) could be varied by fine-tuning the laser along the flanks of the 3 µm water absorption peak. 4.4

Dermatology

The argon laser and the CO2 laser were the first systems with more widespread use in dermatology. However, the multitude of lesions, treated in this specialty today require narrow band irradiation or specific chromophore interaction, in particular for the treatment of portwine stenosis, hemoangioma, tatoo removal, superficial varicosities, etc. As a consequence, short pulsed dye lasers at 510 nm are used for epidermal lesions, as well as Qswitched Nd:YAG lasers at 532 nm and 1064 nm. This source has also been used initially for dermabrasion. The Q-switched ruby laser is used for epidermal or dermal lesions. The CO2 laser has been used preferentially so far for plastic and esthetic surgery and skin rejuvenation with surface ablation. Its short pulses and the scanning application mode reduce the local thermal load and thus the potentially serious effect of deep skin burns. The skin resurfacing procedure resulting in “youthful” regeneration of skin [68] has seen wide acceptance; its disadvantage is prolonged recovery. The Er:YAG laser has been shown to remove skin with high efficiency due to its higher specific absorption in water; the limited penetration depth, beneficial layerselective ablation and optimal depth control, reduce substantially the collateral thermal damage, normally resulting in postoperative pain [69,70]. The Er:YAG’s limited penetration depth means that little thermal energy is conducted distally to the site of ablation with a consecutively narrow zone of devitalized tissue, measuring 5 µm. In contrast, the CO2 laser produces between 50 and 100 µm of ablation with up to 50 µm or more of distal thermal conduction [71,72]. More recently, the CO2 and the Er:YAG have been combined for dual wavelength treatments [73], sometimes associated with surface cooling [74,75]. While both lasers, the CO2 and the Er:YAG, lead to local tissue desiccation, the effect is by far more pronounced with the CO2 laser. Repeated pulses thus create a “heat sink” [72], as desiccated tissue leads

532


to a different pattern of locally different absorption, surface desiccation and even carbonization. Furthermore, heated (tissue-) water has been found to change the absorption maximum from around 3 µm to 2.8 µm. This relatively variable laser–tissue interaction, observable when several passes of laser radiation are applied at the same area, has to meet the tissue histology and anatomy. Skin rejuvenation requires an almost layer-selective precision where primarily the epithelial layer and the papillary dermis are removed. Mild and moderate rhytides can be affected without entering the reticular dermis, lying deeper. The undesirable complications such as scarring, depigmentation and protracted erythema is associated with penetration into the reticular dermis; it can be avoided by limiting the depth of ablation to this mid-papillary layer. Er:YAG surface ablation has only a small hemostatic effect, leading to petechial bleeding which may serve as a “depth gauge” during treatment. Deeper thermal heat conduction beyond the mid-papillary dermis, typical for the use of CO2 lasers, may occur. The deeper thermal injuries stimulate an exaggerated repair process, potentially beneficial to treat deeper rhytides. The greater the collagen remodeling effect, effacing deeper wrinkles has to be balanced against the complications of deeper dermal injuries. In essence, two elements are used for plastic-esthetic skin resurfacing, the surface ablation and the deeper heat conductivity with consecutive collagen shrinkage. These two elements, however, have to take into account the very variable skin thickness [76]. The surface ablation part of skin rejuvenation can be accomplished with less adverse effects using the Er:YAG or another pulsed mid-IR laser. However, the collateral heat, deposited in deeper layers, triggers collagen shrinkage, very much the way we know it from laser thermal keratoplasty. However, this is a fairly uncontrolled heating process, potentially leading to total relaxation within a six months period. The clinical effect can be a return of rhytides which had initially disappeared, due to the said temporary effect. With regard to both elements, the surface ablation effect and the deep thermal induced collagen shrinkage, a more depth-selective procedure in conjunction with variable pulse length could lead to an optimization of both elements in order to improve the overall clinical outcome. The appearance of a “heat sink”, deep thermal conduction and heat pooling explains why CO2 lasers often lead to slight ectropion or eye lid retraction, used as an indicator to discontinue local treatment in the sensitive periocular region with its typically thin skin. Clinical experience tells that this eye lid retraction – in our interpretation, caused by heat induced collagen shrinkage – is reversible and not permanent. Taken together, two elements are used for plastic-esthetic skin resurfacing, surface ablation and collagen shrinkage, due to the deeper heat conductivity leading to stimulated collagen repair. These two elements, however, have to take into account the very variable skin thickness and different local ab-


533

sorbencies. From a biophysical point of view, a future tunable IR laser should thus have substantial advantages over the present technology platforms. 4.5

Urology

In this specialty, lasers are presently used for various entities [77,78]; among them are bladder tumors, condylomina accuminata, urethral stricture, penile carcinoma, laprascopic procedures, interstitial cystitis, uretheral calculy, prostatectomy, vasectomy and partial nephrectomy. Several of them are still under investigation, A typical field for mid-IR lasers is laser assisted prostatectomy, presently utilizing Nd:YAG lasers for coagulation, and hemostatsis. The Ho:YAG is primarily used for tissue removal and recanalization using fiber optic access. Combinations of the two are on the market. Two key procedures which find an increasing interest are TULIP (Trans Urethral Ultrasound guided Laser Induced Prostatectomy [79,80]) involving complex and expensive instrumentation, such as a probe tip incorporating an ultrasound transducer, pressure balloon and a right angle delivery system. Visual laser ablation of the prostate (VLAP) is presently more widely accepted as it uses mostly conventional instruments in conjunction with side firing fibers or contact probes. The prostate is visualized directly during the procedure and the laser is used circumferentially to coagulate or ablate the tissue. Again we find the requirements of local hemostasis together with ablation/cutting, presently using a Ho:YAG/Nd:YAG system. 4.6

Dental Surgery

According to Fig. 3, dental hard substances show two promising absorbencies for mid-IR laser use: one around 3 µm and one around the absorption peak of hydroxyapaitit (9.5 µm). Only the first one has practical relevance so far. Hundreds of publications describe mid-IR laser use under various aspects; Hibst [81] has done extensive biophysical research, in particular on the Er:YAG laser. Presently three elements determine the use of lasers in dentistry, the ablation efficiency in dental hard substances and filling materials, the collateral thermal damage, the temperature gradients in adjacent tissue and the surface roughness parameters, influencing the filling techniques, in particular their mechanical adherence on the surfaces. Attempts have been made also to seal cracks in dental material [82]. Oral mucosa treatment with the Ho:YAG laser has also been suggested [83]. Dentin consists approximately of 25% water, 30% organic material (collagen) and 45% carbonated hydroxyapatit [82]. Comparative studies have been conducted with Er:YAG, Er:YSGG [81,84], Nd:YAG [85] and the picosecond Nd:YLF [86]. The ablation efficiency, measured by the volume of ablated dentin/pulse lies around 0.1 mm3 / s for two subsequent pulses at pulse energies of 300 mJ [81], Enamel has similar ablation rates. Mechanical

534


drills are still more efficient, and achieve 1 mm3 / s, still substantially higher. Photoablative removal, however, causes less pain, a driving element for the introduction of IR photoablation. High ablation efficiencies can be explained by thermal ablation mechanisms, which eject rather large particles of fragmented hard substance. Surface melting, frequently observed in the ablation area as well as on the particles, is secondary. Er:YAG and Er:YSGG have comparable ablation efficiencies. The form and size of the ejected particles also describe – in a first approximation – the roughness of the ablated area if one ignores the collateral melting effect. Surface roughness is essential for fillings and their adherence on the newly created surfaces. The rougher, the better the adherence and the tighter the sealing between filling materials and teeth. Thermal collateral effects are essential limitations for mid-IR laser use. Due to the relatively low water content (25%) repetitive, subablative irradiation increases the thermal damage and leads to surface vitrification and glass-like alterations. Wet ablation on the surface has been investigated also [23] in order to reduce the collateral thermal adverse effects. Another element that needs to be considered is the damage of soft tissue, in particular the pulpa, which suffers irreversible damage if exposed to a temperature increase of more than 12 K above the physiologic temperature. Only the Er:YAG has so far been investigated and clinically introduced into dental surgery. Some Er:YAG systems have been FDA (US Food and Drug Administration) approved for clinical use. Practical limitations are still the availability of stable fiber optic delivery tools; the expensive monocrystalline sapphire fibers have been found to break easily. Taken together, the different concurrential requirements in dental surgery, consisting of high ablation efficiency for short treatment times, minimal collateral thermal damage in hard substances and pulpa as well as surface conditioning by creating the desired roughness for optimal sealing of the filling materials represent the pattern of clinical needs. The availability of shorter pulse lengths, tunability around the 3 µm water absorption peak and better laser delivery systems should favor more extended use in the future. Cavity preparation, ablation in enamel and dentin, and caries removal are fields where extended use of IR lasers can be expected. Postoperative comfort and reduced intra-operative pain are the principal advantages over conventional dissecting types of surgical procedures. Other advantages are the elimination of microorganisms at the surgical site, leading to more successful treatments of periodontal diseases. High temperatures and ejection of material also leads to quasi-sterilization of the root canal and thus successful pulpotomies. Recently, experimental work with FELs has been performed, using wavelengths between 6 and 7 m and at around 9.5 µm, the hydopxyapatit absorption peak [87,88,89]


4.7

535

Ophthalmology

The use of lasers in medicine began historically with retinal coagulations, using the argon laser, still the most used system in ophthalmogy. Mid-IR lasers [90] have been used for photoablation in the treatment of glaucoma for fistulating surgery, increasing the aqueous outflow [91], and various types of trabelectomy [91,92,93,94,95,96]. Typical photothermal applications like thermal keratoplasty, a refractive procedure, described earlier, with the Ho:YAG in the noncontact mode [17], the contact mode [10] as well as DTK using a CW diode laser source with the same absorption as the Ho:YAG in cornea [10,11] have been introduced more recently. As depth control is central for clinical stability, a tunable CW laser source (e.g. Cr:ZnSe) would further optimize DTK/LTK. Photoablative refractive surgery is presently performed only with excimer lasers (193 nm); UV photoablation has opened the era of submicron surgery, enabling the reshaping of the corneal surface, cutting corneal tissue, e.g. for corneal grafting [97,98], etc. Several attempts have been made to use mid-IR lasers to replace the expensive and sophisticated UV technology by more stable solid-state lasers. The HF laser was one historical example [2], followed by systematic investigations of the Er:YAG [24]. Corneal IR photobalation can be accomplished in the 3 and 6 µm water absorption band [33]. Only the Er:YAG has been brought to the level of experimental in vivo corneal reshaping. A free running Er:YAG has been used by Jean and coworkers [7,8], a Q-switched Er:YAG by Seiler and coworkers [9]. OPOs and other approaches using picosecond pulses have only reached the investigational stage [28]. Using nonlinear ablation characteristics with short pulses, tunability, in conjunction with a stable IR solid-state system are options for the future. Photofragmentation for cataract surgery [99] is currently performed with Er:YAG [100] and Ho:YAG lasers [100]. The procedure fragments the opacified lens in situ, saving the lens capsules; an artificial lens is then implanted into the capsule bag to restore vision. The procedure is still in an early stage of applications; it takes longer than conventional ultrasound based lens fragmentation and may be limited by formation of cavitation bubbles. 4.8

Cardiovacular Surgery, Angioplasty

The central goal of angioplasty is recanalization of occluded coronary vessels or laser thrombolyses through catheterization [102,103,104]. After UV 318 nm, the Ho:YAG has been used primarily, followed by experimental use of Er:YAG lasers. The selectivity of material ablation (plaque versus vessel wall) is an essential requirement (Fig. 4). Parameter setting and delivery characteristics are central [105]. The control of collateral adverse effects is essential, in particular pressure wave induced damage; it was considered to trigger local clotting and rethrombosis [106].

