Laser‐based Mid‐infrared Sources and Applications

Laser‐based Mid‐infrared Sources and Applications

Konstantin L. Vodopyanov

University of Central FloridaOrlando, FL, USA

This edition first published 2020© 2020 John Wiley & Sons, Inc.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Konstantin L. Vodopyanov to be identified as the author of this work has been asserted in accordance with law.

Registered OfficeJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA

Editorial Office111 River Street, Hoboken, NJ 07030, USA

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of WarrantyWhile the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging-in-Publication data applied for

ISBN: 9781118301814

Cover Image: © merrymoonmary/Getty ImagesCover Design: Wiley

Set in 10/12pt Warnock by SPi Global, Pondicherry, India

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

http://www.wiley.com/go/permissionshttp://www.wiley.com

In memory of Feliciana Ignatievna Vergunas

vii

About the Author xi Preface xiii

1 Mid‐IR Spectral Range 11.1 Definition of the Mid‐IR 11.2 The World’s Second Laser 31.3 Internal Vibrations of Molecules 4 References 5

2 Solid-state Crystalline Mid‐IR Lasers 72.1 Rare-Earth-based Tm3+, Ho3+, and Er3+ Lasers 7

2.1.1 Tm3+ Lasers 72.1.2 Ho3+ Lasers 102.1.3 Er3+ Lasers 13

2.2 Transition Metal Cr2+ and Fe2+ Lasers 182.2.1 Spectroscopic Properties of Cr2+ and Fe2+ 182.2.2 Lasers Based on Chalcogenide Crystals Doped with Cr2+ 21

2.2.2.1 Broadly Tunable Cr2+ Lasers 212.2.2.2 High-power Continuous-wave Cr2+ Lasers 232.2.2.3 High-power Cr2+ CW Laser Systems Operating at

2.94 μm 232.2.2.4 Gain-switched High-power Cr2+ Lasers 242.2.2.5 Microchip Cr2+ Lasers 252.2.2.6 Waveguide and Thin-disk Cr:ZnSe Lasers 262.2.2.7 Mode-locked Cr:ZnS/Cr:ZnSe Lasers 27

2.2.3 Lasers Based on Chalcogenide Crystals Doped with Fe2+ 302.2.3.1 Free-running Pulsed Fe:ZnSe/ZnS Lasers 302.2.3.2 Gain-switched Regime of Fe2+ Lasers at Room

Temperature 322.2.3.3 Continuous-wave Fe2+ Lasers 332.2.3.4 Tunable Fe2+ Lasers at Room Temperature 352.2.3.5 Ultrafast Amplifier in the 3.8–4.8 μm Range 35

2.3 Summary 35 References 36

Table of Contents

Table of Contentsviii

3 Fiber Mid‐IR Lasers 433.1 Introduction 433.2 Continuous-wave Mid‐IR Fiber Lasers 44

3.2.1 Tm-based Fiber Lasers 443.2.2 Ho-based Fiber Lasers 473.2.3 Er-based Fiber Lasers 493.2.4 Dy-based Fiber Lasers 523.2.5 Raman Fiber Lasers 52

3.3 Q-switched Mid‐IR Fiber Lasers 543.4 Mode-locked Mid‐IR Fiber Lasers 563.5 Summary 60 References 61

4 Semiconductor Lasers 654.1 Heterojunction Mid‐IR Lasers 65

4.1.1 GaSb-based Diode Lasers 664.1.2 Distributed Feedback GaSb-based Lasers 70

4.2 Quantum Cascade Lasers 734.2.1 High Power and High Efficiency QCLs 764.2.2 Single-mode Distributed Feedback (DFB) QCLs 794.2.3 Broadly Tunable QCLs with an External Cavity 824.2.4 Short-wavelength (

Table of Contents ix

5.3 Pulsed Regime 1305.3.1 Pulsed DFG 1305.3.2 Pulsed OPOs 133

5.3.2.1 Broadly Tunable Pulsed OPOs 1335.3.2.2 Narrow-linewidth Pulsed OPOs 1435.3.2.3 High Average Power OPOs 1475.3.2.4 High Pulse Energy OPOs 1505.3.2.5 Waveguide OPOs 152

5.4 Regime of Ultrashort (ps and fs) Pulses 1535.4.1 Ultrafast DFG 1535.4.2 Intra-pulse DFG (Optical Rectification) 1575.4.3 Ultrafast OPOs 161

5.4.3.1 Picosecond Mode 1615.4.3.2 Femtosecond Mode 163

5.4.4 Ultrafast OPGs 1655.4.5 Ultrafast OPAs 167

5.5 Raman Frequency Converters 1685.5.1 Crystalline Raman Converters 1695.5.2 Fiber Raman Converters 1695.5.3 Silicon Raman Converters 1705.5.4 Diamond Raman Converters 1715.5.5 Other Raman Converters 172

5.6 Summary 174 References 174

6 Supercontinuum and Frequency Comb Sources 1896.1 Supercontinuum Sources 189

6.1.1 SC from Lead-silicate Glass Fibers 1916.1.2 SC from Tellurite Glass Fibers 1926.1.3 SC from ZBLAN Fibers 1946.1.4 SC from Chalcogenide Glass Fibers 1966.1.5 SC from Waveguides 2036.1.6 SC from Bulk Crystals 2076.1.7 Other SC Sources 212

