High-Power and Femtosecond Lasers: Properties, Materials and Applications: Properties, Materials and Applications, Nova Science Publishers

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.
High-Power and Femtosecond Lasers: Properties, Materials and Applications : Properties, Materials and Applications, Nova Science Publishers,
Lasers and Electro-Optics Research and Technology Series
HIGH-POWER AND FEMTOSECOND

LASERS: PROPERTIES, MATERIALS AND
APPLICATIONS
No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or
by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no
expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of information
contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in
rendering
High-Power and Femtosecond legal,
Lasers: medicalMaterials
Properties, or any and
other professional
Applications services.
: Properties, Materials and Applications, Nova Science Publishers,
LASERS AND ELECTRO-OPTICS
RESEARCH AND TECHNOLOGY SERIES
High-Power and Femtosecond Lasers: Properties, Materials and Applications
Paul-Henri Barret and Michael Palmer (Editors)
2009. ISBN: 978-1-60741-009-6
Lasers and Electro-Optics Research and Technology Series
HIGH-POWER AND FEMTOSECOND

LASERS: PROPERTIES, MATERIALS AND
APPLICATIONS
PAUL-HENRI BARRET
AND
MICHAEL PALMER
EDITORS
Nova Science Publishers, Inc.

New York
Copyright © 2009 by Nova Science Publishers, Inc.
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical
photocopying, recording or otherwise without the written permission of the Publisher.
For permission to use material from this book please contact us:
Telephone 631-231-7269; Fax 631-231-8175
Web Site: http://www.novapublishers.com
NOTICE TO THE READER

The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or
implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of
information contained in this book. The Publisher shall not be liable for any special,
consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or
reliance upon, this material. Any parts of this book based on government reports are so indicated
and copyright is claimed for those parts to the extent applicable to compilations of such works.
Independent verification should be sought for any data, advice or recommendations contained in
this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage
to persons or property arising from any methods, products, instructions, ideas or otherwise
contained in this publication.
This publication is designed to provide accurate and authoritative information with regard to the
subject matter covered herein. It is sold with the clear understanding that the Publisher is not
engaged in rendering legal or any other professional services. If legal or any other expert
assistance is required, the services of a competent person should be sought. FROM A
DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE
AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS.
LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA
High power and femtosecond lasers : properties, materials, and applications / editors, Paul-Henri
Barret and Michael Palmer.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-62417-044-7 (eBook)
1. High power lasers--Industrial applications. 2. Femtoseconds lasers--Industrial applications. 3.
Laser pulses, Ultrashort. I. Barret, Paul-Henri. II. Palmer, Michael, 1962-
TA1695.5.H54 2009
621.36'6--dc22
2009002422
Published by Nova Science Publishers, Inc.

  New York
CONTENTS
Preface vii
Research and Review Studies xv
Chapter 1 Recent Progress of High Peak Power Solid State Lasers 1
Yung-Sheng Huang, Jung-Sheng Huang and Fang-Ling Chang
Chapter 2 Laser Annealing of Composite Materials with Metal Nanoparticles 27
Andrey L. Stepanov
Chapter 3 Single Crystal Photo-Elastic Modulators 71
F. Bammer
Chapter 4 Diffractive Microoptics for Technological IR-Lasers 125
V. S. Pavelyev, V. A. Soifer, V. I. Konov, V. V. Kononenko and
A. V. Volkov
Chapter 5 Micro- and Nanoscale Heat Transfer in Femtosecond Laser
Processing of Metals 159

Yuwen Zhang, D. Y. Tzou and J. K. Chen
Chapter 6 High Power Femtosecond Laser Machining of Metals in Ambient
Medium 207
S. R. Vatsya, Chengde Li and S. K. Nikumb
Chapter 7 Femtolasers-Mediated Multiphoton Excitation Imaging of Bulk
Ocular Tissues 231
Bao-Gui Wang and Karl-Jürgen Halbhuber
Chapter 8 Optical Properties of Rare Earth Ions Induced by Femtosecond
Laser 279
Lixin Yu and Zhongxin Liu
Chapter 9 Ytterbium Doped Materials for Femtosecond Lasers 315
Pierre-Olivier Petit, Philippe Goldner, Bruno Viana,
Justine Boudeile, Frédéric Druon, Dimitris N. Papadopoulos,
Marc Hanna and Patrick Georges
vi Contents
Chapter 10 Ultrafast Dynamics of Porphyrins in Higher Excited State 335

Dae Won Cho, Mamoru Fujitsuka and Tetsuro Majima
Chapter 11 Femtosecond Excited-State Ultrafast Dynamics of Complex
Molecular Systems: Semiclassical Dynamics Simulations 369
Guang-Jiu Zhao, Yu-Hui Li and Ke-Li Han
Chapter 12 High Intensity Laser-Matter Interactions: Measurements of Ion and
Neutral Emission and Applications 393
A. Daskalova and W. Husinsky
Chapter 13 Improved Thermal Model and its Application in High-Power Laser
Ablation of Target 415
Ranran Fang, Duanming Zhang, Hua Wei and Zhihua Li
Short Communications 439
Dynamics of Excitons Confined in Semiconductor Thin Films 441
Osamu Kojima and Toshiro Isu
High Power Long Pulse Width QCW Laser Diode Bars for Optical
Pumping of Yb-Er Glass Solid State Lasers 455
N. I. Katsavets, V. A. Buchenkov and A. L. Ter-Martirosyan
Index 463
PREFACE
There has been a remarkable upsurge in the importance of high power lasers in the past
decade. This book presents an overall survey of recent advances in high peak power solid
lasers and reviews recent results in the interaction of high power laser pulses with various
dielectric materials containing metal nanoparticles. This book also presents the basics as well
as the theory of a single crystal photo-elastic modulator and a consideration of modern
approaches to the synthesis of diffractive optical elements for technological IR-laser beams
focusing.
During ultrafast laser interaction with metal, the electrons and lattices are not in
equilibrium. This book presents various two-temperature models that can be used to describe
the nonequilibrium heat transfer as well as the results of techniques in high power
femtosecond laser machining of metals. This book presents an overview of the current state of
the art in the field of femtosecond technology with a special emphasis on the research of high-
intensity laser-matter interactions. An investigation of non-linear excitation imaging
technique including two-photon autofluorescence and second harmonic generated signal
imaging is presented to investigate the microstructures of whole-mount corneal, retinal, and
scleral tissues in their native environment. Finally, included is a review of recent

semiclassical excited-state dynamics simulation results of some complex molecular systems
and a report of the dynamics of confined excitons when several exciton states are excited.
Chapter 1 - There has been a remarkable upsurge in the importance of lasers, especially
in high power lasers, on optics in both pure science and in technology in recent years.
Besides, by considering the development of the high power lasers with their rapidly growing
list of applications, the obvious needs for an introduction to this newest field is the primary
reason for this paper. In this article, the authors present an overall survey of recent advances
in high peak power solid state lasers (SSLs). The contents include:
Reviewing how the Degnan’s thermal Boltzman factor affects the inversion population
which is resulted from the non-uniform temperature distribution in the laser crystal. The
Auger effect (or so- called energy transfer up-conversion (ETU) effect) in the laser crystal
could result in thermal loading which is a function of pulse repetition frequency (PRF). So the
theoretical calculation on the thermo-optic effect and thermal focusing length are discussed to
explain the experimental results of SSLs. An introduction to the spatial-time dependence laser
model of rate equations for the active (A-O switch) and passive (Cr 4+ :YAG saturable
absorber and semiconductor GaAs saturable absorber) Q-switched SSLs is presented.
viii Paul-Henri Barret and Michael Palmer
In the second part of this paper, the authors study the E-M transverse modes of laser
output, especially in the generation of high order transverse modes, and calculate the
threshold pump power. In order to obtain a high peak power and single mode laser pulse
output, one can use simultaneously the Q-switching and mode locking technique. In this
article, the authors talk about its newest development in the past few years.
In the third part of this article, the authors talk about the non-linear optical device which
utilizes the intense radiation field of the nonlinear optical response of an optical media (i.e.
KTP or KDP, QPM, and PPLN nonlinear crystals) to generate a new frequency emission
from UV to far Infrared band. Thus, the frequency conversion is a useful technique for
extending the utility of high power SSLs. The authors talk especially about the development
of pulsed optical parametric oscillation (OPO) in SSLs. The discussion materials are based on
some researcher’s recent works. Finally, the OPO tunable source may have impacts on the
development of high power SSLs and important applications in some areas such as gravity
detection, atmospheric detection, DNA detection, global remote sensing…etc.. But the
authors skip this special issue in this article.
Chapter 2 - Recent results on the interaction of high power laser pulses with various
dielectric materials containing metal nanoparticles are reviewed. Original results together
with new publications are observed. In general, the excimer laser pulse modification of silver
and copper nanoparticles synthesized by ion implantation in silicate glasses and sapphire are
considered. One of features of composite samples prepared by the low energy ion
implantation is the growth of metal particles with a wide size distribution in the thin depth
from the irradiated substrate surface. Pulsed laser irradiation makes it possible to modify such
composite layer, improving the uniformity in the size distribution of the nanoparticles.
Changes induced by pulsed laser exposure suggest there are both reductions in average size of
the metal nanoparticles, and some long-range dissolution of metal atoms in the matrix.
Experimental data on laser modification are explained by photofragmentation and melting of
the nanoparticles in the dielectric matrix. Combination of ion implantation and laser
annealing is promising technology for fabrication of novel composite optical materials.
Chapter 4 - Wide use of infrared (IR) lasers in technological applications (metal cutting
and hardening, laser evaporation, laser bending and etc.) have stimulated a research in the
field of synthesis of IR-range diffractive microoptical focusing elements, which could retain
their functionality under continuous irradiation with high power. Successive fulfillment of
this requirement is stipulated both by selection of an appropriate substrate material and by
development of technologies for creation a diffractive microrelief on a substrate as well as
numerical methods for microrelief optimization taking into account possible technological
constraints.
The present chapter contains a consideration of modern approaches to the synthesis of
diffractive optical elements (DOE) for technological IR-laser beams focusing.
Experimental results of produced elements investigation described also.
Different ways, from geometrical optics approximation to stochastic optimization, are
used for calculation of a diffractive microrelief of IR-range optical elements that focus an
illuminating beam into a pre-given 2-dimensional area. Near-IR DOE synthesis on sapphire
substrates by plasmochemical etching method is considered.
Several diamond focusing DOEs for operating at CO2-laser wavelength (λ=10.6 μm) are
produced and presented. The nanosecond UV-laser ablation and plasmochemical lithographic
etching technologies for DOE synthesis on CVD diamond plates are discussed.
Preface ix
Being a perspective optical material, CVD diamond has a high refractive index (n=2.4),
resulted in relative high losses by Fresnel reflection. The present chapter includes
consideration of a realization of antireflecting structures on a diamond film surface by UV-
laser ablation.
The advent of silica and silver-halide IR-range optical fibers for technological
applications has set tasks of optimization methods development and of forming diffractive
microreliefs on optical fiber surface for waveguide beam controlling. The present chapter
includes the results of experimental investigation of diffractive microrelief and antireflecting
structures that are realized on an end-face of IR-fiber.
Chapter 5 - Ultrafast laser material processing has received significant attention due to a
growing need for the fabrication of miniaturized devices at micro- and nanoscales. The
traditional phenomenological laws, such as Fourier’s law of heat conduction, are challenged
in the microscale regime and a hyperbolic or dual phase lag model should be employed.
equilibrium. Various two-temperature models that can be used to describe the nonequilibrium
heat transfer are presented. A semi-classical two-step heating model to investigate thermal
transport in metals caused by ultrashort laser heating is also presented.
The main difference between the semiclassical and the phenomenological two-
temperature models is that the former includes the effects of electron drifting, which could
result in significantly different electron and lattice temperature response from the latter for
higher-intensity and shorter-pulse laser heating.
Under higher laser fluence and/or short pulse, the lattice temperature can exceed the
melting point and melting takes place. The liquid phase will be resolidified when the lattice is
cooled by conducting heat away. Ultrafast melting and resolidification of the thin gold film
and microparticles were investigated. At even shorter pulse width, femtosecond laser heating
on metals produces a blasting force from hot electrons in the sub-picosecond domain, which
exerts on the metal lattices along with the non-equilibrium heat flow.
The author’s work that employs the parabolic two-step heating model to study the effect
of the hot-electron blast in multi-layered thin metal films is also presented.

Chapter 6 - High power ultra-short-pulse lasers offer significant advantages over their
long-pulse counterparts for material processing. Since the duration of the pulse is too short to
establish the temperature equilibrium between the electron and lattice subsystems, little melt
is produced during processing. Ablation is caused by direct removal of the material.
Consequently, the machined surface features conform closely to the intensity profile of the
laser beam. The original laser beam suffers significant distortions during propagation through
a medium such as air. Intensities enhanced further by the self-focusing of the beam are
sufficiently high to cause an optical breakdown of the media, generating plasma. Scattering
effects of plasma deform the laser beam profile. Distortion to the optical beam profile can be
reduced by conducting femtosecond laser machining in vacuum or the gaseous media less
immune to ionization, which restricts its application in the production environment. Removal
of material also generates a plasma plume with the additional scattering effects. For certain
energy range, the competing self-focusing and gas plasma plume supplemented with the
material plasma can combine to cause plasma filamentation and the distortion causing
scattering effects balance out yielding an adequately shaped beam profile. Filament of hot
plasma also acts as an energy source with similar properties. The energy at which
filamentation occurs can be determined by calculating the optical beam profile, which can be
x Paul-Henri Barret and Michael Palmer
controlled experimentally thereby yielding a cleaner fabricated surface. Due to significant

differences in the processes involved, this type of conformity is less pronounced in case of the
dielectrics compared to the metallic materials. This article presents the results on the
applications of such techniques in high power femtosecond laser machining of metals. Some
discussion on non-metallic materials is also included.
Chapter 7 - Currently, femtosecond lasers (femtolasers) are being extensively employed
in diverse research and application fields. Femtolasers-mediated multiphoton excitation laser
scanning microscopy is one of the most exciting recent developments in biomedical imaging
and becomes more and more an inspiring imaging technique in the intact bulk tissue
examinations. In this chapter, the non-linear excitation imaging technique including two-
photon autofluorescence (2PF) and second harmonic generated signal imaging (SHG) was
employed to investigate the microstructures of whole-mount corneal, retinal, and scleral
tissues in their native environment. In details, image acquisition was based on intense
ultrafast femtosecond near-infrared (NIR) laser pulses, which were emitted from a mode-
locked solid-state Ti: sapphire system. By integrating high-numerical aperture diffraction-
limited objectives, multiphoton microscopy/tomography of ocular tissues was performed at a
high light irradiance order of MW-GW/cm2, where two or more photons were simultaneously
absorbed by endogenous molecules located in the thick tissues. As a result, the cellular and
fibrous components of intact scleral and corneal tissues were selectively displayed with the
assistance of the in-tandem detection of 2PF and SHG procedures. Any exogenous dye was
not used. High-resolution optical images of keratocytes in cornea, fibroblasts, mature elastic
fibers and blood capillaries in sclerae as well as of the retina radial Müller glial cells,
ganglion cells, bipolar cells, photoreceptors, and retina pigment epithelial (RPE) cells were
acquired. Furthermore, this promising technique has been proved to be an indispensable tool
in assisting femtolasers intratissue surgery, especially for in situ assessing the obtained
microsurgical effects. Most remarkably, the activated keratocytes, also named myofibroblasts
during wound repair, were in vivo detected using the multiphoton excitation imaging in the
treated animals twenty-four hours after the intrastromal surgery. In this chapter, RESULTS is
divided into five areas of interest: (1) corneal multiphoton imaging; (2) scleral multiphoton
imaging; (3) retinal multiphoton imaging; (4) uses of multiphoton excitation imaging in
intrastromal laser surgery; (5) uses of multiphoton microscopy in detecting activated stromal
cells after surgical laser treatment. Data show that the in-tandem combination of 2PF and
SHG imaging allows for in situ co-localization imaging of various microstructural
components in the whole-mount ocular tissues. Qualitative and quantitative assessment of
microstructures was obtained. The selective displaying merits of tissue components only with
the excitation of different wavelengths is the most exciting development for bulk tissue
imaging, which allows to selectively studying of three-dimensional (3-D) architecture of
cellular microstructures and extracellular matrix arrangement at a substantial depth. Using the
laser power within threshold value, the bulk tissues can be imaged numerous times without
visible photodisruption. Intrinsic emission multiphoton microscopy/tomography is
consequently confirmed to be an efficient and sensitive non-invasive imaging approach,
featured with high contrast and subcellular spatial resolution. The non-linear optical imaging
yields vivid insights into biological specimens that may ultimately find its clinical application
in optical pathological diagnostics. The authors believe that this promising technique will also
find more applications in the biological and medical basic research in the near future.
Potentially, based on its capability of intrinsic fluorescence imaging, the 2PF of cytoplasmatic
Preface xi
NAD(P)H would have a use in studying the dynamics of the mitochondrial bioenergetics and
further utility in conducting nanomedical investigations in this organelle.
Chapter 8 - Ultrafast titanium-sapphire tunable femtosecond (fs) laser can produce fs and
provide high-density to be used as excitation resource, it is ideal tool to study the optical
properties of materials. Rare earth (RE) compounds have many potential and practical
applications in display device, upconverson laser, fluorescent probe, optical communications
and fiber optic amplifiers, etc. The upconversion luminescence (UCL) and fluorescent
changes of RE ions in transparent hosts and complex induced by fs laser have aroused a large
amount of interests. In this review, the authors systematically reported the progresses of UCL
of RE ions (such as Eu3+, Tb3+, Ce3+ and Dy3+, etc) in glass, crystallized hosts, and
complexes, including luminescent characteristic, upconversion mechanisms and electronic
transition processes. The changes of optical properties of RE ions by fs laser irradiation can
achieve valence manipulation of RE ions and microfabrication. Thus in this review the
authors also intensively reported the progresses of this topic, including persistent spectral hole
burning, structural changes, dynamic processes etc.
Chapter 9 - At the end of the last century, three important technological breakthroughs
took place in ultra-short laser pulse generation field. In 1985, Strickland and Mourou first
demonstrated Chirped Pulsed Amplification [1]. Then Moulton, in 1986, developed a new
laser, based on aluminium oxide doped by titanium ions (Ti3+:Al2O3), called Ti:Sapphire [2].
Finally, Spence, Kean and Sibbert carried out the first phase mode-locking (“magic-
modelocking”) in Ti:Sapphire also called Kerr Lens Mode-locking [3]. Since this moment, all
required elements have been demonstrated to build reliable and efficient ultra-short pulses
laser chain.
Chapter 10 - It is well known that in the natural photosynthesis system excitation energy
captured by the light-harvesting complex is transferred to the reaction center after an energy
migration process [1]. In the natural system, higher excitation energy generating the higher
excited states of dyes such as chlorophylls and carotenes is also utilized, despite its very short
lifetime [2]. Inspired by the natural system, artificial energy transfer systems, in which energy
transfer of higher excited-state energy is realized, have been demonstrated by various

researchers [3]. On the other hand, for most molecules, no excited singlet states (S2,
S3…states) have been observed to emit light upon excitation, the same being true for triplet
states (T2, T3…states). This is summarized in Kasha’s rule: The emitting electronic level of a
given multiplicity is the lowest excited level of that multiplicity [4].
Chapter 11 - In past years, the femtosecond lasers have been widely used to study the
time-dependent phenomena in the fields of physics, chemistry, biology, and material sciences.
At the same time, more and more experimental techniques and theoretical methods are
developed to investigate the ultrafast dynamics in the electronic excited states of complex
molecular systems. Nowadays, the time-resolved ultrafast spectroscopy, excited-state
quantum chemical calculations, and excited-state dynamics simulations have been versatile
tools for the study of the electronic excited-state ultrafast dynamics of complex molecular
systems. In this chapter, the authors have reviewed their recent semiclassical excited-state
dynamics simulations results of some complex molecular systems. The detailed changes for
molecular conformations and electronic structures in the femtosecond excited-state ultrafast
dynamics of Tetraphenylethylene (TPE) and 9,9’-bianthryl (BA) molecules are dynamically
simulated. Moreover, the dynamics simulations results are compared with the femtosecond
time-resolved spectroscopic results.
xii Paul-Henri Barret and Michael Palmer
Chapter 12 - This chapter represents an overview of the current state of the art in the field
of femtosecond technology with a special emphasis put on the research on high-intensity
laser-matter interactions. Ultrashort pulse laser ablation provides a promising approach for
processing of almost any kind of material. However, this method has been investigated
mainly by research on solid materials such as metals, semiconductors, etc. Studies on large
biological molecules are still only slightly covered. Following a brief description of the
fundamentals of femtosecond laser technology, an introduction to the physics of the high
intensity laser interaction will be provided. Thorough understanding of heat transfer
mechanism and energy conversion routes is of crucial importance for controlling the resulting
modification of the target material in many practical applications. In the current studies
femtosecond laser ablation was applied to complex biological molecules. Measurements of
TOF mass spectra obtained from the laser–tissue interaction are presented. TOF mass
spectroscopy has been applied to characterize the species produced in the femtosecond (fs)
and nanosecond (ns) regimes. These studies aim to detect and give information on the
photodissociation of these biological molecules and to compare the two ablation regimes. The
morphological changes of the material were investigated using Environmental Scanning
Electron Microscopy. The results will demonstrate the feasibility of the USLA-TOF method
as a high sensitivity analytical technique for elemental analysis and for detection of fragile
biological molecules. Some comments on the potential novel applications of ultra-short laser
radiation are presented, namely efficient implementation in ophthalmology, dentistry,
nanosurgery and innovative trends in femtosecond technology.
Short Communication A - The authors report the dynamics of confined exctions when
several exciton states are excited. The dynamics are measured by using a transient grating
(TG) technique and a degenerate four-wave-mixing (DFWM) technique. In the system of the
quantization of exciton center-of-mass motion, the excitons have several levels in the narrow
energy region. The transient signals generated by the excitation of several exciton levels by
using an ultrashort-pulse laser with a broadband spectrum are different from those generated
by the excitation of single exciton level. The excitons confined in GaAs films with a specific
thickness show enhanced excitonic optical nonlinearity, which enhancement is described by

the nonlocal response theory, and an ultrafast response that is comparable to the pulse width
of the DFWM signals. This result is an important factor in the development of ultrafast
optical devices that exhibit large optical nonlinearity and ultrafast response. Moreover, the
TG signals show a long-lived component, which is longer than that of the single exciton level.
The changes in the response profile can be attributed to the excitation by a spectrally
broadband pulse. Their results will provide impetus to the development of ultrafast optical
devices based on the excitonic optical nonlinearity.
Short Communication B - This publication is devoted to development of 100 W quasi-
continues wave 950 nm long pulse width (5 ms) laser diode bars. Results on optical output
power and degradations characteristics within wide temperature range (from -400C to +850C)
are presented. Developed LDBs are based on advanced single quantum well separate
confinement heterostructures with index-graded waveguide design optimized for generation
wavelength 950 nm at +250C, low threshold current density (70 A/sm2), high characteristic
threshold current temperature (T0=150 K) and good slope efficiency (1,1 W/A).
Estimated LDB lifetime in long pulse width regime is more than 109 and 108 pulses at
heat-sink temperature +250C and +850C, respectively.
Preface xiii
These laser diode bars are developed for effective pumping of Yb-Er glass solid state
lasers (SSLs) with generation within the “eye-safe” spectral range (1.54 µm). Design of Yb-
Er glass SSLs using for optical pumping two 950 nm laser diode bars was developed. Such
SSLs have the pulse energy up to 8mJ at multimode regime, the pulse width 20 ns and
repetition rate up to 20 Hz.
RESEARCH AND REVIEW STUDIES
In: High-Power and Femtosecond Lasers ISBN 978-1-60741-009-6
Editor: Paul-Henri Barret and Michael Palmer © 2009 Nova Science Publishers, Inc.
Chapter 1
RECENT PROGRESS OF HIGH PEAK POWER SOLID

STATE LASERS
Yung-Sheng Huang∗1, Jung-Sheng Huang1 and Fang-Ling Chang2

1
College of Electrical Engineering and Computer Science,
Semiconductors Research Laboratory, I-Shou University, 840 Kaoshiung, Taiwan
2
Research Institute of Computer and Information Science,
University of Oregon, Eugene OR 97405, USA
ABSTRACT
There has been a remarkable upsurge in the importance of lasers, especially in high
power lasers, on optics in both pure science and in technology in recent years. Besides, by
considering the development of the high power lasers with their rapidly growing list of
applications, the obvious needs for an introduction to this newest field is the primary reason
for this paper. In this article, we present an overall survey of recent advances in high peak
power solid state lasers (SSLs). The contents include:
Reviewing how the Degnan’s thermal Boltzman factor affects the inversion population
which is resulted from the non-uniform temperature distribution in the laser crystal. The
Auger effect (or so- called energy transfer up-conversion (ETU) effect) in the laser crystal
could result in thermal loading which is a function of pulse repetition frequency (PRF). So the
theoretical calculation on the thermo-optic effect and thermal focusing length are discussed to
explain the experimental results of SSLs. An introduction to the spatial-time dependence laser
model of rate equations for the active (A-O switch) and passive (Cr 4+ :YAG saturable
absorber and semiconductor GaAs saturable absorber) Q-switched SSLs is presented.
In the second part of this paper, we study the E-M transverse modes of laser output,
especially in the generation of high order transverse modes, and calculate the threshold pump
power. In order to obtain a high peak power and single mode laser pulse output, one can use
∗
Email: yshuang@isu.edu.tw
2 Yung-Sheng Huang, Jung-Sheng Huang and Fang-Ling Chang
simultaneously the Q-switching and mode locking technique. In this article, we talk about its
newest development in the past few years.
In the third part of this article, we talk about the non-linear optical device which utilizes
the intense radiation field of the nonlinear optical response of an optical media (i.e. KTP or
KDP, QPM, and PPLN nonlinear crystals) to generate a new frequency emission from UV to
far Infrared band. Thus, the frequency conversion is a useful technique for extending the
utility of high power SSLs. We talk especially about the development of pulsed optical
parametric oscillation (OPO) in SSLs. The discussion materials are based on some
researcher’s recent works. Finally, the OPO tunable source may have impacts on the
development of high power SSLs and important applications in some areas such as gravity
detection, atmospheric detection, DNA detection, global remote sensing…etc.. But we skip
this special issue in this article.
1. THERMO-OPTIC EFFECTS
In the literature "Heat Generation in Nd:YAG and Yb:YAG"[1] in 1993, T.Y. Fan
reported that a small thermal load was observed in the diode-pumped Nd:YAG and Yb:YAG
with a doped concentration of 1.04 at.% and 6.5 at.%, respectively. Chen et al. [2] gave a
maximum stress fracture of the induced heat in the laser crystal caused by the pumping in
power for a diode pumped optical-coupled laser system. Their research indicated that the
value of the maximum stress fracture was influenced by both Nd-doped concentration and the
absorption coefficient [3]. They therefore determined the ratio of thermal load to pumped
laser mode[4] and the relation between ETU effect and second harmonic generation [5-7].
The non-uniform pumping power causes the non-uniform temperature distribution which
results in the thermal dispersion effect, the lattice distortion, the variation of birefringence,
and the thermal expansion in any direction resulted from thermal stress or displacement of the
lattice site. These are the so-called thermal lensing effect. W. Koechner [8-9] had given both
the theoretical and experimental results of the thermal lensing effect. In early years from 1960
to 1970, Foster[10], Osterink[11], Quelle[12]，Gordon[13] and Baldwin[14] had individually
investigated the thermal lensing effect in the Yd:YAG laser, the limitation of the thermally
induced diffraction loss resulting in thermal distortion in the laser rod, and negative thermal
lensing effect. In 1970, Koechner reported the influence of thermally induced birefringence
on the laser performance [9]。Therefore, the study of the thermal effect on the laser crystals
has been developed very maturely. In 1999, Hardman et al. [15] studied the influence of
thermal lensing effect and ETU effect on the laser performance when laser is lasing and non-
lasing.
J. J. Degnan presented some classical papers in 1989[16], 1995[17] and 1998[18] to work
on the optimization theory of Q-switched solid state laser and discuss the thermal
characteristics. In his research, the dynamical rate equation related to the atomic distributions
on the excited states and ground state in thermal equilibrium according to the Maxwell-
Boltzmann statistical mechanics. Degnan derived a thermal equilibrium distribution function
which he called the inversion reduction factor γ [16-18]. In the textbook of A. Yariv[19], γ
equals to 2. Degnan also measured the fractional distribution function to be a constant at
Recent Progress of High Peak Power Solid State Lasers 3
room temperature [18]. This fractional distribution function also appeared in the model of
passively Q-switched laser, and end-pump quasi-three level Q-switched optimum laser model
[20-21]. These models were developed by R. Beach independently [22]. However, Li et al.
indicated in their recent paper that the equilibrium fractional distribution function derived by
the Maxwell-Boltzmann statistic is not a constant but a function of both temperature and the
lattice site for non-uniform temperature distribution [23]. Therefore, three equilibrium
fractional distribution functions have to be interpreted from another point of view. Li et al.
reported that their model could describe the laser system more accurate, since they considered
the non-uniform thermal distribution in the end-pumped laser crystal. They analyzed the axial
heating model with the cylindrically symmetric crystal and using the surface water cooling
system resulting in the heat-convection diffusion toward its periphery. This boundary
condition was used in their laser thermal model. It was shown that under this model, at certain
cooling condition, the pulse output energy for once or twice circulated back and forth
pumping geometry reduced about 21.7% and 14.5%, respectively. Their simulated results also
suggested that the non-uniform temperature distribution reduced not only the pulse energy but
also the output efficiency. Considering their theoretical approach, the authors adopted more
accurate equations by using non-uniform thermal distribution along the axis of the crystal.
Therefore, the excited and emission Blotzmann distribution factor between the excited and
ground states has to be modified. Since the temperature is not a constant but varies along the
axis of the crystal, the Blotzmann distribution factor depends on the axis of crystal as well.
Therefore, further investigation is needed on how the quantum effects reach equilibrium with
the temperature and thermal effect inside the laser crystal, and then influence the laser pulse
output energy of the resonator under macroscopic condition.
Liu et al. [24] had demonstrated an end-pumped frequency-doubled acoustic-optically Q-
switched Nd:GdVO4 laser from which the green output at power level of several Watts
obtained. The ETU influence and the dependence of the initial and residual inversion density
on the Q-switching repetition frequency had not yet been considered in their theoretical
model. The differences between their theoretical predictions and the experimental results
became significantly remarkable at high power level. Huang et al. proposed a modified model
where the ETU effect and the influence of Q-switching pulse repetition frequency were
investigated [25]. They also reported that the thermal effect degrades the laser performance
greatly especially in high output power.
2. THERMAL FOCAL LENGTH MEASUREMENT

End pumping the solid-state laser is a standard method to produce the laser radiation with a
single transverse mode. For the high power laser operation, the thermal lens of the laser
crystal is an important parameter for optimization of the system. In 1995, B. Ozygus and J.
Erhard [26] measured the beat frequency of adjacent transverse modes to determine the
thermal lensing of end-pumped laser crystals. In 1997, B. Ozygus and Q. Zhang [27]
presented a simple, effective method to determine the focal power of end-pumped lasers. At
certain resonator parameters the transverse structures of the laser beam become a
superposition of many transverse cavity modes of equal resonance frequencies. Because the
resonator parameters depend on the thermally induced lens in the active medium, the strength
of the lens can be determined by means of measuring the shift of the degeneration resonator
length depending on the pumping power.
Osterink et al. [28] presented an analysis model of the thermal effects in Nd YAG.. They
considered the index of refraction variations in the rod due to index changes with temperature
dn/dT, and photo-elastic effects from thermal strains. (But the variations are also undoubtedly
effected by non-uniform pumping and cooling). The temperature distribution in the rod was
computed by using an infinite heat generation and constant surface temperature. Temperature
and hence refractive index were found to vary quadraticly with the radius. Then, the thermal
stress distribution in the rod was determined and the photo-elastic index of refraction
variations were calculated. The results were a quadratic radial variation in refractive index
and an induced uniaxial birefringence whose extraordinary axis always lies in the radial
direction and whose ordinary axis lies in the tangential direction. The focal length varies
inversely with the internally generated heat (pump power), and the focal length is independent
of the coolant temperature (flow rate). The average focal length in the rod is determined only
by the term in the denominator which contains dn/dT, since the photo-elastic term will be
averaged to approximate zero. The cylindrical geometry of laser rod is used generally. The
heat is removed on the circumferential surface of the cylinder, thereby a radial thermal
gradient is generated in the cylindrical geometry. The change in temperature within a laser
rod causes a thermal distortion of the laser beam due to a temperature stress dependent
variation of the refractive index. In addition, the stresses induce birefrigence is generated. An
optical beam propagating along the rod axis suffers a quadratic spatial phase variation. This
perturbation is equivalent to the effect of a spherical lens [29]. According to the early work by
W. Koechner [8] under intense pumping in end-pump configuration, the end effects that
account for the physical distortion of the flatness of the laser rod ends should be considered.
The focal length of the rod caused by an end face curvature is obtained from the thick lens
formula of geometric optics.
Osterink et al.[28], Koechner et al.[8,30-31], and Innocenzi et al.[32] derived the
theoretical analyses that the thermal focal length, f, which is inversely proportional to the
input power. Comparing the results of these theoretical models with the experimental data
measured by Ueda et al.[33] and Mukhopadhyay et al.[50], the calculated thermal focal length
is not fit to the measured data. Ueda et al. considered the thermal load ξ being a constant
(ξ=0.5 is adopted in their calculation) and a characteristic of the crystal. However, Huang et
al. thought that the focal length is proportional to the thermal load, ξ, but it is not a constant.
Instead, ξ is a function of pump power and it is proportional to the pumping power, namely,
ξ(Pin). They used the experimental data measured by Ueda et al. (efficient and compact
intracavity-frequency-doubled Nd:GdVO4/KTP laser end-pumped by a fiber-coupled laser
diode) as an example to verify their model[34]. The results show that the variations of
theξvalue is a linear function under lower input power, however, theξ value approaches to
saturation when the pump power is getting to a large level as shown in Figure 1.
Zhao et al.[35] combined some formulae to present the outline of the theoretical analysis
for the whole thermal focal length of LD end-pumped solid state laser with stable resonator
measuring process. They used slit-scanning method (SSM) to measure the beam quality for
the multi-mode Gaussian beam field.
Figure 1. Measured thermal focal length and the corresponding calculated thermal loading with the
same plot.
Because this method can be operated easily and provide a more reliable beam
information. They measured the beam width at different positions with different pump power
and determined the corresponding TEM00 waist of the optical resonator based on ABCD
matrix theory [36]. Thus the thermal focal length of the gain medium was determined. The
accuracy of this experiment is nearly comparable with the previous researcher.
D. C. Brown indicated that when laser are operating under the condition of high average
power [37], and for the laser crystal, its thermal conductivity , thermal expansion coefficient,
and the variations in index of refraction are all functions of temperature, then thermally
induced variations in the index of refraction are generated. Therefore, the radial and
tangential polarizations are no longer of the quadratic form. Hence a higher order spherical
aberration needs to be considered, and the generated phase aberrations can be expressed as a
Zernicks polynomial series. The author called this effect as nonlinear thermal distortion [38].
3. TRANSVERSE MODES
The advantages of using passive Q-switch in laser resonator are the generation of high
peak power laser, with tiny and compact sizes, and more sturdy and durable. However, there
is a serious problem - the pulse delay on the application of the high peak power laser to
distance measuring or remote sensing. The pulse delay is generated by the pulses time
differences of the different laser transverse modes in the resonator. The results of distance
measuring are affected by the quality of laser transverse mode output. With a laser of good
transverse mode quality, the spot size still remains small after a long distance transmission.
The errors of distance measuring results are caused by the time delays of the pulses arriving at
the target. In general, the practical laser output seldom generating a single TEM00 mode,
together with the imperfect optical system, the resonator without symmetry, the Brewster
angle, the lens defects, the out-of-focus effect … etc., all these effects will generate high
order transverse modes. For a fixed laser spot size, the lower the transverse modes order, the
smaller the divergence angle is, and therefore is more suitable for the application to laser
distance measuring.
The key to the laser output power and the generation of transverse modes is the
overlapping degree between the absorption distribution area of the gain medium and the
intensity distribution area of the pump source in the laser resonator. If the area of the laser
gain medium is small, the output of transverse mode will be a single mode. If the cross
section is large, the output will be multi-modes. Laporta et al.[39] derived a formula
indicating that the better the overlapping degree between the pump modes and the resonator
modes, the lower pump threshold power is. Chen et al.[40] applied Laporta’s formula to
calculate the threshold pumping power as a function of the off-focusing parameter as shown
in Figure 2.
The results indicate that when the off-focusing parameter approaching to zero, the
TEM00 mode has the lowest threshold pumping power. In other words, the TEM00 mode is
generated most easily at lower off-focusing. They also calculated the Hermite Gaussian mode
and Laguerre Gaussian mode spatial distributions. The pure Hermite Gaussian mode
distributions are shown in Figures 3 and 4.
Figure 2. The threshold pump power for fundament and different high order modes varies with different
off focusing parameter.
Figure 3. TEM 00 laser mode in the x-y plane.

Figure 4. TEM 5, 0 laser mode in the x-y plane.
Figure 5. LG 1,5 laser mode in the polar coordinate.

Figure 6. LG 1, 5 with off focusing parameter laser mode in the polar coordinate.
The pure Laguerre Gaussian mode distribution is shown in Figure 5. The results show
that the lowest-order mode on the off-focusing plane has the lowest threshold pump power. In
other words, in this region, the low-order modes are generated more easily than the high-order
modes. They adopted M. Arvidsson’s experimental structures [41] to generate the pure
Hermite-Gaussian mode and Laguerre-Gaussian mode by using the optical fiber coupled off-
axis technique. They also simulated the high-order modes generated from the different off-
focusing and off-axis degrees of the Hermite-Gaussian mode and Laguerre-Gaissian mode as
shown in Figure 6. They used these high-order modes investigating the effects of the time
delay when the pulses arriving at the target at far field. Chen’s simulation results showed that
the time delay is from the minimum, several hundreds ps, to the maximum, 20 ns, and this
cause 6m length errors in the distance measuring.
4. RECENT DEVELOPMENT OF HIGH POWER SIMULTANEOUSLY

Q-SWITCHED AND MODE LOCKED LASERS
Mode locking is a technique which phase-locks the longitudinal modes in the resonator.
The larger the number of phase-locked longitudinal modes, the narrower the output width is.
The electrical field intensity, which is caused by the effect of mode coupling with the
longitudinal modes in the resonator, as a function of time is discussed by A. E. Siegman [36].
If the number of locked modes is increased, the ratio of intensities between the major pulses
and the minor pulses will also be increased, but the widths of the pulse will be decreased.
These pulses are periodic. Therefore, the wider the laser gain bandwidth, the larger the
number of longitudinal modes are contained and the easier the short pulse laser is obtained.
The pulse widths are influenced more by the phase relationship than by the amplitudes
relationship of the longitudinal modes.
It was indicated by Chen et al.[42-43] that if the initial transmission , T0, of the saturable
absorber become smaller, then the pulse widths of the Q-switched lasers become narrower
and the energy of these pulses become larger. The peak power of the mode-locked laser pulse
is larger at small T0 values than at large T0 values. Because the smaller the T0 values, the
shorter the build-up time of the Q-switched laser pulse is. Therefore, the remaining amount of
longitudinal modes is increased after passing through the laser linear gain medium, and this is
of benefit to generate the mode-locking mechanisms. Thus the number of the longitudinal
modes involved in the mode-locking mechanisms is determined by the T0 value as shown in
Figures 7-9. Another method to increase the number of longitudinal modes is by using a
longer resonator. Chen et al. designed a z-configuration resonator to avoid the output
instability caused by the thermal lens effect which is due to the longer resonator.
As mentioned before, another factor which determines the quality of mode locking is the
phases between the longitudinal modes. In general, the phases between the longitudinal
modes are not all the same which causes the background noise signals distribution of the
mode-locking laser output pulses.
Therefore, the better the quality of the mode-locking, the larger the ratio of signal
magnitudes between the mode locking pulses and background noises is. The results of
numerical simulation are consistent with the experimental measurements in the number of
mode-locking laser longitudinal modes inside one Q-switched laser pulse and relative
intensities between mode locking laser pulses.
Figure 7. Calculated simultaneously Q-switched and mode locked laser output under the initial
transparency of saturable absorber T 0 =0.65.
Figure 8. The same conditions as in Figure 7 , except for T 0 =0.5.
Figure 9. The same conditions as in Figure 7, except for T 0 =0.35.
Zhao et al.[44-49] reported a series of models and experimental data of the high power Q-
switched solid state lasers from the year of 2004 to 2008. They considered a folded frequency
doubling cavity system and proposed a new spatial-time dependence rate equation for LD
end-pumped actively Q-switched intracavity frequency doubling laser. The equation includes
the variations of photon densities along the resonant cavity axis because of the long resonant
cavity and also includes the loss caused by the thermal lens effect of the laser crystal. The
calculated and experimental generated green light pulse profile was in agreement quite well
[44]. They replaced the active acoustic optic Q-switch by a GaAs saturable absorber and
considered the single photon absorption (SPA), two-photon absorptions (TPA) and free
carrier absorption (FCA) into rate equations. A model of LD end-pumped passively Q-
switched intracavity frequency doubling Nd:GdVO4/KTP laser was developed. The numerical
solutions of the theoretical calculation agreed with the experimental results [45]. Using the
same procedure, by considering the Gaussian spatial distribution of the intracavity photon
density and initial population inversion density as well as the influence of the AO Q-switch,
Zhao et al. gave a coupled spatial-time dependent rate equations for a diode-pumped doubly
Q-switched laser with both an AO modulator and a Cr4+-doped saturable absorber. These
coupled rate equations were solved numerically. Some curves were generated to explain the
optimal key parameters for the laser design [47]. In general, using a passive-passively double
Q-switched laser with Cr4+:YAG and a GaAs saturable absorber can generate more
symmetric and shorter pulses [48].
Some saturable absorbers, such as the Cr4+:YAG crystal, the GaAs wafer, the
semiconductor saturable absorption mirrors … etc. are employed for simultaneously Q-
switching and mode-locking. The GaAs wafer is more widely used in passively Q-switched
and mode-locked 1-μm lasers than the other saturable absorbers. Combining the GaAs wafer
with the Nd3+-doped laser crystal working at 1.06μm is especially used widely because of its
low price, easy process, and having high threshold damage value. With the same model
formalism mentioned above, Zhao et al. proposed a diode-pumped passively and
simultaneously Q-switched and mode-locked frequency doubling green light laser by using
the GaAs wafer as the saturable absorber [46]. Considering the free carrier absorption (FCA)
process in GaAs wafer, a set of modified rate equations were used to describe the doubly Q-
switched Cr4+:Nd3+:YAG laser with GaAs coupler in a short cavity. Their results showed that
this laser can generate a more symmetric pulse shape, a shorter pulse width, the higher peak
power [48]. In general, comparing with the passively Q-switched and mode-locked laser
(QML), the actively QML generates more stable QML pulses. Zhao et al. established a
relatively simple structure and a short resonant cavity length to generate a high efficiency and
high average power actively Q-switched and mode-locked frequency doubling green light
output [49]. The related researches of this subject can be referred to the recent works of
Mukhopadhyay et al.[50].
5. COATING TECHNIQUE AFFECTING THE LASER OUTPUT BEAM

QUALITY
Coating on the nonlinear crystal surface can increase the structure compactness, decrease
the device loss, and enhance the output power and stability. Figure 10 shows the reflection
coefficient as a function of the wavelength for the three layers anti-reflecting (AR) coatings.
It shows that the reflection coefficient is reduced effectively by the AR coating in the visible
band.
Figure 10. Reflectivity of three layers anti-reflecting (AR) coating in the visible band.
Figure 11. Multi-layers for high reflecting coating in the visible band.
The reflection coefficient as a function of the wavelength is shown in Figure 11 for multi-
layers high reflecting coatings. It shows that the reflection coefficient is increased effectively
by the multi-layers high reflecting coating in the visible band. However, coating also
increases the price and the coated film may be damaged under the high power output. In
frequency doubling technique, the fundamental wave oscillates in optical resonant cavity and
generates second harmonic wave through intracavity nonlinear crystal. If anti-reflecting
coating technique on the output coupler mirror is not good enough, the generated second
harmonic wave will return into the laser gain medium, and results in a non-linear polarized
output wave. Therefore, coating techniques on the optical elements play an important role in
determining the beam quality of the laser output.

Chen et al.[51-52] has demonstrated a 532nm green CW laser of high pumping power, high
repetition rate, and output with Watts, and a 671nm red laser with output level close to Watts.
The system was designed to use an Nd:YVO4 laser crystal which is end-pumped with 808nm
laser diode to radiate 1342nm infrared light source. Applying frequency-doubling technique,
a 671nm red laser output is then obtained. Compared with the semiconductor laser diode, this
all solid state laser with red light output has a much better beam quality than the laser diode.
Experimental results show that when pump power is more than 17W, the red light output
power starts saturation and the optical cavity is getting unstable because of the thermal
lensing effect.
6. QPM TECHNIQUE AND QUASI PHASE MATCHING NONLINEAR

CRYSTAL
Quasi-phase matching (QPM) can offset the phase velocity dispersion problem during the
frequency conversion processes. The phase mismatching between the fundamental waves and
the harmonic waves caused by refraction index dispersion during the frequency conversion
processes is offsetted by the periodic modulation of the crystal nonlinear polarization. The
modulation period can be designed according to the refraction index dispersion of the crystal.
In this way, not only the maximum effective nonlinear coefficients of the crystal can be
chosen, but also QPM can reach the noncritical phase matching at some specific temperature
that is not possible for the birefringent phase matching. Yariv et al. [53] investigated the
theoretical analysis of QPM.
The nonlinear polarization tensors used by PPLN, PPKTP, and PPLT are all d33. The light
beam transmits along the x-axis, and thus the walk-off effect does not exist. Therefore, the
crystals can be fabricated very long to raise the conversion efficiency. The general used
nonlinear optical crystals, such as KTP, BBO, LBO … etc.[54], are cutting along some
specific directions and works at some specific temperature to satisfy the phase matching
condition. And only some specific non-diagonal element coefficients of these nonlinear
crystals can be used because of the differences of the mutually interacted light wave
polarization directions.
During the frequency conversion processes, such as SHG or OPO, the phase mismatching
will be accumulated proportional to the interactive length. After passing through the
coherence length, the conversion efficiency and the output energy are decreasing because of
the phase mismatching. When the crystal length is larger than the coherence length, the
variations of the conversion efficiency from the fundamental waves to the harmonic waves
are fluctuated from zero to the relative peak values according to the crystal mutually
interactive length.
If the nonlinear coefficient can be modulated every double coherent length, (In other
words, the nonlinear coefficient changes the sign after each coherent length.) then the
accumulated phase mismatching is canceled. Therefore, the conversion efficiency will keep
increasing according to the interactive length. Thus the QPM technique is dividing the
nonlinear crystal into several regions. Each region is equal to one coherent length, and the
light axis rotates 180o in each region. Thus, the nonlinear polarization light rotates 180o after
passing through one coherent length. The coherent length is usually equal to several μm long
in the frequency conversion processes.
Periodically poled KTP (PPKTP) is a nonlinear optical crystal of totally new type [55-56].
The PPKTP crystal is fabricated by several etching steps and electric polarized processes, and
its nonlinear optical characteristics is changed permanently. PPKTP has the same transparent
wave ranges as the bulk KTP but has no the phase mistmatching problem which the bulk KTP
has troubles with. Compared with the widely used periodically poled LiNbO3 (PPLN)[52,57-
58], PPKTP has several advantages. Firstly, in the crystal fabrication stage, the required high
electric field is about 21 KV/mm for the LiNbO3 crystal, but it takes only about 2 KV/mm for
the PPKTP to accomplish the polization. Therefore, the maximum crystal thickness is limited
to 0.5 mm for PPLN. But the crystal thickness for PPKTP is at least 1 mm. Secondly, the
damage threshold of PPKTP is far more higher than PPLN. Hence the PPKTP crystal can
operate at room temperature. Thirdly, PPKTP has better performances than PPLN at high
power circumstances. The effective nonlinear coefficient, d33, of PPKTP is 17 pm/V which is
lower than the value, 27 pm/V, of PPLN. However, the damage threshold of PPKTP is larger
than 900 MW/cm2 for 5 ns pulse which is much higher than PPLN. At the same time, PPKTP
has low susceptibility for photorefractive effect. Therefore, it can be operated at room
temperature. Yang et al. [55-56] reported that using PPKTP nonlinear crystal in diode-pump
Nd:GdVO4 laser operating both in CW and active Q-switched, the green laser output was
obtained. Another group reported a model about PPLN OPO in case of CW wavelength with
strong idle absorption condition [59].
7. THEORY OF PARAMETRIC OSCILLATION (OPO)

Consider a visible light impinging onto the cornea of the human eye, being focused by the
eye lens, passing through the aqueous humor, and impinging onto the retina. At the retina the
light is absorbed. The photon energy is converted by the retina into chemical signals which
are transmitted to the brain by the optic nerve and stimulating optical sensations. The parallel
rays of a laser (especially on the visible light range) can be focused through the eye lens to a
point image of about 105 times larger in light intensity than the conventional sources of
illumination. In fact, due to the high intensity of the laser radiation, the retina will be severely
damaged only in a very short exposed time. When the amount of exposure to the laser
radiation exceeds some critical value, the eyes will be damaged no matter what wavelength of
the laser radiation is. The maximum permissible exposure of laser radiation to the eye
depends on the wavelength [60]. Therefore, on the application of remote sensing, optical
communications, and optical detections, using the safer laser wavelengths for human eyes is
desired. Recently, the main technique for the application is to use OPO-generated pulse laser
with 1.5-1.6 μm wavelengths. The pulsewidth is between the nano seconds to pico seconds.
Chen et al. [51-52] have used hemispherical cavity to produce a pulse of 1.572 μm
wavelength and 1.56 W output power at 58.8 KHz repetition rate frequency under 14.5 W
input power.
Louisell, Yariv and Siegman proposed the general quantum theory for the OPO single
mode in 1961 [61-65]. They explained the sources of the quantum noise (or parametric
fluorescence [66]) by using the quantum statistics, and derived its Hamiltonian form. Cassedy
and Jain [67] continued Bjorkholm’s successful research on the injection tuning operation of
OPO. They derived a more maturely theoretical analysis which employed the Gaussian-shape
pump beam to achieve the successful injection and analyzed the time-dynamic behavior of
OPO multi-modes in OPO’s. The theoretical calculation of injection usually stems from the
equation of the single mode OPO quantum theory proposed by Louisell et al.. Ritsch et al.
[68] and Schwob et al.[69] proposed the quantum theoretical analysis of the simultaneous
multi-modes of OPO.
The signal in an OPO is built up from the quantum noise. We use the OPO dynamic
equations to illustrate a numerical example for the temporal evolution of DROPO pumped by
the Gaussian beam profile as shown in Figure 12. It indicates that without seeding a quantity
of the photon source (or the idle source) provided by the parametric fluorescence the signal
wave can not grow up. In pulsed OPO, the macroscopic signal fluctuations are a function of
time, pulse energy, frequency spectrum and transverse mode profile [70-71]. Oppo et al.
proposed the theoretical analysis on the nanosecond SROPO spatial beam quality of output
pulse combined with the idle wave absorption[72]. Smith concluded that the effects on the
instability of CW pumped DROPO operation are due to pump frequency fluctuations causing
the amplitude instability of the oscillator. In other words, the instability comes from the shift
of the pump frequency[73-74].
Figure 12. Temporal profile of OPO. The dash line indicates un-depletion pump field. The blue solid
line represents depleted pump field, and the solid green line is the signal wave.
The spiking phenomenon caused by the behavior of oscillator mode hopping is observed
from the experimental measuring. The oscillator output amplitude fluctuations are also caused
by the variations of the resonant cavity length. Hence the oscillator tuning is resulted from the
pump tuning.
If we want to obtain a single mode output, an etalon element can be inserted inside the
pulsed SROPO resonant cavity. An advantage of DROPO compared with SROPO is that
DROPO has a lower threshold that comes from the resonance of both signal and idle waves in
the optical cavity. Based on this reason, DROPO is widely used for CW operation, and the
theoretical conversion efficiency can reach 100% when pumping power is four times than the
threshold value. Such a low pump power makes it possible to use a semiconductor laser diode
as a pumping source. Thus DROPO has a valuable potential on the commercial applications.
However, the technique for generating DROPO operation is difficult because of the instability
on oscillation frequency.
At CW SROPO operation, there are several hundreds of modes within the bandwidth of
the parametric gain envelope curve. These modes come from the variations of the cavity
length and this results in the frequency tuning – a continuous variation of different cavity
modes in the linear gain region. Through the mode competition mechanism, only the mode
which has the nearest location to the center of gain curve is survived.
DROPO is interesting contrast to SROPO. Many characteristics of DROPO attribute to
its double resonance of the signal and idle waves. The essential condition for DROPO laser
output is that the eigen modes of the signal and idle waves must overlap to each other within
the parametric gain curve. This is because the dispersion relation existed in non-linear crystal
and a little change in the OPO cavity length resulting in large variations between the signal
wave modes and the idle wave modes. Only a few modes will locate within the gain curve.
This phenomenon was demonstrated in Giordmaine-Miller’s diagram[75] and is called “mode
clustering” or “mode-hopping”. And that is the main reason accounting for the unstable
frequency output of DROPO operation.
J Falk suggested reducing the feedback of signal waves via coating techniques on nonlinear
crystal. And this method actually reduces the instability of DROPO, therefore it is called
SROPO[76]. This method arranges the optical elements in the non-collinear geometry of
optical cavity and the resulted signal wave path is not the same as the idle wave path. The
IOPO configuration was also proposed by Falk[77]. The configuration includes any
homogeneous broadened laser oscillator system. The effect of population inversions
dynamics of the laser rode on the OPO performance is considered in Falk’s theory. And the
population inversions dynamics is ignored by Harris[78]. Falk’s numerical solutions predicted
the laser output behavior in a spiking regime, and showed that the repetition rate is increased
as the pump power increases. Falk also used the Green function method to analyze the
focusing effect[79]. This theory includes the diffraction effect and the Poynting vector of
walk-off effect to evaluate the threshold pump power of OPO.
Bjorkholm, Ashkin and Smith predicted that the non-resonant reflectivity of pump wave
feedback into nonlinear crystal for DROPO and SROPO configurations can effectively
improve OPO efficiency and reduce both threshold and build up time[74]. They also proposed
the spatial non-uniform effect of OPO under Q-switched pulsed Gaussian beam profile. The
results of the theoretical calculations and experimental measurements are in good agreement
[80]. Injection tuning technique provides a precisely control of the output spectral frequency
−1
of OPO ( < 1cm in IR range). Bjorkholm and Danielmeyer proposed a theory and
experiments for explanation of the single mode and adjustable frequency output pulse for the
pulsed injection tuned OPO. SROPO with a pump resonance is called PRSRO. This system
has the advantages of single frequency output, mode-hop-free oscillation, and a lower
threshold. These advantages had been theoretical and experimental confirmed by Schneider
[81], Schneider and Mlnek.
Siegman et al.[82] obtained a simple analytical solution for the ultra-short pulse generation
in a degenerate mode-locked OPO pumped synchronously by a mode-locked laser. A very
short pulse (~1 pico second) was predicted at the steady state which is very close to the
threshold where pump depletion is negligible. Hence a mode-locked OPO can generate a
tunable wavelength and ultra-short pulse which is impossible to obtain from the mode-locked
laser. The author used FFT to solve the parametric equation of motion in the frequency
domain for the coupled mode equation.
There are two methods to produce narrow-band OPO. First, using an etalon element
inserts into OPO cavity. This will result in high losses in optical cavity, and thus increases
threshold, also reduces the output pulse energy. The disadvantage of this method is not only
increasing complexity for OPO cavity but also making it more difficult on continue wave
modulation. The second method is via injection seeding into OPO cavity. The advantage of
this method is not only retaining its OPO efficiency but also keeping the simplicity of design
for OPO cavity. It has been demonstrated by the experiment that pulsed seed source with
nano Joules or continuous wave pump power with 1-2 micro Watts are enough to control
wavelength and linewidth of an OPO. And this technique has been successfully applied to the
spectroscopic measuring experiments. Injection seeding of OPO was first verified and given a
quantitative explanation by Bjorkholm and Danielmeyer. A theoretical analysis was proposed
by Cassedy and Jain [67]. This technique can control the spectral properties of an OPO in the
transient behavior of operation and therefore it is applicable only to pulsed OPO. Contrast to
unseeded OPO in which the output pulse energy of each mode shows a strong fluctuation
coming from pulse to pulse. These fluctuations attribute to build-on vacuum oscillation
starting from statistical quantum noise [61]. Fix and Wallenstein gave a detailed analysis on
numerical method and experiments about nano second BBO OPO at seeded sources and
unseeded sources [70]. A special interest is the behavior of transition from seeded to
unseeded operation.
A numerical method for modeling a broad-bandwidth, nano second OPO in the plane wave
approximation has been developed by A.V. Smith[83]. This method based on the split-step
integration method and fast Fourier transform. His theory included diffraction effect and
group velocity dispersion in OPO. A broadband width OPO is helpful to understand the
unseeded spectral property of linewidth, the free running OPO efficiency, and the practical
quantum noise method for injection seeding problem. An injection seeded OPO with walk off
effect and with phase mismatch effect was reported by A.V. Smith and his co-workers [84].
In their model incorporated diffraction effect, birefringent walk off effect , pump depletion,
absorption loss in nonlinear crystal, reflectivity of cavity mirror, and spatial temporal beam
profile into numerical calculation (but the group velocity dispersion effect was not yet
included), so the complexity of program consumed all the digital CPU time. This program is
so-called SNLO. E. Rosencher et al. took into account the cylindrical symmetry of OPO, a
numerical model based on quasi phase matching material (i.e., PPLN) pump by the Gaussian
beam profile was studied [85]. In their model, since absence of the walk off effect, the
isotropic of nonlinear crystal allows 1-D Hankel transform to replace a trditional 2-D
FFT(fast Fourier transform) to evaluate the temporal behavior of cascading effect of SROPO
output pulse.
Peng et al. [86-87] gave a detail analysis on the high average power scaling behavior of
OPOs. They considered thermally induced stress within the nonlinear crystal and the phase
mismatch. Unlike in the solid state laser, the heat generating in OPO comes mainly from
pump, signals, and the absorption of idle waves. The absorption coefficient depends on many
complex factors such as the length of the nonlinear crystal, the wavelength of the resonant
waves, the beam quality, the beam divergence, the length of the resonator, the reflection
coefficient of the output coupling mirrors, and the whole conversion efficiency … etc.. The
author established a multi-pass model to calculate the absorbed resonant-wave power and the
heat generation in an OPO considering all the factors mentioned above.
8. PRACTICAL OPO PERFORMANCE AND RECENT DEVELOPMENT

SROPO usually uses a Q-switched laser with the nano second pulse as a pumping source.
In this situation, high peak power with ns pulse width not only can reach the OPO threshold,
but also can generate a broad bandwidth (usually several nm) spectrum and a multi-mode
output. Rosencher et al. reported a dual-cavity, single mode and stabilized output pulse
DROPO, and derived a model of pulse ns OPO with time-dependent differential
equations[88].Unlike the other models, this model directly adopted classical OPO coupled
mode equations by J. A. Armstrong and N. Bloembergen[89]. Recently, Rosencher et al.
proposed a theoretical model stating the possibility of OPO operation built on isotropic
semiconductor microcavity[90]. Applying this model to the GaAs material, a OPO threshold
condition which is lower than the optical damage can be obtained by modulating the
reflectivity of output coupler. It had been demonstrated by the experiments with over 93%
conversion efficiency of the signal and idle waves for pump depletion OPO devices. A
theoretical simulation program called SNLO was developed by A. V. Smith at Sandia
National Lab[83-84,91]. They also designed and set up two prototypes of the nano-second uv-
sources remote-sensing systems which were supposed to be used in space and operated by the
satellite [92]. They employed the SNLO nonlinear optics software package to design the
system parameters such as the pump beam’s spatial profile, the parameters for the resonant
cavity, the length of the crystal … etc. to generate a 320nm, 10ns pulse. All these parameter
values were obtained by numerical simulation. The systems measuring showed that the pulse
energy was 200mJ and the whole light conversion efficiency from 1064nm to 320nm
wavelengths was more than 20%.
Peng et al. [86-87] developed a high efficiency, high repetition rate, all solid state tunable
intracavity OPO pumped by using Nd:YLF laser. The average OPO output signal power is
more than 1W in the tunable wavelength range from 1650nm to2100nm and in the repetition
rate of 1kHz. The maximum OPO signal energy is 2.5mJ at 1900nm wavelength at which the
pump to signal conversion efficiency is 42%. The maximum average output power is 4W at
3.5kHz.
The widely used OPO configuration is the extracavity OPO configuration which makes use
of Q-switched pulse as the pumping power and energy of OPO, the Q-switched pulse is
required on the refraction limit condition and tight focus state strictly. However, the
intracavity opo configuration was employed by Peng et al. Basically, the OPO locates inside
the resonant cavity. The advantage of the intracavity configuration is generating a high peak
power intensity and a small divergence angle in the laser resonant cavity, and hence it is more
efficient to pump OPO. The numerical simulation of this system was based on the coupled
mode equations proposed by Bass et al..
The quasi phase matching (QPM) has larger conversion coefficient of the nonlinear
coefficient. Because QPM does not have the problem of Poynting vector walkoff in the
process of the nonlinear frequency conversion. Hence, Zayhowski et al. in MIT first reported
using PPLN nonlinear crystal in OPO. Huang et al. [58,93] in Taiwan investigated a
distributed feedback (DFB) diode laser seeding PPLN OPO which generates a pulse of 200ps
and 1549.6nm in wavelength (the IR wavelength range). This PPLN OPO with cavity length
of 2.4cm is pumped by the passively Q-switched ND:YAG laser generating a 4.2ns, 10kW
pulse.
The output conversion efficiency is 22%. Vaidyanathan et al. in US Air Force Wright Lab,
proposed a PPLN OPO operated at continuously tunable 1.3μm to 5μm wavelength range and
10ns pulse.
They employed a long PPLN crystal with 25mm in length. And they reported that at the
condition of high power pumping pulse OPO (more than 20 times of the threshold value),
then the secondary parametric oscillation signal wave was generated. Their experiments
allowed simultaneous multiple wavelength output. The theoretical calculations are agree with
their experimental results. However, it is a technical challenge and formidable for the
practical high average power OPO. R. L. Byer [94-98] reported his works on the laser
interferometer gravitational wave detection (called LISA and LIGO) that were designed to
detect the gravitational waves.
APPENDIX
According to Laporta’s formula[39], we list below some interminable calculations about
the threshold pump power of high order laser transverse modes under different off-focusing
parameter, thereby obtain photon intensity in spatial distribution (see Figures 3-6). The
original source see Chen’s paper[40].
γI sat 1
* Pth (TEM n.0 ) =
ηPL ∫∫∫ S n .0 ( x, y, z )rP ( x, y, z )dv
⎛ 2x ⎞ ⎡ ⎤
⎟ exp ⎢− 2( x +2 y ) ⎥
2 2
2
S n.0 ( x, y, z ) = H n2 ⎜⎜ ⎟
πω 2 n! L
n 2
1 ω
⎝ 1 ⎠ ⎣ ω 1 ⎦
2 α ⎡ 2( x − Δ x ) 2
2y2 ⎤
rP ( x, y, z ) = exp ⎢− − 2 − αz ⎥
πω P (z ) 1 − exp[− αL ] ⎣
2
ω P (z )
2
ω P (z ) ⎦
⇒
2α 2 ⎡ − 2( x − Δx) 2 2 x 2 ⎤
S n.0 ( x, y , z )rP ( x, y , z ) = exp ⎢ − 2 − αz ⎥
π 2 n! L(1 − exp[− αL ]) πω P ( z )ω1
n 2 2
⎣ ω P (z )
2
ω1 ⎦
⎡ 1 1 ⎤ ⎛ 2 x ⎞⎟
× exp ⎢− 2 y 2 ( 2 + 2 )⎥ H n2 ⎜⎜
⎣ ω P (z ) ω1 ⎦ ⎝ ω1 ⎟⎠
2α ⎡ −2 ⎤
=
2
π 2 n! L(1 − exp[− αL ]) πω P ( z )ω1
exp ⎢ 2 2 [
( x 2 − 2 xΔx + Δx 2 )ω12 + x 2ω P2 (z ) ⎥ ]
⎣ ω1 ω P ( z )
n 2 2
⎦
⎡ ⎛ ω 2 (z ) + ω 2 ⎞⎤ 2 ⎛ 2 x ⎞
× exp[− αz ]exp ⎢− 2 y 2 ⎜⎜ P 2 2 1 ⎟⎟⎥ H n ⎜
⎜
⎟
⎟
⎣⎢ ⎝ ω1 ω P (z ) ⎠⎦⎥ ⎝ ω1 ⎠
⎛ ⎡ ⎤ ⎞
⎜ exp ⎢ − 2 [( x 2 − 2 xΔx + Δx 2 )ω12 + x 2ω P2 ( z ) ⎥ ] ⎟
⎜ ⎣ ω1 ω P ( z ) ⎟
2 2
⎦
⎜ ⎟
⎜ ⎡ ⎛ ω P (z ) + ω1 ⎞⎛ 2
2 2
ω1 2
ω1 Δx ⎞⎤
2 2
⎟
⎜ = exp ⎢− 2⎜⎜ ω 2ω 2 ( z ) ⎟⎟⎜⎜ x − 2 xΔx ω 2 (z ) + ω 2 + ω 2 ( z ) + ω 2 ⎟⎟⎥ ⎟
⎜ ⎢⎣ ⎝ 1 P ⎠⎝ P 1 P 1 ⎠⎥⎦ ⎟
⎜ ⎡ ⎛ ω 2 (z ) + ω 2 ⎞ ⎡⎛ 2
⎤ ⎤ ⎟
⎜ ω 2
⎞ ω 4
ω 2
Δx 2
⎟
⎢ ⎜ ⎟ ⎜ ⎟ ⎥⎥
⎜ = exp ⎢− 2⎜ ω 2ω 2 ( z ) ⎟ ⎢⎢⎜ x − ω 2 (z ) + ω 2 Δx ⎟ − 2 Δx + 2
P 1 1 1 2 1
⎜ ⎣ ⎝ 1 P ⎠ ⎣⎝ P 1 ⎠ ω( P ( z ) + ω1)
2 2 ω P ( z ) + ω 2
1 ⎥
⎦⎦⎥ ⎟
⎟
⎜ ⎟
⎡ ⎛ ω 2 (z ) + ω 2 ⎞⎛ ⎞ ⎤
2
⎡ 1 ⎞⎤ ⎟
⎜ ω12 2⎛ ω12
= exp ⎢ − 2⎜ P 1
⎟⎜ x − Δ x ⎟ ⎥ exp ⎢ 2 Δ x ⎜ −
⎜ ω 2 ( z ) ω 2 ( z ) + ω 2 ω 2 ( z ) ⎟⎥ ⎟⎟
⎜
⎜
⎜ 2 2
⎢⎣ ⎝ ω1 ω P ( z ) ⎠⎝
⎟⎜ ω P2 ( z ) + ω12 ⎟⎠ ⎥
⎦ ⎣⎢ ⎝ P ( P )
1 P ⎠⎦⎥ ⎟
⎜ ⎟
⎡ ⎛ ω 2 (z ) + ω 2 ⎞⎛ ⎞ ⎤ 2 ⎛ ω1 − ω1 − ω P ( z ) ⎞
2
⎜ ω12 ⎡ 2 2 2
⎤ ⎟
= exp ⎢ − 2⎜ P 1
⎟⎜ x − Δ x ⎟ ⎥ exp ⎢ 2 Δ x ⎜
⎜ ω 2 ( z ) ω 2 ( z ) + ω 2 ⎟⎥ ⎟
⎜
⎜
⎜ ω 2ω 2 ( z ) ⎟⎜
⎢⎣ ⎝ 1 P ⎠⎝ ω P
2
( z ) + ω 2
1
⎟
⎠ ⎥⎦ ⎣⎢ ⎝ P ( P )
1 ⎠⎦ ⎥
⎟
⎟
⎜ ⎟
⎜ ⎡ ⎛ ω P (z ) + ω1 ⎞⎛
2 2
ω1 2
⎞
2
⎤ ⎡ 2Δx 2
⎤ ⎟
⎢ ⎜ ⎟⎜ ⎟ ⎥
⎜⎜ = exp − 2⎜ ω 2ω 2 ( z ) ⎟⎜ x − ω 2 ( z ) + ω 2 Δx ⎟ exp ⎢− ω 2 ( z ) + ω 2 ⎥ ⎟⎟
⎢⎣ ⎝ 1 P ⎠⎝ ⎠ ⎥⎦ ⎣ 1 ⎦
⎝ P 1 P
⎠
∴ S n.0 ( x, y, z )rP ( x, y, z )
2α 2
=
π 2 n n! L(1 − exp[− αL ]) πω P2 (z )ω12
⎡ ⎛ ω 2 (z ) + ω 2 ⎞⎛ ω12 ⎞ ⎤
2
× exp ⎢− 2⎜⎜ P 2 2 1 ⎟⎟⎜⎜ x − 2 Δx ⎟ ⎥

⎢⎣ ⎝ ω1 ω P (z ) ⎠⎝ ω P (z ) + ω12 ⎟⎠ ⎥
⎦
⎡ 2Δx 2 ⎤ ⎡ ⎛ ω 2 (z ) + ω 2 ⎞⎤ 2 ⎛ 2 x ⎞
× exp ⎢− 2 2 ⎥
exp[− αz ]exp ⎢− 2 y 2 ⎜⎜ P 2 2 1 ⎟⎟⎥ H n ⎜
⎜
⎟
⎟
⎣ ω P ( z ) + ω1 ⎦ ⎣⎢ ⎝ ω1 ω P ( z ) ⎠⎦⎥ ⎝ ω1 ⎠
⇒
γI 1
Pth (TEM n.0 ) = sat
η P L ∫∫∫ S n.0 ( x, y, z )rP ( x, y, z )dv
γI sat π 2 n n! L(1 − exp[− αL]) 1
=
ηP L 2α L
⎡ 2Δx 2 ⎤
∫0 Qn exp⎢⎣− ω P2 (z ) + ω12 − αz ⎥⎦ dz
πγI sat 1 − exp[− αL ] 2 n n!
=
2η P α L
⎡ 2Δx 2 ⎤
∫0 Qn exp⎢⎣− ω P2 (z ) + ω12 − αz ⎥⎦ dz
⎛ Qn ⎞
⎜ ⎟
⎜ 2 ⎡ ⎛ ω 2 (z ) + ω 2 ⎞⎤ ⎛ 2x ⎞ ⎟
⎜ exp ⎢− 2 y 2 ⎜⎜ P 2 2 1 ⎟⎟⎥ dy × H n2 ⎜⎜ ⎟
⎟ ⎟
⎜ πω P (z )ω1
2 2
⎣⎢ ⎝ ω1 ω P ( z ) ⎠⎦⎥ ⎝ ω1 ⎠ ⎟
⎜ = ∫∫ ⎟
⎜ ⎡ ⎛ ω 2 ( z ) + ω 2 ⎞⎛ ω12 ⎞ ⎤
2
⎟
⎜ exp ⎢ − 2 ⎜
⎜ 2 2
P 1
⎟⎜
⎟⎜ x − Δ x ⎟ ⎥ dx ⎟
⎜ ⎢⎣ ⎝ ω1 ω P ( z ) ⎠⎝ ω P2 (z ) + ω12 ⎟⎠ ⎥ ⎟
⎦
⎜ ⎟
⎜ 令 x′ = x − ω1 2
⎟
Δx
⎜ ω P2 ( z ) + ω12 ⎟
⎜ ⎟
⎜ 4 πω P2 (z )ω12 ⎛ 2x ⎞ ⎡ ⎛ ω 2 (z ) + ω 2 ⎞ ⎤ ⎟
⎜= ∫ H n2 ⎜⎜ ⎟ exp ⎢− 2⎜ P 2 2 1 ⎟( x ′)2 ⎥ dx ′
⎜
2
(
π ω P (z )ω1
2 2
) 2
( ( )
2 ω P z + ω1
2 2
) ω
⎝ 1 ⎠
⎟ ⎜ (
⎣⎢ ⎝ ω1 ω P z ⎠ ) ⎟
⎦⎥
⎟
⎟
⎜ ⎟
⎜= 2 1 ⎛ 2⎛ ω2 ⎞⎞ ⎡ ⎛ ω 2 (z ) + ω 2 ⎞ 2⎤ ⎟
∫ H n2 ⎜ ⎜⎜ x ′ + 2 1 2 Δx ⎟⎟ ⎟ exp ⎢− 2⎜⎜ P 2 2 1 ⎟⎟( x ′) ⎥ dx ′
⎜ π ω P2 ( z ) + ω12 ω1ω P ⎜
⎝ ω1 ⎝ ω P (z ) + ω1 ⎟
⎠⎠ ⎣⎢ ⎝ ω1 ω P ( z ) ⎠ ⎦⎥ ⎟
⎟
⎜
⎜ ⎟
⎝ ⎠
⎜ ∫
[ ]
⎛Q exp − x 2 H n2 ( x )dx = 2 n n! π ⎞
⎟
⎜ ⎛ ⎞ ⎟
⎜ ⎜ ⎟ ⎟
⎜ 2 ⎜ 2x′ 2 ⎟ ⎡ ⎛ ω P2 ( z ) + ω12 ⎞ 2 ⎤ ⎟
⎜Q ∫ H k ⎜ exp⎢− 2⎜⎜ 2 2 ⎟ ( x ′ ) ⎥ dx ′ ⎟
ω P2 ( z )ω12 ⎟ ⎢ ω ω ( z ) ⎟
⎠ ⎥⎦ ⎟
⎜ ⎣ ⎝
⎜⎜ 2 ⎟
1 P
⎜ ⎝ ω P ( z ) + ω1 ⎠
2 ⎟ ⎟
⎜ ⎟
⎜# g = 2 2 ′
∗ x → dg = 2 dx ⎟
′
⎜ ω P (z )ω12 ω P2 ( z )ω12 ⎟
⎜ ω P ( z ) + ω1
2 2
ω P ( z ) + ω1
2 2 ⎟
⎜ ⎟
⎜ ⎟
⎜ ⇒ ∫ H k ( g ) exp − g
2
[ 2
] 1
da ⎟
⎜ 2 2 ⎟
⎜ ω P (z )ω1 2
⎟
⎜ ω P2 ( z ) + ω12 ⎟
⎜ ⎟
⎜
⎜ =
1
∫ H k (g ) exp − g da
2 2
[ ] ⎟
⎟
2 2
⎜ ω P ( z )ω1 2 ⎟
⎜ ⎟
⎜ ω P2 ( z ) + ω12 ⎟
⎜ ⎟
⎜ = 1 ⎟
× 2 k! π
k
⎜ 2 2 ⎟
⎜ ω P ( z )ω1 2
⎟
⎜⎜ ⎟⎟
⎝ ω P2 ( z ) + ω12 ⎠
⇒
2 k k! ⎛ 2ω P2 ( z )
k
1 ⎞
Qn = 2 ⎜ ⎟⎟
ω P ( z ) + ω12 2 n ⎜⎝ ω P2 ( z ) + ω12 ⎠
2
⎡(n − k ) 2 ⎛ n ⎞ (k + 2 j )! ⎛ ω 2 ( z ) − ω 2 ⎞
j
⎛ 2ω Δx ⎞⎤
× ⎢ ∑ ⎜⎜ ⎟ ⎜ P 1
⎟⎟ H n − k − 2 j ⎜⎜ 2 1 ⎟⎟⎥
⎢ j =0 ⎝ k + 2 j ⎟⎠ k! j! ⎜⎝ ω P2 (z ) + ω12 ⎠ ⎝ ω P (z ) + ω1
2
⎠⎥⎦
⎣
Q ω P (z ) = ω P = aω1 δ = Δx ω
1
⎛ ω (z ) − ω
1. ⎜⎜
2 2
⎞
j
⎡ω a − 1 ⎤
⎟⎟ = ⎢
⎛ a2 −1⎞
2
( 2
) j j
⎥ = ⎜⎜ 2 ⎟⎟
P 1 1
⎝ ω (z ) + ω
2
P 1
2
⎠ ⎣ω a + 1 ⎦
2
⎝ a + 1⎠
2
1 ( )
⎛ 2ω Δx ⎞ ⎛ 2ω Δx ⎞ 2δ ⎞
2. H n − k − 2 j ⎜⎜ 2 1 ⎟⎟ = H n − k − 2 j ⎜⎜ 2 12 ⎟⎟ = H n − k − 2 j ⎛⎜ 2 ⎟
⎝ ω P ( z ) + ω1
2
⎠ ⎝ ω1 a + 1 ( ) ⎠ ⎝ a + 1⎠
⎛ 2ω 2 (z ) ⎞
k k k
⎛ 2a 2ω12 ⎞ ⎛ 2a 2 ⎞
3. ⎜⎜ 2 P ⎟ = ⎜⎜ 2 2 ⎟⎟ = ⎜⎜ 2 ⎟⎟
2 ⎟
⎝ ω P (z ) + ω1 ⎠ ⎝ ω1 a + 1 ( ) ⎠ ⎝ a +1⎠
4. ω P ( z ) + ω1 = ω1 a 2 + 1
2 2 2
( )
⎡ 2 Δx 2 ⎤ ⎡ − 2Δx 2 ⎤ ⎡ − 2δ 2 ⎤
5. exp ⎢− 2 − α z = exp − α z = exp − αz ⎥
⎣ ω P (z ) + ω1
2 ⎥
⎦
⎢ 2 2
⎣ ω1 a + 1
⎥
⎦
⎢ 2
⎣a +1 ⎦ ( )
NPth (TEM n.0 )
πγI satω12
= Pth (TEM n.0 )
2η P
1 − exp[− αL ] 2 n n! 1 1
=
α 1 ⎡ ⎛ 2δ ⎞⎤ ω1
L 2 2
Qn ∫ exp ⎢− ⎜⎜ 2 + αz ⎟⎟⎥ dz
⎣ ⎝ a +1 ⎠⎦
[ ]
0
2δ 2
1 − exp[− αL ] 2 n n! exp 1+ a2 1
=
α Qn L ω12
∫ exp [− α z ]dz
[ ]
0
(1 − exp[− αL]) exp 2δ

2
1+ a2 2 n n! 1
1
=
α −1
(exp[− αL] − 1) Qn ω1
2
α
= exp 2δ [ 2
] 2 n! 1
1+ a Q ω
2
n
n
2
1
REFERENCES
[1] T. Y. Fan, IEEE J. Quantum Electronics, 29, 1457-1459, 1993.
[2] Y. F. Chen, Y. P. Lan, S. C. Wang, Appl. Phys. B, 71, 827-830, 2000.
[3] Y. F. Chen, IEEE J. Quantum Electronics 35, 234-239, 1999.
[4] Y. F. Chen, J. Opt. Soc., 17, 1835-1840, 2000.
[5] Y. P. Lan, Y. F. Chen, S.C. Wang, Appl. Phys. B, 71, 27-31, 2000.
[6] Y. F. Chen, Y. P. Lan, S. C. Wang, J. Opt. Soc.Am. B, 25, 1016-1018, 2000.
[7] Y. F. Chen, H. J. Kuo, J. Opt. Soc. Am. B, 23, 846-848, 1998.
[8] W. Koechner, Appl. Opt., 9, 2548-2553, 1970.
[9] W. Koechner, D. K. Rice, IEEE J. Quantum Electronics, QE-6, 557-566, 1970.
[10] J. D. Foster, L. M. Osterink, Appl. Phys. Lett., 41, 3656-3663, 1970.
[11] L. M. Osterink, J. D. Foster, Appl. Phys. Lett., 12, 128-131, 1968.
[12] F. W. Quelle, Appl. Opt., 5, 633-637,1966.
[13] J. P. Gordon, R. C. C. Leite, R. S. Moore, S. P. S. Porto, J. R. Whinnery, Appl. Phys.
Lett, 36, 3-8, 1964.
[14] G. D. Baldwin, Appl. Phys. Lett, 38, 2726-2738, 1967.
[15] P. J. Hardman, W. A. Clarkson, G. J. Friel, M. Pollnau, and D. C. Hanna, IEEE J.
Quantum Electronics, 35, 647-655, 1999.
[16] J. J. Degnan, IEEE J. Quantum Electronics, 25, 214-220, 1989.
[17] J. J. Degnan, IEEE J. Quantum Electronics, 31, 1890-1901, 1995.
[18] J. J. Degnan, S. Member, D. B. Coyle, R. B. Kay, IEEE J. Quantum Electronics, 34,
887-899, 1998.
[19] A. Yariv, Optical Electronics in Modern Communications, Oxford University Press,
1997.
[20] F. D. Patel and R. J. Beach, IEEE J. Quantum Electronics, 37, 707-715, 2001.
[21] R. J. Beach, Opt. Commun., 123, 385-393, 1995.
[22] R. J. Beach, IEEE J. Quantum Electronics, 31, 1606-1613, 1995.
[23] C. Li, Q. Liu, M. Gong, G. Chen, P. Yan, Opt. Commun., 231, 331-341, 2004.
[24] J. Liu, C. Wang, Opt. Commun., 188, 155-162, 2001.
[25] Y. S. Huang and F. L. Chang, Opt. Rev., 12, 396-404, 2005.
[26] B. Ozygus and J. Erhard, Appl. Phys. Lett. 67, 1361-1362, 1995.
[27] B. Ozygus and Q. Zhang, Appl. Phys. Lett. 71, 2590-2592, 1997.
[28] L. M. Osterink and J. D. Foster, Appl. Phys. Lett., 12, 128-131, 1968.
[29] D. Marcuse, Light Tansmission Optics, topics: Wave optics of thin lens, Van Nostrand,
New York, 1982.
[30] W. Koechner, Appl. Opt., 9, 1429-1434, 1970.
[31] W. Koechner, J. Appl. Phys., 44, 3162-3170, 1973.
[32] M. E. Innocenzi, H. T. Yura, C. L. Fincher, and R. A. Fields, Appl. Phys. Lett. 56,
1831-1833, 1990.
[33] D. Y. Shen, H. R. Yang, J.G. Liu, S.C. Tam, Y.L. Lam, W.J. Xie, J.H. Gu, K. Ueda,
Appl. Phys. B, 72, 263-266, 2001.
[34] Y. S. Huang, H. L. Tsai, and F. L. Chang, Opt. Commun., 273, 515-525, 2007.
[35] J. Zheng, S. Zhao, and L. Chen, Optics and Laser Technology 34, 439-443, 2002.
[36] A. E. Siegman, Lasers , University Science Book, Mill Valley CA, 1986.
[37] D. C. Brown, IEEE J. Quantum Electronics, 33, 861-873, 1997.