536


Free electron laser investigations over a wide wavelength range have been performed, preferring the 9.5 µm hydroxyapatit peak [107]. Laser recanalization is still an important topic and research is going on with several alternatives including a variety of delivery systems such as bare fibers, hot metal tips, sapphire probes, quartz domed fibers, and hand-held needle delivery systems. Selectivity of the ablation at minimal thermal and pressure wave damage (Fig. 17) as well as the use of shorter pulse lengths are elements to further optimize the laser side of the recanalization procedure; the procedure is frequently combined with mechanical balloon dilation or stent techniques. Transmyocardial revacularization to treat ischemic heart disease had a rather speculative background, expecting vessel growth along a transmural laser perforation; the procedure is currently performed with a CO2 laser. Epicardial and endocardial channels indeed trigger revascularization into the myocard; kilowatt pulses at 30–40 ms from a CO2 laser are used currently [108].

5

Perspective of a Tuneable IR Laser Source

As explained above, optimizing a laser source with regard to its ablation efficiency and collateral adverse effects (pressure wave generation, collateral thermal damage) is a three- dimensional problem: there are numerous indications that heat stress and pressure wave exposure are substantially different in various types of tissue and even cells. Although systematic investigations are still lacking, it would be extremely interesting and rewarding to determine the cell-specific or tissue-specific damage thresholds; this idea could be followed down to subcellular elements: cell membranes and mitochondria do react differently to a combination of thermal and stress exposure, as orientational experiments have revealed [109]. General experience tells us that thermal damage is relative “physiologic” and better tolerated by living material: heat is a more widespread environmental noxe and – under an evolutionary aspect – certainly a very old stress factor. The noxious potential of the pressure wave seems to consist primarily in the pressure gradient over time and not in the absolute pressure amplitude [110]. Figure 17 shows the scheme, relating tissue absorbance, collateral thermal damage and pressure wave exposure. The “optimal” combination is not necessarily the “minimum of all minima”; it may be any point along the pressure or the thermal curves. Tunability along the 3 µm water absorption flank would allow one to fine tune laser tissue interactions, minimizing the collateral effects in a tissue specific manner: Starting at an absorption maximum (e.g. 3 µm), the thermal damage increases with increasing wavelength, while the pressure wave decreases. It is highly unlikely that the damage threshold for temperature and the one for pressure exposure is the same for different tissues and cells or subcellular elements. Tunability along the 3 µm water


537

Fig. 17. Schematic of pressure-induced (curve rising to the left) and temperatureinduced (curve rising to the right) damage as a function of wavelength

absorption flank would allow fine-tuning laser tissue interactions, minimizing the collateral effects in a tissue-specific manner. Tunability for photothermal laser–tissue interaction likewise opens interesting perspectives. So far, the penetration depth for controlled laser heating of tissues is influenced by the tissue absorbencies and the optical focusing pathways applied. A tunable source (Fig. 18) would allow for controlled penetration and thus depth control of coagulation spots that have to be accurately placed in the deep corneal stroma in order to assure permanent refractive correction, while at the same time avoiding thermal damage to the endothelium [10]. Controlling the emission wavelength and thus influencing penetration depth could also be used for optimized photothermal destruction of tumors in very sensitive locations. Photoablation with a tunable source could be optimized with regard to the ablation rate, the dynamic collateral effects like pressure wave and recoil momentum; in vascularized tissues, tunability could be used to allow for a certain amount of wanted or needed surface coagulation for hemostasis. Feedback control options for sophisticated applications, like intraluminal thrombolysis, disc ablation, etc., and photoacoustic online control, described in this chapter could also be used in the contact mode. Another option for future refractive IR surgery are ns–fs pulses for athermal ablation or photospallation, eventually using nonlinear ablation characteristics. A future tunable IR laser source could be ideally complemented by online feedback controls in order to optimize the laser emission parameters, thus

538


Fig. 18. Water absorption and calculated penetration depth as a function of wavelength between 1.8 and 3 µm

optimizing the “magic” triangle of the competing elements of laser surgery: efficiency, selectivity and safety. Acknowedgement The authors are indebted to Prof. Michael Berlin, MD for advice and for reviewing the manuscript.

References 1. K. Winckler, B. Golz, W. Laskowski, T. Bende: Production of photoreactiveable lesion in the yeast s. cerevisiae by irradiation with 193 nm excimer laser light, Photochem. Photobiol. 47, 225 (1988) 514 2. O. Kermani, H. Lubatschowski, T. Ossaner et al. : Q-switched CTE:YAG (2, 69 µm) laser ablation: basic investigations on soft (cornea) and hard (dental) tissues. Lasers Surg. Med. 13, 537 (1993) 514, 535 3. T. Seiler, J. Marshall, S. Rothery: The potential of an infrared hydrogen fluoride (HF) laser (3 µm) for corneal surgery, Lasers Ophthalmol. 1, 49 (1986) 515, 522 4. H. Loertscher, S. Mandelbaum, R. K. Parrish et al.: Preliminary report on corneal incisions created by hydrogen fluoride laser, Am. J. Ophthalmol. 102, 217 (1986) 515 5. InPro, Gehm, Norderstett, Germany; Product information (2002) 6. G. Kahle, X. Daqun, T. Seiler et al. : Wundheilung der Kornea von Neuweltaffen nach fl¨ achiger Keratektomie: Er:YAG-Eximerlaser, Fortschr. Ophthalmol. 88, 380 (1991) 515


539

7. B. Jean, T. Bende, M. Matallana: Corneal photoablation in vivo with the Er:YAG laser – first report, Ophthalmic Technologies V, Proc. SPIE 2393, 69–77 (1995) 515, 535 8. B. Jean, M. Kriegerowski, M. Matallana et al. : Correction of myopia with Er:YAG laser fundamental mode. Photorefractive keratectomy, J. Refract. Surg. 11, 392 (1995) 515, 535 9. T. Seiler, H. Schmidt-Petersen, R. Leiacker et al. : Er:YAG laser photoablation of human cornea. Brief Report, Am. J. Ophthalmol. 120, 668 (1995) 515, 535 10. T. Bende, B. Jean, T. Oltrup: Laser thermokeratoplasty (LTK) using a CW diode laser, J. Refract. Surg. 15, 154–158 (1999) 516, 519, 535, 537 11. R. Brinkmann, N. Koop, G. Geerling et al. : Diode laser thermokeratoplasty: application strategy and dosimetry J. Cataract Refract. Surg. 24, 1195–207 (1998) 516, 519, 535 12. R. Brinkmann, B. Radt, C. Flamm et al.: Influence of temperature and time an thermally induced forces in corneal collagen and the effect on laser thermokeratoplasty, J. Cataract Refract. Surg. 26, 744–54 (2000) 518 13. F. Soergel, B. Jean, T. Seiler, T. Bende, S. M¨ ucke, W. Pechhold, L. Pels: Dynamic mechanical spectroscopy of the cornea for measurement of its viscoelastic properties in vitro, German J. Ophthalmol. 4, 151–156 (1995) 518, 519 14. E. Spoerl, U. Gent, K. Schmalfuß, T. Seiler: Thermo-mechanisches Verhalten der Hornhaut, Klin. Monatsbl. Augenheilkd. 208, 215 (1996) 518 15. T. Bende, T. Oltrup, B. Jean: Dielectric spectroscopy for in vivo measurements on corneal tissue, Proc. SPIE 3591, 39–44 (1999) 518, 519 16. T. Oltrup, T. Bende, K. D. Kramer, B. Jean: Dielektrische Spektroskopie zur nichtinvasiven Untersuchung von Hornhautgewebe, Biomed. Tech. 44, 78–82 (1999) 518, 519 17. J. M. Paral, Q. Ren, G. Simon: Noncontact laser photothermal keratoplasty. I: Biophysical principles and laser beam delivery system, J. Refract. Corneal Surg. 10, 511–518 (1994) 519, 535 18. V. K. Pustovalov, B. Jean: Application of infrared lasers for transscleral cyclophotocoagulation in ophthalmology for the treatment of glaucoma and simulation of the regimes of laser action. Laser Phys. 10, 785–795 (2000) 521 19. Q. Ren, V. Venugopalan et al. : Mid-infrared laser ablation of the cornea: A comparative study, Lasers Surg. Med. 12, 274–81 (1992) 521 20. M. J. Auerhammer, R. Walker, A. F. G. Van der Meer, B. Jean: Dynamic behavior of photoablation products of corneal tissue in the mid-IR: A study with FELIX, Appl. Phys. B 68, 111–117 (1999) 522, 525, 526 21. J. Auerhammer, B. Jean, R. Walker, A. Van der Meer: Dynamic behavior of the ablation products of biological soft tissue measured with FELIX, Laser Tissue Interaction VII, Proc. SPIE 2971, 74–79 (1997) 522 22. T. Bende, B. Jean, M. Matallana, T. Seiler, R. Steiner: Feuchte Hornhautfl¨ achenablation mit dem Er:YAG-Laser, Ophthalmol. 91, 651–654 (1994) 522 23. T. Bende, B. Jean, M. Matallana: Wet areal ablation with the Er:YAG laser (2.94 µm): First results, Lasers Light Ophthalmol. 5, 39 (1992) 522, 534 24. T. Bende, M. Kriegerowski, T. Seiler: Photoablation in different ocular tissues performed with an Er:YAG laser. Lasers Light Ophthalmol. 2, 263 (1989) 522, 535