6.2 Frequency Comb Sources 2136.2.1 Direct Comb Sources from Mode-locked Lasers 2146.2.2 Combs Produced by Spectral Broadening in NL Fibers

and Waveguides 2156.2.3 Combs Produced by Difference Frequency

Generation 2176.2.4 OPO-based Combs 2206.2.5 Combs Based on Optical Subharmonic Generation 2266.2.6 Microresonator-based Kerr Combs 2296.2.7 Combs from Quantum Cascade Lasers 234

Table of Contentsx

6.2.8 Combs from Interband Cascade Lasers 2356.3 Summary 235 References 236

7 Mid‐IR Applications 2477.1 Spectroscopic Sensing and Imaging 247

7.1.1 QCLs for Spectroscopy and Trace-gas Analysis 2487.1.2 Spectroscopy with ICLs 2527.1.3 Spectroscopy with DFG and OPO Sources 2527.1.4 Broadband Spectroscopy with Frequency Combs 2537.1.5 Hyperspectral Imaging 255

7.2 Medical Applications 2587.2.1 Laser Tissue Interactions 258

7.2.1.1 Holmium and Thulium Surgical Lasers 2587.2.1.2 Er:YAG Lasers (λ = 2.9 μm) 2597.2.1.3 Importance of the Spectral Band of 6–7 μm 260

7.2.2 Medical Breath Analysis 2617.2.2.1 Ethane (C2H6) 2627.2.2.2 NO 2627.2.2.3 NH3 2637.2.2.4 CO 2637.2.2.5 OCS 2637.2.2.6 Optical Frequency Comb Spectroscopy for Breath

Analysis 2647.3 Nano‐IR Imaging and Chemical Mapping 2657.4 Plasmonics in the Mid‐IR 2677.5 Infrared Countermeasures 2697.6 Extreme Nonlinear Optics and Attosecond Science 2707.7 Other Applications 273

7.7.1 Laser Wake-field Accelerators 2737.7.2 Laser Acceleration in Dielectric Structures 2747.7.3 Free-space Communications 2747.7.4 Organic Material Processing 275

References 276

Index 287

xi

Konstantin L. Vodopyanov is an Endowed Chair and Professor of Optics and Physics at CREOL, the College of Optics and Photonics at the University of Central Florida. He is a Fellow of The Optical Society of America (OSA), the International Optical Engineering Society (SPIE), the  American Physical Society (APS), and the UK Institute of Physics (IOP). K.L. Vodopyanov is a world expert in mid-IR lasers, laser–matter interactions, nonlinear optics, and laser spectroscopy. He has pioneered several new laser sources, optical para-metric devices, and methods for mid-IR and THz wave generation, including generation of ultrabroadband frequency combs; he is the author of more than 400 technical publications on the subject and a co-editor of the book Solid-State Mid-Infrared Laser Sources (Springer, 2003). His research interests include the use of nonlinear effects for mid-IR and THz generation, spectro-scopic applications of frequency combs, nano-IR spectroscopy, and biomedical applications of lasers.

About the Author

xiii

The field of mid‐infrared (mid‐IR) photonics is rapidly expanding driven by a growing number of applications in fundamental science, technology, defense, medicine, biology, environmental monitoring, among others. The last 25 years have seen remarkable advances in the field of mid‐IR lasers starting with the quantum cascade laser pioneered by Lucent Technologies in 1994 and followed by a growing number of diverse innovative approaches to coherent light generation in this spectral range. Because the last comprehensive book on the subject, Solid‐State Mid‐Infrared Laser Sources, which I coedited with I. Sorokina, was published by Springer in 2003, it was both my and the pub-lisher’s understanding that this material needed to be significantly updated to include the impressive number of new techniques and applications.

The main goal of this book is to introduce the reader to the state‐of‐the‐art technologies used to generate coherent mid‐IR light, and to discuss their most important applications. The book assembles an array of methods developed by several scientific communities, which include solid‐state physics, semiconduc-tor physics, materials science, crystal growth, nonlinear optics, and nanofabri-cation, in their search to create an efficient and inexpensive solid‐state mid‐IR laser source.

The book loosely defines mid‐IR range as 2–20 μm. It examines a variety of state‐of‐the‐art approaches from diverse areas of photonics: solid‐state lasers based on rare‐earth and transition metals; fiber lasers; semiconductor lasers including intra‐ and intersubband cascade lasers; nonlinear‐optical frequency converters including difference frequency generators, optical parametric oscillators and amplifiers, and Raman converters. It also dis-cusses several emerging technologies such as “white light” and frequency combs generation in microresonators, waveguides, and microstructured fib-ers. In the final chapter, the book provides an overview of the most signifi-cant applications of mid‐IR, such as chemical sensing and imaging including nano-imaging, medical and defense applications, plasmonics, extreme non-linear optics, attosecond science, and particle acceleration. Such mature fields as free‐electron lasers, CO2 and CO gas lasers, synchrotron radiation,

Preface

Prefacexiv

and cryogenic lead‐salt semiconductor lasers are outside of the scope of this book, since the reader can find published material on these subjects.