[38] D. C. Brown, IEEE J. Quantum Electronics, 34, 2393-2402, 1998.
[39] P. Laporta and M. Brussard, IEEE J. Quantum Electronics, 27, 2319, 1991.
[40] Y. F. Chen, T. M. Huang, K. H. Lin, C. F. Kao, C. L. Wang, and S. C. Wang, Opt.
Commun., 136, 399-404, 1997.
[41] M. Arvidsson, Opt. Lett., 26, 15, 2001.
[42] Y. F. Chen and S. W. Tsai, IEEE J. Quantum Electron. 37, 580-586, 2001.
[43] Y. F. Chen, J. L. Lee, H. D. Hsieh, and S. W. Tsai, IEEE J. Quantum Electron. 38, 312-
317, 2002.
[44] K. Yang, S. Zhao, G. Li, and H. Zhao, IEEE J. Quantum Electron. 40, 1252-1257,
2004.
[45] G. Li, S. Zhao, K. Yang, and J. Zou, IEEE J. Selected Topic in Quantum Electron. 11,
638-644, 2005.
[46] K. Yang, S. Zhao, G. Li, D. Li, J. Wang, and J. An, IEEE J. Quantum Electron., 42,
683-689, 2006.
[47] D. Li, S. Zhao, G. Li, and K. Yang, IEEE J. Quantum Electron., 42, 500-508, 2006.
[48] K. Yang, S. Zhao, G. Li, D. Li, J. Wang, J. An, and M. Li, IEEE J. Quantum Electrons,
43, 109-115, 2007.
[49] M. Li, S. Zhao, K. Yang, G. Li, D. Li, J. Wang, J. An, and W. Qiao, IEEE J. Quantum
Electrons,. 44, 288-293, 2008.
[50] P. K. Mukhopadhyay, M. B. Alsous, K. Ranganathan, S. K. Sharma, P. K. Gupta, A.
Kuruvilla, and T. P. S. Nathan, Opt. and Laser Tech. 37, 157-162, 2005.
[51] Y. F. Chen and S. W. Tsai, Opt. Lett., 27, 397, 2002.
[52] Y. F. Chen, Y. S. Chen, and S. W. Tsai, Appl. Phys. B, 79, 207, 2004.
[53] S. Somekh and A. Yariv, Opt. Commun., 6, 301-304, 1972.
[54] V. G. Dmitriev, G. G. Gurzadyan, and D. N. Nikogosyan, Handbook of Nonlinear
Optical Crystal, Springer-Verlag, Berlin, 1999.
[55] K. Yang, S. Zhao, and G. Li, J. Opt. Quantum Electronics, 2005.
[56] K. Yang, S. Zhao, and G. Li, Electron. Lett., 41, 191, 2005.
[57] L. E. Myers, G. D. Miller, R. C. Eckardt, M. M. Fejer, and R. L. Byer, Opt. Lett., 20,
52-54, 1995.
[58] Y. C. Huang, A. C. Chiang, Y. Y. Lin, and Y. W. Fang, IEEE J. Quantum Electronis,
38, 1614-1618, 2002.
[59] D. D. Lowenthal, IEEE J. Quantum Electronics, 34, 1356-1366, 1998.
[60] W. Koechner, Solid-State Laser Engineering, Springer-Verlag, Berlin, 1999.
[61] W. H. Louisell, A.Yariv and A. E. Siegman, Phys. Rev. 124, 1646, 1961.
[62] A.Yariv, IEEE J. Quantum Electron. QE-1, 28, 1965.
[63] A.Yariv and W. H. Louisell, IEEE J. Quantum Electron QE-2, 418, 1966.
[64] Yariv, IEEE J. Quantum Electron. QE-2, 30, 1966.
[65] A.Yariv, J. Appl. Phys. 36, 388, 1965.
[66] Y. R. Shen, The Principles of Nonlinear Optics, John Wiley and Sons, 1991.
[67] E. S. Cassedy and M. Jain, IEEE J. Quantum Electron. QE-15, 1290-1301, 1979.
[68] M. Marte, H. Ritsch, L. Lugiato, and C. Fabre, Acta Physica Slovaca 47, 233-242,
1997.
[69] C. Schwob, P. F. Cohadon, C. Fabre, M. A. M. Marte, H. Ritsch, A. Gatti, and L.
Lugiato, Appl. Phys. B, 66, 685-699, 1998.
[70] A. Fix and R. Wallenstein, J. Opt. Soc. Am. B 13, 2484-2497, 1996.
[71] G. Arisholm, J. Opt. Soc. Am. B 16, 117-127, 1999.
[72] S. C. Lyons, G. L. Oppo, W. J. Firth, and J. R. M. Barr, IEEE J. Quantum Electron., 36,
541, 2000.
[73] R. G. Smith, IEEE J. Quantum Electron. QE-9, 530-541, 1973.
[74] J. E. Bjorkholm, A. Ashkin and R.G. Smith, IEEE J. Quatum Electron. QE-6, 797-799,
1970.
[75] J. A. Giordmaine and R. C. Miller, Phys. Rev. Lett. 14, 973-976, 1965.
[76] J. Falk, IEEE J. Quantum Electron. QE-7, 230-235, 1971.
[77] J. Falk, J. M. Yarborough, and E.O. Ammann, IEEE J. Quantum Electron. QE-7, 359-
369, 1971.
[78] S.E. Harris, Proc. IEEE , 57, 2096-2113, 1969.
[79] S. Guha, F. J. Wu, and J. Falk, IEEE J. Quantum Electron. QE-18, 907-912, 1982.
[80] J. E. Bjorkholm, IEEE J. Quantum Electron., QE-7, 109-118, 1971.
[81] S. Schiller, K. Schneider, and J. Mlynek, J. Opt. Soc. Am. B 16, 1512-1523, 1999.
[82] M. F. Becker, D. J. Kuizenga, D. W. Phillion, and A. E. Siegman, J. Appl. Phys., 45,
399-4005, 1974.
[83] A. V. Smith, R. J. Gehr, and M. S. Bowers, J. Opt. Soc. Am. B 16, 609-619, 1999.
[84] A. V. Smith, W. J. Alford, T. D. Raymond, and M. S. Bowers, J. Opt. Soc. Am. B 12,
2253-2267, 1995.
[85] C. Drag, I. Ribet, M. Lefebvre, and E. Rosencher, Appl. Phys. B, 73, 195-200, 2001.
[86] X. Peng, L. Xu, and A. Asundi, IEEE J. Quantum Electronics, 41, 53-61, 2005.
[87] Q. Peng, X. Yang, X. Wang, A. Geng, A. Yao, and Z. Xu, Opt. Rev. 12, 307-312, 2005.
[88] A. Godard and E. Rosencher, IEEE J. Quantum Electronics, 40, 784-790, 2004.
[89] J. A. Armstrong, N. Bloembergen, J. Ducuing, and P. S. Pershan, Phys. Rev. 127, 1918-
1939, 1962.
[90] R. Haidar, N. Forget, and E. Rosencher, IEEE J. Quantum Electronics, 39, 569, 2003
[91] A. V. Smith, D. J. Armstrong, M. C. Phillips, R. J. Gehr, and G. Arisholm, J. Opt. Soc.
Am. B 20, 2319-2327, 2003.

[92] D. J. Armstrong and A. V. Smith, IEEE J. Selected Topic in Quantum Electronics, 13,
721-731, 2007.
[93] A. C. Chiang, T. D. Wang, Y. Y. Lin, C. H. Lau, Y. H. Chen, B. C. Wong, Y. C.
Huang, J. T. Shy, Y. F. Chen, and P. H. Tsao, IEEE J. Quantum Electronics, 40, 791-
799, 2004.
[94] R. L. Byer, IEEE J. Select. Top. Quantum Electronics, 6, 911-930, 2000.
[95] M. M. Choy and R. L. Byer, Phys. Rev. B 14, 1693-1706, 1976.
[96] W. R. Trutna, Y. K. Park, and R. L. Byer, IEEE J. Quantum Electronics, QE-15, 648-
655, 1979.
[97] S.E. Brosnan and R.L. Byer, IEEE J. Quantum Electronics, QE-15, 415-431, 1979.
[98] R.A. Baumgartner and R.L. Byer, IEEE J. Quantum Electronics, QE-15, 432-444,
1979.
Chapter 2
LASER ANNEALING OF COMPOSITE MATERIALS

WITH METAL NANOPARTICLES
Andrey L. Stepanov∗
Laser Zentrum Hannover, 30419 Hannover, Germany
Kazan Physical-Technical Institute, Russian Academy of Sciences,
420029 Kazan, Russian Federation
ABSTRACT
Recent results on the interaction of high power laser pulses with various dielectric
materials containing metal nanoparticles are reviewed. Original results together with new
publications are observed. In general, the excimer laser pulse modification of silver and
copper nanoparticles synthesized by ion implantation in silicate glasses and sapphire are
considered. One of features of composite samples prepared by the low energy ion
implantation is the growth of metal particles with a wide size distribution in the thin
depth from the irradiated substrate surface. Pulsed laser irradiation makes it possible to
modify such composite layer, improving the uniformity in the size distribution of the
nanoparticles. Changes induced by pulsed laser exposure suggest there are both
reductions in average size of the metal nanoparticles, and some long-range dissolution of
metal atoms in the matrix. Experimental data on laser modification are explained by
photofragmentation and melting of the nanoparticles in the dielectric matrix.
Combination of ion implantation and laser annealing is promising technology for
fabrication of novel composite optical materials.
INTRODUCTION
Composite materials, such as dielectrics with embedded metal nanoparticles (MNPs), are
promising optoelectronic materials. An example of their application in optoelectronics is a
prototype of integrated electronic circuit - chip that combines metallic wires as conductors of
∗
e-mail: a.stepanov@lzh.de ; anstep@kfti.knc.ru
28 Andrey L. Stepanov
electric signals with fibers as guides of optical signals. In practice, light guides are frequently
made of synthetic sapphire (Al2O3) or silicon oxide (SiO2), which are deposited on or buried
in semiconductor substrates. In this case, electrooptic emitters and that accomplish electric-to-
optic signal conversion are fabricated inside the dielectric layer. This light signal from a
microlaser is focused in a light guide and then transmitted through the optoelectronic chip to a
high-speed photodetector, which converts the photon flux to the flux of electrons. It is
expected that light guides used instead of metallic conductors will improve the data rate by at
least two orders of magnitude. Moreover, there is good reason to believe that optical guide
elements will reduce the energy consumption and heat dissipation, since metallic or
semiconductor components of the circuits may be replaced by dielectric ones in this case.
Prototype optoelectronic chips currently available are capable of handling data streams with a
rate of 1 Gbit/s, with improvement until 10 Gbit/s in future.
Key elements of dielectric waveguides used for light propagation are nonlinear optical
switches, which must provide conversion of laser signal for pulse duration as short as pico- or
femtoseconds.
Figure 1. A prototypes of optoelectronic chip with a dielectric waveguide combined with silicone
substrate. Ion implantation can be applied to fabricate selective area doped by rear metal ions (marked
by stars) to work as microlaser and to illuminate in waveguide, created by rear-gas ion radiation with
MNPs to form an optical switcher.
Laser Annealing of Composite Materials with Metal Nanoparticles 29
The nonlinear optical properties of MNP-containing dielectrics stem from the dependence
of their refractive index and nonlinear absorption on incident light intensity [1, 2]. This effect
is associated with MNPs, which exhibit an enhancement of local electromagnetic field in a
composite and, as consequence, a high value of the third order nonlinear susceptibility when
exposed to ultrashort laser pulses. Therefore, such MNP-containing dielectric materials may
be used to advantage in integrated optoelectronic devices, for example, as shown in Figure 1.
It is well known [1] that a local field enhancement in MNPs stimulates a strong linear
optical absorption called as surface plasmon resonance (SPR). The electron transitions
responsible for plasmon absorption in MNPs cause also a generation of an optical
nonlinearity of a composite in the same spectral range. As a result, the manifestation of
nonlinear optical properties is most efficient for wavelengths near the position of a SPR
maximum. In practice, to reach the strong linear absorption of a composite in the SPR
spectral region, attempts are made to increase the concentration (filling factor) of MNPs.
Systems with a higher filling factor offer a higher nonlinear susceptibility, when all other
parameters of composites being the same. Usually noble metals and copper are used to
fabricate nonlinear optical materials with high values of third order susceptibility.
There are variety ways to synthesis MNPs in dielectrics, such as magnetron sputtering,
the convective method, ion exchange, sol–gel deposition, etc. One of the most promising
enhanced fabrication methods is ion implantation [2-7] because it allows reaching a high
metal filling factor in an irradiated matrix beyond the equilibrium limit of metal solubility and
provides controllable synthesis of MNPs at various depths under the substrate surface. Nearly
any metal–dielectric composition may be produced using ion implantation. This method
allows for strict control of the doping ion beam position on the sample surface with implant
dose as, for example, in the case of electron- and ion-beam lithography.
Today, ion implantation is widely used in industrial semiconductor chip fabrication.
Therefore, the combination of MNP-containing dielectrics with semiconductor substrates by
same technological approach as ion implanattaion could be reached quite effective. Moreover,
ion implantation can be applied for different steps in optoelectronic material fabrication such
as creation of optical waveguides by implantation with rear gas ions (H+, He+ etc.) [6], a
designing of electric-to-optic signal convectors and microlaser by irradiation of dialectics
waveguides with rear metal ions (Er+, Eu+ etc.) [6, 8] and a synthesis of MNPs (Figure 1).
The history of MNP synthesis in dielectrics by ion implantation dates back to 1973, when
a team of researchers J. Davenas et al. at the Lyons University (France) pioneered this
method to create particles of various metals (silver, sodium, calcium, etc.) in LiF and MgO
ionic crystals [9, 10]. First work on ion-synthesis of noble nanoparticles was done in 1975 by
Arnold in his study of Au-irradiated glasses [11]. Later developments have expanded from the
metal implants to the use of many ions and the active formation of compounds, including
metal alloys and totally different composition precipitate inclusions. In ion implantation
practice MNPs were fabricated in various materials, such as polymers, glass, artificial
crystals, and minerals [12]. Despite these steady advantages the use of ion implantation for
nanoparticle synthesis there have not yet emerged clear mechanisms which allow precisely
controlled particles sizes and depth distributions. Latter has a certain drawback which is the
statistically nonuniform depth of penetration of implanted ions into a material [13, 14]. This
leads to a wide size distribution of synthesized nanoparticles not only in the plane parallel to
the irradiated surface but to a great extent also over the depth of the sample [15, 16].
Dispersion of nanoparticles with respect to sizes leads to a broadening of the SPR optical
absorption band accompanied by a decrease it in the intensity [17]. This is also attributable to
the dependence of the SPR spectral position on the particle size, i.e. the absorption spectrum
in real sample is a superposition of several overlapping less intense bands that corresponding
to particles of various sizes. The concern of the modern task is to increase the uniformity of
the size distribution of MNPs synthesized by ion implantation using an approach of high-
power pulse laser annealing with sequential furnace one. Experience gained from using the
laser annealing techniques for various purposes allowed MNPs to be modified in various solid
state dielectrics.
In general, the interaction of the high-power laser pulses with dielectrics containing
MNPs for modification of their optical properties has been the subject of intense investigation
for the last few years. This interest was conditioned by both a development of
microelectronics based on composites with nanoparticles and also on a need for a
fundamental understanding of the influence of high-power light on properties of the non-
uniform composite media. In particular, such researches are directed to the changes of the
shape, the size, the size distribution, and the structure of the nanoparticles. The table 1
presents data, compiled from the available publications [18-61], on the types of MNPs in
various matrixes and laser conditions for their changes. As seen from the table 1, depending
on the laser wavelength, the power density or the pulse length, composite materials made by
ion implantation can be successfully modified. For example, Nd:YAG laser irradiation in the
visible and near-IR region converts implanted spherical titanium silicide particles into
ellipsoidal ones in silica and soda-lime glass [61], or that there are changes of the internal
crystallographic structure of implanted iron particles after ruby laser (690 nm) treatment [19].
The main feature most of all the experiments described in the table is that the laser light
was applied directly into the spectral region of the transparency of the dielectric matrix, and
consequently, the intense laser pulses were primarily absorbed by the metal particles
[18, 19, 27, 28-31, 34-36, 42-50, 55, 57, 60]. Contrary to that, some years ago in 1993 by
Wood et al. [23] a new approach for annealing was demonstrated, when float soda-lime glass
with silver particles was irradiated by a laser light at wavelengths of glass absorption in the
ultraviolet region. When applying high-power excimer ArF (193 nm) laser pulses, a decrease
of the reflectance intensity of composite samples was observed. It was suggested that the
implanted silver particles in glass can be dissolved and the glass matrix can be modified to be
a silver rich metastable new glass phase. If this is correct then the new phase will be the
potential to be destabilized to precipitate out the new silver particles in a controlled fashion
by furnace. In reality, this new technique gives a variety of possibilities for controlled change
of optical properties of implanted dielectrics, as suggested for the case of the combined laser
annealing of float soda-lime glass implanted with metal ions [37]. However, a number of
questions regarding mechanisms of such modifications and interactions with high-power
excimer laser light with non-uniform composite materials are now just starting to be
understood. The literature of the field is expanding rapidly and the references cited here are
mostly limited to amorphous dielectrics, but with a rapid growth in the range of potential
applications there are a much wider interests of interests in target materials and types of
inclusions.
High-Power and Femtosecond Lasers: Properties, Materials and Applications : Properties, Materials and Applications, Nova Science Publishers, Incorporated,
Table 1. Laser annealing of MNPs in solid dielectric matrix

(ASG – aluminosilicate glass, SLSG – soda-lime silicate glass). SPR – surface plasmon resonance, τ - laser pulse duration
Types of metal Dielectric Synthesis method Laser Laser Main results Authors
nanoparticles matrix of nanoparticles parameters annealing of laser annealing
K MgO Ion implantation and YAG (1000 nm) laser Irradiation close to the Exfoliation of the Rankin et al. 1992 [18]
[100] thermal annealing τ=20 ns, energy per pulse SPR spectral area of surface layer with
10 mJ nanoparticles but of particles from sample
matrix transparence
K MgO Ion implantation and N2 (350 nm) laser Irradiation in spectral There are no Rankin et al. 1992 18]
[100] thermal annealing τ=20 ns energy per pulse area of matrix recognized changes in
10 mJ transparence the sample
α-Fe SiO2 Ion implantation Rubin (690 nm) laser Irradiation in the are of Ordering of the Bukharaev et al. 1991
glass τ=40 ns, energy density matrix transparence crystal structure of [19]
0.1 - 0.3 J/cm2 the particles
Cu Al2O3 Ion implantation KrF (248 nm) excimer Irradiation in the Reduction of the Stepanov et al. 2001
crystal laser, τ=25 ns, 1 Hz absorption bands of particles size [20]
energy density 0.3 J/cm2 matrix Stepanov et al. 2002
[21]
Stepanov 2005 [22]
Stepanov and Khaibullin
2005 [23]
Cu SiO2 Ion implantation KrF (355 nm) excimer Irradiation in the Reduction of the Stepanov et al. 2000
SLSG laser, τ=25 ns, 1 Hz absorption bands of particles size and [24]
glasses energy density 0,2-0.32 matrix oxidation of Cu Stepanov et al. 2002
J/cm2, 1-250 pulses particles [25]
Stepanov and Hole 2002
[16]
Cu SiO2 Ion implantation Nd:YAG (248 nm) laser, Irradiation in the Grow of the particless Masuo et al. 2006 [27]
τ=20 ns, 10 Hz absorption bands of
energy density 0,2 J/cm2, matrix
1-250 pulses
Table 1. (Continued)
Cu SiO2 Ion implantation Nd:YAG (532 nm) laser, Irradiation close to the Dissolution of the Masuo et al. 2006 [27]
τ=20 ns, 10 Hz SPR spectral area of particles
energy density 0,2 J/cm2, nanoparticles but of
1-250 pulses matrix transparence
Cu Polymer Simultaneous Nd:YAG (1064 nm) laser Irradiation in spectral Changes of particle Werner et al. 1994 [28]
plasma in the cw mode, the power area of matrix size and shape by Heilmann et al. 1995
polymerization and from 8.6 to 1.1 W transparence coalescence [29]
metal evaporation Heilmann 2003 [30]
Ag ASG Ion-exchange O-switched Nd:YAG (532 Irradiation close to the Photochromic effect – Akella et al. 1997 [31]
glass nm) laser, τ=10 ns, 10 Hz, SPR spectral area of depletion of metallic
energy density 0.18 and nanoparticles but of particles by
0.6 J/cm2, 100 – 1000 matrix transparence photoionization.
pulses
Ag SiO2 Ion implantation ArF (193 nm) excimer Irradiation in the Dissolution of Wood et al. 1993 [32]
glass laser, 500 pulses, 10 Hz absorption bands of particles into glass Townsend and Olivares
matrix network or a 1997 [33]
separation particles
into smaller ones
Ag SiO2 Co-evaporation Argon and krypton lasers Irradiation close to the Increasing of the Ferrari et al. 1995 [34]
glass metal and glass (350 nm), power SPR spectral area of particles size
0.24-1 W, 5 min nanoparticles
Ag SiO2 Ion-exchange O-switched Nd:YAG (532 Irradiation close to the Increasing of the Gonella et al. 1996 [35]
SLSG and 1064 nm) laser, τ=10 SPR spectral area of particles size and
glasses ns, 10 Hz, energy density nanoparticles but of formation of new
0.3 and 5.0 J/cm2 matrix transparence ones
Ag SiO2 Ion implantation Nd:YAG (532 nm) laser, Irradiation close to the Dissolution of Townsend and Olivares
glass , energy density 0.075 and SPR spectral area of particles into glass 1997 [33]
0.1 J/cm2 nanoparticles but of network or a
matrix transparence separation particles
into smaller ones
Ag SiO2 Ion-exchange O-switched Nd:YAG (532 Irradiation close to the Reduction of the Gonella et al. 1996 [35]
SLSG followed by ion and 1064 nm) laser SPR spectral area of particles size as Battaglin et al. 2003 [36]
glasses irradiation τ=10 ns, 10 Hz, energy nanoparticles but of fragmentation and
density 0.3 and 5.0 J/cm2 matrix transparence formation of Ag2O
shells on Ag particle
core
Ag SiO2 Ion implantation KrF (248 nm) excimer Irradiation in the Reduction of the Stepanov et al. 1998
SLSG laser, τ=25 ns, 1 Hz absorption bands of particles size [37]
glasses energy density 0.2-0.25 matrix followed with Stepanov et al. 2000
J/cm2, 1-250 pulses regrowth of new [24]
particles Stepanov et al. 2002
[38]
Stepanov 2003 [39]
Stepanov and Popok
2004 [40]
Stepanov et al. 2008
[41]
Ag SLSG Ion-exchange Ti:sapphire (400, 500, 550 Irradiation close to the Changes of particle Kaempfe et al. 1999 [42]
glasses nm) laser, τ=<150 fs, 1 SPR spectral area of shape from spherical Kaempfe et al. 2000 [43]
Hz, pulse intensity nanoparticles but of to ellipsoidal with an Kaempfe et al. 2001 [44]
1011 W/cm2 matrix transparence additional halo of Seifert et al. 2000 [45]
linear polarization small particles around Podlipensky et al. 2005
the central one [46]
Stalmashonak et al.
2007 [47]
Ag SiO2 Magnetron Nd:YAG (266 and Irradiation in the Reduction of the Sendova-Vassileva et al.
glass co-sputtering 532 nm) laser, τ=5 ns, absorption bands of particles size 2005 [48]
10 Hz energy density matrix and close to SPR followed their Sendova et al. 2006 [49]
0.35- 2.0 mJ/cm2 area of particles regrowth
Ag SiO2 Ion implantation XeCl (308 nm) excimer Irradiation in the Dissolution of Crespo-Sosa et al. 2007
laser, τ=55 ns, absorption bands of particles into glass [50]
density 0.21-4.0 J/cm2, matrix network or a
10-200 pulses separation particles
into smaller ones
Ag SiO2 Ion implantation Free electron laser (8 μm), Irradiation in the Dissolution of Halabica et al. 2007 [51]
τ=350 ps, 30 Hz, absorption bands of particles into glass
energy 3-10 mJ matrix in infrared area network or a
separation particles
into smaller ones
Ag SiO2 Ion implantation KrF (248 nm) excimer Irradiation in the Redaction of particle Tsang et al. 2008 [52]
laser, τ=25 ns, 1 Hz absorption bands of size and their
energy density 0.25 J/cm2, matrix coalescence followed
5-20 pulses with decreasing of
particle concentration
Ag Polymer Simultaneous Nd:YAG (1064 nm) laser Irradiation in spectral Changes of particle Werner et al. 1994 [28]
plasma in the cw mode, power area of matrix size and shape by Heilmann et al. 1995
polymerization and from 8,6 to 1,1 W transparence coalescence [29]
metal evaporation Heilmann 2003 [30]
Au SiO2 Ion implantation XeCl (308 nm) excimer Irradiation in the Dissolution of Crespo-Sosa et al. 2007
laser, τ=55 ns, absorption bands of particles into glass [50]
density 0.21-4.0 J/cm2, matrix network or a
10-200 pulses separation particles
into smaller ones
Au SiO2 Ion implantation Free electron laser (8 μm), Irradiation in the Increasion of Halabica et al. 2007 [51]
τ=350 ps, 30 Hz absorption bands of nanoparticle size Halabica et al. 2008 [53]
energy 3-10 mJ matrix in infrared area
Au Polymer Simultaneous Argon (514 nm) laser Irradiation close to the Changes of particle Comita et al. 1994 [54]
plasma power 5 W SPR spectral area of shape from irregular
polymerization and nanoparticles but of to spherical
metal evaporation matrix transparence and to large size by
coalescence and
recrystallisation
Au Polymer Simultaneous Ti:sapphire (800 nm) Irradiation close to the Changes of particle Kaempfe et al. 2001 [55]
plasma laser, τ=<150 fs, 1 Hz, SPR spectral area of shape from irregular
polymerization and energy nanoparticles but of to spherical
metal evaporation 100 - 700 μJ matrix transparence and to large size by
coalescence and
recrystallisation
Au Polymer Chemical prparation Ti:sapphire (850 nm) Irradiation close to the Estimated a damage Zijlstra et al. 2007 [56]
laser, τ=100 fs, 82 MHz SPS spectral area of threshold of the
nanoparticles but of composite
matrix transparence
Bi α-SiO2 Ion implantation KrF (248 nm) excimer Irradiation in the Formation of Bi2O3 Park et al. 1998 [57]
crystal laser, τ=20 ns, 10 Hz absorption bands of and removal of
energy density 0.15 J/cm2 matrix particles from the
sample surface
Er Al2O3 Ion implantation ArF (193 nm) excimer Irradiation in the Dissolution of Can et al. 1994 58]
crystal and thermal laser, τ=12 ns, 1 or 5 Hz absorption bands of particles into matrix Can et al. 1995 [59]
annealing air energy density 0.15-0.4 matrix network Townsend and Olivares
J/cm2 1997 [33]
AuCu SiO2 Ion implantation KrF (248 nm) excimer Irradiation close to the Increasing of the Battaglin et al. 2001 [60]
glass and thermal laser, τ=20 ns SPR spectral area of alloy particles size
annealing in gas H2- energy density 1 J/cm2 nanoparticles but of
Ar (5% H2) matrix transparence
AuAg SiO2 Ion implantation KrF (248 nm) excimer Irradiation close to the Increasing of the Battaglin et al. 2001 [60]
glass and thermal laser, τ=20 ns, SPR spectral area of alloy particles size
annealing in gas H2- energy density 1.3 J/cm2 nanoparticles but of
Ar (5% H2) matrix transparence
Ti2S SiO2 Ion-exchange O-switched Nd:YAG (532 Irradiation in the are of Change of particle Battaglin et al. 1998 [61]
SLSG and 1064 nm) laser, τ=10 matrix transparence shape from spherical
glasses ns, 10 Hz, energy density to ellipsoidal and
0.04 and 0.6 J/cm2 increasing Ti in
silicide particles
ION SYNTHESIS OF METAL NANOPARTICLES IN DIELECTRIC MATRIX

BY LOW ENERGY ION IMPLANTATION
Ion implantation is an effective technological tool for introducing single impurities into
the surface layer of the substrate to a depth of several micrometers [6]. The degree of surface
modification of the materials depends on their individual chemical and structural properties,
as well as on variations of implantation parameters, such as the type and energy of an implant,
current density in ion beam, substrate temperature, etc. A most critical parameter is ion dose
F0, which determines the implant amount. Depending on the modification of dielectrics by
irradiation, ion implantation can be conventionally divided into low-dose and high-dose
processes (Figure 2).
In the case of low-dose irradiation (~F0 ≤ 5.0·1014 ion/cm2), the ions implanted, after
stopping and thermalization, are dispersed throughout the volume of the dielectrics and are
well separated from each other. The energy of the implant is transferred to the matrix via
electron shell excitation (ionization) and nuclear collisions. This causes radiation-induced
defects, which, in turn, may reversibly or irreversibly modify the material structure [6].
Various types of crystal structure damage have been observed in practice : extended and point
defects, amorphization and local crystallization, precipitation of a new phase made up of host
atoms or implanted ions, etc. In addition, ion implantation may be accompanied by the
effective sputtering of the surface of the dielectric target, as, for example, observed in the case
of Cr+ and Ti+ implantation into Al2O3 by photon tunnel microscopy (Figure. 3) [62], or Cu+
implantation into the MgAl2O3 spinel matrix and SiO2 glass [63]. Sometimes, sputtering
competes with swelling, as in the case of spinel and glasses exposed to extremely high current
beams of metal ions [63], in the case of inert gas ions embedded in polymers [64] or some
semiconductors irradiated by various ions with high doses [65].
Figure 2. Basic physical processes (from left to right) involved in the formation of nanoparticle from an
implant vs. the ion dose with regard to surface sputtering under irradiation.
The range of high-dose implantation may be divided into two characteristic dose
(irradiation time) sub-ranges (Figure 2).
Figure 3. Three-dimensional photon tunnel microscope image of sapphire implanted with

3.0·1017 ions/cm2 at 160 KeV, at edge of mask showing sputtered region. Fragment of image is taken
with contrast correction from [62].
Figure 4. Micrograph of silver nanoparticles produced by 160-keV Cu+ implantation into SiO2 at a dose
of 6.0·1016 ion/cm2; fragment of picture from [66].
In the range 1015 ≤ ~F0 ≤ 1017 ion/cm2, the implant concentration exceeds the solubility
limit of metal atoms in dielectrics (in particular, in sapphire) and the system relaxes by
nucleation and growth of MNPs, as illustrated in SiO2 glass (Figure 4 [66]). The threshold
dose value (at which MNPs nucleate) depends on the type of the dielectric matrix and
implant. For example, for 25-keV Ag+-ion implantation into LiNbO3, the threshold dose was
found to be F0 ~ 5.0·1015 ion/cm2 [67], for 30-keV silver ions embedded in epoxy resin, F0 ~
1016 ion/cm2 [68], but a formation of ultra-dispersed magnetic phase in polyimide was found
to form under fluencies as high as ≈ 1017 ion/cm2 at 40-keV iron ions irradiation [69].
Figure 5. Micrograph of cobalt labyrinth structures produced by Co+ implantation into epoxy resins at a
dose of 2.5·1017 ion/cm2 [70].
The next subrange of high-dose implantation, ∼F0 ≥1017 ion/cm2, leads to the coalescence
of already existing MNPs with the formation of either MNP aggregates or thin quasi-
continuous films at the dielectric surface (Figure 4). For instance, the irradiation of epoxy
resin by 40-keV cobalt ions at higher-than threshold-nucleation doses favors the formation of
thin labyrinth structures (Figure 5) [70]. The MNP distribution established in the dielectrics
after coalescence or Ostwald ripening may be dramatically disturbed by postimplantation
thermal or laser annealing [6].
INTERACTION OF HIGH-POWER EXIMER LASER PULSES WITH

SODA-LIME SILICATE GLASS CONTAINING ION-SYNTHESISED
SILVER NANOPARTICLES
The works considered in this review were aimed at studying composites consisting of the
dielectrics with isolated metal nanoparticles; i.e., the particles were synthesized by high-dose
(F0 ~ 1017 ion/cm2) ion implantation. Lets start with short characterisation of glasses with
implanted nanoparticles. As it was mentioned in introduction, low energy implantation of
metal ions, as in our case, leads to a non-uniform distribution in the depth of the glass, which
is different from the Gaussian profile traditionally predicted by statistical theory, primarily
because of the effective sputtering from the substrate surface during the irradiation. As seen
in Figure 6 the silver depth concentration in the soda-lime silicate glass (SLSG) derived from
experimental Rutherford backscattering spectrometry (RBS) spattering shows a maximum
near the implanted surface of the sample with some penetration to about 60-65 nm [71]. This
characterises the implanted sample as a depth non-uniform metal/insulator object.
Figure 6. The depth distribution of silver derived from the RBS spectrum for ion implantation with a
dose of 7·1016 ion/cm2 and an energy of 60 keV into the SLSG [71].
An excess of metal atoms in the glass, above the solubility limit, causes nucleation of the
nanoparticles. As usually the size distribution of MNPs appears. Additional to this
distribution in the implanted sample there is distribution particles in the depth. Since the
increase of metal concentration in the depth profile and the sputtering yield depend on
implantation time, then the metal particle nucleation and growth will also vary with time and
depth. It is obvious that during the implantation the size of the growthing particles with depth
is “proportional” to the metal filling factor, because they are both determined by the ion
concentration profile. Consequently, according to the Figure 6 the large silver nanoparticles
(or/and the higher filling factor) in the glass are close to the implanted surface with small ones
in the interior of the implant zone. Similar results were observed by electron microscopy
experiments on Ag-ion implanted SLSG [15] and Co-ion implanted silica [72]. These features
can be recognised in optical spectra of dielectrics with implanted nanoparticles. As example,
transmittance and reflectance data of the non-implanted and implanted glass are presented in
Figure 7 [16]. The transmission spectrum of implanted glass is maximised near 430 nm and
the shape of the spectral curve is almost symmetrical. The reflectance spectra of same sample
are more complex and, although the transmission must be the same whether the glass is
viewed from the implanted or the reverse face, the shapes of the reflectivity curves differ.
Overlapping peaks of reflectance spectra measured from the implant face of the samples
exhibit a shoulder at about 430 nm, on the side of a clearly determined maximum at 490 nm,
whereas reflectivity from the rear face appears to have a simpler peak at longer wavelengths
near 500 nm. The contrast between the spectrum available from transmittance and reflectance
of the implanted glass is recognised as coming from the non-uniform silver profile and size
distribution of the nanoparticles in the implanted glass (Figure 6). Since the distribution
pattern is not symmetric, the reflectivity determined from the ion implanted and rear faces of
the sample differ (Figure 7), and the reflected intensity from a particular layer depends on the
their local absorption in the depth [16].
Figure 7. Optical transmittance and reflectance of the silver implanted and original soda-lime silicate
glass with a dose of 7·1016 ion/cm2 and an energy of 60 keV. Reflectance was measured from both the
implanted and rear faces of the sample [16].
The atomic force microscope (AFM) images of the non-implanted and implanted surfaces
of the samples are shown in Figures 8 and 9 [14, 37]. As seen from these figures, compared
with the initial glass, the implanted surface is smoother (roughness) that is typical for
irradiated glasses at similar conditions. However, there are many hemispherical hills on this
surface with an average diameter of approximately 100-150 nm. There are no such
protrusions on the unimplanted sample. The reason for the existence of surface hills is
assumed to be from a sputtering of irradiated glass during implantation, which leads to
unequal ejection of ions of different elements from the surface, exposing the synthesised
nanoparticles in the sub-surface glass. The sputtered glass thickness is typically of the order
of tens of nanometers for the present ion dose [73], and hence, the synthesised buried MNPs
appear near to the glass surface in the implanted sample (Figure 6).
Consider the excimer laser pulse (25 ns) interaction with Ag-implanted SLSG. In general
the mechanisms of laser processing depend on the parameters of the laser beam and the
physical-chemical properties of the material. The primary interactions between light and
matter are non-thermal and lead to an electronic or vibrational photoexcitation. A prediction
of how the laser processing will develop through thermal (photothermal) or non-thermal
(photochemical) processes depends particularly on the values of the relaxation times
involved. In laser processing for the present composite materials, the relevant excitations and
conversions can be classified into those of the glass-substrate and the metal nanosize
inclusions.
Figure 8. An AFM image as a top view of the surface of SLSG before any treatment [14].
Figure 9. An AFM image as a top view under lateral illumination of the surface of SLSG Ag-implanted
with a dose of 7·1016 ion/cm2 and an energy of 60 keV. The step along the X and Y axes id 100 nm, and
the step along the Z axis is 3 nm [37].
Irradiation with a high power excimer laser at the wavelength of 248 nm (5 eV) couples
directly through the spectral absorption of the SLSG, as the glass has a smaller energy
bandgap (~ 3.5 eV). This creates electron-hole pairs by exciting an electron from the valence
into the conduction band. In metals, both conduction- and valence-band electrons may
participate in laser excitations. The number and energy of absorbed photons immediately
establishes the temperature rise in a laser-irradiated metal, whereas the number of valence-to-
conduction-band excitations establishes the initial density of electron-hole pairs in an
insulator [74]. In a metal, the time between electron-electron collisions is of the order of 10-14
to 10-13 s and electron-phonon relaxations are typically one to two orders of magnitude slower
[75]. In non-metals, interband electronic excitations range from 10-12 to 10-6 s [74]. Overall,
most of these times are short in comparison to the pulse duration (25 ns) of the laser described
here, and so the laser energy can be viewed as relaxing directly into a heating of the
metal/glass composite. Thus, our following consideration of the laser treatments will be based
on the effects of the thermalization processes.
Consequences of the excimer laser pulse with nanosecond pulse width and high beam
intensity are heating, melting and/or vaporisation (ablation) of material on a time scale of
nanoseconds to microseconds. The excimer laser treatment has been applied to many glasses,
for example, [76, 77], but there is less information on high-power pulse laser interaction with
dielectrics containing metal nanoparticles. In the present case the energy density is lower than
the value of the ablation threshold for the SLSG, which was determined to be about 5 J/cm2
[77]. Also, the excimer laser is characterised by a UV-wavelength, which is much longer than
the typical sizes of the nanoparticles formed by ion implantation. Hence, present metal/glass
composites may be considered similar to be a homogeneous material for light propagation
[17]. This is a simplification, which is true generally for low intensity light, but it gives an
estimate of the optical penetration depth (α-1) of the laser pulses into the composite material,
where α is the linear absorption coefficient. An intense laser pulse is absorbed and relaxed
into heat into the surface SLSG layer of a thickness of α-1, which is several microns [33], i.e.,
deeper than the thickness of the implanted layer (Figure 6).
The optical spectral result of pulse laser treatment by 5 pulses of a KrF excimer laser with
pulse length of 25 ns full-width at half-maximum at a wavelength of 248 nm with the total

released energy of 0.2 J/cm2 on the optical spectra of the Ag-implanted glass is presented in
Figure 10 [37]. Applied laser pulses did not change the reflectance and transmittance spectra
of the non-implanted SLSG, but for implanted sample the location of the transmittance
minimum shifts slightly towards shorter wavelengths, and the transmittance in peak position
increases from 16 to 23% (Figure 10a). Remarkable change was found in the reflectance
spectra, where in the case of the implanted surface, the peak of the overlapping bands shifts
continuously from 490 to 450 nm with modifications in the shape of the envelope of the
bands, which become narrower (Figure 10b). The reflectance intensity falls from 38 to 27%.
However, for reflectance from the rear face of the same sample (Figure 10c) there is only a
decrease of the intensity to a 13% maximum, which is at the same initial wavelength as in the
implanted sample.
For composite materials with MNPs effective-medium theories (Maxwell-Garnett,
Bruggeman, etc) can be considered [17]. According to this approach, the measured optical
intensities corresponding to the implanted glass (Figures 7 and 10) are determined mainly by
the near-surface layer, where there is the largest metal filling factor, and also the largest
nanoparticles (Figure 6).
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Laser Annealing of Composite Materials with Metal Nanoparticles 43
Figure 10. Optical spectra of the soda-lime glass after silver implantation as in Figure 7 and the
implantation followed with laser treatment (0.2 J/cm2): (a) transmittance; (b) reflectance measured from
the implanted face; (c) reflectance measured from rear faces of the sample [37].
The smaller filling factors (and smaller nanoparticles) that exist at the other depths
effectively only influence a smearing of the optical spectra. Moreover, consistent with
effective-medium considerations, a decrease of the filling factor is accompanied by
movement of the optical peak position to shorter wavelengths, as is observed in Figure 10.
Thus, a decrease of the filling factor in the composite resulting from laser treatment suggests
there is a size reduction of all the silver nanoparticles, and again the biggest nanoparticles are
localised near the SLSG surface. The difference between reflectance spectra from implanted
and rear faces of laser treated sample (Figure 10b, c) emphasises the existence of a non-
symmetrical depth distribution of the silver nanoparticles as being similar to the distribution
in the implanted sample.
Figure 11. The RBS spectrum for the Ag-implanted SLSG before and after laser treatment as in
Figure 10 [38].
RBS measurements (Figure 11) show that the laser treatment causes a loss ∼25 % of the
entire silver content in the samples, because of some evaporation of silver from the sample.
Reduction of total silver concentration (filling factor) in the glass leads to decreasing of the
reflectance intensities measured both from implanted and rear side of sample (Figure 10).
These data are consistent with another observation of the thermal metal dissolution from the
heated surface of the float glass with implanted silver nanoparticles after excimer laser
irradiation [32], and with a later study of the silica glass with implanted bismuth particles
[57]. At the same time, as seen in Figure 11, the concentration peak position was not shifted,
but the distribution of silver atoms is broadened compared with that of the as-implanted
sample and is explained by diffusion into the depth of the SLSG. All these processes decrease
the filling factor. The assumption about the size-reduction is in agreement with AFM
measurements of laser treated samples. As presented in Figure 12, (note an increased
magnification of ten times in the direction perpendicular to plane of the figure), it is seen that
there are a lot of clearly defined hills (believed to be silver particles) on the glass surface
whose size is one order smaller in contrast to large hills on the implanted glass surface in
Figure 9. The silver accumulate effect on the laser irradiated surface of glass is result of
melting implanted composite layer and some desorption of glass material under the laser
pulses, which exposes the melted metal particles after their solidification. Hence, the first of
the conclusions from the present data on excimer laser treatment of implanted glass is the
reduction of the size of the silver nanoparticles, and second is the existence of some
asymmetry in their depth distribution (Fig 10b, c). To recognise the mechanisms by which the
changes occur for strongly absorbed excimer laser pulses the thermal propagation after the
laser-irradiation must be considered. The laser heating is traditionally characterised by the
heat diffusion length, l(t) = (D⋅ t)1/2 , where D is the heat diffusivity, and t is the laser pulse
duration.
Figure 12. An AFM image as a top view under lateral illumination of the surface of SLSG Ag-
implanted with a dose of 7·1016 ion/cm2 and an energy of 60 keV followed with irradiation with an
excimer laser (0.2 J/cm2) as in Figure 10. The step along the X and Y axes id 100 nm, and the step
along the Z axis is 40 nm [37].
In the present experiment with laser pulses of 25 ns, the heat propagation is
approximately l(t) = 115 nm, that is shorter than the α-1, i.e., the temperature rise is no longer
controlled by the diffusion of the heat. However, the l(t) surpasses the depth of the implanted
silver nanoparticles. As was estimated earlier [33], the temperature at the surface of laser
treated SLSG reaches values exceeding ∼700 °C which is equivalent to the SLSG melting
temperature. Under these temperature conditions there is also a possibility for melting small
silver particles, because their melting temperature is drastically decreased, for example, to
~400 °C for sizes < 30 nm, compared with the bulk metal melting temperature of 960 °C [80].
As was noticed above, the time scale of electronic relaxation and energy transfer to the lattice
vibrations in the metal particles is several orders faster than in the surrounding glass medium.
Therefore, during the interaction of the excimer laser pulse with the metal/glass composite,
the silver nanoparticles are heated and melted more quickly than solidification of the melted
glass can occur. Atomically dispersed silver released from nanoparticles enters into the glass
melt, and immediately diffuse throughout all the heated thickness of the laser treated
substrate. In principle in time this could lead to a uniform metal distribution, where the silver
atom concentration exceeds the solubility value in the solid glass. However, following glass
solidification spreading from the depth to the surface, as heat from the laser pulse penetrates
into the depth of the sample, the cooling part of the annealing cycle will stimulate new
nucleation and regrowth of metal particles. In this case, the possibility of regrowth of metal
particles will depend on competition between regrowth and the cooling speed of the moving
solidification front, resulting in a new non-uniform size distribution of new MNPs over a
depth scale similar to that after ion implantation. Obviously, under some conditions the metal
particles may be dispersed into separate metal ions and/or into such small units that they can
not display nanoparticle type optical properties. The present consideration of the melting of
the MNPs is also suitable for a description of the reflectance modification discussed in
previous article [32]. In alternative interpretations of the result in [32] the authors of [81]
preferred to see the data as evidence for en evaporation of metal atoms from the surface of
nanoparticles, which resulted in shrinkage of the average size. In same time, in [57] for case
of the implanted bismuth nanoparticles in silica glass treated with a KrF excimer laser, the
melting of the particles was ignored, though the melting point of the bismuth nanoparticles is
very low (~140 °C for the size of 20 nm [82]), and primarily the feature was thought to be the
thermal evaporation atoms from the particles. Evaporation is of course one of the thermal
consequences from high power laser processing and leads to some ablation of the metal and
film surface, and dissolution of metal inclusions on non-metal substrates, for example in [81].
However, in case of implanted composite material, the MNPs are not only localised
immediately at the glass surface, but also extend into the depth of heated solid media. Hence,
processes of atom evaporation from the nanoparticles will be very different from those
limited to particles on the outer surface. Therefore the description of the size modification of
the implanted MNPs based on a particle-melting approach seems reasonable. Thus, metal ion
implantation in glass invariably results in a very wide distribution of nanoparticle sizes that
vary with the depth. As firstly shown here, subsequent high power excimer laser pulse
treatments of the ion implanted layer may be used to melt, and/or regrow, the MNPs within
the insulator medium. Overall this results in a tighter distribution of small particles. The laser
treatments have slightly reduced, but not completely removed evidence for a non-symmetric
depth distribution of the silver particles. The silver-insulator composite material is complex,
and so a much wider range of laser pulse conditions, and more data on the cooling rates are
required to fully model the changes in the size distributions which can occur.
MULTIPULSE LASER ANNEALING OF SODA-LIME GLASS WITH

SILVER NANOPARTICLES
Previous example of laser annealing were carried out only with pulses of the lower
energy density ~ 0.2 J/cm2, but the excimer laser gives the possibility of more powerful
treatment of materials. The present description concentrates on experiments with KrF laser
annealing (25 ns), employing a higher level of energy density per pulse (0.25 J/cm2) and a
larger number of pulses until 250 at a repetition rate of 1 Hz to evaluate the modification of
the metal-implanted glasses. The reflectance spectra for Ag-implanted SLSG for different
numbers of pulses are shown in Figure 13 [83]. All spectra of samples, including the annealed
ones, are characterized by selective SPR bands determined by silver particles. As seen from
this figure, approximately 50 pulses of the laser treatment initiates a decrease of reflectance
intensity from 26 to 12 %. The fall in reflectance following laser annealing is qualitatively
consistent with previous example for the case of the lower pulse fluence of 0.2 J/cm2. It is
again suggested there that either the MNPs were dispersed into atoms, or the large particles
were separated into such small units that they no longer interact as particles.
Figure 13. Reflectance of Ag-implanted soda-lime silicate glass after a dose of 4·1016 ion/cm2 treated
with various numbers of excimer laser pulses (0.25 J/cm2) [83].
The salient feature of the present data is a remarkable increase of reflectance when more
than 50 pulses are applied. After about 100 pulses, where the reflectance reached a maximum
value, the intensities decrease again. Such an intensity increase, with dependence on number
laser pulses, has not been detected before for either metal implanted dielectric, as seen in
already mentioned articles, or for different composites of insulators with metal nanoparticles,
for example, particles in aqueous solutions [84] or in melted photochromic aluminosilicate
glass [31]. Plasmon excitations in small metal particles will be described by Mie scattering
theory. According to this, increase in the size of MNPs is accompanied by movement of the
spectral peak of the optical resonance to longer wavelengths. To evaluate the size change of
implanted silver nanoparticles after laser treatment in Figure 13 some selective spectra of
reflectance from Figure 14 are presented in a plane view. It should be mentioned that the
observed spectra are presented as a superposition of plasmon bands influenced from all size-
distributed particles in the sample, and so the exact determination of the particle size from the
width of spectral bands is not suited. As seen in Figure 14 the reduction of reflectance during
of first 50 pulses is accompanied with moving of peak position to shorter wavelength. These
changes induced by laser treatment suggest there is a reduction of the nanoparticles; however,
continuing increase of number of laser pulses shifts the overall SPR bands, which are
broadening, to red spectral region that seems appearance a wide size distribution of silver
particles.
RBS data show that laser treatment modifies the silver profile in the glass by lowering of
the local concentration in the maximum and by inward diffusion in proportion to the number
of pulses.
Figure 14. Selected reflectance spectra from Figure 13 presented in plain view [83].
Figure 15. RBS data for Ag-implanted soda-lime silicate glass after a dose of 4·1016 ion/cm2 treated
with various numbers of excimer laser pulses (0.25 J/cm2) [83].
Measurable changes of profiles appear only if more than 50 pulses are employed. For a
fewer number of laser pulses the RBS curves are very similar and symmetrical. As an
example, in Figure 15 the RBS spectra for the cases of the Ag-implanted sample and ones
corresponding to two anneal with 50 and 100 pulses are presented. The concentration peak
position was not shifted, but the distribution of silver atoms is broadened and is explained by
diffusion into the depth of the SLSG. According to the position of the concentration tail in
RBS spectrum (Figure 15) and the data on silver profile depth in implanted SLSG [16] the
penetration depth of in the sample is assumed to be approximately 60 nm. Inward silver

migration in the sample during annealing shows that there is heating of the treated sample
surface, and as a consequence the diffusion mobility of metal atoms in the glass is increased.
Thus, in the present example of high-energy density laser pulses, the average temperature
of the sample slowly rises after several pulses, reducing the cooling effect. The continuing
heating of the glass with high concentration of silver atoms leads to a new nucleation stage,
and this allows regrowth of metal particles, that immediately results in the increased
reflectance (Figure 13). Note however the size and depth distributions may have been altered.
Laser treatment with more than 150 pulses eventually raises the sample to such a high
temperature that the fast metal diffusion lowers the silver concentration and preventing
further particle nucleation. Metal-ion implantation in glass invariably results in a very wide
distribution of nanoparticle sizes that vary with the depth. As seen from present data,
subsequent high-power excimer laser treatments with high number of pulses of the ion
implanted layer may be used to melt, and/or regrow, the MNPs within the insulator medium.
This is a new multipulse approach for particle modifications and is diferent from previous
ones, which suggested the use of laser annealing, as a first step for a total dissolution of
implanted metal particles, and regrowth then by second treatment in a furnace or with low-
energy laser irradiation [85].
COMBINED LASER AND THERMAL ANNEALING OF SODA-LIME GLASS

WITH SILVER NANOPARTICLES
The concern of the present part of review is to describe how to increase the uniformity of
the size distribution of MNPs synthesized by ion implantation using an approach on
combination of high-power pulse laser annealing with sequential furnace one [85]. As
example, consider a silicate glass implanted by 60 keV Ag+ ions with a dose of
3·1016 ion/cm2. The implanted glasses were irradiated by 5 pulses of a KrF excimer laser with
pulse length of 25 ns full-width at half-maximum at a wavelength of 248 nm with the total

released energy being kept equal to 0.2 J/cm2. After laser treatment, the samples were
annealed in a furnace according to the conventional procedure at a temperature of 350oC for 1
h in air.
As it was shown earlier the no-uniformity in size distribution of metal particles over the
depth in the sample leads to dissimilar absorption and reflection of layers containing different
in sized nanoparticles. This was observed by the pronounced discrepancy between the
reflection spectra recorded on the implanted side and rear one (Figure 7). From Figure 16a it
is seen that the reflection recorded from the implanted side of the sample is characterized by a
broad band with a maximum at 460 nm while the reflection maximum is near 485 nm for the
rear side. Thus, there is a possibility to use an optical spectroscopy for registration of the
particles size uniformity in a sample. Optical reflection spectra of the implanted samples after
the pulsed laser annealing are presented in Figure 16b.
Figure 16. Optical reflectance measured from the implanted side (1) and from the rear side (2) for the
samples after ion implantation (a), implantation followed with laser treatment (b), and implantation
with the sequential laser and thermal annealing (c).
It is seen that after laser irradiation the spectra corresponding to reflection from both
surfaces of the sample undergo the changes. The spectrum after the laser annealing is
characterized by a narrower reflection band compared to the spectrum of the as-implanted
sample measured on the implanted surface (curve 1 in Figure 16a and 16b). The reflection
maximum shifts toward the short wavelengths up to 450 nm. A similar shift of the reflection
band is also observed for the spectrum recorded on the rear side, with the maximum turning
out to be at 475 nm. However, despite the coherent shift of the bands toward UV region, the
pronounced spectral difference between different sides of the sample still persists, thus
characterizing the available no-uniform size distribution of the silver nanoparticles over the
depth.
As a control instrument the Mie theory can be applied for the modeling of optical spectra.
This theory allows the calculation of an optical extinction cross section σext for a single
isolated nanoparticle embedded into different media. Simulated optical spectra in present here
are shown as spectral dependencies of σext using the well known Mie equations. Dielectric
functions of Ag and Ag2O nanostructured materials for calculations are taken from [86] and
[87], respectively.
Figure 17. Calculated extinction spectra of Ag nanoparticles in the glass as a function of the size [23].
In Figure 17 modeled Mie extinction spectra of silver nanoparticles in the glass

environment with dependence on a particle size from 5 to 50 nm, which corresponds to
typical values [6], are presented. As seen in the figure, decrease of the nanoparticles in size is
accompanied by movement of the SPR peak to shorter wavelengths with reduction of the
intensity. These peak shift is caused by the phase retardation of the electromagnetic waves
and the influence of higher multipoles [17]. Comparison of this calculation with the
experimental data (Figure 16) also suggests that the laser treatment of the glass with Ag
nanoparticles stimulates a decrease of their size. As was mentioned before, this redaction in
size is explained by the fact that laser emission in the spectral region of glass absorption
causes heating-up and melting of the near-surface layer within only several nanoseconds.
Laser heating-up of the glass matrix to a temperature higher than the melting point of metal
nanoparticles, the particles undergo melting that results in decrease of their sizes up to
complete dissociation. On the other hand, in the experiments with Nd:YAG laser treatment of
Ag implanted nanoparticles in silicate glasses [35] the oxidation of the metal particle was
assumed as a reason for the degradation of optical extinction of the composites. To check
such a possibility for our case the extinction spectra of core Ag particles with Ag2O shells in
the glass were calculated using “recursion formalism” [88-90]. In this modeling the diameter
of the metal core was taken to be constant, equal to 10 nm and the thickness of the shell was
varied monotonically from 0 to 10 nm (Figure 18). It is shown that the increase in thickness
of the shell shifts the extinction peak to longer wavelengths and the intensity of the SPR band
decreases. This is contrary to the current observations (Figure 16). Thus, the calculations
allow one to infer that after the excimer laser annealing of the implanted glass, a decrease in
the Ag particle size is more probable than the particle oxidation.
Figure 18. Calculated extinction spectra of Ag nanoparticles with Ag2O shell in the glass depending on
the shell thickness. The core size is constant and equal to 10 nm [23].
Hence, in laser treatment conditions are realized under which the near-surface layer of the
glass turns out to be saturated with silver atoms and minute metal particles which can serve as
nuclei for synthesis of an ensemble of new nanoparticles characterized by more uniform size
distribution. For this purpose, after the ion implantation and pulsed laser irradiation the
samples can be furnace annealed.

Optical reflectance spectra of the samples after the implantation, laser and furnace
annealing are presented in Figure 16c. The distinctive feature of the optical spectra recorded
both on the implanted and a rear side is the exact coincidence of the maxima positions at
475 nm and the spectral shapes. A slight difference in the intensities of the bands can be
attributed to the thickness of the glass backing and associated light losses in recording the
reflection on the side of the backing. The observed similarity of the spectra is indicative of the
fact that after the sequential pulsed laser irradiation and thermal annealing the composite
layer with the uniform size distribution of the MNPs over the depth is formed This effect is
explained by the increase of the diffusional mobility of silver atoms during the annealing of
the sample in a furnace, which is accompanied by enlargement of MNPs with simultaneous
averaging of their sizes practically throughout the entire implanted layer.
Narrowing of the size distribution function of the MNPs should lead to an increase in the
absorption coefficient near the SPR maximum, that is shown in Figure 19. From the presented
optical transmission spectra it is seen that for the as-implanted sample the intensity of the
band corresponding to absorption caused by silver nanoparticles is markedly lower than in the
sample after the sequential laser and furnace treatment.
Figure 19. Transmittance of the glass after ion implantation (1), implantation followed with laser
treatment (2), and implantation with the sequential laser and thermal annealing (3) [83].
This increase of absorption does not give permission for the suggestion on oxidation of
silver particles by annealing in furnace. It was already mentioned that the increase of the
linear absorption of glasses with MNPs as a consequence of SPR inevitably causes an
increase of the third order susceptibility of the material. Moreover, from Figure 19 it is seen
that the maximum of the transmission band of the composite glass after the combined
treatment lies between the more long-wave maximum of the implanted glass and the more
shortwave maximum corresponding to the glass after the implantation and laser irradiation.
According to the Mie theory, for MNPs with a diameter less than λ/20, where λ is the
wavelength of light, the position of the absorption maximum of the material is determined by
the mean size of nanoparticles in it. Consequently, the mean size of Ag nanoparticles in the
glass after final heat treatment can come, according to our estimations, to 5-10 nm.
Using silver ion implantation into the SLSG in combination with high-power pulsed
irradiation by an excimer laser (in the absorption band of the glass matrix) and furnace
annealing, one can create composite materials characterized by a high filling factor and
uniform size distribution of the MNPs throughout the entire implanted layer of the material.
The approach developed makes it possible to fabricate more advanced elements of nonlinear
optics and also to control their characteristics by simple methods of optical spectroscopy.
LASER ANNEALING OF SIPPHIRE WITH ION-SYNTHESISED

COPPER NANOPARTICLES
In the case of implantation by Cu+ ions at relatively low energy 40 keV, nuclear
collisions prevail during interaction of accelerated ions with atoms of substrate, but along
with this, target Al2O3 atoms also effectively lose electrons. As a result, the Cu+ ions deionize
with the formation of neutral copper atoms Cu0. They displace atoms in the sapphire matrix
and break some of chemical bonds of substrate. Basically, copper atoms may produce
chemical bonds with unfettered matrix atoms, specifically, with oxygen atoms. However, Cu–
Cu bonding is energetically more favorable, as can be judged from the change in the Gibbs
free energy in comparison with copper–oxygen reaction. A similar situation is observed with
silver atoms in silver-doped silica glass [91]. In view of the fact that copper atoms readily
combine with each other, an excess of the copper amount over the solubility limit of metal in
Al2O3 leads to the formation of copper nanoparticles in the ion-implanted layer.
An increase in the absolute copper ion concentration in the depth profiles and the surface
sputtering coefficient depend on the implantation time (or the time of implant accumulation).
As in the case of Ag+-ion implantation into SLSG, in present low-energy irradiation the
larger copper particles will be observed closer to the Al2O3 surface, while finer nanoparticles
will penetrate deeper into the matrix. Copper distribution inwards to the Al2O3 matrix have
been corroborated experimentally. As example, Figure 20 shows the RBS copper atom
profiles near the surface of the sapphire corresponding to ion implantation with a dose of
1.0·1017 ion/cm2 at two ion current densities of 5 and 12 μA/cm2. Note that the RBS method
gives the distribution of only copper atoms over the volume, ignoring the possible presence of
the volume metal phase. As seen in figure, in the experimental distributions obtained at two
ion current values, the copper concentration reaches a maximum near the surface (∼ 10-20 nm
in depth), decreasing monotonically with depth down to 60 nm.
Figure 20. Depth profile of the Cu+-ion distribution for 40 keV implantation into Al2O3 at dose of
1.0·1017 ion/cm2 and two ion beam current densities, evaluated from RBS spectra [21].
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Laser Annealing of Composite Materials with Metal Nanoparticles 55
Figure 21. Optical reflectance spectra of Al2O3 samples after Cu+ ion implantation with fixed dose of
1.0·1017 ion/cm2 at various ion current densities (thin lines) and spectra taken from the same samples
subjected to post-implantation laser annealing (thick lines) [20].
Measured optical reflectance spectra of Cu-implanted Al2O3 for different ion current
densities are represented in Figure 21 from [20]. The wide reflection bands in the visible
range directly indicate the formation of copper nanoparticles in the sapphire volume. These
bands are attributed to SPR localized in the small metal particles [17]. By comparing electron
microscopic and optical data for MNPs synthesized by different techniques, it was shown that
this optical resonance, which is due to collective oscillations of free electrons, is observed in
the visible range when the size of copper particles lies between ∼2 nm and several tens of
nanometers [17]. Such reflection spectra are typical of copper particles in dielectric matrix.
As seen in Figure 21, for the lowest current density of 2.5 μA/cm2, the reflectance peak is
positioned at 610 nm with an intensity of 27 %, implying the formation of small copper
particles. Increasing the ion current density shifts the reflectance peaks to longer wavelengths
(up to ∼640 nm) and intensity reaches 41 % for 12.5 μA/cm2. These spectral changes
correspond to formation larger Cu particles. Similar optical effects were detected in the case
of silver nanoparticles in SiO2 fabricated at different current densities [92].
Figure 22. AFM image of the surface of Al2O3 before ion implantation. The values plotted in the X and
Y axes are given in nanometers. The step in the Z direction is 20 nm [21].
Figure 23. AFM image of a fragment of the surface of Al2O3 implanted with Cu+ ions at an ion current
density of 7.5 μA/cm2. The values plotted in the X and Y axes are given in nanometers. The step in the
Z direction is 24 nm [21].
The AFM images of the non-implanted Al2O3 and same substrate after Cu+ implantation
at an ion beam current density of 7.5 µA/cm2 are shown in Figure 22 and 23 [21]. The
formation of copper nanoparticles at the implanted surface is observed in AFM as semi-
spherical hillocks of mean size 20 nm, which are absent on the nearly smooth surface of
virgin Al2O3. These hillocks seen in Figure 22 result from surface sputtering of Al2O3 during
ion implantation and are merely the striped tops of spherical-shaped MNPs nucleated in the
near-surface layer of the substrate. It is assumed that sapphire is removed (sputtered) with a
higher rate than the metallic phase. Estimates [13, 14] made for high-dose copper
implantation into sapphire show that the Al2O3 layer sputtered may be several nanometers
thick for the doses and energies considered in this work. For Cr+ and Ti+ implantation into
sapphire, the thickness of the layer sputtered from the sapphire surface was the value of same
order (Figure 3) [62]. Similar surface morphology with MNPs were earlier detected by AFM
on the samples after low-energy (< 60 keV) high-dose ion implantation, for example: Au+
into Al2O3 [93] and mica [94]; Fe+ into SiO2 [95]; Ag+ into Al2O3, Ta2O3, Si3N4, SiO2
[92, 96, 97] and in SLSG (Figure 9).The metal phase of nanoparticles in the hillocks on the
surface was identified by glancing incidence X-ray diffraction [97].
The Mie equations were be used to simulate the extinction spectra for copper
nanoparticles in Al2O3. In the calculations described below, we used the complex optical
constants for Al2O3 [98] and copper [99] corrected for size effects (limitations imposed on the
free path of electrons) observed in particles of size ranging from 1 to 200 nm. Simulated
extinction spectra for copper nanoparticles embedded in Al2O3 matrix for comparison them
with experimental data (Figure 21) are shown in Figure 24 [20]. These spectra feature a wide
band, which covers the entire spectral range. In the given range of particle sizes, the position
of the SPR absorption maximum shifts toward longer wavelengths with increasing particle
size. Simultaneously, the intensity of the extinction band grows and the spectra somewhat
broaden, which is in agreement with the changes in the experimental spectra with increasing
ion current (Figure 21). Such behavior of the reflection spectra confirms the statement that
copper particles in sapphire grow with ion beam current density.
Figure 24. Calculated spectra of optical extinction cross-section for copper nanoparticles embedded in
Al2O3 vs. particle size [20].
The most plausible reason for the increasing of the particle sizes is that implantation rises
the temperature of the Al2O3 matrix. For example, as was shown before [40], when low-
energy Ag+ ions were implanted into preheated (from 20 to 60°C) silica glasses at a moderate
ion current density 3 µA/cm2, larger silver particles were formed in warmer samples. It is
obvious that an increase in the substrate temperature accelerates diffusion mobility of the
implants and increases a probability for nucleation and growth of MNPs.
Main feature in the experiments presented above was that the laser pulses were applies in
ultraviolet spectral area, which falls into the range of strong absorption by dielectrics, for
example, in the case of SLSG containing implanted silver nanoparticles. Such annealing
usually leads to melting of the glass near surface layer, including the implanted MNPs,
followed by rapid cooling. As a result, the size of the silver nanoparticles diminishes, because
the heat of the matrix, which intensely heats up, melts the particles. Melting is favored by a
relatively low melting point of the SLSG (≈750°C), application of high-power laser pulses,
and also the fact that the melting point of small silver particles is close to the melting point of
the glass. In present case, the similar approach to copper nanoparticles implanted in Al2O3 for
modification using an excimer laser was applied. However, the melting point of artificial
sapphire is higher (1400°C) than that of SLSG and sapphire is more trans-parent at the KrF
laser radiation wavelength of 248 nm (Figure 25) [23]. Therefore, one can assume that the
laser radiation absorption in sapphire is lower than in the glass and consider the annealing
process as direct interaction between laser light and MNPs irrespective of the substrate
temperature. As seen in Figure 25, after ion implantation the optical transmission of the Al2O3
in the ultraviolet spectral range decreases because of induced radiation defects of matrix [6].
Yet, the matrix remains quite transparent (35–40%) near the laser wavelength especially
versus the transmission of SLSG of the same thickness.
Figure 25. Optical transmission spectra of Al2O3 before and after copper implantation for different
values of the ion beam current density. The spectrum for non-irradiated SLSG is given for comparison
[23].
It should be noted that synthesized copper nanoparticles also slightly absorb ultraviolet
light due to interband transitions just as in the bulk metal [17].
The reflection optical spectra of the implanted samples subjected to laser annealing are
shown by thick lines in Figure 21. These selective bands indicate that the MNPs are present in
the sapphire after the annealing as well. However, in all the samples, the maxima of these
reflection SPP bands turn out to be shifted toward shorter wavelength and they are less
intense than in the spectra taken immediately after implantation. These changes are more
pronounced in the case of high-current ion implantation; in other words, for samples with
large copper particles are more sensitive to laser radiation that fine particles, which were
synthesized at low current densities. Redaction of the intensity of the spectral bands for
annealed samples in present study is smaller than it was earlier registered for copper and
silver nanoparticles formed by ion implantation in SLSG and treated by laser under similar
conditions. In terms of the Mie theory, which describes the optical properties of small metal
particles (Figure 24), the shift of the reflection maxima to shorter wavelength means that laser
annealing of sapphire decreases the mean size of the copper particles. This conclusion is
corroborated by AFM examinations. In fact, the hillocks seen in the surface image after ion
implantation Figure 23) are large than those observed after ion implantation and laser
annealing (Figure 26) at least by one order of magnitude. As was mentioned, in the case of
metal-dielectric composite, photoprocesses of excitation and relaxation can be distinguish
between effects in sapphire matrix and in the MNPs. Despite of this sapphire matrix has an
energy bandgap at ∼ 9.9 eV, irradiation with a high power excimer laser at the wavelength of
248 nm (5 eV) couples directly through the spectral absorption edge of Al2O3 (Figure 25).
Therefore, in Al2O3 the radiation generates electron–hole pairs by direct exiting electrons
from the valence into the conduction band. In the copper particles, both conduction- and
valence-band electrons may participate in the laser excitations at the same wavelength. In
other words, the number and energy of absorbed photons immediately establishes the
temperature rise in a laser-irradiated metal.
Figure 26. AFM image of a fragment of the surface of Al2O3 implanted with Cu+ ions at an ion current
density of 7.5 μA/cm2 after laser annealing. The values plotted in the X and Y axes are given in
nanometers. The step in the Z direction is 48 nm [21].
For long pulse duration (25 ns) of the laser described here, the laser energy transfers
directly into a heating of the MNPs. In came time, although Al2O3 is considered as almost
transparent to the laser light (Figure 25), a fraction of the laser radiation is still absorbed by
the sapphire matrix especially with matrix defects stimulated during ion implantation.
Nevertheless, it may be assumed that most of the energy of the laser radiation is absorbed by
the MNPs and the transparent substrate has no time to heat up. Thus, if we exclude melting of
MNPs as in case of SLSG, then the decrease in the metal nanoparticle size observed in the
experiment (Figures 21and 26) may be due to photofragmentation as suggested for laser
irradiated small metal particles [100]. This process means a decrease in the total number of
the particles in the substrate when some of them disintegrate into atoms or tiny molecule-like
clusters, which do not exhibit SPR-related absorption. Clearly, large MNPs, which have a
larger volume and cross-sectional area and, therefore, absorb a greater amount of the laser
energy, are more prone to fragmentation. Especially, the larger particles are responsible for
the reduction of the SPR band intensity and their shift to shorter wavelength when they
disintegrated. Earlier, metal particle fragmentation was observed in colloidal solutions
exposed to high-power picosecond or nanosecond laser shorts at wavelengths outside the SPR
spectral range of MNPs [100–102]. It was assumed that fragmentation takes place when the
laser radiation causes electrons to move to the periphery of the particles, which thereby
acquire an appreciable surface charge (Figure 27) [100].
On the other hand, the conversion of the laser radiation in metal particles may be treated
as the photon energy transformation (relaxation) directly into heating of the MNPs in
composite. The temperature may rise to the melting point of copper particles, since it is
known that the melting point of a metal particle may drop substantially when its size
decreases to the nanometer scale from the bulk [103, 104]. Consequently, when the
temperature of the particles exceeds their melting point, they melt, diminish, and may even
collapse. In general, melting of nanoparticles is a nontrivial process and its correct description
requires that several sequential stages be considered: surface atom migration (surface
premelting), structure fluctuations (quasi-melting), and the formation of mixed (liquid–solid)
phases. This effect is dependent on the particle size, and so smaller particles will decrease
faster than larger ones. In general, a common decrease in the particle sizes of all particles by a
given melting mechanism will be reflected in the optical spectra in the same way as in
photofragmentation. Therefore, it is as yet difficult to decide between the two mechanisms of
particle decrease at laser annealing. The difference between them is that fragmentation breaks
down primarily larger particles, while fine particles are more prone to melting. Fragmentation
seems to be a more vigorous process and possibly is responsible for the changes observed at
laser annealing.
Figure 27. Stages of metal particle fragmentation with laser excitation. A transient aggregate formed via
the photoejection of electrons is considered to be a precursor for complete fragmentation of the particle.
LASER ANNEALING OF SILICA WITH ION-SYNTHESISED