540


25. T. Bende, T. Seiler, J. Wollensak: Photoablation mit dem Er:YAG-Laser an okul¨ aren Geweben. Fortschr. Ophthalmol. 88, 12–16 (1991) 26. O. Kermani, H. Lubatschowski, T. Asshauer: Q-switched CTE:YAG (2.69 µm) laser ablation: Basic investigations on soft (corneal) and hard (dental) tissues. Laser Surg. Med. 13, 537 (1993) 522 27. T. Bende, B. Jean, M. Matallana et al. : Einfluß der Pulsl¨ ange des Er:YAG Lasers auf die Photoablation in okul¨ arem Gewebe (Kornea und Sklera). Klin. Monatsbl. Augenheilkd. 202, 52 (1993) 522 28. W. B. Telfair, H. J. Hoffman: Evaluation of corneal ablation by an optical parametric oscillator (OPO) at 2.94 micron and an Er:YAG laser and comparison to ablation by a 193 nm excimer laser, Proc. SPIE 3246 (1998) 522, 523, 535 29. S. L. Jacques: Laser-tissue interactions: photochemical, photothermal and photomechanical, Lasers Gener. Surg. 72, 531–558 (1992) 522 30. R. S. Dingus, R. J. Scammon: Gruneisen-stress induced ablation of biological tissue, Proc. SPIE 1427, 45–54 (1991) 522 31. D. Stern, C. A. Puliafito, E. T. Dobi: Infrared laser surgery of the cornea: studies with a Raman-shifted neodymium:YAG laser at 2.80 and 2.92 micron. Ophthalmol. 95, 1434–1441 (1988) 522 32. T. Bende, R. Walker, B. Jean: Thermal collateral damage in porcine corneas after photoablation with free electron laser, Refract. Corneal Surg. 11, 129 (1995) 523 33. B. Jean, T. Bende: Photoablation of gelatin with the free-electron laser between 2.7–6.7 µm. Refract. Corneal Surg. 10, 433 (1994) 523, 525, 535 34. R. Walker: Infrarot Photoablation am Freien Elektronen Laser, Dissertation, Univ. T¨ ubingen (1998) 524, 525 35. R. Walker: Photoablation in biologischen Geweben als Funktion der Wellenl¨ ange am Elektronenstrahllaser. Thesis, MSc, Univ. T¨ ubingen (1994) 524 36. B. Jean: Medical and surgical applications of FELS. IEEE Proc. 1995 Particle Accelerator Conference, Vol. 1, 75–79 (1995) 525 37. M. Ostertag, R. Walker, H. Weber, L. Van der Meer, J. McKinley, N. Tolk, B. Jean: Photoablation in teeth with the free electron laser around the absorption peak of hydroxyapatit (9.5 µm) and between 6 and 7.5 µm, Lasers in Dentistry II, Proc. SPIE 2672, 181–192 (1996) 525 38. M. Ostertag, R. Walker, T. Bende, B. Jean: Optimizing photoablation parameters in the mid IR – a predictive model for the description of experimental data, Laser–Tissue Interaction VI, Proc. SPIE 2391, 138–149 (1995) 525 39. S. Jacques: Laser–tissue interactions – photochemical, photothermal, and photomechanical, Lasers Gener. Surg. 75, 531 (1992) 525 40. K. Nordwald, A. Holschbach, S. Lohmann et al. : Determination of acoustic shock waves generated by an (fundamental mode) Er:YAG laser in corneal photoablation, Invest. Ophthalmol. Vis. Sci. 37, S571 (1996) 525 41. B. Jean, T. Bende, M. Matallana: Noncontact photoacoustic spectroscopy during photoablation with a 193 nm excimer laser, German J. Ophthalmol. 2, 404–408 (1993) 525 42. K. Nahen, A. Vogel: Investigations on acoustic on-line monitoring of IR laser ablation of burned skin, Lasers Surg. Med. 25, 69–78 (1988) 525 43. G. Edwards, R. Logan, M. Copeland: Tissue ablation by a free-electron laser tuned to the amide II band, Nature 371, 416 (1994) 525


541

44. H. Specht, T. Bende, B. Jean, W. Fruehauf: Non contact photoacoustic spectroscopy (NCPAS) for photoablation control – data acquisition and analysis using cluster analysis, Proc. SPIE 3591, 33–38 (1999) 528, 529 45. BMBF Projekt 13N 6035 Final Report, Photablative Verfahren zur Keratorefraktions-Chirurgie mit gepulsten Lasern, T¨ ubingen (1995) 528 46. J. E. Adducci: Gynecologic surgery using the CO2 laser, Int. Surg. 63, 72–74 (1978) 529 47. J. Fanning: Laser vaporization conisation, J. Reprod. Med. 37, 534–536 (1992) 529 48. G. T. Absten: Physics of light and lasers, Obstet. Gynecol. Clin. North Am. 18, 407–27 (1991) 529 49. D. E. Townsend, E. J. Marks: Cryosurgery and the CO2 laser, Cancer 48, Suppl. 2, 632–637 (1981) 529 50. D. E. Brenner: Carcinoma of the cervix – a review, Am. J. Med. Sci. 284, 31–48 (1982) 529 51. M. S. Baggish, J. H. Dorsey, M. Adelson: A ten-year experience treating cervical intraepithelial neoplasia with the CO2 laser, Am. J. Obstet. Gynecol. 161, 60–8 (1989) 529 52. G. A. McIndoe, M. S. Robson, J. A. Tidy, W. P. Mason, M. C. Anderson: Laser excision rather than vaporization: the treatment of choice for cervical intraepithelial neoplasia, Obstet. Gynecol. 74, 1965–8 (1989) 529 53. L. Kjellberg, G. Wadell, F. Bergman, M. Isaksson, T. Angstr¨ om, J. Dillner: Regular disappearance of the human papillomavirus genome after conization of cervical dysplasia by carbon dioxide laser, Am. J. Obstet. Gynecol. 183, 1238–42 (2000) 529 54. B. Bl¨ umel, J. Nieder, O. Stefanovic, W. Weise: Lasereinsatz an der Cervix uteri, CO2 - oder KTP-Laser; Laser use in the uterine cervix, CO2 or KTP laser, Z. Gyn. 118, 458–461 (1996) 529 55. A. Bar-Am, Y. Daniel, I. G. Ron, J. Niv, M. J. Kupferminc, J. Bornstein, J. B. Lessing: Combined spectroscopy, loop conization, and laser vaporization reduces recurrent abnormal cytology and residual disease in cervical dysplasia, Gynecol. Oncol. 78, 47–51 (2000) 529 56. D. E. Townsend, R. U. Levine, C. P. Crum, R. M. Richart: Treatment of vaginal carcinoma in situ with the carbon dioxide laser, Am. J. Obstet. Gynecol. 143, 565–568 (1982) 529 57. T. M. McGee, E. A. Diaz-Ordaz, J. M. Kartush: The role of KTP laser in revision stapedectomy, Otolaryngol. Head Neck Surg. 109, 839–43 (1993) 530 58. C. L. Strunk, Jr, F. B. Quinn, Jr: Stapedectomy surgery in residency: KTP-532 laser versus argon laser, Am. J. Otol. 14, 113–117 (1993) 530 59. S. Kodali, S. A. Harvey, T. E. Prieto: Thermal effects of laser stapedectomy in an animal model: CO2 versus KTP, Laryngoscop. 107, 1445–50 (1997) 530 60. S. G. Lesinski, A. Palmer: Lasers for otosclerosis: CO2 vs. argon and KTP-532, Laryngoscope 99, Suppl. 45, 1–8 (1989) 530 61. M. Barbara, A. Caggiati, F. Attanasio, R. Filipo: Effect of mechanical trauma on the stapedial footplate after stapedectomy. A scanning electron microscopic study, ORL J. Otorhinolaryngol. Relat. Spec. 52, 286–291 (1990) 530 62. B. Kotecha, S. Paun, P. Leong, C. B. Croft: Laser assisted uvulopalatoplasty: An objective evaluation of the technique and results, Clin. Otolaryngol. All. Sci. 23, 354–359 (1998) 530

542


63. C. Nueruntarat: Laser-assisted uvulopalatoplasty: Short-term and long-term results, Otolaryngol. Head Neck Surg. 124, 90–93 (2001) 530 64. M. Littner, C. A. Kushida, K. Hartse, W. M. Anderson, D. Davila, S. F. Johnson, M. S. Wise, M. Hirshkowitz, B. T. Woodson: Practice parameters for the use of laser-assisted uvulopalatoplasty: an update for 2000, LARYA8 530 65. S. G. Lesinski: Lasers for otosclerosis – which one if any and why, Lasers Surg. Med. 10, 448–457 (1990) 530 66. M. Kautzky, A Tr¨ odhan, M. Susani, P. Schenk: Infrared laser stapedotomy, Eur. Arch. Otorhinolaryngol. 248, 449–451 (1991) 530 67. G. D. Casper, L. L. Mullins, V. Hartmann: Laser-assisted disc decompression: A clinical trial of the holmium:YAG laser side-firing fiber, J. Clin. Laser Med. Sur. 13, 27–31 (1995) 531 68. R. Kaufmann, A. Hartmann, R. Hibst: Cutting and skin-ablative properties of pulsed mid-infrared laser surgery, J. Dermatol. Surg. Oncol. 20, 112–118 (1994) 531 69. S. B. Jian, V. J. Levine, K. S. Nehal, M. Baldassano, H. Kamino, R. A. Ashinoff: Er:YAG laser for the treatment of actinic keratoses, Dermatol. Surg. 26, 437–440 (2000) 531 70. J. B. Newman, J. L. Lord, K. Ash, D. H. McDaniel: Variable pulse erbium:YAG laser skin resurfacing of perioral rhytides and side-by-side comparison with carbon dioxide laser, Lasers Surg. Med. 26, 208–14 (2000) 531 71. B. S. Biesman: Cutaneous facial resurfacing with the carbon dioxide laser, Ophthalmol. Laser. 27, 685–698 (1996) 531 72. R. E. Fritzpatrick: Depth of vaporation and residual thermal damage using multiple passes of ultrapulse CO2 laser, Lasers Surg. Med. 21, Suppl. 9, 31 (1997) 531 73. C. Weinstein, M. Scheflan: Simultaneously combined Er:YAG and carbon dioxide laser (derma K) for skin resurfacing, Clin. Plast. Surg. 27, 273–285 (2000) 531 74. B. Majaron, W. Verkruysse, K. M. Kelly, J. S. Nelson: Er:YAG laser skin resurfacing using repetitive long-pulse exposure and cryogen spray cooling: II. Theretical analysis, Lasers Surg. Med. 28, 131–137 (2001) 531 75. B. Majaron, K. M. Kelly, H. B. Park, W. Verkruysse, J. S. Nelson: Er:YAG laser skin resurfacing using repetitive long-pulse exposure and cryogen spray cooling: I. Histological study, Lasers Surg. Med. 28, 121–30 (2001) 531 76. J. E. Kopelman: Aesthetic facial skin resurfacing, the Er:YAG Laser versus the ultrapulsed carbon dioxide laser, Ophthalmic Clin. North Am. 11, 257– 266 (1998) 532 77. K. M. Bhatta: Lasers in urology, Lasers Surg. Med. 16, 312–330 (1995) 533 78. M. Kitagawa, H. Furuse, K. Fukuta, Y. Aso: HO:YAG laser resection of the prostate versus visual laser ablation of the prostate and transurethral ultrasound-guided laser induced prostatectomy: A retrospective comparative study, Int. J. Urol. 5, 152–156 (1998) 533 79. H. Schulze: TULIP, Transurethral ultrasound-guided laser-induced prostatectomy, World J. Urol. 13, 94–97 (1995) 533 80. W. Hochreiter, C. Hugonnet, U. E. Studer: Transurethrale Resektion der Prostata mit dem Holmium-Kontaktlaser. Ein Fortschritt in der Behandlung der BPH? Transurethral resection of the prostate with the Holmium contact laser. Progress in treatment of benign prostatic hypertrophy, Urologe A 38, 156–61 (1999) 533