The book is based on the short courses that I taught at major laser confer-ences, including the Conference on Lasers and Electro‐Optics (CLEO) and SPIE Photonics West. Each chapter begins with a self‐contained description of the underlying principle for a given method, and gradually brings the reader to the discussion of the latest achievements. I made every effort to make the narrative comprehensible to a broader community. However, it is assumed that the reader is familiar with basic concepts of laser physics, such as population inversion, Q‐switching, and mode‐locking, as well as of nonlinear optics, such as frequency mixing and nonlinear refraction.

The book should be useful to students, academics, researchers, and engi-neers, and to those who would like to learn about state‐of‐the‐art and major trends in the development of mid‐IR laser sources, and their current and upcoming applications.

I would like to thank Dr. Sergey Vasilyev, Prof. Sergey Mirov, Prof. Ken Schepler, Prof. Stuart Jackson, Prof. Gregory Belenky, Prof. Leon Shterengas, Dr. Jerry Meyer, Dr. Igor Vurhaftman, Prof. Arkadiy Lyakh, and Prof. Jerome Faist for reading the book chapters and making valuable suggestions. Finally, my wife Mila has earned my endless gratitude for her optimism, continuous support, helpful edits of the text, and for her bearing my spending long late‐night and weekend hours on writing this book.

January 2020 Konstantin L. VodopyanovOrlando, Florida

Laser-based Mid-infrared Sources and Applications, First Edition. Konstantin L. Vodopyanov. © 2020 John Wiley & Sons, Inc. Published 2020 by John Wiley & Sons, Inc.

1

1

1.1 Definition of the Mid‐IR

Infrared radiation was unknown before the year 1800 when Friedrich Wilhelm Herschel  –  a German‐born musician, who moved to England to work as a music band conductor, but later became obsessed with astronomy and eventu-ally landed the position of the King’s Astronomer − discovered infrared radia-tion. He made this finding while exploring sunlight, dispersed into its colors by a glass prism, with the aid of a liquid thermometer with a blackened bulb to absorb radiation (a prototype of a modern microbolometer). His experimenta-tion led to the conclusion that there must be an invisible form of light beyond the visible spectrum [1].

Further experiments showed that this invisible radiation is electromagnetic radiation with a lower frequency than the red in the visible spectrum. Modern science further divides the infrared spectral region into near‐infrared, mid‐infrared, and far‐infrared.

According to the Encyclopedia Britannica, the “middle infrared” (mid‐infra-red or mid‐IR) region of the electromagnetic spectrum covers, in wavelength, the portion between 2.5 and 50 μm (6–120 THz in frequency or 200–4000 cm−1 in wavenumbers).1 (The wavenumber is the inverse of the vacuum wavelength, λ, expressed in cm−1; it is also equal to the optical frequency divided by the speed of light, ν/c.)

However, the definitions of the “mid‐IR” vary substantially in the technical literature, depending on a field‐specific community. For example, the detec-tor‐based community subdivides the IR into four spectral bands, based on transmission windows of the atmosphere,2 as can be seen in Figure  1.1: shortwave infrared (SWIR), 1–3 μm; mid‐wave infrared (MWIR), 3–5 μm;

1 https://www.britannica.com/science/infrared-radiation2 https://en.wikipedia.org/wiki/Infrared_vision

Mid‐IR Spectral Range

1 Mid‐IR Spectral Range2

longwave infrared (LWIR), 8–14 μm; and very‐long wave infrared (VLWIR), 14–30 μm.

Also, it is not uncommon in the current literature to refer to mid‐IR as “multi‐terahertz range,” especially when the authors generate few‐cycle mid‐IR transients combined with electro‐optic methods of their detection, which is typical for terahertz science.

This book loosely defines the mid‐IR range as 2–20 μm. This definition allows, on the short‐wavelength side, to encompass a few categories of solid‐state and fiber lasers, as well as certain types of microresonator‐, nonlinear fiber‐, and waveguide‐based sources. On the long‐wavelength side, 20 μm is a suitable practical limit set by the atmospheric transparency.

Heat energy is often transferred in the form of infrared radiation, which is given off from an object as a result of atomic and molecular motion. The mid‐IR region overlaps with the spectral range of heat (blackbody) radiation at temperatures close to room temperature. Based on the Planck’s law, the peak of the infrared radiation (in terms of power per unit wavelength) emit-ted by a human body at 310 K is at λ ≈ 9.35 μm. Overall, our body emits 52 mW of mid‐IR radiation per square centimeter; that radiation can be easily detected by a thermal microbolometer‐based camera, as shown in Figure 1.2.

However, this book is about coherent laser sources, and the difference between the diffuse light of a heated body and a monochromatic laser‐like beam is that the latter has a well‐defined frequency and phase.

50

1H O H O

CO CO HDOH O H O

0.8

0.6

Tran

smis

sion

0.4

0.2

0500 1000 1500 2000

Wavelength (μm)

Wavenumber (cm–1)

2500 3000 3500 4000 4500 5000

20 10 5

Atmosphere,L = 100 m

3 2

Figure 1.1 Transmission spectrum of a 100‐m path in the “standard” atmosphere (spectral resolution 4 cm−1). The plot uses data from the HITRAN database (US model, mean latitude, summer, and zero elevation) [2]. The labels indicate the molecules that are responsible for transmission dips in the corresponding spectral regions.