COPPER NANOPARTICLES
The questions of the ability of MNPs to absorb laser radiation energy independently and
effectively and of the related practical possibilities of modifying composite materials remain
unanswered. Therefore additional experiments were done investigate the ion synthesis of
copper nanoparticles in SiO2 with subsequent annealing of the composite material by an
excimer KrF laser [24-26, 105]. The application of an SiO2 (fused silica) substrate
characterized by a higher melting point (∼1175°C) and a higher degree of transparency at the
lasing wavelength (248 nm), as compared with SLSG, makes it possible to prevent strong
absorption of laser radiation and, consequently, to study the interaction of nanoparticles with
the laser beam independently of the melting of the dielectric matrix. Note that absorption of
SiO2 at wavelength of 248 nm lower than in Al2O3. Implantation was carried out by 50-keV
Cu+ ions with a dose of 8·1016 ion/cm2 at a current density in an ion beam of 10 µA/cm2.
Laser treatment of the implanted SiO2 was carried out by pulses of an excimer KrF laser of
duration 25 ns (248 nm) at a frequency of 1 Hz, with the energy released in a pulse being
equal to 0.21 J/cm2. The number of pulses was varied from 1 to 150.
The nucleation and growth of the metallic nanoparticles in the SiO2 during implantation
occur on exceeding the solubility limit by the introduced copper atoms. The threshold dose is
∼1·1016 ion/cm2 for an energy of 50 keV. The formation of Cu nanoparticles is supported by
the appearance of a broad selective SPR band with a minimum at 600 nm in the optical
transmission spectrum of the implanted glass (Figure 28).
Figure 28. Transmission spectra of SiO2 before ion implantation (1). Spectra measured after Cu-ion
implantation with a dose of 8·1016 ion /cm2 of SiO2 (2) and SLSG (3) [105].
It should be mentioned that the spectral position of the SPR band is determined by the
particle size, and for the implanted layer, which, as has been shown, is characterized by
variations in the nanoparticle sizes, the experimental optical spectrum is a superposition of the
absorption of differently sized particles. At the same time, despite the variations in the
nanoparticle sizes, it is precisely the large particles that make the greatest contribution to the
absolute intensity of the absorption and reflection band, while the small particles contribute
mainly to the broadening of the spectrum. Radiation defects in a glass increase its absorption
at the edge of the fundamental band in the ultraviolet region of the spectrum [17], which can
significantly diminish the transparency of the composite material in the range of wavelengths
at which the irradiation is carried out by an excimer laser and thus make the solution of the
problem at hand difficult. However, as is seen from Figure 28, the chosen conditions of
implantation, on the one hand, have made it possible to form Cu nanoparticles in the surface
layer of the glass, but, on the other hand, although they caused additional absorption in the
UV region, they have not led to the complete limitation of the transmission of the glass at a
wavelength of 248 nm (T = ∼30%). Here it should be taken into account that a portion of light
is absorbed by the Cu nanoparticles because of their high concentration. For comparison, we
present the typical transmission spectrum of a SLSG implanted with Cu+ ions, which shows
that transmission is completely absent for a wavelength < 300 nm.
Since only a small area of the surface of the specimen was exposed to light in the present
experiment because of the limitation of the laser beam cross section, it was analyzed with the
use of only optical reflection spectroscopy. The measured reflection spectra of the glass
subjected to implantation and subsequent laser annealing (depending on the number of pulses
used) are shown in Figure 29.
Figure 29. Reflectance of Cu-implanted SiO2 treated by various numbers of excimer laser pulses with
an energy density of 0.21 J/cm2 [26].
It is seen that after the first laser pulse, the intensity in the maximum of the reflection
band of the Cu nanoparticles decreases from 22 to 19%. A similar effect was observed earlier
for Ag nanoparticles in a SLSG (Figure 13 and 14) and SiO2 irradiated at the wavelengths of
the SPR of nanoparticles in the visible region with the aid of an Nd:YAG laser [35]. The
increase in the number of laser pulses in our case (Figure 29) leads to the increase in the SPR
band intensity, which reaches its maximum (21%) after 10 pulses, and then to the monotonic
decrease of the optical reflection coefficient to 12.5% (150 pulses), i.e., approximately
proportionally to the laser pulse energy.
Figure 30 presents the spectra in a two-dimensional form from Figure 29 in order to make
a detailed analysis of the changes in the spectral position of the maxima of the reflection
bands and in their shape caused by the laser action. The main distinguishing features of these
groups of spectra is that in the first case (Figure 30a) the irradiation does not change the
spectral position of the reflection band and the intensity in the maximum increases with
increase in the number of pulses.
Figure 30. Selected reflectance spectra from Figure 29 presented in plain view [105].
In the second case (Figure 30b), the intensity decreases and the reflection band shifts to
the side of the long wavelengths (from 600 to 620 nm). It should be emphasized that earlier,
in the case of exposure of SLSG with silver particles to an excimer laser with ∼10 pulses, the
increase in the number of pulses in the process of irradiation also led to the decrease in the
reflection intensity, but the spectrum itself shifted to the side of shorter wavelengths. This
behavior of the spectra corresponded to the decrease in the sizes of nanoparticles in the
implanted layer according to the Mie theory.
Using the Mie calculation approach for estimating the dynamics of the change in the
particle sizes and starting from the position of the experimental maxima in the spectra (Figure
30a), it may be suggested that the mean size of the Cu nanoparticles does not change after the
first laser pulse. The decrease in the reflection intensity is likely to point to the fact that the
total number of particles in the layer decreases because of the breakdown of a portion of the
nanoparticles into individual atoms or small-size molecule-like clusters that do not show
plasma resonance. On irradiation by a single nanosecond pulse, the main portion of the laser
radiation is absorbed by the nanoparticles, and the transparent glass substrate has no time to
heat up. Evidently, the largest nanoparticles of large volume and cross section were subjected
to fragmentation. It is precisely the breakdown of the large nanoparticles that causes the
decrease in the intensity of the maximum in the reflection spectrum. Similar fragmentation of
metallic nanoparticles was observed earlier in colloidal solutions irradiated by high-power
picosecond or nanosecond laser pulses at wavelengths beyond the range of the plasma
resonance spectral position [100-102].
Simultaneously with the breakdown of the large particles, the implanted layer becomes
saturated or oversaturated with the copper atoms that earlier were part of these particles. After
application of several additional pulses, the implanted glass layer is heated due to both the
absorption of a portion of the laser radiation by the structural radiation defects and the
transfer of heat from the heated metallic particles to the glass substrate. It is known [106] that
an increase in the temperature of the medium containing differently sized nanoparticles leads
to the enlargement of the particles due to the so-called diffusion growth, when they are joined
by the metal atoms from the medium. This gives rise to competing processes: fragmentation
of nanoparticles, on the one hand, and their enlargement, on the other. When the temperature
of the matrix increases, the second process becomes dominant. This explains the increase in
the reflection intensity on increase in the number of laser pulses from 1 to 5 (Figure 30a).
Subsequent laser irradiation (more than 10 pulses) can cause heating of the glass
substrate due to the absorption of light by radiation defects and, along with it, the heating of
the Cu nanoparticles to higher temperatures. At the same time, the observed monotonic
decrease in the reflection band intensity on increase in the number of pulses points to the
gradual and complete disappearance of the metallic particles in the implanted layer. This can
be explained by effective melting of the nanoparticles. However, in the experiments
conducted we have not revealed melting of the glass surface, and therefore the mechanism
proposed for the melting of the Ag nanoparticles in SLSG due to the melting of the glass
itself cannot be used to obtain an adequate explanation of the present experiment. Thus, for
the Cu nanoparticles synthesized in SiO2, the following mechanism of their transformation in
laser annealing can be proposed. Cu nanoparticles acted on by a sequence of pulses are heated
as a result of direct absorption of laser radiation and its transformation into heat energy. The
surrounding glass matrix serving as a thermostat enables a particle to effectively accumulate
the energy obtained and store it for a time exceeding the interval between the laser pulses. But
here, one or several laser pulses cannot provide the temperature required for the melting of a
particle; however, the successive effect of tens of pulses makes it possible to reach the
melting point and leads to the irreversible destruction of the nanoparticles. It should be noted
that this mechanism can be dominant but not unique. The melting and dissociation of the
nanoparticles can be favored by the effects of local melting of the glass matrix in the
implanted layer and also by the above-mentioned effects of fragmentation occurring as a
result of excitation of the electrons of the metallic particles by laser radiation and their
transforming into a nonequilibrium state.
The observed shift of the reflection spectrum maximum to the side of longer wavelengths
(Figure 30b) cuold be explained by the formation of a CuO or a CuO2 shell coating the copper
nucleus, which appears at the intermediate stage between the heating of the metallic
nanoparticles and their melting. In the previous experiments with Ag nanoparticles in a
SLSG, the formation of oxide shells had not been detected. It seems that the melting of the
glass itself hindered the formation of metal oxide on the particles.
Thus, here the results of modeling the distribution profiles of Cu+ ions implanted into
SiO2 over its depth and experimental data on the investigation of the processes of formation
of Cu nanoparticles and the changes in their parameters under the action of high-power pulses
of an excimer laser are presented. It has been established that in the absence of effective
absorption of laser radiation by the glass substrate, the dynamics of the change in the metallic
nanoparticles is determined by the number of irradiation pulses. At the initial stage of laser
treatment fragmentation of the largest nanoparticles occurs, which is followed by their
agglomeration as a result of the heating of the glass to a certain level. A prolonged laser
action (tens to hundreds of pulses) leads to effective accumulation of energy in the
nanoparticles and, as a consequence, to their melting and dissociation into small clusters and
individual atoms.
Reflectance of Cu-implanted SLSG for various numbers of laser pulses are presented in
Figures 31. As seen from figure, Again, the intensity of SPP reflectance is increasing, when
the number of pulses increases (from 1 until 5-10 in the present case), similar to the rise in
reflectance in Ag-implanted SLSG (Figure 13 and 14). Continued laser treatment with more
pulses leads to a decreasing reflectance. In same time, there is a difference in laser annealing
for SiO2 and SLSG with Cu nanoparticles. The reflectance reduction accrues more quickly
(for a smaller number of pulses) in the SiO2 compared with the SLSG. To correctly describe
the influence of type of the substrate on reflectance changes during laser treatment requires
answers to several questions.
One needs to know the role of non-uniform absorption of laser light, the metal
distribution after implantation, the diffusion coefficient of metal atoms in various materials
and so on. All these factors have to be taken into account and haven’t be evaluated at present
study yet.
However, note that the observed effect of increasing reflectance, via increases with the
number of pulses, as a result of regrowth of nanoparticles, must not be considered as specific
to copper or silver only. It is probably a typical behavior of MNPs in insulators under high-
power laser treatment. Although this effect was demonstrated using only an excimer laser, the
use of the transparent silica matrix in these experiments allows the possibility of annealing
with Nd:YAG and others lasers.
Figure 31. Reflectance of Cu-implanted SLSG treated by various numbers of excimer laser pulses with
an energy density of 0.21 J/cm2 [13].
REFERENCES
[1] C. Flytzanis, F. Hache, M.C. Klein, D. Ricard, and P. Rousignol Nonlinear Optics in
Composite Materials (Elsevier Science, Amsterdam, 1991).

[2] A.L. Stepanov, I.B. Khaibullin, P.D. Townsend, D.E. Hole, and A.A. Bukharaev, Rus.
Feder. Patent No. 2156490, 2000.
[3] P.D. Townsend, and L. Massarelli, US Patent No. 5102736, 1992.
[4] J.M. Hampikian, and E.M. Hunt, US Patent No. 6294223, 2001.
[5] N. Kishimoto, Y. Takeda, and S. Okubo, Jap. Patent No. JP2004091817, 2004.
[6] P.T. Townsend, P.J. Chandler, and L. Zhang, Optical Effects of Ion Implantation
(Cambridge Univ. Press: Cambridge, 1994).
[7] A.L. Stepanov, and D.E. Hole, in Recent Research Development Applied Physcs;
Transworld Research Network: Kuala, 2002, Vol. 5, pp. 1-26.
[8] A. Polman, A. J. Appl. Phys. 82, 1 (1997).
[9] J. Davenas, A. Perez, P. Thevenard, and C.H.S. Dupuy, Phys. Stat. Sol. A 19, 679
(1973).
[10] M. Treilleux, P. Thevenard, G. Ghassagne, and L.H. Hobbs, Phys. Stat. Sol. A 48, 425
(1978).
[11] G.W. Arnold, J. Appl. Phys. 46, 4466 (1975).
[12] A.L. Stepanov, in Metal-Polymer Nanocomposites, edited by L. Nicolais, G.
Carotenuto (John Wiley and Sons Publ, London, 2004), p. 195.
[13] A.L. Stepanov, V.A. Zhikharev, D.E. Hole, P.D. Townsend, and I.B. Khaibullin, Nucl.
Instr. Meth. B 166/167, 26 (2000).
[14] A.L. Stepanov, V.A. Zhikharev, and I.B. Khaibullin, Phys. Sol. State 43, 766 (2001).
[15] L.C. Nistor, J. van Landuyt, J.B. Barton, D.E. Hole, N.D. Skelland, and P.D.
Townsend, J. Non-Cryst. Solids. 162, 217 (1993).
[16] A.L. Stepanov, D.E. Hole, and P.D. Townsend, Nucl. Instr. Meth. B 161/163, 913
(2000).
[17] U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters (Springer, Berlin,
1995).
[18] J. Rankin, P. Thevenard, L.J. Romana, L.A. Boatner, C.W. White, C.J. McHargue, and
L.L. Horton, Surf. Coat. Technol. 51, 471 (1992).
[19] A.A. Bukharaev, A.V. Kazakov, R.A. Manapov and I.B. Khaibullin , Sov. Phys. Solid
State 33, 578 (1991).
[20] A.L. Stepanov, U. Kreibig, D.E. Hole, R.I. Khaibullin, I.B. Khaibullin, and V.N.
Popok, Nucl. Instr. Meth. B 178 ,120 (2001).
[21] A.L. Stepanov, V.N. Popok, D.E. Hole, and I.B. Khaibullin, Appl. Phys. A 74, 441
(2002).
[22] A.L. Stepanov, Tech. Phys. 50, 1 (2005).
[23] A.L. Stepanov, and I.B. Khaibullin, Rev. Adv. Mater. Sci. 9, 109 (2005).
[24] A.L. Stepanov, D.E. Hole, and P.D. Townsend, Nucl. Instr. Meth. B 166/167, 882
(2000).
[25] A.L. Stepanov, D.E. Hole, and P.D. Townsend, Nucl. Instr. Meth. B 191, 468 (2002).
[26] A.L. Stepanov, and D.E. Hole, Surf. Coat. Tech. 158/159, 526 (2002).
[27] K. Masuo, O.A. Plaksin, Y. Fudamoto, N. Okubo, Y. Takeda, and N. Kishimoto, Nucl.
Instr. Meth. B 247, 268 (2006).
[28] J. Werner, A. Heilmann, and O. Stenzel, Int. J. Electronics 77, 945 (1994).
[29] Heilmann, J. Werner, F. Homilius, and F. Müller, J. Adhesion Sci. Technol. 9, 1181
(1995).
[30] Heilmann, Polymer Films with Embedded Metal Nanoparticles (Springer, Berlin,
2003).
[31] A. Akella, T. Honda, A.Y Liu, and L. Hesselink , Opt. Lett. 22, 967 (1997).
[32] R.W. Wood, P.D. Townsend, N.D. Skelland, D.E. Hole, J. Barton, and C.N. Afonso, J.
Appl. Phys. 74, 5754 (1993).
[33] P.D. Townsend, and J. Olivares, Appl. Surf. Sci. 109/110, 275 (1997).
[34] M. Ferrari, L.M. Gratton, A. Maddalena, M. Montagna, and C. Tosello, J. Non.-Cryst.
Solids 191, 101 (1995).
[35] F. Gonella, G. Mattei, P. Mazzoldi, E. Cattaruzza, G.W. Arnold, G. Battaglin, P.
Calvelli, R. Polloni, R. Bertoncello, and R.H. Haglund, Jr., Appl. Phys. Lett. 69, 3101
(1996).
[36] G. Battaglin, E. Cattaruzza, F. Gonella, R. Polloni, F. D’Acapito, S. Colonna, G.
Mattei, C. Maurizio, P. Mazzoldi, s. Padovani, C. Sada, A. Quaranta, A. Longo, Nucl.
Instr. Meth. B 200, 185 (2003).
[37] A.L. Stepanov, D.E. Hole, A.A. Bukharaev, P.D. Townsend, and N.I. Nurgazizov,
Appl. Sur. Sci. 136, 298 (1998).
[38] A.L. Stepanov, D.E. Hole, and A.A. Bukharaev, Vacuum 64, 169 (2002).
[39] A.L. Stepanov, Rev. Adv. Mater. Sci. 4, 45 (2003).
[40] A.L. Stepanov, V.N. Popok, Surf. Coat. Technol. 185, 30 (2004).
[41] A.L. Stepanov, B.N. Chichkov, V.F. Valeev, V.I.Nuzhdin, I.A. Faizrakhmanov, Tech.
Phys. Lett. 34, 184 (2008).
[42] M. Kaempfe, T. Rainer, K.-J. Berg, G. Seifert, and H. Graener, Appl. Phys. Lett. 74,
1200 (1999).
[43] M. Kaempfe, H. Hofmeister, S. Hopfe, G. Seifert, and H. Graener, J. Phys. Chem. B
104, 11847 (2000).
[44] M. Kaempfe, G. Seifert, K.-J. Berg, H. Hofmeister, and H. Graener, Eur. Phys. J. D 16,
237 (2001).
[45] G. Seifert, M. Kaempfe, K.-J. Berg, and H. Graener, Appl. Phys. B 71, 795 (2000).
[46] A. Podlipensky, A. Abdolvand, G. Seifert, and H. Graener, Appl. Phys. A 80, 1647
(2005).
[47] A. Stalmashonak, G. Seifert, and H. Graener, Opt. Lett. 32, 3215 (2007).
[48] M. Sendova-Vassileva, M. Sendova, and A. Troutt, Appl. Phys. A 81, 871 (2005).
[49] M. Sendova, M. Sendova-Vassileva, J.C. Pivin, H. Hofmeister, K. Coffey, and A.
Warren, J. Nanosci. Nanotechnol. 6, 748 (2006).
[50] A. Crespo-Sosa, P. Schaaf, J.A. Reyes-Esqueda, J.A. Seman-Harutinian, and A. Oliver,
J. Phys. D: Appl. Phys. 40, 1890 (2007).
[51] Halabica, R.H. Magruder III, and R.F. Haglund Jr. Proc. SPIE 6458, 64581O (2007).
[52] W.M. Tsang, A.A.D.T. Adikaari, V. Stolojan, and S.R.P. Silva, J. Vac. Sci. Technol. B
26, 860 (2008).
[53] A. Halabica, J.C. Idrobo, S.T. Pantelides, R.H. Magruder III, S.J. Pennycook, and R.F.
Haglund Jr., J. Appl.Phys. 103, 083545 (2008).
[54] P.B. Comita, E. Kay, R. Zhang, and W. Jacob, Appl. Surf. Sci. 79/80, 196 (1994).
[55] M. Kaempfe, H. Graener, A. Kiesow, and A. Heilmann, Appl. Phys. Lett. 79, 1876
(2001).
[56] P. Zijlstra, J.W.M. Chon, and M. Gu, Opt. Express 15, 12151 (2007).
[57] S.Y. Park, T. Isobe, M. Senna, R.A. Weeks, and R.A. Zuhr, Appl. Phys. Lett. 73, 2687
(1998).
[58] N. Can, P.D. Townsend, and D.E. Hole, Appl. Phys. Lett. 65, 1871 (1994).
[59] N. Can, P.D. Townsend, D.E. Hole, H.V. Snelling, J.M. Ballesteros, and C.N. Afonso,
J. Appl. Phys. 78, 6737 (1995).
[60] G. Battaglin, E. Cattaruzza, C. de Julian Fernandez, G. De Marchi, F. Gonella, G.
Mattei, C. Maurizio, P. Mazzoldi, A. Motello, C. Sada, and F. D’Acapito, Nucl. Instr.
Meth. B 175/177, 410 (2001).
[61] G. Battaglin, E. Borsella, E. Cattaruzza, F. Gonella, R.F. Haglund Jr., G. Mattei, P.
Mazzoldi, D.Z. Osborne Jr, and R. Polloni, Nucl. Instr. Meth. B 141, 274 (1998).
[62] J.D. Demaree, S.R. Kirkpatrick, A.R. Kirkpatrick, and J.K. Hirvonen, Nucl. Inst. Meth.
B 127/128, 603 (1997).
[63] C.G. Lee, Y. Takeda, N. Kishimoto, and N.J. Umeda, J. Appl. Phys. 90, 2195 (2001).
[64] V.B. Odzhaev, I.P. Kozlov, V.N. Popok, and D.V. Sviridov, Ion Implantation into
Polymers, (Belarusk. State. Univ.: Minsk, 1998). [in Russian].
[65] G.G. Zakirov, I.B. Khaibullin, and M.M. Zaripov, Sov. Phys. Semmicond. 17, 150
(1983).
[66] R.H. Magruder III, R.F. Haglund Jr., L. Yang, J.E. Wittig, and R.A. Zuhr, J. Appl.
Phys. 76, 708 (1994).
[67] S. Deying, Y. Saito, and S. Suganomata, Jpn. J. Appl. Phys. 33, L966 (1994).
[68] A.L. Stepanov, S.N. Abdullin, R.I. Khaibullin, V.F. Valeev, Yu.N. Osin, V.V. Bazarov,
and I.B. Khaibullin, Mater. Res. Soc. Symp. Proc. 392, 267 (1995).
[69] V.N. Popok, R.I. Khaibullin, A. Toth, V. Beshliu, V. Hnatowicz, and A. Mockova,
Surf. Sci. 532/535, 1034 (2003).
[70] A.L. Stepanov, R.I. Khaibullin, S.N. Abdullin, Yu.N. Osin, and I.B. Khaibullin, Pros.
Inst. Phys. Conf. Ser. 147, 357 (1995).
[71] A.L. Stepanov, D.E. Hole, and P.D. Townsend, J. Non.-Cryst. Solids 244, 275 (1999).
[72] A. Cintors-Gonzalez, D. Muller, C. Estournes, M. Richard-Plouet, R. Poinsot, J.J. Grob,
and J. Guille, Nucl. Instr. Meth. B 178, 144 (2001).
[73] D.E. Hole, A.L. Stepanov, and P.D. Townsend, Nucl. Instr. Meth. B 148, 1054 (1998).
[74] D. Baüerle, Laser processing and Chemistry (Springer, Berlin, 1996).
[75] J.-Y. Bigot, J.C. Merle, O. Cregut, and A. Daunois, Phys. Rev. Lett. 75, 4702 (1995).
[76] K. Arai, H. Imai, H. Hosono, Y. Abe, and H. Imagawa, Appl. Phys. Lett. 53, 1891
(1988).
[77] Buerhop, B. Blumenthal, R. Weissmann, N. Lutz, and S. Biermann, Appl. Surf. Sci. 46,
430 (1990).
[78] G. Mie, Ann. Phys. 25, 377 (1908).
[79] C.F. Bohren, and D.R. Huffman, Absorption and Scattering of Ligth by Small Particles
(John Willey and Sons, New York, 1983).
[80] T. Castro, R. Reifenberger, E. Choi, and R.P. Andres, Phys. Rev. B 42, 8548 (1990).
[81] J. Basbach, D. Martin, F. Stietz, T. Wenzel, and F. Trager, Appl. Phys. Lett. 74, 2605
(1999).
[82] H. Itoigawa, T. Kamiyama, and Y. Nakamura, J. Non.-Cryst. Solids 210, 95 (1997).
[83] A.L. Stepanov and D.E. Hole, Phil. Mag. Lett. 82, 149 (2002).
[84] A. Takami, H. Yamada, K. Nakano, and S. Koda, Jap. J. Appl. Phys. 35, L781 (1996).
[85] A.L. Stepanov, D.E. Hole, and P.D. Townsend, Nucl. Instr. Meth. B 149, 89 (1999).
[86] M. Quinten, Z. Phys. B 101, 211 (1996).
[87] L.A.A. Pettersson and P.G. Snyder, Thin Solid Films 270, 69 (1995).
[88] Aden and M. Kerker, J. Appl. Phys. 22, 1242 (1951).
[89] J. Sinzig, U. Radtke, M. Quinten, and U. Kreibig, Z. Physik D 26, 242 (1993).
[90] J. Sinzig and M. Quinten, Appl. Phys. A 58, 157 (1994).
[91] P.W. Wang, Appl. Surf. Sci. 120, 291 (1997).
[92] A.L. Stepanov and V.N. Popok, Surf. Sci. 566/568, 1250 (2004).
[93] D.O. Hendersen, R. Mu, M.A. George, A. Burger, S.H. Morgan, C.W. White, R.A.
Zhur, and R.H. Magruder III, J. Vac. Soc. Technol. B 13, 1198 (1995).
[94] D.O. Hendersen, R. Mu, A. Ueda, Y.S. Tung, C.W. White, R.A. Zhur, and J.G. Zhu, J.
Non.-Cryst. Solids 205/207, 788 (1996).
[95] A.A. Bukharaev, V.M. Janduganov, E.A. Samarsky, and N.V. Berdunob, Appl. Surf.
Sci. 103, 49 (1996).
[96] G. Steiner, M.T. Pham, C. Kuhne, and R. Salzer, Fresenius. J. Anal. Chem. 362, 9
(1998).
[97] M.T. Pham, W. Matz, and Seifarth, Anal. Chem. Acta. 350, 209 (1997).
[98] E.D. Palik, Handbook of Optical Constants of Solids (Academic Press, London, 1997).
[99] P.B. Johnson and R.W. Christy, Phys. Rev. B 6, 4370 (1972).
[100] P.V. Kamat, M. Flumian, G.V. Harland, J. Phys. Chem. B 102, 3123 (1998).
[101] S. Link, C. Burda, M.B. Mohamed, B. Nikoobakht, and M.A. El-Sayed, J. Chem. Phys.
B 103, 1165 (1999).
[102] S. Link and M.A. El-Sayed, Annu. Rev. Phys. Chem. 54, 331 (2003).
[103] K. Dick, T. Dhanasekaran, Z. Zhang, and D. Meisel, J. Am. Chem. Soc. 124, 2312
(2002).
[104] Dalacu and L. Martinu, Appl. Phys. Lett. 77, 4283 (2000).
[105] A.L. Stepanov, V.N. Popok, V.B. Odzhaev, J. Appl. Spectr. 69, 97 (2002).
[106] A.C. Zettlemoyer, Nucleation (Pergamon, New York, 1969).
Chapter 3
SINGLE CRYSTAL PHOTO-ELASTIC MODULATORS
F. Bammer
Vienna University of Technology,
Institute for Forming and High Power Laser Technology, Vienna, Austria
ABBREVIATIONS
VIS visible light
NIR near infrared light
MIR mid infrared light
PEM Photo-Elastic Modulator
SCPEM Single Crystal Photo-Elastic Modulator
DMSCPEM Dual Mode Single Crystal Photo-Elastic Modulator
PRF Pulse Repetition Frequency
1. BASICS OF SINGLE CRYSTAL PHOTO-ELASTIC MODULATORS

(SCPEM)
1.1. Conventional Photo-Elastic Modulators
Conventional photo-elastic modulators (PEMs, Figure 1) modulate the polarization of a

light beam and are mainly used in ellipsometry in the form of Kemp-modulators
[1, 2, 30, 32].
They are made of a piece of optical glass which is glued to a quartz-crystal. Both pieces
are adjusted to have the same longitudinal resonance frequency. When the quartz-crystal is
electrically excited with the proper frequency the system will start a strong resonant
oscillation.
72 F. Bammer
Figure 1. A conventional photo-elastic modulator (PEM) [1].
Since damping is low the strain amplitudes become even at low voltage amplitudes so
large that a significant artificial birefringence modulation due to the photo-elastic effect is
caused. Polarized light passing the glass will experience a strong modulation of its
polarization.
High precision is needed to adjust the system such that it oscillates with high merit.
Furthermore the device is very large compared to its useful aperture.
In [13] an advanced design for a conventional PEM is proposed, where small actuators
are glued on the sides of the glass piece.
1.2. Definition and Description of a SCPEM
A Single Crystal Photo-Elastic Modulator (SCPEM) is a piezoelectric optical transparent

crystal that is electrically excited on one of its resonance frequencies. A further necessary
feature is that the polarization of light of any wavelength passing this crystal must not be
changed when the crystal is at rest. This for example cannot be fulfilled with the quartz
crystal of a conventional PEM, due to its optical activity [17].
In [20] many possible configurations are described. SCPEMs for the MIR are based on
43m-crystals and are experimentally and theoretically described in [36]. The most
advantageous configuration for visible and NIR-light is sketched in Figure 2.
It is based on a crystal of the uniaxial crystal symmetry class 3m. The light travels along
the optical axis ( = z-axis), the exciting electrical field points into the y-direction, and the
longitudinal x-oscillation is used. In most cases a polarizer (analyzer) oriented at 45° is placed
behind the modulator. Figure 2 further shows a photo diode to get a transmission signal and a
resistor (here with 100 Ω) to generate a measure for the piezo-electric current generated by
the crystal.
Single Crystal Photo-Elastic Modulators 73
Figure 2. A Single Crystal Photo-Elastic Modulator (SCPEM) made of a 3m-crystal. This choice of the
axes and the polarizer orientation is the standard configuration in this work.
We define now for monochromatic light with the wavelength λ the retardation as (for a
more general description consult appendix 5.3.1, Eq. A-4)
L (n x − n y )
δ = 2π (1)
λ
where L is the z-dimension of the crystal and nx, ny are the refractive indices for x- and
y-polarized light. They are calculated in chapter 2.2 for the photo-elastic effect and in
chapter 2.6 for the electro-optic effect.
The transmission T(δ) for the situation of Figure 2 is
T(δ) = cos2(δ/2) = (1 + cos δ)/2 . (2)
This is deduced in appendix 5.4.

45°-polarized input is used for Q-switching, pulse picking and optical switching.
For left or right circularly polarized input one has simply to replace δ by δ ± π/2 in Eq. 2:
T(δ) = (1 + cos (δ ± π/2))/2 (3)
This is depicted in Figure 13. For 0.25 < T < 0.75 a nearly linear dependence on δ is
found. Hence in this region a change in transmission is proportional to a change of
retardation, which is linear dependent on deformation. We obtain ΔT = ± Δδ/2 at δ = 0. Hence
small variations of the transmission are directly proportional to the deformation of the crystal.
Circularly polarized input is used for time-multiplexing and for measurements on the
SCPEM [3, 21].
74 F. Bammer
1.3. Two Candidates for a SCPEM: LiTaO3 and LiNbO3
The two most important exponents of the crystal class 3m are Lithium-Niobate (LiNbO3)
and Lithium-Tantalate (LiTaO3). Both are suited for the application as a SCPEM. The left
picture in Figure 3 shows a LiTaO3-crystal with dimensions 13.2 x 7.1 x 5.5mm and
resonance frequency 199 kHz and its holder.
Figure 3. left: SCPEM made of LiTaO3, 13.2x7.1x5.5mm, first x-resonance at 199 kHz; right: LiNbO3-
SCPEM mounted such that only the x-oscillation is present, 15.7x7.1x5.05mm, 192.3 kHz.
LiNbO3 (on the right in Figure 3 in a fixed mounting that allows only the longitudinal
mode to oscillate) has the disadvantage of having a much higher natural birefringence than
LiTaO3.
Therefore the acceptance angle is much smaller for LiNbO3. Its only advantage is that it
offers a ~20% higher velocity of longitudinal acoustic waves. In applications where high
resonance frequencies in the MHz-region are required (e.g. pulse-picking, chapter 4.3)
LiNbO3 crystals can be cut larger than LiTaO3-crystals and therefore allow better handling.
However, as discussed in chapter 2.9, the acceptance angle of LiTaO3 is much higher
such that the product of aperture and acceptance angle (the analog to the beam parameter
product) is much more favorable for LiTaO3.
Resonant operation of LiNbO3 is used as an acousto-optical modulator in [15] for active
laser mode-locking. There the diffractive effect of a high order longitudinal y-oscillation
(chapter 2.3) was used.
2. THEORY OF A SCPEM
In the following a 1-dimensional model for a SCPEM is discussed. Two cases will be
analyzed, namely first that the exciting electrical field is parallel to the direction of
propagation of the excited eigenmode (as it is the case in chapter 4.1 [3]) and second that both
directions are perpendicular to each other (as in Figure 2).
2.1. Reduction to One Dimension: Y-Excited X-Oscillation
We consider a y-excited oscillation in x-direction (Figure 2). For this purpose it is

necessary to reduce the constitutive equations A-1 in appendix 5.1 to a 1-dimensional
representation on the base of plausible but not fully fulfilled assumptions. The crystal is cut
along the x-, y- and z-axes and the largest dimension is found in x-direction 1. In this case all
tensions besides σ1 can be set approximately equal to zero. The calculation in chapter 2.4 will
show that the first set of constitutive equations in Eq. A-1 directly leads to the solution. In
these 9 equations with altogether 18 dependent and independent variables we set σ2-6 = 0.
Further it is assumed that the electrodes are placed only at the y-facets and hence only the
electric displacement in y-direction is unequal to 0. Therefore we set D1 = D3 = 0. Herewith
only 11 variables remain. Further we solve the 9 equations of state for the 9 variables σ1, D2,
ε2-6, E1, E3, which are calculated in dependence on the two left variables ε1, E2. At the end the
solutions for σ1 and D2 are important, which yields the following equations:
tension: σ1 = C ε1 – e E2 [N/m²] (4)
electric displacement: D2 = e ε1 + χ E2 [C/m²]
Here C, e, χ are the searched coefficients. In case of LiTaO3 the above procedure yields:
C = 202.813 GPa, e = 1.59482 C/m², χ = 4.62033 10 −10 As/Vm.
The stiffness C is the equivalent to the E-modulus for the standard solution of
longitudinal oscillations in a bar made of isotropic material. Further e is a piezo-electric and χ
is a dielectric coefficient.
For the calculation of the photo-elastic effect in chapter 2.2 the other strains are needed,
too. For LiTaO3 they are given by:
y-strain: ε2 = −0.116 ε1 + 6.95 10 −12 E2 (5)
z-strain: ε3 = −0.261 ε1 − 2.05 10 −12 E2 (6)
yz-strain: ε4 = 0.175 ε1 + 24.5 10 −12 E2 (7)
zx- and xy-strain: ε5 = ε6 = 0 (8)
The electro-optic contribution of E2 will be neglected in the further discussion since in

resonance the photo-elastic contribution is by three orders of magnitude higher (chapter 2.5
vs. 2.6).
The above calculation is needed in the chapters 2.2 and 2.5. A similar procedure should
be carried out for the case considered in chapter 2.3.
1
Otherwise the x-oscillation is governed by nontrivial plate oscillation modes and surface waves. Further
significant tensions perpendicular to the x-axis are found [7].
76 F. Bammer
2.2. The Effect of an X-Oscillation on Birefringence
We consider now the photo-elastic effect for a resonant x-oscillation under the
assumptions made in the previous chapter. We use the syntax explained in appendix 5.3. The
effect of strain on birefringence is described by Eq. A-5 in appendix 5.3.2. For the considered
resonance ε5 = ε6 = 0 holds. The electro-optic effect can be neglected and we set therefore
E = 0. Further we use the photo-elastic matrix PE for the crystal class 3m [17, 18, 27]. The
artificial birefringence vector B for a resonant x-oscillation is now given by:
⎛ B1 ⎞ ⎛ p11E p12E p13E p14E 0 ⎞ ⎛ ε 1 ⎞ ⎛ p11E ε 1 + p12E ε 2 + p13E ε 3 + p14E ε 4 ⎞

0
⎜ ⎟ ⎜ E ⎟⎜ ⎟ ⎜ ⎟
⎜ B2 ⎟ ⎜ p12 p11E p13E − p14E
0 0 ⎟ ⎜ ε 2 ⎟ ⎜ p12E ε 1 + p11E ε 2 + p13E ε 3 − p14E ε 4 ⎟
⎜ B3 ⎟ ⎜ p E 0 ⎟⎟ ⎜ ε 3 ⎟ ⎜⎜ ⎟
p31E p33E 0 0 p31E ε 1 + p31E ε 2 + p33E ε 3 ⎟
B=⎜ ⎟=⎜ E ⎜ ⎟=
31
⎜ B4 ⎟ ⎜ p 41 − p 41E
0 p 44E
0 0 ⎟ ⎜ε 4 ⎟ ⎜ E
p 41 ε 1 − p 41E ε 2 + p 44E ε 4 ⎟
⎜B ⎟ ⎜ E E ⎟⎜ ⎟ ⎜ ⎟
⎜ 5⎟ ⎜ 0 0 0 0 p 44 p 41 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟
⎜B ⎟ ⎜ 0
⎝ 6⎠ ⎝ 0 0 0 p14E p 66E ⎟⎠ ⎜⎝ 0 ⎟⎠ ⎜⎝ 0 ⎟
⎠
⎛ 1 ⎞ ⎛ 1 ⎞ ⎛ 1 ⎞
Indicatrix : x 2 ⎜⎜ 2 + B1 ⎟⎟ + y 2 ⎜⎜ 2 + B2 ⎟⎟ + z 2 ⎜⎜ 2 + B3 ⎟⎟ + 2 B4 y z = 1
n
⎝ o ⎠ n
⎝ o ⎠ n
⎝ e ⎠
The original indicatrix is an ellipsoid of evolution with the rotational axis in z-direction.
In z-direction the extension is ± ne and the radius of the circular cross-section in the xy-plane
is equal to no (appendix 5.3.1).
The coefficient 2B4 causes in the yz-plane a small rotation by the angle
B4 /(no−2 − ne−2) (9)
This, however, has no influence for the following calculation and will be treated later.
For z-parallel rays the coefficients B1 and B2 are decisive. The corresponding cross-
section-ellipse (appendix 5.3.1) is described by
⎛ 1 ⎞ ⎛ 1 ⎞
x 2 ⎜⎜ 2 + B1 ⎟⎟ + y 2 ⎜⎜ 2 + B2 ⎟⎟ = 1
⎝ no ⎠ ⎝ no ⎠
Figure 4 shows its shape for the deformed and undeformed case.
According to appendix 5.3.1 in z-direction only x- and y-polarized light can exist. The
refractive index for x-polarized light going in z-direction is the square root of the inverse of
the coefficient of x2 (assumption: B1 << 1):
nx = −2
1
= no
1
2
2
( 2
≈ no 1 − no B1 ≈ no 1 − no B1 / 2 = no − no B1 / 2) 3
no + B1 1 + no B1
The analog holds for ny. The difference of the refractive indices is now given by:
no ³
Δn = n x − n y = (B2 − B1 )
2
=
no ³
2
(
( p12E − p11E )(ε 1 − ε 2 ) − 2 p14E ε 4 ) (10)
n ³
(
= o 1.116 ( p12E − p11E ) − 0.35 p14E ε 1
2
)
Here we used ε2 = −0.116 ε1 (Eq. 5) and ε4 = 0.175 ε1 (Eq. 7), which holds for LiTaO3.
For B4 we obtain
B4 = p 41
E
(
ε 1 − p 41E ε 2 + p 44E ε 4 = 1.116 p 41E + 0.175 p 44E ε 1 ) (11)
Light, linearly polarized at 45°, will be split in the crystal in an x- and a y-component of
equal amplitude but with different velocities c0/nx and c0/ny. The path difference between the
two waves is equal to L Δn (L...path length). The superposition of the two waves at the exit
surface of the crystal yields in general a light wave with elliptic polarization. If the path
difference is equal to the quarter (half) of the wavelength circularly (−45° linear) polarized
light results.
Figure 4. Deformation of the indicatrix in the xy-plane by an x-strain.
2.3. Longitudinally Excited Oscillation
It is assumed that the electrical field points in y-direction and excites a longitudinal
y-oscillation in the crystal class 3m 2 (Figure 5). The crystal extension in x-, y- and z-direction
is l, b, L.
2
If the field points in x-direction an x-propagating shear oscillation (used in [3]), which is mathematically
equivalent, can be excited. (appendix 5.2)
78 F. Bammer
Figure 5. Excitation of a y-longitudinal oscillation.
The following equations of state are advantageous:
y-tension: σ=Cε–hD [N/m²] (12)
electrical field (in y-direction): E = −h ε + χ−1D [V/m] (13)
ε = vy … y-strain [-],
v(t, y) … y-shift [m],
C … y-stiffness constant [N/m²],
U (E = Uy)…electric potential (voltage) [V]
h… piezo-electric stress modulus [N/C],
χ…dielectric coefficient [As/Vm],
D…electric y-displacement [C/m²]
D is independent of y, i.e. Dy = 0, because every dy-layer generates a current in y-direction.

According to Kirchhoff this series connection of all dy-layers must result in the same current
for all layers.
Further we define with the crystal density ρ [kg/m³] an acoustic velocity c [m/s] with
c2 = C/ρ.
Herewith a partial differential equation for v(t, y) can be derived. The starting point is:
density x acceleration = − y-derivative of the tension
ρ vtt = −σy = − C vyy − h Dy = − C vyy Æ vtt = c2 vyy (14)
We obtain the 1-dimensional linear wave-equation. Here the advantage of the chosen
constitutive equations becomes obvious. This choice introduces Dy, which is zero, into the
partial differential equation in Eq. 14. Another choice of constitutive equations would
incorporate Ey, which in this case is not zero.
In case of an alternating excitation D writes with the complex amplitude D0 as 3:
D(t, y) = D(t) = D0 exp( j ω t), (15)
The boundary conditions are:
y= b/2: σ(t, b/2) = −M vt and U(t, b/2) = U0 exp( j ω t) (16)
y = −b/2: σ(t,−b/2) = M vt and U(t,−b/2) = 0 (17)
The friction coefficient M describes external damping. Inner damping will not be
considered. U0 and ω are the amplitude and the angular frequency of the harmonic alternating
voltage at the electrodes with electrode area AE = L l.
For ω = 0 a static strain εstat, a static shift vstat = b εstat/2 of the crystal ends, a static charge
Qstat, and a static electric displacement Dstat = Qstat /AE are caused. Further σ = 0 holds. This,
placed into Eq. 12 and Eq. 13, leads to
y-tension: 0 = 2 C vstat/b – h Qstat/AE ,
Electrical field (in y-direction): U0/b = −2 h vstat/b + χ−1Qstat/AE .
This, solved for Qstat and vstat, yields:
U0 χ AE χ h2 (18)
v stat = , Qstat = U 0 = C stat U 0 with H=
2 (−h + χ C h )
−1 −1
1− H b C
The dimensionless number H must fulfill H < 1, because otherwise Qstat can become
infinite. In fact H << 1 holds such that it is safe to use 1 + H instead of 1/(1 − H).
Further a static (low frequency) capacity Cstat was defined. Additionally we define the
dynamic (high frequency) capacity Cdyn = AE χ /b.
For the dynamic analysis we define a normalized angular frequency
ωb
Ω := .
πc
This is related to the first resonance frequency fR1, which we expect at fR1 = c/(2b). The
corresponding angular resonance frequency is ωR1 = 2π fR1 = π c/b and hence ΩR1 = 1.
The constitutive Eq. 12 and the boundary condition Eq. 16 yield
C vy(t, b/2) + M vt(t, b/2) = h D(t). (19)
We use now the following ansatz for the y-deformation v(t, y), which fulfills the wave
equation 14 (Ω := ω b/(π c)...normalized angular frequency):
3
Of course in Eq. 15 and in all similar definitions, that use the term exp( j ω t), the real part is meant.
80 F. Bammer
sin(ω y / c)
v(t , y ) = v0 (Ω) exp( j ω t ) (20)
sin(Ω π / 2)
At y = b/2 this yields v(t, b/2) = vA(Ω) exp( j ω t). v0 is the complex oscillation amplitude
of the zx-surfaces. Ansatz 20, put in Eq. 19, yields:
v0 1 1 c
= = mit r := M (21)
v stat π ⎛ ⎛ π ⎞ ⎞ f ( Ω, r ) C
Ω ⎜⎜ cot ⎜ Ω ⎟ + j r ⎟⎟
2 ⎝ ⎝2 ⎠ ⎠
The dimensionless figure r is a measure for the damping rate. f (Ω, r) ≈ 1 holds for small
Ω and therefore v0 = vstat for small excitation frequencies. At the resonance frequencies
Ω = 1,3,5...the displacement amplitude v0 becomes infinite for r = 0.
If Eq. 13 is y-integrated from 0 to l/2 one obtains with Eq. 18:
U0 /2 = −h v0 + χ-1 D0 b / 2 Æ D0 = χ (U0 + 2 h v0)/b (22)
with the complex current amplitude I0(Ω) we write for the current I(t) generated by the crystal
I(t) = AE dD(t)/dt = I0(Ω) exp( j ω t)
We obtain with Eq. 21 and Eq. 22
I 0 (Ω ) ⎛ H 1 ⎞
= j Ω ⎜⎜1 + ⎟⎟
ω R1 C dyn U 0 ⎝ 1 − H f (Ω, r ) ⎠
The current amplitude, too, has resonances at Ω = 1,3,5... . Figure 7 shows the typical
course of amplitude and phase of the normalized deformation and current at the first
resonance.
2.4.Transversally Excited Oscillation
Now we assume an electrical field in y-direction and a wave propagation in x-direction

(Figure 6). The configuration is the same as in Figure 5. Whether the excitation is transversal
or longitudinal depends here on the eigenmode that is excited and herewith on the frequency
applied to the crystal electrodes. The crystal extension in x-, y- and z-direction is l, b, L.
The following constitutive equations are advantageous:
x-tension: σ = C ε – e E [N/m²] (23)
Electric y-displacement: D = e ε + χ E [C/m²] (24)
ε = vx … strain [-],
U ... electrical potential [V],
E = Uy … electrical y-field [V/m],

χ…dielectric coefficient [As/Vm],
v … x-displacement [m],
C … x-stiffness [N/m²],
e…piezo-electric coefficient [N/(Vm) or C/m²]
Obviously ε and D (for the same reason as in the previous case), and due to Eq. 24 the
electric field E, too, do not depend on y. Additionally we assume that, like the potential U, the
electrical field E does not depend on x. Therefore we get Ex = Ey = 0. Similar to the previous
chapter one obtains:
ρ vtt = − σx = − C vxx + e Ex = −C vxx Æ vtt = c2 vxx (25)
with c2 = C/ρ ; ρ…density (26)
The boundary conditions are:
x = l/2: σ(t, l/2) = −M vt (27)
x = −l/2: σ(t, −l/2) = M vt (28)
y = b/2: U(t, b/2) = U0 exp( j ω t) (29)
y = −b/2: U(t, −b/2) = 0 (30)
The friction coefficient M describes external damping. U0 and ω are the amplitude and
the angular frequency of the alternating voltage at the electrodes.
Figure 6. Transversal excitation of a longitudinal x-oscillation.
For ω = 0 a static strain εstat, a static shift vstat = l εstat/2 of the crystal ends, a static charge
Qstat, and a static electric displacement Dstat = Qstat /L/l are caused. Further σ = 0 holds. This,
placed into Eq. 25 and Eq. 26, leads to
82 F. Bammer
x-tension: 0 = 2 C vstat/l – e U0/b
electric y-displacement: Qstat/L/l = 2 e vstat/l + χ U0/b
This, solved for Qstat and vstat, yields:
e l/2 lL e2
v stat = U0 , Qstat = (1 + Λ) χ U 0 = C stat U 0 with Λ = (31)
C b b Cχ
The dimensionless characteristic number Λ is a measure for the piezo-electric coupling.

Further a static (low frequency) capacity Cstat was defined. Additionally we define the
dynamic (high frequency) capacity Cdyn = l L χ /b.
For the dynamic analysis we define a normalized angular frequency
ωb
Ω := (32)
πc
This is related to the first resonance frequency fR1, which we expect at fR1 = c/(2l). The
corresponding angular resonance frequency is ωR1 = 2π fR1 = π c/l and hence ΩR1 = 1.
In case of an alternating voltage the electrical field is described by:
E(t, x, y) = E(t) = E0 exp( j ω t) = U0 exp( j ω t)/b
Eq. 23 and the boundary condition Eq. 27 yield:
C v x (t , l / 2) + M vt (t , l / 2) = e E (t ) (33)
In Eq. 33 we use for the shift v(t, x) an ansatz that fulfills Eq. 25 and has the additional
property v(t, l/2) = v0 exp(j ω t). (Ω:= ω l/(π c)...normalized angular frequency):
sin ω x / c
v(t , x) = v0 exp( j ω t ) (34)
sin Ω π / 2
With this the amplitude v0 of the shift v(t, x) in dependence on Ω is immediately obtained:
v0 1 1 c
= = mit r := M (35)
v stat π ⎛ ⎛ π ⎞ ⎞ f ( Ω, r ) C
Ω⎜⎜ cot⎜ Ω ⎟ + j r ⎟⎟
2 ⎝ ⎝2 ⎠ ⎠
The dimensionless characteristic number r is again a measure for the damping rate. Note
that v0 = 0 for high damping r → ∞.
For the application as a PEM the strain ε(x, t) = ε0(x) exp( j ω t) = dv(x, t)/dx in x-direction
is needed. This yields
ω cos(ω x / c )
ε 0 ( x ) = v0 .
c sin (Ω π / 2 )
At the first resonance with Ω = ΩR1 = 1 and ω = ωR1 = π c/l and f (1, r) = π r/2 we obtain
2 v stat
ε 0 R1 ( x) = cos(π x / l ) (36)
lr
This will be used in chapter 2.12 to calculate the useful aperture of the SCPEM.
For the dielectric displacement D(t, x) Eq. 24 yields
⎛ ω cos ω x / c U 0 ⎞
D(t , x) = ⎜⎜ e v0 − ⎟ exp( j ω t ) .
⎝ c sin Ω π / 2 b ⎟⎠
with the complex current amplitude I0 we write for the current I(t) generated by the crystal
l/2
dD(t , x)
I (t ) = L ∫
−l / 2
dt
dx = I 0 exp( jω t ) .
The evaluation of this leads to
1 I0 1 1 ⎛ Λ ⎞
= = j Ω ⎜⎜1 + ⎟ (37)
ω R1 C dyn U 0 ω R1 C dyn Z ⎝ f (Ω, r ) ⎟⎠
We defined here the impedance Z = U0/I0 of the crystal. It is a common measure to

describe the electric behavior of piezo-electric oscillators (chapter 2.7, [23]).
2.5. Discussion and Evaluation
For both cases the same solutions are obtained at the end if Λ and H/(1−H) are exchanged
with each other. Since usually H << 1 it is possible to use H instead of H/(1−H). In the
following we will point out some interesting properties of the solutions.
The Ω-derivative of the normalized current (Eq. 37) at Ω = 0 yields a gradient of 1 + Λ,
while at high frequencies, where the resonances are hardly seen due to damping, the gradient
is equal to 1. Hence at low frequencies the capacity of the crystal is higher by a factor 1 + Λ,
which is found in Eq. 31 for the static charge Qstat, too.
Resonance frequencies are found at Ω = 1,3,5... . We discuss the first one at Ω = 1. The
resonance enhancement of the deformation is equal to 2/π/r (Eq. 34: Ω → 1), for the current it
is 2Λ/π/r (Eq. 37: Ω → 1).
The full width at half maximum (FWHM) is given by 2r. At Ω = 1 ± r the amplitude of
the shift and of the current is only half of the maximum value at Ω = 1. This is important
84 F. Bammer
since a simple frequency measurement at the maximum current amplitude and at the half of
this maximum yields the damping rate r. This is easier than to deduce a value from literature.
Typically our crystals show a full width at half maximum of 1/5000 of the resonance-
frequency. Therefore we assume r = 10−4.
Within the FWHM-bandwidth the phase of the displacement covers the largest part of its
way from phase 0° previous to the resonance to phase –180° behind resonance.
The current course shows at 4
Ω = 1 + 4Λ /(4Λ + π2)
a significant minimum equal to the inverse of the resonance enhancement r π/Λ/2. This is the
so called anti-resonance.
At the main and anti-resonance the phase of the current crosses zero. Within this
frequency range the phase is < 0 in the vicinity of –90°. Previous and behind to that the phase
quickly goes to 90°, i.e. away from the resonance the current advances the voltage by 90°
which indicates purely capacitive behavior.
We analyze now the solution for the longitudinal x-oscillation for LiTaO3 (Figure 2). In
chapter 2.1 the values C = 202.813 GPa, e = 1.59482 C/m² and χ = 4.62033 10−10 As/Vm
were found for this case. Herewith one obtains with Eq. 31 Λ = 0.027143. This yields a
resonance enhancement equal to 2/π/r = 6366 for the displacement and 2Λ/π/r = 173 for the
current amplitude. Figure 7 shows the frequency dependence of the normalized displacement
and the normalized current. The curves are plotted with an increased damping rate r = 2000,
since otherwise the resonance peaks become so sharp that they cannot be nicely depicted
together with the antiresonance. The characteristic course of the amplitudes and the phases is
experimentally well documented (chapter 3, [21]).
Up to now no information of the crystal dimensions has entered the calculations. These
are necessary only if real values, e.g. the resonance frequency or the crystal capacity, are
calculated. The dimensions of the specimen made of LiTaO3 (left picture of Figure 3) are
l = 13.2mm, b = 7.1mm and L = 5.5mm and its density is ρ = 7465 kg/m³. With that the static
displacement of the end surfaces is
v stat e l/2
= = 7.31 pm/V.
U0 C b
Eq. 36 yields for the resonance amplitude of the x-strain ε0R1 at x = 0:
ε 0 R1 (0) 2 v stat e
= = [V-1]
U0 l rU0 C b r
4
This is an approximate formula for small r and small Λ. It is derived with the series expansion of the current
amplitude course at Ω = 1 and r = 0. Without damping the current is exactly zero at the anti resonance. Hence
the first two terms of this expansion are set equal to zero, which immediately leads to the result. Since the
damping number r is very small in all cases of interest we neglect the small r-dependence.
Figure 7. Resonance behavior of a LiTaO3 crystal with Λ = 0.027143 and r = 1/2000; amplitude and
phase of the deformation amplitude (left, Eq. 35) and of the current amplitude (right, Eq. 37).
Herewith the necessary voltage amplitude can be calculated to obtain in the resonance at
x = 0 a retardation amplitude δ0 = π (half wave retardation). For that we use Eq. 10 from
chapter 2.2 with ε = ε1.
Note here that δ, Δn depend as ε on x and t. The complex amplitude at the first resonance
is indexed with 0R1 and depends only on x. The additional index λ/2 means that half wave
retardation is achieved.
For the retardation amplitude δ0R1 at the first resonance at x = 0 the following is found:
3
δ 0 R1 (0) = 2 π
L
λ
Δn0 R1 (0) = 2 π
L no
λ 2
(
1.116( p12E − p11E ) − 0.35 p14E
eU 0
Cbr
)
Herewith we obtain a half wave amplitude U0R1,λ/2 (i.e. δ0R1(0) = π) for λ = 1064nm 5,
no = 2.18, p11E = −0.081, p12E = 0.081, p14E = −0.026, p41E = −0.085, p44E = 0.028
(appendix 5.5.3):
Cbr λ 1 1
U 0 R1,λ / 2 = 3
= 8.66 V
1.116 e L no p11 − p12E
E
Indeed a resonance retardation of δ0R1(0)/U0 = 0.4 V−1 was measured (chapter 3, [21]) 6.
Therefore the half wave amplitude is equal to π /0.4 = 8 V, which is a good coincidence.
With U0R1,λ/2 = 8.66 V we obtain the half wave x-strain ελ/2 = 0.97 10−4 and the half wave
x-stress σλ/2 = C ελ/2 = 20 MPa or N/mm² 7.
The dynamic capacity for high frequencies far above the first resonance frequency is
5
Here we choose the Nd:YAG-laser wavelength λ = 1064nm, since we used an Nd-YAG-fiber laser in [5, 6].
6
The 1st peak in 3rd graph in Fig. 6 in [1] shows a transmission amplitude T0/U0 = 0.2 [V−1]. Since in this linear
region T0 = ± δ0/2 holds (last lines of chapter 1.2), we obtain δ0/U0 = 0.4 [V−1].
86 F. Bammer
lL
C dyn = χ = 4.72 pF
b
and the static one is higher by a factor 1 + Λ and equal to Cstat = 4.85 pF.
The acoustic velocity of the longitudinal wave in x-direction is
C
c= = 5212 m/s
ρ
and the first theoretic resonance frequency is:
c
f res = = 197.44 kHz .
2l
The measured first resonance is found at 198.89 kHz. This small difference can be
explained firstly with diffuse literature values, and secondly with not fully fulfilled
assumptions for the calculation of C, e and χ in chapter 2.1.
2.6. Comparison with a Pockel’s Cell
The polarization of light can be controlled with a Pockel’s cell, as well. Such cells use the
electro-optic effect, i.e. the change of the index ellipsoid (appendix 5.3.2) of a medium by an
external electrical field [16, 17, 22, 27, 28]. One candidate for this application is LiNbO3 in
the configuration of Figure 2. Important dimensions of the crystal are the height b in
y-direction and the length L in z-direction. Appendix 5.3.2 (Eq. A-6) shows that, when the
crystal is free from tensions (T = 0) and when a static voltage U and herewith a static
electrical field E2 = U/b is applied, the change of the indicatrix is described by (the chosen
structure of the electro-optic matrix RT is that of the crystal class 3m)
⎛ B1 ⎞ ⎛ 0 − r22T r31T ⎞ ⎛ − r22T E2 ⎞

⎜ ⎟ ⎜ ⎟ ⎜ T ⎟
⎜ B2 ⎟ ⎜ 0 r22T r31T ⎟ ⎜ r22 E2 ⎟
⎜B ⎟ ⎜ 0 ⎟⎛ 0 ⎞ ⎜ ⎟
0 r33T ⎟ ⎜ ⎟ ⎜ 0 ⎟
⎜ 3⎟=⎜ ⎜ E2 ⎟ =
⎜ B4 ⎟ ⎜ 0 r15T 0 ⎟ ⎜ ⎟ ⎜ r15T E2 ⎟
⎜B ⎟ ⎜ T ⎟⎝ 0 ⎠ ⎜ ⎟
⎜ 5 ⎟ ⎜ r15 0 0⎟ ⎜ 0 ⎟
⎜ B ⎟ ⎜ − rT 0 0 ⎟⎠ ⎜ 0 ⎟
⎝ 6 ⎠ ⎝ 22 ⎝ ⎠
⎛ 1 ⎞ y2 ⎛ 1 ⎞ z2
Indikatrix : x 2 ⎜⎜ 2 + B1 ⎟⎟ + 2 ⎜⎜ 2 + B2 ⎟⎟ + 2 + 2 B4 y z = 1
⎝ no ⎠ no ⎝ no ⎠ ne
7
The half wave x-strain and x-stress depend on the material, on the length L of the crystal, on the wavelength λ, and
on the assumptions made in chapter 4.1 for a longitudinal x-oscillation. They do not depend on x, on an
oscillation or on the presence of a resonance. Therefore the index 0R1 is not used here.
The rijT are the electro-optic coefficients, here for the case that the tensions are constant
(in praxis equal to zero).
The coefficient B4 causes a small rotation of this ellipsoid in the yz-plane, which can be
neglected for z-paraxial rays in the first order.
For the effect of the coefficients B1 and B2 we can use the first line in Eq. 10:
3
no
Δn = n x − n y = (B2 − B1 ) = no 3 r22T E 2 = no 3 r22T U
2 b
Further the retardation δ is given by:
L L U T
δ = 2π Δn = 2 π no
3
r22
λ λ b
Now we set δ = π to obtain the action of a half wave plate and get with the dimensions of
the crystal on the left in Figure 3 (λ = 1064nm, L = 5.5mm, b = 7.1mm, LiNbO3: no = 2.244,
r22T = 6.7 10−12 m/V):
1 b λ 9.71 kV
Uλ/2 = 3 T
=
2 no L r22
This voltage is by three orders of magnitude higher than that needed by a SCPEM. Note
further that Uλ/2 is even higher for LiTaO3 since it shows in this configuration a much weaker
electro-optic effect than LiNbO3 (LiTaO3: r22T = 0.33 10−12 m/V −> Uλ/2 = 196 kV for the
crystal on the left in Figure 3).
It must be mentioned here that usually the length of a Pockel’s cell is much higher. With
e.g. L = 30mm the half wave voltage is 1.66 kV for LiNbO3.
2.7. Electrical Model
For calculations and simulations (e.g. with PSPICE) of the electrical interaction of the
crystal with an external circuit it is useful to define an electrical model as given in Figure 8,
which reflects the electrical behavior of the crystal at a resonance [23] 8.
The resistor Rind simulates a mechanical damping of the crystal while RΩ results from the
electrical resistance of the wires and the electrodes, which was not considered in the
analytical calculation before. The impedance Z of this system, i.e. the quotient of the complex
excitation amplitude and the resulting complex current amplitude for a given excitation
angular frequency ω is equal to:
8
In this citation [23] the capacity Cdyn is stated to be the static capacity. This is obviously wrong since at low
frequencies the inductivity acts as a short circuit such that the static capacity is the sum of both single
capacities. For high frequencies the inductivity blocks such that only Cdyn remains.
88 F. Bammer
1 ⎛ 1 ⎞
Z = RΩ + || ⎜⎜ + j ω Lind + Rind ⎟⎟
j ω C dyn ⎝ j ω (C stat − C dyn ) ⎠
|| means the total impedance of a parallel connection of the impedances left an right of ||,
i.e. Z1||Z2 = Z1Z2/( Z1+Z2). The resonance of the model is found at
1 .
f res =
2π Lind (C stat − C dyn )
Since Cstat, Cdyn and fres are known, Lind can be immediately calculated with this result. In
a similar way Rind is determined by the resonance peak.
If a replacement circuit for more resonance frequencies is needed, one simply connects
all corresponding electrical model circuits in parallel.
Figure 8. Electrical model of a piezo-electric crystal.
2.8. Mechanical Model

Alternatively a simple spring-mass-system can describe a resonance. It is assumed that

the spring is prepared with electrodes and elongates or contracts when a voltage is applied.
Hence the force F of the spring depends both on the deformation x and on the applied voltage
U. Additionally an electrical charge Q is generated. The analogs to Eq. 12 (or 23) and Eq. 13
(or 24) are
F=kx–eU
Q = e x + Ck U,
with k...spring constant, x...spring deformation or motion of the mass m, e...piezo-electric

coefficient, U...applied voltage, Ck...electrode capacity. The time-derivative of the charge
yields the current I
dQ/dt = Q’ = I = e x’ + Ck U’ ,
with ( )’ = d( )/dt. With the equation for the spring force F, which acts on the mass m, one
derives out of “mass x acceleration = force” the differential equation for the motion of the
mass m:
m x’’+ r x’ + k x = e U
Here an additional dissipative term r x’ is included. With the excitation U = U0 exp( j ω t) and
the ansatz
x = x0 exp( j ω t) and I = I0 exp( j ω t) ,
with x0, I0…complex deformation- and current amplitude and ω…excitation angular
frequency, one immediately obtains the amplitude- and phase course of current and
deformation. If the values for the constants are chosen properly one derives exactly the same
impedance as with the electrical model circuit. The advantage of the mechanical model is that
one additionally obtains information about the deformation in a simple way, which is decisive
for the use of a photo-elastic modulator.
2.9. Acceptance Angle of a SCPEM, 1-Dimensional View
A photo-elastic modulator needs exact adjustment. The maximum possible deviation of

optical rays from the optical axis is limited. We try to understand the effect of a skew passage
of an optical ray. For that we consider a paraxial (i.e. small deviation from the z-axis) light
ray in an uniaxial crystal.
Figure 9. Cross-section through the indicatrix of an uniaxial crystal.
Figure 9 shows the cross-section trough the indicatrix (here for the crystal class 3m an
ellipsoid having two equal equatorial semi-axes = spheroid) in the zx-plane and a light ray
(arrow), which travels in the zx-plane and includes with the z-axis an angle α. On the right the
corresponding cross section ellipse (definition in appendix 5.3.1) is shown.
90 F. Bammer
Which index of refraction is now relevant for this ray ? For this purpose position and
length of the main axis must be determined. Obviously one semi-axis is directed along the
y-direction and has the length no. Hence the velocity of y-polarized light is c0/no (c0…vacuum
light velocity) independent of α. The other main axis shows approximately in x-direction (for
α <<1) and its length is determined with the equation for the indicatrix in the zx-plane:
x2 z2
2
+ 2
=1
no ne
For the z-coordinate zα of the considered peak points zα = α no holds in first order and the
x-coordinate xα is then given by
α 2 no 2 .
xα = no 1 − 2
ne
2 2
The length no + Δn is now equal to xα + zα and for Δn we obtain
⎛ n 2 ⎞ α2 ⎛ no 2 ⎞
xα + zα − no = no 1 + α 2 ⎜⎜1 − o 2 ⎟ − no ≈ no ⎜1 − 2 ⎟.
2 2
Δn = ⎟ ⎜ n ⎟
⎝ ne ⎠ 2 ⎝ e ⎠
The retardation between x- and y-polarization is
L L ⎛ n 2⎞
δ = 2π Δn = π no α 2 ⎜⎜1 − o 2 ⎟⎟ .
λ λ ⎝ ne ⎠
Now we consider the case that the impinging light is linearly 45°-polarized. The x- and
y-component is then equal and Eq. 2 can be used. The latter shows that for the standard
configuration (Figure 2) and for small retardation values the reflection R is described by 9
R(δ) ~ δ2/4 for δ << 1.
This yields:
2
⎛π L ⎛ n 2 ⎞⎞
R(α ) = R(δ (α ) ) = δ / 4 = ⎜ no α 2 ⎜⎜1 − o 2 ⎟⎟ ⎟
2
(38)
⎜2 λ ⎟
⎝ ⎝ ne ⎠ ⎠
In case of a maximal admissible reflection Rmax for skew incidence one obtains:
9
It is assumed here that a polarizing beam splitter cube is used as a polarizer. For a polarizer made of a sheet of
plastic absorption would have to be used instead of reflection.
λ 2 Rmax
α max =
L π no 1 − no 2 / ne 2
This is the maximum possible angle inside the crystal. For the angle included with the z-
axis outside of the crystal we write α’. In this paraxial case Snell’s law writes in first order as
no α = α’. Herewith one obtains finally for the acceptance angle:
2
λ 2 no Rmax λ no Rmax (39)
α ' max = no α max = ≈
L π 1 − no 2 / ne 2 L π ne − no
This is for a given Rmax the maximal admissible divergence angle. For e.g. Rmax = 5%,
λ = 1064nm and L = 5mm we calculate α’max = 1.7° for LiNbO3 (no = 2.244, ne = 2.1605) and
α’max = 8.3° for LiTaO3 (no = 2.1787, ne = 2.1821). The same values hold for an angular
deviation of the optical ray in the yz-plane.
In case of left or right circular polarized light the reflection is given, as explained at the
end of chapter 1.2, in first order by
1 δ 1 π L ⎛ n2⎞
R(α ) = R(δ (α ) ) ≈ ± = ± no α 2 ⎜⎜1 − o 2 ⎟⎟ .
2 2 2 2λ ⎝ ne ⎠
For a given allowed deviation Rmax of the reflectance in the initial position the formula for
the acceptance angle is now
λ λ no Rmax .
2
2 no Rmax
α ' max = no α max = ≈

L π 1 − no 2 / ne 2 L π ne − no
with the same values as above the acceptance angle for circular polarized light is 0.38° for
LiNbO3 and 1.86° for LiTaO3.
2.10. Rotation of the Main Axis of the Index Ellipsoid
For LiTaO3 in the standard configuration Eq. 11 holds. With the parameters
p41E = −0.085, p44E = 0.028 (appendix 5.5.3) we obtain
(
B4 = 1.116 p 41
E
+ 0.175 p 44
E
)
ε 1 = −0.09 ε1 .
For half wave retardation a strain amplitude ε1λ/2 = 0.97 10−4 was found (for LiTaO3,
λ = 1064nm and an optical path length L = 5.5mm, chapter 2.5). This leads to B4λ/2 = 8.73 10−6
and with Eq. 9 we get an angle of rotation in the yz-plane of 0.76°. For an external observer
92 F. Bammer
this rotation is by a factor no larger, which yields 1.66° as the final angle of rotation. This
holds when the crystal is elongated or compressed in x-direction, such that half wave
retardation is achieved. This is sufficiently smaller than the acceptance angle 8.3° (for
Rmax = 5%), which was calculated in the previous chapter 2.9 for 45°-polarized light.
2.11. Acceptance Angle of a SCPEM, 2-Dimensional View
The previous analysis can be extended to the transmission of linearly polarized light rays,
that deviate in the crystal by the angle α in x-direction and β in y-direction from the optical
axis. Then they pass a polarizer which is oriented such that the incoming polarization is
transmitted. The artificial birefringence is characterized by the 6x1-vector B (appendix 5.3.2).
B = 0 means the undisturbed crystal. We present an overview how this case can be treated and
refer for a more detailed treatment to [7].
In Figure 10 an uniaxial crystal is illuminated along its optical axis (z-axis) with
divergent x-polarized light which then passes a polarizer (analyzer) and then hits a screen.
The initial polarization is vertical such as the orientation of the analyzer. For B = 0 the typical
Maltese-cross in case of uniaxial crystals can be seen. This is used for the adjustment of
Pockels-cells.
One single ray going in a direction (α, β) is split up in the crystal into two perpendicular
polarized partial rays, which theoretically travels on different paths. This latter effect is
neglected, since we assume that the mutual walk-off is much smaller than the lateral
coherence length, which is of course only valid in case of z-paraxial rays.
Figure 10. Setup to visualize the birefringence of an uniaxial crystal.
n1 and n2 are the refractive indices seen by the two partial rays in dependence on B, α, β.
The corresponding retardation is
δ(B, α, β) = 2π L (n1(B, α, β) − n2(B, α, β))/λ
The angle between the x-axis and the smaller semi-axis (the fast axis) of the cross section
ellipse is called φ. In [24] and in appendix 5.4 the following formula is derived for given φ
and δ 10:
10
Note that for φ = 45° this formula becomes identical to Eq. 2
2
⎛ ⎧ δ ( B, α , β ) ⎫ ⎞
T ( B, α , β ) = 1 − ⎜⎜ sin (2 ϕ ( B, α , β ) ) sin ⎨ ⎬ ⎟⎟ (40)
⎝ ⎩ 2 ⎭⎠
The determination of n1, n2, φ needs the two eigenvalues and eigenvectors of the
2x2-matrix M2x2 built from the coefficients of the equation of the cross section ellipse, which
is calculated out of the 3x3-matrix of the coefficients of the indicatrix (appendix 5.3.2):
⎛ no −2 + B1 B6 B5 ⎞
⎜ −2
⎟
M 2x2 = P12 ⋅ Bβ ⋅ Aα ⋅ ⎜ B6 no + B2 B4 ⎟ ⋅ Aα T ⋅ Bβ T ⋅ P12 T
⎜⎜ ⎟
+ B3 ⎟⎠
−2
⎝ B5 B4 ne
Aα, Bβ ∈ SO(3) 11 describe rotations by α, β around the x- and y-axis. P12 is the 2x3-
projection matrix to extract from a 3x3 matrix the 2x2 matrix of the positions 11, 12, 21, 22.
The square root of the inverted eigenvalues are the refractive indices n1, n2. The shorter one of
the two orthogonal eigenvectors includes the angle φ with the x-axis. Now Eq. 40 can be
used. In that way for every point (α, β) on the screen the transmission and herewith the
brightness is calculated. Hereby it is to consider that α, β are the angles within the crystal.
When the rays leave the crystal the divergence angles are increased by a factor no according
to the paraxial approximation of Snell’s law.
Figure 11. Interference pattern for LiNbO3, λ = 1064nm, L = 5.5 mm and B = 0 (left) and for
B = (0, 0, 0, 0, 0, λ/L/no3)T (right). The axis coordinates are the angles measured inside the crystal. The
circle corresponds to a maximal divergence angle αmax = 0.76° according to Eq. 39 with Rmax = 5%.
Figure 11 shows two results of the calculation for angles inside the crystal. On the left the
pattern for the undisturbed crystal is shown and on the right the pattern that is found when the
6th component of B is set equal to
11
SO(3).Lie group of 3-dimensional rotations
94 F. Bammer
λ
B6 = .
L no ³
This situation appears when the half wave voltage, calculated in chapter 2.6 for the
electro-optic effect of a static electrical y-field, is applied to electrodes on yz-surfaces such
that the electrical field points in the x-direction. This produces along the optical axis a 90°-
turn of the polarization (initially oriented along the x- or y-axis) and hence in the
configuration of Figure 10 an analyzer transmission equal to 0 along the optical axis.
The pictures obviously illustrate the limitation of the divergence from the optical axis and
that a good collimation and adjustment of the beam is necessary.
2.12. Useful Aperture of a SCPEM
In chapter 2.4 it was found that for the configuration in Figure 2 the x-strain and herewith
the retardation δ depends on x and is proportional to cos(π x/l) at the first resonance (Eq. 36).
If the voltage amplitude is set such that at x = 0 half wave retardation is achieved then
δ(x) = π cos(π x/l)
This, put into Eq. 2, yields for z-parallel 45° linearly polarized light a transmission
T(x) = T(δ(x)) = (1 + cos δ)/2 = (1 + cos(π cos(π x/l)))/2
Obviously T(0) = 0. If now a certain transmission Tmax is allowed for light rays passing
the crystal at x ≠ 0, then
Tmax = (1 + cos(π cos(π xmax/l)))/2
or
xmax/l = π −1arcos(π −1arcos(2Tmax − 1)) .
This yields xmax/l = 0.11 for Tmax = 1% and xmax/l = 0.17 for Tmax = 5%.
Our model reveals no y-dependence of the retardation. However numerical and
experimental work show that for conventional PEMs in a similar configuration a slight
y-dependence is present, as well [30].
2.13. Description of Spatial Varying Artificial Birefringence
An additional increase in complexity is found for SCPEMs with more complicated

eigenmodes since the deformation and the corresponding artificial birefringence B varies then
along the optical path, too.
In this case a finite element simulation is necessary, which calculates the NF -dimensional
displacement vector uV(t) (NF...degree of freedom of the FE-model).
Here for one certain kind of excitation (harmonic voltage course at the electrodes with
amplitude 1 Volt) with a fixed excitation angular frequency ωA (chosen such that a resonance
frequency is hit) the corresponding displacement vector is calculated with a harmonic
analysis. The ansatz
uV(t) = Re[uV0 exp( j ωA t)]
is put into the equations of motion. The term exp( j ωA t) can be eliminated such that a linear
system of equations with the complex solution vector uV0 results. Due to the linearity of the
problem this vector is direct proportional to the amplitude of the excitation.
Figure 12 shows an example of an eigenmode for a crystal as used in [3, 7].
For a given path and a given time t the FE-program can calculate the strains along this
path with internal spline functions. These, and the resulting artificial birefringence vector
(Eq. A-5, A-6) B(p) then depend on the path coordinate p 12.
The polarization of the light changes now continuously along the path. Within an
infinitesimal length dp the vector B(p) is assumed to be constant.
The polarization is now described with the Stokes-vector SV = (S0, S1, S2, S3)T [7, 25], with
S0…overall intensity, S1…the intensity of horizontal and vertical polarized light, S2…intensity
of + or −45° linear polarized light and S3…intensity of right or left circular polarized light.
For polarized light S02 = S12 + S22 + S32 holds.
Figure 12. ANSYS-simulation of a shear eigenmode of a LiNbO3 crystal with x-excitation and
dimensions 2x5x30mm in x-, y-, z-direction [7].
12
Here the time needed for the passage through the crystal is neglected.
96 F. Bammer
The impact of an optical element on the polarization is described with the Müller-matrix.
Its multiplication with the input-Stokes-vector yields the output-Stokes-vector. The Müller-
matrix Mret for an ideal retardation-plate of thickness L is [25]
⎛1 0 0 0 ⎞
⎜ ⎟
⎜ 0 cϕ + sϕ cos δ sϕ cϕ (1 − cos δ ) − sϕ sin δ ⎟
2 2
M ret =⎜
0 sϕ cϕ (1 − cos δ ) sϕ + cϕ cos δ cϕ sin δ ⎟
2 2
⎜ ⎟
⎜0 sϕ sin δ − cϕ sin δ cos δ ⎟⎠
⎝
13
with cφ = cos2φ, sφ = sin2φ, φ…angle between the first possible polarization and the x-
axis, δ = 2π L (n1 – n2)/λ …retardation with n1,2 … refractive indices for the two possible
polarizations. All these variables depend on the path length p.
If the length L in the definition of δ is replaced by the differential dp, a dδ is defined and
an infinitesimal Müller-matrix dMret can be written down, in which only the first order of dδ
is considered:
⎛1 0 0 0 ⎞
⎜ ⎟
⎜0 1 0 − s ϕ dδ ⎟
dM ret =⎜ + O ( dδ 2 )
0 0 1 cϕ dδ ⎟
⎜ ⎟
⎜ 0 s dδ − cϕ d δ 1 ⎟⎠
⎝ ϕ
The change of the Stokes-vector after passing the path element dp is:
dSV = SV ( p + dp ) − SV ( p ) = dM ret SV ( p) − SV ( p )
⎛1 0 0 0 ⎞⎛ S 0 ⎞ ⎛ S 0 ⎞ ⎛ S0 ⎞ ⎛ S0 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜0 1 0 − sϕ dδ ⎟⎜ S1 ⎟ ⎜ S1 ⎟ ⎜ S1 − sϕ dδ S 3 ⎟ ⎜ S1 ⎟
=⎜
0 0 1 ⎟⎜
cϕ dδ S 2 ⎟ −⎜ ⎟ = ⎜
S2 S 2 − sϕ dδ S 3 ⎟ −⎜S ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ 2⎟
⎜ 0 s dδ
⎝ ϕ − cϕ dδ 1 ⎟⎠⎜⎝ S 3 ⎟⎠ ⎜⎝ S 3 ⎟⎠ ⎜⎝ − sϕ dδ S1 + cϕ dδ S 2 + S 3 ⎟⎠ ⎜⎝ S 3 ⎟⎠
⎛ 0 ⎞ ⎛0 0 0 0 ⎞
⎜ ⎟ ⎜ ⎟
⎜ − sϕ S 3 ⎟ 2π 0 0 0 − sϕ ⎟
=⎜ dδ = (n1 − n2 )⎜⎜ S V ( p ) dp
− sϕ S 3 ⎟ λ 0 0 0 cϕ ⎟
⎜ ⎟ ⎜ ⎟
⎜− s S + c S ⎟
⎝ ϕ 1 ϕ 2⎠
⎜0 s
⎝ ϕ − cϕ 0 ⎟⎠
This yields a system of coupled ordinary differential equations that describe the evolution
of the Stokes-vector:
13
Obviously the intensity S0 remains constant.
⎛0 0 0 0 ⎞
⎜ ⎟
dSV 2π ⎜0 0 0 − sϕ ⎟
= (n1 − n2 )⎜ SV
dp λ 0 0 0 cϕ ⎟
⎜ ⎟
⎜0 s − cϕ 0 ⎟⎠
⎝ ϕ
These are three coupled ordinary differential equations (the first one with dS0/dp = 0 is
trivial), which can be integrated with any numerical equation solver, if the initial conditions
are given and if n1, n2, φ are calculated as functions of p out of B(p) according to the
procedure described in chapter 2.11.
Now the tools are developed to calculate the output polarization of any input light ray or
light pulse into a SCPEM. Based on this [7] provides a detailed analysis of oscillating crystals
of the kind used in publication [3] with the finite-element-program ANSYS and with
MATHEMATICA.
3. EXPERIMENTAL ANALYSIS OF A SCPEM