543

81. R. Hibst: Untersuchungen zur Klinischen Anwendbarkeit der Gewebeablation, in Angewandte Lasermedizin II, H. P. Berlien, G. M¨ uller (Eds.) (Econ, Munich 1989) 533 82. G. I. D‘yakonow, V. I. Konov, V. A. Mikhailov, S. K. Pak, I. A. Shcherbakov, Cr, Er:YSGG lasers as an instrument for dental surgery, in S. I. O’Brien, D. N. Dederich, H. Wigdor, A. M. Trent (Eds.), Lasers in orthopedic, dental and veterinary medicine. Proc. SPIE 1424, 81–86 (1991) 533 83. M. Kautzky, M. Susani, P. Schenk: Ho:YAG infrarot Laser und UV-Excimer, Laser Effekte auf orale Schleimhautgewebe, Laryngorhinootol. 71, 347–352 (1992) 533 84. K. K.-F. Roth: Die Bearbeitung von Zahnhartgeweben mit Lasern der infraroten Spectralbereiches. Habilitationschrift, FB Medizin, Univ. Hamburg (1991) 533 85. E. Tasev, L. Delacretaz, L. W¨ oste: Drilling in human enamel and dentin with lasers: A comparative study. in: S. N. Joffe, K. Atsumi (Eds.), Laser Surgery: Advanced Characterization, Therapeutics and Systems II, Proc. SPIE 1200, 437–445 (1990) 533 86. M. H. Niemz, L. Eisenmann, T. Pioch: Vergleich von drei Lasersystemen zur Abtragung von Zahnschmelz, Schweiz. Monatsschr. Zahnmed. 103, 1252–1256 (1993) 533 87. K. U. Koster: Photoablation mit dem FEL an Zahnhartsubstanzen (Dentin, Schmelz) – Untersuchung der Ablationsrate als Funktion der Wellenl¨ ange. Dissertation, Univ. T¨ ubingen (1998) 534 88. M. Ostertag, R. Walker, H. Weber, L. Van der Meer, J. Mc Kinley, N. Tolk, B. Jean: Ablation in teeth with the free electron laser around the absorption peak of hydroxyapatit (9.5 µm) and between 6.0 and 7.5 µm, Laser in Dentistry II, Proc. SPIE 2672, 61–67 (1996) 534 89. K. Schwenzer: Zahnhartsubstanz Ablation mit dem Freien Elektronenlaser im Infrarot. Dissertation, Univ. T¨ ubingen (1996) 534 90. Q. Ren, V. Venugopalan, K. Schomacker, T. F. Deutsch, T. J. Flotte, C. A. Puliafito, R. Birngruber: Mid-infrared laser ablation of the cornea: A comparative study, Lasers Surg. Med. 12, 274–281 (1992) 535 91. T. S. Dietlein, P. C. Jacobi, G. K. Krieglstein: Ab interno infrared laser trabecular ablation: Preliminary shortterm results in patients with open-angle glaucoma, Graefes Arch. Clin. Exp. Ophthalmol. 235, 349–353 (1997) 535 92. S. A. Ozler, R. A. Hill, J. J. Andrews, G. Baerveldt, M. W. Berns: Infrared laser sclerostomies, Invest. Ophthalmol. Vis. Sci. 31, 2498–2502 (1991) 535 93. T. S. Dietlein, P. C. Jacobi, G. K. Krieglstein: Er:YAG laser ablation on human trabecular meshwork by contract delivery endoprobes, Ophthalmol. Surg. Laser. 27, 939–945 (1996) 535 94. T. S. Dietlein, P. C. Jacobi, R. Schr¨ oder, G. K. Krieglstein: Experimental Er:YAG laser photoablation of trabecular meshwork in rabbits: An in vivo study, Exp. Eye Res. 64, 701–706 (1997) 535 95. J. Kampmeier, M. Klafke, R. Hibst, S. Wierschin, E. Sch¨ utte, R. Steiner: Modifizierte Strahlungsapplikation bei der Er:YAG-Laser-abexterno-Sklerostomie, Klin. Monatsbl. Augenheilkd. 211, 48–52 (1997) 535 96. P. C. Jacobi, T. S. Dietlein, T. Colling, G. K. Krieglstein: Photoablative lasergrid trabeculectomy in glaucoma filtering surgery: Histology and outflow facility measurements in porcine cadaver eyes, Ophthalmol. Surg. Lasers 31, 49–54 (2000) 535

544


97. H. Loertscher, S. Mandelbaum, J. M. Parel, R. K. Parrish 2nd : Noncontact trepination of the cornea using a pulsed hydrogen fluoride laser, Am. J. Ophthalmol. 104, 471–475 (1987) 535 98. K. P. Thompson, E. Barraquer, J. M. Parel, H. Loertscher, S. Pflugfelder, T. Roussel, S. Holland, K. Hanna: Potential use of lasers for penetrating keratoplasty, J. Cat. Refract. Surg. 15, 397–403 (1989) 535 99. L. T. Sperber, J. M. Dodick: Laser therapy in cataract surgery, Curr. Opin. Ophthalmol. 5, 105–109 (1994) 535 100. B. S. Ross, C. A. Puliafito: Erbium YAG and holmium YAG laser ablation of the lens, Lasers Surg. Med. 15, 74 (1994) 535 101. W. Wetzel, R. Brinkmann, N. Koop et al. : Photofragmentation of lens nuclei using the Er:YAG laser: Preliminary report of an in vitro study, J. Ophthalmol. 5, 281 (1996) 102. H. J. Geschwind, J. I. Dubois-Rande, D. Murphy-Chutorian, T. Tomaru, R. Zelinsky, D. Loisance: Percutaneous coronary angioplasty with mid-infrared laser and a new multifibre catheter, Lancet 336, 245–246 (1990) 535 103. R. R. Heuser, S. S. Mehta: Holmium laser angioplasty after failed coronary balloon dilation: Use of a new solid-state, infrared laser system, Cathet. Cardiovasc. Diagn. 23, 187–189 (1991) 535 104. H. J. Geschwind, T. Tomaru, F. Nakamura, J. Kvasnicka: Holmium YAG laser coronary angioplasty with multifiber catheters, J. Intervent. Cardiol. 4, 171– 179 (1991) 535 105. O. Topaz, M. Mclvor, G. W. Stone, M. W. Krucoff, E. C. Perin, A. E. Foschi, J. Sutton, R. Nair, E. deMarchena: Acute results, complications, and effect of lesion characteristics on outcome with the solid-state, pulsed-wave, midinfrared laser angioplasty system: Final multicenter registry report. HO:YAG Laser Multicenter Investigators, Lasers Surg. Med. 22, 228–239 (1998) 535 106. L. van Erven, T. G. van Leeuwen, M. J. Post, M. J. van der Veen, E. Velema, C. Borst: Mid-infrared pulsed laser ablation of the arterial wall. Mechanical origin of “acoustic” wall damage and its effect on wall healing, J. Thorac. Cardiovasc. Surg. 104, 1053–1059 (1992) 535 107. B. Jean, R. Walker, M. Ostertag, T. Bende, M. Wehrmann, L. Van der Meer, K. Karsch: Photoablation in atherosclerotic plaque at 9.5 µm with the free electron laser FELIX, Proc. SPIE 2681, 245–251 (1996) 536 108. R. A. Hartman, P. Whittaker: The physics of transmyocardial laser revascularization, J. Clin. Laser Med. Surg. 15, 255–259 (1997) 536 109. B. Jean, G. Valet: unpublished data 536 110. D. C. Lamb, L. Reinisch, J. Tribble, R. H. Orsoff, T. J. Flotte, A. G. Doukas: The FEL as the ideal stress-wave generator, SPIE Intern. Symp. Biomed. Optics, (suppl), 2391, 192–301 (1995) 536

Index

ablation efficiency, 522, 523, 530, 531, 533, 534, 536 absorber, 511, 512 aqueous outflow, 535 artherosclerotic plaques, 513 carbonization, 522, 523, 529, 532 cataract surgery, 535 cavity preparation, 534 cervical intraepithelial neoplasia, 529 chromophores, 511, 512, 516 collagen shrinkage, 521, 532 collateral thermal damage, 514, 523, 524, 531, 533, 534, 536 condylomata accuminata, 529 corneal tissue, 513, 516, 518, 523, 524, 526, 535 coronary vessels, 535 CW irradiation, 518, 519, 521 cyclophotocoagulation, 516, 521 dental surgery, 511, 513, 524, 533, 534 dentin, 513, 533, 534 dermatology, 511, 531 desoxyhemoglobin, 512 dielectric spectroscopy, 518, 519 enamel, 513, 533, 534 endometrial ablation, 529 Er:Cr:YSGG laser, 514 Er:YAG laser, 514 esthetic surgery, 531 feedback laser control, 527 fistulating surgery, 535 free electron laser (FEL), 511, 522–526, 534 glaucoma, 535

gynecology, 511, 529 HF laser, 515, 522 histo-micrometry, 523 Ho:YAG laser, 516, 518, 519, 529–531, 533, 535 IR photothermal ablation, 521, 525 laser – assisted uvula palatoplasty, 530 – colposcopy, 529 – Er:Cr:YSGG, 514 – HF, 522, 535 – Nd:YAG, 516, 530, 531, 533 – thermo keratoplasty, 516 leukoplakia, 529 mechanical spectroscopy, 518, 519 melanin, 512, 521 Monte Carlo calculation, 518 Nd:YAG laser, 516, 530, 531, 533 neurosurgery, 511, 516, 530 online therapy control, 528 ophthalmology, 511, 535 optimized spiking mode, 515 otorhinolaryngology, 511, 530 oxyhemoglobin, 512 percutaneous discectomy, 531 photoacoustic spectroscopy (PAS), 528 photofragmentation, 535 photomechanical interaction, 525 photospallation, 522, 523, 537 photothermal laser tissue interaction, 516 pulse length, 522, 529, 532, 534, 536