1.2 The World’s Second Laser 3

1.2 The World’s Second Laser

Interestingly, the world’s second laser – after the Maiman’s ruby laser – was a mid‐IR solid‐state laser based on trivalent uranium‐doped calcium fluoride (U3+:CaF2) [3]. It was operating at a wavelength λ = 2.49 μm and was developed by Peter Sorokin and Mirek Stevenson at the IBM research labs, in the same year as the Maiman’s ruby laser, 1960.

The laser was pumped by a pulsed flashlamp and was cooled by liquid helium. The energy‐level diagram of U3+:CaF2 is shown in Figure  1.3. Broadband pumping in the visible part of the spectrum causes transitions to excited U3+ bands. These pumping transitions are followed by rapid, nonra-diative transitions to the two metastable upper laser levels. The thick arrow shows the 2.49‐μm laser transition observed by the authors. The laser oscil-lation takes place in a transition from a metastable state to a level that is approximately 515 cm−1 above the ground state. At liquid helium tempera-tures, this state is depopulated by at least a factor of 1010 relative to the ground state. Hence, this was the first demonstration of a four‐level solid‐state laser.

Figure 1.2 This is what the author’s laboratory looks like in the mid‐IR at 8–12 μm.

1 Mid‐IR Spectral Range4

Furthermore, the 1960 work by Sorokin and Stevenson coined a few signifi-cant keywords used in this book: trivalent metal cation, and the 4‐level system.

1.3 Internal Vibrations of Molecules

Molecules typically have a characteristic absorption spectrum in the mid‐IR, which is often used for identifying their structure. This results from the fact that mid‐IR frequencies coincide with the strongest main‐tone vibrational (strictly speaking, rotational–vibrational) frequencies of most of the molecules. In the gas phase, molecules possess dozens of distinct, sharp, and strong absorption features (with the exception of symmetric diatomic molecules like nitrogen, N2, whose vibrations are not infrared active). This makes the mid‐IR range especially important for chemical sensing, molecular

U3+:CaF2

≈ 120 cm–1

≈ 515 cm–1

2.49 μm(4016 cm–1)

2.15 μm(4651 cm–1)

2.42 μm(4132 cm–1)

2.21 μm(4525 cm–1)

Figure 1.3 Energy‐level diagram of trivalent uranium in calcium fluoride [3]. Broadband pumping light applied in the blue and green visible spectrum causes transitions to excited bands. These pumping transitions are followed by rapid, nonradiative transitions to the two metastable upper laser levels. The thick arrow shows the 2.49 μm laser transition observed by the authors. Source: reproduced from figure 1 of [3], with permission of APS.

References 5

spectroscopy, and molecular fingerprinting. Figure  1.4 shows a rotational–vibrational band of the CO molecule, located between 4.5 and 5 μm in wave-length. The band is represented by a regular sequence of sharp and extremely strong absorption peaks. For example, a 1‐mm path of pure CO gas at 1 atm pressure would absorb >99.5% of the incoming mid‐IR light, given that the light is tuned to one of the resonances. In theory, these resonances can even serve as a reference for high‐precision molecular clocks for time and frequency metrology.

Characteristic vibrational transitions in the mid‐IR are also present in the solid and liquid phases of matter, and also in 2D materials with exotic proper-ties, such as graphene [4].

References

1 Herschel, W. (1800). Experiments on the refrangibility of the invisible rays of the sun. Philos. Trans. R. Soc. Lond. 90: 284.

2 Gordon, I.E., Rothman, L.S., Hill, C., Kochanov, R.V., Tan, Y., Bernath, P.F., Birk, M., Boudon, V., Campargue, A., Chance, K.V., Drouin, B.J., Flaud, J.-M.,

Wavelength (μm)

CO

Abs

orpt

ion

stre

ngth

per

mol

ecul

e (c

m2 /

cm)

5 4.81E–18

1E–19

1E–20

1E–21

1E–22

1E–23

1E–241950 2000 2050

Wavenumber (cm–1)

2100 2150 2200 2250

4.6 4.4

Figure 1.4 Rotational–vibrational absorption band of the CO molecule at 4.5–5 μm. The plot uses data from the HITRAN database [2].

1 Mid‐IR Spectral Range6

Gamache, R.R., Hodges, J.T., Jacquemart, D., Perevalov, V.I., Perrin, A., Shine, K.P., Smith, M.-A.H., Tennyson, J., Toon, G.C., Tran, H., Tyuterev, V.G., Barbe, A., Császár, A.G., Devi, V.M., Furtenbacher, T., Harrison, J.J., Hartmann, J.-M., Jolly, A., Johnson, T.J., Karman, T., Kleiner, I., Kyuberis, A.A., Loos, J., Lyulin, O.M., Massie, S.T., Mikhailenko, S.N., Moazzen-Ahmadi, N., Müller, H.S.P., Naumenko, O.V., Nikitin, A.V., Polyansky, O.L., Rey, M., Rotger, M., Sharpe, S.W., Sung, K., Starikova, E., Tashkun, S.A., VanderAuwera, J., Wagner, G., Wilzewski, J., Wcisło, P., Yu, S., and Zak, E.J. (2017). The HITRAN 2016 molecular spectroscopic database. J. Quant. Spectrosc. Radiat. Transf. 203: 3.

3 Sorokin, P.P. and Stevenson, M.J. (1960). Stimulated infrared emission from trivalent uranium. Phys. Rev. Lett. 5: 557.