The main purpose of an experimental analysis of a SCPEM is to determine the resonance
frequencies and the corresponding current and retardation amplitudes related to the voltage
amplitude. The basic experimental setup is sketched in Figure 2. Note that here it is
advantageous to use circular polarized light since in this case the relation between
transmission and retardation is linear (Figure 13b, Eq. 3) in a limited range. Hence the
retardation amplitude can be directly seen from the transmission amplitude. In Figure 13d and
Figure 14 experimental data of the first resonances of the crystal in the left picture of Figure 3
are shown.
Figure 13. Relations between (a) retardation course, (b) transmission over retardation, (c) transmission
over time. (d)…realization of the predicted transmission course with U(t)…applied voltage with
amplitude 3V, I(t)…resulting current and T(t)…transmission course.
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. 98 F. Bammer
Figure 14. Amplitude and phase of current I (upper two pictures) and transmission T (lower two
pictures) related to amplitude and phase of the applied voltage U in the frequency range 180-430 kHz.
In Figure 14 we see for the frequency region 180-430 kHz the amplitudes and phases of
current and transmission related to amplitude and phase of the applied harmonic voltage
course. In the current amplitude spectrum three main resonance’s at 199, 348 and 377 kHz are
found.
The first one is the longitudinal x-oscillation as described theoretically and quantitatively
in the chapters 2.4 and 2.5.
The second one is the longitudinal y-oscillation as described theoretically in the
chapter 2.3.
The third one with the largest current peak is a yz-shear oscillation and is theoretically
equivalent to the case discussed in chapter 2.3.
The phase courses of the current behave according to the prediction of Figure 7 showing
at every resonance a transition from advancing to lagging and back. The current signal can be
used in a feedback-loop to the input of the amplifier generating the voltage course. When this
is done with our setup the system oscillates at 377 kHz, since the current answer is the
strongest at this frequency. With a filter every resonance frequency can be selected.
The amplitude course of the transmission T shows further characteristic peaks besides the
main resonance’s, which cannot be explained up to now. For this numerical simulations will
be performed. One important finding is the extremely low voltage amplitude to achieve a
significant modulation of transmission. E.g. in the main resonance at 199 kHz the peak value
of the transmission is found to be 0.2 V-1. This corresponds to a retardance amplitude of
0.4 V-1, hence with a voltage amplitude of only ~2.5 V a retardance amplitude of ~ π/2 can be
achieved. The phase course of the transmission T corresponds nicely to the phase course of
deformation in Figure 7, which shows in the resonance a transition from 0° to –180°.
Correspondingly in every resonance a transition from 180° to 0° or from 0° to −180° is
observed for the phase of transmission, depending on the sign of the effect.
Of course the spectrum shows many significant peaks at higher frequencies, i.e. at higher
harmonics. One significant peak is at 1 MHz, a further one at 2.94 MHz.
4. APPLICATIONS OF SCPEMS
4.1. Time-Multiplexing of Pulsed Laser Diodes [3, 8, 9, 11]
The idea of time-multiplexing in this context is to guide high power laser pulse sequences
from several sources onto one common optical path. The resulting laser beam has the same
beam quality than each single pulse sequence while the beam power increases with the
number of combined sources. This is a way to create high power laser sources with high
brilliance, which is particularly interesting for diode laser systems [12] because of their very
low beam quality.
Figure 15 schematically shows the setup which combines two branches that can in
principle be driven independently from each other (which is not desirable of course).
In each branch two laser diodes are alternately emitting laser pulses, which are
polarization multiplexed and afterwards transformed to circular polarized light with a quarter
wave plate. Each LiNbO3-crystal mechanically oscillates on an eigenfrequency in a shear
100 F. Bammer
mode and influences the polarization due to the photo-elastic effect. In one of its two extreme
positions during an oscillation period the effect of the quarter wave plate is doubled, leading
to a 90° turn of polarization. In the other extreme position the effect of the quarter wave plate
is annihilated, thus the polarization is changed back to its initial one before the quarter wave
plate.
When the emission of the laser pulses takes place at the extreme positions of the
oscillation a kind of polarization rectification at the output of the crystal is obtained. This
allows further polarization multiplexing (with the polarization filter PF3) of the output of the
two branches. When the crystals, which are assumed to oscillate in phase, cancel the effect of
the quarter wave plate the laser diodes LD1 and LD3 emit a pulse and when the effect is
doubled the diodes LD2 and LD4 are addressed.
Figure 15. Scheme for time-multiplexing of four laser diodes: LD1-4…laser diodes,
PF1-3…polarization filters 1-3, Q1-2…quarter-wave-plates, C1-2…LiNbO3-crystals with electrodes on
the yz-surfaces, H1-2…half-wave-plates, xyz1-2…local crystal coordinate systems.
To make things clearer consider the first branch for the laser diodes LD1 and LD2. The
two laser diodes alternately emit laser pulses. The light of laser diode LD1 is horizontally
polarized and goes straight through the polarization filter PF1. Then it passes the quarter
wave plate Q1 to become circularly polarized (it does not matter whether left or right). The
timing must be now such that the crystal C1 is passed when it is in the extreme position,
where it acts as a further quarter wave plate (the deformation amplitude and hence the
amplitude of the exciting voltage must be well adjusted to achieve this) but with opposite sign
as the first fixed quarter-wave-plate Q1 such that the net retardation is zero and the light
leaves the crystal C1 with the initial horizontal polarization. The light of laser diode LD2 is
horizontally polarized, too, but is turned to vertical polarization by the half-wave-plate H1 to
enable reflection in the desired direction in the polarization filter PF1. The laser diode LD2
pulses now when the crystal C1 is in the other extreme position such that the net retardation
of the crystal and the quarter-wave-plate Q1 is that of a half-wave-plate and the vertical
polarization is turned back to horizontal polarization. Hence the pulses of both laser diodes
leave the branch with horizontal polarization. In the other branch the same happens in an
analog way but synchronized such that the output polarization is vertical, so that the output of
both branches can be polarization multiplexed via the polarization filter PF3.
4.1.1. Beam Switching by Harmonic Modulation

For optical time-multiplexing of periodically produced laser pulses in the manner
described above a polarization switch is needed, which periodically switches between two
orthogonal linear polarizations. Now in Figure 13c the graph, which is drawn through, shows
this characteristics. There a configuration as in Figure 2 with circular polarized input is
assumed and with a proper retardation course (Figure 13a) switching between +45° and −45°
happens. In this experiment the retardation amplitude is a bit too high, such that counter peaks
evolve where the transmission should be 0 or 1. With a slightly smaller amplitude the desired
characteristics is well approximated, but with the problem that the orientations of the crystal,
of the polarizer and of the polarizations are different to that needed in Figure 15.
Now in Figure 15 the crystals are excited with an electrical x-field (instead of a y-field in
Figure 2) and are cut and oriented such that an x-propagating shear mode (polarized in the yz-
plane) 14 is excited, which affects the refractive indices of +45° and −45° polarized light
(instead of 0° and 90° polarized light in Figure 2). Hence polarization switching happens
between 0° and 90°-polarized light and with a horizontally oriented polarizer the achieved
transmission course is again that of Figure 13. The advantage of this new arrangement is that
the cross-section of the laser diode beam fits better to the aperture of the crystal.
With a crystal thickness of 1.8mm the corresponding switching frequency is 1 MHz. The
opening times (T = 0 or 1) last then ~300ns and the proposed switching scheme can be
performed.
4.1.2. The Prototype

Figure 16 shows the compact realization of time-multiplexing of four laser diodes from
JDS Uniphase specified for 4 W @ 960nm mounted on C-mounts. Their M² value after fast
axis collimation (which typically increases the fast axis beam parameter product by a factor of
two) is 30. At 1 MHz each diode emits laser pulses of 300ns duration and with peak power
values up to 15 W. The average power is 3.6 W corresponding to 90 % of the specified cw-
power to maintain life time. A long time test of 160 hours of one laser diode with pulsed
operation at 4 W average power showed no failure or degradation. We assume that due to the
high pulse repetition rate the degradation is like in cw-operation.
Cooling is done by three thermo-electric coolers (Melcor, PE-Series, nominal cooling
power: 51.4 W) below the setup. These are fixed on a heat sink with cooling fins supported by
a little blower.
The dimensions of the crystals are 1.8 x 5 x 30 mm (x-y-z-values). They oscillate in
phase with 1 MHz in a shear mode. It is notable that a second usable resonance exists at 1.14
MHz corresponding to a second shear oscillation mode with a motion direction perpendicular
to the first one. The amplitude of the excitation voltage is 10 V, twice as high as in the
numerical simulation. This is due to additional damping caused by the friction of the crystal
in its mount (the crystals are gently clamped between massive aluminum electrodes).
The experiment generates 10.5 W laser power at the output. The total optical power
emitted by all laser diodes is 4 x 3.6 = 14.4 W, hence an overall optical efficiency of
10.5/14.4 = 73 % is achieved. This rather low value is due to several trade-offs regarding
adjustment, optical elements and most important the slightly different resonance behavior of
14
This is not the shear mode mentioned at the end of chapter 3.
102 F. Bammer
the two crystals. The switching efficiency of the crystals can reach values of up to 95 % but a
small detuning necessary for full operation of our experiment causes a decrease to 85 %.
Figure 16. The prototype (dimensions 115 x 115 mm) for time–multiplexing of four laser diodes.
The beam quality of this diode laser is theoretically the same as that of one single laser
diode. Two issues may affect this negatively in praxis. First a pulsed laser diode may have a
larger divergence and hence a lower beam quality than a continuous driven laser diode. We
examined this at various power levels and found a decreasing beam quality at higher power
levels but no difference in the beam quality of both operation modes. Second the accuracy of
adjustment of the laser diodes limits the overall beam quality since the axes of their beams
must be perfectly aligned to the optical output path. Final alignment was done via tilting of
the polarization filters. We achieved eventually collinear laser beams (at least within the
accuracy of far field spots observed via a NIR-sensitive camera), i.e. an overall beam quality
indeed being the same as that of one laser diode.
4.2. Q-Switching [5, 6, 14, 29, 37]
Pulsed laser operation is of utmost importance for many applications. One important
method to produce a short and strong laser pulse is Q-switching. This method uses an optical
switch in the laser quality, which blocks the optical light path while energy is pumped to the
gain medium. When it suddenly opens all stored energy is rapidly converted to a strong light
pulse. Usual optical switches use the acousto- or electro-optic effect.
A SCPEM in standard configuration (Figure 2) can do this job with very low cost. During
one oscillation period the SCPEM passes two times through its undeformed shape, where the
transmission is 100%. Hence two laser pulses are emitted every period.
If we assume now that the standard configuration is placed inside a laser cavity with a
round-trip-frequency fc (= pulse repetition frequency PRF) and round-trip-transmission Tc the
resulting time dependent photon life time is given by [4]
t p (t ) = − 1 ln(Tc T (δ (t ) ))
fc
(41)
This variable plays a significant role in the laser rate equations, which will be treated
later. In Figure 17 an evaluation of Eq. 2 and Eq. 41 is shown for a harmonic retardation
course for the parameters fc = 150 MHz, Tc = 0.96 and a retardation amplitude of δ0 = 1.1 π. It
can be seen that the transmission and the photon life time shows significant and rather short
peaks when the PEM crosses its zero-position. During the rather long times of maximal
extension or compression of the PEM the light path is blocked since the PEM acts then as a
half-wave-plate. This behavior can obviously be used for Q-switching with constant
frequency.
This can be investigated in a first step by a numerical solution of the basic laser rate
equations for photon number P and inversion number n shown in Eq. 42 [4, 14]
P& (t ) = c0 σ P(t ) n(t ) − P(t ) / t p (t ) + M n(t ) / t n

, (42)
n& (t ) = p − n(t ) / t n − c0 σ P(t ) n(t )
with c0…velocity of light, σ [m²]…cross-section of induced emission, M…fraction of light

emitted by spontaneous emission which travels in a direction that contributes to the laser
mode, tn….inversion life time, p…pumping rate (number of excited states per m³ and s).
Figure 17. Typical courses of (a) retardation with δ0 = 1.05 π, (b) retardation dependent transmission of
a PEM/polarizer-configuration, (c) resulting time dependent transmission of a PEM/polarizer-
configuration and (d) resulting time dependent photon life time of a laser cavity with a PEM/polarizer-
configuration inside.
104 F. Bammer
Figure 18 shows a solution for the parameters σ = 10-20m², tn = 100µs, M = 8 10-6,

p = 1.23 1023 m-3s-1 and the modulator frequency f0 = 10 kHz. A clean pulse sequence is the
theoretical result of the chosen parameters. That the pulse energy is constant can be seen by
the height of the jumps in the course of inversion number since the course of the photon
number is not precisely plotted due to numerical problems.
With other choices of parameters typical problems like pre- and post-lasing and spiking
are encountered. It is therefore of importance to further reduce the time of high transmission.
This can be achieved by adding a third harmonics to the retardation course as depicted in
Figure 19. It must be noted that the 3rd longitudinal harmonics in the initial oscillation
direction has not exactly a frequency three times the basic frequency f0 due to the lateral
extension of the bar.
Figure 18. Typical solution of the laser rate equations (Eq. 42) for a course of photon life time like in
Figure 17d.
Figure 19. Adding a third harmonics in the retardation course: Typical courses of (a) retardation with
δ0 = 1.24π, (b) retardation dependent transmission of a PEM/polarizer-configuration, (c) resulting time
dependent transmission of a PEM/polarizer-configuration and (d) resulting time dependent photon life
time of a laser cavity with a PEM inside.
On the other side a perpendicular thickness oscillation, i.e. in y-direction (Figure 2) can
be excited with the same electrodes and can be adjusted via the height of the crystal to the
desired frequency.
The same is possible with a yz-shear oscillation mode, which frequency is mainly
determined by the smaller dimension in y- or z-direction. These eigenmodes can be used to
generate the desired 3rd harmonics [6]. We use for such a Dual Mode Single Crystal
Modulator the term DMSCPEM.
Figure 20 shows one possible scheme for a simple PEM-Q-switched laser. In this case a
double pass configuration reduces the necessary retardation amplitude by a factor of two. A
polarizing cube beam splitter is chosen as a polarizer and acts here as the out-coupler, i.e. a
kind of cavity dumping is realized.
To summarize, Q-switching with photo-elastic modulators seems to be a simple and
cheap method to realize a robust pulsed laser with constant pulse repetition frequency, as
needed e.g. in case of laser range finders or in some rapid prototyping methods.
In [5, 6] and in [37] a realization of fiber-laser Q-switching with mono-mode and with
dual-mode SCPEMS is described. Peak powers of 1100W (FWHM pulse duration: 70ns,
PRF: 127 kHz) were achieved for a laser designed for 10W average power [37].
Particularly interesting is the use of 43m-crystals (e.g. GaAs, ZnSe) for Q-switching of
CO2-lasers, since for these lasers no satisfying Q-switches are available [36]. This crystal
class is optical isotropic and therefore offers the same acceptance angle as conventional
PEMs. Alas, since their piezo-electric coefficients are rather small and since the necessary
retardation amplitude δ0 scales with the wavelength, high driving amplitudes in the order of
100 V would be necessary.
Figure 20. Setup of a laser with Q-switch and cavity dumping.
One disadvantage of this method is the fixed pulse repetition frequency such that only
few applications like LIDAR or laser marking (e.g. glass scribing) can profit from this
method.
106 F. Bammer
Figure 21. Transmission scheme for pulse picking. The vertical lines display the incoming pulse
sequence. On the left (δ0 = π and fSCPEM = PRF/4) every second and on the right (δ0 = 1.15π and
fSCPEM = PRF/6) every third pulse can pass the pulse picker.
4.3. Pulse-Picking
This application addresses passively mode locked laser systems. Such a multiple folded
laser with, for example, a cavity length of LC ~7m has a pulse repetition frequency of
PRF = c0/2/LC ~ 21 MHz (c0…vacuum light velocity) [31]. Usually the pulse energy is
increased with a further optical amplifier. In many applications a much lower pulse repetition
frequency is desired. This can in principle be achieved with a larger cavity length, which is
cumbersome, first due to an increasingly difficult adjustment, and second because the pulse
duration would be increased, too. Usually costly pulse pickers based on Pockels-cells are used
to pick single pulses out of the pulse train. Now, since the pulses are usually linearly
polarized, the standard configuration (Figure 2) together with an analyzer can be used for the
reduction of the pulse repetition frequency. Figure 21 shows for a fixed pulse sequence how
every second or every third pulse can be picked with transmission courses similar to that used
for Q-switching. They are described by Eq. 2 with a retardation course (δ0 …retardation
amplitude, fSCPEM …oscillation frequency of the SCPEM)
δ(t) = δ0 sin(2π fSCPEM t) . (43)
4.4. Optical Switch for Continuous Wave Operation
Here a very simple and robust realization of an optical switch is presented. It uses again
the standard configuration (Figure 2). A continuous incoming light beam is assumed. If now
the retardation course given by Eq. 43 is put into Eq. 2 for the transmission T(δ), then the
following series can be developed [2] (here we use ω = 2π fSCPEM) :
1 + J 0 (δ 0 )
T (t ) = cos(δ 0 sin(ω t ) / 2) = + J 2 (δ 0 ) cos(2ω t ) + J 4 (δ 0 ) cos(4ω t ) + ...
2
J0, J2, J4 are the Bessel-functions of first kind. In the average only the first time-
independent term remains such that we can write now for the averaged transmission T and
reflection R:
1 + J 0 (δ 0 ) 1 − J 0 (δ 0 )
T (δ 0 ) = , R(δ 0 ) = .
2 2
Figure 22 shows the temporal course of the transmission (left) and its temporal average
(right). For a retardation amplitude δ0 = 1.2π the transmission takes a minimum equal to 30%.
Between this and 100% the transmission can be freely chosen simply via the driving voltage
amplitude. With the use of a DMSCPEM (chapter 4.2) the setting range is 18-100%.
T T
1 1
0.75 0.75
0.5 0.5
0.25 0.25
wt A
p 2p 3p p 2p 3p
Figure 22. Transmission through the SCPEM-analyzer configuration of Figure 2 (input 45° linear
polarized); left: temporal course with amplitudes δ0 = π/2, π, 1.2π; right: average transmission.
4.5. Ellipsometry
This application is fundamentally different from the others, since it belongs to optical
metrology, i.e., in most cases it is used to determine the parameters of an optical coating
[2, 16]. For that the change of polarization of reflected light is measured and quantified with
the two ellipsometric angles Ψ and Δ (0 ≤ Ψ ≤ 90°, 0 ≤ Δ < 360°) [16]. For the fast and exact
determination of these two variables a method using PEM-modulated polarized light is used
[1], which is shown in Figure 23.
Light, modulated in its polarization with a PEM, is reflected by the sample and passes
through an analyzer.
A detector measures the resulting varying intensity. A Lock-In-Amplifier identifies the
mean value and the first and second harmonics of this signal. With these values the
ellipsometric angles can be immediately calculated. The advantage of this method against
others like e.g. zero-ellipsometry is that it is extremely fast.
For visible light (VIS) and the near infrared (NIR) this conventional PEM can now be
easily replaced by a SCPEM made of LiTaO3. In this way the system becomes smaller and
cheaper. One disadvantage is the limited acceptance angle due to the natural birefringence of
3m-crystals (chapter 2.9).
108 F. Bammer
Figure 23. Principle of a PEM-based ellipsometer.
5. APPENDIX
5.1. Elasticity and Piezo-Electricity
The elastic deformation of an infinitesimal small volume in a linear elastic body is

described by the tension vector T and the strain vector S. Both have six components, the first
three are the main diagonal of the tension tensor T ' or the strain tensor S ', the last three are
the off-diagonal entries. This takes the following form for a 3-dimensional displacement field
uk(x1, x2, x3) (k = 1,2,3):
⎛ σ 11 σ 12 σ 13 ⎞ ⎛ ε 11 ε 12 ε 13 ⎞
⎜ ⎟ ⎜ ⎟ ∂u ∂u j
T ' = ⎜ σ 12 σ 22 σ 23 ⎟ , S ' = ⎜ ε 12 ε 22 ε 23 ⎟ mit ε ij = i +
⎜σ x xi
σ 33 ⎟⎠ ⎜ε ⎟ j
⎝ 13 σ 23 ⎝ 13 ε 23 ε 33 ⎠
⎛ σ 11 ⎞ ⎛ σ 1 ⎞ ⎛ ε 11 ⎞ ⎛ ε 1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ σ 22 ⎟ ⎜ σ 2 ⎟ ⎜ ε 22 ⎟ ⎜ ε 2 ⎟
⎜σ ⎟ ⎜σ ⎟ ⎜ ε ⎟ ⎜ε ⎟
T =⎜ ⎟=⎜ 3⎟ , S =⎜ ⎟=⎜ 3⎟
33 33
⎜ σ 23 ⎟ ⎜ σ 4 ⎟ ⎜ 2ε 23 ⎟ ⎜ ε 4 ⎟
⎜ σ ⎟ ⎜σ ⎟ ⎜ 2ε ⎟ ⎜ ε ⎟
⎜ 31 ⎟ ⎜ 5 ⎟ ⎜ 31 ⎟ ⎜ 5 ⎟
⎜ σ ⎟ ⎜σ ⎟ ⎜ 2ε ⎟ ⎜ ε ⎟
⎝ 12 ⎠ ⎝ 6 ⎠ ⎝ 12 ⎠ ⎝ 6 ⎠
Here we used the convention for reduced indices with the assignments
11 Æ 1, 22 Æ 2, 33 Æ 3, 23 Æ 4, 31 Æ 5, 12 Æ 6,
which allows a representation of the tensions and strains in vector form. This has the
disadvantage that for the vector representation there is no simple rule for a transformation into
a new coordinate system. Within the tensor representation the transformation rule is simple,
e.g. for the tension: σij” = aik σkl alj . Here the aik are the coefficients of the rotation matrix of
the coordinate transform xi” = aik xk [17].
The strain S and tension T for isotropic materials or crystals without piezo-electricity are
coupled by T = c S or S = s T . Here c is the 6 x 6-elasticity-matrix and s = c –1 is the
corresponding compliance matrix.
For isotropic materials these matrices are determined by only two constants, e.g. the
Young’s modulus EM and the Poisson-ratio ν or the shear modulus G [33]. In that case they
take the following form:
⎛1 − ν ν ν 0 0 0 ⎞
⎜ ⎟
⎜ ν 1 −ν ν 0 0 0 ⎟
EM ⎜ ν ν 1 −ν 0 0 0 ⎟
c= ⎜ ⎟
(1 + ν )(1 − 2ν ) ⎜ 0 0 0 1 − 2ν 0 0 ⎟
⎜ 0
⎜ 0 0 0 1 − 2ν 0 ⎟⎟
⎜ 0
⎝ 0 0 0 0 1 − 2ν ⎟⎠
⎛ 1 −ν −ν 0 0 0 ⎞
⎜ ⎟
⎜ − ν 1 − ν 0 0 0 ⎟
1 ⎜⎜ − ν − ν 1 0 0 0 ⎟
s= ⎟
EM ⎜ 0 0 0 1 +ν 0 0 ⎟
⎜ 0
⎜ 0 0 0 1 +ν 0 ⎟⎟
⎜ 0
⎝ 0 0 0 0 1 + ν ⎟⎠
For the widely used optical glass BK7 the values EM = 81 GPa and ν = 0.208 are found.
For crystals the number of independent components increases with decreasing degree of
symmetry and the structure of the matrices becomes more complex (Figure 24, [17, 34],
appendix 5.5.2/3).
Figure 24. Structure of the constitutive matrices for the crystal class 3m.
110 F. Bammer
Certain crystal classes show piezo-electric and at the same time electro-optical behavior.
Additionally to S or T the electrical field vector E or the electric displacement vector D is
necessary to determine the actual state of the material [17] 15. Hence four pairs of constitutive
vectors are available: (S, E), (S, D), (T, E) or (T, D). They are connected by the following
equations:
⎛ T ⎞ ⎛⎜ c E − e ⎞⎛ S ⎞ ⎛ T ⎞ ⎛⎜ c D − h ⎞⎟⎛ S ⎞
T T
⎜⎜ ⎟⎟ = ⎟⎜ ⎟ ⎜⎜ ⎟⎟ = −1 ⎜
χ S ⎟⎠⎜⎝ E ⎟⎠ ⎜ ⎟⎟
⎝ D ⎠ ⎜⎝ e ⎝ E ⎠ ⎜⎝ − h χ S ⎟⎠⎝ D ⎠ (A-1)
⎛ S ⎞ ⎛⎜ s E d ⎞⎛ T ⎞ ⎛ S ⎞ ⎛⎜ s D g ⎞⎛ T ⎞
T T
⎜⎜ ⎟⎟ = ⎟⎜ ⎟ ⎜⎜ ⎟⎟ = ⎟ ⎟
−1 ⎜
⎝ D ⎠ ⎜⎝ d χ T ⎟⎠⎜⎝ E ⎟⎠ E
⎝ ⎠ ⎝ ⎜ − g χ T
⎟⎜⎝ D ⎟⎠
⎠
The 3x6 matrices e, h, d, g describe the piezo-electric effect, the 3x3 matrices χS, χT the
relation between the electrical field and the electric displacement. Here the lower T and S
mean the cases of constant mechanical tension T and strain S.
In the same way the index of the elasticity matrices cE, cD and the compliance matrices
sE, sD means constant electrical field E (electrodes shorted) or constant electric displacement
D (electrodes open).
In the tensor representation the piezo-electric effect is described by 3rd-rank tensors
e ', h ', d ', g ' and elasticity by 4th-rank tensors c 'E, c 'D, s 'E, s 'D.
Obviously all four descriptions in Eq. A-1 are equivalent and all material-matrices cE, cD,
sE, sD, χT, χS, e, h, d, g can be determined, if the material matrices of one of the four systems
of equations is known. If, e.g., the matrices cE, e, χS are known, the remaining matrices will be
calculated as follows [22]:
−1
sE = cE , d = e sE , χT = χ S + ed ,
T
(A-2)
−1 −1 −1
h = χ S d c E , g = χ T d , c D = c E + e h, s D = c D
T
For the crystal class 3m the structure of the material matrices is shown in Figure 24. In
appendix 5.5.2/3 values for the two most important exponents LiNbO3 and LiTaO3 are given.
5.2. Acoustical Waves [18]
Now the following question will be answered for a given direction (m1, m2, m3)T with
m1 + m22 + m32 = 1:
2
Which acoustical waves are possible for this direction in an infinite crystal ?
For this purpose an ansatz for a plane wave that travels into the given direction is put into
the equation of motion. For a clear representation of the possible solutions we define
Γil = c ijkl m j m k , γ i = ekij m j mk , χ = χ jk S m j m k .

E
15
Temperature and entropy should be taken into account, too. We ignore that and assume isentropic conditions.
Here the cijklE represent the entries of the stiffness tensor c’E, out of which the stiffness-
matrix cE for constant electrical field (appendix 5.1) is constructed with the reduced
coordinates (chapter 2.1). The ekij are the entries into the piezo-electric tensor e’ and the χSjk
are the entries in the dielectric tensor χS at constant strain. The possible polarizations of the
wave are now the eigenvectors of the tensor
_
γi γl
Γ il = Γ jl +
χ
The corresponding eigenvalues are the wave velocities. The three possible polarizations
are perpendicular to each other and are sorted such that the first one is the quasi-longitudinal
wave with the highest velocity. Its polarization vector is approximately parallel to the
traveling direction. Both other polarizations belong to two shear waves, where the motion of
the material is approximately perpendicular to the traveling direction.
If now for the three polarizations in every spatial direction the inverse of the velocity is
plotted, three closed surfaces will be defined („slowness surfaces“). Cross sections of them
are shown for LiNbO3 (very similar to LiTaO3) in Figure 25. The additional broken lines are
the slowness surfaces when the piezo-electric effect is neglected (i.e. ekij = 0).
A direction (m1, m2, m3)T in Figure 25 defines crossing points with the slowness surfaces.
For every kind of wave the perpendicular on the tangential plane is the corresponding
direction of energy flow that travels with the group velocity 16. If this direction coincides with
the initially chosen direction then a main wave direction is found. The perpendicular on the
wave front and the direction of energy flow are identical here 17. Such a direction should be
chosen for an acoustic resonator, because in case of a reflection no one of the two other wave
types is generated.
In Figure 25 obviously the x-direction is a main traveling direction for all polarizations
since in the xy- and in the zx-plane the x-axis is perpendicular to all surfaces. Now we want to
know whether these waves can be excited by the own piezo-electric effect. The following rule
holds: For a given direction and polarization the difference between the two slowness
surfaces with and without the piezo-electric effect is a measure for the possibility to excite
this acoustic wave with an electrical field parallel to the traveling direction.
Figure 25 shows, that shear-waves in x-direction can be well excited by an electric field
in x-direction, as used for time-multiplexing (chapter 4.1, [3]). This is not possible for the
longitudinal wave. Further an electrical y-field can excite a longitudinal y-oscillation as
described in chapter 2.3. If one takes a look on the piezo-electric matrices in Figure 24 (and
appendix 5.5.2/3), then it is seen that an electrical field in y-direction can excite a longitudinal
wave in x-direction, as well 18. This is the ideal kind of oscillation for an acoustic resonator of
the kind shown in Figure 2. Another exceptional direction for an optimized piezo-effect is
seen in the yz-plane in Figure 25 This direction includes an angle of 36° with the y-axis and is
16
The direction of energy flow is the real traveling direction of an acoustic beam as produced by a transducer.
17
We remark that in the case of optical waves similar relations hold, but only two polarizations exist. The analog to
the slowness surfaces are the refractive index surfaces, that can be derived form the indicatrix (appendix
5.3.1).
18
This important fact is not mentioned by the high quality source [18]
112 F. Bammer
used for longitudinal wave transducers. For shear-wave-transducers a direction including with
the y-axis an angle of 163° is used.
Y Z
1 1
€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€
10000 10000
X Y
1 1
€€€€€€€€€€€€€€€ €€€€€€€€€€€€€€€€
10000 10000
1
€€€€€€€€€€€€€€€
10000
X
1
€€€€€€€€€€€€€€€
10000
Figure 25. Cross sections of the slowness surfaces for LiNbO3 with (continuous line) and without
(broken line) the piezo-electric effect. Inner graphs: quasi-longitudinal waves.
5.3. Electro- and Elasto-Optics
5.3.1. General Description of Natural Birefringence

Crystals are generally birefringent, some, as SiO2 out of the crystal group 32, are
additionally optical active (circular birefringence), which is not considered here. Ordinary
birefringence is described by the so called indicatrix (or index-ellipsoid) [17].
2 2 2
x1 x2 x3 (A-3)
2
+ 2
+ 2
=1
n1 n2 n3
n1, 2, 3...main refractive indices. Here a selected coordinate system is assumed, where this
advantageous representation is possible and the ellipsoid is oriented to the main axes. Figure
26 shows this with a direction of light input OP and corresponding polarizations OA and OB.
Figure 26. The indicatrix and the construction of the polarizations OA, OB for a given wave direction
PO [17].
The meaning of the indicatrix is: If a wave vector (k1, k2, k3)T for the electromagnetic
wave is now chosen (direction PO), then only two certain linear polarizations are possible.
These are found in the cross-section of the indicatrix with the plane that is perpendicular to
the wave vector and goes through the origin O. This results in an ellipse, called “cross section
ellipse” here. Its main axis OA and OB are the two possible polarizations. The light wave in
this approach is described in terms of the electric displacement:
Di (t , x1 , x 2 , x3 ) = Di exp j (ν t − k l xl ) with ( k1 , k 2 , k 3 ) T || OP and ( D1 , D2 , D3 ) T || OA or OB

ν = 2 π c0/λ …optical angular frequency, (k1, k2, k3)T…wave vector
The electrical field components are found via Di = χil El, where the dielectric coefficients
take other values than in the description of piezo-electricity in appendix 5.1, since optical
frequencies are present. In case of uniaxial crystals and z-paraxial light waves the difference
between the direction of D and E can be neglected.
114 F. Bammer
The length of the half axis OA or OB is the corresponding refractive index nA or nB. The
axis with the lower refractive index is the fast axis.
If both possible waves are present in a given direction, then after a traveling way L a
phase difference between the two waves develops. Usually it is described as retardation by
δ = 2 π L(nB - nA)/λ . (A-4)
At the exit of the crystal the superposition of the two waves yields in general an
elliptically polarized wave. The precise value of δ and the amplitude of the partial waves
decide about the resulting polarization.
5.3.2. Additional Artificial Birefringence

If an external electrical field or strain is present the indicatrix (Eq. A-3) is modified:
2⎛ 1 ⎞ 2⎛ 1 ⎞ 2⎛ 1 ⎞
x1 ⎜⎜ 2 + B1 ⎟⎟ + x 2 ⎜⎜ 2 + B2 ⎟⎟ + x3 ⎜⎜ 2 + B3 ⎟⎟ + 2 B4 x 2 x3 + 2 B5 x3 x1 + 2 B6 x1 x 2 = 1
⎝ n1 ⎠ ⎝ n2 ⎠ ⎝ n3 ⎠
All Bi are collected in a 6x1-vector B and are calculated for example by:
B = PE S + RS E , (A-5)
or
B = ΠE T + RT E . (A-6)
PE...6x6-matrix of the photo-elastic strain-coefficients for constant electrical field

ΠE...6x6- matrix of the photo-elastic tension-coefficients for constant electrical field
RS...electro-optic 6x3-matrix for constant strain
RT... electro-optic 6x3-matrix for constant tension
The structure of the matrices PE and ΠE is similar to the structure of the stiffness and
compliance matrices cE, cD, sE, sD and the structure of the electro-optic matrices is always
equal to that of the piezo-electric e, h, d, g-matrices.
The relation between RS and RT is found when S = sE T + d T E (Eq. A-2 left) is put into
A-5.
B = PE (sE T + d T E) + RS E = PE sE T + (PE d T + RS) E
If this is compared with Eq. A-6 one obviously obtains
RT = PE d T + RS and ΠE = PE sE .
Alternatively the following calculation of B , which uses the pairs (S, D) or (T, D), is
possible:
B = PD S + QS D ,
B = ΠD T + QT D .
If no electrodes are placed on the crystal, only the matrices PD und ΠD can be determined,
since in this case the electric displacement D is constant. The relations between PD and PE or
ΠE and ΠD can be found in a similar way as for RS and RT . One obtains
PD = PE − RS h and ΠD = ΠE − RT g.
The electro-optical matrices QS and QT are not used and were only introduced for a
consistent representation here 19.
5.4. Transmission In Dependence on δ and φ
To get some insight into the formulas for polarization and polarizing elements we derive
Eq. 40 of chapter 2.11. We calculate the transmission for linearly x-polarized light going
through a retardation plate and an x-oriented polarizer. The retarder parameters are the
retardation δ (Eq. A-4) and the angle φ, included between the x-axis and the fast axis of the
retarder.
We write for the electrical field strength at one point on the input path
Ein(t) = sin νt .
(here ν = 2π c0/λ...angular frequency of the light). In the retardation plate the wave is split in
linear polarized partial waves. The polarizations are oriented along the fast and the slow axis
of the retarder. The temporal courses of the two waves after passing the retarder are described
by
sin φ sin νt and cos φ sin (νt + δ) .
The polarizer transmits of both waves only the x-polarized parts with the amplitudes
sin2φ sin νt and cos2φ sin (νt + δ) .
Behind the polarizer the superposition of both waves yields the electric field amplitude
Eout(t) = sin2φ sin νt + cos2φ sin (νt + δ) .
The intensity is proportional to the temporal integral of the electrical field amplitude. To
determine the transmission T related to the incoming intensity we write
tS
T=
∫ 0
Ein (t ) 2 dt
tS
∫ 0
Eout (t ) 2 dt
19
[16, 17, 18, 22] fails in doing that.
116 F. Bammer
The integration time tS depends on the type of sensor used. This is ~100ms for the human
eye and 1 ns for a fast photo detector. One oscillation period of visible light lasts ~1 fs
(= 10−15s). Therefore even within one nanosecond ~106 cycles happen and the following can
be used:
tS
tS tS ∫ sinν t cosν t dt
∫ sin 2 (ν t + δ ) dt = ∫ sin 2 ν t dt
0
, tS
=0
∫
0 0
sin 2 ν t dt
0
Herewith the transmission T and the reflection R (if the polarizer is a polarizing beam
splitter cube) in dependence on δ can be calculated:
∫ (sin ϕ sinν t + cos ϕ sin(ν t + δ )) dt

2 2 2
T (δ ) =
∫ sin ν t dt
2
∫ (sin ϕ sin ν t + cos ϕ sin (ν t + δ ))dt + 2 sin ϕ cos ϕ ∫ (sinν t sin(ν t + δ ) )dt
4 2 4 2 2 2
=
∫ sin ν t dt
2
(sin ϕ + cos ϕ ) ∫ sin ν t dt + 2 sin ϕ cos ϕ ∫ (sin ν t cos δ + sinν t cosν t sin δ )dt
4 4 2 2 2 2
=
∫ sin ν t dt
2
= (sin 4ϕ + cos 4ϕ ) + 2 sin 2ϕ cos 2ϕ cos δ

= (sin 4ϕ + cos 4ϕ ) + 2 sin 2ϕ cos 2ϕ (1 − 2 sin 2 (δ / 2)) = 1 − 4 sin 2ϕ cos 2ϕ sin 2 (δ / 2)
= 1 − sin 2 (2ϕ ) sin 2 (δ / 2)
This calculation yields Eq. 40. If φ = 45° as it is found in the standard PEM-configuration
(Figure 2) 20, then the result is reduced to Eq. 2.
5.5. Material Coefficients for BK7, LiNbO3 and LiTaO3
In the following we provide the elastic, dielectric, piezo-electric, elasto- and electro-
optical matrices for three materials that are significant in this work. Here the consistency of
the data is important. In appendix 5.1 (Eq. A-2) the equations to change between different
representations are given. For a complete set of material matrices for a piezo-electric material
one possible choice is for example cE, e, εS, RS, PE. All other matrices of the different
representations of elasticity, piezo-electricity and artificial birefringence can be calculated out
of this set (Eq. A-3). In literature these data sets, when taken from different sources, usually
do not agree to each other very well. It is therefore necessary to calculate for every used
material one consistent data set, if one wants to work in this area.
20
The situation here is that of Fig. 1, if the coordinate system in Fig. 1 is turned around the z-axis by 45° such that
the x-axis is parallel to the input polarization .
Regard that the optical parameters (n, no, ne, P, Π, PE, PD, RT, RS, Π) all depend on the
wavelength. However it is very difficult or even impossible to find the wavelength
dependence of these parameters. Usually only the refractive indices are tabulated vs.
wavelength.
5.5.1. BK7 21
Density ρ = 2510 kg/m³
Refractive index no = 1.507 (@ 1060nm)
⎛ 90.93 23.88 23.88 0 0 0 ⎞
⎜ ⎟
⎜ 23.88 90.93 23.88 0 0 0 ⎟
9
Stiffness matrix: c = ⎜ 23.88 23.88 90.93 0 0 0 ⎟ 10 N/m²
⎜ ⎟
⎜ 0 0 0 33.53 0 0 ⎟
⎜ ⎟
⎜ 0 0 0 0 33.53 0 ⎟
⎜ 0
⎝ 0 0 0 0 33.53 ⎟⎠
⎛ 12.35 − 2.57 − 2.57 0 0 0 ⎞

⎜ ⎟
⎜ − 2.57 12.35 − 2.57 0 0 0 ⎟
-12
Compliance matrix: s = ⎜ − 2.57 − 2.57 12.35 0 0 0 ⎟ 10 m²/N
⎜ ⎟
⎜ 0 0 0 29.83 0 0 ⎟
⎜ 0 0 0 0 29.83 0 ⎟⎟
⎜
⎜ 0 0 0 0 0 29.83 ⎟⎠
⎝
⎛ 0.284 1.86 1.86 0 0 0 ⎞

⎜ ⎟
⎜ 1 . 86 0.284 1.86 0 0 0 ⎟
-12
Piezo-elastic matrix: ∏ = ⎜ 1.86 1.86 0.284 0 0 0 ⎟ 10 m²/N
⎜ ⎟
⎜ 0 0 0 − 1.576 0 0 ⎟
⎜ 0
⎜ 0 0 0 − 1 .576 0 ⎟⎟
⎜ 0 0 0 0 0 − 1.576 ⎟⎠
⎝
⎛ 0.115 0.22 0.22 0 0 0 ⎞

⎜ ⎟
⎜ 0.22 0.115 0.22 0 0 0 ⎟
Photo-elastic matrix: P = ⎜ 0.22 0.22 0.115 0 0 0 ⎟
⎜ ⎟
⎜ 0 0 0 − 0.053 0 0 ⎟
⎜ 0 0 0 0 − 0.053 0 ⎟⎟
⎜
⎜ 0 0 0 0 0 − 0.053 ⎟⎠
⎝
5.5.2. LiNbO3 [22] (stoichiometric)
Density ρ = 4700 kg/m³, refractive indices no = 2.244, ne = 2.1647 (@960nm)
Stiffness-matrix for constant electrical field:
21
The matrices of BK7 are included here only to be able to compare the matrices of LiNbO3 and LiTaO3 with a
typical isotropic material.
118 F. Bammer
⎛ 203 53 75 9 0 0⎞
⎜ ⎟
⎜ 53 203 75 − 9 0 0⎟
cE = ⎜ 109 N/m²
75 75 245 0 0 0⎟
⎜ ⎟
⎜ 9 −9 0 60 0 0 ⎟
⎜ 0 0 0 0 60 9 ⎟⎟
⎜
⎜ 0 0 0 0 9 75 ⎟⎠
⎝
Stiffness-matrix for constant electric displacement:
⎛ 219.1 37.2 76.1 − 14.6 0 0 ⎞
⎜ ⎟
⎜ 37.2 219.1 76.1 14.6 0 0 ⎟
cD = ⎜ 109 N/m²
76.1 76.1 251.9 0 0 0 ⎟
⎜ ⎟
⎜ − 14.6 14.6 0 94.9 0 0 ⎟
⎜ 0 0 0 0 94.9 − 14.6 ⎟⎟
⎜
⎜ 0 0 0 0 − 14.6 90.9 ⎟⎠
⎝
Compliance matrix for constant electrical field:

⎛ 5.77 − 1.01 − 1.46 − 1.02 0 0 ⎞
⎜ ⎟
⎜ − 1.01 5.77 − 1.46 1.02 0 0 ⎟
sE = ⎜ 10−12 m²/N
− 1.46 − 1.46 4.97 0 0 0 ⎟
⎜ ⎟
⎜ − 1.02 1.02 0 16.97 0 0 ⎟
⎜ 0 0 0 0 16.97 − 2.04 ⎟⎟
⎜
⎜ 0 0 0 0 − 2.04 13.58 ⎟⎠
⎝
Compliance matrix for constant electric displacement:
⎛ 5.2 − 0.44 − 1.44 0.87 0 0 ⎞
⎜ ⎟
⎜ − 0.44 5 .2 − 1.44 − 0.87 0 0 ⎟
sD = ⎜ 10−12 m²/N
− 1.44 − 1.44 4.84 0 0 0 ⎟
⎜ ⎟
⎜ 0.87 − 0.87 0 10.8 0 0 ⎟
⎜ 0 0 0 0 10.8 1.73 ⎟⎟
⎜
⎜ 0 0 0 0 1.73 11.27 ⎟⎠
⎝
Piezo-electric e-matrix:
0 3.7 − 2.5 ⎞
e = ⎛⎜
0 0 0
⎟ C/m²
⎜ − 2.5 2.5 0 3.7 0 0 ⎟
⎜ 0.2 0.2 1.3 0 0 0 ⎟⎠
⎝
Piezo-electric d-matrix:
67.89 − 41.48⎞
d = ⎛⎜ 0 0 0 0
⎟ C/N or m/V
⎜ − 20.74 20.74 0 67.89 0 0 ⎟
⎜ − 0.94 − 0.94 5.88 0 0 0 ⎟⎠
⎝
Piezo-electric h-matrix:
0 9.43 − 6.37 ⎞
h = ⎛⎜ 0 0 0
⎟ 109 N/C or V/m
⎜ − 6.37 6.37 0 9.43 0 0 ⎟
⎜ 0.82 0.82 5.32 0 0 0 ⎟⎠
⎝
Piezo-electric g-matrix:
90.87 − 55.52 ⎞
g = ⎛⎜
0 0 0 0
⎟ 10-3 m²/C
⎜ − 27.76 27.76 0 90.87 0 0 ⎟
⎜ − 3.74 − 3.74 23.38 0 0 0 ⎟⎠
⎝
Dielectric matrix for low frequency or constant tension 22:
χT = ⎛⎜ 84.38 0 0 ⎞
⎟ 8.854 10−12 As/Vm
⎜ 0 84.38 0 ⎟
⎜ 0 0 28.42 ⎟⎠
⎝
Dielectric matrix for high frequency or constant strain:

⎛ 44.3 0 0 ⎞
χS = ⎜ ⎟ 8.854 10−12 As/Vm
⎜ 0 44.3 0 ⎟
⎜ 0 0 27.6 ⎟⎠
⎝
Electro-optic matrix for low frequency or constant tension:

32.05 − 6.65 ⎞
RTT = ⎛⎜
0 0 0 0
⎟ 10−12 m/V
⎜ − 6.65 6.65 0 32.05 0 0 ⎟
⎜ 9.71 9.71 30.09 0 0 0 ⎟⎠
⎝
Electro-optic matrix for high frequency or constant strain:

⎛ 0 0 0 0 17 − 3.4 ⎞
RST = ⎜ ⎟ 10−12 m/V
⎜ − 3.4 3.4 0 17 0 0 ⎟
⎜ 9 9 30 0 0 0 ⎟⎠
⎝
Photo-elastic matrix for constant electrical field:

⎛ − 0.025 0.08 0.13 − 0.08 0 0 ⎞
⎜ ⎟
PE = ⎜ 0.08 − 0.025 0.13 0.08 0 0 ⎟
⎜ 0.17 0.17 0.07 0 0 0 ⎟
⎜ ⎟
⎜ − 0.15 0.15 0 0.13 0 0 ⎟
⎜ 0 0 0 0 0 . 13 − 0. 15 ⎟
⎜ ⎟
⎜ 0 0 0 0 − 0.08 − 0.0525 ⎟⎠
⎝
Photo-elastic matrix for constant electric displacement:

⎛ − 0.054 0.094 0.082 − 0.048 0 0 ⎞
⎜ ⎟
PD = ⎜ 0.094 − 0.054 0.082 0.048 0 0 ⎟
⎜ 0.145 0.145 − 0.09 0 0 0 ⎟
⎜ ⎟
⎜ − 0.042 0.042 0 − 0.03 0 0 ⎟
⎜ 0 0 0 0 − 0.03 − 0.042 ⎟⎟
⎜
⎜ 0 0 0 0 − 0.048 − 0.074 ⎟⎠
⎝
5.5.3. LiTaO3 [26, 35]

Density ρ = 7465 kg/m³, refractive indices no = 2.1787, ne = 2.1821 (@633nm)
Stiffness matrix for constant electrical field:
22
The factor 8.854 10−12 As/Vm is the dielectric field constant.
120 F. Bammer
⎛ 233.1 46.4 85.2 − 16.1 0 0 ⎞

⎜ ⎟
⎜ 46.4 233 . 1 85. 2 16. 1 0 0 ⎟
cE = ⎜ 109 N/m²
85.2 85.2 276.1 0 0 0 ⎟
⎜ ⎟
⎜ − 16.1 16.1 0 102.9 0 0 ⎟
⎜ 0 0 0 0 102.9 − 16.1⎟⎟
⎜
⎜ 0 0 0 0 − 16.1 93.35 ⎟⎠
⎝
Stiffness matrix for constant electric displacement:
⎛ 241.93 37.64 84.62 − 28.67 0 0 ⎞
⎜ ⎟
⎜ 37.64 241.93 84.62 28.67 0 0 ⎟
cD = ⎜ 109 N/m²
84.62 84.62 286.27 0 0 0 ⎟
⎜ ⎟
⎜ − 28.67 28.67 0 120.87 0 0 ⎟
⎜ 0 0 0 0 120.87 − 28.67 ⎟⎟
⎜
⎜ 0 0 0 0 − 28.67 102.15 ⎟⎠
⎝
Compliance matrix for constant electrical field:
⎛ 4.96 − 0.55 − 1.36 0.86 0 0 ⎞
⎜ ⎟
⎜ − 0. 55 4. 96 − 1 . 36 − 0 .86 0 0 ⎟
sE = ⎜ 10−12 m²/N
− 1.36 − 1.36 4.46 0 0 0 ⎟
⎜ ⎟
⎜ 0.86 − 0.86 0 9.99 0 0 ⎟
⎜ 0 0 0 0 9.99 1.72 ⎟⎟
⎜
⎜ 0 0 0 0 1.72 11.01⎟⎠
⎝
Compliance matrix for constant electric displacement:
⎛ 4.8 − 0.44 − 1.29 1.24 0 0 ⎞
⎜ ⎟
⎜ − 0 . 44 4 . 8 − 1 . 29 − 1. 24 0 0 ⎟
sD = ⎜ 10−12 m²/N
− 1.29 − 1.29 4.25 0 0 0 ⎟
⎜ ⎟
⎜ 1.24 − 1.24 0 8.86 0 0 ⎟
⎜ 0 0 0 0 8.86 2.49 ⎟⎟
⎜
⎜ 0 0 0 0 2.49 10.49 ⎟⎠
⎝
Piezo-electric e-matrix:
⎛ 0 0 0 0 2.63 − 1.84 ⎞
e= ⎜ ⎟ C/m²
⎜ − 1.84 1.84 0 2.63 0 0 ⎟
⎜ − 0.11 − 0.11 1.93 0 0 0 ⎟⎠
⎝
Piezo-electric d-matrix:
⎛ 0 0 0 0 23.1 − 15.73 ⎞
d= ⎜ ⎟ C/N or m/V
⎜ − 7.86 7.86 0 23.1 0 0 ⎟
⎜ − 3.11 − 3.11 8.91 0 0 0 ⎟⎠
⎝
Piezo-electric h-matrix:
⎛ 0 0 0 0 6.83 − 4.78 ⎞
h= ⎜ ⎟ 109 N/C or V/m
⎜ − 4.78 4.78 0 6.83 0 0 ⎟
⎜ − 0.3 − 0.3 5.27 0 0 0 ⎟⎠
⎝
Piezo-electric g-matrix:
⎛ 0 0 0 0 48.67 − 33.14 ⎞
g= ⎜ ⎟ 10-3 m²/C
⎜ − 16. 57 16 . 57 0 48. 67 0 0 ⎟
⎜ − 8.09 − 8.09 23.19 0 0 0 ⎟⎠
⎝
Dielectric matrix for low frequency or constant tension 23:

⎛ 53.6 0 0 ⎞
⎜ ⎟ 8.854 10−12 As/Vm
χT = ⎜ 0 53.6 0 ⎟
⎜ 0 0 43.4 ⎟⎠
⎝
Dielectric matrix for high frequency or constant strain:

⎛ 43.47 0 0 ⎞
⎜ ⎟ 8.854 10−12 As/Vm
χS =⎜ 0 43.47 0 ⎟
⎜ 0 0 41.38 ⎟⎠
⎝
Electro-optic matrix for low frequency or constant tension:

21.98 − 0.33 ⎞
RT T = ⎛⎜ 0 0 0 0
⎟ 10−12 m/V
⎜ − 0 .33 0 .33 0 21.98 0 0 ⎟
⎜ 7.83 7.83 29.35 0 0 0 ⎟⎠
⎝
Electro-optic matrix for high frequency or constant strain:

20 − 1⎞
RS T = ⎛⎜
0 0 0 0
⎟ 10−12 m/V
⎜ −1 1 0 20 0 0⎟
⎜ 7 7 30.3 0 0 0 ⎟⎠
Photo-elastic matrix for constant electric field:

⎛ − 0.081 0.081 0.93 − 0.026 0 0 ⎞
⎜ ⎟
PE = ⎜ 0.081 − 0.081 0.93 0.026 0 0 ⎟
⎜ 0.089 0.089 − 0.044 0 0 0 ⎟
⎜ ⎟
⎜ − 0.085 0.085 0 0.028 0 0 ⎟
⎜ 0 0 0 0 0.028 − 0.085 ⎟⎟
⎜
⎜ 0 0 0 0 − 0.026 − 0.081⎟⎠
⎝
Photo-elastic matrix for constant electric displacement:

⎛ − 0.084 0.088 0.056 − 0.019 0 0 ⎞
⎜ ⎟
PD = ⎜ 0.088 − 0.084 0.056 0.019 0 0 ⎟
⎜ 0.098 0.098 − 0.204 0 0 0 ⎟
⎜ ⎟
⎜ 0.011 − 0.011 0 − 0.109 0 0 ⎟
⎜ 0 0 0 0 − 0.109 0.011 ⎟⎟
⎜
⎜ 0 0 0 0 − 0.019 − 0.086 ⎟⎠
⎝
23
The factor 8.854 10−12 As/Vm is the dielectric field constant.
122 F. Bammer
REFERENCES
[1] www.hindsinstruments.com.
[2] J. C. Kemp, 1987, Polarized Light and its Interaction with Modulating Devices,
Information sheet, Hinds Instruments.
[3] F. Bammer, B. Holzinger, T. Schumi, 2006, Time multiplexing of high power laser
diodes with single crystal photo-elastic modulators, Optics Express, Vol. 14, No. 8.
[4] Siegmann, 1986, Lasers.
[5] F. Bammer, R. Petkovsek, 2007, Q-switching of a fiber laser with a single crystal
photo-elastic modulator, Optics Express, Vol. 15, No. 10.
[6] F. Bammer, R. Petkovsek, D. Schuöcker, J. Mozina, 2008, Q-switching a fiber laser
with a dual mode single crystal photo-elastic modulator, SPIE-Proceeding of Photonics
Europe 2008, Solid State Lasers and Amplifiers III, Vol. 6998.
[7] T. Schumi, 2006, Analyse von photoelastischen Modulatoren aus Lithiumniobat,
Diploma Thesis, Vienna University of Technology.
[8] F. Bammer, B. Holzinger, 2007, Time-multiplexing generates a diode laser beam with
high beam quality, Optics and Laser Technology, 39, p. 1002 - 1007.
[9] F. Bammer, B. Holzinger, 2006, A compact high brilliance diode laser, SPIE-
Proceeding of Photonics West 2006, High-Power Diode Laser Technology and
Applications IV, Vol. 6104.
[10] D. Schuöcker, 1998, Handbook of the EuroLaser Academy, Chapman and Hall.
[11] F. Bammer, 2005, Verfahren und Vorrichtung zur Schaffung eines Laserstrahls hoher
Brillanz, Patent, EP1588462.
[12] R. Diehl, 2000, High-Power Diode Lasers, Springer Publishing Company
[13] J.C. Canit, J. Badoz, 1983, New Design for a Photoelastic Modulator, Appl. Opt. 22,
592
[14] J. Eichler, H.J. Eichler, 2006, Laser, Springer Publishing Company.
[15] L. Turi, Cs. Kuti, and F. Krausz, 1990, Piezoelectrically Induced Diffraction
Modulation of Light, IEEE Journal of Quantum Electronics, Vol. 26, No. 7.

[16] M. Bass, 1995, Handbook of Optics, McGraw Hill.
[17] J. F. Nye, 1985, Physical Properties of Crystals, Oxford University Press.
[18] D. Royer, E. Dieulesaint, 2000, Elastic Waves in Solids, Springer-Verlag.
[19] Melles Griot, 2000, Product catalogue, Melles Griot.
[20] F. Bammer, 2007, Photoelastischer Modulator und Anwendungen, Patent, AT 502 413.
[21] F. Bammer, B. Holzinger, T. Schumi, 2007, A single crystal photo-elastic modulator;
SPIE-Proceeding of Photonics West 2007, Optical Components and Materials IV,
Vol. 6469.
[22] K. Wong, 2002, Lithium Niobate, INSPEC, London.
[23] E. Böhmer, 2000, Elemente der angewandten Elektronik, Vieweg Publishing Company.
[24] Bergmann-Schäfer, 1993, Optik, Lehrbuch der Experimentalphysik 3, de Gruyter,
Berlin.
[25] Collett E, 2003, Polarized Light in Fiber Optics, The PolaWaveGroup.
[26] www.roditi.com/SingleCrystal/Lithium_tantalate/LiTaO3_Properties.html.
[27] Yariv, 1984, Optical Waves in Crystals, New York, Wiley.
[28] T. A. Maldonado, 1995, Electro-optic modulators, Handbook of optics, McGraw-Hill
[29] W. Koechner, 1999, Solid State Laser Engineering, Springer Publishing Company.
[30] D. Yang, C. Canit, E. Gaignebet, 1995, Photoelastic Modulator: Polarization
modulation and phase modulation, J. Optics (Paris), vol. 26, n° 4, pp. 151-159.
[31] www.highqlaser.at
[32] J.C. Kemp, Piezo-Optical Birefringence Modulators, J. Opt. Soc. Am. 59, 950, 1969.
[33] Dubbel, 1986, Taschenbuch für den Maschinenbau, Springer Publishing Company.
[34] N. W. Ashcroft, N. D. Mermin, 1986, Solid State Physics.
[35] R. T. Smith, 1967, Elastic, Piezoelectric, and Dielectric Properties of Lithium
Tantalate, Applied Physics Letters, 11, (5): pg. 146-148.
[36] R. Weil and D. Halido, 1974, Resonant-piezoelectro-optic light modulation, Journal of
Applied Physics, Vol. 45, No. 5.
[37] F. Bammer, R. Petkovsek, B. Schulz, M. Frede, Q-switching with a Dual Mode Single
Crystal Photo-Elastic Modulator, SPIE-Proceeding of GCL/HPL 2008, Novel
Approaches, to be published.
Chapter 4
DIFFRACTIVE MICROOPTICS FOR TECHNOLOGICAL

IR-LASERS
V. S. Pavelyev1, V. A. Soifer1, V. I. Konov2, V. V. Kononenko2 and

A. V. Volkov1
1
Image Processing Systems Institute RAS, Samara, Russia
Samara State Aerospace University, Samara, Russia
2
Natural Sciences Center of General Physics Institute, Moscow, Russia
ABSTRACT
Wide use of infrared (IR) lasers in technological applications (metal cutting and
hardening, laser evaporation, laser bending and etc.) have stimulated a research in the
field of synthesis of IR-range diffractive microoptical focusing elements, which could

retain their functionality under continuous irradiation with high power. Successive
fulfillment of this requirement is stipulated both by selection of an appropriate substrate
material and by development of technologies for creation a diffractive microrelief on a
substrate as well as numerical methods for microrelief optimization taking into account
possible technological constraints.
The present chapter contains a consideration of modern approaches to the synthesis
of diffractive optical elements (DOE) for technological IR-laser beams focusing.
Experimental results of produced elements investigation described also.
Different ways, from geometrical optics approximation to stochastic optimization,
are used for calculation of a diffractive microrelief of IR-range optical elements that
focus an illuminating beam into a pre-given 2-dimensional area. Near-IR DOE synthesis
on sapphire substrates by plasmochemical etching method is considered.
Several diamond focusing DOEs for operating at CO2-laser wavelength (λ=10.6 μm)
are produced and presented. The nanosecond UV-laser ablation and plasmochemical
lithographic etching technologies for DOE synthesis on CVD diamond plates are
discussed.
Being a perspective optical material, CVD diamond has a high refractive index
(n=2.4), resulted in relative high losses by Fresnel reflection. The present chapter
126 V.S. Pavelyev, V.A. Soifer, V.I Konov et al.
includes consideration of a realization of antireflecting structures on a diamond film

surface by UV-laser ablation.
The advent of silica and silver-halide IR-range optical fibers for technological
applications has set tasks of optimization methods development and of forming
diffractive microreliefs on optical fiber surface for waveguide beam controlling. The
present chapter includes the results of experimental investigation of diffractive
microrelief and antireflecting structures that are realized on an end-face of IR-fiber.
INTRODUCTION
Development of methods for synthesis of diffractive optical elements (DOEs) – optical
wafers coated with a double- or single-sided diffractive microrelief – has created new
functionalities with regard to laser radiation control previously unattainable by means of
conventional optical elements.
These include focusing a laser beam into a designed area, analysis of the transverse mode
content of the beam, generation of a beam with designed transverse mode content, and the
like. The diffractive optical elements can be implemented for operation both in transmission
and in reflection modes.
Due to many technical applications, focusing the laser beam into a designed two-
dimensional (2D) domain has become a most popular task. The formulation of the problem of
synthesis of such a DOE is depicted in Figure 1.
In early 80-s, first DOEs to focus light into designed areas (a circle, a rectangular, a
transverse line, an axial line, and so on) were described [1,2] and patented [3-5]. Such
elements were designed using the geometric optics approximation. In a number of cases,
DOEs designed in the geometric optics approach allow a high-quality focusing of the initial
beam into the desired area. Results of characterization of a DOE to focus the Gaussian beam
of the CO2-laser (wavelength λ=10.6 µm) into a rectangle with >98% energy efficiency and
<10% error were presented in Ref. [6]. The intensity distribution in the DOE focal plane is
shown in Figure 2.
v
η
u ξ
W0
z z
I0
τ
Figure 1. The formulation of the problem of focusing a laser beam with complex amplitude W0(u,v) in
the cross-section into a designed area in (ξ ,η)-plane.
Diffractive Microoptics for Technological IR-Lasers 127
Figure 2. The intensity distribution generated by gauss-to-rectangle focusing DOE.
Successful characterization results of reflecting DOEs designed in the geometric optics

approach to focus the radiation of the high-power industrial CO2-laser (up to 3 kW) into a
circle and a line were reported in Ref [7]. However, designing the diffractive microrelief
within the geometric optics approach gives no way of accounting for diffraction effects,
generally resulting in deviations from the desired intensity distribution. With regard to
focusing the high-power laser radiation, such deviations are extremely undesirable from the
ecological point of view.
Advances in the computing facilities have given an impetus to the rapid development of
iterative algorithms for designing focusing DOEs in the scalar diffraction theory. The iterative
synthesis of DOEs has been discussed in numerous publications, e.g. [8-10]. The choice of
the optical substrate material and a diffractive microrelief fabrication technique is of utter
importance for the high-power laser. The material of choice must have high radiation
resistance and good transmission (reflection) in the required spectral range. The fabrication
technology is supposed to ensure the required accuracy of the resulting microrelief. Inevitable
systematic errors inherent in the technology chosen must be accounted for in the algorithm of
microrelief calculation.
This chapter is devoted to the synthesis of DOEs intended to focus beams of industrial IR
lasers. The use of optical materials, such as sapphire and CVD diamond films, that show
much promise for IR range is discussed. Topics related to generating the diffractive
microrelief on the surface of silver-halide waveguides for transmitting power IR laser light
are considered.
Development of iterative procedures for designing optical elements to focus the laser
light into required two-dimensional areas with due regard for microrelief fabrication errors is
discussed.
1. DIAMOND DIFFRACTIVE OPTICS FOR CO2-LASERS

1.1. Set of the Problem
The fabrication of diamond-based optical elements for high-power CO2 lasers is of

particular interest because of the low optical absorption coefficient of this material in
combination with it’s very high thermal conductivity and the weak temperature dependence
of refractive index [11]. Recent advances in gas-phase synthesis have made it possible to
fabricate polycrystalline CVD diamond films (DF) whose optical and thermal properties are
close to those of single crystal diamond material (see Table 1), whereas they are far cheaper.
As a result, these sophisticated materials are applied more and more to tasks dominated till
now by other materials. Such examples for this are windows for high-power CO2 lasers in the
5 – 20 kW domain [11] and beam-splitters [12]. However, the use of high-quality CVD
diamond plate as a substrate of more complicated optical element is limited by substantial
Fresnel-reflection losses (about 30%) and small thickness (up to 1-2 mm) which restricts an
aperture of a classical optical element. DF surface antireflective structuring using the
selective ablation of the surface of the diamond-like substrate by the excimer UV laser light is
discussed [13,14]. UV laser ablation has been successfully applied for manufacturing the
diffractive lens micro-relief [15,16]. The multilevel cylindrical lens of wavelength λ=10.6 μm
was used as an example [15] in studying the possibilities of laser-assisted generation of
microrelief for diamond DOEs operating in the far IR range.
The manufactured diamond diffractive optical elements (DOEs) to focus the CO2 laser
light are studied experimentally [17,18] and the results are discussed.
However, the technology [15-18] does not allow the diffractive micro-relief to be
generated with lateral resolution better than 10-20 μm, which certainly decreases the potential
fields of the technology application.
The first results of ion-chemical etching technology application for diffractive grating
formation on the surface of diamond film have been presented [19,20].

The results of diamond micro-lenses and sub-wavelength antireflection structures
formation by plasma etching were presented also. The present paper is devoted to further
development of considered technologic approaches [15-18,19].
Table 1.
Physical parameter CVD diamond films ZnSe

Refractive Index, n 2.38-2.42 2.4
Transmission at 10,6 μm 0.68-0.71 0.71
Thermal conductivity 300 K, (W/cm K) 18-20 0.16
Adsorbtion at 10,6 μm (1/см) 5*10-2 5*10-4
Hardness (GPa) 81±18 1.05
Coefficient of linear thermal Expansion (10-6K-1) 1.0 7.1
at 300K
Dn/dT (1/K) 9.6*10-6 57*10-6
1.2. Laser Microprocessing of Diamond Film Surface
Diamond films 300-400 µm thick were grown on polished Si substrates by a CVD

technique using a microwave plasma chemical reactor (Model ASTeX-PDS19, 5 kW power,
2.45 GHz frequency)[21]. Upon separation from the substrates, the resulting free-standing
diamond plates were cut with a laser, mechanically polished, and then were used for laser
patterning experiments.
For selective-area material removal a KrF excimer laser (model EMG 1003i «Lambda
Physik»), operating at 248 nm, was used as the laser source in a projection optical scheme.
The pulse duration is 15 ns, and the laser pulse energy is typically ~200 mJ, although only a
small fraction of the output energy is utilized. The pulse repetition rate is 50 (Hz), and the
pulse-to-pulse instability of the laser energy was 7%. The image of a mask (square) was
projected onto the sample surface by a short-focal length objective with linear
demagnification of 1:15.
A diamond sample placed on a computer-driven X-Y stage was translated controllably so
that a selected region of given coordinates on the diamond surface was irradiated by a certain
number of laser shots to achieve the resulting surface profile close to the calculated one. The
etching depth was controlled by the laser fluence and the number of laser shots. A surface
profiler “Zygo” (model New View 5000) based on phase-shifting interferometry was used to
examine the topology of original and laser-irradiated surface.
One of the main tasks in the laser ablation study of diamond is to determine the ablation
characteristics for a particular diamond plate to be further used in laser patterning
experiments.
The first reason is that the ablation character (value of the ablation threshold and the
ablation rate-on-fluence behavior) depends strongly on the physical properties of a CVD
diamond material [11]. And, secondly, since the ablation depth is determined as a product of
the ablation rate and the number of laser shots, the exact data on the ablation rate-on-fluence
dependence is needed to reach the resulting surface profile of a DOE with a high accuracy.
The ablation rate was measured as a function of the KrF laser fluence for the laser spot
size of 40x40 μm2.
Figure 3 shows the dependence of the ablation depth on the number of laser shots for the
energy density varied from 10 to 60 J/cm2. It is seen that at a given fluence the depth
increases linearly with the number of shots (at least, to the depth of 25-30 µm). These
experimental plots are taken to determine the average ablation rate at a fixed laser fluence
(Figure 3, down).
It can be inferred from the ablation rate-on-fluence dependence that, if the ablation depth
were regulated only by variation in the number of shots, the required accuracy of the
microstructure height would be provided in the range of low fluences; e.g., for E=10.7 J/cm2
the etching rate is ν=95 nm/pulse. However, in this case the productivity of laser treatment is
rather low, making it time consuming to fabricate a diamond DOE with the parameters shown
in Table 2.
30
2
56 J/cm
25
2
27 J/cm
20 2
10 J/cm
Depth, μm
15
10
0
0 10 20 30 40 50 60 70 80 90 100
Number of pulses
350
Etching rate, nm/pulse
300
250
235 nm/pulse
200
150
100
50
10 20 30 40 50 60
2
Laser fluence, J/cm
Figure 3. Characteristics of laser ablation of CVD diamond plate by KrF excimer laser: (up) ablation
depth vs number of pulses at different fluences, and (down) ablation rate vs laser fluence.
Table 2.
Fresnel lens “Gauss-To-Square Contour” focuser

Focal length 127 mm 100 mm
Operating wavelength 10.6 μm 10.6 μm
Level amount of quantization 8 8
Aperture 13 mm 6.64×6.64 mm
Discretization step 50 μm 40 μm
Calculated efficiency 95% 75%
Measured efficiency 86% 62%
Method of calculation Analytically Adaptive iterative procedure
To increase the processing rate requires higher ablation rates (higher laser fluences) but,
if ν>0.1·hmax/M, a proper optimization of irradiation condition is needed for a simple equality,
ν-N ≅ hmax/M, be fulfilled, where hmax – maximal etching depth, M – number of relief
quantization levels. For manufacturing of the 2-D Fresnel lens, we selected the etching rate of
235±5 nm/pulse obtained at E=30.5 J/cm2. The scatter in the ablation rate values is defined by
the accuracy of the etching depth measurements and variations in the pulse energy. It should
be noted that the ablation rate measurements (Figure 3) were performed for the number of
laser shots in the range N=30-100, whereas the formation of the first-level relief structures
requires only four laser shots. It is clear that the influence of the pulse-to-pulse instability
(7%) of laser energy on fluctuations in the ablation depth increases when a small number of
laser shots is applied, and this can cause some deviations of the depth of individual craters as
compared with the average depth of the structures of the same approximation level. The
degree of the influence of laser instability on the depth fluctuations of the M-level structures
can be found from optical tests of a final element.
After optimization of the laser irradiation conditions, the designed diffractive microrelief
of the lens or beam shaper were reproduced on the surface of diamond plates. The diamond
sample was discretely scanned back and forth, and the laser-induced relief consisted of a
raster of square regions with various depths as shown in Figure 4a.
a) b)
c)
Figure 4. Typical fragment of the realized diffractive microrelief (a) and DOE (Fresnel lens) view
before (b) and after (c) annealing in ambient air.
During laser treatment, a thin graphite-like layer is formed at the ablated diamond
surface. Evidently, this layer must be completely removed, as even a very thin layer of the
absorbing graphitic material decreases both the optical transmission and damage threshold of
the optical element by CO2 radiation. It was reported [22,23] that this absorbing surface layer
can be effectively removed either by hydrogen plasma etching or by oxidation in air at
elevated temperatures. In the present study, the laser-structured diamond plates were annealed
in ambient air at 580°C.
The laser-ablated surface (inside single ablated square) is smooth, an observation which
was confirmed by comparative roughness measurements of original and laser-etched diamond
surfaces using an AFM. The initial roughness of the mechanically polished diamond surface
was σ0=20 nm. As a result of laser etching, the roughness increased to σ=40+100 nm for
different depths. Optical losses due to surface scattering can be estimated as 0.3-1.5%. This
estimation takes into account only incoherent light scattering. The presence of the periodic
substructures in the relief between adjacent ablated squares may be responsible for additional
diffractive optical losses, which should be pronounced in the increased intensity of non-zero
diffractive orders. Note that specific „borders“ have places between neighboring pixels as a
result of different character of interaction of UV radiation with diamond at the borders of
light square-like spot and in the central part of it. It was suggested that light diffraction at
these borders will not give significant income in forming of focal image because their feature
size was estimated less then wavelength. However, such effect leads to certain decreasing of
DOE diffractive efficiency [17,18].
1.3. Investigation of Diamond Doe Realized by Laser Microstructuring
Most of the optical tests of the spherical lenses were performed using a multimode CO2
laser with the quasi-Gaussian distribution of the intensity (beam size w1=1.55 mm). In the
focal region of the lens the intensity distribution I(u) retained the Gaussian form (Figure 5). A
computer calculation of the intensities in various planes of the focal region of the lens was
performed by numerical calculation of the Fresnel-Kirchhoff integral [24] for the fundamental
TEM00 mode of laser beam illuminating an optical element. The data of optical testing of the
diamond spherical lens is in good agreement with the results of computer simulations as well
as with optical tests of a conventional spherical KCl lens with the same focal distance (Figure
5). The experimental value of the diffractive efficiency of the Fresnel lens was 86% (after
antireflecting coating) which is close to the theoretical estimation.
The setup used for experimental investigation of the field formed by fabricated diamond
beam shaper (“Gauss-To-Square Contour” focuser) as well as of the incoming beam consists
of 5W CO2 laser (single-mode TEM00) with a divergence of 3.84 mrad (full angle) and a
beam-waist diameter of 3.5 mm (1/e2), and IR camera with 120x120 pixels, a pixel size of
100 μm, and an 8-bit intensity resolution. The experimental determination of the diffraction
efficiency was realized when the camera was replaced by the head of a power meter.
The structure of the beam read by computer-controlled IR camera is illustrated in Figure
6. Comparison with theoretical results manifests on one hand a very good correspondence,
regarding the formed intensity distribution.
DOE
1,0 lens (KCl)
theory
0,8
Intensity, (a.u.)
0,6
0,4
0,2
0,0
-400 -200 0 200 400

u, μm
Figure 5. The distribution of the CO2 laser radiation intensity I(u) in the focal plane of the spherical
lenses: experimental characteristics of the investigated DOEs (squares), of a conventional KCl lens with
the same focal distance (triangles), and corresponding results of computer simulation for ideal
diffractive elements (solid curve).
a)
b)
c)
Figure 6. Calculated (left) and measured (right) intensity distribution formed by the fabricated Gauss-
to-Square Contour focuser at different distance z between the DOEs plane and the plane under
investigation: (a) z=90 mm, (b) z=f=100 mm, (c) z=110 mm.
On the other hand in experiment a certain deviation of the rectangle intensity distribution
is visible, which is not predicted by theory for the ideal element. This deviation can be
explained by technological errors connected with certain UV laser nonstability during

ablation process.
The energy efficiency in this experiment was found to be e=62%, which is somewhat less
than the theoretical estimations mentioned in the Table 2. This difference may be accounted
for by influence of borders between neighbouring pixels.
Summarizing, we can conclude that technology of diamond DOE manufacturing by laser
ablation allow to form pregiven intensity distribution as well as wavefront with good quality.
1.4. Investigation of Dry Etching Technology Capabilities
The feasibility of generating a desired microrelief on the diamond film surface using ion-
chemical etching [25-29] was studied in two stages [20].
For the first stage, it was established that the microrelief on the DF surface can be
generated using the following etching techniques:
a) ion etching in an argon medium with the use of an high-frequency (HF)- system with
an independent ion source. The etching is anisotropic and non-selective;
b) ion-chemical etching (ICE) with the use of HF- systems with cathode or/and anode
coupling. The following mixtures can be used as the working gases: (Ar +O2), O2,
(N2 +O2), (He +O2). The etching is anisotropic and selective.
From the analysis conducted, it follows that due to its anisotropy and selectivity, the ion-
chemical etching (item b)) shows the best promise for generating the micro-relief on DFs.
1.5. Ion-Chemical Etching of Diamond Films

At present, at Image Processing Systems Institute of the Russian Academy of Sciences

(IPSI RAS), the microrelief on quartz, silicon, saphire, etc. substrates is generated using the
INTRA unit for ion-chemical etching of plates in the HF-discharge in the magnetic field for
lithography with feature size of 0.1-0.2 microns [18].
When choosing the specific modes of DF ICE on the INTRA unit, the gas mixtures
(Ar+O2) and (N2+O2) are considered most suitable [25-29]. Since the masses of ionized atoms
of Ar (40) and N (28) are significantly different, other conditions being equal, the physical
vaporization coefficient and, hence, the etching rate, will be higher for the (Ar+O2) mixture.
In addition, when using the (N2+O2) mixture for ion-chemical etching, harmful
compounds such as CN can be formed. Thus, an (Ar+O2) mixture was selected as the working
gas [19].
The total consumption of Ar and O2 in the course of experiments did not exceed 2.5
1/hour. The content of O2 in the (Ar+O2) mixture was varied from several percent to 45-50%.
When the O2 proportion in the (Ar+O2) mixture was less than 10-15%, the etching rate was no
greater than 14 nm/min.
An increase in the O2 content in the (Ar+O2) mixture up to 50 % and the application of
the longitudinal magnetic field with respect to the reactor's horizontal axis allows the rate of
the DF ICE to be enhanced up to 75 nm/min.
Use of higher etching rate appears inexpedient due to necessity to provide possibly
accurate approximation of the continuous microrelief function by a quantized function.
The HF-power supplied to the reactor in the course of the experimental adjustment of the
etching mode was within the range 550-600 W. The HF-power variations within the above-
mentioned range were specified by the variation of the total impedance of the (Ar+O2)
mixture plasma. Thus, an optimal (for the given conditions) etching medium was fitted — an
(Ar+O2) mixture with about 50% content of O2. In addition, the major technological modes of
the DF ICE were determined and a fairly large etching rate (about 75 nm/min) was attained.
The analysis of lithography capabilities and adjustment of the ICE modes have made it
possible to proceed to the next stage in solving the problem, i.e. choosing a mask material that
is stable against the (Ar+O2) plasma. The lifetime of such a mask should be no less than 100
minutes (to achieve the DF etching depth of 7-8 μm). At this stage, the sital (CT50-05), glass
(K8), and polycor substrates were coated in vacuum by the following materials: chrome (Cr),
niobium (Nb), niobium oxide (Nb2O5), and a resistant alloy PC 5402 (containing 54% of
nikel, 2% of chrome, with the rest being silicon). Although the methods for coating the above
materials are different and include, respectively, explosion evaporation, resistive heating by
sublimation, and HF-magnetron deposition, all the parameters were fitted in such a manner
that all the films were of the same thickness (in the range from 150 to 170 nm).
Experiments on etching the fabricated samples of the masking materials using the ICE on
the INTRA unit in the DF etching modes in the (Ar+O2) medium have shown that the metal
films of niobium and its oxide, Ni2O5, prove to be most stable against the (Ar+O2) plasma.
It was also established that there was no need in specially creating a Nb2O5 film (for
example, by anode deposition) because the oxide on the Nb mask surface unavoidably
emerged in the course of the ICE of DF in the (Ar+O2) medium due to the fairly large oxygen
content. Another argument in favor of using Nb as the masking material for the DF is that Nb
is readily etched by the ICE in CF4 medium.
The technological scheme for fabricating the diffractive microrelief on the DF is as
follows:
1. After preliminary cleaning, the DF surface is coated with a Nb film by vacuum

deposition.
2. The fabricated film is coated by a photoresistant film by centrifuging.
3. Using the familiar photolithography techniques and applying a desired photomask, a
photoresist mask is created on the surface of the metallic Nb film. The sample is
placed in the INTRA unit reactor and is etched in the CF4 medium through the
fabricated photoresist mask. Thus, a Nb mask is created on the DF surface (with the
remains of the photoresist mask).
4. With the sample left in the reactor, the CF4 mixture is replaced by the (Ar+O2)
mixture with a 50% of O2, and the DF is etched to a desired depth.
5. With the sample still in the reactor (without de-vacuuming the chamber), the (Ar+O2)
is again replaced by CF4 and the remains of the Nb2O5 and Nb films are etched off
for 5-6 minutes. The process of diffractive microrelief formation by etching is
schematically shown in Figure 7.
Figure 7. Sequence of technological operations of generating a diffractive microrelief.
The microrelief parameters of the DF surface were checked using techniques of white
light microinterferometry and microscopy.
The interferometrical profilometer WLI (Fraunhofer-Institute for Applied Optics and
Precision Engineering, Jena, Germany) and the scanning probe microscope Solver PRO were
used.
1.6. Analysis of Results
The surface profile of the test grating made by scanning probe microscope is shown in
Figure 8. Parameters for the grating microrelief were chosen as follows: grating period is
d=12 μm, relief height is h=3.5 μm.
The proposed technique for fabricating the desired microrelief using ion-chemical etching
was employed also to manufacture a rectangular grating structure of period d=200 μm and the
calculated etching depth h=7.5 μm on a DF substrate about 400 μm thick.
The surface profile of this grating made by white light micro-interferometry is shown in
Figure 9.
The etching depth of 7.5 μm was chosen to be equal to the maximal height of the
diamond diffractive microrelief (the refractive index is n=2.4) for the wavelength λ=10.6 μm:
hmax = λ ( n − 1) .
The initial results of the study of the test structures (Figure 8-10) have demonstrated that
the proposed technique for ion-chemical etching shows promise for generating the DF surface
microrelief.
Note that on the interfaces of different-depth domains there are no "ledges" or “borders”
that emerge when the DF microrelief is generated using the laser ablation [18].
Note also that there is good agreement between the generated structure depth and the
required etching depth.
Figure 8. Microrelief of realized test grating.