546

Index

recanalization, 533, 535, 536 recoil momentum, 523, 525, 526, 537 refractive surgery, 511, 518, 535

trabelectomy, 535 transmyocardial revacularization, 536 TULIP, 533

skin rejuvenation, 531, 532 specific absorbers, 512, 525 spectroscopy – photoacoustic (PAS), 528

urology, 511, 533

tissue interaction, 516

xantophyll, 512

water absorption, 516, 525, 530, 531, 534–537

Index

4-dimet hylamino-N-methyl-4st ilbazolium-tosylate (DAST), 417 A19,368-370, 379 ablation efficiency, 522, 523, 530, 531, 533, 534, 536 absorber, 511,512 absorption, 222 224,228 - coefficient, 224, 228 combination, 451 fundamental, 97 line, 450, 451, 462, 463, 487, 488, 490, 494,495 linewidth, 448 loss, 68 - overtone, 451 spectroscopy, 462, 463, 471, 473, 474, 477, 483, 489, 494 spectrum, 471, 488 acceptance angle, 109, 110 acceptance bandwidth, 109 acoustic phonon, 370, 379 acousto-optic modulator (AOM), 476, 487 acousto-optical Q-switch, 367 ADP, 106, 120 AgAsS3, 120 AgGaS2 (AGS), 106, 120, 153, 154 AgGaSe2 (AGSe), 106, 120, 153, 154 AgSbS3, 106, 120 air-broadening coefficient, 466 airborne tunable diode laser system, 486 alexandrite, 358 A1GaAs/GaAs quasi-CW diode laser, 367 ambient air, 453, 473, 484, 490

ambient temperature, 492, 496 ammonia, 88, 132 concentration, 480 - mixture, 476 sensor, 490 amplitude modulation, 445, 470 amplitude-modulated phasemodulation spectroscopy (AMPM), 470 analo~digital converter, 488 angle tuning, 108, 457 angular acceptance bandwidth, 189 anti-Stokes, 364, 379 antimonide diode laser, 450, 451 aperture distance, 113 aqueous outflow, 535 ArF excimer laser, 383 artherosclerotic plaques, 513 astigmatic beam profile, 451,453, 458 astigmatic mirror Herriott cell, 492 astigmatic mirror Inulti-pass cell, 472 Auger recombination, 4, 9 11, 13, 16, 17, 41, 47 -

/3-BaB204 (BBO), 106, 120 Ba2NaNbsO15, 106 Ba2TiO3, 106 Ba(NOa)2, 355 357, 359, 360, 368, 370374, 378, 379 balanced beam, 467 balanced ratiometric detection (BRD), 467, 469 ballistic electron, 432 BaMoO4, 370, 371 band alignment, 6-8 type-I, 6, 7, 14 18 type-II, 6, 14 18, 28, 41 - type-III, 6, 18, 19, 45

546

Index

band offset, 7, 8, 14, 15, 20, 24, 26, 28, 32, 33 bandwidth, 459, 460, 468,480, 491,492 acceptance, 289 relative, 255 base-emitter curve, 469 BaWO4, 370 374, 376, 379 beam bounce pattern, 472 beam cleanup, 359, 360 beam divergence, 445,450, 451 beam pointing stability, 458,461,473 beam profile, 451, 453, 455, 458, 461, 462 beam steering, 90 beamsplitter/combiner (BS), 455 Beer-Lambert absorption law, 463, 464, 471,472, 478 BeSO4, 106 biharmonic pumping, 368 birefringent nonlinear crystal, 457 bleaching, 231 Bloch oscillation, 432 Boltzman distribution, 495 Bragg grating, 381-384, 386, 452 Bragg reflector, 75,449,493 branching ratio, 228 Brewster angle window, 495 broad ridge waveguides, 77 broadening (lifetime), 227 broadly tunable monolithic integrated nmlti-section diode laser chips, 456 buried heterostructure design, 70 buried heterostructure technique, 85 12CO2, 494 13CO2, 494 C2H2, 477, 481 C2H6, 481 CaCO3, 355, 356 CaMoO4, 370 carbonization, 522, 523, 529, 532 carrier lifetime, 426 428 cascade lasing, 237, 239 cascaded Raman fiber laser, 382 cataract surgery, 535 cavity dumping, 359 cavity eigenmode, 476 cavity finesse, 476 cavity preparatio m 534

cavity ring-down spectroscopy (CRDS), 130, 475,476 cavity transmission spectrum, 486 cavity-enhanced absorption spectroscopy (CEAS), 476, 477, 486 CaWO4, 365, 370, 379 CdGeAs2 (CGA), 106, 120, 153, 161 CdS, 106 CdSe, 106, 120, 160 ceramics, 222, 223 cervical intraepithelial neoplasia, 529 CH20, 450, 465, 480, 481, 483-486, 492494 CH4, 451, 477, 480 chalcogenide, 222, 243 glass, 232, 243, 244 characteristic temperature, 2, 19, 23 26, 30 32, 37, 44, 48 charge oscillation, 431, 432 Chernin multi-pass cell, 472 chromium-doped zinc selenide chalcogenide, 454 chromophores, 511,512, 516 cladding, 384 - pumped, 223 CO (carbon monoxide), 132, 447, 451, 477, 479 481,489, 490 CO laser, 86, 98 CO2, 132, 447, 477, 479 481,488,495 CO2 laser, 86,496 co-dopant, 240 co-doped, 231, 236, 240 coefficient effective nonlinear optical, 104, 111 nonlinear optical, 103, 106, 116, 117, 120 coherence length, 115, 193 Coherent Anti-Stokes Raman Scattering (CARS), 368 collagen shrinkage, 521,532 collateral thermal damage, 514, 523, 524, 531,533, 534, 536 collisional de-excitation, 478 combination-overtone m-vibrational line, 462 combined broadening, 464 comparison grating emission/facet emission, 74

Index conduction band, 447 condylomata accuminata, 529 confinement factor, 5, 9, 23, 42, 43, 46 continuous wave, 448, 450, 462 continuous wavelength tunability, 452 conversion efficiency, 110, 355 Cooper pairs, 431 core pumped, 223,226,228 corneal tissue, 513, 516, 518, 523, 524, 526, 535 coronary vessels, 535 coupled cavities, 359,361 coupled-wave equations, 182 Cr2+:Cdl ~Mn~Te~ 308 Cr2+:CdSe, 309 Cr2+:ZnS, 308 Cr2+:ZnSe, 157, 304, 376 Cr:YAG saturable absorber, 380 critical phase matching (CPM), 110, 152, 189 critical thickness, 17 cross-relaxation, 231, 233, 235-237, 240, 273 cross-section, 228, 238 cryogenic cooling, 447, 449,451,453 crystal, 106, 117, 120, 121 biaxial, 106 biaxial nonlinear, 105 centrosymmetric, 105 class, 105 group, 112, 113 monoclinic, 105 negative birefringent, 104, 112 - orthorhombic, 105 positive birefringent, 104, 112 trielinic, 105 uniaxial, 104, 106 uniaxial negative, 104 CS2,481 CsH2AsO4 (CDA), 106 CsLiB6010 (CLBO), 117, 120 CsTiOAsO4 (CTA), 106, 120 CW dye laser, 470 CW irradiation, 518, 519, 521 cyclophotocoagulation, 516, 521 -

-

-

D2,382 damage threshold, 118, 120 DAST, 417

547

DAST-DFG, 418 Davydov splitting (DS), 370, 379 dcc-coupled Peltier-cooled HgCdTe (MCT), 492 de-tuned grating, 73 degeneracy factor, 184 dental surgery, VII, 511, 513, 524, 533, 534 dentin, 513,533,534 dephasing, 368 370, 379 rate~ 368 time, 352, 368,371,372 dermatology, VII, 511, 531 desoxyhemoglobin, 512 detection limit, 133 detection scheme, 129 diamond, 368, 369 diamond heatsink, 71, 85 Dieke diagram, 261 dielectric plasmon mode, 452 dielectric spectroscopy, 518, 519 difference frequency generation (DFG), 102, 107, 108, 122, 416, 417, 419, 455, 457, 461,490, 492 beam profile, 491 system, 122 difference-frequency mixing, 429, 437 differential optical absorption spectroscopy (DOAS), 466 differential quantum efficiency, 23-25, 30, 31 diode pumping, 362 dipole moment, 100 direct diode-pumping, 455 dispersion, normal, 114 distributed Bragg reflector (DBR), 449, 456 distributed feedback diode laser (DFB), 458 divergence, 354 divergent beam profile, 451,453 domain length, 115 dopant gain region, 457 doped solid-state bulk material, 446 Doppler cross-section, 464 double heterostructure (DH) laser, 3-5, 14, 22, 40 double QW gain region, 66 -

548

Index

double refraction, 457 double-clad fiber, 224, 226,240 fiber lasers (DCFL), 381,387 geometry, 239 .... pump, 224 double-pass regime, 378 double-phonon-resonance gain region, 66 doubly resonant optical parametric oscillator (DROPO), 169, 170 dual chamber absorption cell, 495 dual channel absorption laser, 482 dual signal-wave quasi-phase-matched optical parametric oscillation (DSW-QPM-OPO), 416, 417 dual-beam detection, 468 dual-beam long-path absorption spectroscopy, 492 duty cycle, 71 Dy 3+, 243, 244 dye lasers, 357

equivalent noise bandwidth (ENBW), 468 Er 3+, 223, 227, 232, 243 Er/Yb fiber amplifier, 494 Er:Cr:YSGG laser, 157, 158, 163, 164, 293, 514 Er:YAG laser, 319, 320, 514 erbium fiber amplifier, 455 esthetic surgery, 531 ethane, 132 excited-state absorption (ESA), 228, 230, 238,239, 271 exciton, 432 external cavity diode laser (ECDL), 456, 458, 496 external pumping, 356, 364,366, 379 extraction efficiency, 66 eye-safe radiation, 360 Raman laser, 359,360 spectral region, 366 wavelength, 358

edge-emitting laser, 72 effective mass, 438 effective nonlinear coefficient, 104, 111 efficiency saturation, 372 electric-dipole approximation, 101 transition, 228 electrical pa~ssivation, 69 electrooptical Q-switching, 367 ellipsoidal off;axis mirror (OAE), 454 elliptical beam profile, 451,453, 458 emission cross-section, 222, 223, 228, 230 emission spectra, 228 enamel, 513, 533, 534 end-caps, 288 end-coupling, 381 endometrial ablation, 529 energy conservation, 103, 114, 128 energy migration, 231, 232 energy recycling, 241, 242 energy transfer, 230 232, 235, 237, 240 - upconversion (ETU), 231,240-242 energy-level scheme, 227, 233, 235, 237, 239, 240 environment, 131

F +, 357 F~-, 357 Fabry P6rot interferometer, 34 Fabry P6rot laser, 5, 9, 18, 22, 26, 30, 31, 43, 448, 451, 452, 456, 477 far field, 74 feedback laser control, 527 ferroelectric crystal, 459, 461 ferroelectric oxide, 145 fiber amplifier, 479, 491,494 fiber based pump source, 490 fiber deterioration, 380 fiber optic amplifier, 491 first Stokes - radiation, 355, 357, 359, 361, 362, 364, 365, 367, 378, 380, 385 wave, 359 fistulating surgery, 535 flat-flat resonator, 356 fluoride, 230 fluoride fiber , 222 fluoride glass, 222 fluoride, 222 FM spectrometer, 471 FM spectroscopy, 471 formaldehyde, 132,450