4 Mak, K.F., Ju, L., Wang, F., and Heinz, T.F. (2012). Optical spectroscopy of graphene: from the far infrared to the ultraviolet. Solid State Commun. 152: 1341.

7

Laser-based Mid-infrared Sources and Applications, First Edition. Konstantin L. Vodopyanov. © 2020 John Wiley & Sons, Inc. Published 2020 by John Wiley & Sons, Inc.

2

Crystalline mid‐IR lasers are direct sources of coherent light in the sense that they require a minimal number of energy conversion steps. When combined with laser diode pumping, these lasers are efficient, simple, and compact. The gain medium of a crystalline laser is a host crystal doped with active ions. These active ions (also referred to as impurity ions) doped into a crystalline matrix acquire, due to energy‐level splitting, characteristic set of energy levels, not present in free ions. For rare earths, the primary cause of energy‐level split‑ting is the interaction of electron spins of the dopant ion with the orbital angu‑lar momentum of electrons (spin–orbit interactions), while in transition metal ions, it is mostly due to the interaction of the optically active electron with the crystalline electric field of the host (the Stark effect).

The most common active media for mid‐IR crystalline lasers are based on triply ionized rare‐earth thulium (Tm), holmium (Ho), and erbium (Er) ions in yttrium aluminum garnet (Y3Al5O12 or YAG), yttrium lithium fluoride (LiYF4 or YLF), yttrium‑scandium‑gallium garnet (Y3Sc2Ga3O12 or YSGG), or other crystalline hosts. Alternatively, transition‐metal‐doped (Cr2+, Fe2+) II–VI zinc chalcogenide crystals (ZnSe, ZnS) or other chalcogenides (CdSe, CdS, ZnTe, and CdMnTe) can serve as active media for mid‐IR lasers with an extremely broad gain bandwidth.

2.1 Rare-Earth-based Tm3+, Ho3+, and Er3+ Lasers

2.1.1 Tm3+ Lasers

The energy‐level diagram for the trivalent thulium ion is shown in Figure 2.1a. Tm‐doped crystalline lasers can provide tunable operation in two spectral regions: 1.8–2.2 μm using the 3F4−3H6 transition and 2.2–2.4 μm using the 3H4−3H5 transition.

Solid-state Crystalline Mid‐IR Lasers

2 Solid-state Crystalline Mid‐IR Lasers8

At first glance, a laser that uses 3F4−3H6 transition appears as a three‐level system. In such a system the lower laser level is the ground state. Nevertheless, it should be mentioned that both 3F4 and 3H6 levels consist of manifolds of energy levels split due to the Stark effect in the electric field of the crystalline lattice, so that the lower laser level is not necessarily the lowest energy state. Due to the fast energy relaxation between these sublevels, the system becomes virtually a four‐level system. An additional effect is that the presence of par‑tially overlapping manifolds of Stark levels within the upper and lower laser state broadens the bandwidth of fluorescence, resulting in broad emission linewidths. Furthermore, thulium lasers (e.g. Tm:YAG and Tm:YSGG) are characterized by large phonon broadening. The phonons (vibrations of the ions of the host crystal) modulate the crystal field at the site of the “lasing” dopants (thulium ions in this case), which in turn broadens the energy levels [1]. Both of these effects allow laser tunability over several hundreds of nanom‑eters. The room‐temperature (RT) fluorescence spectrum from the 3F4 state to the 3H6 ground state of Tm3+ in YAG is shown in Figure 2.1b.

YAG crystal is one of the most commonly used host materials for thulium because of its unique thermal–mechanical and optical properties. Typically, Tm3+‐doped solid‐state lasers are pumped (3H6 → 3H4) by commercially avail‑able high‐power AlGaAs diode bars at ∼800 nm. There is a “two‐for‐one” cross‐relaxation process that can lead to pumping quantum efficiencies approaching a factor of two. (The pumping quantum efficiency indicates how many laser photons are emitted per one absorbed pump photon.) The essence of this effect (Figure 2.2) is that because of the fortuitous proximity (resonance) of the energy spacing 3H4−3F4 and 3F4−3H6, the 3F4 upper laser level is populated through the cross‐relaxation process 3H4 + 3H6  →  3F4 + 3F4 [1]. The effectiveness of this

3H4

(a) Tm3+

1.5 ms

6.8 ms

3H5

Laser2.3 μm

Laser2 μm

3F4

3H6

Pump790 nm

(b)1.0

0.8

0.6

0.4

Inte

nsity

(ar

b. u

nits

)

0.2

0.01.6

Wavelength (μm)1.8 2.0 2.2

Figure 2.1 (a) Energy‐level diagram for the trivalent thulium. Wavy arrows indicate nonradiative phonon‐assisted decay. Laser upper‐state lifetimes are also indicated. (b) Spectrum of fluorescence from the 3F4 level of Tm3+ in YAG. Source: reproduced from figure 1 of [1], with permission of OSA, The Optical Society.

92.1 Rare‐Earth‐based, Tm3+, Ho3+, and Er3+ Lasers

two‐body cross‐relaxation process increases with Tm3+ doping concentrations; it becomes significant, typically at >3 at.% concentration. For example, it was shown that at Tm3+ concentration of 12% for Tm:YAG and Tm:YSGG, the cross‐relaxation totally dominates the decay of the 3H4 state. As a result, a slope efficiency of as high as 59% has been demonstrated in Tm:YAG, considerably larger than the 39% maximum expected from the quantum defect alone [1]. (The quantum defect is defined as the ratio of the energy of the lasing photon to that of the pump.)