Figure 9. Microrelief of realized grating.
Figure 10. Microrelief of realized grating (cross-section).
2. DEVELOPMENT AND STUDIES OF A METHOD FOR GENERATING A

DOE MICRORELIEF IN SAPPHIRE SUBSTRATES
Nowadays, the wide use of high power lasers has brought to the forefront the urgent
necessity for harnessing DOEs operating in high-intensity range of radiation. This imposes
specific requirements on the DOEs substrate materials, including
1. High transmittance of the substrate in the DOE operating wavelength range.

2. High allowable operating temperatures of the DOE substrate material.
3. Chemical inertness of the substrate material when operating in high-temperature air
environment.
4. High thermal conductivity of the DOE substrate material.
5. Good processibility of the substrate material allowing a designed microrelief to be
produced on the surface.
It may appear fairly difficult to pick out a DOE substrate material complying with all of
the above-listed requirements, with a number of materials satisfying them within varying
limits, like substrates fabricated from natural or artificial diamonds, sapphire, and different
types of fused silica. Results on the characterization of DOEs fabricated on diamond-like
films, glass and fused silica were discussed in quite some detail in this Chapter and in [29-
30]. In the present authors’ opinion, diamond and diamond-like substrates are most promising
materials for fabricating DOEs for high-power IR lasers. Substrates fabricated from diamond
films show a number of unique qualities, namely, the highest known thermal conductivity
(Table 1), a fairly broad transparency interval, high refractive index (n=2.38÷2.42 for λ=10.6
µm), chemical inertness, and so on. However, these substrates have major disadvantages of
high cost and limited size. The use of the various types of fused silica as the DOE substrates
has its limitations owing to the low thermal conductivity thereof (<1 W/mK). At the same
time, the fused-silica-based substrates can be large enough, showing good surface quality (as
high as 14 and 15 grade) and acceptable optical characteristics. Sapphire can be used as an
alternative to the above-listed materials of the DOE substrate. Sapphire presents a natural or
synthetic single crystal α-aluminum oxide - Al2O3, structurally classified with trigonal
symmetry class, 3 m (Figure 11). Because of this, the values of a number of critical
characteristics, like the thermal conductivity coefficient, permittivity, thermal coefficient of
linear expansion, and so on, depend on the choice of the crystallographic orientation of the
bulk substrate. Leucosapphire is a colorless synthetic single crystal Al2O3 (sapphire)
containing no more than 10-4% of impurities [31].
The state-of-the art technological processes enable growing large-sized leucospphire
crystals with the strictly oriented crystallographic axes, thus ensuring the manufacture of
large-sized substrates with reproducible physical parameters.
Out of the familiar techniques of growing leucosapphire crystals, three major processes
that presently define the state-of-the-art manufacture technology for bulk leucosapphire
substrates worth special mentioning: a horizontal directed crystal growth process, Kiropulos
method (growth from the melt using seed), and Stepanov method [32].
The method of horizontal directed crystal growth from the Al2O3 melt enables production
of large leucosapphire wafers (up to 300х300х50 mm), which are well suited for the
fabrication of standard-size rectangular substrates. This method features the implementation
simplicity and good reproducibility, with the industrially produced crystals comparing well in
quality with the natural ones.
The Kiropulos method makes it possible to grow large bullet-shaped leucosapphire
crystals reaching up to 300 mm in diameter and 150-170 mm in height. Being a variety of
Chokhralsky method (growth of single crystal silicon bullets), this method allows large
crystals of high-quality leucosapphire to be grown.
Figure 11. Crystal structure of aluminum oxide (Al2O3).
Stepanov method allows one to grow shaped leucosapphire crystals by the film-fed growth
from the aluminum oxide melt using a Мо or W mould. With this method, leucosapphire bands
50÷80 mm wide, 0.5÷2,5 mm thick, and up to 1 m long can be produced.
The methods for growing crystals described above make it possible to form structurally
perfect single crystals of leucosapphire. High precision substrates in a wide variety of sizes,
perfect surface, and the enhanced stability of physiochemical properties are produced using such
crystals. Besides, due to solidity, high thermal and chemical stability, the leucosapphire substrates
can operate in rather severe conditions.
The optical and physiochemical properties of the leucosapphire substrates are as follows:
transparency range: 0.17 – 6.5 µm; refractive index: n=1.756 at λ=1µm and n=1.766 at
λ=0.633 µm [33]; Mohs hardness: 9 units.; density: ρ=3.98 g/cm3 (at 20°С); melting
temperature: 2040°С; admissible operating temperature in air, vacuum, argon, and nitrogen -
up to 1900°С [32]. The heat conductivity of α-Al2O3 crystals is also reasonably high [33],
with the maximum found at Т=40К. At the liquid nitrogen temperature, the heat conductivity
of α-Al2O3 is even higher than that of copper.
The technology of mass production of high-quality leucosapphire substrates has been
developed sufficiently well, offering much cheaper substrates compared with CVD diamond.
The proposed in [34] technique for fabricating a DOE microrelief in sapphire substrates
involves several stages. At the first stage, on the chemically pre-cleaned sapphire disk of
diameter 25 mm and thickness 3 mm, a solid chromium film of thickness d≈0.3 µm was
thermally sprayed in vacuum at Р=8⋅10-6 Torr (1,1⋅10-3 Pa) from a resistive molybdenum
spraying gun in the UVN-74P3 unit. Then, an FP-051k photoresist was applied by spinning at
n=3000 rev/min on the generated chromium film. Next, the resulting photoresist layer was
dried in a thermostat at Т=80÷90°С for 15÷20 min. The photoresist layer was then exposed
through a chromium photomask in the EM-5006 unit. The photoresist was developed in a
standard weak KOH solution. Following the hardening of the photoresist mask at Т=140°С
for 40÷50 min, the chromium was etched through the resulting photoresist. The solution of
cerium sulfate Ce(SO4)2 was used as the etching agent. Thus, as a result of the first stage, a
two-layer protective mask composed of Cr of thickness dCr≈0.3 µm and a photoresist of
thickness dph≈1 µm was fabricated in the sapphire substrate.
At the second stage, the ion-chemical etching of the sapphire was conducted on the unit
UTP PDE-125-009 (“INTRA”). To do a preliminary selection of etching parameters and the
matching of stable operation modes of the high-frequency generator (HFG), namely, the
mode matching with the reaction chamber, attainment of the maximal output power and
magnetic field strength, studies were conducted using a polycrystalline substrate Al2O3
(polycor). As a result of that quest, the etch rate of the polycor of about 13 nm/min in the СF4
(freon-14) medium was achieved. The etch modes of the polycrystalline substrate Al2O3 we
had found made it possible to save material and time costs when matching the sapphire
etching modes. Based on the research results, the etching of the sapphire substrate was
conducted using the ion-chemical method on the “INTRA” unit in the Freon -14 (СF4)
medium. Note that the (before-etching) residual pressure in the reaction chamber of the unit
was not higher than 1.5⋅10-2 Pa, the working pressure of the СF4 gas was at 2.4⋅10-1 Pa, the
automatic bias voltage across the specimen stage was U≈300 V, the electromagnet current
was I≈0.53 А, and the high-frequency useful power fed to the reaction chamber was РHF≈550
W at fHF=13.56 MHz.
Under the above conditions, the sapphire was etched through a double mask for about 30
min, with the substrate etch depth being slightly less than 0.3 µm. Then, the chromium mask
remains were removed in the cerium sulfate solution, thus producing the designed (binary)
microrelief on the sapphire substrate surface.
Using the above-described approach, a binary microstructure in the sapphire substrate
was fabricated, operating at λ=0.45 µm as a diffractive lens with the focal length of about 50
mm. The sapphire substrate was 3 mm thick and 26 mm in diameter.
The resulting microrelief was characterized using a scanning tunnel microscope Р4-SPM-
MDT. A fragment of the microrelief generated in the sapphire substrate is shown in Figure
12.
Figure 13 shows the profilogram of the binary microrelief obtained in the profilometer-
profilograph mode of the scanning microscope, whereas Figure 14 presents a fragment of the
profilogram of a binary lens microrelief synthesized for operation in the near IR range (λ=1
µm). The height of the fabricated microrelief of the near-IR diffractive lens was in excess of
0.6 µm, the period being 6.3 µm. The etch process duration was 65 min, the etch rate of the
fabricated substrate being about 10 nm/min.
Based on the studies conducted an approach has been offered and technological regimes
have been worked out for generating a DOE microrelief in the sapphire substrates. The
analysis of the generated binary microrelief suggests that sapphire is a promising material for
fabricating DOEs for various wavelength ranges, from ultraviolet to infrared.
Our further studies will be concerned with generating a multi-level diffractive optical
microrelief in the sapphire substrates.
Figure 12. Microrelief fragment generated in the sapphire substrate.
Figure 13. Profilogram of the microrelief in Figure 12.

Figure 14. Profilogram of the diffractive microrelief, generated on a scanning probe microscope (Zone
width: 6.29 µm. Etching angles: left walls: tan = 1.01; tan = 1.005. right walls: tan = 1.037; tan = 0.87.
Top step width: 2.36 and 2.19 µm. Bottom width: 2.77 µm).
3. REALISATION AND INVESTIGATION OF DIFFRACTIVE

MICRORELIEF ON THE END FACE OF SILVER-HALIDE WAVEGUIDE
A major drawback of the available optical systems to couple light in/out of fibers is that
alongside the waveguide’s end there are other refractive surfaces that result in radiation loss.
The use of immersion fluids, e.g. in reflectometers where the problem is especially pressing,
fails to solve it in full measure. This difficulty could be obviated by directly fabricating a
diffractive microrelief on the end face of the waveguide, in this way also making the structure
much more miniature. Besides, the diffractive microrelief on the waveguide end face allows
one to do without the alignment procedure, which is needed when, for example, the
diffractive microrelief is used for waveguide mode excitation [35,36]. There have been
publications devoted to fabricating a diffractive microrelief on the waveguide end face to
focus the output light into a designed region [37] and to control the output light phase [38].
However, the mentioned studies were conducted for fused silica fibers with use of
microlithography technology. In addition to being highly laborious, another drawback of this
approach is that such structures are difficult to replicate for mass production. For silver halide
(polycrystalline) IR fibers (PIR-fibers) that are widely used for spectrometry and laser
technology tasks [39–41], the microlithography is hardly applicable due to challenges related
to the chemical processing of materials. Besides, with each fiber processed individually, the
reproducibility of the result, especially of multi-level microreliefs, is very low.
It has been known that the extrusion technique is widely used just for fabricating silver
halide (polycrystalline) waveguide structures [42], suggesting that an appropriate method for
fabricating the diffractive microrelief on the polycrystalline fiber surfaces is by hot stamping.
The first results on the use of the hot stamping technique for fabricating a beam splitting
diffraction grating on the end face of the PIR-fiber were reported in [43].
As a test structure, a binary diffraction grating was chosen (Figure 15.).
Considering the optical density of the PIR-fiber core (n=2.15) and the wavelength of the
radiation channeled (λ=10.6 µm), the estimated microrelief height was h=λ/2(n-1)) = 4.6 µm.
The period d of the fabricated binary grating in the experiment was chosen so that to ensure
the spatial separation of diffraction orders. The grating microrelief was fabricated in two
stages: fabrication of a matrix using photolithography and transfer of the microrelief from the
matrix to the PIR-fiber end face by hot stamping. The matrix microrelief was generated using
a photolithographic process [44, 18]. The image reproduced in the photoresist layer was
transferred onto a glass substrate by liquid chemical etching in the hydrofluoric acid. The
solution concentration was fitted in such a manner as to ensure a sufficiently high rate of glass
etch while maintaining a maximal photoresist stability. During etching, the solution was
being stirred. The process was periodically controlled by an analyzer. At the final stage of the
matrix preparation, the photomask was eliminated with ethyl acetate and the substrate was
rinsed in distilled water.
Figure 15. Fabrication of a binary microrelief on waveguide end face.
To stamp the relief on the PIR-fiber end face, a setup was assembled consisting of a
microscope, a TV-camera, a three-axis table, and a heater with digital temperature control.
The stamping involved the alignment of the matrix and fiber’s end, followed by embossing
the relief in the fiber’s end face. The stamping pressure was chosen to be 0.3 kg/ mm2 for a
fiber of radius R=700 µm. The temperature of the heater was in the range 200 – 390º. Note
that due to a finite heat conductivity and the heat exchange with the environment, the
temperature in the stamping area somewhat decreased. The quality of the surface fabricated
on the fiber’s end face was checked using an automated interferometer Zygo.
To explore the beam splitting properties of the grating we conducted an optical
experiment in which a fiber with an embossed single-end microrelief was excited by coupling
the CO2 laser light (λ=10.6 µm) through the structure-free end face. The intensity distribution
of output light produced by the diffractive microrelief was registered by a photosensitive
paper. The result of the experiment is shown in Figure 16.
The result of the experiment has shown that the grating fabricated shows good beam
splitting properties. However, microrelief fabrication errors (in particular, the matrix material
wall angle, somewhat larger-than-designed etch depth of the matrix microrelief, and
stamping-related bending of the end surface) significantly increased the energy portion going
to the zero (axial) diffraction order (see Figure 16).
Further technology improvement, in particular, the optimization of the matrix etching
process and the choice of the stamping temperature regime has made it possible to enhance
the quality of the microrelief fabricated. Note that fabricating more complex diffractive
optical microreliefs (like Fresnel lenses, mode filters, diffractive optical elements (DOEs) to
focus fiber radiation into designed regions) directly on the PIR-fiber end face requires the
solution of the problem of alignment of the matrix center and the fiber optical axis.
To solve this problem, an optical-kinematical setup schematically shown in Figure 17
was assembled.
Figure 16. The result of interaction between the fiber light and the photosensitive paper.
Figure 17. Optical-kinematical setup of the stamping device.
Step motor 1, which is computer-controlled through a matching device, is connected with

precision adjustment screw 2, which is rigidly fixed on a base. The displacement controlled
by indicator 3 is transferred to warm stage 7 with the matrix, through adaptor unit 6 making it
possible to observe the operating area on microscope 4 and a TV-camera. The fiber under
processing 8 can be precisely positioned with two-axis table 9 to provide for the possibility of
fabricating axially symmetric structures. The heater placed on the warm stage is controlled by
unit 5 with digital temperature indication. A kinematical peculiarity of the setup is that the
micrometric screw’s precision nut is fastened rigidly, while the step motor is sliding, thus
minimizing system’s mechanical play.
The unit has the following parameters:
• Micrometric screw’s step- 500 µm.

• Matrix holder’s (warm stage) displacement- 2.5 µm/pulse.
• Maximal matrix displacement- 20 mm.
• Clock pulse frequency range- 1 -200 Hz.
• Displacement force on the matrix holder- >5 kg
• Fiber displacement (XY)- +- 10 mm.
• Matrix holder’s temperature- 190…450º C.
The displacement direction, clock-pulse frequency, and the magnitude of matrix holder’s
displacement can be specified by programming means. The zero reading (the instant of time
at which the matrix gets in contact with fiber) was identified visually on the microscope. To
achieve the best stamping results, the heater’s temperature was chosen in the range 200º C –
270ºC.
Shown in Figure 18 is the resulting microrelief of Fresnel lens fabricated on the PIR-fiber
end face using the installation designed.
Figure 18. Resulting Fresnel lens microrelief on the fiber end face.
Regarding the application of silver halide PIR-fibers, of big concern are large Fresnel
reflection losses resulting from a high refractive index of the material (n=2.15). In this work,
the developed technique and installation for the hot stamping of silver halide PIR-fiber
surfaces were used for applying antireflection structures. Matrices for applying the
antireflection coatings were fabricated in fused silica substrates by plasmo-chemical etching
[18].
The matrix profile obtained using a scanning probe microscope (SPM) is shown in Figure
19.
Figure 19. The SPM-profile of the stamping matrix.
Figure 20. (1) – transmission spectra of a fiber with two-sided antireflection microrelief;. (2) -
transmission spectra of a fiber without microrelief , (3) – T% - ratio (1) to (2), (4) – spectral region of
antireflection effect.
The microrelief was explored using an SPM Solver-Pro manufactured by the firm NT-
MDT (the city of Zelenograd, Russia).
The efficiency of the fabricated antireflection structures was studied experimentally by
comparing the spectral transmission of a fiber with antireflection coatings on both ends with
that of an identical fiber without antireflection coatings. The result of the experiment is shown
in Figure 20.
The analysis of the experimental results (Figure 20) suggests that using the above-
described approach to reducing Fresnel-reflection losses the transmission of a PIR-fiber
sample can be increased by nearly 18% (maximum) in the waveband of 4.5 – 18 µm.
However, to obtain the antireflection effect for shorter wavelengths (< 4.5 µm) the technology
of lithographical fabrication of stamping matrix microrelief needs to be improved.
Undoubtedly, this result can be further improved by optimizing the technology of the
diffractive microrelief fabrication on the matrix, which is subsequently transferred to the
fiber. Further development of the approach discussed will make it possible to produce more
sophisticated subwavelength structures, e.g. photonic crystal devices for medium IR range.
Available methods for modifying the refractive index of silver halides using ion
bombardment in combination with the method for microrelief fabrication on PIR-fiber
surfaces presented here offer wide prospects for developing integrated-optical devices for IR
range.
4. OPTIMIZATION OF DIFFRACTIVE MICRORELIEF

4.1. Different Approaches for Doe Optimization
With enhanced performance of computing facilities the iterative design of phase

diffractive elements (DOEs) [8-10] has become widely accepted. A great number of up-to-
date technologies for DOE fabrication make use of the approximation of the commonly
continuous DOE phase function by a piecewise continuous (quantized) function [18]. This is
the reason why constructing iterative procedures for the design of quantized DOE (DOEs
with quantized phase function) has become topical [8,9].
The iterative process of designing DOEs can be realized through either the consecutive
construction of generalized mathematical projections onto closed sets corresponding to the
limitations imposed on the complex amplitude in the focal and DOE planes or gradient
procedure. In [45] it was shown that iterative procedure of phase restoration based on
generalized projections onto closed sets (IFTA-algorithm [8] or Fienup algorithm [10]) can
not guarantee convergence to global optimum. Gradient procedure also can not guarantee
convergence to global optimum [9].
Besides, designing quantized DOEs using Fienup-type iterative algorithms is hampered
by the necessity to solve the inverse task at every iteration [8,9]. The use of stochastic
procedures [46-48] does not make it necessary to solve the inverse task. Thus, the DOE phase
function may be sought for directly over a set of “technologically implemented” functions (in
our case – quantized phase functions), allowing the quantization errors to be avoided [48].
However, constructing a stochastic optimization procedure for a real DOE calls for solving a
great number of direct tasks, which in general may result in an unpractical computational
effort. It seems worthwhile to consider how the stochastic DOE phase optimization
procedures can be used when solving the direct problem does not require great computational
effort (e.g. for a radially symmetric DOE) [48].
4.2. Optimization of the Radially Symmetric Doe Phase Using a Genetic

Algorithm
The genetic algorithm (GA) is an iterative stochastic algorithm operating a set of

“individuals” (the population). Every individual represents a potential solution to the
problem. Every individual represents a potential solution to the problem [46]. By way of
illustration, when designing a radially symmetric DOE as a separate individual, we can take
such a DOE phase function, ϕ(ρ), that the corresponding microrelief
h(ρ ) = λ ϕ(ρ ) 2π(n − 1) , (1)
where n is the refractive index of the substrate and λ is the wavelength of light, can be
implemented using available technological facilities. With the available technological
facilities for fabricating binary (two-level) DOEs, it is expedient to consider as the individuals
the phase functions taking at each point one of the following values: 0 or π (or π and 2π). In
the general case of designing a DOE with M quantization levels, the phase function takes the
values
ϕ(ρ ) ∈ {(2πm / M ) : m ∈ [1, M ]} . (2)
Note, that the use of “deterministic” methods [8,9] in designing quantized DOEs with
small M (M=2-8) requires additional complicated modifications of the design procedure. On
the contrary, for stochastic optimization the decreased number of quantization levels means a
decrease in the power of the set of admissible solutions and, hence, decreased computational
effort. Prior to starting the algorithm, the population is generated at random. Then, using
some “fitness” (cost) functional Ф(ϕ), every individual ϕ is assigned an individual measure of
value with respect to the problem in question. The value of Ф(ϕ) is used by the algorithm to
direct the search. In our case, the particular form of the functional Ф(ϕ) is determined by the
DOE prospective application. At every iteration, the population is increased artificially by
means of a definitely introduced procedure of “cross-breeding” of initial population
individuals. The best individuals selected (in terms of the introduced fitness functional Ф(ϕ))
are used as the initial population for the next algorithm iteration. It has been shown in before
that if the previous-iteration best individual is permanently preserved and there are
“mutations” (which are stochastic variations of pixels of the function ϕ(ρ) with some
predeterminated probability p) such a procedure converges to a global optimum. The
particular form of the functional Ф(ϕ) is determinated by the DOE application and its
calculation does not necessarily require solving the direct task of diffraction theory. By way
of illustration, if we need to design a DOE generating a beam with desired mode composition,
the calculation of the functional Ф(ϕ) will require the decomposition of the light field in terms
of the modes in the plane immediately behind the optical element.
Let us consider in more detail constructing the genetic optimization procedure for a
radially symmetric quantized DOE of radius R. The initial population is a set Ngen of
N −1
stochastically generated 1-D arrays ⎧⎨ϕ ⎫⎬ .
j
⎩ ⎭ j =0
Every array corresponds to a single realization of the DOE phase, ϕ(ρ), taken along the
radius: ϕ j = ϕ( jΔ ) ∈ {2πm / M } where Δ = R / (N − 1) and R is the DOE radius. Upon
generating the initial population (“parent” population) a loop is organized, with the following
operations being consecutively performed within it:
*crossbreeding: generating Nchild additional new individuals-arrays, with the pixel value
being chosen for each of them as the value of the corresponding pixel of the
stochastically chosen “parent”-array.
*mutation: a stochastic change in the value of randomly chosen pixels of the resulting new
arrays. The value of mutation probability, p, determines the rate at which the procedure
will be converging.
*selection: identifying the next “parent” population via selecting Ngen best (in terms of the
problem to be solved) individuals using the computational functional Ф(ϕ) for all
Ngen+Nchild individuals.
Thus, at each iteration we must find (Ngen+Nchild) values of the functional Ф(ϕ). The
number of iterations, Nit, it takes to optimize the DOE phase was chosen on the basis of the
number of iterations it takes to optimize a test functional, for which the value of the global
optimum is known. To illustrate the capabilities of the proposed approach we designed a
DOE to focus the Gaussian beam into a circle (flat-top). This type of DOE has a wide range
of potential applications. While choosing the DOE phase optimization functional, the
following requirements should be taken into account:
− the functional should be adequate to the physical problem;

− the required computational efforts should be minimized.
Let us consider the synthesis of a DOE of radius R which we intend to generate from the
plane Gaussian beam
I0 (ρ) = exp⎛⎜ − 2ρ2 σ2 ⎞⎟ (3)

⎝ ⎠
the focal intensity distribution (Figure 21) of which is given by
⎧I , 0 ≤ r ≤ R1
I (r ) = ⎪⎨ 1 (4)
⎪⎩ 0, r > R1
The intensity I1 may be derived from the energy conservation law:
R
2
I1 = C M ∫ I 0 (ρ)ρdρ (5)
R12
where СМ is a constant describing energy losses related with the phase quantization in M
levels (for M=4 СМ=0.81). The DOE is designed as an addition to the external spherical lens.
Figure 21. Synthesizing a DOE to focus the beam into a flat-top.
As the minimizing functional Ф(ϕ), it is expedient to choose
N −1
l
Φ (ϕ ( ρ ) ) = ∑ ( I ( rl ) − I1 )
2
, (6)
l =0
where N1 = R1 Δ1 is the radial size of the focal domain in diffraction spots, Δ1 = λf 2R

is the diffraction spot size, rl = (l + 0,5)Δl and I (r , ϕ(ρ)) is the focal intensity distribution
generated by the DOE with phase function ϕ(ρ).
The light field complex amplitude, w(r), in the lens focal plane may be expressed as a
sum of amplitudes resulting from the diffraction of a plane wave by a ring-like aperture
( ) ( )×
N −2
w(r ) = ∑ exp − ρ 2 / σ 2 exp iϕ
j +1 j +1
j =0
( )
× ⎡⎢ ρ j +1 J1 kr ρ j +1 f / r − ρ j J1 kr ρ j / f / r ⎤⎥
⎣ ⎦( ) (7)
where, ρ j = jΔ , J1 (* ) is the Bessel function of the first order. Hence, the intensities at
points rl may be derived as follows:
N −2
I (rl ) = ∑ exp⎛⎜ − ρ 2j +1 /σ 2 ⎞⎟exp iϕ j +1 ×
⎝ ⎠
( )
j =0
2
× ⎡⎢ρ j +1J1⎛⎜ krlρ j+1 f ⎞⎟ / rl − ρ jJ1⎛⎜ krlρ j / f ⎞⎟ / rl ⎤⎥ , (8)
⎣ ⎝ ⎠ ⎝ ⎠ ⎦
or
( )C
N −2
( )
I rl = ∑ exp iϕ
j +1
l
j
, (9)
j =0
where
⎛ ⎞
C lj = exp⎛⎜ − ρ 2 σ 2 ⎞⎟⎜ ρ J ⎛⎜ kr ρ f ⎞⎟ / r − ρ J ⎛⎜ kr ρ / f ⎞⎟ / r ⎟ . (10)
⎝ j + 1 ⎠⎝ j + 1 1 ⎝ l j + 1 ⎠ l j 1 ⎝ l j ⎠ ⎠ l
Note that the coefficients C lj are independent of ϕ(ρ) and can be calculated prior to
starting the optimization procedure. Once the optimization procedure is completed, the energy
efficiency is calculated by the formula
R R
l
e= ∫ I ( r )rdr ∫ I 0 (ρ)ρdρ (11)
0 0
and the root-mean-square deviation of the intensity from the average value I is derived from
R
1 2 l
δ= 2
I Rl ∫
2
[
I (r ) − I rdr , ] (12)
0
R
2 l
I = 2 ∫ I (r )rdr .
Rl
0
In the numerical simulation the following parameters were used: the illuminating beam
wavelength is λ=10.6 μm, the illuminating beam radius is σ=1.85 mm, the external lens focal
distance is f=100 mm, the aperture radius is R=3.7 mm, the focal radius is R1=0.6 mm, the
number of the phase function pixels is N=100, the number of quantization levels is M=4, and
the number of iterations of the stochastic function is Nit=6000. Figure 22 shows the calculated
intensity in the DOE focal plane. From the numerical simulation, the calculated energy
efficiency amounted to e=80.3%, the r.m.s. deviation of the intensity being δ=6%.
Figure 22. Radial intensity distribution generated by the designed DOE in the focal plane (the
numerical simulation result).
Thus, the use of the stochastic iterative procedure with functional (6) enables us to design
four-level DOEs to focus the Gaussian beam into a circle with energy efficiency close to the
theoretical limit value of 81% and a low r.m.s. deviation. Also note the modest computational
efforts required for calculating the element, because once the coefficients in (9) and the
intensities in (6) have been preliminary calculated, the calculation of the functional becomes
rather simple.
4.3. Results of the Experimental Investigation (4-Level Doe)
The testing DOE designed [49] using 1000000 iterations of the procedure [48] as a
complement to a spherical lens has following parameters: the focus of the external lens f =
300 mm, the wavelength of the illuminating beam λ = 543 nm, the aperture radius R=1.75
mm, the focal circle radius R1=0.6 mm, the radius of the illuminating Gaussian beam o=0.525
mm, the number of pixels of the phase function along the radius N=50, the number of
quantization levels is 4. The DOE has a radially symmetrical phase structure with 14 zones.
Figure 23 depicts an estimated relief (radial section) and the result of the profilometry of the
implemented microrelief.
The He-Ne laser, operating at the 543-nm wavelength with a divergence M2 ≅ 1.2, was
used as a radiation source. Figure 24 shows cross-sections of the generated beam, obtained
with the aid of a CDD-camera in the lens output plane and in the plane, located at 20 mm
from the lens output plane.
To compare, Figure 24 depicts a cross-section of the estimated intensity distribution in a
lens focus for a beam with a limited diffraction (M2=1). The presence of higher-order modes
in the illuminating beam provides an explanation for the larger size of the generated focal
circle as compared to the estimated one.
Figure 25 represents 2D intensity distributions, formed by the DOE and measured in the
absence of the optical element. Note that for a 4-level element, implemented for a given set of
physical parameters, a plane with a lesser non-uniformity of intensity distribution is located

behind the lens output plane.
Figure 23. Estimated profile of the DOE (daggers) as compared to the implemented one (firm line) after
the measurements taken.
Figure 24. Comparison between the estimated intensity distribution in the lens output plane (dots), the
measured intensity distribution in the lens focus (firm line), and the measured intensity distribution in
the plane, located at 20 mm behind the lens output plane.
A numerical integration with regard for the results of measurements, taken in the lens
output plane with the aid of a CDD-camera, produced the values of energy efficiency of 77.6
and 79.6 percent for the values of focal circle radii R1=0.6 mm and R1=0.66 mm,
respectively, what is in good agreement with the values, obtained in the course of
computational experiment (81.3 and 81.6 percent, respectively). In the real experiment, a ratio
of the energy, focused by the DOE into a focal circle, to the total energy amount of the light
beam, having passed through the plate with an applied DOE microrelief, was designated as
energy efficiency.
Thus, the measured efficiency of the 4-level element was found to be about 97 percent of
theoretical limit at the high uniformity of the generated intensity distribution.
a)
Figure 25. Continues on next page.
b)
c)
Figure 25. Intensity distribution measured in the lens output plane: (a) in the absence of the DOE; (b) in
the presence of the DOE; and (c) in the plane, located at the distance of 20 mm behind the lens output
plane.
CONCLUSION
CVD diamond diffractive optical elements (Fresnel spherical lense, beam shaper
(focuser)) have been fabricated by laser structuring of the diamond surface. The diffraction
efficiency of the CVD diamond Fresnel lens was found to be 86%. The correlation between
the experimental characteristics of the elements and the results of numerical simulation
suggests that the laser processing provides adequate accuracy for the formation of DOE
microrelief. Thus, our investigations have shown that a laser ablation technique is a promising
method for manufacturing of low loss multilevel diamond IR diffractive elements. The
potential use of dry etching of polycrystalline diamond films for fabricating high-power
DOEs intended to control the CO2-laser light was analyzed also.
Based on the studies conducted an approach has been offered and technological regimes
have been worked out for generating a DOE microrelief in the sapphire substrates. The
analysis of the generated microrelief suggests that sapphire is a promising material for
fabricating DOEs for various wavelength ranges, from ultraviolet to infrared.
The present chapter includes the results of experimental investigation of diffractive
microrelief and antireflecting structures that are realized on an end-face of IR-fiber.
Available methods for modifying the refractive index of silver halides using ion
bombardment in combination with the proposed method for microrelief fabrication on PIR-
fiber surfaces presented here offer wide prospects for developing integrated-optical devices
for IR range.
Different ways, from geometrical optics approximation to stochastic optimization, can be
used for calculation of a diffractive microrelief of IR-range optical elements that focus an
illuminating beam into a pre-given 2-dimensional area. The use of stochastic procedures does
not make it necessary to solve the inverse task. It is shown that, the DOE phase function may
be sought for directly over a set of “technologically implemented” functions, allowing the
quantization errors to be avoided.
REFERENCES
[1] Golub, M.A.; Karpeev, S.V.; Prokhorov, A.M.; Sisakyan, I.N.; Soifer,V.A. Soviet
Technical Physics Lett. 1981, 7, 264-266.
[2] Danilov, V.A.; Popov, V.V.; Prokhorov, A.M.; Sagatelyan, D.M,; I.N.; Soifer,V.A.
Soviet Technical Physics Lett. 1982, 8, 351-353.
[3] Prokhorov A.M., Sisakyan I.N., Golub M.A., Soifer V.A., Karpeev S.V., Goncharsky
A.V. Optical phase element for focussing monochromatic radiation. Patent of Great
Britain, No.2185126 prior.30.04.1985, Printing 8.07.1987, patent publ.24.05.1989.
[4] Prokhorov A.M., Sisakyan I.N., Golub M.A., Soifer V.A., Karpeev S.V., Goncharsky
A.V. Procede de focalisation d'un rayonnement monochromatique et element optique de
dephasage mettant en oeuvre ledit procrde. Patent of France, No. Publication 2585854,
No. Enregistrement national 8511759 Date de delivrance 07.12.1987, Bopi delivrance
No.51 du 18.12.1987, Classt 4 G02F 1/19/.
[5] Danilov V.A., Popov V.V., Prokhorov A.M., Sisakian I.N., Sagatelian D.M., Soifer
V.A., Sisakian E.V., Naumidi L.P., Danileiko J.K., Terekhin J.D., Akopian V.S.,
Murzin S.P., Shorin V.P., Mordasov V.I. Device for laser treatment of an object. United
States Patent No US005103073A Int. Cl. B23K 26/00. Date of Patent: Apr. 7, 1992.
[6] Duparre’, M.; Golub, M.A.; Ludge, B.; Pavelyev, V.S.; Soifer, V.A.; Uspleniev, G.V.;
Volotovskii, S.G. Applied Opt. 1995, 34(14), 2489-2497.
[7] Golub, M.A.; Sisakyan, I.N.; Soifer, V.A. Optics and Lasers in Engineering 1991, 15,
297-309.
[8] Wyrowski, F. Journ. Opt. Soc. Amer. 1990, 7, 961-963.
[9] Soifer, V.A.; Kotlyar, V.V.; Doscolovich, L.L. Iterative Methods for Diffractive Optical
Elements Computation; Taylor and Francis Ltd.: London, 1997.
[10] Fienup, J.R. Optical Engineering 1980, 19, 297-303.
[11] Harris, D.C. Infrared Window and Dome Materials; SPIE Tutorial Text Series; 1992;
Vol. TT10.
[12] Brasunas, J.C.; Cushman, G.M.; Lakew, B. Applied Optics 1998, 37(19), 4226- 4229.
[13] Kononenko, T.V.; Kononenko, V.V.; Konov, V.I.; Pimenov, S.M.; Garnov, S.V.;
Tischenko, A.V.; Prokhorov, A.M.; Khomich, A.V. Appl. Phys. A. 1999, 68(1), 99-102.
[14] Kononenko, V.V.; Kononenko, T. V.; Konov, V. I.; Pimenov, S. M.; Garnov, S. V.;
Tishchenko, A. V.; Prokhorov, A. M.; Khomich, A. V. Quantum Electronics 1999,
26(2), 158 -162.
[15] Kononenko, V.V.; Konov, V.I.; Pimenov, S.M.; Prokhorov, A.M.; Pavelyev, V.S.;
Soifer, V. A. Quantum Electronics 1999, 26(1), 9-10.
[16] Kononenko, V.V.; Konov, V.I.; Pimenov, S.M.; Prokhorov, A.M.; Pavelyev, V.S.;
Soifer, V.A. New Diamond and Frontier Carbon Technology 2000, 10, 97-107.
[17] Pavelyev, V.S.; Soifer, V.A.; Golovashkin, D.L.; Kononenko, V.V.; Konov, V.I.;
Pimenov, S.M.; Duparre, M.; Luedge, B. Proc. SPIE 2004, 5456, 209-219.
[18] Golovashkin, D.L.; Kazanskiy, N.L.; Soifer, V.A.; Pavelyev, V.S.; Solovyev, V.S.;
Usplenyev, G.V.; Volkov, A.V. In Methods for Computer Design of Diffractive Optical
Elements; Soifer, V.A.; Ed.; John Wiley and Sons, Inc.: New-York, US, 2002; 267-345.
[19] Volkov, A.V.; Kazansky, N.L.; Kostyuk, G.F.; Pavelyev, V.S. Optical Memory And
Neural Networks 2002, 11(2), 135-138.
[20] Pavelyev, V.S.; Borodin, S.A.; Kazanskiy, N.L.; Kostyuk, G.F.; Volkov, A.V. Optics
and Laser Technology 2007, 39 (6), 1234-1238.
[21] Ralchenko, V.G.; Smolin, A.A.; Konov, V.I.; Sergeichev, K.F.; Sychov, I.A.; Vlasov,
I.I.; Migulin, V.V.; Voronina, S.V.; Khomich, A.V. Diamond Relat. Mater. 1997, 6,
417-421.
[22] Pimenov, S.M.; Smolin, A.A.; Ralchenko, V.G.; Konov, V.I.; Sokolina, G.A.;
Bantsekov, S.V.; Spitsyn, B.V. Diamond and Relat. Mater. 1993, 2, 291-297.
[23] Pimenov, S.M.; Kononenko, V.V.; Ralchenko, V.G.; Konov, V.I.; Gloor, S.; Lüthy, W.;
Weber, H.P.; Khomich, A.V. Applied Physics A 1999, 69 (1), 81-88.
[24] Born, M.; Wolf, E. Principles of Optics; Pergamon Press: Oxford, UK, 1965.
[25] Golub, M.A.; Rybakov, O.E.; Usplenjev, G.V.; Volkov, A.V.; Volotovsky, S.G. Optics
and Laser Technology 1995, 25(4), 215-218.
[26] Tachi, S.; Okudaria, S. J. Vac. Technol. В. 1986, 49(5), 459-487.
[27] Oehrlein, G.S.; Braight, A.A.; Robey, S.W. J. Vac. Sci. Technol. A. 1988, 6(3), 1989-
1993.
[28] Einsprach, N.; Brown, D. Plasma Processing for VLSI, Academic Press, Inc., 1984.
[29] Kazanskiy, N.L.; Uspleniev, G.V.; Volkov, A.V. Proc SPIE 2000, 4316, 193-199.
[30] Volkov, A.V.; Doskolovich, L.L.; Kazanskiy, N.L.; Uspleniev, G.V.; Zanelli, A.
Computer Optics 2000, 20, 84-89.
[31] In Large Soviet Encyclopedia; Prokhorov, A. M.; Ed.; Sovetskaya Entziklopediya
Publishers: Moscow, SU, 1991; Vol. 2, pp. 654.
[32] Batygina, E.I.; Batygin, V.N.; Bravinsky, V.G.; Kotyurgina, O.A.; Kulikov, V.I.
Reviews in Electronic Engineering; Series 1. Microwave Electronics; Electronika
Publishers: Moscow, SU, 1980; Vol. 4(703).
[33] In Handbook of Electrotechnology Materials; Koritzky, Yu.V.; Pasynkov, V.V.; Tokarev,

B.M.; Ed.; Energiya Publishers: Leningrad, SU, 1976; Vol. 3, pp. 748, 772-773.
[34] Volkov, A.V.; Istinova, O.G.; Kazanskiy, N.L.; Kostyuk, G.F. Computer Optics (in
Russian) 2002, 24, 70-73.
[35] Karpeev, S.V.; Pavelyev, V.S.; Duparre, M.; Luedge, B.; Rockstuhl, C.; Schroeter, S.
Optical Memory And Neural Networks 2003, 12(1), 27-3.
[36] Karpeev S.V., Pavelyev V.S., Khonina S.N., Kazanskiy N.L., Gavrilov A.V., Eropolov
V.A. Journal of Modern Optics 2007, 54(6), 833-844.
[37] Prasciolu, M.; Cojoc, D.; Cabrini, S.; Businaro, L.; Candeloro, P.; Tormen, M.; Kumar,
R.; Liberale, C.; Degiorgio, V.; Gerardino, A.; Gigli, G.; Pisignano, D.; Di Fabrizio, E.;
Cingolani, R. Microelectronic Engineering 2003, 67-68, 169-174.
[38] Johnson, E.G.; Stack, J.; Suleski, T.J.; Koehler, C.; Delaney, W. Applied Optics 2003,
42(5).
[39] Butavina, L. In Infrared Fiber Optics; Sanghera, J.; Aggarwal, I.; Ed.; CRC press: Boca
Raton, (UK), 1998; 209-249.
[40] Petrich, W. Appl. Spectrosc. Rev. 2001, 36, 181-237.
[41] Wilson, R.H.; Tapp, H.S. Trends Anal. Chem. 1999, 18, 85-93.
[42] Artjushenko, V.G.; Bochkarev, E.P.; Golovanov, P.F. Soviet Journal of Quantum
Electronics 1981, 8, 398-400.
[43] Pavelyev, V.S.; Moiseev, O.Y.; Volkov, A.V.; Eropolov, V.A.; Dmitriev, S.V.;
Karpeev, S. V.; Artyushenko, V. G.; Kashin, V.V. Proc SPIE 2008, 6994, 69940Q.
[44] Levinson, H.J.; Principles of Lithography; SPIE Press: Washington (US), 2004.
[45] Levi, A.; Stark, H. J. Opt. Soc. Am. 1984, 1(2), 932-943.
[46] Lu, C.Y.; Liao, H.Z.; Lee, C.K.; Wang, J.S. Applied Optics 1997, 36(20), 4702-4712.
[47] Znou, G.; Yuan, X.; Dowd, P.; Lam, Y.-L.; Chan, Y.-C. Optics Letters 2000, 25(17),
1288-1290.
[48] Pavelyev, V. S. Proc of SPIE 2005, 5962, 596230.
[49] Pavelyev, V.S.; Soifer, V.A.; Glaeser, T.; Schroeter, S.; Puehlmann, R.; Bartelt, H.
Computer Optics (in Russian) 2002, 22, 53-55.
Chapter 5
MICRO- AND NANOSCALE HEAT TRANSFER IN

FEMTOSECOND LASER PROCESSING OF METALS
Yuwen Zhang, D. Y. Tzou and J. K. Chen

Department of Mechanical and Aerospace Engineering
University of Missouri, Columbia, MO 65211, USA
ABSTRACT
Ultrafast laser material processing has received significant attention due to a growing
need for the fabrication of miniaturized devices at micro- and nanoscales. The traditional
phenomenological laws, such as Fourier’s law of heat conduction, are challenged in the
microscale regime and a hyperbolic or dual phase lag model should be employed.
equilibrium. Various two-temperature models that can be used to describe the
nonequilibrium heat transfer are presented. A semi-classical two-step heating model to

investigate thermal transport in metals caused by ultrashort laser heating is also
presented.
The main difference between the semiclassical and the phenomenological two-
temperature models is that the former includes the effects of electron drifting, which
could result in significantly different electron and lattice temperature response from the
latter for higher-intensity and shorter-pulse laser heating.
melting point and melting takes place. The liquid phase will be resolidified when the
lattice is cooled by conducting heat away. Ultrafast melting and resolidification of the
thin gold film and microparticles were investigated. At even shorter pulse width,
femtosecond laser heating on metals produces a blasting force from hot electrons in the
sub-picosecond domain, which exerts on the metal lattices along with the non-
equilibrium heat flow.
Our work that employs the parabolic two-step heating model to study the effect of
the hot-electron blast in multi-layered thin metal films is also presented.
160 Yuwen Zhang, D. Y. Tzou and J. K. Chen
NOMENCLATURE
Ae materials constants for electron relaxation time [1/K2-s]
Bl materials constants for electron relaxation time [1/K-s]
Be Coefficient for electron heat capacity [J/m3-K2]
c thermal propagation speed [m/s]
C heat capacity, [J/m3-K]
cp specific heat, [J/kg-K]
E Young’s modulus [Pa]
f distribution function of the free electrons
Liquid Fraction in Control Volume
G electron-lattice coupling factor [W/m3-K]]
H nondimensional electron-phonon coupling factor
Hl enthalpy of lattice per unit volume [J/m3]
hm latent heat of fusion [J/kg]
I unit tensor
J heat source fluence (J/m2)
k thermal conductivity [W/m-K]
kB Boltzmann constant [J/K]
l thickness [m]
L dimensionless thickness
M molar mass [kg/kmol]
N number density of atom [m-3]
P pressure [Pa]
q′′ heat flux vector [W/m2]
R
reflectivity
Rg gas constant [J/kg-K]
s solid-liquid interfacial location [m]
S intensity of the internal heat source [W/m3]
t time [ s ]
tp FWHM (Full Width at Half Maximum) pulse width [s]
T temperature [K]
Tm melting point [K]
Tl,I lattice temperature at solid-liquid interface [K]
us solid-liquid interfacial velocity [m/s]
u electron velocity vector [m/s]
displacement vector of metal lattice [m]
U dimensionless displacement
v mean (drift) velocity vector [m/s]
V0 interfacial velocity factor [m/s]
x coordinate [m]
Micro- and Nanoscale Heat Transfer in Femtosecond Laser Processing of Metals 161
Greek Symbols
α thermal diffusivity [m2/s]

β dimensionless time
δ optical penetration depth [m]
δb ballistic range [m]
η thermomechanical coupling factor
κε thermal expansion coefficient of strain
λ Lamé constant
μ Lamé constant
ν Poisson’s ratio
θ dimensionless temperature
ρ density [kg/m3]
τ relaxation time [s]
τq phase lag for the heat flux vector [s]
τT phase lag for the temperature gradient [s]
ξ electron energy density [J]
dimensionless coordinate
Subscripts
e electron
F Fermi
i Initial
l lattice
l liquid
s solid
1. INTRODUCTION
Micro- and nanoscale heat transfer has become one of the very active areas drawing
considerable attention from many researchers due to its importance in many areas of
advanced manufacturing and materials processing. The traditional phenomenological laws,
such as Fourier’s law of heat conduction, are challenged in the microscale regime because (1)
the characteristic lengths of the various heat carriers are comparable to each other and to the
characteristic length of the system considered, and/or (2) the characteristic times of the
various heat carriers are comparable to the characteristic energy excitation time (Wang and
Prasad, 2000). Thus, microscale heat transfer can be referred to as heat transfer occurring on
both the micro-length and micro-time scales. Microscale heat transfer finds applications in
thin film (micro- length scale) as well as ultra-short pulsed laser processing (micro- time
scale). Chirped-pulse amplification was first accomplished over two decades ago (Strickland
and Mourou, 1985). Development of femtosecond lasers followed, then accelerated over the
past decade, because feasibility studies on metals (Momma et al., 1996; Hopkins and Sibbett,
2000), high bandgap materials (Fan et al., 2002), and biological materials (Riemann et al.,
2004) revealed high precision and high-quality surfaces. Among demonstrated potentials was
the capability of femtosecond lasers to produce “clean cuts,” namely the removal of material
with minimal collateral damage. This may be the most striking feature that has not been
matched by any other material processing means. Laser intensity/fluence must be increased
simultaneously, to assure sufficient processing power, as the pulse duration (heating time) for
each pulse shrinks into the femtosecond domain (Yamanouchi et al., 2007). To date, with the
processing power carried by each pulse entering pettawatts (1015 W) in less than 100
femtoseconds, the femtosecond lasers remove virtually any type of material in a mere tenth of
a picosecond while confining the process zone to within ten nanometers. This extremely short
heating time is a key element that limits undesirable spread of the heat-affected zone, assuring
minimal collateral damage on neighboring materials when clean cuts are made with
femtosecond lasers.
When the laser pulse is reduced to a nanosecond (10-9 sec) or less, the heat flux of the
laser beam can be as high as 1012 W/m2. For femtosecond pulse lasers, the laser intensity can
even be up to 1021 W/m2. Compared to long-pulsed laser processing, short-pulsed laser
processing enables users to precisely control the size of the heat-affected zone, the heat rate,
and the interfacial velocity. Griffith et al. (2003) investigated femtosecond laser machining of
steel and developed the femtosecond laser micro-fabrication capability in the microscale
regime. Klein-Wiele et al. (2003) introduced a new technology of laser processing via
ablation for nano-fabrication of solid materials with femtosecond laser. Heat affected zone of
metals ablated with femtosecond laser was investigated by Hirayama and Obara (2003).
During laser-metal interaction, the laser energy is first deposited into electrons on the
metal surface, where two competing processes occur (Hohlfeld et al., 2000). One is ballistic
motion of the excited electrons into deeper parts of the metal with velocity close to the Fermi
velocity (~106 m/s). Another process is collision between the excited electrons and electrons
around the Fermi level – an electron temperature is defined upon establishment of equilibrium
among hot electrons. These hot electrons are then diffused into deeper part of the electron gas
at a speed (<104 m/s) much lower than that of the ballistic motion. Meanwhile, the hot
electrons are cooled by transferring their energy to the lattice through electron-phonon
coupling. If the laser pulse width is shorter than the time required for the electron and lattice
to achieve thermal equilibrium (thermalization time), the electrons and lattices can no longer
be treated as being in thermal equilibrium (Grigoropoulos and Ye, 2000). The energy
equations for the electrons and lattice must be specified separately and coupled through a
coupling factor. The nonequilibrium between electrons and lattice has been observed
experimentally (Eesley, 1986; Elsayed-Ali et al., 1987) and can be described by the two-
temperature model, which was originally proposed by Anisimov et al. (1974) and rigorously
derived by Qiu and Tien (1993) from the Boltzmann transport equation. The nonequilibrium
electron and lattice temperature model can also be derived using the dual-phase-lag model by
considering lagging behavior of different energy carrier (Tzou, 1997; 2006). Chen et al.
(2006) proposed a semi-classical two-step heating model to investigate thermal transport in
metals caused by ultrashort laser heating.
Most existing two-temperature models dealt with the case that lattice temperature is well
below the melting point and only pure conduction is considered. Under higher laser fluence
and/or short pulse, the lattice temperature can exceed the melting point and melting takes
place. The liquid phase will be resolidified when the lattice is cooled by conducting heat
away. Short-pulsed laser melting of thin film involves the following three steps (Kuo and
Qiu, 1996): (1) absorption of photon energy by free electrons, (2) energy transfer between the
free electrons and the lattice, and (3) phase change of the lattice due to the propagation of
energy. The rapid phase change phenomena induced by ultrashort pulse laser are controlled
by nucleation dynamics at the interface, not by interfacial energy balance (Von Der Linde, et
al., 1987). The solid-liquid interface can be heated well above the melting point during a rapid
melting process, in which case the solid becomes superheated. Similarly, the solid-liquid
interface can be cooled far below the melting point in the rapid solidification process, in
which case the liquid becomes undercooled. Both superheated solid and undercooled liquid
are thermodynamically metastable states. Once phase change is triggered in a superheated
solid or undercooled liquid, the solid-liquid interface can move at an extremely high velocity
(on the order of 10 to 103 m/s). At even shorter pulse width, ultrafast laser heating on metals,
from a mechanical point of view, disturbs the equilibrium energy state between electrons and
phonons. When such a disturbance is active in the femtosecond domain, the lattices remain
almost thermally undisturbed the gradient of the excessive disturbed energy over the
equilibrium energy produces a net force exerting on the cold metal lattices. Magnitude of
such a force increases with the power intensity of the laser, which can be sufficiently strong
to destroy the metal lattices before they become noticeably heated (Falkovsky and
Mishchenko, 1999). Severe deformation and the subsequent failure of the lattices, therefore,
occur near the equilibrium temperature that can never be realized by conventional
thermoelasticity based on lattice expansion/contraction developed in a non-uniform
temperature flied. Such a hot-electron blast depends on both temperature and temperature
gradient in the electron gas, resulting in pronounced effects in multi-layered metal films due
to discontinuous heat transfer and load transmission across the interface.
The rest of this chapter will include a review of heat transfer models that include classical
Fourier’s conduction, hyperbolic, and Dual-Phase Lag (DPL) models; these are followed by
several classical two-step heating models as well as a semiclassical two-step heating model.
Very detailed analyses of ultrafast melting and resolidification of thin metal film and particles
in micron size are discussed next. Finally, the hot electron blast, which may be the very cause
for clean cuts produced by femtosecond lasers, will be discussed.
2. NONEQUILIBRIUM MODELS
2.1. Classical Heat Conduction and Its Extension
Classical Heat Conduction

The energy equation that is valid for any spatial and temporal scales can be expressed as
∂T
C = −∇ ⋅ q′′ + S (1)
∂t
where C is the heat capacity (J/m3-K), q′′ is the heat flux vector, and S is the internal heat
source (W/m3) that can be caused by volumetric heating or coupling between electron and
phonon (for ultrafast heat transfer in metal). For classical engineering problems, the heat
capacity is C = ρ c p and the heat flux is related to the temperature gradient by the Fourier’s
law:
q′′ = −k ∇T (2)
Substituting Eq. (2) into Eq. (1) results in the following very well known heat conduction
equation:
∂T
ρcp = ∇ ⋅ ( k ∇T ) + S (3)
∂t
If the thermal conductivity is independent from temperature, Eq. (3) can be further
simplified to
1 ∂T S
= ∇ 2T + (4)
α ∂t k
where α = k /( ρ c p ) is thermal diffusivity. Equation (4) is a parabolic equation since the

first order derivative with respect to time and the second order derivative with respect to
spatial variables appeared in the equation.
Hyperbolic Conduction Model

The classical heat conduction theory based on Fourier’s law assumes that thermal
disturbance propagates with an infinite speed. As heat conduction is accomplished by
successive collision of the energy carriers (phonons or electrons), the prorogation of thermal
disturbance is always at a finite speed. This is particularly important for those processes
involving extremely short times, cryogenic temperatures, or high heat fluxes. To account for
the finite propagation speed of thermal wave, the Cattaneo-Vernotte thermal wave model can
be used
∂q′′
q′′ + τ = − k ∇T (5)
∂t
where τ is the relaxation time that can be interpreted as the time scale at which intrinsic length
scale of thermal diffusion ( α t ) is equal to the intrinsic length scale of thermal wave ( ct )
(Tzou and Li, 1993; Tzou, 1997), where c is the thermal propagation speed. Thus the
relaxation time is
τ = α / c2 (6)
Substituting Eq. (5) into Eq. (1) and assuming constant thermal conductivity, the
following energy equation is obtained
1 ∂T τ ∂ 2T 1⎛ ∂S ⎞
+ = ∇ 2T + ⎜ S + τ ⎟ (7)
α ∂t α ∂t 2
k⎝ ∂t ⎠
which is referred to as the hyperbolic heat conduction model because the second order
derivative of temperature with respect to time appeared. With appropriate relaxation times,
Eq. (7) can be used to describe temperatures of different energy carriers – such as phonon and
electron temperatures – as well be discussed later.
Mathematically, Eq. (7) is a thermal wave equation and the thermal diffusivity appears as
a dumping effect of thermal propagation. In addition to the initial temperature, the initial rate
of temperature, ∂T ( x, 0) / ∂t must also be known. Consider a semi-infinite solid with initial
conditions of T = 0 and ∂T / ∂t = 0 at t = 0 . If a temperature surface temperature x = 0 is
suddenly increased to Tw, the temperature distribution in the semi-infinite solid is shown in
Figure 1. It can be seen that a sharp wave front advances in the physical domain which
separates the heat affected zone from the thermally undisturbed zone. Across the wave front,
the temperature presents a finite jump while the temperature gradient presents a singularity.
At β = 10 and thereafter, the wave front becomes weak and the results predicted by wave
theory collapses onto those predicted by the classical model based on the Fourier’s law.
Figure 1. Temperature distribution in a semi-infinite solid at different times (Ozisik and Tzou, 1992)
Dual-Phase Lag Model

Equation (5) can be viewed as the first order approximation of the following equation,
q′′(r, t + τ ) = −k ∇T (r, t ) (8)
which indicates that there is a delay between the heat flux vector and the temperature
gradient. For the same point in the conduction medium, the temperature gradient is
established at time t, but the heat flux vector will be established at a later time t + τ , i.e., the
relaxation time, τ , can be interpreted as the time delay from the onset of the temperature
gradient to the heat flux vector. While the thermal wave model assumes that the temperature
gradient always precedes the heat flux, Tzou (1997) proposed a dual-phase lag model that
allows either the temperature gradient (cause) to precede heat flux vector (effect) or the heat
flux vector (cause) to precede the temperature gradient (effect), i.e.,
q′′(r, t + τ q ) = − k ∇T (r, t + τ t ) (9)
where τ q is the phase lag for the heat flux vector, while τ T is the phase lag for the
temperature gradient. If τ q > τ T , the local heat flux vector is the result of the temperature
gradient at the same location but an early time. On the other hand, if τ q < τ T , the
temperature gradient is the result of the heat flux at an early time. The first order
approximation of Eq. (9) is:
∂q′′ ⎡ ∂ ⎤
q′′ + τ q = − k ⎢∇T + τ T (∇T ) ⎥ (10)
∂t ⎣ ∂t ⎦
Substituting Eq. (10) into Eq. (1), the energy equation based on the dual-phase lag model
is
1 ∂T τ q ∂ 2T ∂ 1⎛ ∂S ⎞
+ = ∇ 2T + τ T (∇ 2T ) + ⎜ S + τ q ⎟ (11)
α ∂t α ∂t 2
∂t k⎝ ∂t ⎠
which reduces to the parabolic conduction equation (3) if τ q = τ T . In absence of phase lag
for temperature gradient ( τ T = 0 ), Eq. (11) is reduced to the hyperbolic conduction model,
Eq. (7). As the values of τ q and τ T increase, Eq. (10) may not be accurate enough because
the higher order terms were neglected. For the case that τT t , up to the first order, and
τ q < t , up to the second order, Eq. (10) becomes
∂q′′ τ q ∂ 2q′′
2
⎡ ∂ ⎤
q′′ + τ q + = − k ⎢∇T + τ T (∇T ) ⎥ (12)
∂t 2 ∂t 2
⎣ ∂t ⎦
substituting Eq. (12) into Eq. (1), the heat equation with linear effect of τ T and the second
order effect of τ q is obtained:
1 ∂T τ q ∂ 2T τ q ∂ 3T 1⎛ ∂S τ q ∂ 2 S ⎞
2 2
∂ 2
+ + = ∇ 2
T + τ (∇ T ) + ⎜ S + τ + ⎟ (13)
α ∂t α ∂t 2 2α ∂t 3 k ⎜⎝ ∂t 2 ∂t 2 ⎟⎠
T q
∂t
where the third order derivative of temperature with respect to time appeared. In addition, the
second order derivative of the heat source also appeared in the heat equation.
Figure 2 shows a typical response of thermal lagging with β = t / τ q and
B = τ T / τ q = 100 , which are typical values for metals. It can be seen that the dual-phase lag
model with the τ q effect displays a wave behavior in heat propagation. The wave front that is
2
located at x = ( )
2ατ T / τ q t , separates the heat affected zone from the thermally unaffected
zone. The linear DPL model, on the other hand, exhibits high-order diffusion behavior. While
the DPL- τ q model results in high temperature than the linear DPL model in the heat affected
2
zone, the temperature obtained from the linear DPL model is higher than that obtained by the
DPL- τ q model in the thermally unaffected zone.
2
Figure 2. Comparison among Fourier diffusion, thermal wave, and DPL models (Tzou, 2006).
The wave propagation speed for DPL- τ q2 model is 2ατ T / τ q , which is 2τ T / τ q times
faster than the classical Cattaneo-Vernotte (CV) wave. Obviously, the speed of DPL- τ q
2
model will be slower if τ T / τ q < 0.5 . The result of classical Fourier conduction can be
obtained by setting τ q = τ T . The temperature obtained from the classical Fourier conduction
is between the results obtained from the CV wave and DPL models.
2.2. Two-Step Models
For ultrafast laser materials processing, the laser pulse duration is much shorter than the
time it takes for the electron and lattice to reach to thermal equilibrium. Various two-step
heating models can be employed to describe such nonequilibrium phenomena between the
electrons and lattices.
Parabolic Two-Step Model

Assuming heat conduction in the electron can be described by the Fourier’s law and
neglecting heat conduction in the lattice, the energy equations of the free electrons and
lattices (phonons) are
∂Te
Ce = ∇ ⋅ ( ke∇Te ) − G (Te − Tl ) + S (14)
∂t
∂Tl
Cl = G (Te − Tl ) (15)
∂t
where the volumetric heat capacity of lattice is Cl = ρcp, and the volumetric heat capacity of
electrons is
π 2 ne k B
Ce = Te = BeTe (16)
2μ F
where ne is the number density of electrons, kB is the Boltzmann constant, and μF is Fermi
energy. Equation (16) indicates that the volumetric heat capacity of the electron is
proportional to the electron temperature. It should be noted that the volumetric heat capacity
of electron is much less than that of the lattice even at very high electron temperature. At
nonequilibrium condition, thermal conductivity of the electrons depends on the temperatures
of both electrons and lattice, i.e.,
⎛T ⎞
ke = keq ⎜ e ⎟ (17)
⎝ Tl ⎠
where keq(T) is the thermal conductivity of the electron when the electrons and lattice are in
thermal equilibrium. The electron-phonon coupling factor, G in W/m3K, is a new property in
microscale heat transfer, which describes the time-rate of exchange of thermal energy
between electrons and phonons per unit volume. It can be estimated by
9 ne k B2TD2 vF
G= (18)
16 Λ (Tl )Tl μ F
where TD is Debye temperature, vF is Fermi velocity, and Λ is the electron mean free path.
Neglecting conduction in the lattice is justified by the fact that the thermal conductivity of the
lattice is two orders of magnitude smaller than that of the free electrons (Klemens and
Williams, 1986). The heat conduction model represented by Eqs. (14) and (15) is referred to
as a parabolic two-step model because Fourier’s law was used to describe heat conduction in
the electron gas.
If the thermophysical properties of electrons and lattices can be assumed to be constant,
Eqs. (14) and (15) can be combined to eliminate the electron temperature and obtain the
following equation for the lattice temperature
Ce + Cl ∂T CeCl ∂ 2T C ∂ S
+ = ∇ 2Tl + l (∇ 2T ) + (19)
ke ∂t Gke ∂t 2
G ∂t ke
where the subscript “l” for lattice has been dropped for notation. Comparing Eq. (19) with the
energy equation for the dual-phase lag model, Eq. (11), it is apparent that they have almost
identical form except the partial derivative of heat source with respective to time is not
present in Eq. (19). The thermophysical properties in the dual-phase lag model is related to
the properties appeared in the two-temperature model by
ke C CeCl
k = ke , α = , τT = l , τ q = (20)
Ce + Cl G G (Ce + Cl )
The ratio of two phase-lag times is
τ T Ce + Cl C
= = 1+ l (21)
τq Ce Ce
which indicates that τ T is always greater than τ q .
Hyperbolic Two-Step Model

If we considered the hyperbolic effect on the conduction in the electron gas, the energy
equation for the electron gas is
∂Te
Ce = −∇ ⋅ q′′ − G (Te − Tl ) + S (22)
∂t
where
∂q′′
q′′ + τ e = − ke∇Te (23)
∂t
while the energy equation for the lattice is still Eq. (15). Equations (22) and (23) can be
combined to yield
∂Te ∂ 2T
Ce + Ceτ e 2e = ∇ ⋅ (ke∇Te )
∂t ∂t
∂ ∂S
−G (Te − Tl ) − τ e ⎡⎣G (Te − Tl ) ⎤⎦ + S + τ e (24)
∂t ∂t
The conduction model represented by Eqs. (24) and (15) is referred to as a hyperbolic
two-step model. Qiu and Tien (1993) simulated picosecond laser-metal interaction using
different models. Figure 3 shows comparison of the front surface electron temperature change
of a 0.1μm- thick gold film subject to a 96 fs-laser pulse irradiation. It can be seen that the
parabolic one-step (POS) model (classical Fourier’s conduction) and hyperbolic one-step
(HOS) model (C-V wave model) significantly over predict the electron temperature change.
The parabolic two-step (PTS) model predicted the general temperature response but it failed
to predict the finite speed of energy propagation. Therefore, they suggested that the
hyperbolic two-step model can provide better accuracy for ultrafast laser interaction with
metal.
Figure 3. Comparison of electron temperature change (Qiu and Tien, 1993).
Dual-Parabolic Two-Step Model

The contribution of heat conduction in phonon was neglected in the above two models. If
it is assumed that the heat conduction in the phonons can be modeled using the classical
Fourier’s law, the energy equations of the lattices (phonons) are
∂Tl
Cl = ∇ ⋅ ( kl ∇Tl ) + G (Te − Tl ) (25)
∂t
The bulk thermal conductivity of metal measured at equilibrium, keq, is the sum of
electron thermal conductivity, ke, and the lattice thermal conductivity, kl. Since the
mechanism for heat conduction in metal is diffusion of free electron, ke is usually dominate.
For gold, ke is 99% of keq, while kl only contributes to 1% of keq (Klemens and Williams,
1986). Since both Eqs. (14) and (25) are parabolic, this model is referred to as dual-parabolic
two-step model. For the case that phonon temperature gradient is significant, inclusion of
conduction in phonon is essential.
Dual-Hyperbolic Two-Step Model

For the case that heat conduction in both electrons and phonons need to be considered
using hyperbolic model, the energy equation for the lattice becomes
∂Tl
Cl = −∇ ⋅ q′′l + G (Te − Tl ) (26)
∂t
where
∂q′′l
q′′l + τ l = −kl ∇Tl (27)
∂t
Combining (26) and (27) to eliminate q′′l yield
∂Tl ∂ 2T ∂
+ Clτ l 2l = ∇ ⋅ (kl ∇Tl ) + G (Te − Tl ) + τ l ⎡⎣G (Te − Tl ) ⎤⎦
Cl (28)
∂t ∂t ∂t
Equation (28) together with Eq. (24) become governing equations for the dual-hyperbolic
two-step model. Chen and Beraun (2001) applied the dual-hyperbolic two-step model to
simulate ultrashort laser pulse interactions with metal film. They found that the electron
temperatures obtained from the dual-hyperbolic model and the hyperbolic model are very
close. However, the lattice temperatures obtained from the two models differ significantly.
3. SEMICLASSICAL TWO-STEP HEATING MODEL

In a non-uniformly heated conductor in the absence of current there arises an electric
field (Ashcroft and Mermin, 1976). On the other hand, substantial electron heating will occur
if electric field and carrier gradient are excessively large (Snowden, 1986). When an opaque
solid medium is heated by ultrashort laser pulses, very high carrier gradient and, in turn, a
strong electric field can be generated due to the fact that interaction between photons and
electrons and the subsequent non-equilibrium thermal transport are essentially confined to a
very small region where the laser light penetrates. There is a need to investigate whether the
electric field and carrier gradient caused by ultrashort laser pulses can influence thermal
response of electrons and the lattice.
In this section, a semiclassical two-step heating model that includes the conservation
equations of number density, momentum and energy for electron subsystem is presented.
These equations are derived from the first three moments of the Boltzmann transport
equation. For the phonon subsystem the thermal transport equation remains the same as that
used in phenomenological two-temperature (2T) models (Anisimov et al., 1974; Qiu and
Tien, 1993; Chen and Beraun, 2001). The main difference between the semiclassical and the
phenomenological 2T models is that the former includes the effects of the electron drift
velocity resulting from non-uniform electric temperature field.
3.1. Model Formulation
Let f (r, u, t ) be a distribution function of the free electrons at time t in the phase space
with position vector ( r ) and velocity vector (u). If the Lorentz force due to magnetic field is
negligible, the Vlasov equation becomes (Kruer, 1988)
df ∂f e ⎛ ∂f ⎞
= + u ⋅∇ r f + E ⋅∇ u f = ⎜ ⎟ (29)
dt ∂t m ⎝ ∂t ⎠c
where the del operators ∇ r = ∂ / ∂xi and ∇u = ∂ / ∂ui (i = 1, 2, 3 refers to the direction);
eE / m is the Lorentz force resulting from electric field E with e and m denoting the charge
and mass of an electron. The term (∂f /∂t)c is the time-rate of change of f due to the electron
collision.
To derive the conservation equations of number density, momentum and energy for the
electron subsystem, electron number density n(r, t ) , mean velocity vector v (r, t ) , kinetic
pressure dyad P% (r, t ) , and energy flux vector q′′(r, t ) are defined below (Kruer, 1988):
n(r, t ) = ∫ f (r, u, t )du (30)
1
n∫
v (r, t ) = uf (r, u, t )du (31)
P% (r, t ) = m ∫ ( u − v )( u − v ) f (r, u, t )du (32)
m
( u − v ) ⋅ ( u − v )( u − v ) f (r, u, t )du
2∫
q′′(r, t ) = (33)
The continuity equation can be derived by integrating Eq. (29) over velocities
⎛ ∂f e ⎞ ⎛ ∂f ⎞
∫ ⎜⎝ ∂t + u ⋅∇ r f+
m
E ⋅∇ u f ⎟ du = ∫
⎠
⎜ ⎟ du
⎝ ∂t ⎠c
(34)
Since r and u are independent, Eq. (34) can be re-expressed as
∂n ⎛ ∂n ⎞
+ ∇r ⋅ (nv) = ⎜ ⎟ (35)
∂t ⎝ ∂t ⎠c
In the above simplification the fact that f → 0 as v → ∞ and the condition that E is
independent of u are used.
Multiplying Eq. (29) by the velocity u and integrating the equation over the velocity
space results in the equation of the conservation of momentum
∂v 1 ⎛ ∂v ⎞
m + mv ⋅∇ r v − eE + ∇ r ⋅ P% = m ⎜ ⎟ (36)
∂t n ⎝ ∂t ⎠c
In absence of current the electric field E = β ∇Te with β = −1.42 × 10

−4
(Te / TF ) V/K
for free electrons (Ashcroft and Mermin,, 1976). Assume that the pressure is isotropic, i.e.,
P% = pI% = nk BTe I% , where kB is the Boltzmann constant and I% the unit dyad. The collision
term m ( ∂v / ∂t )c is approximated as (Snowden, 1986)
⎛ ∂v ⎞ v
m ⎜ ⎟ = −m (37)
⎝ ∂t ⎠c τk
where τk = mμoTl/eTe is the momentum relaxation time (Baccarani and Wordeman, 1985)
with μo being the mobility of electrons at room temperature.
Similarly, the energy equation can be obtained by multiplying Eq. (29) by the kinetic
energy mu ⋅ u / 2 , averaging the result over the entire velocity space, and utilizing the
continuity and momentum conservation conditions. It is as follows:
∂ξ 1 1 ⎛ ∂ξ ⎞
+ v ⋅∇ rξ + ∇ r ⋅ (nk BTe v ) + ∇ r ⋅ q′′e − eβ v ⋅∇Te = ⎜ ⎟ (38)
∂t n n ⎝ ∂t ⎠c
2
where ξ = m v / 2 + 3k BTe / 2 is the electron energy density. The collision term (∂ξ/∂t)c,
primarily due to electron-phonon scattering, is approximated as (∂ξ/∂t)c = (ξ-ξl)/τe-p where ξl
denotes the equilibrium energy at Te = Tl and τe-p the characteristic time for electrons and
phonons to reach equilibrium (Snowden, 1986). By making use of the momentum equation
and including the optical-electron scattering due to laser excitation, Eq. (38) is re-written in
the form
⎛ ∂T 2 ⎞
Ce ⎜ e + v ⋅∇ rTe + Te∇ r ⋅ v ⎟ + ∇ r ⋅ q′′e = −G (Te − Tl ) + S (r, t ) (39)
⎝ ∂t 3 ⎠
where G = Ce/τe-p is the electron-phonon coupling factor.
3.2. Electron Kinetic Pressure
If one atom can give one valence electron to the free electron gas, the number density of
the excited electrons at temperature Te << TF is given by (Kittle, 1967)
1 T
n = π 2N e (40)
3 TF
This leads to the renowned linear electron heat capacity
3
Ce = nk B = CeoTe (41)
2
with Ceo = π Nk B / 2TF . Substitution of Eq. (31) for nk B into the gradient of the isotropic
2
electron kinetic pressure dyad yields

⎛2 ⎞
∇ r ⋅ P% = ∇ ⎜ CeoTe2 ⎟ (42)
⎝3 ⎠
The above ∇ r ⋅ P% is the so-called hot-electron blast force ∇ gCeoTe ( 2

) that was derived
by Falkovsky and Mishchenko (1999) with undetermined constant g ~ 1. In fact, this form of
the hot-electron blast force is only adequate for Te < 0.1TF. A general form of the hot-electron
blast force should be 2∇(CeTe)/3.
Figure 4 gives the temperature-dependent heat capacity of electrons for gold. The
following relations are approximated from the result of Jiang and Tsai (2005):
⎧ BeTe Te < TF / π 2
⎪
⎪ 2 B T /3 + Ce′ / 3 TF / π 2 ≤ Te < 3TF / π 2
Ce = ⎨ e e (43)
⎪ Nk B + Ce′ / 3 3TF / π 2 ≤ Te < TF
⎪ Te ≥ TF
⎩3 Nk B / 2
Figure 4. Heat capacity of gold (Chen et al. , 2006).
where
3Nk B / 2 − BeTF /π 2
Ce′ = BeTF /π 2 +
TF − TF / π 2 ( Te − TF / π 2 ) (44)
with the relationships ∇ r ⋅ P% = ∇ r ( nk BTe ) , E = β ∇Te , and ( ∂v / ∂t )c = eTe v / μoTl , Eq.