Index four-level CW laser, 234 fiber laser, 237 laser, 240 transition, 238 Fourier components, 74 Fourier transform IR spectrometer, 71 Fourier transform limit, 87 free carrier absorption, 9, 10, 23, 31 loss, 23, 46, 73 free electron laser (FEL), 511, 522--526, 534 free spectral range (FSR), 475, 477 free-space optical data link, 89 frequency conversion process, 100 frequency doubling, 364 frequency locking technique, 457 frequency mixing, 358 frequency modulation spectroscopy (FMS), 470, 471 fused quartz, 368 FWHM, 289, 448 GaA1As quasi-CW diode laser, 380 GaAs, 106, 171,428,429, 433, 437, 438, 440 gain coefficient, 352, 369 gain factor, 183 gain region design, 64 GaInAsSb/AiGaAsSb, 2 4, 11, 13 40 GaInAsSb/GaSb, 2, 7, 11, 15, 16, 20, 29 GaInSbAs quantum well, 451 GMnSbAs/GaSb, 451 GaLaS (GLS), 221, 222, 243 GaP, 106, 171 garnet, 262 gas sensing, 122, 131 GaSe, 106, 120, 153, 160, 161, 164, 173, 440,441 Gaussian, 464 466 Gaussian beam quality, 458 Gaussian fit, 491 GEISA databank, 466 germanosilicate fiber, 381, 382, 384 glass fiber, 380 glaucoma, 535 grating coupling coefficient, 68 grating spectrometer, 71

549

ground-state, 231 bleaching, 228, 240 transition, 233, 236 ground-state absorption (GSA), 228, 231, 235, 238, 271 guided-wave parametric process, 460 gynecology, VII, 511,529 H2-sensitized fiber, 382 H202, 481 heat, 226 heavy hole, 432 heine oxygenase, 489 Henry linewidth parameter, 87 Herriot multipass cell, 87 Herriott multi-pass cell, 472, 477, 482, 484 HF laser, 515, 522 HgCdTe, 2, 4, 11, 13 detectors, 71 HgOa2S4, 120 HgS, 120 high inherent mid-IR wavelength stability, 492 high sensitivity, 490 high speed modulation, 89 high thermal conductivity, 70 high-power continuous-wave DFG source, 493 higher Stokes components, 365 histo-micrometry, 523 HITRAN database, 88, 98, 463, 466, 472,495 Ho a+, 232,243,244 Ho:YAG laser, 516, 518, 519, 529-531, 533, 535 holography, 69 homogeneous line broadening, 368,369 homogeneous width, 368 homonuclear diatomic molecule, 463 HWHM, 452,453,465 hydride vapor phase epitaxy (HVPE), 459 hydrogen fluoride (HF), 494 hydrogen-bonded cluster, 475 idler, 102, 181, 399, 402,457, 460, 461 in situ measurement, 467 InAs, 437, 438

550

Index

InAs/A1Sb, 5, 7, 49 InAs/GalnSb, 4, 7, 11, 41, 45-48 InAs/GaSb, 7, 18, 19 InAsSb/InAs, 7, 40, 41 InAsSb/InAsP, 40, 44 InAsSb/InAsPSb, 3, 40 index coupling, 74 index ellipsoid, 104 indicatrix, 104 indium-antimonide photovoltaic detector, 482 inhomogeneous broadening, 354, 368 injection efficiency, 66 injection seeding, 403-410 injection-seeded TPG (is-TPG), 403 410 injector region, 64 InP, 106, 437 InSb, 437 integral cross-section, 354, 369, 371, 379 integrated cavity output spectroscopy (ICOS), 476, 477, 486 interaction length, 111,352,354,380 interband cascade laser, 3, 4, 12, 13 interfacial roughness scattering, 68 interminiband transition, 434 internal quantum efficiency, 24, 28, 30 intersubband cascade laser, 4, 13 intersubband electron scattering time, 67 intersubband transition, 64, 451 interwell intersubband transition, 433 intracavity pumping, 359,360, 362 Raman oscillation, 364 -- SRS generation, 366 ion host interaction, 261 IR photothermal ablation, 521,525 IV VI semiconductor, 3, 5, 7

KGW, 355,364 366 KH2PO4 (KDP), 106, 120 KNbO3, 106, 117, 120, 193 KTiOAsO4 (KTA), 106, 120, 145, 146, 149, 151 153, 160, 173 KTiOPO4 (KTP), 106, 117, 120, 145, 146, 149-152, 155, 156 KY(WO4)2, 376

-

-

-

-

-

Jahn Teller effect, 267 Jamin interferometer, 479 junction down mounting, 70 KaLi2NbsO15, 106 KD~PO4 (KD*P), 106, 120 Kerr cell, 351 KGd(WO4)2, 355, 357, 371, 376

lanthanide, 227 contraction, 263 laser, 234 assisted uvula palatoplasty, 530 beam, 457, 473, 478480, 487, 489 cavity, 351 CO, 86, 98 Co 2+, 296 CO2, 86, 98, 496 color-center, 310 314 colposcopy, 529 Cr 2+ , 300

damage threshold, 355-357, 362,366, 380 DCFL, 387 double heterostructure, 3 5, 14, 22, 40 ECDL, i00, 121, 126, 127, 456, 458, 496 Er 3+, 319 Er:Cr:YSGG, 157, 158, 163, 164,293, 514 external cavity diode, i00, 121, 126, 127 Fabry P~rot, 448, 451,452, 456, 477 Fe 2+ , 309 HF, 522,535

Ho3+, 314 interband cascade, 3, 4 - intersubband cascade, 4, 13 lead salt, 2 4, 11, 13, 100 line, 351 Nd:YAG, 126, 147-156, 159, 162, 166-169,358, 404, 408,412, 417, 516, 530, 531,533 Ni 2+, 295 296 power scaling, 287 pumping, 285 quantum cascade (QCL), 3, 64 68, 99, 447, 451,453, 475, 481,489

Index radiation filtering, 476 Raman fiber, 380 383, 386, 387 nanosecond, 379 picosecond, 361,367 ruby, 351, 358 self-terminating, 231 signal, 102, 125 source, 351, 445 447, 453, 455 457, 459, 475,476, 478 480, 491, 492,494, 496 spectrum, 351 surface-emitting, 74 surface-skimnfing, 68 thermo keratoplasty, 516 thin-disk, 288 three-level, 234, 238 Tm a+, 316 transition-metal, 295. 310 transition-metal-ion-doped, 226 - tunable, 357 tuning, 289 292 vibronic, 260 lasing threshold, 487 lateral current injection, 68 lattice phonon, 368,370 lead salt laser, 2 4, 11, 13 lead-salt diode laser, 447, 449, 450, 481, 492,496 lens ducts, 287 leukoplakia, 529 LiBaO5 (LBO), 106, 117, 120 Libbrecht design, 452 LiF, 357 LiF:F2, 372 LiF:F2, 366 color center laser MALSAN-201,366 color center tunable lasers, 368 lifetime, 33, 40, 228, 230, 231 quenching, 226 light detection and ranging (LIDAR), 358, 360, 361, 466 light hole, 432 LiIOa, 106, 120, 361, 362 LiNbOa, 106, 117, 120, 141, 145, 146, 149, 153, 155, 163 167, 367, 398, 399, 401,402,460, 492

551

periodically poled (PPLN), 120, 126, 128, 142, 145 149, 153, 155, 157-162, 166-173,417, 455, 459, 461,462 linear high-finesse optical cavity, 488 lineshape, 464, 465 liquid nitrogen dewar, 448,450, 481 - optical cryostat, 487 liquid SRS media, 353 LIRF, 468 Lissajous pattern, 482 LiTaOa, 106, 117, 120, 399, 459, 460 LO-phonon scattering, 433, 434 lock-in amplifier, 468, 470, 471, 483 logarithmic conformance, 469 long optical path length spectroscopy, 472 Lorentz force, 437 Lorentzian, 466 loss, 221, 223, 224, 243, 244 loss coupling, 74 low-loss single mode fiber, 496 LT-GaAs, 428 -

MAK, 132 mechanical spectroscopy, 518, 519 medicine, 131 melanin, 512, 521 MEMS, 456 methane, 126, 132 methanol, 480 MgO:LiNb3, 120, 163, 402 microphone array, 87 microphone responsivity, 88 mid-IN detector, 469 mid-IN spectroscopic application, 480 minimal detectable CH20 concentration, 493 minimum detectable concentration, 467 minimum detectable fractional absorption, 468 minimum relative detectable concentration, 468 miscibility gap, 21, 22, 31, a3, 44 mode field diameter (MFD), 384 molecular beam, 475 molecular beam epitaxy (MBE), 450, 459 molecular beam epitaxy growth, 68

552

Index

molecular collision, 464, 465 molecular rot ationM-vibrational (ro-vibrational) state, 462 molecule, 462, 463, 465, 468, 471, 472, 478,481,486 momentum conservation, 107 monochromatic emission spectrum, 470 monolithic non-planar ring oscillator (NPRO), 126, 4.57 Monte Carlo calculation, 518 MOSFET, 488 multi-pass cell, 472,473,477, 486,489, 492, 496 multimode waveguide, 381,384 multiphonon decay, 368 relaxation, 220, 222, 226, 228, 230, 354 multiple harmonics, 470 multiple linear regression approach, 484 multiple quantum-well, 433 nmltiple quantum-well heterostructure, 451 multiple spatial mode, 460 multiplet, 227, 228 (NO3)-, 357 N2, 463, 476, 488 NaNO3, 355, 356, 368 nanosecond pumping, 366 nanosecond Raman laser, 379 nanosecond SRS, 366 ~- amplification spectroscopy, 368 narrow-linewidth diode laser source, 491 Nd DCFL, 381 Nd 3+, 223, 225, 243, 244 Nd3+:PbWO4, 380 ND4H2PO4, 106 Nd:GGG, 372 Nd:KGd(WO4)2 (Nd:KGW), 362, 367 Nd:KGW, 364 367 Nd:KLa(MoO4)2, 364 Nd:KY(WO4)2 (Nd:KYW), 362 Nd:KYW, 364, 365 Nd:NaLa(MoO4)2, 364 Nd:YAG laser, 126, 147 156, 159, 162, 166 169,358, 404, 408, 412, 417, 516, 530, 531, 533

near-Gaussian intensity distribution, 492 neurosurgery, VII, 511,516, 530 NH3,451, 477, 479, 481,490 NH4H2PO4, 106, 120 NO, 476, 481, 486488 NO-cGMP, 489 noise reduction, 466,467, 469-471,477, 491 noise-immune cavity-enhanced optical heterodyne spectroscopy (NICE-OHMS), 477, 478 non-critical phase-matching (NCPM), 146, 152, 157, 159, 160, 174, 189 non-phase-matched, 440 non-phase-matched process, 440 non-planar ring oscillator (NPRO), 126, 457 non-radiative (SHR) coefficient, 9, 46 non-resonant probe laser beam, 479 noncollinear phase matching, 399, 402, 410 nonlinear figure of merit (FOM), 114, 186, 267 nonlinear intracavity dumping, 372 nonlinear optical coefficient, 103, 106, 116, 117, 120 effective, 111, 112 nonlinear optical figure of merit (NLO FOM), 143, 155, 173 nonlinear processes, 399, 400 nonlinear refractive index, 353 nonradiative decay, 228,230,269 nonstationary pumping, 362 numerical aperture, 224 off-axis cavity alignment, 477 off-axis mirror, 482 online therapy control, 528 OPG-OPA, 142, 167, 169 ophthalmology, 511, 535 OPO cavity, 143 145, 148, 150 optical absorption spectrum, 462 optical breakdown, 354 optical fiber, 446 optical gain, 9, 42 maximum gain, 9, 42 modal gain, 9, 42, 43, 46 optical isolator, 493