With a 785‐nm laser pumping, broadly tunable continuous wave (CW) laser emission over the ranges 1.87–2.16 μm in Tm:YAG and 1.85–2.14 μm in Tm:YSGG has been reported [1]. Similarly, tunable Tm laser emission using a different host crystal – Tm:YALO (Tm:YAlO3) – was observed over the range 1.93–2.0 μm [2].

The 3F4−3H6 transition in Tm is especially attractive for high‐power applica‑tions because of the ability to use highly efficient AlGaAs diodes or diode stacks operating around 800 nm as a pump source. A compact diode‐pumped Tm:YAG laser capable of generating 115 W of CW power at 2.01 μm has been demon‑strated at the 805‐nm pumping power of 360 W [3]. Another high‐power CW Tm:YAG laser used a linear laser cavity with three laser rods, each side‐pumped by arrays of laser diode bars with central wavelength of 785 nm, arranged in fivefold symmetry around each laser crystal (Figure 2.3). The laser was water‐cooled at 8 °C and yielded a maximum output CW power of 267 W at 2.07 μm, with the total laser‐diode pump power of 1.3 kW. The corresponding optical‐to‐optical conversion efficiency was 20.7%, with slope efficiency of 29.8% [4].

The Q‐switching performance of Tm:YAG near 2 μm is facilitated by a long fluorescence lifetime of 11 ms, which leads to a high energy storage capability [5].

Pum

p

Lase

r

Lase

r

Ion 1 Ion 2

3H6

3F4

3H5

3H4Tm3+

Upper laser level

Figure 2.2 Resonant pumping diagram for the 2‐μm Tm3+ laser. The “two‐for‐one” cross‐relaxation process leads to the efficient transfer of all absorbed pump energy to the excitation of Tm ions to the 3F4 level.

2 Solid-state Crystalline Mid‐IR Lasers10

For example, Q‐switched pulses at 2.016 μm with 20.4 mJ energy and 69 ns dura‑tion were demonstrated at 500 Hz pulse repetition rate; a Tm:YAG ceramic slab laser was end‐pumped by a diode laser (the absorbed pump power 53 W) [6]. In general, because of the low gain due to low stimulated emission cross section of Tm ions, Q‐switched Tm lasers operate by necessity at high intracavity energy density (fluence), close to the material damage threshold [5]. The reason is that in order to achieve the laser threshold in this low‐gain medium, one needs high population inversion; this in turn results in a high amount of energy stored in the medium that is eventually released as an energetic Q‐switched laser pulse that may cause material damage.

2.1.2 Ho3+ Lasers

Figure 2.4a represents the energy‐level diagram for trivalent holmium. With Ho3+ as a laser ion, one can obtain oscillation using the 5I7−5I8 transition in the 1.95–2.15 μm range (the corresponding emission spectrum for this transition is shown in Figure 2.4b), as well as 5I6−5I7 transition at 2.85–3.05 μm, 5I5−5I6 transition at 3.94 μm, and 5S2−5F5 transition near 3.2 μm. Holmium lasers are more favorable (as compared to thulium lasers) to operation in the Q‐switched mode due to high stimulated emission cross section (high gain); also, they have long (~10 ms, similar to thulium) fluorescence lifetime for the upper laser energy manifold 5I7.

Since there are no convenient schemes for direct diode pumping of Ho3+ lasers (at least with available high‐efficiency AlGaAs or InGaAs laser diodes), holmium laser crystals are often codoped (sensitized) with other ions. For example, crystals doped with a combination of Tm3+ and Ho3+ ions have proven to be efficient sources for diode‐pumped laser action in the 2‐μm region using

Lasermodule 1

Lasermodule 2

Lasermodule 3M1 M2

Figure 2.3 High‐power (267 W) Tm:YAG laser system at 2.07 μm based on a linear laser cavity containing three laser rods. The laser cavity was formed by two plane mirrors: M1 with reflectivity R > 99.5%, and the outcoupler mirror M2 with transmission T = 5% around 2 μm. Each rod (Tm concentration of 3.5 at.%, 4 mm in diameter, and 69 mm in length) was side‐pumped by an array of laser diode bars at 785 nm. Undoped YAG end caps were bonded to end faces of the rods to reduce thermal effects as well as reabsorption losses in the unpumped regions. Source: reproduced from figure 1 of [4], with permission of OSA, The Optical Society.

11

the following mechanism: thulium ions provide absorption of the AlGaAs diode‐laser pump around 780–790 nm and permit efficient population of the 3F4 state of Tm3+. This excitation is transferred to the 5I7 energy level of Ho3+ through the dipole–dipole interaction between nearby Tm3+ and Ho3+ ions, and laser action takes place on the Ho3+ 5I7−5I8 transition at 2.1 μm [8].