(36) can be re-written in terms of velocity and temperature
∂v ⎡ ⎛ T ∂Ce ⎞ ⎤ eTe v
m + mv ⋅∇ r v + ⎢ k B ⎜1 + e ⎟ − eβ ⎥ ∇Te = − (45)
∂t ⎣ ⎝ Ce ∂Te ⎠ ⎦ μoTl
Te ∂Ce
For low temperatures = 1 since Ce = CeoTe.
Ce ∂Te
The semiclassical two-step heating model includes: (i) the momentum equation (45) and
energy equation (39) for electron subsystem; (ii) the energy equation (26) for lattice
subsystem; and (iii) the constitutive equations for heat fluxes in electrons and the lattice, Eqs.
(23) and (27). For a 1D problem, for example, the above five equations are satisfied by the
five unknowns, qex′′ , Te, vx, Qlx and Tl, while vx is excluded from the phenomenological 2T
models.
3.3. Laser Heat Source in Two-Temperature Models
A laser heat source term widely used in 2T models is given as follows:
⎧ ⎡⎛ t − 2t ⎞
2
⎤⎫
4 ln 2 (1 − R )φ ⎪ x ⎥ ⎪⎬
exp ⎨− − 4 ln 2 ⎢⎜
p
S ( x, t ) = − ⎟⎟ (46)
π δ tp ⎜
⎪⎩ δ ⎢⎝ t p
⎣ ⎠ ⎥⎪
⎦⎭
where φ is fluence, R is surface reflectivity, δ is optical penetration depth, and tp is pulse

duration. Lasing is assumed to start at t = 0 and end at t = 4tp. Equation (46) is derived
directly from the Beer’s law. When it is used in the energy equation for an electron
subsystem, an intrinsic assumption that the excited electrons are immediately and fully
thermalized has been applied.
Immediately after a metal solid is illuminated by an ultrashort laser pulse, two competing
processes take place. The photo-excited electrons move ballistically with a velocity close to
the Fermi velocity. Therefore, the laser heat source used in 2T models should be modified by
changing the spatial distribution of the laser heat source according to the electron ballistic
motion. Meanwhile, those excited electrons continue to thermalize, through collision, into a
Fermi-Dirac distribution in which the electron temperature is measurable. This also alters the
time function of the laser heat source.
To incorporate the ballistic electron effect, an effective absorption depth was introduced
by including an electron ballistic range to the optical penetration depth (Wellershoff et al.,
1999; Chen et al., 2002; Ibrahim et al., 2004). For instance, a range of 105 nm was assumed
for gold, based on the postulation that the excited, nonequilibrium electrons move at the
Fermi velocity (VF) for about 100 fs (Wellershoff et al., 1999). Another approach considered
that the ballistic electrons scatter randomly at the Fermi velocity for a finite period of time (tb)
(Chen et al., 2006). For a uniform 1D mesh of q points, for example, the spatial function of
(
the absorbed laser energy is modified with the initial values S j = exp[− x / δ ] as follows:
( ( ( ( ( ( ( ( (
S j = ( S j −1 + S j +1 ) / 2 for interior points and S1 = ( S1 + S 2 ) / 2 and S q = ( Sq −1 + S q ) / 2
(
for the boundary points. All S j on the left-hand sides are updated with the previous values.
The above redistribution continues for N times, where N is determined from NΔx/VF = tb
with Δx being grid spacing.
A system response function was proposed to account for the delayed rise time of electron
temperature (Sun et al., 1993; Ibrahim et al., 2004). For the electrons excited at time t ' the
delay function is given by (Ibrahim et al., 2004)
⎧⎪ ⎡ ⎛ t − t ' ⎞ 2 ⎤ ⎫⎪
Ψ (t , t ') = u (t − t ') ⎨1 − exp ⎢ − ⎜ ⎟ ⎥⎬ (47)
⎪⎩ ⎢⎣ ⎝ t R ⎠ ⎥⎦ ⎪⎭
Completion of the electron thermalization, according to Eq. (47), is 0, 63.2%, and

99.99% when t − t ' = 0, tR and 3tR, respectively.
By combining both the effects of electron ballistic motion and internal thermalization, the
volumetric laser heat source (Ŝ) employed in a 2T model becomes
Sˆ ( x, t + Δt )
4 ln 2 (1 − R)φ ( t '=t −Δt ⎡ ⎛ t '− 2t p ⎞

2
⎤
= S ( x) ∑ exp ⎢ −4 ln 2 ⎜ ⎟⎟ ⎥ [ Ψ (t + Δt , t ') − Ψ (t , t ') ] (48)
π δ tp ⎢ ⎜ t ⎥
t '= 0
⎣ ⎝ p ⎠ ⎦
3.4. Numerical Results
For numerical analysis a 1D version of the governing equations together with the
following initial and boundary conditions are solved with a central difference method:
Te ( x, 0) = Tl ( x, 0) = T o ; vx ( x, 0) = 0 (49)
qex′′ (0, t ) = qex′′ ( L, t ) = 0 (50)
where To is set at 300 K. The thermal conductivity, ke, and electron-phonon coupling factor,
G, (Chen et al., 2005) are temperature-dependent:
(ϑe2 + 0.16)5 / 4 (ϑe2 + 0.44)ϑe

ke = χ (51)
(ϑe2 + 0.092)1/ 2 (ϑe2 + ηϑl )
⎡A ⎤
G = GRT ⎢ e (Te + Tl ) + 1⎥ (52)
⎣ Bl ⎦
In Eq. (51) θ e = Te / TF and θl = Tl / TF . The material properties used are as follows

unless otherwise mentioned: Ceo = 70 Jm-3K-2, kl = 0, TF = 6.4 ×104 K, N = 5.9×1028 m-1, μo =
4.8×10-3 m2s-1V-1, R = 0.93, δ = 15.3 nm, Ae = 1.2×107 K-2s-1, Bl = 1.23×1011 K-1s-1, GRT =
2.2×1016 Wm-3K-1, χ = 353 Wm-1K-1, η = 0.16, and tb = 100 fs.
Figure 5 shows the normalized change of electron temperature at the front surface (x = 0)
of an 80-nm gold film irradiated by a 2.8 mJ/cm2, 800-nm, 150-fs laser pulse. The values of
R, δ and tR used in this calculation are 0.967, 12.7 nm and 500 fs, respectively. The result
computed with tR = 0 (i.e. all the excited electrons complete thermalization instantaneously) is
also compared. The simulated maximum electron temperature is 335.9 K for the case of tR =
500 fs and 357.7 K for tR = 0, occurring at 0.988 ps and 0.358 ps, respectively. Clearly, the
result computed with tR = 500 fs agrees well with the experimental data (Ibrahim et al., 2004),
but the case of tR = 0 fails to capture the electron temperature response. At this low level of
laser fluence the difference between the semiclassical and the phenomenological 2T models is
insignificant.
Figure 5. Comparison of the change in electron temperature at the front surface of an 80-nm gold film
irradiated by a 2.8 mJ/cm2, 800-nm, 150-fs laser pulse (Chen et al., 2006).
35
Semiclassical 2T model
Phenomenological 2T model
30
25
Temperature (103 K)
20
15 Te
10
5
Tl
0
0 5 10 15 20
Time (ps)
Figure 6. Time history of temperatures at the front surface of a 1.0-μm gold film heated with a 1.0 J/m2,
560-nm, 100-fs laser pulse (Chen et al., 2006).
Figure 6 shows the electron and lattice temperatures at the heated surface of a 1.0-lm gold
film illuminated by a 1.0 J/cm2, 630-nm, 100-fs laser pulse. At this level of laser fluence, the
difference between the semiclassical and phenomenological 2T models is discernable. Further

analysis shows that the difference can be immaterial for fluence 0.1 J/cm2 and lower.
Figure 7 compares the simulated damage fluence threshold with the experimental data
(Stuart et al., 1996) for gold irradiated by two 1053-nm laser pulses of durations 600 fs and
800 ps, respectively. It is assumed here that initiation of damage occurs when the front
surface completes the solid-to-liquid phase change (melting point Tm = 1337 K and fusion
heat Hm = 6.275×104 Jkg-1). At this wavelength the skin depth is 3 nm. The surface
reflectivity used in the simulation is 0.9260 for the 600-fs pulse and 0.9546 for the 800-ps
pulse. For the case of the 800-nm film heated by the 600-fs laser pulse, for example, the
damage threshold obtained from the semiclassical model is about 12% higher than that
obtained from the phenomenological model. The difference reduces to 7% for the longer,
800-ps pulse.
Figure 7. Comparison of the simulated and measured damage thresholds for a gold film heated with
600-fs and 800-ps laser pulses (Chen et al., 2006).
As given in the energy equation (40) for electron subsystem, the two terms involving the
electron drift velocity are v ⋅∇ rTe and 2Te ∇ r ⋅ v / 3 . The numerical results show that the
latter has much more impact on the change of electron temperature than the former. The
difference is on the order of six (6). Accordingly, the term v ⋅∇ rTe can be neglected without
significant change of the result.
4. ULTRAFAST MELTING AND RESOLIDIFICATION

The models presented in the preceding two sections are for the case that the lattice
temperature is below the melting point so that only pure conduction needs to be modeled.
When the lattice temperature at the heating surface reaches melting point, melting takes place
on the heating surface while the deeper part of the metallic target is still at room temperature
(Kuo and Qiu, 1996). After the laser pulse is off, the melted layer near the heating surface
will be resolidified. Such rapid melting and resolidification can find their application in
selective laser sintering of metal powders using pulsed laser (Fischer et al., 2002; Konrad et
al., 2007). During ultrashort laser pulse interaction with metal target, the solid-liquid interface
velocity is controlled by nucleation kinetics, instead of energy balance (Kuo and Qiu, 1996).
In this section, a fixed grid interfacial tracking method is presented to solve kinetics
controlled rapid melting and resolidification during ultrashort pulse laser interaction with a
free-standing metal film. Nonlinear electron heat capacity and temperature-dependent
electron-lattice coupling factor on the rapid phase change are also investigated. Finally,
ultrafast melting and resolidification of a submicron gold particle subject to pico- to
femtosecond laser pulse are studied.
4.1. Interfacial Tracking Method
Many numerical models for melting and solidification of various Phase Change Materials
(PCMs) have been developed. The numerical models can be divided into two groups (Vollor,
1997): deforming grid schemes (or strong numerical solutions) and fixed grid schemes (or
weak numerical solutions). Deforming grid schemes transform solid and liquid phases into
fixed regions by using a coordinate transformation technique. The governing equations and
boundary conditions are complicated due to the transformation. However, the disadvantage of
deforming grid schemes is that it requires significant amount of computational time. On the
contrary, the fixed grid schemes use one set of governing equations for the whole
computational domain including both liquid and solid phases, and solid-liquid interface is
later determined from the temperature distribution. This simplicity makes the computation
much faster than deforming grid schemes, while it still provides reasonably accurate results
(Sasaguchi et al., 1996). There are two main methods in the fixed grid schemes: the enthalpy
method and the equivalent heat capacity method. The enthalpy method (Binet and Lacroix,
2000) can solve heat transfer in mushy zone but has difficulty with temperature oscillation,
while the equivalent heat capacity method (Morgan, 1981; Hsiao, 1984) requires large enough
temperature range in mushy zone to obtain converged solution. Cao and Faghri (1990)
combined the advantages of both enthalpy and equivalent heat capacity methods and
proposed a Temperature Transforming Model (TTM) that could also account for natural
convection.
Kuo and Qiu (1996) solved rapid melting and resolidification of gold film subject to
picosecond laser heating. The energy equation for electrons was solved using a semi-implicit
Crank-Nicholson scheme, while the energy equation and phase change for lattices was solved
using an explicit enthalpy model. While explicit scheme is very easy to apply, a very small
time step is required to ensure stability. Development of an implicit scheme for enthalpy
model that can outperform explicit scheme is numerically challenging (Voller, 1997). The
temperature transforming model (Cao and Faghri, 1990) and the enthalpy linearization model
(Swaminathan and Voller, 1993) are two very efficient models that can be easily discretized
using implicit scheme. These two models assume phase change occurs in a range of
temperature near melting point, which prevents inclusion of superheat in solid in melting
stage and undercooling of liquid in the solidification stage. The interfacial tracking method
proposed by Zhang and Chen (2008a) can be used to accurately predict the kinetics controlled
melting and resolidification processes.
Figure 8. Physical model for laser melting of thin film (Zhang and Chen, 2008a).
The physical model of the problem under consideration is shown in Figure 8. A metal
film with a thickness of L and an initial temperature of Ti is subjected to a temporal Gaussian
laser pulse with a FWHM (Full Width at Half Maximum) pulse width of tp and fluence of J
(J/m2) from the left surface (x = 0). The problem can be approximated to be one-dimensional
because the radius of the laser beam is significantly larger than the metal film thickness. The
dual parabolic two-step model will be used to describe the ultrafast laser-metal thin film
interaction. The energy equations of the free electrons and lattice can be obtained by
simplifying Eqs. (14) and (25), i.e.,
∂Te ∂ ⎛ ∂Te ⎞
Ce = ⎜ ke ⎟ − G (Te − Tl ) + S (53)
∂t ∂x ⎝ ∂x ⎠
∂Tl ∂ ⎛ ∂T ⎞
Cl = ⎜ kl l ⎟ + G (Te − Tl ) (54)
∂t ∂x ⎝ ∂x ⎠
Equation (53) is valid in the entire computational domain. On the contrary, Eq. (54) is
valid in both solid and liquid phases but not at the solid-liquid interface. The heat capacity of
electron is proportional to the electron temperature as indicated by Eq. (16). The thermal
conductivity of the electrons is obtained from Eq. (17). The lattice thermal conductivity, kl ,
is taken as 1% of the thermal conductivity of bulk metal, keq, since the mechanism of heat
conduction in metal is mainly by electrons , i.e.,
kl = 0.01keq (55)
The source terms in Eq. (53) can be described by the following equation:
1− R ⎡x ⎛ t ⎞
2
⎤
S = 0.94 J exp ⎢ − 2.77 ⎜ ⎟⎟ ⎥ (56)
t pδ ⎢δ ⎜t ⎥
⎣ ⎝ p ⎠ ⎦
where R is reflectivity of the thin film, tp is laser pulse duration (s), δ is the optical penetration
depth (m), and J is laser pulse fluence (J/m2).
The energy balance at the solid-liquid interface is (Faghri and Zhang, 2006)
∂Tl , s ∂Tl ,l
kl , s − kl , l = ρ l hmus x = s (t ) (57)
∂x ∂x
where the additional interfacial velocity due to the density change during melting and
solidification has been considered.
For the conventional melting process, the velocity of the solid-liquid interface is obtained
by energy balance as specified by Eq. (57). However, this is not the case for rapid
melting/solidification processes, because the velocity of the interface is dominated by
nucleation dynamics. For short-pulsed laser melting of gold, the velocity of the solid-liquid
interface is described by (Kuo and Qiu, 1996)
⎡ ⎛ h T − T ⎞⎤
us = V0 ⎢1 − exp ⎜ − m l , I m ⎟ ⎥ (58)
⎢⎣ ⎜ RT Tl , I ⎟⎠ ⎥⎦
⎝ g m
where V0 is the maximum interface velocity, Rg is the gas constant for the metal, and Tl,I is the
interfacial temperature. The interfacial temperature, Tl,I, is higher than melting point, Tm,
during melting and lower than melting point during solidification.
The time t = 0 is defined as the time when the peak of a laser pulse reaches the film
surface. Therefore, the initial conditions of the problem are
Te ( x, −2t p ) = Tl ( x, −2t p ) = Ti (59)

The boundary conditions of the problem can be specified by assuming that the heat loss
from the film surface can be neglected, i.e.,
∂Te ∂Te ∂Tl ∂Tl

= = = =0 (60)
∂x x =0 ∂x x =l ∂x x=0 ∂x x =l
The computational domain (0, L) is discretized with non-uniform grids. The implicit
finite difference equations are obtained by integrating Eqs. (53) and (54) in each control
volume and time step (Pantankar, 1980). For the control volume that contains solid-liquid
interface, the lattice temperature, Tl,P, is numerically set as interfacial temperature, Tl,I. The
energy equation for lattice, Eq. (54) is valid for solid and liquid phase only, but not at the
interface. The energy equation for the control volume that contains solid-liquid interface (see
Figure 9) can be written in the enthalpy form:
∂H l ∂ ⎛ ∂T ⎞
= ⎜ kl l ⎟ + G (Te − Tl ) (61)
∂t ∂x ⎝ ∂x ⎠
The volumetric enthalpy can be expressed as
Tl ,I
Hl = ∫ Cl , s (Tl )dTl + f ρ l hm (62)
0
where the first term is enthalpy of solid phase at the interfacial temperature, and the second
term is the latent heat due to partial melting. Substituting Eq. (62) into Eq. (61), one obtains
∂Tl , I ∂f ∂ ⎛ ∂T ⎞
Cl , s (Tl , I ) + ρl hm = ⎜ kl l ⎟ + G (Te − Tl ) (63)
∂t ∂t ∂x ⎝ ∂x ⎠
Figure 9. Grid system for control volume with interface.
which is applicable in the control volume that contains solid-liquid interface. The liquid
fraction, f, is related to the location of the solid-liquid interface by (see Figure 9).
∂f 1 ds us
= = (64)
∂t (Δx) P dt (Δx) P
where ( Δx) P is the size of the control volume P, and s is the solid-liquid interfacial location.
Substituting Eq. (64) into Eq. (63) and integrating the resulting equation in the control
volume that contains solid-liquid interface, the interfacial velocity is obtained as
1 ⎡ kl , w (Tl ,W − Tl , I ) kl ,e (Tl , I − Tl , E )
us = ⎢ −
ρ l hm ⎣ (δ x) w − (0.5 − f P )(Δx) P (δ x)e + (0.5 − f P )(Δx) P (65)
(Δx) P ⎤
+ GP (Te, P − Tl , P )(Δx) P − Cl , s (Tl , I )(Tl , I − Tl ,0I )
Δt ⎦⎥
where kl,w and kl,e are lattice thermal conductivities at faces w and e of control volume,
0
respectively, and Tl , I is the solid-liquid interfacial temperature at the previous time step. The
third and fourth terms in the bracket at the right-hand side of Eq. (65) represent the effects of
electron-lattice interaction and change of the interfacial temperature on the solid-liquid
interfacial temperature.
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. 184 Yuwen Zhang, D. Y. Tzou and J. K. Chen
Figure 10. Comparison of surface lattice temperatures and interfacial locations (Zhang and Chen,
2008a).
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Micro- and Nanoscale Heat Transfer in Femtosecond Laser Processing of Metals 185
Figure 11. Comparison of interfacial temperatures and velocities (Zhang and Chen, 2008a).
Equations (65) will be used together with Eq. (58) to determine the solid-liquid phase
interfacial velocity and temperature. The interfacial location at each step can be determined
once the converged interfacial velocity is obtained.
Figure 10 shows comparison of the surface lattice temperature and interfacial location for
a 1000-nm gold film under irradiation of a laser pulse with fluence of J = 0.3 J/cm2 and a
pulse width of t p = 20 ps .
The grid number is 2502 and the time step is Δt = 0.05 ps. The peak surface lattice
temperature is 1562 K and occurs at t = 27 ps, while the maximum melting depth is 12.59 nm
and occurs at t = 209 ps. The peak temperature and the maximum melting depth obtained in
the present paper are 1.4% and 6% lower than those obtained by Kuo and Qiu (1996). The
duration of phase change in the present paper is 9% shorter than that obtained by Kuo and
Qiu (1996).
The interfacial temperature and velocity during the picosecond laser pulse and gold film
interaction are shown in Figure 11. Strong superheating in the melting stage and undercooling
in the solidification stage can be observed in Figure 11(a), although the degrees of
superheating and undercooling in the present paper are slightly lower than that in Kuo and
Qiu (1996). The peak interfacial velocity in the present paper is lower than that in Ref. Kuo
and Qiu (1996), which is consistent with the result in Figure 10(b). The interfacial velocity
reached to its peak at very early stage of phase change ( t = 23.5 ps ), and the phase change
continues until t = 575 ps . Since interfacial velocity decreases in most part of phase change
process ( 23.5 ps < t < 575 ps ), the melting depth obtained by the present implicit scheme is
smaller than that obtained by explicit scheme.
4.2. Melting and Resolidification of Gold Film Irradiated by Nano- to

Femtosecond Lasers
The thermal conductivity of the electrons given by Eq. (17) is valid when the electron
temperature is much lower than the Fermi temperature, TF, which is 6.42×104 K for gold. If
the electron temperature is comparable to Fermi temperature, the thermal conductivity need to
be calculated by Eq. (51). For high electron temperature, ϑe 1 , Eq. (51) results in the well-
known dependence ke Te5/ 2 that is the characteristics of low-density plasma. On the other
hand, Eq. (51) will reduce to Eq. (17) under low electron temperature limit, ϑe 1 . When
the more sophisticated electron thermal conductivity model, Eq. (51), was employed, the
interfacial velocity and temperature were significantly increased due to the strong deviation
between the two thermal conductivity models at high temperatures (Zhang and Chen, 2008a).
While Eq. (56) is widely used in many existing works to describe the internal heat source
induced by ultrafast laser, Wellershoff et al. (1999) and Hohlfeld et al. (2000) suggested that
the ballistic motion and diffusion of the hot electrons spread the absorbed laser energy into
much greater depth of electrons, especially for the s/p-band metals. This hot electron bath
should be initially localized within either the ballistic range (for s/p-band metals) or the
optical penetration depth. Therefore, it is necessary to incorporate the effect of the ballistic
motion and hot electron diffusion by adding the ballistic range, δ b , to the optical penetration
depth in Eq. (56), i.e.,
1− R ⎛ x ⎛ t ⎞ ⎞
2
S = 0.94 J exp ⎜ − 2.77 ⎜ ⎟ ⎟ (66)

t p (δ + δ b )[1 − e − L /(δ +δb ) ] ⎜ δ + δb ⎜t ⎟ ⎟
⎝ ⎝ p⎠ ⎠
− L /(δ +δ b )
where [1 − e ] is to correct the film thickness effect. Consideration of the ballistic
range of electron motion resulted in penetration of laser energy into deeper parts of the metal
film and thus lowers interfacial velocity and temperature (Zhang and Chen 2008a).
The heat capacity of electron is assumed to be proportional to the electron temperature as
it was indicated by Eq. (16). This assumption is valid for Te < 0.1TF where TF is the Fermi
temperature (Ashcroft and Mermin, 1976). When the electron temperature is greater than the
Fermi temperature, the electron heat capacity can be obtained by ideal gas model, i.e., Ce =
3NkB/2, where N is number density of atoms, and kB is Boltzmann constant. Jiang and Tsai
(2005) obtained electron heat capacity by numerically integrating the relation
⎛∂ ε ⎞
Ce (Te ) = ne ⎜ ⎟ (67)
⎝ ∂Te ⎠V
where ne is the number density of free electrons, ε is average kinetic energy per electron in
J, and V is the volume. The results showed that the heat capacity can be accurately
represented by Eq. (16) for low temperature while ideal gas model is valid for higher
temperature. Chen et al. (2006) analyzed the results of Jiang and Tsai (2005) and suggested
that the electron heat capacity can be approximated by Eq. (43).
Most of the ultrafast laser heating analysis has been carried out with a constant electron-
phonon coupling factor (G). Due to the significant changes in the electron and lattice
temperature caused by high-power laser heating, G could be temperature-dependent. The
phenomenological temperature-dependent G proposed by Chen et al. (2005) is adopted by
Zhang and Chen (2007):
⎡A ⎤
G = GRT ⎢ e (Te + Tl ) + 1⎥ (68)
⎣ Bl ⎦
where GRT is the coupling factor at room temperature; Ae and Bl are materials constants for
electron relaxation time. The coupling factor for liquid phase is enlarged by 20% from the
result obtained from Eq. (68), because electrons collide more frequently with liquid atoms
than with the atoms in the solid crystals.
Figure 12 shows comparison of interfacial locations, temperatures, and velocities
obtained using different electron heat capacities and coupling factors.
Since the highest electron temperature is less than TF/π2, the electron heat capacity
obtained by Eqs. (16) and (43) are identical. The differences exhibited in Figure 12 are
attributed to different electron-phonon coupling factor only. It can be seen that the maximum
melting depth is increased from 10.28 nm (occurring at t = 199 ps) for constant coupling
factor to 17.30 nm (occurring at t = 233.5 ps) for temperature-dependent coupling factor. The
maximum interfacial temperature in the melting stage is increased from 1499 K to 1581 K
when the temperature-dependent coupling factor is considered. The peak interfacial velocity
is increased from 151 m/s to 209 m/s due to the effect of temperature-dependent coupling
factor. The results shown in Figure 12 indicate that strong electron-lattice coupling results in
higher lattice temperature, more rapid melting, and longer duration of phase change.
Figure 12. Comparison of interfacial location, temperatures and velocities (J = 0.3 J/cm2, tp = 20 ps;
Zhang and Chen, 2007).
Figure 13. Effects of pulse width on the interfacial locations, temperatures, and velocities (J = 0.3
J/cm2; Zhang and Chen, 2007).
Figure 13 shows the effects of laser pulse width on the interfacial locations, temperatures,
and velocities for a 1-μm gold film when the laser fluence is held at J = 0.3 J/cm2 . The grid
number is 2502 and the time step varies from is 0.5 fs to 2 ps depending on the pulse widths.
Although total laser energy delivered to the gold film are the same for all cases, concentration
of laser energy into short pulse results in dramatic increase of the electron temperature. Since
the electron temperature for t p = 5 ps and t p = 100 fs exceeded TF / π 2 , it is essential to use
nonlinear electron heat capacity and temperature dependent coupling factor to model ultrafast
melting and resolidification. It can be seen that both the maximum melting depth and duration
of phase change increase with decreasing pulse width although the total laser energy
delivered to the metal film are the same for all cases. The maximum melting depths for four
different pulses width are 5.19 nm, 17.30 nm, 26.88 nm, and 33.32 nm, respectively. The
maximum interfacial temperatures in the melting stage are 1378 K, 1585 K, 1718 K to 1808
K, respectively, and the corresponding peak interfacial velocities are 44 m/s, 211 m/s, 289
m/s and 332 m/s, respectively.
Numerical simulation was also performed for laser pulse of tp = 1 ns, but no melting takes
place because the peak surface lattice temperature was only 1080 K.
4.3. Microparticles
Selective Laser Sintering (SLS) is a rapid manufacturing/tooling technology that can

build functional parts from powdered material via layer-by-layer sintering (for amorphous
powder, such as polycarbonate) or melting (for crystalline powder, such as metal) by a
directed laser beam (Beaman et al., 1997). Fabrication of the metal part is very challenging
because the temperature required to bind the metal powder particles is much higher than that
to bind the amorphous powders particles. The single-component metallic parts can be
fabricated by melting the skin of the powder particles and join together the non-melted solid
cores by resolidifying the liquid layer (Tolochko et al., 2003a,b). SLS of a single component
metal powder via partial melting is very challenging due to the very small window of the
processing parameters (Xiao and Zhang, 2006, 2007). Inappropriate combination of scanning
velocity and laser beam intensity may result in non-melting or complete melting of metal
powder particles.
It has been considered as a drawback of SLS technique that the fabricated metal parts are
always porous and require post-processing to achieve fully densified (Das et al., 1998).
However, porosity is not always undesirable for some applications such as in wick structure
of heat pipes, electrodes of fuel cells, aerospace, and bioengineering. Porous metallic
materials, which are consistent of metal skeleton and beneficial pores, possess useful
properties that the corresponding bulk materials do not have. Instead of avoiding the porosity
in the product, it is possible to utilize and control the porosity in the final product by
controlling the laser pulse width, repetition rate, laser intensity, and scanning velocity. While
nanosecond laser can be used to sinter metal particles (Fischer et al., 2002; Konrad, et al.,
2007), application of pico- to femtosecond lasers in sintering of metal powder particles may
allow more accurate control of porosity of the sintered part via controlling the degree of
partial melting. Because the pulse duration of femtosecond lasers is shorter than the relaxation
times of most materials, and they have extremely high irradiance, the phenomena associated
with femtosecond laser interaction with matter are very different from those associated with
longer pulses. Sintering of metallic submicron or nanoparticles with femtosecond lasers has
not been investigated.
Zhang and Chen (2008b) modeled ultrafast melting and resolidification of submicron
particles irradiated by pico- to femtoseconds laser using the interfacial tracking method. The
diameter of the metal particle is much smaller than the diameter of the laser beam, which is in
turn much smaller than the dimension of the sintered part. Since the laser radiation penetrates
the powder bed over a distance of several powder-sphere diameters, it can be assumed that
multiple scattering of the radiation leads to a nearly homogeneous distribution of the heat flux
within the optically penetrated layer (Fischer et al., 2002), which leads to an almost normal
incidence of the radiation on the surfaces of the grains in the underlying layers. Figure 14
shows the physical model of melting and resolidification under consideration. A particle with
a radius of ro (diameter of dp) and an initial temperature of Ti is subjected to a temporal
Gaussian laser pulse with a FWHM (Full Width at Half Maximum) pulse width of t p and
fluence of J (J/m2) from the surface (r = ro). Due to symmetry of the particle, the model can
be simplified to be 1-D in the r -direction. The origin of time is chosen as the time when the
laser pulse is at its maximum.
Figure 14. Ultrafast melting and resolidification of metal particle (Zhang and Chen, 2008b).
The energy equations of the free electrons (e) and lattice (l) can be obtained by
simplifying Eqs. (14) and (25), i.e.,
∂Te 1 ∂ ⎡ 2 ∂Te ⎤
Ce = r ke − G (Te − Tl ) + S (69)
∂t r 2 ∂r ⎢⎣ ∂r ⎥⎦
∂Tl 1 ∂ ⎡ 2 ∂Tl ⎤
Cl = r kl + G (Te − Tl ) (70)
∂t r 2 ∂r ⎢⎣ ∂r ⎥⎦
The heat conduction in both electrons and lattice is assumed to be parabolic since the
laser pulses considered are 100 fs or longer. The heat capacity of electron is obtained by Eq.
(43). The thermal conductivity of the electrons, which depends on the temperatures of both
electrons and lattice, is obtained form Eq. (51). The temperature-dependent electron-phonon
coupling factor, G, was obtained from Eq. (68). The coupling factor for liquid phase is
enlarged by 20% from the result obtained from Eq. (68), because electrons collide more
frequently with liquid atoms than with the atoms in the solid crystals. The laser heat source
term in Eq. (69) can be described by the following equation:
1− R ⎛ r −r ⎛ t ⎞
2
⎞
S = 0.94 J exp ⎜ o − 2.77 ⎜ ⎟⎟ ⎟ (71)
t p (δ + δ b )[1 − e − ro /(δ +δ b )
] ⎜ δ + δb ⎜t ⎟
⎝ ⎝ p ⎠ ⎠
Figure 15. Effects of pulse width on surface electron and lattice temperatures (J = 0.3 J/cm2, dp = 1 μm;
Zhang and Chen, 2008b).
Figure 16. Effects of pulse width on the interfacial locations, temperatures, and velocities (J = 0.3
J/cm2, dp = 1 μm; Zhang and Chen, 2008b).
The time t = 0 is defined as the time when the peak of a laser pulse reaches the film
surface. Therefore, the initial conditions of the problem are
Te ( x, −2t p ) = Tl ( x, −2t p ) = Ti (72)
The boundary conditions of the problem can be specified by assuming that the heat loss
from the film surface can be neglected, i.e.,
∂Te ∂Te ∂Tl ∂Tl

= = = =0 (73)
∂r r =0 ∂r r = ro ∂r r =0 ∂r r = ro
Figure 15 shows the effects of laser pulse width on the electron and lattice temperatures
at the particle surface for a 1-μm gold particle when the laser fluence is held at J = 0.3 J/cm2 .
The grid number is kept at 2502 while the time step varies from 1 fs to 2 ps depending on the
pulse widths. For the three different pulse widths studied, the peak electron surface
temperatures are 6,355 K, 11,391 K, and 16,983 K, respectively. With the same laser fluence
of J = 0.3 J/cm 2 , the total laser energy delivered to the gold particle are the same for all
cases. However, concentration of laser energy into short pulse duration results in dramatic
increase of the electron temperature. The peak lattice temperatures at the surface for the three
different pulse widths are 1666 K, 1836 K, and 1953 K, respectively. It can be seen that the
surface lattice temperature is much lower than the electron temperature, especially for shorter
pulse width. As the pulse width decreases, there is longer delay from the time at which peak
electron temperature occurs to the time at which peak lattice temperature occurs.
Figure 16 shows the interfacial locations, temperatures, and velocities for the three pulse
widths studied. With decreasing pulse width, both the maximum melting depth and duration
of phase change increase. The maximum melting depths for the three pulse widths are 21.98
nm, 32.51 nm, and 38.99 nm, respectively. The maximum interfacial temperatures in the
melting stage are 1609 K, 1749 K to 1836 K, respectively, and the corresponding peak
interfacial velocities are 227 m/s, 304 m/s and 345 m/s. Thus, larger melting depth and higher
interfacial velocity can be achieved by concentrating the same amount of laser energy into
short pulse duration.
5. HOT-ELECTRON BLAST
5.1. Origin of the Electron Blast Force
The ways in which heat and load are transmitted through materials play a central role in
understanding the refined mechanisms during ultrafast heating and deformation in the
femtosecond domain. To inspire further interest in modeling the various competing
mechanisms in the ultrafast processes, grooving on stainless steel by a nanosecond laser (a)
and by a femtosecond laser (b) is shown in Figure 17. Prolonged heating by nanosecond
lasers causes melting, resulting in rough surfaces when the liquid phase reconsolidates back
into the solid phase, as shown on the left. On the other hand, the same material processed by a
femtosecond laser, as shown on the right, results in a strikingly smooth surface in the channel.
The smooth surface may result from two different processes. In the first process, there is a
spontaneous formation of bubbles in the melt during the femtosecond heating, which
drastically increases from about 1 bubble/cm3 to 1026 bubbles/cm3 due to dynamic nucleation
(Rethfield et al., 2002; Xu and Song 2000; Dyer et al., 2003).
Figure 17. Grooving on stainless steel by different lasers (Photos Courtesy of Lawrence Livermore
Laboratory, the University of California).
The melt is sputtered away as these bubbles rise altogether. The second process, the hot-
electron blast (Tzou et al., 2005; Chen et al., 2006), results from the nonequilibrium heating
of electrons in the femtosecond domain. Fast expansion of the electron gas results in a high
pressure that is exerted on the surfaces of the metal lattices, resulting in separations of the
lattices as the blasting force becomes excessive. In reality, femtosecond processes may be a
combination of both, with the latter being a mechanical process that does not attribute to the
phase change of the material.
The mechanical force generated by femtosecond lasers is counterintuitive in nature.
Unlike the continuous or long-pulse laser heating where damage by the thermal process
prevails, femtosecond lasers can produce high quality surfaces with minimal collateral
damage by the ways the mechanical forces are produced and their intrinsic effects on the
ultrafast deformation of the lattices.
With the fundamental understanding developed in the previous works (Tzou et al., 2005)
during the sub-picosecond transient, this work reinstates the effects of volumetric
expansion/contraction of the metal lattices, thermal wave effect for heat propagation in
phonons, and thermomechanical coupling through the temperature gradient across the metal
lattices.
These physical mechanisms follow the hot-electron blast developed in the femtosecond
domain, becoming highly activated shortly after the laser pulse. The thermomechanical model
incorporates all these effects is therefore able to describe the ultrafast thermomechanical
response covering approximately three orders of magnitude, from tens of femtoseconds
(where electron-to-phonon interactions dominate) to tens of picoseconds (where phonon-to-
phonon interactions take over).
While accommodating the temperature dependent thermal properties of electrons, the
thermophysical parameters that characterize the ultrafast transient are extracted in terms of
nondimensional groups. Their values shall be varied over one to two orders of magnitude in
studying their effects on the ultrafast heating and lattice deformation. The finite-difference
differential formulation shall be developed to tackle the nonlinearly coupled partial
differential equations (PDEs), and the formulation shall be extended to study the effect of
repetitive pulsing, as well as the potential effect of plasma shielding due to the surface
ionization induced by the UUL.
5.2. Ultrafast Deformation – Electron Blast
Due to the significant deviation of the lattice temperature from the reference temperature,
particularly beyond 10 picoseconds (threshold value of the relaxation time for phonons)
where the metal lattices become highly heated by the hot electrons, the linear theory of
thermoelasticity breaks down and nonlinear coupling between the thermal and mechanical
fields become intrinsic in ultrafast thermomechanics. With all these effects interwoven during
the ultrafast transient, the field equations governing the ultrafast response are extremely
complicated and nonlinear.
Ultrafast heating is highly nonequilibrium, necessitating separate considerations of
energy transport in electrons and phonons with the electron-phonon thermal coupling. The
energy equations for electrons and lattices are:
∂Te
Ce = ∇ ⋅ ( ke∇Te ) − G (Te − Tl ) + S (74)
∂t
∂Tl ⎛ Cl η ⎞ ∂
Cl +⎜ ⎟ (∇ ⋅ u) = −∇ ⋅ q′′l + G (Te − Tl ) (75)
∂t ⎝ κε ⎠ ∂t
where the heat flux for the lattice, q′′l , satisfies Eq. (27), τ the phonon-to-phonon relaxation
time, η the thermomechanical coupling factor, and κε the thermal expansion coefficient of
strain. Thermal relaxation of electrons is neglected since the electron-to-electron relation time
is only for a few femtoseconds. On the other hand, consumption of thermal energy in support
of the thermal expansion/contraction of the lattice volume has now been included in Eq. (75)
due to the high time-rate of change of the lattice volume ∂( ∇ ⋅ u )/∂t in the femtosecond
domain. This effect is important because during the fast transient in the femtosecond domain,
a volumetric strain ∇ ⋅ u of the order of nanometers would result in a strain rate ∂( ∇ ⋅ u )/∂t
on the order of 106. Under constant phonon properties, the heat flux vector ql′′ in phonons can
be eliminated from Eq. (75) by considering Eq. (27), resulting in
⎛ ∂T ∂ 2T ⎞ ⎡ ⎛ ∂Te ∂Tl ⎞ ⎤
kl ∇ 2T = Cl ⎜ l + τ 2l ⎟ − G ⎢(Te − Tl ) + τ ⎜ − ⎟ +
⎝ ∂t ∂t ⎠ ⎣ ⎝ ∂t ∂t ⎠ ⎥⎦
Cl η ⎡ ∂ ∂2 ⎤
⎢ ( ∇ ⋅ u ) + τ 2 (
∇ ⋅ u )⎥ (76)
κε ⎣ ∂t ∂t ⎦
A number of high-order derivatives result due to the high-rate responses, including the
lagging behavior describing the delayed response between the heat flux vector and the
temperature gradient in heat transport (Tzou, 1997). Consequently, special methods
developed in classical thermoelasticity assuming Fourier’s law heat diffusion, such as the
Goodier’s potential or Boussinesq-Papkovich functions (Boley and Weiner, 1960), no longer
hold.
Electrons can be viewed as a special type of gas surrounding the metal lattices (phonons)
(Tzou et al., 2005; Chen et al., 2006). When highly excited, the electron gas rapidly expands,
which could result in excessive pressure being exerted on the surfaces of the metal lattices.
The equations describing the motion of the metal lattices can be written as
∂ 2u
ρ 2 = ∇ ⋅ σ + ∇Pe (77)
∂t
where σ is Cauchy stress tensor expressed as
σ = λ(∇ ⋅ u)I + μ ⎡⎣(∇u) + (∇u)T ⎤⎦ − κ σ (Tl − T0 ) I (78)
where ρ is the mass density, u is the displacement vector of the metal lattice, I is unit tensor,
λ and μ are the Lamé constants, κσ is the thermal expansion coefficient of stress. Kinetic
pressure of the electron gas is represented by:
π2 N ⎛ Te ⎞ π2 Nk B
Pe = nk BTe with n = ⎜ ⎟ ⇒ Pe = ΛTe
2
with Λ = (79)
3 ⎝ TF ⎠ 3TF
with n representing the number density of electrons proportional to the electron temperature
(Te) and number density of atoms (N) and kB representing the Boltzmann constant. Neglecting
the quantum effect, Eq. (79) can be viewed as the equation of state for the electron gas, where
the kinetic pressure of electrons is proportional to the electron temperature squared.
Substituting Eq. (79) into Eq. (77) results in
∂ 2u
ρ = ∇ ⋅ σ + 2ΛTe∇Te , (80)
∂t 2
which displays a driving force due to the rapid expansion of hot electrons. Such a blasting
force exerting on the lattice surfaces is proportional to both the temperature and temperature
gradient of the electron gas, which will be significant in the early stage of electron heating as
described by Eq. (74). Combination of the Eqs. (77) and (78) gives
∂ 2u
ρ = (λ + μ)∇ ( ∇ ⋅ u ) + μ∇ 2u + ΛTe∇Te − κ σ∇Tl (81)
∂t 2
which includes explicit effects from both Te (through the hot-electron blast) and Tl
(thermomechanical coupling) on the motion of the metal lattices.
Equations (74), (76), and (81) were derived from the Boltzmann transport equation (Qiu
and Tien, 1993) and the nonequilibrium partition functions for electrons (Falkovsky and
Mishchenko, 1999). Rather than solving them in the microstructural frameworks for electrons
and phonons, the continuum formulation as shown absorbs the microstructural effects in
additional terms led by the new coefficients. They are the relaxation time (τ) and electron-
phonon coupling factor (G) in Eq. (76) and the electron pressure constant (Λ) in Eq. (81).
This formulation should hold as long as the physical ground for stress (and consequently the
elastic moduli in general) and temperature (hence the thermal conductivity or diffusivity)
holds, which however will breakdown as the conductor is composed of only a few lattices.
5.3. One-Dimensional Example
To demonstrate the capability of the above thermomechanical model, Tzou and Pfautsch
(2008) presented a finite difference solution of a one-dimensional example. Unique features
in ultrafast heating and deformation, as shown by Equations (74), (76), and (81), lie in the
thermomechanical response in time. Consideration of a one-dimensional case, in space,
therefore, is sufficient to develop the full resolution during the ultrafast transient. Considering
the lattice motion in a one-dimensional film with lateral constraints, the plane strain condition
applies and the one-dimensional forms of Eq. (74), (76), and (81) are
∂Te ∂ ⎛ ∂Te ⎞
CeTe = ke − G (Te − Tl ) + S (82)
∂t ∂x ⎜⎝ ∂x ⎟⎠
∂ 2Tl ⎛ ∂Tl ∂ 2Tl ⎞ ⎡ ⎛ ∂Te ∂Tl ⎞⎤

kl = Cl ⎜ + τ ⎟ − G ⎢(Te − Tl ) + τ ⎜ − ⎟⎥
∂x 2 ⎝ ∂t ∂t 2 ⎠ ⎣ ⎝ ∂t ∂t ⎠⎦
Cl η ⎛ ∂ 2u ∂ 3u ⎞
+ ⎜ + τ ⎟ (83)
κε ⎝ ∂t ∂x ∂t 2 ∂x ⎠
∂ 2u ∂ 2u ∂T ⎛ ∂T ⎞
ρ − E1 + E2 κσ l = 2Λ ⎜ Te e ⎟ (84)
∂t 2
∂x 2
∂x ⎝ ∂x ⎠
where
E (1 − ν) E
E1 = , E2 = . (85)
(1 + ν )(1 − 2ν) 1 − 2ν
The spatial variable x in Eqs. (82) – (84) describes the distance measured from x = 0, the
surface subject to the laser irradiation. The heat capacity and thermal conductivity of the
electrons gas are obtained from Eq. (16) and (17), respectively. The internal heat source in
Eq. (82) is obtained by Eq. (56). Heat loss from the film surfaces is negligible during the
femtosecond transient,
∂Te
= 0, ql′′ = 0 at x = 0 and x=l (86)
∂x
with L denoting the film thickness. The front and rear surface of the film is assumed strain-
free,
∂u
= 0 at x = 0 and x=l, (87)
∂x
whereas the film is assumed to be heated from a stationary state:
∂Tl ∂u
Te = T0 , Tl = T0 , = 0, and = 0 as t = t0 = −2t p (88)
∂t ∂t
The strain-free boundary conditions described in Eq. (87) are employed to tackle the
ultrafast response in a simpler situation. For stress-free conditions, Eq. (87) is replaced by
∂u ∂u ⎛ E2 ⎞
σ = E1 − E2 κεTl = 0, or = ⎜ ⎟ κ εTl at x = 0 and x=l (89)
∂x ∂x ⎝ E1 ⎠
which relate the surface strains to the surface temperatures of phonons.
Defining the following dimensionless variables and parameters:
x t Te (l ) u E /ρ t
ξ= , β = , θe ( l ) = , U= , Y = 0 2 , β0 = 0 ,
δ tp T0 ΛT0 t p / ρδ
2 2
(
(δ / t p ) tp )
⎛ 1− R ⎞ Gt p α 0t p C l k
S = 0.94 J ⎜ ⎟, H = , K = 2 , C = e0 , L = , α0 = e0
⎝ δCe 0T0 ⎠ Ce 0 δ Cl δ Ce 0
η τ E1t 2p Eκ
M= , Γ= , Y1 = 2 , Y2 = 2 ε , (90)
⎛ ρδ2 κε ⎞ tp ρδ ΛT0
⎜⎜ ⎟⎟
⎝ ΛT0t p
2
⎠
Eqs. (82) – (84) become
∂θe ∂ ⎛ θ ∂θ ⎞
θe = K ⎜ e e ⎟ − H ( θe − θl ) + S exp ⎡⎣ −ξ − 4 ln(2)β2 ⎤⎦ (91)
∂β ∂ξ ⎝ θl ∂ξ ⎠
∂ 2 θl ∂ 2 θl ∂θl ⎡ ∂ ⎤
KC = Γ + − HC ⎢( θe − θl ) + Γ ( θe − θl ) ⎥
∂ξ 2
∂β 2
∂β ⎣ ∂β ⎦
⎛ ∂ 2U ∂ 3U ⎞
+M⎜ +Γ 2 ⎟ (92)
⎝ ∂β∂ξ ∂β ∂ξ ⎠
∂ 2U ∂ 2U ∂θ ∂θ
− Y1 2 + Y2 l = 2θe e (93)
∂β 2
∂ξ ∂ξ ∂ξ
Ultrafast heating and deformation is thus characterized by 7 parameters: K

(nondimensional thermal diffusivity of electrons), H (nondimensional electron-phonon
coupling factor), S (nondimensional laser fluence), C (heat-capacity ratio), Γ (nondimensional
relaxation time of phonons), M (nondimensional group of thermomechanical coupling), Y1
(nondimensional elastic modulus), and Y2 (nondimensional thermoelastic modulus).
Equations (91) – (93) provide three equations to be solved for three unknowns, θe, θl, and U,
subjected to the initial and boundary conditions:
∂θl ∂U
θe = θl = 1, = 0, and = 0 as β = −2; (94)
∂β ∂β
∂θe ∂θl ∂U
= 0, = 0, = 0 at ξ = 0 and x = L. (95)
∂ξ ∂ξ ∂ξ
Numerical solutions are obtained for a gold film with the following thermal and optical
properties: ke0 = 315 W/mK, Ce0 = 2.1 × 104 J/m3K, τ = 10 ps, Cl = 2.5 × 106 J/m3K, g = 2.6 ×
1016 J/m3K, δ = 15.3 nm, R = 0.93, T0 = 300 K, and l = 0.02 μm (Tzou, 1997). Mechanical
properties for gold are E = 79 GPa, ν = 0.42, κε = 14.2 × 10-6, and Λ = 105 J/m3K. The laser
fluence is taken to be J = 732 J/m2. The nondimensional parameters are thus H ≅ 0.124, K ≅
6.41, C ≅ 8.4 × 10-3, Γ ≅ 100, M = 6.4915 × 10-5, Y1 = 4.46376 × 10-4, Y2 = 0.27156, S0 = 500,
and L = 1.307. These values shall be used throughout the numerical analyses unless stated
otherwise.
Figure 18 describes the straining patterns developed in the gold film as the parameter H
varies from 0 (no energy exchange between electrons and phonons) to 12.4. The hot electron
blast, 2θe (∂θe / ∂ξ) in Eq. (93), introduces compressive strains (U′ < 0) on the initial contact
(β = 0) with the film (near ξ = 0), as indicated in Figure 18(a). The peak of the compressive
strain is located between the characteristic lines of thermal and mechanical waves (Tzou et
al., 2005). After the initial compressive strain near the front surface, two tensile ripples
follow, in the direction of increasing time (β increases) or in the direction toward the interior
of the film (ξ increases). Another compressive ripple is resulted near the back surface of the
film (ξ > 0.8) at longer times (β > 20). Such alterations between compressive and tensile
strains prevail in the temporal and spatial directions, which become even more exaggerated at
longer times due to reflections of the displacement waves from the front (ξ = 0) and rear (ξ =
1.307) surfaces of the film. As the value of H increases, the peak values of compressive and
tensile strains slightly increase. All qualitative features of the straining patterns, however,
remain the same.
Figure 18. Straining patterns developed in the gold film (Tzou and Pfautsch, 2008).
CONCLUSION
Micro- and nanoscale heat transfer in femtosecond laser processing of metals is reviewed
in this chapter. The classical heat conduction theory based on Fourier’s law assumes that
thermal disturbance propagates with an infinite speed. As heat conduction is accomplished by
successive collision of the energy carriers (phonons or electrons), the prorogation of thermal
disturbance is always at a finite speed. To account for the finite propagation speed of thermal
wave, the Cattaneo-Vernotte thermal wave mode can be used. While the thermal wave model
assumes that the temperature gradient always precedes the heat flux, the dual-phase lag model
allows either the temperature gradient (cause) to precede heat flux vector (effect) or the heat
flux vector (cause) to precede the temperature gradient (effect). The nonequilibrium between
electrons and lattice has been observed experimentally and can be described by the classic
two-temperature model. Various two-step heating models can be employed to describe such
nonequilibrium phenomena between the electrons and lattices.
A semi-classical two-step heating model to investigate thermal transport in metals caused
by ultrashort laser heating is also discussed. The semiclassical two-step heating model
includes the conservation equations of number density, momentum and energy for electron
subsystem. The main difference between the semiclassical and the phenomenological two-
temperature models is that the former includes the effects of electron drifting, which could
result in significantly different electron and lattice temperature response from the latter for
higher-intensity and shorter-pulse laser heating. The volumetric laser heat source used in the
model was modified for accommodating the electron ballistic motion and the delayed rise
time of the electron temperature. The temperature fields in the electrons and lattice were
solved with the equations of the conservation of momentum and energy in electron subsystem
and the equation of heat conduction in the lattice subsystem. It was shown that for higher-
intensity and shorter-laser pulses, the semiclassical two-temperature model could result in
different thermal response than the phenomenal two-temperature model.
melting point and melting takes place. The liquid phase will be resolidified when the lattice is
cooled by conducting heat away. A fixed grid interfacial tracking method is presented to
solve kinetics controlled rapid melting and resolidification during ultrashort pulse laser
interaction with a free-standing metal film. Melting and resolidification were modeled by
considering the interfacial energy balance and nucleation dynamics. An iterative solution
procedure was employed to determine the elevated melting temperature and depressed
solidification temperature in the ultrafast phase-change processes. Nonlinear electron heat
capacity and temperature-dependent electron-lattice coupling factor on the rapid phase change
are also investigated. Ultrafast melting and resolidification of a submicron gold particle
subject to pico- to femtosecond laser pulse are studied.
At even shorter pulse width, femtosecond laser heating on metals produces a blasting
force from hot electrons in the sub-picosecond domain, which exerts on the metal lattices
along with the non-equilibrium heat flow. Physical phenomena of the hot-electron blast
induced by ultrafast, ultraintense lasers have been further explored to include the finite speed
of heat propagation in phonons, energy consumption in support of the volumetric
expansion/contraction of the metal lattices, thermomechanical coupling, and thermal
relaxation of phonons. Nonequilibrium heating emphasizes different temperatures of phonons
and electrons before they come to thermal equilibrium. Inertia effect during the lattice motion
and effect of volumetric expansion rate that have often been neglected in classical
thermoelasticity have been found important in ultrafast deformation.
ACKNOWLEDGEMENT
The work presented in this chapter is supported in part by the US National Science
Foundation under grant number CBET -0730143.
REFERENCES
Anisimov, S. I. Kapeliovich, B. L., Perel’man, T. L., 1974, “Electron Emission from Metal
Surface Exposed to Ultrashort Laser Pulses,” Sov. Phys. JETP 39, 375-377.
Ashcroft, N.W., Mermin, N.D., 1976, Solid State Physics, Saunders College, Philadelphia,
PA.
Baccarani, G., Wordeman, M.R., 1985, An investigation of steady-state velocity overshoot in
silicon, Solid-State Electron. 28, 407–416.
Beaman, J.J., Barlow, J.W., Bourell, D.L., Crawford, R.H., Marcus, H.L., and McAlea, K.P.,
1997, Solid Freeform Fabrication: A New Direction in Manufacturing, Kluwer,
Bordrecht.
Binet, B. and Lacroix, M., 2000, “Melting from Heat Sources Flush Mounted on a
Conducting Vertical Wall,” Int. J. Num. Meth. Heat and Fluid Flow 10, 286-306.
Boley, B.A., and Weiner, J.H., 1960, Theory of Thermal Stresses, Wiley, New York.
Cao, Y., and Faghri, A., 1990, “A Numerical Analysis of Phase Change Problem including
Natural Convection,” ASME J. Heat Transfer 112, 812-815.
Brorson, S.D., Fujimoto, J.G., and Ippen, E.P., 1987, “Femtosecond Electronic Heat Transfer
Dynamics in Thin Gold Film,” Phys. Rev. Lett., 59, 1962-1965.
Chen, J. K., and Beraun, J. E., 2001, “Numerical Study of Ultrashort Laser Pulse Interactions
with Metal Films,” Numerical Heat Transfer A: Applications 40, 1-20.
Chen, J. K., Latham, W.P., and Beraun, 2002, “Axisymmetric modeling of femtosecond pulse
laser heating on metal films,” Numer. Heat Transfer A 42, 1–17.
Chen, J. K. Latham, W. P., and Beraun, J. E., 2005, “The Role of Electron-Phonon Coupling
in Ultrafast Laser Heating,” J. Laser Applications 17, 63-68.
Chen, J. K., Beraun, J. E., Tzou, D. Y., 2006, “A Semiclassical Two-Temperature Model for
Ultrasfast Laser Heating,” Int. J. Heat and Mass Transfer 49, 307-316.
Das, S., Wohlert, M., Beaman, J.J., and Bourell, D.L., 1998, “Producing Metal Parts with
Selective Laser Sintering/Hot Isostatic Pressing,” Journal of Materials 50, 17-20.
Dyer, P.E., Maswadi, S.M., and Walton, C.D., 2003, “VUV F2 Laser Ablation of Sodium
Chloride,” Appl. Phys. A 76, 817-822.
Eesley, G. L., 1986, “Generation of Non-equilibrium Electron and Lattice Temperatures in
Copper by Picosecond Laser Pulses,” Physical Review B 33, 2144-2155.
Elsayed-Ali, H. E., Norris, T. B., Pessot, M. A., and Mourou, G. A., 1987, “Time-Resolved
Observation of Electron-Phonon Relaxation in Copper,” Physical Review Letters 58,
1212-1215.
Faghri, A., and Zhang, Y., 2006, Transport Phenomena in Multiphase Systems, Elsevier,
Burlington, MA.
Falkovsky, L.A., and Mishchenko, E.G., 1999, “Electron-Lattice Kinetics of Metals Heated
Ultrashort Laser Pulse,” J. Exp. Theoret. Phys. 88, 84-88.
Fan, C. H., Sun, J., and Longtin, J. P., 2002, “Plasma Absorption of Femtosecond Laser
Pulses in Dielectrics,” ASME J. Heat Transfer 124, 275-283.
Fischer, P., Karapatis, N., Romano, V., and Weber, H.P., 2002, “A Model for Interaction of
Near-Infrared Laser Pulse with Metal Powders in Selective Laser Sintering,” Applied
Physics, A. 74, 467-474.
Griffith, M. L., Ensz, M. T., and Reckaway D. E., 2003, “Femtosecond Laser Machining of
Steel,” Proceedings of SPIE 4977, 118-122.
Grigoropoulos, C.P., and Ye, M., 2000, “Numerical Methods in Microscale Heat Transfer:
Modeling of Phase-Change and Laser Interactions with Materials,” Advances of
Numerical Heat Transfer, 2, eds. W.J. Minkowycz and E.M. Sparrow, 227-257, Taylor
and Francis, London, Great Britain.
Hirayama, Y., and Obara M., 2003, “Heat Affected Zone of Metals Ablated with
Femtosecond Laser Pulses,” Proceedings of SPIE 4977, 417-425.
Hohlfeld, J., Wellershoff, S. S., Gudde, J., Conrad, U., Jahnke, V., and Matthias, E., 2000,
“Electron and Lattice Dynamics Following Optical Excitation of Metals,” Chemical
Physics 251, 237-258.
Hopkins, J. M. and Sibbett, W., 2000, “Ultrashort-Pulse Lasers: Big Payoffs in a Flash,” Sci.
Amer. 283, 72-79.
Hsiao, J. S., 1984, “An Efficient Algorithm for Finite Difference Analysis of Heat Transfer
with Melting and Solidification,” ASME paper No. 84-WA/HT-42.
Ibrahim, W.M.G., Elsayed-Ali, H.E., Bonner, C.E., Shinn, M., 2004, Ultrafast Investigation
of Electron Dynamics in Multi-layer Metals, Int. J. Heat Mass Transfer 47, 2261–2268.
Jiang, L., and Tsai, H.L., 2005, Energy transport and material removal in wide bandgap
materials by a femtosecond laser pulse, Int. J. Heat Mass Transfer 48, 487–499.
Kittle, C., 1967, Introduction to Solid State Physics, John Wiley, New York, NY.
Klein-Wiele, J., Bekesi J., Ihlemann, J., and Simon, P., 2003, “Nanofabrication of Solid
Materials with Ultraviolet Femtosecond Pulses,” Proceedings of SPIE 5399, 139-46.
Klemens, P. G., and Williams, R. K., 1986, “Thermal Conductivity of Metals and Alloys,”
Int. Metals Review 31, 197-215.
Konrad, C., Zhang, Y., and Shi, Y., 2007, “Melting and Resolidification of Subcooled Metal
Powder Particle Subjected to Nanosecond Laser Heating,” International Journal of Heat
and Mass Transfer 50, 2236-2245.
Kruer, W.L., 1988, The Physics of Laser Plasma Interactions, Addison-Wesley, Redwood
City, CA.
Kuo, L. S., and Qiu, T., 1996, “Microscale Energy Transfer During Picosecond Laser Melting
of Metal Films,” ASME National Heat Transfer Conference, 1, 149-157.
Morgan, K., 1981, “A Numerical Analysis of Freezing and Melting with Convection,” Comp.
Meth. App. Eng. 28, 275-284.
Momma, C., Chichkov, B. N., Nolte, S. Alvensleben, F. V., Tunnermann, A., Welling, H.,
and Wellegehausen, B., 1996, “Short-Pulse Laser Ablation of Solid Targets,” Optics
Comm. 129, 134-142.
Ozisik, M.N., and Tzou, D.Y., 1992, “On the Wave Theory in Heat Conduction,”
Fundamental Issues in Small Scale Heat Transfer, ASME HTD-Vol 227, 13-27.
Patankar, S. V., 1980, Numerical Heat Transfer and Fluid Flow, McGraw-Hill, New York.
Qiu, T. Q., and Tien, C. L., 1993, “Heat Transfer Mechanisms During Short-Pulse Laser
Heating of Metals,” Journal of Heat Transfer 115, 835-837.
Rethfield, B. , Sokolowski-Tinten, K., and von der Linde, D., 2002, “Ultrafast Thermal
Melting of Laser-Excited Solids by Homogeneous Nucleation,” Phys Rev (B) 65,
092103:1-092103:4.
Riemann, I., Fischer, P., and Koenig, K., 2004, “Photodynamic therapy and knocking out of
single tumor cells by multiphoton excitation processes,” SPIE 5462, 103-109.
Sasaguchi, K., Ishihara, A. and Zhang, H., 1996, “Numerical Study on Utilization of Melting
of Phase Change Material for Cooling of a Heated Surface at a Constant Rate,”
Numerical Heat Transfer, Part A 29, 19-31.
Snowden, C.M., 1986, Introduction to Semiconductor Device Modeling, World Science,
Singapore.
Strickland, D. and Mourou, G., 1985, “Compression of Amplified Chirped Optical Pulses,”
Optics Comm. 56, 219-221.
Stuart, B.C., Feit, M.D., Herman, S., Rubenchik, A.M., Shore, B.W., Peery, M.D., 1996,
Optical ablation by high-power short-pulse lasers, J. Opt. Soc. Am. B 13, 459–468.
Sun, C.K., Vallee, F. Acioli, L., Ippen, E.P., Fujimoto, J.G., 1993, Femtosecond investigation
of electron thermalization in gold, Phys. Rev. B 48, 12365–12368.
Swaminathan, C. R., and Voller, V. R., 1993, “On the Enthalpy Method,” Int. J. Heat Fluid
Flow 3, 233-234.
Tolochko, N. K., Arshinov, M. K., Gusarov, A.V., Titov,V.I., Laou, T., and Froyen, L.,
2003a, “Mechanism of selective laser sintering and heat transfer in Ti powder”, Rapid
Prototyping Journal 9, 314-326.
Tolochko, N., Mozzharov, S., Laoui, T., Froyen, L., 2003b, “Selective laser sintering of
single- and two-component metal powders”, Rapid Prototyping Journal 9, 68-78.
Tzou, D. Y., Macro- to Microscale Heat Transfer, Taylor and Francis, Washington, DC,
1997.
Tzou, D. Y., “Computational Techniques for Microscale Heat Transfer,” in Minkowycz, W.
J., Sparrow, E. M., and Murthy, J. Y., Eds., Handbook of Numerical Heat Transfer, 2nd
ed., John Wiley and Sons, Hoboken, NJ, 2006.
Tzou, D.Y., Chen, J.K., and Beraun, J.E., 2005, “Recent Development of Ultrafast
Thermoelasticity,” J. Thermal Stresses 28, 563-594.
Tzou, D.Y., and Li, J., 1993, “Thermal Wave Emanating from a Fast-Moving Heat Source
with a Finite Dimension,” ASME Journal of Heat Transfer 115, 526-532.
Tzou, D.Y., and Pfautsch, E. J., 2008, “Ultrafast Heating and Thermomechanical Coupling
Induced by Femtosecond Lasers,” J. Eng. Math. 61, 231-247.