Index optical loss, 9, 42, 44-46, 383, 384, 386 internal loss, 9, 23, 24, 28, 30, 41, 44 mirror loss, 9 optical parametic amplifier (OPA), 185 optical parametric amplification (OPA), 121, 127 optical parametric generation (OPG), 102, 121, 127 optical parametric generator (OPG), 143, 145, 161, 163-165, 185, 403 optical parametric oscillation (OPO), 462 optical parametric oscillator (OPO), 141--170, 173, 185, 455,457, 462, 476, 479 optical pathlength, 467, 472,476, 482, 489, 490 optical pump field, 181 optical rectification (OR), 103,437,440 optimized spiking mode, 515 optimum orientation, 365 orientation patterned GaAs (OPGaAs), 171 174 oscillator strength, 268 otorhinolaryngology, VII, 511,530 overgrown grating, 70 overheating of active region, 80 overtone ro-vibrational line, 462 oxides, 230, 243 oxyhemoglobin, 512

-

P205, 383,384, 386 p-n junction, 447 P = O bonds, 383 parabolic off-axis mirror (OAP), 454 parametric conversion source, 475 parametric fluorescence, 183 parametric frequency conversion, 446, 447, 455, 456 parametric frequency down-conversion, 142 parametric gain, 143, 145, 146, 153, 164 parametric lmninescence, 183 parametric noise, 143, 145, 183 parity, 228 passive Q-switching, 355, 367 passive mode locking, 364 Pb(NO:~)2, 355, 356 Pb-salt diode laser source, 448 450, 481

553

PbEuSeTe, 449 PbMoO4, 370 PbSnTe, 449 PbWO4, 370, 379, 380 Pd/A1203, 484 peak cross-section, 369, 370 Peltier cooling, 89, 453, 480 percutaneous discectomy, 531 period length, 117 periodic poling, 115, 193 PFBHA, 493 phase matching, 105, 129, 142, 184, 361, 367, 401 angle, 108 birefringent, 107 condition, 457 critical, 109, 110, 152, 189 first-order quasi-, 117 non-critical, 109, 146, 152, 157, 159, 160, 174, 189 non-perfect, 111 noncollinear, 399, 402, 410 perfect, 111 quasi-, 114 - type I, 107, 112 - type II, 107, 112 type III, 107 phase mismatch, 107 parameter, 182 phase-matched optical rectification, 440, 441 phonon absorption, 425,442 phonon energy, 220-222, 230,232, 243 phosphosilicate fibers, 382, 383, 386 photo-Dember field, 437 photoacoustic spectroscopy (PAS), 87, 467,478-480, 528 photoconductive switch, 425, 426, 428, 429 photofragmentation, 535 photomechanical interaction, 525 photon (idler), 398 photon (pump), 398 photospallation, 522, 523, 537 photothermal absorption spectroscopy, 479 photothermal laser tissue interaction, 516 -

554

Index

phototherlnal spectroscopy, 478 picosecond pumping, 362, 378 picosecond Raman gain, 357, 376 picosecond Raman laser, 361,367 pigtailed LD array, 384 plane-concave (Raman) laser cavity, 367, 379 platelet aggregation, 489 polariton, 399 polarization, 364, 365 dielectric, 100 extra-ordinary, 107 nonlinear, 101 of light, 107 - ordinary, 107 rotator, 468 polyatomic molecule, 463 population inversion, 354, 434, 442 portable gas sensor, 490 potassium titanyl phosphate (KTiOPO4, or KTP), 106, 117, 120, 145, 146, 149-152, 155, 156, 459 PP KTA, 119, 120, 146 151, 192 PP KTP, 119 121, 146, 149, 150, 192 195 PP RTA, 120, 121, 146, 149-150, 192 ppbv, 463, 476, 479, 480, 484, 490, 492 PPLN, 120, 121,126, 128, 142,145 149, 153, 155, 157 162, 166 173, 191 212, 417, 455, 459, 461,462,491 ppmv, 463, 467, 480, 490 pptv, 450, 463, 467, 472, 486, 493 Pr a+, 240, 243, 244 Pra+-co-doped, 240 probe beam, 468,472,479 processing steps, 68 pulse compression, 365 pulse length, 522, 529, 532, 534, 536 pump, 102, 125, 126, 224,399,402,404 pump absorption, 224, 226,228 pump cladding, 223, 224 Q-switching, 351 QPM-PPLN, 490, 492,494 quantum cascade laser (QCL), 3, 64 68, 447, 451,453,475,481,489 quantum efficiency, 220, 222, 226, 230, 234, 240

quantum well, 432, 449, 451,452 quantum yield, 284, 354, 357 quartz, 106, 120 quasi-CW, 362 quasi-phase-matched material (QPM), 193, 447 quasi-phase-matching (QPM), 114, 142, 143, 145 148, 150, 151, 162, 170 174, 459, 461 first-order, 117 property, 459 quenching, 231, 233,240,244 radiation trapping, 275 radiative (spontaneous) coefficient, 9, 46 radiative decay, 228, 230 radiative recombination, 6, 9, 17, 19 radiative transition, 228 radioisotope counting technique, 489 Raman amplifier, 364 Raman fiber amplifier, 381,386 Raman fiber laser, 380 383,386, 387 Raman frequency shift, 353, 354, 362, 374, 382 Raman gain, 354 coefficient, 362, 386 picosecond, 357, 376 transient, 353 Raman laser - eye-safe, 359,360 fiber, 380 383, 386, 387 nanosecond, 379 picosecond, 361,367 Raman line broadening, 353, 371,379 Raman linewidth, 352 Raman scattering (RS), 351 cross-section, 352 integral cross-section, 352, 371 - peak cross-section, 353 Raman spectral line, 352 rapid background subtraction, 483 rapid sweep integration, 471 rare-earth ion, 222, 226 228, 231, 243, 244 rare-earth solubility, 222 rare-earth-doped chalcogenides, 222 fiber laser, 232

Index heavy-metal oxide, 243 solid-state laser, 230, 231 rate constant, 228, 230 RbH2PO4 (RDP), 106 RbTiOAsO4 (RTA), 106, 120, 145, 146, 149 151 reabsorption, 226, 230, 234 recanalization, 533, 535,536 recoil momentum, 523, 525,526,537 recycling, 239, 240 reference detector, 469 reflecting objective (RO), 454 refractive index, 101, 222,224, 352, 381 extraordinary, 104 fluctuation, 224 modulation, 452 ordinary, 104 refractive surgery, 511, 518, 535 relaxation rate, 370 resonant excitation beam, 479 resonator, 225, 230 ring-down decay, 476 to-vibrational line, 462 room temperature CW operation, 84 ruby laser, 351, 358 sampled grating distributed bragg reflector diode laser (SG-DBR-DL), 458 saturable absorber, 367 SBD, 406 scheelite~ 370 structure, 370, 379 Schottky barrier diode (SBD), 404 Se, 106 second harmonic, 357, 367 second harmonic generation (SHG), 101, 107, 127, 355 second Stokes radiation, 355, 360, 361,374 wave, 359 selectivity, 445, 465, 467, 471, 479,480, 495 self-conversion, 364, 374 self-focusing, 353 self-frequency conversion, 367 self terminating laser, 231 self-terminating transition, 237 semiconductor surface, 425, 437

555

sensitivity, 445, 467-469,471,476, 477, 479, 480,486,490, 492, 496 sensitizing, 231, 235 237, 244 series resistance, 5, 20, 37 shear stress, 85 short-path absorption cell, 469 Si-prism coupler, 411,415 side pmnping, 367 signal, 181 signal enhancement, 466, 467, 472 474, 477 480 signal laser, 102, 125 silica, 221 fiber, 381,386 glass, 232 silicate, 222 single beam interferometer, 471 single longitudinal mode (SLM), 165, 166, 168 170, 173 single phonon bridge processes, 370 single-mode fiber core, 381 single-pass difference-frequency generation: 490 single-pass pumping, 357 single-pass Raman shifting scheme, 366 single-pass scheme~ 355,357, 366 singly resonant oscillator (SRO), 147~ 167 SiO2, 106, 120, 383, 386 skin rejuvenation, 531, 532 slope efficiency, 232, 236, 238 242 solid-state laser spectrometer, 358 solubility, 220, 222 specific absorbers, 512, 525 spectral acceptance, 110 bandwidth, 189 spectral-line broadening, 222, 227,228 spectroscopic gas sensor, 476, 486 spectroscopy, 97 cavity ring-down (CRDS), 130, 475, 476 FM, 471 gas, 132 photoacoustic, 129 photoacoustic (PAS), 87, 467, 478 480,528 transmission, 129 SrMoO4~ 370, 371

Index heavy-metal oxide, 243 solid-state laser, 230, 231 rate constant, 228, 230 RbH2PO4 (RDP), 106 RbTiOAsO4 (RTA), 106, 120, 145, 146, 149 151 reabsorption, 226, 230, 234 recanalization, 533, 535,536 recoil momentum, 523, 525,526,537 recycling, 239, 240 reference detector, 469 reflecting objective (RO), 454 refractive index, 101, 222,224, 352, 381 extraordinary, 104 fluctuation, 224 modulation, 452 ordinary, 104 refractive surgery, 511, 518, 535 relaxation rate, 370 resonant excitation beam, 479 resonator, 225, 230 ring-down decay, 476 to-vibrational line, 462 room temperature CW operation, 84 ruby laser, 351, 358 sampled grating distributed bragg reflector diode laser (SG-DBR-DL), 458 saturable absorber, 367 SBD, 406 scheelite~ 370 structure, 370, 379 Schottky barrier diode (SBD), 404 Se, 106 second harmonic, 357, 367 second harmonic generation (SHG), 101, 107, 127, 355 second Stokes radiation, 355, 360, 361,374 wave, 359 selectivity, 445, 465, 467, 471, 479,480, 495 self-conversion, 364, 374 self-focusing, 353 self-frequency conversion, 367 self terminating laser, 231 self-terminating transition, 237 semiconductor surface, 425, 437