Although Tm and Ho lasers operate in the similar wavelength range near 2 μm, there are at least two reasons why Ho lasers are preferable, especially in the Q‐switched regime: Ho3+ has higher stimulated‐emission cross section (σ  =  1.2 × 10−20 cm2 at λ  =  2.09 μm in Ho:YAG) compared with that in Tm3+ (σ = 1.5 × 10−21 cm2 at λ = 2.01 μm in Tm:YAG) [7, 9]. This enables Q‐switching at high repetition rates and makes compact low‐threshold devices feasible. Besides, Ho lasers operate at slightly longer wavelengths that make them more suitable for coherent Doppler light detection and ranging (LIDAR) applica‑tions because of much smaller absorption in the atmosphere at λ > 2 μm (see Figure 1.1).

Ho lasers can be directly pumped by Tm‐doped lasers (including both solid‐state and fiber‐based lasers). For example, while the absorption spectrum of Tm:YLF falls well within the emission spectrum of commercially available laser diodes emitting in the 792–793 nm range, its emission spectrum at ~1.91 μm aligns well with the absorption spectrum of Ho:YAG [9]. An efficient diode‐pumped cascaded Tm−Ho system (both CW and Q‐switched) was reported in [10]. Using a high‐brightness 1.91‐μm output from a diode‐pumped Tm:YLF laser, the authors resonantly populated the 5I7 manifold in Ho:YAG. An end‐pumped geometry was adopted for both the Tm (1.91 μm) and Ho (2.09 μm) lasers (Figure 2.5). The Tm laser used two laser crystals to enable the use of four

5S2

5F5

5I4

5I5

5I6

5I7

5I8

Laser 3.2μm

(a) (b)

Laser 3.94μm

1.6

1.2

0.8

Emission spectrumτrad =14.3 ms

0.4

01800 1900 2000

Wavelength (nm)

Em

issi

on c

ross

sec

tion

(10–

20 c

m2 )

2100 2200

Laser 2.9μm

Laser2.1μm

Pump1.9μm

Ho3+

Figure 2.4 (a) Energy‐level diagram for the trivalent holmium. Red arrows indicate laser transitions. (b) Emission spectrum of Ho3+ (the 5I7−5I8 transition) in LiYF4 host. Source: reproduced from figure 3c of [7], with permission of IEEE.

2.1 Rare‐Earth‐based, Tm3+, Ho3+, and Er3+ Lasers

2 Solid-state Crystalline Mid‐IR Lasers12

fiber‐coupled laser diodes as the pump and to distribute the thermal load. With this design, 36 W of Tm laser output was achieved with the optical‐to‐optical conversion efficiency (diode‐to‐Tm laser) of 32%. The secondary holmium laser was pumped by the Tm:YLF laser and used a 5‐mm‐diameter, 20‐mm‐long Ho:YAG crystal. With 36 W of Tm laser pump, 19 W of continuous‐wave output at 2.09 μm was demonstrated [10]. This corresponds to Tm:YLF to Ho:YAG optical‐to‐optical efficiency of 56% and gross laser‐diode‐to‐Ho:YAG conversion efficiency of 18%.

The Q‐switched mode of Ho:YAG in the above‐described experiment was achieved by inserting an acousto‐optic modulator in the Ho laser cavity. In the repetitively Q‐switched (9–50 kHz) configuration of Ho:YAG pumped by the CW Tm:YLF laser, the average power reached 16 W with the corresponding Tm:YLF to Ho:YAG optical‐to‐optical efficiency of 50% (overall diode‐to‐Ho:YAG conversion efficiency of 15%). At a Q‐switch frequency of 15 kHz, the beam propagation factor was M2 < 1.5, the pulse duration 35 ns, and the energy per pulse 1.1 mJ [10].

Another laser material, Ho‐doped YLiF4 (Ho:YLF) is an excellent gain medium for obtaining high‐energy Q‐switched 2‐μm pulses. Ho:YLF has a weak thermal lensing and is naturally birefringent, which allows one to main‑tain polarization. A setup that delivered single‐frequency pulses with up to 330‐mJ pulse energy (pulse width 350 ns, λ = 2.064 μm, repetition rate 50 Hz) was reported [11]. A double‐pass Ho:YLF slab amplifier was end‐pumped with a diode‐pumped Tm:YLF slab laser at 1.89 μm (a wavelength that matched one

Fiber-coupledlaser diodes

CL

Fiber-coupledlaser diodes

Ho:YAGlaser

Tm:YLFlaser

2.09 μm

HoOC

AO Q-SWHo:YAG

HoHR

1.9 μm

TmOC

TmHR

45 Di

Tm

:YLF

Tm

:YLF

Figure 2.5 Ho:YAG laser (λ = 2.09 μm) resonantly pumped by a Tm:YLF laser (λ = 1.91 μm) [10]. HR, high reflector; 45Di, 45° dichroic mirror; CL, coupling lens; AO Q‐SW, acousto‐optic Q‐switch (not used in the CW experiment); OC, output coupler mirror.

13

of the strong absorption peaks of Ho:YLF), and seeded with 50‐mJ single‐fre‑quency pulses. The seed laser in this case was a single‐frequency ring‐cavity Ho laser system, pumped with an 80‐W 1940‐nm Tm:fiber laser.