Voller, V. R., 1997, “An Overview of Numerical Methods for Solving Phase Change
Problems,” in W. J. Minkowycz and E. M. Sparrow (Eds.),” Advances in Numerical Heat
Transfer, 1, Taylor and Francis, Basingstoke.
Von Der Linde, D., Fabricius, N., Danielzik, B., and Bonkhofer, T., 1987, “Solid Phase
Superheating During Picoseond Laser Melting,” Mat. Res. Symp. Proc. 74, 103-108.
Wang, G. X., Prasad, V., 2000, “Microscale Heat and Mass Transfer and Non-equilibrium
Phase Change in Rapid Solidification,” Mater. Sci. Eng. A292, 142-148.
Wellershoff, S., Hohlfeld, J., Güdde, J., Matthias, E., 1999, “The Role of Electron-Phonon
Coupling in Femtosecond Laser Damage of Metals,” Appl. Phys. A: Mater. Sci. Process.
69(Suppl.), 99–107.
Xiao, B., and Zhang, Y., 2006, “Partial Melting and Resolidification of Metal Powder in
Selective Laser Sintering,” AIAA J. Thermophysics and Heat Transfer 20, 439-448.
Xiao, B., and Zhang, Y., 2007, “Analysis of Partial Melting in a Metal Powder Bed with
Constant Heat Flux,” Heat Transfer Engineering 28, 472-483.
Xu, X., and Song, K.H., 2000, “Phase Change Phenomena during High Power Laser-
Materials Interaction,” Mater. Sci. Eng. A292, 162 - 168.
Yamanouchi, K., Chin, S.L., Agostini, P., and Ferrante, G., eds., Progress in Ultrafast Intense
Laser Science, Vol. I, II, and III, Berlin, Springer, 2006 (I), 2007 (II and III).
Zhang, Y., and Chen, J. K., 2007, “Melting and Resolidification of Gold Film Irradiated by
Nano- to Femtosecond Lasers,” Applied Physics A 88, 289-297.
Zhang, Y., and Chen, J. K., 2008a, “An Interfacial Tracking Method for Ultrashort Pulse
Laser Melting and Resolidification of a Thin Metal Film,” ASME J. Heat Transfer 130,
062401.
Zhang, Y., and Chen, J. K., 2008b, “Ultrafast Melting and Resolidification of Gold Particle
Irradiated by Pico- to Femtosecond Lasers,” Journal of Applied Physics 104, 054910.
Chapter 6
HIGH POWER FEMTOSECOND LASER MACHINING OF

METALS IN AMBIENT MEDIUM
S. R. Vatsya1, Chengde Li2 and S. K. Nikumb1,∗

1
Industrial Materials Institute, National Research Council of Canada
London, Ontario, Canada, N6G 4X8
2
Department of Applied Physics, School of Science,
Beijing Institute of Technology, Beijing, 100081, China
ABSTRACT
High power ultra-short-pulse lasers offer significant advantages over their long-pulse
counterparts for material processing. Since the duration of the pulse is too short to
establish the temperature equilibrium between the electron and lattice subsystems, little
melt is produced during processing. Ablation is caused by direct removal of the material.
Consequently, the machined surface features conform closely to the intensity profile of
the laser beam. The original laser beam suffers significant distortions during propagation
through a medium such as air. Intensities enhanced further by the self-focusing of the
beam are sufficiently high to cause an optical breakdown of the media, generating
plasma. Scattering effects of plasma deform the laser beam profile. Distortion to the
optical beam profile can be reduced by conducting femtosecond laser machining in
vacuum or the gaseous media less immune to ionization, which restricts its application in
the production environment. Removal of material also generates a plasma plume with the
additional scattering effects. For certain energy range, the competing self-focusing and
gas plasma plume supplemented with the material plasma can combine to cause plasma
filamentation and the distortion causing scattering effects balance out yielding an
adequately shaped beam profile. Filament of hot plasma also acts as an energy source
with similar properties. The energy at which filamentation occurs can be determined by
calculating the optical beam profile, which can be controlled experimentally thereby
yielding a cleaner fabricated surface. Due to significant differences in the processes
involved, this type of conformity is less pronounced in case of the dielectrics compared to
the metallic materials. This article presents the results on the applications of such
∗
Corresponding author. Tel: +1 519 430 7058; Fax: +1 519 430 7064. E-mail address:Suwas.nikumb@nrc.ca
208 S. R. Vatsya, Chengde Li and S. K. Nikumb
techniques in high power femtosecond laser machining of metals. Some discussion on

non-metallic materials is also included.
1. INTRODUCTION
Most manufacturing applications involving laser machining processes require high
quality precise features, satisfactory material removal rate, and minimal post-processing
procedures. Material ablation with longer-pulse lasers, in the nanosecond and picosecond
ranges, proceeds with the creation of a melt pool. As more energy is absorbed, some of the
molten material evaporates and the remainder re-solidifies. This process results in
deformations due to the fluid flow processes and variations in the flow conditions during re-
solidification [1-6]. Also, due to longer pulse duration, thermal conduction effects are
significant, which influence the material properties and the features outside the machined area
[1,2]. Ultrashort pulse duration in the femtosecond range is far below the time needed for
thermal equilibrium to establish in the electron and lattice subsystems. Therefore thermal
diffusion tends to be minimal, although not absent [7]. With short pulses photo-ionization and
impact ionization create an imbalance of the electromagnetic forces leading to the material
removal. Consequently, undesired thermal effects are greatly reduced and a cleaner surface
structure can be obtained [8-10]. For these reasons, ultrashort pulse laser processing of
materials is becoming a preferred technique for various machining applications such as hole
drilling, cutting, fabrication of complex patterns and property modification [9,10].
Femtosecond laser processing of glass, polymers and other transparent materials, e.g.,
fabrication of wave-guides by changing the refractive index of glass, has been successfully
demonstrated [8,11-14]. In addition, experiments with machined features significantly smaller
than the diffraction limit, i.e., sub-micron features inside transparent bulk materials [8] and
nanometer scale features on the surface of thin metal films [13] by taking advantage of the
threshold effect have been successfully conducted [14].
There exists an energy threshold below which there is a high probability of re-absorption
of the ionization electrons by the lattice. At such energies, little or no ablation takes place as
there is no net repulsive force in the material. Even above the threshold energy, the ablated
material may not have sufficient energy to escape and may re-settle on the surface generating
dross, debris and deformation. Re-absorption and agglomeration due to re-settling of the
removed material also transfers energy to the lattice generating heat and subsequent melt.
High powers are also needed for efficiency in the manufacturing environment. For these
reasons, applications of low power, ultrashort pulse lasers in the production environment
remained limited. With the rapid development of compact, stable, regeneratively amplified
Ti:Sapphire lasers in recent years, the applications of high power femtosecond lasers in the
industrial sector are on the increase.
While ultrashort pulse high power lasers offer significant advantages for processing of
materials, some difficulties must be overcome in order to glean their benefits fully. Optical
radiation propagating through gaseous medium experiences self-focusing, Kerr effect,
resulting mainly from the nonlinearity in the propagation property. High energy enhanced
further by self-focusing is sufficient to ionize the gases with low ionization potentials such as
air, generating the gas plasma. Energies insufficient to ionize the gas are too low to ablate the
material with satisfactory machining results. If the density of the plasma generated by gas
High Power Femtosecond Laser Machining of Metals in Ambient Medium 209
ionization exceeds certain critical value in the fabrication region, proportionate photon-
plasma scattering effects diffuse and distort the original beam profile. In case of the air, at
moderate energies the plasma density, 2 x 1019 cm−3, is lower than the critical density for 775
nm radiation, which is about 2.5 x 1021 cm−3. Shock waves generated by a rapid material
ablation supplies energy to the surrounding medium to transform it into ionized plasma. Up to
two orders of magnitude increase in the plasma density induced by CO2 laser beams observed
in various gases has been attributed to the shock waves [15]. For the normal experimental
parameters in high power femtosecond laser fabrication, plasma density in excess of the
critical density is expected and observed in the vicinity of the focal point, due to direct
ionization of air and the shock waves. At higher energies, these effects are enhanced further.
Photon-plasma scattering effects defocus the beam neutralizing the self-focusing effect to
some extent but at the cost of distorting the original beam profile [16,17].
Due to direct ablation, the machined surface profile conforms with the beam intensity
profile [17-24]. Such a correspondence is observed to a significant extent in fabrication with
the nanosecond pulse lasers as well [20]. Primarily due to different processes responsible,
such a correspondence is expected to a larger extent in case of the metallic materials
compared with the dielectrics, although not absent. Calculated beam profile in an ionizable
medium varies rapidly in the vicinity of the focal point along the beam as well as
perpendicular to it, with corresponding impact on the machined surface features [17]. While a
low energy plasma distorts the beam without compensating effects, high-energy particles in
plasma, which participate in the material ablation, have a positive effect in achieving better
quality machined surfaces [21]. Therefore, to obtain a machined surface with precisely
defined features, a sufficiently high energy optical beam with an adequately shaped intensity
profile is required.
To reduce the distortions to the beam profile, femtosecond laser material processing is
often conducted in vacuum [22]. Alternatively, a gaseous medium with high ionization
potential may be used to reduce the distortion. Ionization potential of air is about 14.9 eV.
Among the feasible media, Helium has the highest ionization potential of 24.6 eV, rendering
it the most suitable among the gases [23]. Requirement to conduct machining in vacuum or in
Helium as the medium limits its application in the production environment. Alternatively, if
the profile of a part of the beam is adequately shaped away from the focal point, which is
possible, it can be used for fabrication. This technique has been successfully used to improve
the quality of the machined surfaces [24].
As described above, high power short pulse laser fabrication of materials involves
complex mechanisms with various competing phenomena. In some experimental situations,
these competing effects can neutralize the adverse effects of each other almost exactly
creating a long filament maintaining a steady profile of the optical beam, although not the
same as in vacuum and the balance is fragile. Eventually, the filamentation ends where the
energy loss from ionization, inherent radiation, self-phase modulation, and supercontinuum
generation breaks up the process. The observations of longer light filaments in air created
with collimated high power ultrashort laser pulses [25] generated considerable interest in the
studies of ultra-intense laser pulse propagation in the transparent media [26–33]. For the case
of converging geometry, recent experiments and simulations have focused on femtosecond
pulses propagating in water [34] and to applications in laser processing of metals [35,36]. The
energy range can be calculated to create conditions for filamentation, which determines the
necessary experimental parameters. While the balance is delicate, it is robust enough for its
exploitation to obtain a high energy optical beam with adequately shaped intensity profile.
Machining of metals under such experimental conditions yields surfaces with significantly
improved features [18,36].
In the present article, we describe the modeling and methods to calculate the intensity
profile of an optical beam propagating in an ionizable medium. A method to calculate the
fabricated surface features from the knowledge of the intensity profile is also developed.
Machining results for various metals are presented and compared with the calculated values.
Some discussion of the non-metallic materials is included.
2. LASER-MATERIAL INTERACTION
Due to a faster energy deposition, ablation process with femtosecond laser pulses, in
general differs from the long-pulse laser ablation. Both, the properties of the laser pulse and
the material determine the ablation process. In this section, we briefly describe the basic
processes involved in the ablation of metals, semi-metals, semi-conductors and dielectrics, by
the laser pulses of varying widths.
2.1 Metals
Metals are characterized by an abundance of the conduction band electrons. The electrons
absorb the laser energy due to inverse bremsstrahlung phenomenon or Joule heating. The
absorption is followed by a fast energy relaxation within the electronic subsystem. Energy
transfer to the lattice takes place by diffusion, collisions and the electron-phonon coupling.
Relaxation time lies in the picosecond range. Therefore, in case of the pulse-widths in
nanosecond range, the temperature equilibrium is established during the initial phases of the
pulse. For the remainder of the pulse, the energy in excess of the diffused heat, melts and
evaporates the material. The surface profile is determined by the vapor pressure [1-4] and the
pressure and flow in the melt pool [5,6], which yields a rough ablated surface. For pulses in
the picosecond ranges, these effects are somewhat diminished but the ablation is still quite
adequately described by the melting and evaporation processes.
For the ultrashort pulse widths, in the femtosecond range, although the heat diffusion into
the lattice is not completely absent, its effect is diminished considerably. In this pulse-width
range, the electron-phonon coupling effects become significant [4]. Since the relaxation time
is too long, no meaningful temperature can be assigned, particularly to the lattice subsystem.
Thus, the ablation process in the most part is described by the energy absorbed by the free
electrons, which is then transferred to the lattice by collisions, breaking the molecular and
atomic bonds. Finally, the material is expelled by the electro-magnetic repulsive forces. Since
little melt is present and most of the material is removed directly, the surface closely
conforms to the beam shape. If the intensity profile of the beam is adequately shaped, a
cleaner cut results [18]. If the energy is sufficiently high, the ablated plasma has enough
energy to escape reducing the debris and dross. Energy is transferred to the lattice by the
phonons. Some energy is transferred to the lattice by its collisions with the plasma, which
continues even after the pulse is gone. This yields a small amount of melt, resulting in some
roughness of the surface and formation of debris [2,4]. However, an adequately shaped
plasma plume, which results from filamentation, reduces these effects. At still higher
energies, the melt is created during the initial phases of the pulse. Further energy transfer
from the pulse and subsequent energy transfer by the collisions with plasma increases the
melt and evaporation conditions, rendering the processes similar to the long pulse ablation.
Therefore cleaner fabricated features can be obtained with the energy in an intermediate,
filamentation range.
2.2. Semi Metals and Semi Conductors
Semi-metals and semi-conductors are characterized by low density of free electrons and
small band-width. As a consequence, one photon collision is sufficient to ionize a molecule or
the atom by lifting an electron from the bound state to the conduction band or even removing
it. Thus, during the initial phases of a pulse, significant number of electrons are generated in
the conduction band for the material to behave as a metal. Therefore, the ablation process for
these materials is described mostly by the processes similar to the metals, more so for the
semi-metals than for the semiconductors. Major difference results from a diminished rate of
conduction, which does not change the basic processes. The results for the metals describe the
response of the semi-metals and semi-conductors to a laser pulse to a significant extent.
2.3. Dielectrics
Ablation process in the dielectrics differs from the metals and semiconductors in a
fundamental way. The dielectrics are characterized by an almost complete absence of the free
electrons and a large electronic band gap. Therefore, photo-ionization requires multi-photon
absorption, which is a low probability phenomenon. Consequently, ablation cannot rely on
direct bond breaking by the photons. Instead, the energy is absorbed by the few free electrons
available by the Joule heating, which is diffused into the electron subsystem by collisions.
When some of the free electrons acquire energy equal to or greater than the band gap, the
lattice subsystem can be ionized by the impact ionization. The seed electrons are sometimes
provided by added impurities; otherwise the initial multi-photon ionization remains their only
source. These processes take place in the metallic materials and semiconductors also but their
contribution differs significantly from that in case of the dielectrics. After sufficient
conduction band electrons have been generated, the energy is transferred by similar processes
as in the metallic materials and semiconductors. For the longer pulses, the photo-energy
absorbed by the conduction band electrons is transferred to the lattice in the form of heat,
which can melt the material and the dielectric acquires somewhat metal-like properties.
Energy transfer by the collisions still continues. For shorter pulses, no thermal equilibrium
establishes for the duration of the pulse. After the photo-energy transfer ceases, the energy
diffusion among the conduction band electrons and to the lattice by impact continues. During
this period, thermal equilibrium can develop and heat generated in the lattice. During the
ablation phase, the energy is transferred to the lattice by the phonons. Due to the diffusion of
energy within the electronic subsystem and then to the lattice, ablation process continues after
the pulse no longer transfers the energy to the free electrons. In fact, most of the material
removal in the dielectrics occurs during the post-pulse period [37-39].
For shorter pulses, the energized phonons transfer energy to the lattice. This can cause
damage to the material directly and by energizing the electrons in the bound states, which in
turn are removed by impact ionization. Damage to the dielectrics is lesser for the short pulses
compared to the long pulses, as is the case with the metals and semiconductors. However,
damage has been observed in the dielectrics for all pulse-widths. Most of the damage to the
dielectrics for short pulses results from high density of the free electrons. Avalanche effect
can generate the free electrons at an enormous rate. As their number density reaches the
threshold, which is at about 1021 per unit volume, they are known to cause melt and other
damage, which results from the impact ionization and the phonon energy. Melt results from
the fact that at such high density of free electrons, the dielectric behaves as metal. At these
energies, ablation takes place mostly by melting and vaporization.
As described above, two mechanisms are involved in the ablation of the dielectrics by
short pulse lasers: Coulomb explosion or gentle ablation [40-43] and thermal vaporization or
strong ablation [44-46]. Both mechanisms coexist to some extent in all short pulse ablation of
the dielectrics. During the initial stages, Coulomb explosion dominates and after reaching the
critical density, strong ablation dominates. Thus, a cleaner fabricated surface results at the
fluences in an intermediate range, which is high enough for an effective ablation but
sufficiently low so that the critical density is not reached and strong ablation does not take
place.
3. OPTICAL BEAM PROPAGATION

3.1 Modeling and Method
The amplitude A of an optical beam propagating in a gaseous medium satisfies a

nonlinear Schrödinger-type equation [16,17,23,32, 47-49]:
∂A 1
i + ∇ r′ 2A + V ( I ) A = 0 , (1)
∂z ′ 2k
with the intensity given by I =| A |2 ; ∇ r′ is the Laplacian in the plane r′ = ( x′, y′)
2
perpendicular to the axis of the beam z ′ ; k = (2π / λ ) is the wave number with λ being the
wavelength; and
⎡ 2π e 2 N e ( I ) ⎤
V ( I ) = ⎢ n2 kI − ⎥ (2)
⎣ kme c 2 ⎦
is the potential term, where the first term describes the self-focusing effect in terms of the
Kerr coefficient n2 ; the second term describes the generation and effect of plasma resulting
from the interaction of high intensity radiation with the material in terms of the electron
charge e , electron mass me and the speed of light c in the gas. Free electron density N e
induced by the laser pulse by field ionization is determined by the rate equation:
∂N e ( I (t ))
= N 0 Rion ( I (t )) , (3)
∂t
where N 0 is the number density of the neutral air molecules; Rion is the ionization rate and
t is the retarded time given by t = t ′ − z ′ / c with t ′ being the time.
The derivations of Eq. (1) are based on a variety of arguments that include the reasoning
parallel to Feynman’s path integral formulation of quantum mechanics [50,51], Maxwell’s
equations in the slowly varying envelope approximation, modification of the paraxial
equation, and their combinations, which can then be related to the path-integral approach by
expressing the solutions in terms of the Green’s function [52].
Among the methods available to solve Eq. (1), finite difference and the procedures based
on the Fourier transforms have frequently been used. Although these and the other methods
have been successfully applied, they suffer from some undesirable features as follows. All of
the methods experience difficulties when applied to the case of a rapidly converging or
diverging beam. In the case of a focused beam, the beam radius of the order of millimeters at
the lens may converge on to the radius of the order of micrometers at the focal point. The
resulting difficulty is overcome at times by questionable or cumbersome adjustments. In
addition, the finite difference schemes [17,32,47,48] accumulate significant round offs. Finite
element methods suffer from similar difficulties. Evaluation of the Fourier transforms by the
fast Fourier transform ( FFT ) routines is quite efficient as well as convenient [16,23,49] but
the difficulty of managing a rapidly converging or diverging beam remains. The procedure
described below eliminates the problem of a rapidly changing beam radius and improves the
efficiency and accuracy of the methods based on the Fourier transforms further [51].
It is more convenient to consider the dimensionless form
∂A 1
i + ∇ r 2 A + ω0 V ( I ) A = 0 , (4)
∂z 2
of Eq. (1), resulting from the transformations r = r′ k / ω0 = ( x, y ) and z = z ′ / ω0 , where

ω0 is the beam radius at some convenient location. For propagation in vacuum, the
perturbation term drops out reducing Eq. (4) to the exactly solvable Helmholz equation, with
the solution A0 given by
(ω 2f + 2i[ z0 − z f ]) ⎡ | r |2 ⎤
A0 (r; z ) = exp ⎢ − ⎥, (5)
(ω 2f + 2i[ z − z f ]) ⎢⎣ (ω f + 2i[ z − z f ]) ⎥⎦
2
where ω f is the beam radius at the focal point z f with respect to the transformed units and
the solution is normalized so that A0 (0; z0 ) = 1 . In the experimental settings, usually a
collimated beam is focused at a desired location with a convex lens. It is usually convenient
to let z0 be the lens location.
With respect to the variables ρ = ( ρ x , ρ y ) and ξ defined by
1 ⎡ 2 (z − z f ) ⎤
ρ = r / ω ( z ) and ξ = tan −1 ⎢ ⎥,
2 ⎢⎣ ω 2f ⎦⎥
and the representation
A = cos(2ξ ) exp ⎡⎣i ρ 2 tan(2ξ ) ⎤⎦ u , (6)
u satisfies the equation
∂u 1
i = − ∇ ρ 2u + V u , (7)
∂ξ 2
where V = − (2 ρ 2
+ ω0ω 2 (ξ )V ) . For V ( I ) = 0 , the solution u 0 is given by
exp [2i (ξ0 − ξ )]

u0 (ρ ,ξ ) = exp [ − ρ 2 ] . (8)
cos (2ξ 0 )
ρ
Eq. (7) is similar to Eq. (4) with a basic difference: Parallel rays with respect to
conform to the characteristic curves of the unperturbed equation determined by A0 .
Consequently, an unperturbed Gaussian beam with varying width with respect to (r; z )
coordinate system is collimated with respect to ( ρ ; ξ ) and the corresponding solution given
by Eq. (8) is a soliton. For mild perturbations, as is the case presently, the deviation from a
collimated beam is not significant. Consequently, the adjustments required to solve Eq. (4),
e.g., a variable mesh, are not needed to solve Eq. (7). Due to such simplifying features,
solving Eq. (7) for u is more convenient. The solution A of Eq. (4) can then be obtained by
substituting for u in Eq. (6).
Let A0 be the self-adjoint realization of − ∇ ρ / 2 and let A = ( A0 + V ) . Then Eq.
2
(7) is exactly solved by
u (ξ + ε ) = exp[−iAε ] u (ξ )
for all ε . The solution accurate up to the second order in ε is given by [49]
u (ξ + 2ε ) = e − iA0ε e −2iV ε e − iA0ε u (ξ ) = P0 (ε )e −2iV ε P0 (ε ) . (9)
Thus the problem reduces to an evaluation of the two-dimensional propagator P0 (ε ) ,

expressible as [51]:
∞ ∞
1 ⎡ i 2⎤ ⎡ i 2⎤
( P0 (ε ) g(ξ )) (ρx , ρ y ;ξ + ε ) = ∫ d ρ′y exp ⎢⎣ 2ε (ρ y − ρ′y ) ⎥⎦ −∞∫ d ρx′ exp ⎢⎣ 2ε (ρx − ρx′ ) ⎥⎦ g(ρx′ , ρ′y ;ξ )
2π iε −∞
= [ p0 (ε )]2 g (ξ ),
(10)
which can be obtained from the one-dimensional propagator p0 (ε ) given by
∞
1 ⎡ i ⎤
( p0 (ε ) g (ξ ) ) ( ρ x ; ξ + ε ) = ∫ d ρ ′ exp ⎢⎣ 2ε ( ρ
x x − ρ x′ ) 2 ⎥ g ( ρ x′ ; ξ ) . (11)
2π iε −∞ ⎦
Eq. (11) can be obtained in the framework of Feynman’s path integral formulation of
Quantum Mechanics or by an exact evaluation of one of the Fourier transforms [51].
It is clear that
( p0 (ε ) g (ξ ) ) ( ρ x ; ξ + ε ) = ( M (ε ) [ p′ (ε )] M (ε ) g (ξ ) ) ( ρ ;ξ + ε ) ,
0 x
(12)
where M (ε ) is the operation of multiplication by exp[i ρ x /(2ε )] and p0′ (ε ) is the

2
Fourier transform:
∞
1 ⎡ i ⎤
f (ρx ;ξ + ε ) = ( p0′ (ε ) g (ξ ) ) ( ρ x ; ξ + ε ) = ∫ d ρ ′ exp ⎢⎣− ρ x ρ x′ ⎥ g ( ρ x′ ; ξ ) . (13)
2π iε −∞
x
ε ⎦
Quite accurate values of the Fourier transforms can be obtained efficiently by the FFT
schemes. The resulting procedure is convenient if the sets of grid points for the functions g
and f are reciprocals of each other. To determine the evolution of the solutions for the case
presently under consideration, it is desirable to have the same input and output grid structures
with respect to the space variable, since the solution is advanced indefinitely. This is possible
but cumbersome and compromises accuracy, which is therefore not preferred. Major
complicating feature is the rapidly oscillating integrand resulting from a small value of ε ,
which is desirable for accuracy. It is shown below that this disadvantage can be exploited by
constructing a suitable grid structure reducing the effort required to evaluate f in Eq. (13).
The integral in Eq. (13) can be approximated by restricting the interval of integration to
[−a, a] with a sufficiently large but finite value of a . The value of the integrand outside this
interval is assumed to be negligible. By the mid-point integration with equi-spaced grid points
{a ( 2n / N − 1)}n =0 , the integral is further approximated by
N
( p0′ (ε ) g (ξ ) ) ⎛⎜ a (2 ⎞ ⎡ ⎤
N
n 2a m
⎝ N
− 1); ξ + ε ⎟ =
⎠ N 2π iε
∑ g ⎣⎢ a(2 N − 1); ξ ⎦⎥
m =1
⎡ 4a 2 ⎛ nm n + m 1 ⎞ ⎤
× exp ⎢ − i⎜ 2 − + ⎟⎥
⎣ ε ⎝N 2N 4 ⎠⎦
n = 1, 2,...., ( N − 1);
f ( a; ξ + ε ) = f ( − a; ξ + ε ) = 0.
(14)
Let a = π N ε / 2 . Since N is typically a large integer and ε is a small number, this

condition yields a moderate value of a , which is commensurate with the decay properties of
a large class of solutions. Substitution from Eq. (14) into Eq. (12) yields
⎡π i
( p0 (ε )g(ξ )) ⎛⎜[a(2 ⎞ ⎤ N
n 2a m
−1)];ξ + ε ⎟ = exp ⎢ (n −1)2 ⎥ ∑ g[a(2 −1);ξ ]
⎝ N ⎠ N 2π iε ⎣N ⎦ m=1 N
⎡ (n −1)(m −1) ⎤ ⎡π i ⎤
× exp ⎢−2π i ⎥ exp ⎢ (m −1)2 ⎥ . (15)
⎣ N ⎦ ⎣N ⎦
n = 1,2,....,( N −1);
( p0 (ε )g(ξ )) [−a;ξ + ε ] = ( p0 (ε ) g(ξ )) [a;ξ + ε ] = 0.
Except for elementary operations of multiplication, use of Eq. (15) reduces the number of
numerical operations required to about half [51] in comparison with the earlier evaluations
[49] of the propagator by the Fourier transform method.
While the condition a = π N ε / 2 can be implemented to accommodate various
practical situations with other choices, a convenient choice is as follows. Let
1
a = 2γ π , γ = 0,1, 2,..., and ε = , ν = 1, 2,...,
2ν
which determine N :
N = 2μ , μ = (ν + γ + 1) .
The interval of integration can be selected first to be sufficiently large by setting a

appropriately. Then ε can be set sufficiently small to ensure the accuracy of the propagation
with respect to ξ . The value of N is determined by a and ε .
3.2. Calculated Intensity Profiles
Optical beam profiles were calculated at the focal point corresponding to a Clark-MXR
CPA 2010 femtosecond laser, which includes a 35 MHz SErF fiber oscillator and a chirped-
pulse-amplification, Ti: Sapphire regenerative amplifier, delivering 150 fs pulses of

wavelength centered at 775 nm with average power of 1 W corresponding to the maximum
pulse energy of 1 mJ, at a repetition rate of 1 kHz. The pulse energy could be controlled with
a set of ND 705 filters, collimated with a beam expander and focused with a chromatic
objective (CVI Laser Co.) with NA = 0.23 with focal length of 21 mm. The beam diameter at
the lens was adjusted to obtain the desired focal spot size.
For the calculations and observations, the amplitude at the lens location was taken to be
Gaussian:
⎡ r ′2 kr ′2 ⎤ ⎡ t2 ⎤
A( x′, y′, z′, t ) = A0 exp ⎢ − 2 − i ⎥ exp ⎢ −κ 2 ⎥ , (16)
⎣ w 2f ⎦ ⎣⎢ τ p ⎦⎥
which is the same as in Eq. (5), modulated with a Gaussian shape in time. In Eq. (16), w is
the waist radius at z ′ ; f is the focal length; τ p is the pulse width; t = t ′ − z ′ / c is the
retarded time with c being the speed of light and κ = 2 log(2) is a laser specific constant.
The peak intensity I 0 = A at z ′ can be calculated from the pulse energy E pulse from the
2
0
relation
∞ ∞ ∞
I0
E pulse =
τP ∫
−∞
dt ′ ∫ dx′ ∫ dy′ exp[−2( x′2 + y′2 ) / w2 − 2κ (t / τ P ) 2 ]
−∞ −∞
(17)
1 π
= π w2 I 0 .
2 2κ
The needed values of the Kerr coefficients and the ionization potentials for the gases are
listed in Table 1.
Table 1. Physical properties of the gases
Gas Kerr Ionization potential

coefficient (eV)
(m2/W)
Air 5.0 × 14.9
10-23
Nitrogen 4.5 × 15.6
10-23
Neon 1.0 × 21.6
10-24
Helium 3.5 × 24.6
10-25
Calculated results shown in Figures 1 and 2 correspond to the pulse energy of 200 μJ
with the beam propagating in air and the rate Rion , which depends on the ionization potential,
was calculated from a polynomial fit to the experimental values for O2 (20%) and N2 (80%)
[53]. Figure 1 shows the self-focusing effect of air and consequent impact of the plasma
generated on a Gaussian signal at a spot radius corresponding to a high intensity. Fluence is
normalized in units of the peak input fluence. Intensities for a beam propagating in vacuum
are sufficiently higher to cause some air breakdown but with little impact on the beam profile.
However, the intensity is increased substantially by self-focusing, causing a commensurate
distortion to the beam profile. For comparison, calculations were carried out for Neon and
Helium also. Each one of these gases has lower Kerr coefficient and higher ionization
potential than air. Both of these properties tend to reduce the distortion to the beam profile.
Neon showed little self-focusing effect and the effect of Helium was found to be unnoticeable
at this pulse energy and the other parameter values, with commensurate impact of plasma.
Figure 1. Normalized fluence in vacuum (dashed line), in air with self-focusing (dotted line) and with
consequent plasma effects included (solid line) at the focal radius of 1.5 μm.
Figure 2. Calculated intensity profile of the optical beam propagating in air with spot radius of 2.5 μm
(dashed line), 2.0 μm (dotted line) and 1.5 μm (solid line).
Figure 2 investigates the onset of the distortion to the beam profile. With the indicated
parameter values, noticeable distortion is observed as the spot radius is decreased below 2.5
μm . Thus, a clean fabrication can be expected for the spot radii of greater than 2.5 μm, as
long as it is not so large that the deformations due to the consequent low intensity become
significant. In any case, it is desirable to fabricate in the highest possible energy range, i.e., at
slightly larger spot radius than 2.5 μm with some margin allowed for the material plasma
having a minor impact on the results, which is not included in the present calculations.
4. SURFACE PROFILE OF METAL-TYPE MATERIALS

4.1 Modeling and Method
Calculation method described below is based on the assumption that the material removal
takes place continuously by direct energy transfer from the optical beam and subsequent
ablation for the duration of the phase. For the long pulse ablation, the material removal takes
place by melt and vaporization, and in case of the shorter pulses, most of the ablation takes
place during the post-pulse period when the energy diffusion among the conduction band
electrons and the transfer of energy by impact dominate. In both of these categories, energy is
not directly transferred with the consequent immediate removal of the material. Therefore, a
close conformity with this assumption is not so obvious.
While the fluid movement does take place in the melt causing diffusion of the original
beam profile for longer pulses, it is expected that the energy distribution and the directions of
the momenta retain the memory of the profile to a large extent. In cases of the shorter pulses,
a similar memory is expected in the energized electrons modulated by the energy diffusion in
the electron subsystem for an adequately shaped beam which is associated with equally
suitable plasma filament. For a distorted beam, which produces diffused and distorted plasma
plume, transfer of the energy is more complex and diffused. This view is supported by the
success of the threshold effect that is used to fabricate craters with smaller diameters by
reducing the beam energy, since the special dimensions of the beam above threshold are
closely maintained in the fabricated crater. Therefore, this assumption was expected to
produce a reasonable first approximation to the profile of the fabricated surface. As will be
shown in the foolowing, this method predicts surface profile quite accurately in all cases of
the laser fabrication of metallic materials, with adequately shaped beam. The method
described below is material independent [20]. The material properties impact upon the energy
needed for ablation.
Starting with an arbitrary point P (t ) = ( x, y, z ( x, y, t )) at time t, on the ablated surface
S (t ) , the purpose is to determine the corresponding point at an infinitesimal time later. Since
the initial surface is known, this is sufficient to determine it at all later times.
The tangent space TP (t ) to S (t ) at P (t ) is spanned by the basis vectors ex and e y ,
given by
∂z ∂z
e x = ( 1, 0, ) , and e y = ( 0, 1, ).
∂x ∂y
To a good degree of accuracy, a laser beam incident and reflected at P (t ) can be

assumed to transfer the momentum and energy in the direction of the outward normal
ν ( x, y , t ) to the space TP (t ) , given by,
(ey × ex ) 1 ⎛ ∂z ∂z ⎞
ν ( x, y, z ( x, y, t )) = = , , −1⎟ , (18)
| (ey × ex ) | η ( x, y, z ( x, y, t )) ⎜⎝ ∂x ∂y ⎠
where
⎡ ⎛ ∂z ⎞
2
⎛ ∂z ⎞ ⎤
2
η ( x, y, z (t )) = ⎢1 + ⎜ ⎟ + ⎜⎜ ⎟⎟ ⎥ .
⎣⎢ ⎝ ∂x ⎠ ⎝ ∂y ⎠ ⎦⎥
Let aθ be the absorption coefficient, where θ is the angle between the reflected beam
and the inward normal. The intensity I ( x, y, z ( x, y, t )) dt at P (t ) from time t to (t + dt )
can be calculated by the method described in Sec. 3.1. The absorbed energy I abs per unit area
is then given by I abs = [ aθ I ( x, y, z ( x, y, t )) dt ] . The absorption coefficient is assumed to be
directly proportional to the transferred momentum, i.e.,
aθ ( x, y, z ( x, y, t )) = a0 cos(θ ) = a0 (− zˆ •ν [ x, y, z ( x, y, t )]) = a0 / η ( x, y, z ( x, y, t ))
where ẑ is the unit vector in the upward normal direction with respect to the un-ablated
surface and the dot denotes the scalar product. Consequently,
I ( x, y, z ( x, y, t ))dt
Ι abs = a0 dt .
η ( x, y, z ( x, y, t ))
Let J dt be the energy per unit area lost to the material from time t to (t + dt ) . Then
the energy available to melt and evaporate the material is equal to ( I abs − J dt ) . Thus, the
depth of ablation σ ( x, y , z (t )) dt in the direction of the outward normal is equal to
I abs − J dt a I ( x, y , t ) / η ( x, y , z ( x, y , t ) − J
σ ( x, y, z (t )) dt = = 0 dt
EH EH
(19)
J ( x, y, z ( x, y, t ))
= dt ,
EH η ( x, y, z ( x, y, t ))
where E H is the energy required to remove the unit volume of the material. From Eqs. (18)
and (19), the point P (t + dt ) = [ P (t ) + dP (t )] , on the surface at time (t + dt ) ,
corresponding to P (t ) , is given by,
σ ( x, y, z ( x, y, t ), t ) dt ⎛ ∂z ∂z ⎞
dP (t ) = (dx, dy, dz ( x, y, t )) = ⎜ , , −1⎟ dt , (20)
η ( x, y, z ( x, y, t )) ⎝ ∂x ∂y ⎠
Eq. (20) provides the x, y and z components of the vector representing the change in
the position of a point P (t ) , yielding
dx σ ( x, y, z ( x, y, t ), t ) ∂z dy σ ( x, y, z ( x, y, t ), t ) ∂z
= , = ,
dt η ( x, y, z ( x, y, t )) ∂x dt η ( x, y, z ( x, y, t )) ∂y
and
dz ⎡ ∂z dx ∂z dy ∂z ⎤ σ ( x, y, z ( x, y, t ), t )
= ⎢ + + ⎥ = − ,
dt ⎣ ∂t dt ∂x dt ∂y ⎦ η ( x, y, z ( x, y, t ))
i.e.,
∂z
= − (ση )( x, y, z ( x, y, t ))
∂t (21)
yielding the following equation for time evolution of the ablated surface:
∂z J ( x, y, z ( x, y, t ))
+ = 0. (22)
∂t EH
Let tmin be the time t when the material removal starts, which is equal to the point when
the available energy crosses the ablation threshold, and let it end at tmax , the time when the
laser energy is no longer sufficient to maintain the material removal process and the losses to
the material. The surface profile is obtained by solving Eq. (22), from tmin to tmax with
z ( x, y, tmin ) describing the initial profile, which is assumed to be known. In our calculations,
the forward Euler method provided quite satisfactory solution of Eq. (22).
For the metal type materials, the energy needed to remove the unit volume EH , is
determined in a straightforward manner by the specific heats in its solid and liquid states, its
melting and evaporation points and its latent heats of melting and evaporation. This is clear
for the longer pulse lasers as the material removal takes place by evaporation. For the
femtosecond lasers with energy in the intermediate range, although the material is removed
directly due to an imbalance of the electromagnetic forces, the amount of energy needed to
remove a material volume is still quite accurately approximated by the energy needed to melt
and evaporate the volume, and the losses due to the diffusion can be ignored. For the longer
pulse lasers, the thermal diffusion equation describes the melting and evaporation processes.
Its solutions determine tmin , tmax and the heat loss by diffusion [20]. Since the material melts
early in the pulse and evaporation continues until close to the pulse end, tmin and tmax have
been found to be quite close to zero and the pulse-width, respectively. For the femtosecond
lasers, the beam profile, modulated by the diffusion in the electronic subsystem, is transferred
to the conduction band electrons from tmin to tmax , which are even closer to zero and the
pulse-width, respectively. Since this memory is retained to a large extent and there is no
energy transfer to the electron subsystem after the pulse, zero and the pulse-width can replace
tmin and tmax with a good degree of accuracy. In spite of an intuitive nature and a crudeness
of these arguments, the corresponding approximations have been found to be quite
compatible with the usual experimental accuracy, in all pulse-width ranges. More accurate
calculations would be too complex and cumbersome without a commensurate improvement in
describing the fabricated surface.
4.2. Surface Profile
Surface profile calculated by the method of Sec. 4.1 with the intensities calculated by the
method of Sec. 3.1 was compared with the surface of the craters fabricated on thin sheets of
copper for the same experimental setting as for Figures 1 and 2. Pulse delivery was adjusted
to ensure that each crater was produced by a single pulse. The craters produced were not
identical to each other due to noticeable random variations. For comparison with the
theoretical values calculated for a radially symmetric system, the depths along X and Y - axes
were averaged to obtain the profile with respect to the radial distance and measured data was
also averaged over a number of craters to minimize the effects of the statistical variations.
Since the representative experimental profile was obtained by averaging, it is radially
symmetric and due to the cancellations of statistical variations, it is smoother than the
individual craters. Since the beam and the entire system is radially symmetric, lack of
symmetry and statistical variations are of a random nature with undeterminable experimental
origin. In spite of various simplifying assumptions and approximations, the calculated and the
observed profiles are in close agreement as shown in Figure 3, indicating that the impact of
various other processes involved has a secondary impact and the shape of the fabricated
surface is largely determined by the intensity profile of optical beam above the ablation
threshold for the duration of pulse.
Effect of the material plasma is not included in these models. The complexity of the
phenomena describing the formation of the material plasma exceeds by far the value of the
additional information as the expected impact on the beam profile and the surface features is
estimated to be small in comparison. A better strategy is to develop an on-line monitoring
system to make minor adjustments to the calculated parameters if desired [36].
Figure 3. Calculated (solid line) and averaged experimental (circles) profile of the surface machined in
air with spot radius at focal point of 2.5 μm. Experimental depths were averaged over a number of
craters to minimize the impact of statistical variations.
5. RESULTS AND DISCUSSION

Typical dynamics of merger of the air plasma and material plasma are illustrated in
Figures 4(a) to 4(c). The plasma column region was imaged using a high resolution MEIJI
EMZ-TR stereo-microscope, with working distance of 93 mm and available magnification
range from 2.1X to 270X, together with a color CCD camera from the side. The microscope
due to its long working distance, resolution and magnification was found to be quite suitable
to observe the plasma columns. Also, the high sensitivity of CCD camera is better suited to
detect faint plasma. The images were captured and analyzed by using the software Vision
Gauge. Figures 4(a) to 4(c) show the plasma column produced by the air breakdown moving
and merging with the plasma plume produced by laser ablation, in the focal region with focal
spot size of 10 µm and peak input power density of 3.7 × 1011 W/cm2, where the plasma
formation is maximum and more visible in the energy range considered in the experiments
[24]. Minimum energy required for producing the plasma columns was found to be 50 µJ,
delivering the peak input power density of 1.85 × 1010 W/cm2. Figure 4(a) shows the two
columns before merging, Figure 4(b) shows the onset of merging and Figure 4(c) shows the
two in a merged state. The plume is seen to expand transversely as the two merge together.
Figure 5 shows the craters punched by 3000 shots of pulses with τ = 150 fs and the peak
input intensity of 24 J/cm2 in air, Nitrogen, Neon and Helium, at ambient pressure in copper
with material surface at the focal point. The intensity at the focal point is sufficiently high to
ionize the air which has the lowest ionization potential and suffers from higher self-focusing
effect due to the highest value of the Kerr coefficient, as shown in Table 1. Both of these
effects decrease progressively from air to Nitrogen and Neon, reduced to minimum for
Helium.
Figure 4. Total plasma-density variation during merger of the plasma columns due to air breakdown
with the ablation plume at peak power density of 3.7x1011 W/cm2. The diameter and the length of the
air plasma are ~ 88 µm and 294 µm, respectively.
Figure 5. Holes in copper punched by 3000 shots of pulses with τ = 150 fs and intensity of 24 J/cm2 in
(a) air, (b) N2, (c) Ne, and (d) He at ambient pressure.
As shown in Figure 5, quality of the fabricated craters is commensurate with the beam
shape, i.e., as the distortion to the beam reduces, the quality of surface features improves.
Figure 6 shows the calculated beam profile at various spot diameters maintaining the
peak input power density of 3.7 × 1010 W/cm2, to determine the onset of distortion. As
indicated, the distortion starts at the spot radius of about 5.0 μm. Taking into account the
ignored effects of material plasma and experimental variations, 5.5 to 6.0 μm is a safe spot
radius for a clean fabrication. About the same profile can be obtained with 5.0 μm spot radius
at about 50 μm past the focal point, and at the focal point with the lens power density of about
3.2x1010 W/cm2. Thus the basic characteristic determining the experimental parameters for a
clean machined surface is to maintain a moderate intensity at the fabrication point.
Experimentally, a convenient procedure to this effect is to determine the aperture at the lens
to result in the desired spot radius and adjust the power by changing the beam energy in the
vicinity of the calculated value for a clean fabrication. This parallels the procedure used in
fabricating the features smaller than the wavelength by taking advantage of the threshold
effect.
Figure 6. Calculated intensity profile of the optical beam propagating in air with spot radius of 6.0 μm
(dashed line), 5.5 μm (dotted line) and 5.0 μm (solid line) and power density of 3.7x1010 W/cm2 at the
lens.
Figure 7 shows the craters fabricated with 2000 femtosecond and nanosecond pulses on a
steel plate. A nanosecond, 532 nm YAG laser (Spectra Physics) was used for comparison
with the femtosecond machining. The power density of 3.7x1010 W/cm2 at the lens was
sufficiently high to produce a distorted beam profile at the focal point with spot radius of 5.0
μm as shown in Figure 5. Femtosecond machining of Figure 7(a) was conducted by placing
the fabrication point 60 µm past the focal point yielding the cleanest features, which is close
to the calculated value of 50 µm. Similar results were achieved either by increasing the spot
radius to 6.0 μm or by reducing the power density to 3.2x1010 W/cm2 and machining at the
focal point. In all three cases, the beam is adequately shaped with the corresponding
filamentation effect. Figure 7(b) shows the result for the femtosecond machining at the focal
point with the power density at the lens of 3.7x1010 W/cm2 with spot radius of about 5.0 μm.
Due to the distorted beam profile and defused plasma, the dimensions of the fabricated crater
in (b) do not conform as closely to the spot size as in (a). For Figure 7(c), the location and
spot radius are the same as in Figure 6(a) with pulsewidth of 40 ns. As shown, the cleanest
surface results with the femtosecond pulses and an adequately shaped beam, as expected. The
crater of Figure 7(a) also has almost vertical walls compared to the others, which have conical
shapes. With a number of femtosecond pulses in the vicinity of each other, heat generated
with each pulse accumulates creating some thermal equilibrium and melt, as well as a small
amount of debris and dross. As a result, the fabricated surface features with many pulses are
not as clean compared to the single pulse but still considerably cleaner than longer pulses.
(a) (b)
(c)
Figure 7. Comparison of the holes drilled with 150 femtosecond pulses and with 40 nanosecond pulses.
The figure shows the 1000× SEM of the craters drilled by 2000 pulses with peak input power density of
3.7 × 1010 W/cm2 at pulse repetition rate of 1 kHz, in each case. (a) femtosecond pulses of 150 fs with
moderate intensity, (b) femtosecond pulses of 150 fs with high intensity, (c) nanosecond pulses of 40 ns
with moderate intensity.
(a) (b)
Figure 8. Comparison of channels machined with 150 femtosecond pulses with peak input power
density of 6.1 × 1010 W/cm2. (a)With moderate intensity and (b) With high intensity. The cutting speed
was 200 mm/min and 50 passes resulted in a depth of about 160 µm and width of 35 µm in (a) and 30
µm in (b).
Figure 8 shows the results of femtosecond machining of channels in steel. The peak input
power density in Figure 8 is higher than in Figure 7 but with a correspondingly larger spot
radius. As a result the intensity at the work piece surface in Figure 8 (a) was comparable with
that of Figure 7(a), and in Figure 8(b), which has a smaller spot radius compared to (a), the
intensity was comparable with Figure 7(b). Thus, the machined cut of Figure 8(a) was
obtained with adequately shaped beam and Figure 8(b), with a distorted beam profile and
plasma plume. As seen, an adequately shaped beam produces a cleaner cut with almost
vertical wall with aspect ratio of about 5, and no debris and dross while the distorted beam
results in a conical shape with debris and dross accumulated. It was noted further that little
dross and re-solidification may deposit on the cut periphery even with an adequately shaped
beam if the cutting speed is very low. At a low feed rate, a larger number of pulses deposit
energy in closer proximity of each other creating effects somewhat similar to many pulses
fabricating a single crater. High feed rate reduces the depth of the fabricated features but it
can be increased by increasing the number of passes. For the present experiments, a cutting
speed of 100 mm/min to 300 mm/min was found to be adequate.
a) b)
c)
Figure 9. Clean cutting in (a) Copper (b) Shape memory alloy NiTi (c) Platinum
Figure 9 shows clean cutting of channels obtained in copper, NiTi shape memory alloy
and platinum using adequately shaped optical beam profile at a cutting speed of 100 mm/min.
As indicated earlier, the optimal fluence at the fabrication point depends on the input fluence
at the lens and the spot diameter. The input fluence was adjusted in each case depending on
the width of the channel to obtain a moderate fluence at the fabrication point.
CONCLUSION
In this article, we have presented a theoretical and experimental study of femtosecond
laser machining of metals and compared with the longer pulse ablation. Laser material
interaction for the metals, semimetals, semiconductors and dielectrics for the pulse widths
covering from the nanosecond to femtosecond range is discussed to place the present study in
perspective. An improved method to calculate the intensity profile of the optical beam and a
method to calculate the fabricated surface features of the material are described. Calculated
features of a crater punched with a single femtosecond pulse closely agree with the machined
surface features. Craters and channels fabricated with multiple pulses are also presented.
Theoretically and experimentally, it is found that there is an intermediate energy range
where the effects of scattering between the photon and plasma balance out resulting in an
adequately shaped beam profile, which is associated with plasma filamentation. At higher
energies, these effects distort the beam profile with commensurate impact on the surface
features and at lower energies; the material plasma has too little energy resulting in resettling
and agglomeration generating debris and dross. The adequate energy range can be calculated
yielding suitable experimental parameters for cleaner fabrication. If the input power density is
high enough to distort the beam profile at the focal point, adequate fluence can be achieved
by increasing the spot radius rendering the focal point suitable for fabrication. Alternatively,
fabrication can be conducted at a point away from the focal point. Both approaches result in
an increased spot radius. With a fixed spot radius, adequately shaped beam can be achieved
and machining be conducted at the focal point by decreasing the input power density at the
lens as calculated with some adjustment in its vicinity, if needed. This scheme provides better
control over the dimensions of the fabricated surface and appears to be more convenient
experimentally.
REFERENCES
[1] Vatsya, S.R.; Virk, K.S. J. Lase. Appl. 2003, 15, 273-278.
[2] Nolte, S.; Momma, C.; Jacobs, H.; Tunnermann, A.; Chichkov, B. N.; Wellegehausen,
B.; Welling, H. J. Opt. Soc. Am. 1997, 14, 2716-2722.
[3] Corkum, P. B.; Brunel, F.; Sherman, N. K. Phys., Rev. Lett. 1988, 61, 2886-2889.
[4] Chichkov, B. N.; Momma, C.; Nolte, S.; von Alvensleben, F.; Tunnermann, A. Appl.
Phys. A: Mater. Sci. Process. 1996, 63, 109-115.
[5] Schwarz-Selinger, T.; Cahill, D. G.; Chen, S. C.; Moon, S. J.; Grigoropoulos, C. P.
Phys. Rev. B 2001, 64, 15 5323.
[6] Vatsya, S. R.; Nikumb, S. K. Phys. Rev. B 2003, 68, 035410-1 - 035410-5.
[7] Nedialkov, N. N.; Imamova, S. E.; Atanasov, P. A.; Heusel, G.; Breitling, D.; Ruf, A.;
Hügel, H.; Dausinger, F.; Berger, P. Thin Solid Films 2004, 453, 496-500.
[8] Marcinkevičius, A.; Juodkazis, S.; Watanabe, M.; Miwa, M.; Matsuo, S.; Misawa, H.;
Nishii, J. Opt. Lett. 2001, 26, 277-279.
[9] Buerle, D. Laser Processing and Chemistry; 3rd Ed.; Springer-Verlag Berlin, 2000.
[10] Craig, B. Laser Focus World 1998, 34, 79-88.
[11] Hirao, K.; Miura, K. J. Non-Cryst. Solids 1998, 239, 91-95.
[12] Sudrie, L.; Franco, M.; Prade, B.; Mysyrowicz, A. Opt. Commun. 2001, 191, 333-339.
[13] Venkatakrishnan, K.; Tan B.; Stanley, P.; Lim, L. E. N.; Ngol, B. K. A. Optical
Engineering 2002, 41, 1441-1445.
[14] Pronko,P.; Dutta, S.; Squier, J. Opt. Commun. 1995, 114, 106-110.
[15] Hermann, J.; Le Floch, T. J. Appl. Phys. 2004, 96, 3084-3094.
[16] Sun, J., Longtin; J. P. J. Opt. Soc. Am. B 2004, 21, 1081-1088.
[17] Chin, S. L.; Aközbec, N.; Proulx, A.; Petit, S.; Bowden, C. M. Opt. Commun. 2001,
188, 181-186.
[18] Vatsya, S. R.; Nikumb, S. K. J. Phys.: Conference Series 2007, 59, 704-707.
[19] Vatsya, S. R.; Li, C., Nikumb; S. K. J. Appl. Phys. 2005, 97, 034912.
[20] Vatsya, S. R.; Bordatchev, E. V.; Nikumb, S. K. J. Appl. Phys. 2003, 93, 9753-9759.
[21] Li , C.; Nikumb, S. K. Appl. Opt. 2003, 42, 2383–2387.
[22] Preuss, S.; Demchuk, A.; Stuke, M. Appl. Phys. A: Mater. Sci. Process. 1995, 61, 33–
37.
[23] Sun, J.; Longtin, J. P. J.Appl. Phys. 2001, 89, 8219-8224.
[24] Li, C.; Vatsya, S. R.; Nikumb, S. K. J. Laser Appl. 2007, 19, 26-31.
[25] Braun, A.; Korn, G.; Liu, X.; Du, D.; Squier, J.; Mourou, G. Opt. Lett. 1995, 20, 73–75.
[26] Nibbering, E. T. J.; Curley, P. F.; Grillon, G.; Prade, B. S.; Franco, M. A.; Salin, F.;
Mysyrowicz, A. Opt. Lett. 1996, 21, 62–64.
[27] Brodeur, A.; Chien, C. Y.; Ilkov, F. A.; Chin, S. L.; Kosareva, O. G.; Kandidov, V. P.
Opt. Lett. 1997, 22, 304–306.
[28] Mlejnek, M.; Wright, E. M.; Moloney, J. V. Opt. Lett. 1988, 23, 382–384.
[29] Schwarz, J.; Rambo, P.; Diels, J. C.; Kolesik, M.; Wright, E. M.; Moloney, J. V. Opt.
Commun. 2000, 180, 383–390.
[30] Kasparian, J.; Sauerbrey, R.; Chin, S. L. Appl. Phys. B: Lasers Opt. 2000, 71, 877–879.
[31] Tzortzakis, B.; Berge, L.; Couairon, A.; Franco, M.; Prade, B.; Mysyrowicz, A. Phys.
Rev. Lett. 2001, 86, 5470–5473.
[32] Couairon , A.; Berge, L. Phys. Rev. Lett. 2002, 88, 135003.
[33] Zhao, X. M.; Diels, J. C.; Wang, C. Y.; Elizondo, J. M. IEEE J. Quantum Electron.
1995, 31, 599–612.
[34] Liu, W.; Kosareva, O.; Golubtsov, I. S.; Iwasaki, A.; Becker, A.; Kandidov, V. P.;
Chin, S. L. Appl. Phys. B: Lasers Opt. 2003, 76, 215–229.
[35] Tzortzalis, S.; Sudrie, L.; Franco, M.; Prade, B.; Mysyrowicz, A. Phys. Rev. Lett. 2001,
87, 213902.
[36] Zheng, H.Y.; Deng, Y. Z.; Vatsya S. R.; Nikumb, S. K. Appl. Surf. Sci. 2007, 253,
3408-3412.
[37] Vatsya, S. R.; Nikumb, S. K. J. Appl. Phys. 2002, 91, 344-351.
[38] Apostolova, T.; Hahn, Y. J. Appl. Phys. 2000, 88, 1024-1034.
[39] Stuart, B. C.; Feit, M. D.; Herman, S.; Rubenchik, A. M.; Shore, B. W.; Perry, M. D.
Phys. Rev. B 1996, 53, 1749-1761.
[40] Wang, X.; Xu, X. J. Heat Transfer 2002, 124, 265-274.
[41] Stoian, R.; Rosenfeld, A.; Ashkenashi, D.; Hertel, I. V.; Bulgakova, N. M.; Campbell,
E. E. B. Phys. Rev. Lett. 2002, 88, 097603-097606.
[42] Brabec, T.; Krausz, F. Rev. Mod. Phys. 2000, 72, 545-591.
[43] Krainov, V. P.; Roshchupkin, A. S. Phys. Rev. A 2001, 64, 063204-063208.
[44] Glezer, E. N.; Siegal, Y.; Huang, L.; Mazur, E. Phys. Rev. B 1995, 51, 9589-9596.
[45] Miatello, A.; Kelley, R. Appl. Phys. A Suppl. 1999, 69, S67-S73.
[46] Toulemonde, M.; Dufour, C.; Meftah, A.; Paumier, E. Nucl. Instrum. Methods Phys.
Res. B 2000, 167, 903-912.
[47] Rae, S. C. Opt. Commun. 1994, 104, 330-335.
[48] Chiron, A.; Lamouroux, B.; Lange, R.; Ripoche, J. F.; Franco, M.; Prde, B.; Bonnaud,
G.; Riazuelo, G.; Mysyrowicz, A. Eur. Phys. J. D 1999, 6, 383-396.
[49] Subbarao, D.; Singh, H.; Uma, R.; Bhaskar, S. J. Plasma Phys. 1998, 61, 449-467.
[50] Feynman, R. P.; A. R. Hibbs, Quantum Mechanics and Path Integrals; McGraw-Hill:
New York, NY, 1965; pp 75-80.
[51] Vatsya, S. R. J. Opt. Soc. Am. B 2005, 22, 2512-2518.
[52] Hawkins, R. J. Appl. Opt. 1987, 26, 1183–1188.
[53] Talebpour, A.; Yang, J.; Chin, S. L. Opt. Commun. 1999, 163, 29-32.
Chapter 7
FEMTOLASERS-MEDIATED MULTIPHOTON
EXCITATION IMAGING OF BULK OCULAR TISSUES
Bao-Gui Wang∗ and Karl-Jürgen Halbhuber

Lasermicroscopy Research Unit, Institute of Microscopic Anatomy (Anatomy II),
Medical Faculty, Friedrich-Schiller University Jena, Germany
ABSTRACT
Currently, femtosecond lasers (femtolasers) are being extensively employed in
diverse research and application fields. Femtolasers-mediated multiphoton excitation
laser scanning microscopy is one of the most exciting recent developments in biomedical
imaging and becomes more and more an inspiring imaging technique in the intact bulk
tissue examinations. In this chapter, the non-linear excitation imaging technique
including two-photon autofluorescence (2PF) and second harmonic generated signal
imaging (SHG) was employed to investigate the microstructures of whole-mount corneal,

retinal, and scleral tissues in their native environment. In details, image acquisition was
based on intense ultrafast femtosecond near-infrared (NIR) laser pulses, which were
emitted from a mode-locked solid-state Ti: sapphire system. By integrating high-
numerical aperture diffraction-limited objectives, multiphoton microscopy/tomography of
ocular tissues was performed at a high light irradiance order of MW-GW/cm2, where two
or more photons were simultaneously absorbed by endogenous molecules located in the
thick tissues. As a result, the cellular and fibrous components of intact scleral and corneal
tissues were selectively displayed with the assistance of the in-tandem detection of 2PF
and SHG procedures. Any exogenous dye was not used. High-resolution optical images
of keratocytes in cornea, fibroblasts, mature elastic fibers and blood capillaries in sclerae
as well as of the retina radial Müller glial cells, ganglion cells, bipolar cells,
photoreceptors, and retina pigment epithelial (RPE) cells were acquired. Furthermore,
this promising technique has been proved to be an indispensable tool in assisting
femtolasers intratissue surgery, especially for in situ assessing the obtained microsurgical
effects. Most remarkably, the activated keratocytes, also named myofibroblasts during
wound repair, were in vivo detected using the multiphoton excitation imaging in the
∗
Address: Institute of Anatomy, Teichgraben 7, 07743 Jena, Germany. Phone: ++49 3641 938550; fax: ++49 3641
938552. Emails: Karl.Halbhuber@mti.uni-jena.de; Baogui.Wang@mti.uni-jena.de
232 Bao-Gui Wang and Karl-Jürgen Halbhuber
treated animals twenty-four hours after the intrastromal surgery. In this chapter,
RESULTS is divided into five areas of interest: (1) corneal multiphoton imaging; (2)
scleral multiphoton imaging; (3) retinal multiphoton imaging; (4) uses of multiphoton
excitation imaging in intrastromal laser surgery; (5) uses of multiphoton microscopy in
detecting activated stromal cells after surgical laser treatment. Data show that the in-
tandem combination of 2PF and SHG imaging allows for in situ co-localization imaging
of various microstructural components in the whole-mount ocular tissues. Qualitative and
quantitative assessment of microstructures was obtained. The selective displaying merits
of tissue components only with the excitation of different wavelengths is the most
exciting development for bulk tissue imaging, which allows to selectively studying of
three-dimensional (3-D) architecture of cellular microstructures and extracellular matrix
arrangement at a substantial depth. Using the laser power within threshold value, the bulk
tissues can be imaged numerous times without visible photodisruption. Intrinsic emission
multiphoton microscopy/tomography is consequently confirmed to be an efficient and
sensitive non-invasive imaging approach, featured with high contrast and subcellular
spatial resolution. The non-linear optical imaging yields vivid insights into biological
specimens that may ultimately find its clinical application in optical pathological
diagnostics. We believe that this promising technique will also find more applications in
the biological and medical basic research in the near future. Potentially, based on its
capability of intrinsic fluorescence imaging, the 2PF of cytoplasmatic NAD(P)H would
have a use in studying the dynamics of the mitochondrial bioenergetics and further utility
in conducting nanomedical investigations in this organelle.
Keywords: optical • non-linear • near-infrared (NIR) • 2PF • SHG • autofluorescence •

tomography • cornea • sclera • retina • epithelium • Bowman’s layer • keratocytes •
endothelium • fibroblasts • capillary • elastic fibers • collagen • Müller cells • bipolar
cells • RPE • laser surgery • myofibroblasts.
1. INTRODUCTION
The advent of femtolaser sources led to the experimental implementation of multiphoton

excitation microscopy, providing a non-invasive approach to examine the structures and
functions of living cells and tissues. This optical imaging technique, based on simultaneous
absorption of two or more photons under high light irradiance, is one of the most exciting
recent technical developments in biomedical imaging [Denk et al. 1990, So et al. 1998, Vogel
et al. 1999, König 2000a, Campagnola et al. 2002, Zoumi et al. 2002, Halbhuber et al. 2003,
Zipfel et al. 2003a, Williams et al. 2005, Han et al. 2006]. This image acquisition is based on
non-linear optical response to endogenous and exogenous fluorophores arising from their
asymmetric distribution of the electronic charges in the molecular ground and excited state.
The feasibility of simultaneous absorption of two photons was predicted by Goeppert–
Mayer in her dissertation in 1931. Multiphoton microscopy, including two-photon
fluorescence (2PF) and second harmonic generation (SHG), only occurs when the irradiance
reaches an order of MW-GW/cm2 [Göppert, 1931]. This technique met with the interests of
biomedical scientists after the introduction of convenient lasers. The first two-photon
excitation of electronic states was found for blue fluorescence of CaF2:Eu2+ crystals based on
the availability of lasers in 1961 [Kaiser et al. 1961]. To date, tunable mode-locked
Femtolasers-Mediated Multiphoton Excitation Imaging of Bulk Ocular Tissues 233
Ti:sapphire lasers operating at high repetition rate (typically 80–90 MHz) and a ultrashort
pulse duration (less than 200 fs) are typically used as NIR sources in multiphoton microscopy
[Zipfel et al. 2003b, König et al. 2003, Rubart 2004, Wang et al. 2006a]. The required
transient high irradiance can be obtained by tightly focusing ultrashort laser pulses within a
subfemtoliter focal volume provided by a special objective with high numerical aperture.
Endogenous fluorophores such as NAD(P)H, flavin, lipofuscin, and porphyrins in
biological tissues were revealed. In single-photon confocal laser scanning microscopy
(CLSM), most of the fluorophores can be stimulated by UV radiation to emit longer
wavelengths with majority of the spectral intensity between 400 and 600 nm. Multiphoton
microscopy is well suited for the high-resolution imaging of intrinsic molecular signals in
living specimens. The NAD(P)H in mitochondria is the predominant source of endogenous
fluorophore in cells. This intrinsic fluorophore can be evoked at wavelengths of 720–780 nm
in a two-photon absorption process [Zipfel et al. 2003b]. In addition, the lipofuscin found in
retina pigment epithelial (RPE) cells [Han et al. 2006] and the elastin located in bulk tissues
[König et al. 2005a] can also be evoked under two-photon excitation condition.
However, multiphoton microscopy requires enormously high light intensities that, if
continuous, would almost instantly vaporize the specimen. This problem has been
circumvented by using such lasers that produce amazingly brief pulses (fs, ~ 10–15 s) at a high
repetition rate, thus generating a very high instantaneous energy but low average energy. In a
two-photon excitation process, the rate of excitation is proportional to the average squared
photon density; the fluorophore must simultaneously absorb two photons per excitation
process. Only in the focused volume is there sufficient intensity to generate appreciable
excitation [König et al. 2000b, Williams et al. 2001, Halbhuber et al. 2003, Vogel et al. 2003,
Helmchen et al. 2005]. As a result, the multiphoton imaging is spatially confined to the
minute sub-femtoliter focal volume resulting from the focusing of special microscope
objectives; consequently is thereby no essential out-of- focus photobleaching.
This multiphoton imaging technique, in conjunction with the pulsed infrared light,
provides other distinct advantages over single-photon CLSM: (i) No pinhole aperture is
required for multiphoton microscopy. Under CLSM microscope, not only the in-focus and
also the out-of-focus molecules can be excited [Halbhuber et al. 1998]. The emitted
fluorescent signals of the both areas are collected. The pinhole aperture, an indispensable
element of CLSM, functions to rejecting the out-of-focus fluorescence signals. The
development of multiphoton excitation greatly minimized this problem because only the in-
focus molecules can be stimulated under two-photon excitation condition. Meanwhile, the
focused volume can be further adjusted through the depth of the sample and thus to achieve
non-linear optical sectioning in three-dimensions; (ii) Samples can be examined with
subcellular resolution in ex vivo even in vivo states [König et al. 1996, Halbhuber et al. 2003,
Wang et al. 2007a]. Conventional fluorescence microscopy yields poor images, because in-
focus structures are washed out by the fluorescence signals outside plane of focus. In general,
most of the biological one-photon based microscopic techniques require labelling with
conventional fluorophores or fluorescent proteins. Thus, the in situ microstructural basis for
intact tissue properties without any histological assistance is difficult to be revealed under
single-photon CLSM. By contrast, the endogenous molecules in deep bulk tissues can be
probed under multiphoton microscope. Therefore, this new technology provides an
encouraging examination approach for living cells and tissues; (iii) In addition, owing to the
use of short (high-energy) excitation wavelengths, photodamage and photobleaching
unavoidably takes place in CLSM; the use of UV light also limited imaging depth penetrating
into biological tissues due to its high absorption and scattering coefficients of the incident
radiation. By contrast, multiphoton microscopy in conjunction with NIR light allows for
reduced photodamage and deeper sensing depth.
These advantages of multiphoton microscopy, along with its 3D demonstration, has been
employed as an encouraging imaging method in studying of living cells [König et al. 1996,
Piston 1999], mapping ligand-gated ion channel distributions [Denk 1994], organelles and
DNA imaging [Malak et al. 1997, König et al. 2000c, Dedov et al. 2001], imaging
mammalian embryos [Squirrell et al. 1999], skin probing [Masters et al. 1997, So et al. 1998,
König et al. 2003], arterial imaging [Zoumi et al. 2004], detecting of microtubules in neuronal
cell [Dombeck et al. 2003] and deep brain tissue [Levene et al. 2004, Brecht et al. 2004,
Nitsch et al. 2004]. Cellular interactions were as well revealed using this non-linear image
acquisition technique [Stoll et al. 2002]. Membrane imaging and visualizing of biomolecular
array were reported [Moreaux et al. 2000, Campagnola et al. 2003]. Fluorescence
spectroscopy of NAD(P)H and flavoprotein under two-photon absorption were furthermore
addressed [Huang et al. 2002]. The benefits of this emerging optical technique indicate that it
may provide significant progress in the assessment of cellular and extracellular components
in ocular tissues.
In this chapter, the non-linear excitation imaging technique including 2PF and SHG
signal imaging, realized with the converging ultrafast NIR femtolasers, was employed to
investigate the microstructures of whole-mount corneal, retinal, and scleral tissues in their
native environment. We paid considerable attentions to elucidating the cellular and fibrous
components with subcellular spatial resolution using in-tandem detection with the two non-
linear optical imaging approaches, especially to quantitative and qualitative acquisition of
microstructures in different layers of the intact bulk tissues.
2. MULTIPHOTON IMAGING SYSTEM INCLUDING TWO MAIN PARTS:

A MULTIPHOTON MICROSCOPE AND THE FEMTOLASER CHAMELEON

Multiphoton microscopy of ocular tissues was performed with an inverse laser scanning
microscope (Carl Zeiss, Jena, Germany), to which a tunable compact mode-locked
Ti:sapphire laser Chameleon (Coherent Inc., Santa Clara, CA, USA) was coupled (Figure 1).
The laser pulses were featured by 80 MHz to 90 MHz repetition rate, 690 nm to 1040 nm
wavelength spectrum, and 140 fs pulse duration (Figures 1 and 3).
A diffraction-limited 40x objective (1.3 N.A., oil, Zeiss Plan Neofluar) with a work
distance penetrating into the samples up to 140 µm and a 20x objective (0.95 N.A., water)
with a work distance of more than 1000 µm in bulk tissue were used. A 650 nm beam splitter
(Chroma Technology, Brattleboro VT, USA) in the backward direction was indispensable for
separating the emitted light from the reflected incident light. Another band-pass filter 435-485
nm for 2PF imaging was additionally built up before a photomultiplier tube (PMT). For SHG
signal imaging, additional narrower band-pass 420 ± 10 nm was used. PMT was an essential
component in the image acquisitions. The pixel-wise detecting photosensitive device, rapidly
responding with high sensitivity and efficiency, was externally installed under the baseport of
the multiphoton microscope to modulate and magnify the emission fluorescence and SHG
signals.
Figure 1. Ex vivo 2PF/SHG signal imaging tomography in diverse ocular tissues. The schematic (upper
figure) shows that the multiphoton imaging system included an inverted confocal microscope and the
laser Chameleon. The optical apparatus deflecting laser beams is not shown in this drawing. The
photogram (lower figure) taken from upper-side focused on the sample alignment stage depicting on-
site laser microscopy of a porcine eye globe. (A). A special experimental sample holder including a
tissue chamber, a 170-micron-thick glass, and two parts of metal covers which fulfilled an essential
function to keeping the eyeball moist during image acquisitions (Reprint Lasers Surg Med 2008)
The emission radiances of autofluorescence and SHG were measured with the assistance
of another highly sensitive polychromatic META-detector and of a special lambda-scanning
program installed in this microscopic system (Carl Zeiss, Jena, Germany). The emission
radiances were separated into various wavelengths and then the entire spectrum was projected
onto 32 channels of the META detector. LSM software can display the images acquired from
the META detector channels as an image series from shortest to longest wavelengths. The
lambda-scanning fulfilled a function to determining the emission wavelength optimum of a
fluorophore.
Figure 2. Examination of eye globes before multiphoton excitation imaging using a Zeiss slit lamp. (A).
Overview of the all-rounder Zeiss slit lamp 120 equipped with a digital camera; (B). On-site photogram
of corneal examination pre-treatment. Only the clear corneas were forwarded to take further
examinations.
Figure 3. Corneal multiphoton microscopy with living animals. (A). As illustrated on the schematic
drawing, the ultrashort laser pulses were deflected into the multiphoton microscope and then conveyed
into samples through a scanning unit built up in the microscopic system. The emitted lights including
autofluorescence and second harmonic generated signals were separated from the incident lights by a
filter cube and then guided into PMT in a backward direction; (B). On-site photogram for overview of
multiphoton imaging system. The animal was being imaged under the inverse multiphoton microscope.
The optical mirrors deflecting the laser beams are illustrated in a simplified way on this photogram. a)
compact mode-locked tunable 90 MHz Ti: sapphire laser Chameleon; b) multiphoton microscope based
on the inverse Zeiss LSM; c) animal during imaging; d) wavelength tuner; e) imaging monitor; (C).
Frontal photogram showing a rabbit being imaged under the multiphoton microscope. The animal was
narcotized by intravenous narcotics ketamine and xylacine. The blood oxygen concentration and pulse
frequency of the being treated animal was totally monitored from begin to end during the non-linear
optical imaging by a Draeger Dialogue 2000. Insert (D). Image taken from underside of the sample
stage showing the animal eye being treated.
Image acquisition of cellular autofluorescence and collagen SHG was performed at

different excitation wavelengths respectively. The wavelengths are mentioned in the
following sessions with the corresponding images. Owing to the high NIR penetration and the
minute focus volume (0.1 fl), serial optical transversal sections with low phototoxicity to the
surrounding region were obtained in certain fine z-steps. The high numerical aperture
objectives used allowed the emission lights to be efficiently collected.
The laser powers used for multiphoton imaging were set between 20 mW and 35 mW
with respect to different depths (Figure 4) [Wang et al. 2007a].
Figure 4. Diagram illustrating the discrepancy of optimized laser power for multiphoton excitation
imaging and of laser threshold power for intrastromal nanosurgery in different depths of cornea with
living animals. The laser powers set for multiphoton imaging were between 20 mW and 35 mW in
dependence on the imaging depth. If the laser intensity goes beyond threshold value, photodisruptive
effects could occur. Photodisruptive effects mean nanosurgical effects in this chapter. For instance, at a
depth of 90 micrometers in cornea, the optimized laser power for multiphoton imaging was 26 mW and
the threshold value for photodisruption was 98 mW (Reprint Cell Tissue Res 2007).
For image preparations, Zeiss LSM 510 Image Brower (Carl Zeiss, Jena, Germany) and
Adobe Photoshop 5.0 (Adobe Systems Incorporated, San Jose, CA, USA) were used for
image transfer and image improvement in an acceptable manner.
3. SUBJECTS AND EXAMINATIONS

The ex vivo multiphoton imaging was performed with porcine eyes since the porcine
eyeballs are of a size and structure similar to the human ones. The fresh samples were
removed from ocular orbit 20 minutes after slaughtering and transported in a nutrition
solution (Sigma-Aldrich, UK) at a constant temperature (+ 4oC) to the experimental
laboratory.
In general, the whole-mount scleral, corneal, and retinal tissues were not stained or fixed
before multiphoton excitation imaging. In order to minimize the autolysis post mortem,
examination time with each sample was kept not longer than 10 minutes. Glucose was
especially added to PBS in order to keep the ocular tissues better in vitality.
3.1. Corneal and Scleral Examinations
The examinations of cornea and sclera were performed with whole eye globes. The entire
globe was laid on the round 170-micron-thick glass set in a 65 mm-diameter dish. In order to
hold eye globes in a physiologically similar state, two drops of the nutrition solution (Sigma-
Aldrich, UK) were rendered to the entire globes during the examinations. In addition, the
entire eye globe was covered with a special chamber, which was made of rustless steel and
comprised two parts: one was a cylinder with 2.6 cm-diameter and 2.5 cm-height, the other
was a cover set to the cylinder. During the imagings, the laser beams were directed into
sample tissues from their outer surfaces. The emission radiance was subsequently collected
and then immediately projected into a photomultiplier tube (PMT, Hamamatsu, Bridgewater,
New Jersey, USA) [Wang et al, 2005a and 2008a].
3.2. Retinal Examinations
To make a preparation for retinal specimens, a posterior portion of eye globe was
explanted transsclerally with the assistance of a 8 mm-diameter biopsy punch (Stiefel Inc.,
Offenbach, Germany), then mounted onto the central point of a 65 mm-diameter plastic dish
bottomed with a 170-micron-thick glass coverslip (Karl Hecht KG, Germany); meanwhile,
the retinal side of the explant was laid downwards to contact the glass. One drop of the
nutrition solution (Sigma-Aldrich, UK) was rendered to the whole sample in order to keep its
moisture. One drop of Aqua bidest was added between the glass and eye globe in order to
minimize the light dispersion. The laser beams penetrated the internal limiting membrane into
retina. The retinal tissues examined were located in situ 2 mm from the optic disk on the
temporal side of the eyeballs [Wang et al. 2006b and 2008a].
The in vivo corneal multiphoton imaging was performed with rabbits (Figures 2 and 3).
New Zealand Albino rabbits (Charles River Wiga, Sulzfeld, Germany) with an average age of
6 ± 0.8 months and a mean weight of 4.1 ± 0.6 kg were used. Animal care and procedures in
this chapter were in accordance with the guidelines of the “German Law on the Protection of
Animals” and the “Principles of Laboratory Animal Care”. Animal permission was granted
by the Thuringian Ministry for Research and Culture (Nr. 02-17/04 Thuringia).
Corneas were examined preoperatively by a Zeiss SL 120 Slit lamp (Carl Zeiss, Jena,
Germany) and a Zeiss Stemi DV 4 stereomicroscope (Carl Zeiss, Jena, Germany). Only the
clear corneas were forwarded to take optical imaging. The conjunctival sac of the treated eye
was rinsed with phosphate-buffered saline (PBS) pre- and post-treatment. After narcotization
by the intravenous narcotics ketamine (Atarost, Twistringen, Germany) and xylacine (Bayer
Vital, Leverkusen, Germany), the animals were positioned on an animal bed linked to the
multiphoton microscope. The eye globe was laid on a 170 µm–thick glass coverslip lying in a
specimen chamber specialized for in vivo animal experiments. The life parameters of the
being treated animals such as the blood oxygen concentration and pulse frequency were
monitored on a back extremity with a Draeger dialogue 2000 (Draegerwerke, Lübeck,

Germany) [Wang et al. 2006c and 2007c].
4. CORNEAL MULTIPHOTON IMAGING

In vivo microscopic evaluation of ocular structures has always been a challenge for
ophthalmic clinicians and microscopical researchers. Although the inherent transparency of
cornea was exploited initially by clinic instruments such as the slit lamp biomicroscope and
ophthalmoscope, high magnification ocular observations still presented many difficulties.
Clinical confocal microscopy is thought to improve the in vivo ocular imaging. While
confocal microscopy can achieve cellular imaging in the cornea, it cannot detect the
collagenous component of stroma. In addition, this technique uses a visible light source,
which causes more scattering through the media. Generally, its image quality including
spatial resolution and signal to noise is unsatisfied. So far, the micro- and nanostructures of
corneal tissue have been mostly limited to investigations with conventional histological and
electron microscopical methods [Wilson 1970, Boyde 1974, Maser et al. 1977, Sekundo et al.
1994, Bentley et al. 2001]. In these methods, fixation, sectioning and staining have to be
prepared before imaging examination. The needs to the exhaustive procedures are prone to
introduce artefacts; in the case of pathological examination, the methods do not permit in situ
sequential images of disease evolving. The in vivo studies of biological tissues using the
conventional imaging approaches are not meet the needs for basic research and clinical
diagnostics.
The introduction of multiphoton microscopy promises to lead to significant progress in
situ even in vivo investigation of tissue component with high precision and contrast
[Brakenhoff et al. 1996, König et al. 1996, Diaspro et al. 2000, Zipfel et al. 2003a, Levene et
al. 2004, Wang et al. 2005b]. Based on this technique, visualization of the intracellular
mitochondria [Barzda et al. 2005] and quantification of intracellular serotonin [Maiti et al.
1997] and NAD(P)H [Patterson et al. 2000] have been reported. It makes high-resolution
imaging of endogenous fluorophores possible and is well suited to intrinsic molecular
imaging in biological specimens.
In this chapter, the corneal tissues were firstly ex vivo examined and thereafter imaged
with living animals under multiphoton microscopy.
4.1. Optical Overview of Corneal Architecture with Subcellular Resolution

Based on Multiphoton Excitation Microscopy/Tomography
In order to make a better understanding of corneal optical imaging, histology of cornea is

briefly described in the following words. Generally, cornea comprises five layers: epithelium,
Bowman’s layer, stroma, Descemet’s layer and endothelium (Figure 5). Epithelium involves
three sublayers: (i) 2-3 layers of superficial, flat squamous cells named epithelial squamous
cells; (ii) 3-4 wing-shaped polygonal cells named epithelial wing cells; (iii) single layer of
columnar basal cells named epithelial basal cells. Oxygen can diffuse across the three
sublayers and penetrate into stroma.
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. 240 Bao-Gui Wang and Karl-Jürgen Halbhuber
Figure 5. Histological cross-section (A) and 3-D illustration of microstructural components (B) in
cornea. The figure (A) is an imitated drawing according to the histological cross-section stained with
HE showing the microstructure five layers: a, epithelium; b, Bowman’s layer; c, stroma; d, Descemet’s
layer; e, endothelium. Bowman’s layer is located underneath the basal cells and connects to the
upperside of stroma. The thin Descemet’s layer functions as the normal basement membrane of
endothelial cells and forms the scaffolding on which the endothelial cells spread themselves. The figure
(B): a, epithelial squamous cells; b, epithelial wing cells; c, epithelial basal cells; d, collagenous
lamellas; e, stromal keratocytes; f, nerve fibres. Epithelium involves three sublayers: (i) 2-3 layers of
superficial, flat cells named epithelial squamous cells; (ii) 3-4 wing-shaped polygonal cells named wing
cells; (iii) single layer of columnar basal cells named epithelial basal cells. Keratocyte is the
predominant cell and collagen is the main protein structure in extracellular matrix of corneal stroma.
The cells of endothelium are arranged in a continuous monolayer. The subbasal nerve fibres stem from
the peripheral limbus and run through stroma and Bowman’s layer and finally form the nerve terminal
endings between epithelial cells (Reprint, Dissertation, Wang, 2006).
Bowman’s layer is located directly under the basal cells and connects to the upperside of
stroma. Stroma is comprised of stromal cells and extracellular matrix involving collagens and
amorphous substance. Keratocyte, a flat cell sandwiched between the collagen lamellas, is the
predominant cell responsible for maintaining the integrity and regeneration of corneal fibrous
components. Each cell always extends five to seven cell processes connecting to adjacent
keratocytes. Descemet’s layer is the normal basement membrane of endothelial cells and
forms the scaffolding on which the endothelial cells spread themselves. The cells of
endothelium are arranged in a continuous monolayer, functioning predominantly to
maintaining the hydration and dehydration balance in stroma. Hence, endothelium is an
essential part for keeping the corneal transparency [Scharenberg 1955, Maurice 1984]. The in
situ non-linear optical detecting of the deepest-located endothelial cells was reported in this
chapter.
The 2PF of intrinsic fluorophores contained in corneal cells and the SHG of stromal
collagens can be realized when the photon intensity reaches to MW-GW/cm2 [Fine et al.
1971, So et al. 1998, Campagnola et al. 2002]. The high photon intensity was induced in the
following studies by the ultrashort laser pulses in a sub-femtoliter intrastromal focus volume
obtained with diffraction-limited objectives.
4.1.1. Multiphoton Autofluorescence Tomography of Corneal Epithelial Layers (At

Depths of 10, 25, and 41 Microns, Respectively)
Intact cornea imaging was obtained under multiphoton microscope. The three layers can
be optically differentiated based on their different cells morphology. The epithelial cells,
replenished by cell moving centrally from the limbus and anteriorly from the basal layer of
epithelium, represent a barrier against fluid loss and exogenous pathogens. As illustrated in
epithelial cells (Figure 8), the cells are closely packed and tightly adhered to each another.
Based on this feature, the cellular outline can be defined clearly. A smooth, intact, and
healthy corneal epithelium is a prerequisite for maintaining corneal transparency.
The cells can be studied with intracellular observations under the multiphoton
microscopy and show following characteristics: the fluorescent cytoplasm and the non-
fluorescent nuclei. The fluorescent cytoplasmatic granules of mitochondria encircle non-
fluorescent nuclei. Owing to the absence of suitable fluorophore at the excitation wavelength
of 760 nm, the nuclei are darkly displayed. The intracellular autofluorescence was based on
two-photon excitation of NAD(P)H, and with an emission wavelength range between 460 nm
and 480 nm [Wang et al. 2008a].
The 3-D tomography of epithelial cells was depicted (Figures 6, 7, 9). The squamous
cells are determined by flat cell bodies and oval to round non-fluorescent nuclei surrounded
by the cytoplasmic fluorescent granules stemmed from NAD(P)H located in mitochondria.
The wing cells have 3-4 polygonal cellular contours and are located in the middle epithelial
layer. The single columnar basal cells contact the submicron-scale basal membrane, which
connects the upperside of the Bowman’s layer. The black portion on this optical section was
the basement membrane of epithelial basal cells [Wang et al. 2007a]. The keratocytes in
stroma can be detected in a further depth under the Bowman’s layer (Figure 10).
The thickness of epithelium of New Zealand rabbit was determined to be approximately
38.6 ± 5.8 microns with the assistance of the multiphoton microscope.
Figure 6. In vivo intrinsic 2PF image of epithelial squamous cells in cornea. (A). Epithelium at a
corneal depth of 10 micrometers with high subcellular spatial resolution and high signal-to-noise ratio
at an excitation wavelength of 760 nm. The squamous cells, possessing flat cell bodies and oval to
round nuclei, are characterized by the fluorescent cytoplasmatic granules stemming from mitochondria
and non-fluorescent nuclei displayed darkly on the optical section. The fluorescent granules encircle
non-fluorescent nuclei; (B). Observation of one cell with high magnification. The endogenous
fluorophore was stimulated substantially under two-photon excitation. Scale bar: 10 microns (Reprint
Lasers Surg Med 2007).
Figure 7. In vivo intrinsic 2PF imaging of epithelial wing cells in cornea. (A). Epithelial cells at a
corneal depth of 25 microns excited at a wavelength of 760 nm. The fluorescence stemming from
mitochondria encircle non-fluorescent nuclei. Based on the multiphoton tomography, the cells were
determined to have 3-4 wing-shaped polygonal cellular contours and were located in the middle
epithelial layer. The fluorescence based on NAD(P)H in mitochondria was verified by emission
spectrum. (B). Observation of one cell with high magnification. Scale bar: 25 microns (Reprint Lasers
Surg Med 2007).
Figure 8. Emission spectrum of intrinsic 2PF images of epithelial cells at a depth of 28.4 microns in a
cornea with high signal-to-background ratio. (A). The fluorescent granules of mitochondria and the
non-fluorescent dark nuclei can be identified easily without any staining assistance; (B). The high
magnification of cells contour and fluorescent granules, revealing the bioenergetics of mitochondria in
a certain living period; Histogram (C) reveals that the predominant spectral profile was located between
460 nm and 480 nm (y-axes: amplitude), which corresponds to the eliciting intracellular NAD(P)H
fluorescence. Knowledge of the spectrum provide a novel signature for further analysis of molecular
entities emitting these fluorescent signatures and would have a use in differentiating the NAD(P)H
autofluorescence from other intrinsic fluorophores such as flavin, lipofuscin, and porphyrins in tissues
and cells. Scale bar: 30 microns (Reprint Lasers Surg Med 2008).
Figure 9. In vivo intrinsic 2PF image of epithelial basal cells at a corneal depth of 41 micrometers
excited at a wavelength of 760 nm. The single columnar basal cells contact the basement membrane,
which connects the upperside of the stroma. The non-fluorescent portions of this image are
corresponding to the collagen in basement membrane, which forms the scaffolding on which the basal
cells spread themselves. Scale bar: 20 microns (Reprint Lasers Surg Med 2007).
The emission spectra of intracellular autofluorescence were measured. Knowledge of the

spectrum provides an interesting signature for further analysis of the molecular entities
emitting these fluorescent signatures and would have a use in differentiating the NAD(P)H
autofluorescence from other intrinsic fluorophores such as flavin, lipofuscin, and porphyrins.
4.1.2. Revealing of Stromal Intact Cellular and Collagenous Networks Based on

Multiphoton Excitation Imaging
(i) Selective Displaying of Intratissue Keratocyte Networks in Stroma with

Intracellular Spatial Resolution
Keratocytes are the principal corneal stromal cell. By far, the morphology of the cells was
examined by electron microscopy [Nishida et al. 1988]. Keratocyte networks were visualized
using vital dyes [Poole et al. 1993]. However, none of these methods allows direct
visualization of keratocytes in the intact bulk corneal tissue.
With the assistance of the excited endogenous cytoplasmatic fluorophore, the in situ
visualization of intratissue cells becomes realizable, which overcomes the artefacts always
occurring during the histological preparation. The artefacts would involve disturb of the
topography and morphology of microstructures. More interesting, the sample tissues can be
scanned numerous times without observable morphological photo-impairment under
multiphoton microscope.
Owing to the absence of collagen lamellas at the excitation wavelength of 730 nm, the
stroma cells can be studied selectively (Figures 10_A, 11, 12, 13). The keratocytes can be
identified through the fluorescent cytoplasm and dark non-fluorescent nuclei. Owing to the
high spatial resolution, each cell can be studied with morphological details. The cells are
connected with each other through the fluorescent processes. Keratocytes, the modified tissue
fibroblasts, possess large, flattened and eccentric nuclei (Figure 11). The cell bodies had a
well-defined satellite shape with long, fine and generally non-branched cell processes
projecting from their apices [Wang et al. 2006d]. The processes of neighbouring fibroblasts
contacted with each other to form an extensive and continuous cellular network structure.
The cellular processes functioning as communication of signal and exchange of nutritive
substances displayed in Figure 12_A are comparable to the conventional histology in Figure
12_B. The cellular network appears to constitute an integrated system, which may be
involved in the synchronized regulation of the metabolic and physiologic homeostasis for the
maintenance of corneal transparency. Based on the ultrahigh spatial resolution and 3-D
visualization, the size of keratocytes was determined. Mean value of keratocytes based on
measurement of more than 1000 cells in 50 ex vivo porcine eye globes: 23.48 ± 3.66 x 16.58
± 2.68 x 11.8 ± 2.8 microns (only the cell bodies were measured, up to 140 µm depth of
stroma with z-adjusting; eye globes of adult porcine; measurement within one hour after
slaughtering). The cell morphology was also in vivo studied with the high-resolution optical
images.
The mean volume value of keratocytes was determined based on measurements of more
than 100 cells: 18.48 ± 3.46 x 13.53 ± 2.87 x 9.7 ± 1.8 microns (only cell bodies were
measured, at depths of up to 140 microns, rabbits) [Wang et al. 2007a and 2008a].
Figure 10. Multiphoton excitation images of corneal stroma performed with different excitation
wavelengths at a depth of 118.7 micrometers (A, B, C). (A). Keratocytes image based on the
cytoplasmatic NAD(P)H in mitochondria at an excitation wavelength of 730 nm; Owing to the absence
of collagen lamellas, the stromal cells can be selectively studied. One keratocyte with three cellular
processes is marked with an arrow head. Using this type of high-resolution non-linear optical imaging,
the cellular details were in vivo studied. The mean size of keratocytes based on a measurement of more
than 100 cells: 18.48 ± 3.46 x 13.53 ± 2.87 x 9.7 ± 1.8 microns (measurement of cell body without
processes, up to 140 µm depth in stroma, rabbits); (B). Collagen lamellas based on SHG detection at an
excitation wavelength of 840 nm. SHG enables the non-centrosymmetric biological structures to be
directly imaged. The intrastromal collagen lamellas and filaments can be selectively visualized by SHG
signal imaging due to the deficiency of keratocytes autofluorescence at this excitation wavelength; (C).
Overlaying of 730 nm and 840 nm excitation images discloses the microstructural co-localization and
topography of keratocytes and their hosting collagen lamellas in situ (with pseudocolours). Green:
autofluorescence; Red: SHG. The keratocytes, playing an important role in the metabolic and
physiologic homeostasis, are folded into the labyrinths of collagen fibrils. Only based on two different
wavelengths, keratocytes and collagen lamellas yield two distinct features for a same region; (D). A
transmission electron microgram (TEM) of corneal stroma was enclosed to complementarily reveal the
ultrastructural topography between cellular and extracellular components. Keratocytes, incorporating
into corneal stroma, are displayed that they are sandwiched loosely in part among the collagen lamellas.
Scale bars: 40 microns (upper) for images A, B, C; 500 nm (lower) for EM (Reprint Lasers Surg Med
2007).
Figure 11. High spatially-resolved stromal keratocytes at a depth of 80 µm. The stromal cell can be
easily identified through the fluorescent cytoplasm and dark non-fluorescent nuclei. Based on the
deficiency of collagen signal at this excitation wavelength, cellular boundaries can be clearly
determined. The cellular density can be studied. Eight cells in 100 x 120 µm2 area are counted. Thanks
to the high spatial resolution, each cell can be morphologically studied in details. The cells possess
large, flattened and eccentric nuclei; the cell bodies had a well-defined satellite shape with long, fine
and generally non-branched cell processes projecting from their apices. The processes of neighbouring
fibroblasts came into contact with each other to form a continuous network structure. Remarkably, no
cellular photobleaching or photodisruptive phenomena took place during the high-zoomed imaging
under multiphoton microscope. Scale bar: 20 microns (Reprint Ann Anat 2006).
Figure 12. Network structure of keratocytes formed by the cellular processes elucidated by two
approaches: non-linear optical imaging (A) and conventional histology of gold impregnation (B). The
cell bodies had a well-defined satellite shape with long, fine and generally non-branched cell processes
projecting from their apices. The processes of neighbouring fibroblasts contacted with each other to
form an extensive and continuous network structure. The in situ selectivity of non-linear optical
imaging nicely coincides with the in vitro conventional histology. The cellular network appears to
constitute an integrated system which may be involved in the synchronized regulation of the metabolic
and physiologic homeostasis for the maintenance of corneal transparency. Scale bar: 30 microns
(Reprint Histochem Cell Biol 2006).
(ii) SHG Signal Imaging of Collagen Lamellas in Corneal Stroma

Collagen is the main protein structure in corneal stroma. The collagen filaments can be
excited under multiphoton microscope at wavelengths between 800 and 860 nm, with
emission lights equal to half of the incident lights [Williams et al. 2005]. The intrastromal
collagen lamellas and filaments can be selectively visualized by second harmonic imaging
[Wang et al. 2005c]. SHG enables the anisotropic non-centrosymmetric biological structures,
possessing large hyperpolarizabilities under high laser irradiance, to be directly imaged [Roth
et al. 1979, Freund et al. 1986]. Owing to its own non-centrosymmetric structure of collagen
fibrils, a triple helix composed of three protein chains, collagen is thereby a well-documented
source of tissue SHG imaging. Corneal stroma contains various types of collagen, each one
being distinguished from the others by the biochemical composition of three protein chains.
The predominant collagen in normal cornea is type I and VI; type IV is characteristic of
basement membranes; type VII constitutes the anchoring fibrils of epithelium; type III
becomes much more prominent during wound healing. Some other types, such as collagens
XI and XVIII, which do not exist in the normal physiological state of cornea, appear only
during the wound reparative phase [Light 1979, Tseng et al. 1982, Nakayasu et al. 1986,
Zimmermann et al. 1986, Kato et al. 2003]. Collagen has the intrinsic capability of frequency
doubling under high light irradiance based on its special biochemical composition. Membrane
imaging and visualizing of biomolecular array with the assistance of SHG signal imaging
were reported [Moreaux et al. 2000, Campagnola et al. 2002]. Similar to the occurring
conditions of two-photon excited fluorescence, SHG takes place in a minute focal volume at
high laser irradiance, permitting high resolution 3-D optical sectioning of collagenous
component in thick intact tissues [Campagnola et al. 2003].
As illustrated in Figure 10_B, distribution and orientation of the collagen lamellas are
clearly revealed. The lamellas are densely packed with characteristic fluctuations. The in vivo
optical imaging provides reliable information about the characteristics of the cornea, the
details of which coincide with the conventional histological architecture [Freemann et
al.1978]. With the assistance of SHG, the in vivo 3-D tomography of collagen lamella in
cornea with high resolution becomes feasible for the first time (Figure 14). The 3-D
architecture of collagen lamellas can be visualized in conjunction with the z-positioning. The
collagen can be selectively visualized due to the deficiency of keratocyte autofluorescence at
the wavelength 830 nm. The lamellas run parallel to the corneal surface. However, in the x-y
dimension, they show different orientations, especially on the surface of keratocytes [Wang et
al. 2007a].
As a rather simple and straightforward imaging method, SHG signal imaging is
invaluable for studying the assembly of collagen fibrils in bulk tissues. In addition to the
fundamental study of collagen fibrils in different tissues, SHG imaging may have a clinical
use for early diagnostics of cornea-related diseases such as stromal edema, which is
characterized by disturbing the regularity of collagen fibrils and the groove-like order of
collagen lamellas. Based on its selectively displaying function, SHG could also be of great
merit to in vivo monitor collagen regeneration in the wound healing process after laser
refractive surgery. Spectral imaging verified that both cornea and sclera have the SHG
emission wavelength at 420 nm (Figure 18). In comparison with sclera, collagens in cornea
are arranged in a rather organized fashion of lamellas packing with a depth-dependent
variation (Figure 14 and Figure 18).
Figure 13. Three-dimensional non-linear optical tomography of stromal keratocytes with vertical z-
interval of 5 micrometers based on the two-photon autofluorescence imaging at a wavelength of 760
nm. Mean value of keratocytes based on 3-D measurement of more than 1000 cells in ex vivo porcine
eye globes: 23.48 ± 3.66 x 16.58 ± 2.68 x 11.8 ± 2.8 microns (only the cell bodies were measured, up to
140 micrometers depth of stroma with z-adjusting; eye globes of adult porcine; measurement within
one hour after slaughtering). Scale bar: 15 microns (Reprint Ann Anat 2006).
The spaces of inter- and intra-collagenous lamellas are always filled with keratocytes and
their processes. The imaging specificity of SHG allowed to differentiating the both ocular
tissues. More interesting, it offered a high degree of contrast enhancement in the topography
between cellular and collagenous components.