555

sensitivity, 445, 467-469,471,476, 477, 479, 480,486,490, 492, 496 sensitizing, 231, 235 237, 244 series resistance, 5, 20, 37 shear stress, 85 short-path absorption cell, 469 Si-prism coupler, 411,415 side pmnping, 367 signal, 181 signal enhancement, 466, 467, 472 474, 477 480 signal laser, 102, 125 silica, 221 fiber, 381,386 glass, 232 silicate, 222 single beam interferometer, 471 single longitudinal mode (SLM), 165, 166, 168 170, 173 single phonon bridge processes, 370 single-mode fiber core, 381 single-pass difference-frequency generation: 490 single-pass pumping, 357 single-pass Raman shifting scheme, 366 single-pass scheme~ 355,357, 366 singly resonant oscillator (SRO), 147~ 167 SiO2, 106, 120, 383, 386 skin rejuvenation, 531, 532 slope efficiency, 232, 236, 238 242 solid-state laser spectrometer, 358 solubility, 220, 222 specific absorbers, 512, 525 spectral acceptance, 110 bandwidth, 189 spectral-line broadening, 222, 227,228 spectroscopic gas sensor, 476, 486 spectroscopy, 97 cavity ring-down (CRDS), 130, 475, 476 FM, 471 gas, 132 photoacoustic, 129 photoacoustic (PAS), 87, 467, 478 480,528 transmission, 129 SrMoO4~ 370, 371

556

Index

SrWO4, 370 374,379 standing-wave resonators, 360 Stark, 227, 233,238 component, 238 level, 234 steady-state gain coefficient, 358 steady-state nanosecond oscillation, 357 steady-state regime, 352, 353, 355, 362, 369, 371, 378 steady-state RS peak cross-section, 355 stimulated emission, 228,230 stimulated Raman scattering (SRS), 351 -active molecular gases, 353 amplifier, 364,365,371 converters, 368 gain, 354, 357, 370, 371 gain coefficient, 356, 357, 366 lasers, 358 peak cross-section, 371 shifters, 358 threshold, 353, 355 357, 361, 362, 365, 366, 368, 369, 371, 374 stoichiometry, 448 Stokes efficiency, 234,238 240 Stokes shift, 364 366, 382, 386 Stokes wave, 352,359 stopband, 73, 76 -width, 73, 76 strain-compensated material, 67, 77 sub-ppb detection sensitivity, 494 subpicosecond pumping, 371 subpicosecond regime, 365 subthreshold, 72 sulfide fiber, 244 sulfide glass, 222 sum frequency generation (SFG), 102, 107, 127 superlattice, 432, 434 gain region, 66 superradiant emission, 380 surface-depletion field, 437 surface-emitting laser, 74 surface-plasmon mode, 452 surface-skimming laser, 68 susceptibility linear, 100 nonlinear, 101

symmetrical vibrations, 357 synchronous pumping, 196 Tanabe Sugano diagram, 265 Tb 3+ , 243 telecommunication, 89 telluride, 106 TEM00, 359, 455,458, 460, 462, 488 temperature acceptance bandwidth, 110, 128, 189 temperature difference, 80 temperature tuning, 109 thermal conductivity, 85,221 223 measurement, 81 thermal management, 70 thermal motion, 464, 465 thermal resistance, 30, 37 thermal roll-over, 78 third harmonic generation, 127 third Stokes, 374, 381,382, 386 three QW gain region, 66 three-level laser, 234, 238 threshold, 228, 230,234 236, 240, 243, 282 of an OPO, 145 of a single-pass OPG, 145 THz wave, 397, 399, 402,411,416 parametric generator (TPG), 402 parametric oscillator (TPO), 402, 412, 413, 416 Ti:sapphire, 358 tissue interaction, 516 Tm 3+, 232, 235 237, 243 toroidal off-axis mirror (OAT), 454 totally symmetric Raman active vibration, 368 trabeleetomy, 535 trace gas detection, 445,447, 449,462, 467, 478,479,490,496 transient Raman gain, 353 transient regime, 352,353,369 transition-metal ion, 223, 226 transition-metM-ion-doped laser, 226 transmyocardial revacularization, 536 transparency, 220 222, 232 range, 118-120 triggering circuit (TC), 488 TULIP, 533 tunability, 234, 235, 238

Index of lasers, 228 tunable diode laser absorption spectroscopy (TDLAS), 2, 34~ 38 tunable laser, 357 tunable optical parametric oscillator, 461 tunable solid-state laser characteristics, 455 tunable solid-state laser source, 447, 486 tuning ~talon, 291,455 acousto-optical, 291 angle, 108 birefringent filter, 290 electro-optical, 292 grating, 289 prism, 289 temperature, 109 wavelength, 109 tuning characteristic, 128 two-photon resonance~ 352 two-tone frequency-modulation (TTFM), 470 type-I laser, 2, 11, 14, 17, 18, 20, 23 29, 34, 40--42 type-II laser "W" laser, 1, 3, 4, 11, 13, 41, 46 47 interband laser, 1, 2, 4, 41 intersubband laser, 1, 49 MQW laser, 4, 13, 28 33, 41 44 type-III laser, 18, 45 48 ultra-high reflective spherical mirror, 473 unipolar semiconductor injection laser, 451 upconversion, 273, 370 urology, VII, 511,533 variable-ratio beamsplitter, 468 vascular smooth muscle cells (VSMCs), 489 vertical cavity surface emitting laser (VCSEL), 26, 27, 456, 494

557

vibrational excitations, 354, 368 vibronic laser, 260 vibronic mode, 374 vibronically broadened transition, 454 Voigt profile, 465, 466 W W distance, 370 walk-off angle, 113 walk-off distance, 113 Wannier Stark (WS) state, 432 water absorption, 516, 525, 530, 531, 534 537 waveguide layer, 6, 10, 42, 46 waveguide loss, 68 waveguide phase matching, 460 wavelength division multiplexer (WDM), 381,382, 493 wavelength modulation, 481 spectroscopy (WMS), 470, 471,493 wavelength tunability, 118, 445, 456, 462 wavelength tuning, 76, 109 wavenumber time domain, 475 wavevector propagation collinear, 108 wet chemical etching, 69 White multi-pass cell, 472 working place concentration, 131 maximum permissible, 132 xantophyll, 512 YAG material, 457 Yb DCFL, 382,386 Yb 3+, 223, 225, 237 ZBLAN, 221 223, 228,230, 232,243 zero-background subtraction detection, 467 zincblende, 171 ZnGeP2 (ZGP), 106, 120, 153, 155 160, 163 -173 ZnSe, 171

Topics in Applied Physics 72 Glassy Metals III Amorphization Techniques, Catalysis, Electronic and Ionic Structure By H. Beck and H.-J. Giintherodt (Eds.) 1994. 145 figs. XI, 259 pages 73 Hydrogen in Metals III Properties and Applications By H. Wipf (Ed.) 1997. 117 figs. XV, 348 pages 74 Millimeter and Submillimeter Wave Spectroscopy of Solids By G. Griiner (Ed.) 1998. 173 figs. XI, 286 pages 75 Light Scattering in Solids VII Christal-Field and Magnetic Excitations ByM. Cardona and G. Giintherodt (Eds.) 1999. 96 figs. X, 31o pages 76 Light Scattering in Solids VIII C6o, Semiconductor Surfaces, Coherent Phonons By M. Cardona and G. Gfintherodt (Eds.) 1999. 86 figs. XII, 228 pages 77 Photomechanics By P. K. Rastogi (Ed.) 20o0, 314 Figs. XVI, 472 pages 78 High-Power Diode Lasers By R. Diehl (Ed.) 2ooo, 26o Figs. XIV, 416 pages 79 Frequency Measurement and Control Advanced Techniques and Future Trends By A. N. Luiten (Ed.) 2OOl, 169 Figs. XIV, 394 pages 80 Carbon Nanotubes Synthesis, Structure, Properties, and Applications By M. S. Dresselhaus, G. Dresselhaus, Ph. Avouris (Eds.) 2OOl, 235 Figs. XVI, 448 pages 81 Near-Field Optics and Surface Plasmon Polaritons By S. Kawata (Ed.) 2OOl, 136 Figs. X, 21o pages 82 Optical Properties of Nanostructured Random Media By Vladimir M. Shalaev (Ed.) 2002, 185 Figs. XIV, 45o pages 83 Spin Dynamics in Confined Magnetic Structures I By B. Hillebrands and K. Ounadjela (Eds.) 2002, 166 Figs. XVI, 336 pages 84 Imaing of Complex Media with Acoustic and Seismic Waves By M. Fink, W. A. Kuperman, J.-P. Montagner, A. Tourin (Eds.) 20o2, 162 Figs. XII, 336 pages 85 Solid-Liquid Interfaces Macroscopic Phenomena - Microscopic Understanding By K. Wandelt and S. Thurgate (Eds.) 2oo3, 228 Figs. XVIII, 444 pages 86 Infrared Holography for Optical Communications Techniques, Materials, and Devices By E Boffi, D. Piccinin, M. C. Ubaldi (Eds.) 2003, 90 Figs. XII, 182 pages 87 Spin Dynamics in Confined Magnetic Structures II By B, Hillebrands and K. Ounadjela (Eds.) 2003, 179 Figs. XVI, 321 pages 88 Optical Nanotechnologies The Manipulation of Surface and Local Plasmons By J, Tominaga and D. E Tsai (Eds.) 2003, 168 Figs. XII, 212 pages 89 Solid-State Mid-Infrared Laser Sources By I. T. Sorokina and K. L. Vodopyanov (Eds.) 2oo3, 263 Figs. XVI, 557 pages

Printing: Saladruck Berlin Binding: Sttirtz AG, Wtirzburg

Solid-State Mid-Infrared Laser Sources Topics in Applied Physics

Photomechanics (Topics in Applied Physics)

Solid-State Mid-Infrared Laser Sources (Topics in Applied Physics 89)

Turbulent Reacting Flows (Topics in Applied Physics)

Mossbauer spectroscopy (Topics in applied physics)

Problems in Laser Physics

Problems in laser physics

Laser Physics

Laser Physics

Physics of Ferroelectrics: A Modern Perspective (Topics in Applied Physics)

Physics of Ferroelectrics: A Modern Perspective (Topics in Applied Physics)

Topics in Applied Probability

Topics in Atomic Physics

Topics in Atomic Physics

Topics In Atomic Physics

Topics in Polymer Physics

Theory of Defects in Semiconductors (Topics in Applied Physics)

Advanced topics in applied mathematics

Applied Stratigraphy (Topics in Geobiology)

Strong Field Laser Physics

Laser physics and applications

Applied Stratigraphy (Topics in Geobiology)

Integrated Optics, Second Edition (Topics in Applied Physics)

Infrared Holography for Optical Communications Topics in Applied Physics

Microscale and Nanoscale Heat Transfer (Topics in Applied Physics)

Integrated Optics (Topics in Applied Physics, Vol 7)

Dye Lasers, Third Edition (Topics in Applied Physics, Vol. 1)

Microscale and Nanoscale Heat Transfer (Topics in Applied Physics)

Thermal Nanosystems and Nanomaterials (Topics in Applied Physics, 118)

Silicon Photonics II: Components and Integration (Topics in Applied Physics)

Materials Science with Ion Beams (Topics in Applied Physics)

Solid-State Mid-Infrared Laser Sources Topics in Applied Physics

Photomechanics (Topics in Applied Physics)