There have been a number of reports on fiber‐bulk hybrid lasers combining Q‐switched Ho:YAG and Ho:YLF laser system with CW Tm‐doped fiber laser (1908 and 1940 nm, respectively) as an optical pump. This approach enables the efficient (>50%) conversion of a low‑cost and reliable high‑power fiber laser’s output into high‑energy nanosecond pulses. The recent achievement on the fiber‐bulk hybrids including 1‐kHz, nanosecond Ho:YAG and Ho:YLF systems with 52 and 35 mJ pulse energy, respectively, are summarized in [12]. The Ho:YLF hybrids with pulse energy in excess of 100 mJ were also reported [13].

Codoping holmium with ytterbium as sensitizer offers the possibility of using well‐developed InGaAs (970 nm) diode lasers for pumping Ho lasers. In [14], the researchers used Yb (at 20 at.%) and Ho (1 at.%) codoped YSGG as the active material to achieve laser action in holmium near 3 μm. YSGG was chosen as the host material because of the lower (as compared to YAG) phonon energies and thus longer lifetime of the upper laser level (5I6) of the 3‐μm transition. First, the Yb3+ absorbs a pump photon at 970 nm; this excitation is then transferred to the 5I6 upper laser level of the Ho3+. Experiments were performed under quasi‐CW excitation. A high‐power pulsed laser diode array (500 W peak power) was used at a pulse length of 0.8 ms and repetition rate 15 Hz. A maximum pulse energy of 10.5 mJ was obtained at the laser wavelength 2.844 μm (peak power 13 W). In the CW regime, the output power was a few milliwatts; the reason for the low power is a typical self‐terminating behavior of the laser transition because of the long‐lived lower laser level 5I7 [14]. (Ideally, the lifetime of the lower laser level should be shorter than that of the upper level to avoid a “bottleneck” effect.)

More information on Ho lasers doped with multiple (Cr3+, Er3+, and Tm3+) dopants, cascaded lasing at 2.1 and 2.9 μm, and on flashlamp‐pumped Ho lasers can be found in [15]. Flashlamp‐pumped 2‐μm holmium lasers at low repetition rates are widely used in medicine. For example, commercially avail‑able Ho:YAG lasers with free‐running pulse energies exceeding 3 J are com‑monly exploited for renal stone removal [16]. The 2‐μm radiation has numerous other applications. Strong absorption at this wavelength by human tissue is attractive for laser surgery (see Chapter 7); low atmospheric absorption makes this system useful for range‐finding, remote sensing, and wind‐shear detection onboard an aircraft with coherent laser radar. Also, 2‐μm lasers are widely used for pumping optical parametric oscillators to reach longer mid‐IR wavelengths (see Chapter 5).

2.1.3 Er3+ Lasers

Considerable interest in the lasers operating in the 2.7–3 μm range, based on trivalent erbium (Er:YAG, Er:YSGG, and others), has been observed in the last  30 years. The reason is that these lasers are extremely appealing to

2.1 Rare‐Earth‐based, Tm3+, Ho3+, and Er3+ Lasers

2 Solid-state Crystalline Mid‐IR Lasers14

specialists from different branches of medicine (dentistry, laser therapy, sur‑gery, and microsurgery), due to the fact that the laser wavelength near 2.94 μm corresponds to the strongest absorption peak of water (absorption strength >104 cm−1), and water is present in abundance in soft and hard biological tis‑sues. Furthermore, 3‐μm lasers are suitable pump sources for mid‐IR nonlinear optical frequency downconversion (see Chapter 5).

A great majority of literature before 1994 is dedicated to flashlamp‐pumped Er:YAG (2.94 μm) and Er:YSGG (2.8 μm) lasers operating in the free‐running [15, 17], Q‐switched (see [18] and references therein), and actively mode‐locked [19] regimes. Q‐switched operation with pulse energies of 10–100 mJ in the fundamental TEM00 mode, and mode‐locked operation with pulse energies up to 4 mJ in 100‐ps pulse duration are accessible with these lasers. In the free‐running pulsed mode, flashlamp‐pumped erbium lasers can deliver several joules of pulse energy and are commercially available from a variety of vendors.

The Er3+ energy‐level diagram relevant to the 3‐μm laser transition is shown in Figure 2.6a. The laser transition is between the 4I11/2 (upper) and the 4I13/2 (lower) states (not to be confused with more common 1.5‐μm Er laser transi‑tion, which is between 4I13/2 and 4I15/2 energy levels). For the 3‐μm transition, the upper laser state 4I11/2 can be directly pumped at 970 nm, or at 800 nm (in the latter case the upper laser level is populated via a fast nonradiative transi‑tion from 4I9/2 to 4I11/2). The corresponding emission spectrum for the 3‐μm transition is shown in Figure 2.6b. The lifetime of the lower laser level 4I13/2 (6.4 ms in Er:YAG) is much longer than the lifetime of the upper laser level 4I11/2 (0.1 ms in Er:YAG), which would lead, under normal circumstances, to self‐termination of the laser transition. However, at high Er3+ concentrations

Pump970 nm

4I15/2

Laser 2.9 µm

(a) (b)

4I13/2

4I11/2

4I9/2Er3+

Pump800 nm

Wavelength (μm)

Flu

ores

cenc

e in

tens

ity

2.5 2.6 2.7 2.8 2.9 3.0

Figure 2.6 (a) Energy‐level diagram for the trivalent erbium. (b) Emission spectrum of Er3+ in GSGG (Gd3Sc2Ga3O12) host at room temperature. Source: reproduced from figure 2 of [20], with permission of OSA, The Optical Society.