The in situ knowledge of collagen is of crucial importance in the anatomy and pathology
of the connective tissues. SHG microscopy has promise for imaging collagen organization
and provides high resolution and more detailed images of collagen distribution. Multiphoton
microscopy has thereby the potential to become one of the most valuable tools in situ even in
vivo microscopy of tissue collagen and could be even employed to inspect collagen-related
alterations such as corneal fibrosis and cellular infiltration in melanoma.
(iii) Microstructural Co-Localization and Topography of Keratocytes and Their

Hosting Collagen Lamellas in Corneal Stroma Based on 2PF and SHG
The microstructural co-localization of corneal stroma was displayed with different
excitation wavelengths in a depth of 118.7 microns (Figure 10_C). Keratocytes and collagen
lamellas can be selectively displayed only with the in-tandem assistance of two different laser
wavelengths based on the non-linear optical effects. The keratocytes, folded into the
labyrinths of collagen fibrils and playing an important role in the metabolic and physiologic
homeostasis, are major cellular components of the corneal stroma. They can be identified
through the fluorescent cytoplasm and non-fluorescent nuclei in the multiphoton imaging.
Figure 14. In vivo corneal collagen lamellas 3-D tomography in conjunction with a z-interval of 5
micrometers based on the SHG signal imaging in a backward direction. SHG takes place in a minute
focal volume, permitting high vertical resolution of collagenous component in thick intact tissues. The
distribution and orientation of the collagen lamellas is clearly revealed. In correlation with the z-
positioning, the 3-D architecture of intratissue collagen can be studied. The lamellas are densely packed
with characteristic fluctuations and run parallel to the corneal surface. However, they show different
orientations in the x-y dimension, especially on the surface of keratocytes. In comparison with SHG
image in sclera (Fig.18), corneal collagen is arranged in an organized pattern of lamellas. As a rather
simple and straightforward imaging method, the SHG tomography can be used clinically for evaluating
the cornea fibrosis and stromal edema. Scale bar: 10 microns (Reprint Ann Anat 2006).
The fluorescent cytoplasm was based on intrinsic NAD(P)H in mitochondria excited at

730 nm. Thus, the high-resolution optical images acquisition allowed the cell morphology to
be in situ studied precisely. In details, their long cytoplasmic processes came into contact with
those of neighboring keratocytes. The cytoplasmatic autofluorescence revealed the
bioenergetics of mitochondria in their different living periods [Wang et al. 2008a].
As seen in Figure 10_D, a transmission electron microgram (TEM) of corneal stroma was
enclosed to complementarily elucidate the ultrastructural topography between cellular and
extracellular components in this relationship. Keratocytes, incorporating into corneal stroma,
are displayed in this figure that they are sandwiched loosely in part among the collagen
lamellas.
4.1.3. Multiphoton Excitation Detection of Endothelial Cells at a Corneal Depth of 369

Microns with Living Animals
Endothelium is located in the deepest layer of cornea. The thickness of cornea of the New
Zealand rabbit was determined to be 380 ± 18 microns measured with a pachometer (Corneo-
Gagetm plus 2, Sonogage, U.S.). The objective (20x 0.95 N.A., water) with a work distance
of more than 1000 micrometers in bulk tissue used in this chapter allowed the deep
endothelial layer of cornea to be visualized (Figure 15).
Figure 15. In vivo intrinsic 2PF imaging of endothelial cells. (A). Autofluorescence image of
endothelial cells taken at an excitation wavelength of 760 nm. The objective (20x 0.95 N.A., water)
with a work distance of more than 1000 microns in bulk tissue used in this chapter allowed the deepest
endothelial layer of cornea to be clearly visualized. The cells are featured by the filmy cytoplasm and
large round to oval nuclei. The paving-stone mosaic pattern of the hexagonal endothelial cells shows
only slight variation in size and shape based on the in vivo non-linear optical tomography. The
thickness of the continuous monolayer endothelial cells was determined to be 8.4 ± 2.1 microns. The
homogenous cells have an average value of 7.38 ± 1.25 microns in diameter. The optical imaging of
endothelial cells makes it possible to on-line monitor endothelial cells during the intrastromal surgery;
(B). Observation with high magnification. Scale bar: 10 microns (Reprint Lasers Surg Med 2007).
The paving-stone mosaic pattern of the hexagonal endothelial cells under multiphoton
microscope in vivo shows only slight variation in size and shape. The volume of endothelial
cells was measured with the non-linear optical tomography. The thickness of the continuous
monolayer endothelial cells was determined to be 8.4 ± 2.1 microns. The homogenous cells
have an average value of 7.38 ± 1.25 microns in diameter. However, the values of in situ
optical measurement here are not in agreement with the conventional histological results. The
endothelial cells with thickness of 4–6 micrometers and diameter of 8–11 micrometers were
reported [Maurice, 1984]. The reasons for this difference might involve (i) different values
between in vivo (non-linear optical tomography) and in vitro (histology); (ii) different species
between human (histology) and rabbit (in this chapter); (iii) the refractive index mismatch of
optical measurement. As to the phenomenon of refractive index mismatch, geometric
distortions between the sample and the medium play a role in the image deterioration. In the
fact, this optical mismatch interferes with the real size of cells. The vertical mismatch is yet
not corrected automatically by the microscopic operation system. We measured the thickness
of cells located in tissues between utmost-upper point and utmost-lower point of the cells.
The size of cells presented here provides the examination outcomes based on laser
microscopy [Wang et al. 2007a].
The optical imaging of endothelial cells makes it possible to on-line monitor endothelial
cells during the intrastromal surgery and to conduct follow-up observations on these cells
postoperatively. To date, deep lamellar endothelial keratoplasty (DLEK) has been used as an
innovative approach to treat Fuchs’ endothelial dystrophy (Terry et al. 2005). The endothelial
cells from a donor are implanted to a recipient, whose endothelium is functionally insufficient
to keep the corneal hydration and dehydration balance. The grafting endothelial cells should
be pre- and post-operatively monitored. Assisted by multiphoton microscopy, the in vivo
evaluation and monitoring of the implanted endothelium with subcellular spatial resolution
become easier.
4.2. Optical Determination of Bowman’s Layer with in-Tandem Combination

Assistance of 2PF and SHG Signal imaging
Bowman’s layer was described in humans by William Bowman in 1947. The author
defined it as a homogeneous layer between epithelium and substantia propria in light
microscopy, a lusterless but perfectly transparent, continuous, homogenous lamina that did
not appear to have any internal structure. Camber et al. (1987) found that in the case of the
pig, Bowman’s layer was unequal and corrugated in form, thus making it difficult to
distinguish it from the surrounding fine collagenous fibres. Stained with hematoxylin-eosin
(HE-staining), Merindano et al. (2002) reported that the pig appears to possess no Bowman’s
layer.
On contrary, Hayashi et al. (2002) reported that pig possess Bowman’s layer in cornea.
The border between Bowman’s layer and the substantia propria is clearly defined in guinea
pig. The diameters of the collagen fibrils in Bowman’s layer and the substantia propria are
almost the same. However, the existing morphological studies of cornea offer little
information about the morphometry of Bowman’s layer. The nonlinear optical imaging was
used here to detect this layer and conduct the morphometric study.
Bowman’s layer is comprised of collagen [Maurice 1984]. The epithelial basal cells
connecting to the upper side of Bowman’s layer are fluorescent at the excitation wavelength
of 730 nm (Figure 16_A). When the laser focus was adjusted further deeper to the stromal
layer, the keratocytes would be displayed based on its autofluorescence.
On Figure 16_B, the collagen components in Bowman’s layer between basal cells and
stroma are demonstrated based on SHG signal imaging of collagen at the incident wavelength
of 840 nm. With the assistance of tissue optical tomography based on in-tandem detection
with 2PF and SHG, the Bowman’s layer can be optically defined, where SHG is there but
2PF absent.
By contrast, basal cells show 2PF but no SHG; Stroma has SHG and also 2PF. As a
result, the collagen depth without cellular structure should be the Bowman’s layer located
between basal cells and keratocytes. The thickness of this collagen component between
epithelial basal cells and stroma was determined to be approximately 18.6 ± 3.6 microns
(porcine eye ex-vivo, cornea in situ).
Figure 16. Optical determination of Bowman’s layer and nerve fibres in cornea. The determination of
Bowman’s layer based on the in-tandem detection with 2PF (A) and SHG signal imaging (B). (C). The
nerve fibres of subbasal plexus can be displayed only based on the reflection optical imaging. The three
figures (A-C) were taken at a same intracorneal plane at a depth of 48.88 µm. Portion a in (C)
corresponds to the portion a in (A); portion b in (C) corresponds to the portion b in (B). (D). High
magnification of nerve fibres showing the beadings (arrow heads on D) of subbasal nerve fibres. With
the assistance of the three figures (A-C) based on the different imaging approaches, the collagen layer
between the epithelial basal cells and stroma can be definitely determined. The predominant component
of Bowman’s layer is collagen. The epithelial basal cells are fluorescent at the excitation wavelength of
730 nm (Figure A). When the laser focus was adjusted further deeper to the stromal layer, the
keratocytes would be imaged based on its autofluorescence. In Fig.B, the collagen component in
Bowman’s layer between basal cells and stroma is demonstrated based on SHG signal imaging of
collagen at the incident wavelength of 840 nm. With the assistance of tissue optical tomography based
on in-tandem detection with the both approaches, the Bowman’s layer can be optically defined, where
SHG is there but 2PF absent. By contrast, basal cells have 2PG but no SHG; Stroma has SHG and also
2PF. As a result, the collagen layer without cellular structure should be the Bowman’s layer located
between basal cells and keratocytes. The thickness of this collagenous component between epithelial
basal cells and stroma was determined to be approximately 18.68 ± 3.66 microns (porcine eye ex-vivo,
cornea in-situ). The nerve fibres in cornea were studied with 2PF, SHG, and reflection (C, D) imaging
approaches as well as gold chloride-staining (E). Arrow heads in figure (E) show the nerve fibres in
stroma. It was determined that the nerve fibers in cornea can not be excited under non-linear optical
imaging (A and B). The beading-like subbasal nerve fibres are displayed on the reflexion image (Figure
C). The nerve fibres can be followed three-dimensionally based on optical tomography. They stem from
the peripheral limbus and run through stroma and Bowman’s layer and finally form the nerve terminal
endings between epithelial cells which can no more be observed under this light technique. Most of
these visible fibres distribute in the anterior one third corneal thickness. In the posterior half of cornea,
these fibres were seen sparsely. The subbasal nerve fibres comprise beading area and interbeading area
(Figure D). The size of the fibres in beading area is proximately 1.5-2 folds larger than that of the
interbeading area. The mean diameter of subbasal nerve fibres was measured with a mean value of 1.88
± 0.39 microns (subbasal nerve fibres, 50 porcine eyes ex-vivo, reflexion imaging). Scale bars: 30
microns for A-C; 50 microns for D; 20 microns for E.
4.3. Optical Observations of Nerve Fibres in Cornea
The nerve fibers in cornea were examined with three approaches: reflexion, 2PF and
SHG. Through comparison of the three figures (Figures 16 _A, B, C), it has been postulated
that neither at the incident wavelength of 730 nm nor at 840 nm can the nerve fibres of
subbasal plexus be displayed under the condition of multiphoton absorption. The beading-like
subbasal nerve fibres are displayed only in the reflexion image of Figure 16_C. The nerve
fibres can be visualized three-dimensionally based on this imaging approach. They stem from
the peripheral limbus and run through Bowman’s layer and form the nerve terminal endings
between the epithelial cells. Most of these visible fibres distribute in the anterior one third
corneal thickness. In the posterior half of cornea, these fibres were sparsely seen. The
subbasal nerve fibres comprise beading area and interbeading area (Figure 16_D). The fibre
diameter in the areas of beading and in the areas between beadings was measured in this
work. The size of fibres in beading area is proximately 1.5-2 folds larger than that of the
interbeading parts. The mean diameter of subbasal nerve fibres was measured with mean
value of 1.88 ± 0.39 microns (subbasal nerve fibres, 50 porcine eyes ex vivo, reflexion
imaging). To be concluded, the nerve fibers in cornea can not be displayed under non-linear
optical imaging condition.
Non-linear optical images of the full-thickness of cornea were obtained with multiphoton
microscopy based on 2PF and SHG signal imaging. Epithelial cells, stroma, and endothelial
cells were identified and displayed with the assistance of the non-linear optical imaging
technique. Qualitative and quantitative studies of the corneal cells were conducted. In
particular, corneal Bowman’s layer was optically determined in situ. Our results of non-
destructive selective visualization of various microstructural components in corneal tissue
show that multiphoton microscopy is becoming a powerful tool in advancing our
understanding of corneal optical tomography and biomechanics. By means of the in vivo
feasible microscopy, the cell reaction during wound repair in laser intrastromal nanosurgery is
investigated and reported in another following session. In further studies, the fluctuation
distribution of the strain collagen in tissue will be computerized and tried to find the
correlation with the morphometry of individual filaments and collagenous lamellas in cornea.
5. SCLERAL MULTIPHOTON IMAGING: IN SITU NON-LINEAR

OPTICAL BIOPSY OF WHOLE-MOUNT SCLERAL TISSUE
5.1. Selective Displaying of Fibrocytes Based on 2PF at an Excitation
Wavelength of 730 nm
Two-photon autofluorescence image of fibroblasts, the main cellular composition in

scleral tissue, was acquired at an excitation wavelength of 730 nm (Figure 17_A).
Figure 17. Multiphoton excitation imaging at a depth of 93.9 micrometers in scleral tissue performed
with different excitation wavelengths to elucidate fibrocytes, capillaries, elastic fibres, and collagen
filaments. The images give snapshots of cell morphology and extracellular matrix organization. The
imaging plane was parallel to the sclera surface. The emitted lights, including 2PF and SHG, were
deflected backward in the direction towards PMT so that the non-linear optical images were recorded.
(A). Fibroblasts and mature elastic fibers based on its blue/green autofluorescence at an excitation
wavelength of 730 nm were detected. The non-fluorescent nuclei of fibroblasts displayed darkly on the
optical sections are surrounded by the fluorescent granules. The cellular boundary can be determined
based on the fluorescent cytoplasm and the non-stimulated extracellular collagens. Using z-positioning,
the vertical dimension of the cells was determined. The size of the cells on this site was 17.68 ± 2.36 x
14.33 ± 3.89 x 8.7 ± 1.5 microns (z dimension). Besides the fibrocytes, the intratissue elastic fibers and
capillaries were also in situ displayed based on 2PF signal imaging; (B). Collagenous fibers image
based on SHG detection at an excitation wavelength of 840 nm were obtained by ultraviolet/violet
radiation in a backward direction. Most of collagen bundles run parallel to the sclera’s surface. They
can be selectively demonstrated due to the autofluorescence deficiency of fibroblasts and elastic fibres
at the excitation wavelength of 840 nm. The imaging specificity of SHG allowed collagen orientation to
be well identified; (C). Co-localization of cellular and fibrous microstructures. Overlaying of 730 nm
and 840 nm excitation images reveals the in situ state topography of elastic fibers, fibroblasts and
collagen filaments without the assistance of any staining or slicing (with pseudocolours). Green:
autofluorescence at the wavelength of 730 nm; Red: SHG at 840 nm. The rare elastic fibers were
embedded in the hosting framework of collagen filaments. The two figures (A and B) were imaged at a
same optical plane with a time-shifting of 8 seconds. Scale bars: 40 microns. The symbols on Fig. A:
fb:fibroblasts; f:elastic fibers. The novel non-invasive imaging opens the general possibility of high-
resolution in situ 3-D imaging of elastic fibers and collagen structures in connective tissues. This
chapter demonstrated that the contrast for both intratissue cells and collagen fibers was sufficiently
generated only based on the in-tandem detection with 2PF and SHG. The high-degree contrast offers a
snapshot of clear topography between cellular and collagenous components. This proof-of-principle
examination has already started to reveal intricate features that are otherwise lost in conventional
histology studies, suggesting the potential of using non-linear optical imaging as a routine tool for more
advanced studies (Reprint Lasers Surg Med 2008).
The imaging plane was parallel to the sclera surface. The intracellular autofluorescence
was based on the two-photon process of NAD(P)H, which was located in the mitochondria of
fibroblasts. With intracellular observations, the non-fluorescent nuclei displayed darkly on the
optical sections surrounded by the fluorescent mitochondria can be clearly seen. Thus, the
cellular boundary can be clearly determined based on the fluorescent cytoplasmatic granules
and the non-stimulated extracellular collagens. Furthermore, the cytometric measurement of
fibroblasts based on these image features was done. Using z-positioning, the vertical
dimension of cells could be determined. The cells on this site have a average size of 17.68 ±
2.36 x 14.33 ± 3.89 x 8.7 ± 1.5 microns (z) [Wang et al. 2008a].
In addition to the usual cellular component of fibroblasts, a few melanocytes are
generally present in the deep layer of sclerae. These rare pigment-containing cells,
interspersed between amounts of fibroblasts, have similar morphology to fibroblasts such as

dendritic or tripolar cell bodies and prominent nuclei with little cytoplasm. These cells could
be identified with the assistance of the special spectral imaging based on the contained
pigment granules. The cells displayed in Figure 17 were determined to be the fibroblasts since
our optical plane was located at a one-third depth of sclera.
5.2. Selective Displaying of Collagen Based on SHG at an Excitation

Wavelength of 840 nm
SHG signal image of the collagen, in which the fibroblasts host, was obtained at an
excitation wavelength of 840 nm (Figure 17_B). The sclera SHG image acquisition at a depth
of 93.9 microns in a backward direction is displaying the in situ collagen filaments and
bundles. Most of collagen bundles run parallel to the sclera’s surface. They can be selectively
demonstrated due to the autofluorescence deficiency of fibroblasts at the excitation
wavelength of 840 nm. With the assistance of SHG signal imaging, the in vivo tomography of
collagen fibres in sclerae with high resolution can also be performed (Figure 18).
Figure 18. SHG signal imaging tomography and emission spectrum in sclera. The tomogram was
performed with a vertical interval of 5 microns in a backward direction showing collagenous filaments
at an excitation wavelength of 840 nm. A: at a depth of 76.9 micrometers. Histogram showing the SHG
emission wavelength of 420 nm (y-axes: SH amplitude). Scale bars: 20 microns. SHG 3-D tomography
of collagenous filaments was taken at depths of between 76.9 µm and 91.9 µm in the whole-mount
scleral tissue. The optical sectioning was performed parallel to the coverslip surface. Distribution and
orientation of the collagen bundles were revealed clearly. The cross-linking collagen fibres (marking
with rounds in Fig.B) with different extending directions (arrows in Fig.B) are packed randomly and
densely. Comparing to the SHG images in cornea (Fig.14), the morphological difference is significant.
The non-linear optical tomography shows that no collagen lamellas are imaged in sclerae. The non-
homogeneous tubelike structures of collagen filaments and irregular arrangement of collagen fibers in
sclerae would most likely contribute to this striking SHG signal image discrepancy between sclera and
cornea (Reprint Lasers Surg Med 2008).
Thus, the 3-D of collagen architectures can be studied in correlation with z-positioning
[Wang et al. 2008a]. 3-D tomography of collagenous component in sclerae with a vertical
distance of 5 microns was acquired (Figure 18). SHG images of collagen fibrils were taken in
a backward direction at depths of between 76.9 and 91.9 microns in the whole-mount scleral
tissue. Distribution and orientation of the collagen bundles were revealed clearly. The cross-
linking collagen fibres, lacking of lamellar arrangement like in cornea, are packed randomly
and densely.
5.3. Non-Linear Optical Comparison of Collagen Organisation in Scleral and

Corneal Tissues
Spectral imaging verified that both cornea and sclera have the SHG emission wavelength
at 420 nm (Figure 18).However, significant SHG image discrepancies between the
collagenous structure in sclera and cornea were observed (comparing Figure 18 in sclera with
Figure 14 in cornea). In contrast to collagen lamellas in cornea, the backward SHG signals in
sclerae were more significant. The cornea collagen is arranged in an organized fashion
packing with a depth-dependent variation. The spaces of inter- and intra-collagenous lamellas
are always filled with keratocytes and their processes. The collagen fibres in sclera showing
different direction connect to each other.
The SHG radiation pattern is mainly determined by the phase-matching condition. The
part of the backward SHG signals may be attributed to the backscattering of the forward SHG
signals from the highly scattered scleral tissue [Han et al. 2005]. The non-homogeneous
collagen fibrils and the irregular arrangement of collagen fibers in sclerae would most likely
contribute to this striking non-linear optical image discrepancy between sclera and cornea.
5.4. In Situ Studies of Intratissue Capillary Based on Multiphoton Excitation

Imaging
The capillaries imaged were embedded in scleral tissue located in a physiologic-similar

circumstance.
Two-photon autofluorescence imaging provides us with qualitative and quantitative
assessment of the capillaries (Figure 19). As regards to a peripheral circulation, blood flowing
from arteries narrows into arterioles, and then narrows further into capillaries. The emitted
2PF lights were deflected backward in the direction towards the PMT so that the non-linear
optical images could be recorded. The intact capillary displayed has a diameter of 40.63 ±
4.65 microns and the other optically-sliced one, in which some of the blood cells are located
and fluorescent, has a diameter of 25.53 ± 2.98 microns. The flat and pancake-shaped
monolayer endothelial cells, a layer of fairly thin specialized epithelium, are featured by the
filmy cytoplasm and the large round nuclei. The size of the endothelial cells was measured
with values of 10.46 ± 1.34 x 8.35 ± 1.67 x 5.7 ± 0.7 microns (z) [Wang et al. 2007a].
The in situ imaging of intratissue capillary in intact bulk tissue would be interesting for
the morphological study of microangiopathy such as in the pathophysiology of diabetes
mellitus.
Figure 19. Intrinsic 2PF imaging of intratissue blood capillaries. The image was performed at a depth of
93.9 microns with an excitation wavelength of 730 nm. The elastic fibers and fibrocytes are also
displayed under the multiphoton microscopy. The sparse elastic fibers distributed perivascularly with a
density of less than 20 pieces in a 200 x 200 µm2 area. Collagen filaments were absent in this figure.
The capillaries imaged were embedded in scleral tissue located in a physiologic-similar circumstance.
The two-photon autofluorescence imaging provides us with qualitative and quantitative assessment of
the intratissue capillaries. The intact capillary displayed has a diameter of 40.63 ± 4.65 microns and the
other optically-sliced one, in which some of the blood cells were located and fluorescent, has a diameter
of 25.53 ± 2.98 microns. The wall of each capillary comprises monolayer of endothelial cells. The flat
and pancake-shaped endothelial cells, a layer of fairly thin specialized epithelium, are featured by the
filmy cytoplasm and the large round nuclei. The size of the endothelial cells was measured with size of
10.46 ± 1.34 x 8.35 ± 1.67 x 5.7 ± 0.7 microns (z dimension). The symbols: f:elastic fibers;
e:endothelial cells of blood capillaries; fb:fibroblasts. Scale bar: 30 microns (Reprint Lasers Surg Med
2008).
5.5. In situ Optical Differentiation of Intrascleral Elastic Fibers from

Collagens Based on Multiphoton Excitation Imaging
The sparse mature elastic fibres embedded in collagenous networks were also discernible
based on its emission blue/green autofluorescence under multiphoton microscope. By
contrast, collagen in the tough layer had an emission of violet radiation (420 nm) by SHG
imaging (Figures 17 and 20). Elastic fibers are one of the main fibrous components of
connective tissues. They, providing elastic recoil to tissues, are an abundant and integral part
of many extracellular matrices, in which they also provide resilience and deformability to
tissues.
Figure 20. Autofluorescence of elastic fibers in cryosection of sclera at an excitation wavelength of 730
nm. Elastic fibers stemming from sclera run centrically and penetrate finally into corneal limbus. Arrow
heads just show the sparse elastic fibers. Scale bar; 20 microns.
The major component of elastic fibers is the protein elastin, an insoluble rubberlike
fibrous polymer responsible for the characteristic elastic properties. It plays an important role
in providing mechanical strength for connective tissues. Collagens are the framework of
connective tissues and provide the main determinants of its mechanical behaviour in tension.
The elastic function complements collagen fibrils, which impart the connective tissues tensile
strength. Ultrastructurally, elastic fibers possess complex structures comprising an amorphous
core of extensively cross-linked elastin, surrounded by fibrillin-rich microfibrils. Cross-
linking between collagen and elastin provides necessity for the maintenance of proper
function of different tissues [Brown-Augsburger et al. 1994].
By far, most of our knowledge about elastic fibers is limited to in vitro studies, which
require the enucleated biopsies to be further processed: histological slicing, embedding,
fixation and finally staining with different procedures such as Silverman-Movat pentachrome
[Elbadawi 1976], Weigert's resorcin–fuchsin, Hart’s elastica-staining or immunogold
antibody labelling [Farguharson et al. 1991]. With the advent of fluorescence microscope and
laser scanning microscopic technique, the autofluorescence of elastic fibers can be recorded
using ultraviolet (UV) excitation radiation based on single photon absorption. However, this
approach is unable to differentiate the elastic fibers in bulk connective tissues from the
framework of collagen filaments based on their intrinsic fluorescence. The novel non-invasive
multiphoton imaging technology opens the general possibility of high-resolution in situ
imaging of elastic fibers in connective tissues without any assistance of staining. Recently, it
is reported that the intratissue elastic fibers can be detected in heart valves based on
multiphoton autofluorescence by near infrared femtosecond laser pulses [Schenke-Layland et
al. 2004, König et al. 2005a]. We attempted to circumvent this limitation of single photon
laser scanning microscopy by using this high technology to differentiate the elastic fibers
from the collagen fibrils distributing in bulk scleral tissue.
The co-localization of elastic fibers and collagen filaments has been displayed under
multiphoton microscope (Figure 17_C). High quality images of intraocular tissues were
produced, enabling both qualitative and quantitative analysis of the fibrous microstructures
within eye globes. As shown in Figure 17_A, the elastic fibers in scleral tissue are detected at
an excitation wavelength of 730 nm based on its blue/green autofluorescence emission,
meanwhile the collagen fibrils are displayed at an excitation wavelength of 840 nm based on
the SHG signal imaging [Wang et al. 2008a]. The elastic fibers can be clearly distinguished
from the collagenous structures because SHG emits ultraviolet/violet radiation. In a sample
sentence, the image selectivity was based that the collagen SHG can be obtained at 840 nm
excitation instead of at 730 nm excitation and the 2PF of elastic fibers can be acquired at 730
nm excitation instead of at 840 nm excitation. The rare elastic fibers were embedded in the
framework of collagen fibrils and distributed perivascularly with a density of less than 20
pieces in a 200 x 200 µm2 area (Figure 19). Based on the same settings, such fibrous
component in cornea was not observed at the excitation wavelength of 730 nm.
Investigations by multiphoton microscopy have the advantages of enabling examination
of the elastic fibers in its physiological state, avoiding the artefacts induced by histological
processing, and allowing multiple examinations with the same sample tissue. After the optical
imaging, the same scleral samples were then prepared for cryosection or paraffin-embedding
and then stained with both Weigert's resorcin–fuchsin and Hart’s elastica methods.
Unexpectedly, the elastic fibers were not found under light microscopic observations.
Presumably, the reason would involve that the in situ non-linear optical approach is more
sensitive for detection of sparse microstructures in bulk tissues than the histological staining.
Our research suggests that the multiphoton microscopy is a highly sensitive imaging
technique for studying fibrous microstructures in bulk tissues. Based on the 2PF imaging of
fibrocytes, endothelial cells of capillary, elastic fibres and the SHG imaging of collagen in
scleral tissue, the optical co-localization and microstructral topography of different
components in dense connective tissue became feasible. The 3-D images give snapshots of
cell morphology and cell-extracellular matrix organization. The novel multiphoton imaging
approach opens the general possibility of highly-resolved in situ co-localization of elastic
fibers and collagen structures in connective tissues without staining [Wang et al. 2007a].
In a conclusion, the fibrocytes and their hosting collagen can be selectively displayed
only with the assistance of two different laser wavelengths. The intratissue elastic fibers can
be in situ differentiated from the collagen filaments with the assistance of multiphoton
excitation imaging.
6. RETINAL MULTIPHOTON IMAGING: RETINA CELLULAR

INTERPRETATION WITH SUBCELLULAR SPATIAL RESOLUTION AND
HIGH SIGNAL-TO-BACKGROUND RATIO
A clear view of the retinal cellular microstructures is essential for the prophylaxis and
early diagnosis of retinal diseases. In terms of the retinal image acquisition, an optical
coherence tomography (OCT) system is being currently presented in clinic. OCT is based on
the detection of backscattering photons from the ocular structures. It provides retina imaging
in axial sectioning. The actual resolution of the commercially available OCT systems is
typically limited to 10-15 micrometers, which is insufficient to resolve the nerve fibers and
the intracellular microstructures of retinal cells [Huang et al. 1991; Li et al. 2001; Fujimoto et
al. 2003]. Advances in non-linear optical imaging are now allowing direct observations of
retinal structure including cells and nerve fibres. 2PF, selectively conducting the transversal
images of retinal cells with subcellular resolution, represents a more interesting tool for the
investigation of quantitative and spectral characteristics of retinal cells. Based on the
experience with the corneal and scleral tissues using multiphoton microscopy, further
investigations on retinal tissue with the same imaging settings were performed. The coronal
non-linear optical sectioning of retinal layers with subcellular spatial resolution was acquired.
Taking advantage of the layered structure of retina, most of the cells in retina can be
identified in different depths with the multiphoton imaging technique. A paraffin cross-
section of retina (Figure 21_E) reveals that the ganglion cells (the output neurons of the
retina) lie innermost in the retina closest to the lens, and the photoreceptors (rods and cones)
lie outermost against the RPE cells. Owing to this cellular arrangement, light must first pass
through the whole retina before striking and activating the cone photoreceptors. And
thereafter, the photons are absorbed by the photoreceptors and translated further into a
biochemical message. Subsequently, an electrical signal that can stimulate all the succeeding
neurons of the retina is transformed.
The nerve fibers, the axon of ganglion cells functioning as the third ganglion of optic
nerves, guide signals produced in the photoreceptor further to the visual processing center
located in the occipital lobe of cerebral cortex. These nerve fibers are presented (Figure
21_A). Müller cells are the radial glial cells of retina. The internal limiting membrane of the
retina is composed of the contacting Müller cells’ end feet, which are located between the
nerve fibers and the associated basement membrane. The mitochondria-enriched Müller glial
cells can be clearly seen based on its intrinsic fluorescence (Figs. 21_A and B). Despite the
high resolution and contrast, the interplexiform cells such as dopaminergic starburst amacrine
cells and horizontal cells can not be differentiated from bipolar cells in this figure. Rods and
cones are displayed in Figure 21_C. As showing in Figure 21_D, most of the RPE cells were
found to have regular hexagon-shaped cell bodies and to possess six immediate neighboring
cells. The mosaic pattern of these cells showed only slight variation in size and shape. The
distribution of cytoplasmic mitochondrial granules allows the non-fluorescent cell membranes
of RPE cells to be accurately delineated. The fluorescent cytoplasm surrounding the dark
nuclei was based on the mitochondrial NAD(P)H autofluorescence. With the assistance of
optical tomography and based on their morphological features, the thickness of the
continuous monolayer pigment cells was determined to be 6.6 ± 1.3 microns (z dimension).
The homogenous cells had an average value of 9.88 ± 1.65 microns in diameter (x-y
dimension) [Wang et al. 2008a].
Figure 21. Intrinsic 2PF tomography of retinal layers at an excitation wavelength of 760 nm. Each panel
(A, B, C, D) shows a non-linear optical transversal section of retina at different depths. Most of the
cells in different layers of retina can be identified with this two-photon imaging technique. The imaging
plane was parallel to the outer layer of tunica fibrosa bulbi. (A). Starting at a depth of 15.8 micrometers
showing the nerve fiber layer (nf), ganglion cells and Müller glial cells (M); (B). Bipolar cells and
Müller glial cells at 42.9 micrometers in depth; Müller cells are the radial glial cells of retina. The
internal limiting membrane of the retina is composed of the contacting Müller cells’ end feet, which are
located between the nerve fibers and the associated basement membrane. To the best of our knowledge,
ours is the first report on the Müller cells in different layers of intact retinal tissue based on 2PF with
intracellular resolution (Wang et al. 2008, Lasers Surg Med); (C). Photoreceptor cells at 96.3
micrometers in depth; (D). Mosaic pattern of the hexagonal RPE cells at 108.8 micrometers in depth.
This figure was acquired with the laser beams penetrating through the whole layers of retina (ganglion
cells, bipolar cells, photoreceptors) and then reaching the RPE cells using the high-numerical aperture
diffraction-limited objective 40 x 1.3 oil, which had a work distance up to 140 micrometers in tissue.
The sample examined was an intact whole-mounted eye globe wall involving retina, choroid and sclera.
A same precise filter set as used in cornea was used to acquire the intrinsic fluorescence of unstained
RPE cells. Most of the RPE cells were found to have regular hexagon-shaped cell bodies and possess
six immediate neighboring cells. These cells showed only slight variation in size and shape. In
particular, the distribution of mitochondrial granules allows the non-fluorescent cellular membranes of
RPE cells to be accurately delineated. The fluorescent cytoplasm surrounding the dark nuclei was based
on mitochondrial NAD(P)H autofluorescence. With the assistance of 3-D optical tomography, the
thickness of the continuous monolayer pigment cells was determined to be 6.6 ± 1.3 microns (z
dimension). The homogenous cells had an average value of 9.88 ± 1.65 microns in diameter (x-y
dimension). The middle figure (E) is a paraffin cross-section of a porcine retina located at periphery of
fovea, stained by conventional histological technique of Hematoxylin-Eosin (HE). Scale bars: 50
microns for the four optical images (Reprint Lasers Surg Med 2008).
The multiphoton excitation imaging of retinal lipofuscin has been recently reported [Han
et al. 2006]. Lipofuscin, a ubiquitous material presenting in the RPE cell, appears to be the
product of the peroxidation of unsaturated fatty acids, and may be symptomatic of membrane
damage, or damage to mitochondria and lysosomes [Eldred et al. 1993].
Lipofuscin accumulation is a major risk factor implicated in macular degeneration. Once
formed, the RPE cell apparently has no means either to degrade or release the lipofuscin
granular into the extracellular space. Excessive accumulation of lipofuscin occurs as a result
of lifelong phagocytosis of photoreceptor outer segments. The accumulation represents a
common pathogenetic pathway in various monogenetic and complex retinal diseases
including the age-related macular degeneration (AMD). Under one-photon CLSM, RPE
lipofuscin can be excited at 364, 488, 568, and 633 nm [Marmorstein et al. 2002].
The distribution of human RPE lipofuscin granules was also recorded under multiphoton
microscope excited by wavelength of 800 nm with emission spectrum of 500 nm to 550 nm
wavelengths. The typical diameter of lipofuscin granules was found to be below one
micrometer [Han et al. 2006, Bindewald-Wittich et al. 2006]. The two-photon excitation of
melanin was examined typically with the wavelength of 800 nm (Teuchner et al. 2000,
Hoffmann et al. 2001).
In this chapter, the RPE cells were excited and imaged with the same settings for imaging
of corneal and scleral cells (excitation wavelength of 760 nm using 435-485 nm filter cube).
The autofluorescence was mainly based on the cytoplasmatic NAD(P)H.
To the best of our knowledge, ours is the first report on the non-linear imaging of Müller
cells in different layers of intact retinal tissue based on 2PF with intracellular observation
[Wang et al. 2008a]. The 3-D co-localization images give snapshots of cell-extracellular
matrix organisation, providing highly-resolved visualization of the Müller cells in their native
environment. Multiphoton microscopy proved to be a novel tool suitable for visualizing
retinal autofluorescence with subcellular spatial resolution and would contribute to assessing
the pathophysiology of age- and disease-associated organelles alterations in retina.
7. USE OF MULTIPHOTON EXCITATION IMAGING IN ASSISTING AND

ASSESSING INTRASTROMAL SURGERY: MULTIPHOTON MICROSCOPY
FOR MONITORING INTRATISSUE FEMTOLASER SURGICAL EFFECTS
The NIR femtosecond pulsed lasers were used as a combined tool for multiphoton
excitation imaging and microsurgery [Wang et al. 2007d]. Femtolasers provide transient high
peak power that can be used in conjunction with focusing optics possessing high numerical
aperture to produce light intensities at high terawatt values per square centimetre enough for
intratissue photoablation. In the low-repetition-rate regime, material processing is induced by
a single pulse based on the formation of high-density plasma. By contrast, the focused laser
pulses in a high-repetition-rate femtosecond laser system result in an increase in temperature
through heat accumulation, which induces the formation of low-density plasma. As the
number of pulses increases, the threshold energy for photodisruption decreases due to a
cumulative effect [Stern et al. 1989, Stuart et al. 1996, Juhasz et al. 1996, Vogel et al. 1999,
König et al. 2002].
The use of a femtosecond oscillator with high repetition-rate enables laser surgery at low
pulse energy on an order of sub-nanojoule and nanojoule [König et al. 2005b]. A
femtosecond laser pulse can produce localized energy absorption because the pulse width is
shorter than the time required for heat to diffuse out of the focal volume. Thus, the
photodisruption occurs only in the focal volume and no photon damage takes place outside
the focal femtoliter volume. Another reason for the no-photon-damage outside focus is that
biological tissues virtually possess no efficient cellular absorbers in the 700–840 nm spectral
region. The non-amplified femtolasers show its advantages in nanoprocessing and
multiphoton absorption without compromising viability in the surrounding region. NIR
femtolasers have recently attracted a large amount of attention in cellular nanosurgery and
nanoprocessing in biological materials. Several researchers have adopted this non-amplified
NIR femtosecond laser pulses for precise nanosurgery of cells and sub-cellular organelles
such as knock-out single organelles in living cells [König et al. 1999, Watanabe et al. 2004],
nanodissection of human chromosomes [König et al. 2001], membrane optoporation [Kohli et
al. 2005]. Photoablation in human chromosomes with a full width at half maximum (FWHM)
cut size of less than 70 nm have been demonstrated with femtolaser pulses of less than 3 nJ
pulse energy [König et al. 2001]. NJ femtolasers possess the capability to perform precise
intratissue and intracellular photodissection on a submicrometer precision.
Intrastromal surgery was performed with the same multiphoton imaging system but with
a higher laser power in comparison with the laser power for imaging. In general, the laser
power used for multiphoton imaging was set between 20 mW (mean power, corresponding to
0.25 nJ pulse energy and 1.60 kW peak power) and 35 mW (mean power, corresponding to
0.44 nJ pulse energy and 2.80 kW peak power) with respect to different depths in the cornea
with a frame scan (2PF and SHG). For corneal microsurgery, laser power increased to
between 130 mW (mean power, corresponding to 1.64 nJ pulse energy and 10.40 kW peak
power) and 165 mW (mean power, corresponding to 2.08 nJ pulse energy and 13.20 kW peak
power) with a line scan at a wavelength of 800 nm.
Intrastromal microsurgery at a depth of 100 micrometers in cornea was visualized using
the multiphoton excitation imaging (Figure 22). The in situ differentiation of corneal layers
makes it possible to determine the target region before the intrastromal microsurgery. Non-
linear optical imaging helped making it possible to determine the targeted region for
intrastromal surgery. Guided by the optical overview of cornea (Figures 6, 7, 9, 10, 15), the
laser was directed on a region of interest for further surgical procedures. After intrastromal
treatment was performed with a laser power of 130mW at a wavelength of 800 nm by a line
scan (zoom=1.3, t=1s, at a depth of 100 microns), the optical images (SHG at an excitation
wavelength of 830 nm in Figure 22_A and autofluorescence images at 760 nm in Figure
22_B) were immediately taken (within 10s). As seen in the optical sections, four laser cuts
were acquired with submicron size scale of 0.4-0.6 microns. Some transient subtle bubbles
were detected (mean diameters of 2-3 microns, lifetimes less than 2 seconds) along the cuts.
With the multiphoton excitation imaging, the surgical performance can be clearly visualized
and precisely evaluated within seconds. Multiphoton microscopy/tomography consequently
proved capable of optically evaluating the intrastromal laser surgical effects [König et al.
2004, Wang et al. 2007a].
Figure 22. Postoperative multiphoton optical evaluating of intrastromal femtolasers surgery performed
at a depth of 100 micrometers. A. SHG image at an excitation wavelength of 830 nm in a backward
direction; B. Autofluorescence image at an excitation wavelength of 760 nm. The non-linear optical
imaging/tomography of cornea was used as a precise means for definition of interest of region
preoperation as well as for inspection and confirmation of the surgical results in real time immediately
after the intrastromal nanosurgery. Scale bar: 45 microns. After visualizing of cornea with assistance of
the 3-D non-linear optical tomography, the targeted region for intrastromal surgery was determined. On
accomplishment of the intrastromal nanosurgery performed with a laser power of 130 mW, at a
wavelength of 800 nm, with line scan (zoom=1.3, t =1 s, at a depth of 100 µm), the optical images were
immediately taken (within 10 s). As seen in the optical sections, four laser cuts were acquired at sub-
micron size of 0.4–0.6 microns. The overlying and underlying layers remain intact. During the
treatment, some transient subtle bubbles were detected (mean diameters of 2-3 microns, lifetimes less
than 2 seconds) along the incision. The surrounding keratocytes and collagen lamellas remain intact and
are even comparable to those microstructures in physiological stats (Fig.10). The surgical performance
can be simultaneously visualized and evaluated in real time, which makes the multiphoton excitation
imaging an indispensable assessing component of the laser intratissue surgery (Reprint Lasers Surg
Med 2007).
The data obtained in this chapter suggest distinct advantages of nJ femtosecond lasers
compared with microjoule (µJ) femtosecond lasers. First, the absence of large intratissue
bubbles by nJ femtolasers makes it feasible to ablate the tissue more precisely. The sharp
boundary between the treated and untreated areas indicates the high precision of the surgical
treatment (Figure 22 and Figure 23_G).
Figure 23. In vivo multiphoton excitation imaging tomography of myofibroblasts based on their
intrinsic 2PF at an excitation wavelength of 760 nm with a vertical z-interval of 2 micrometers twenty-
four hours after the laser operation (A-F). (G). In vivo reflection imaging of the treated stromal region
twenty-four hours after the intrastromal surgery. The activated cells are marked with the symbol c in
Fig.G. Scale bars: 20 microns for A-F and 35 microns for G. Optical imaging (reflexion) of the treated
cornea 24 h postoperation showing the ablated area and the myofibroblasts at a distance of 50
micrometers from the intrastromal laser lesions. Surgery was performed within the cornea at depths of
between 90 µm and 100 µm with vertical z-interval of 2 µm using high-powered laser beams between
150 mW (at the depth of 90 µm) and 165 mW (at the depth of 100 µm) with line scanning. Observed 24
h after laser treatment, the luminescence along the cut edges became noticeably weaker compared to
that immediately after the laser operation. 3-D 2PF tomography of the activated cells was performed to
study the cellular response to the lesions in the laser-treated corneas during the postoperative wound
healing process. Remarkable, cell aggregates were detected in the surrounding area of treated region.
Due to the deficiency of collagen SHG signals at the excitation wavelength of 760 nm, the
morphological details of the activated cells can be selectively studied with three dimensions and
intracellular observations. Based on their altered morphology observed -- (i) hypertrophic cell bodies
with diameters of 25-30 microns (the normal cells have a size of less than 20 microns, Fig.10); (ii) the
appearance of augmented cell processes; and (iii) nuclei round or oval in shape -- these cells were
determined to be the activated intrastromal keratocytes, called myofibroblasts. Control corneas
exhibited no such cell aggregates. The rapid transformation of keratocytes seems to be the first cellular
reaction after intrastromal laser surgery (Reprint Lasers Surg Med 2007).
Collateral effects, such as tissue streaks induced by the µJ femtolasers in intracorneal

surgery, were reported [Lubatschowski et al. 2000]. Streaks, created inside the corneal tissue
irradiated with 160 fs laser pulses at a pulse energy of 2 μJ, were visualized on histological
sections. In TEM check, the streaks can be noticed as a dark staining, crossing the picture in
vertical direction. The diameter of the streaks was in the range of 200–500 nm. The
intrastromal streaks, still seen for some days postoperatively, would have an effect on the
wound healing process [Arnold et al. 2005]. Such collateral phenomenon was not observed
with nJ femtolasers based on histology, optical imaging or EM examinations.
The nJ ultrashort lasers were employed not only as probing beams but as also a surgical
tool. The outcomes of intratissue microsurgery in animal studies have been on-line displayed
with the most convenient imaging combination being available. Conventional histological
methods have difficulty in showing the intratissue minute cuts because the two cut edges
merge immediately after the absorption of the tiny bubbles. Although electron microscopy
can show the microsurgical results, the procedure is complicated and time-consuming. The
non-linear optical tomography of the cornea is perfectly suited for determining the targeted
region preoperatively. The corneal layers and depth confirmed by this optical tomography
ensure the precision of femtosecond laser surgery. This imaging technique is indispensable
for immediately evaluating intratissue microsurgery. Therefore, integrating the two
techniques in one system, as was done in this chapter, is currently the most efficient and
convenient way to carry out intratissue surgery with non-amplified femtolasers.
This chapter reports on the in vivo application of optical corneal imaging based on
multiphoton microscopy/tomography in the intrastromal microsurgery performed with NIR nJ
femtolasers. The surgical performance can be simultaneously visualized and evaluated in real
time, which makes the multiphoton excitation imaging an indispensable component for laser
intratissue surgery. The use of multiphoton imaging was found also in the generation of
intratissue ablation [Wang et al. 2007b] and intracorneal lenticules [Wang et al. 2008b].
With the assistance of the non-linear optical imaging, multiphoton-mediated intratissue
ablation was generated while excluding severe damage and mechanical effects in the
surrounding tissue. In conclusion, multiphoton microscopy is an outstanding tool in assisting

in performing intratissue laser surgery and ensures predictable and reproductive surgical
outcomes. This non-linear optical imaging technique continues to find an increasing number
of applications in biological and medical laser surgery. A most interesting potential
application of the multiphoton excitation imaging technique would be in assisting the
intracellular organelle nanosurgery.
8. MULTIPHOTON MICROSCOPY OF ACTIVATED CELLS IN STROMA

AFTER SURGICAL TREATMENT: APPLICATION OF MULTIPHOTON
EXCITATION IMAGING TO DETECTING THE STROMAL CELL
ACTIVATION DURING WOUND REPAIR AFTER CORNEAL LASER
SURGERY
Keratocyte is the predominant cell responsible for maintaining the integrity and
regeneration of corneal components. The cells are connected to each other by gap junctions
through their own cell processes. This cell in physiological state is inactive in phagocytosis.
Under stress such as laser surgery, the cells can transform to myofibroblasts in the area
around the corneal wound, and secret new ECM that ultimately reconstitutes the components
of cornea [Jester et al. 1999, Wilson 2002, Mohan et al. 2003]. The factors, inducing cascades
of wound repair, are secreted firstly by the activated cell [Funderburgh et al. 2001, Berryhill
et al. 2002]. Multiphoton excitation imaging was used to detect the stromal cell activation.
Surgery was performed within the cornea at depths of between 90 µm and 100 µm with
vertical z-step of 2 micrometers using high-powered laser beams between 150 mW (at a depth
of 90 µm) and 165 mW (at a depth of 100 µm) assisted by the line scanning (Figure 23_G).
Observed 24 hours after the laser treatment, part of the cut spaces were filled with
migrating cells. Remarkable, cell aggregates were detected in the surrounding region within
50 μm of the intrastromal lesion sites. Based on their special morphology observed -- (i)
hypertrophic cell bodies with diameters of 25-30 microns (the normal cells have a size of less
than 20 microns, Figure 10); (ii) the appearance of augmented cell processes; and (iii) the
round or oval nuclei were surrounded by the intensified cytoplasmic autofluorescence
granules -- these cells most likely seemed to be activated intrastromal keratocytes
(myofibroblasts) (Figure 23). Control corneas exhibited no such cell aggregates [Wang et al.
2007a].
Interestingly, we have been able to in vivo detect stromal myofibroblasts after intratissue
ultrashort laser surgery. The myofibroblasts are defined in this research by their special
morphology and intensifier intracellular endogenous fluorescence. There are a large number
of studies related to the transformation of fibroblasts. Masur et al. [1996] have studied the
transformation of keratocytes with a smooth muscle α-actin stain. Whereas transforming
growth factor-β1 stimulates myofibroblasts differentiation, fibroblast growth factor-2 and
platelet-derived growth factor inhibit such transformation [Maltseva et al. 2001, Baldwin et
al. 2002]. Cells transformed from keratocytes resemble satellite cells, which are highly
reflective and characterized by intense cytoplasmic fluorescence. The rapid transformation of
keratocytes during the course of laser-induced wound repair seems to be the first cellular
reaction after the intrastromal laser surgery. Thereafter, myofibroblasts synthesize
chemokines and chemotaxis, leading to characteristic inflammatory cascade reactions during

wound repair. The contributions of myofibroblasts involving wound contraction and
phagocytosis in the healing process have been discussed elsewhere [Funderburgh et al. 2001,
Chakravarti et al. 2004].
2PF signal imaging was performed to study the cellular response to the lesions in the
laser-treated corneas during the postoperative wound healing. In vivo imaging allowed the
intracorneal myofibroblasts to be detected based on its intrinsic NAD(P)H autofluorescence at
the wavelength of 760 nm.
CONCLUSION
Our aim was to use spectrally sensitive multiphoton excitation microscopy to elucidate
microstructures of the whole-mount ocular tissues. Here, data show that 2PF can be used to
identify cellular components, whereas SHG is valuable in displaying of collagenous
microstructure in bulk tissues. Thick, dense, and non-fixed bulk tissues can be easily probed
based on their endogenous fluorophore and collagen. Contrast specifically for both intratissue
cells and their hosting ECM was sufficiently generated only based on the in-tandem detection
with 2PF and SHG signal imaging. The 3-D co-localization images give snapshots of cell
morphology and cell-extracellular matrix organization. The quantitative analysis of tissue
microstructures including cellular and fibrous compositions was obtained. Multiphoton
microscopy allows for the in vivo non-invasive study of biological specimens in three
dimensions with submicron resolution at depths of up to some hundreds of micrometers.
Because the examinations in this chapter do not involve any assistance of exogenous dyes or
the need for tissue preparation or fixation, it avoids artefacts induced by additional sample
staining and slicing procedures. This proof-of-principle examination has already started to
reveal the intricate features in bulk tissues that are otherwise lost in conventional histology
studies, suggesting the potential of using non-linear optical imaging as a routine tool for more
advanced optical histology.
Several other advantages of multiphoton microscopy can be immediately confirmed and
inferred from our experimental data: (i) Improved depth penetration in thick specimens due to
greatly reduced scattering of longer wavelengths. In the NIR illumination wavelengths of
700–840 nm, there are virtually no other efficient tissue absorbers except for water. With a
suitable objective 20x 0.9 N.A. used in the chapter, the deepest layer of corneal tissue can be
even optically visualized in situ [Wang et al. 2007a]); (ii) A promising imaging approach for
in vivo examinations with biomedical samples since no cellular and tissue staining or slicing
is required; (iii) Tremendous improvements in optical localization and z-resolution because
the majority of all multiphoton excitations was confined to a minute volume of 0.1 fl focal
point; the smallest z-step reaches to the high precision of 25 nm; (iv) Dramatically promoted
signal-to-noise ratio since the large separation between the incident NIR excitation and the
subsequent visible emission wavelengths, which makes it straightforward to achieve a high
signal to background ratio.
More recently, a novel organ image acquisition technique called two-photon tissue
cytometry has been posed. The criteria for the development of 3-D tissue cytometry include
spatial resolution, depth penetration and molecular specificity. The imaging features of
multiphoton microscopic technique are perfectly in coincidence with these requests. Two-
photon tissue cytometry, imaging the microstructural components in organs and tissues with
sufficient details to reveal 3-D sample morphology, offers all the benefits of fluorescence-
based approaches including high specificity and sensitivity as well as appropriateness for
molecular imaging. Coupled with automated histological sectioning, an entire heart has been
imaged with subcellular resolution based on two-photon microscopy [Ragan et al. 2007]. This
new sensitive imaging technique was demonstrated to be capable of quantifying 3-D
morphology of cardiac myocytes and of evaluating microvasculature remodelling. In an
analogous manner, the development of imaging entire eye globes based on multiphoton
excitation microscopy would also be encouraging. The data achieved in this chapter provide
the primary anatomical knowledge of eye globes for the further development of two-photon
ocular tissue cytometry.
Multiphoton excitation images obtained from various tissues show providing excellent
structural and biomechanical characterization. Intrinsic 2PF has been used to distinguish
dysplastic and neoplastic tissues from normal tissues, to identify tumorous tissue [Guo et al.
1999, Brown et al. 2003] and the decrepit brain tissues associated with Alzheimer’s disease
[Hensley et al. 1998, Christie et al. 2001]. The nonlinear optical tomography also provides a
new approach to study of dysplasia in squamous epithelium and abnormal collagen in
keratoconus [Lyubovitsky et al. 2006, Morishige et al. 2007]. Owing to emerging

technologies such as fluorescence resonance energy transfer (FRET) and fluorescence
correlation spectroscopy, multiphoton technique is now being further widely applied to probe
single protein molecules and to investigate the interactions of their functions [Xu et al.1996,
Combs et al. 2001, Biskup et al. 2007]. Imaging of SHG signals was employed as a sensitive
tool for detection of corneal fibrosis and then further to be used to assess the effects of
antifibrotic therapy on corneal wound healing [Farid et al. 2008].
Data show that the combination of 2PF and SHG allows for the selective visualization of
various microstructural components in the whole-mount scleral, corneal and retinal tissues. In
situ quantitative and qualitative assessment of microstructures in different layers of the intact
bulk tissues was obtained. It enables high-resolution deep tissue imaging of endogenous
fluorophores and is well suitable for molecular imaging in living biological specimens.
Intrinsic emission multiphoton microscopy/tomography is consequently confirmed to be an
efficient and sensitive non-invasive imaging approach. The non-linear optical imaging can be
used as a versatile biomedical tool for optical diagnostic imaging, yielding vivid insights into
biological specimens, which will ultimately find its clinical application in optical pathological
diagnostics. We also believe that this promising technique will find more applications in the
biological and medical basic research in the near future. Potentially, the autofluorescence of
NAD(P)H provides a valid basis and would have a use in studying the dynamics of the
mitochondrial bioenergetics and further in conducting nanomedical investigations in this
organelle.
ACKNOWLEDGMENTS
The cooperation with Drs. Karsten König and Iris Riemann from the Fraunhofer Institute
of Biomedical Technology, Harald Schubert from the Institute of Laboratory Animals
Science, Shuping Song from the Institute of Molecular Cell Biology, Dietrich Schweitzer and
Martin Hammer from the Eye Clinic, Annett Eitner from the Institute of Anatomy II, as well
as with the companies of Carl Zeiss Jena and JenLab in Germany is gratefully acknowledged.
This research was supported in part by DFG.
REFERENCES
Arnold, CL; Heisterkamp, A; Ertmer, W; Lubatschowski, H. Streak formation as side effect
of optical breakdown during processing the bulk of transparent Kerr media with ultra-
short laser pulses. Appl. Phys. B, 2005; 80(2):247-253.
Baldwin, HC; Marshall, J. Growth factors in corneal wound healing following refractive
surgery: a review. Acta . Scand, 2002; 80 (3):238-247.
Barzda, V; Greenhalgh, C; Aus der Au, J; Elmore, S; Van Beek, J; Squier, J. Visualization of
mitochondria in cardiomyocytes by simultaneous harmonic generation and fluorescence
microscopy. Opt. Express, 2005; 13:8263-8276.
Bentley, E; Abrams, GA; Covitz, D; Cook, CS; Fischer, CA; Hacker, D; Stuhr, CM; Reid,
TW; Murphy, CJ. Morphology and immunohistochemistry of spontaneous chronic
corneal epithelial defects in dogs. Invest Ophthal Vis. Sci, 2001; 42(10):2262-2269.
Berryhill, BL; Kader, R; Kane, B; Birk, DE; Feng, J; Hassell, JR. Partial restoration of the
keratocyte phenotype to bovine keratocytes made fibroblastic by serum. Invest
Ophthalmol. Vis. Sci, 2002; 43 (11):3416-3421.
Bindewald-Wittich, A; Han, M; Schmitz-Valckenberg, S; Snyder, SR; Giese, G; Bille, JF;
Holz, FG. Two-photon-excited fluorescence imaging of human RPE cells with a
femtosecond Ti:Sapphire laser. Invest Ophthalmol Vis Sci, 2006; 47:4553-4557.
Biskup, C; Zimmer, T; Kelbauskas, L; Hoffmann, B; Klöcker, N; Becker, W; Bergmann, A;
Benndorf, K. Multi-dimensional fluorescence lifetime and FRET measurements. Microsc
Res. Tech., 2007; 70(5):442-451.
Boyde, A. (1974). Scanning Electron Microscopy: Three-dimensional aspects of SEM images
(Well, O.C.) (ed.). New York, McGraw-Hill.
Brakenhoff, GJ; Squier, J; Norris, T; Bliton, AC; Wade, WH; Athey, B. Real-time two-
photon confocal microscopy using a femtosecond, amplified Ti:sapphire system. J.
Microsc, 1996; 181:253-259.
Brecht, M; Fee, MS; Garaschuk, O; Helmchen, F; Margrie, TW; Svoboda, K; Osten, P. Novel
approaches to monitor and manipulate single neurons in vivo. J. Neurosci, 2004;
24(42):9223-9227.
Brown-Augsburger, P; Broekelmann, T; Mecham, L; Mercer, R; Gibson, MA; Cleary, EG;
Abrams, WR; Rosenbloom, J; Mecham, RP. Microfibril-associated glycoprotein binds to
the carboxyl-terminal domain of tropoelastin and is a substrate for transglutaminase. J.
Biol Chem, 1994; 269(45):28443-28449.
Brown, E; McKee, T; Ditomaso, E; Pluen, A; Seed, B; Boucher, Y; Jain, RK. Dynamic
imaging of collagen and its modulation in tumors in vivo using second-harmonic
generation. Nat Med, 2003; 9:796-800.
Camber, O; Rehbinder, C; Nikkila, T; Edman, P. Morphology of the pig cornea in normal
conditions and after incubation in a perfusion apparatus. Acta Vet Scand, 1987;
28(2):127-134.
Campagnola, PJ; Millard, AC; Terasaki, M; Hoppe, PE; Malone, CJ; Mohler, WA. Three-
dimensional high-resolution second harmonic generation imaging of endogenous
structural proteins in biological tissues. Biophys J., 2002; 82(1):493–508.
Campagnola, PJ; Loew, LM. Second-harmonic imaging microscopy for visualizing
biomolecular arrays in cells, tissues and organisms. Nat. Biotechnol., 2003; 21(11):1356-
1360.
Chakravarti, S; Wu, F; Vij, N; Roberts, L; Joyce, S. Microarray studies reveal macrophage-
like function of stromal keratocytes in the cornea. Invest. Ophthalmol. Vis. Sci, 2004; 45
(10):3475-3484.
Christie, RH; Bacskai, BJ; Zipfel, WR; Williams, RM; Kajdasz, ST; Webb, WW; Hyman,
BT. Growth arrest of individual senile plaques in a model of Alzheimer’s disease
observed by in vivo multiphoton microscopy. J. Neurosci, 2001; 21(3):858-864.
Combs, CA; Balaban, RS. Direct Imaging of Dehydrogenase Activity within Living Cells
Using Enzyme-Dependent Fluorescence Recovery after Photobleaching (ED-FRAP).
Biophys. J., 2001; 80(4):2018-2028.
Dedov, VN; Cox, GC; Roufogalis, BD. Visualisation of mitochondria in living neurons with
single- and two-photon fluorescence laser microscopy. Micron, 2001; 32(7):653-660.
Denk, W; Strickler, JH; Webb, WW. Two-photon laser scanning fluorescence microscopy.
Science, 1990; 248 (4951):73-76.
Denk, W. Two-photon scanning photochemical microscopy: mapping ligand-gated ion
channel distributions. PNAS, 1994; 91(14):6629-6633.
Diaspro, A; Robello, M. Two-photon excitation of fluorescence for three-dimensional optical
imaging of biological structures. J. Photochem Photobiol. B, 2000; 55(1):1-8.
Dombeck, DA; Kasischke, KA; Vishwasrao, HD; Ingelsson, M; Hyman, BT; Webb, WW.
Uniform polarity microtubule assemblies imaged in native brain tissue by second-
harmonic generation microscopy. PNAS, 2003; 100(12):7081-7086.
Elbadawi, A. Hexachrome modification of Movat's stain. Biotechnic and Histochemistry,
1976; 51(5):249-253.
Eldred, GE; Lasky, MR. Retinal age pigments generated by self-assembling lysosomotropic
detergents. Nature, 1993; 361(6414):724-726.
Farguharson, C; Robins, SP. Immunolocalization of collagen type I, III, and IV, elastin and
fibronectin within the heart of normal and copper deficient rats. J. Comp Pathol., 1991;
104(3):245-255.
Farid, M; Morishige, N; Lam, L; Wahlert, A; Steinert, RF; Jester, JV. Detection of Corneal
Fibrosis by Imaging Second Harmonic Generated Signals (SHG) in Rabbit Corneas
Treated with Mitomycin C Following Excimer Laser Surface Ablation. Invest
Ophthalmol Vis. Sci, 2008; DOI: 10.1167/iovs.08-1983.
Fine, S; Hansen, WP. Optical second harmonic generation in biological systems. Appl. Opt,
1971; 10(10):2350–2353.
Freeman, IL. Collagen polymorphism in mature rabbit cornea. Invest Ophthalmol Vis. Sci,
1978; 17(2):171-177.
Freund, I; Deutsch, M; Sprecher, A. Connective tissue polarity. Optical second-harmonic
microscopy, crossed-beam summation, and small-angle scattering in Rat-tail tendon.
Biophys J, 1986; 50:693-712.

Funderburgh, JL; Funderburgh, ML; Mann, MM; Corpuz, L; Roth, MR. Proteoglycan
expression during transforming growth factor beta-induced keratocyte-myofibroblast
transdifferentiation. J. Biol. Chem., 2001; 276(47):44173–44178.
Fujimoto, JG. Optical coherence tomography for ultrahigh resolution in vivo imaging. Nat
Biotechnol., 2003; 21(11):1361-1367.
Göppert-Mayer, M. Über Elementarakte mit zwei Quantensprüngen. Ann Phys, 1931; 9:273-
294.
Guo, Y; Savage, HE; Liu, F; Schantz, SP; Ho, PP; Alfano, RR. Subsurface tumor progression
investigated by non-invasive optical second harmonic tomography. PNAS, 1999;
96(19):10854-56.
Halbhuber, KJ; Krieg, W; König, K. Laser scanning microscopy in enzyme histochemistry.
Visualization of cerium-based and DAB-based primary reaction products of
phosphatases, oxidases and peroxidases by reflectance and transmission laser scanning
microscopy. Cell Mol. Bio (Noisy-le-grand), 1998; 44: 807-826.
Halbhuber, KJ; König, K. Modern laser scanning microscopy in biology, biotechnology and
medicine. Ann. Anat, 2003; 185(1):1-20.
Han, M; Giese, G; Bille, JF. Second harmonic generation imaging of collagen fibrils in
cornea and sclera. Opt. Express, 2005; 13:5791-5797.
Han, M; Bindewald-Wittich, A; Holz, FG; Giese, G; Niemz, MH; Snyder, S; Sun, H; Yu, J;
Agopov, M; Schiazza, OL; Bille, JF. Two-photon excited autofluorescence imaging of
human retinal pigment epithelial cells. J. Biomed Opt, 2006; 11(1):010501.
Hayashi, S; Osawa, T; Tohyama, K. Comparative observations on corneas, with special
reference to Bowman's layer and Descemet's membrane in mammals and amphibians. J
Morphol, 2002; 254(3):247-58.
Helmchen, F; Denk, W. Deep tissue two-photon microscopy. Nat. Methods, 2005;
2(12):932-940.
Hensley, K; Maidt, ML; Yu, Z; Sang, H; Markesbery, WR; Floyd, RA. Electrochemical
analysis of protein nitrotyrosine and dityrosine in the Alzheimer brain indicates region-
specific accumulation. J. Neurosci., 1998; 18(20):8126-8132.
Hoffmann, K; Stücker, M; Altmeyer, P; Teuchner, K; Leupold, D. Selective Femtosecond
Pulse-Excitation of Melanin Fluorescence in Tissue. J. Invest Dermatol., 2001; 116:629-
630.
Huang, D; Swanson, EA; Lin, CP; Schuman, JS; Stinson, WG; Chang, W; Hee, MR; Flotte,
T; Gregory, K; Puliafito, CA; Fujimoto, JG. Optical coherence tomography. Science,
1991; 254(5035):1178-1181.
Huang, S; Heikal, AA; Webb, WW. Two-photon fluorescence spectroscopy and microscopy
of NAD(P)H and flavoprotein. Biophys J., 2002; 82(5):2811-2825.
Jester, JV; Petroll, WM; Cavanagh, HD. Corneal stromal wound healing in refractive surgery:
the role of myofibroblasts. Prog. Retin Eye Res, 1999; 18(3):311-356.
Juhasz, T; Kastis, GA; Suárez, C; Bor, Z; Bron, WE. Time-resolved observations of shock
waves and cavitation bubbles generated by femtosecond laser pulses in corneal tissue and
water. Lasers Surg Med, 1996; 19:23-31.
Kaiser, W; Garrett, CGB. Two-Photon Excitation in CaF2: Eu2+. Phys. Rev Lett, 1961;
7:229-231.
Kato, T; Chang, JH; Azar, DT. Expression of type XVIII collagen during healing of corneal
incisions and keratectomy wounds. Invest Ophthalmol Vis. Sci, 2003; 44:78-85.
Kohli, V; Elezzabi, AY; Acker, JP. Cell nanosurgery using ultrashort laser pulses:
Applications to membrane surgery and cell isolation. Lasers Surg. Med, 2005; 37:227-
230.
König, K; Simon, U; Halbhuber, KJ. 3D-resolved two-photon fluorescence microscopy of
living cells using a modified confocal laser scanning microscope. Cell Mol Biol (Noisy-
le-grand), 1996; 42(8):1181-1194.
König, K; Riemann, I; Fischer, P; Halbhuber, KJ. Intracellular nanosurgery with near infrared
femtosecond laser pulses. Cell Mol Biol (Noisy-le-grand), 1999; 45(2):195-201.
König, K. Multiphoton Microscopy in Life Sciences. J. Microscopy, 2000(a); 200(Pt 2):83-
104.
König, K. Laser tweezers and multiphoton microscopes in life sciences. Histochem Cell Biol,
2000(b); 114(2):79-92.
König, K; Riemann, I; Fischer, P; Halbhuber, KJ. Multiplex FISH and three dimensional
DNA imaging with near infrared femtosecond laser pulses. Histochem Cell Biol, 2000(c);
114(4): 337-345.
König, K; Riemann, I; Fritzsche, W. Nanodissection of human chromosomes with near-

infrared femtosecond laser pulses. Opt. Lett, 2001; 26(11):819-821.
König, K; Krauss, O; Riemann, I. Intratissue surgery with 80 MHz nanojoule femtosecond
laser pulses in the near infrared. Opt Express, 2002; 10:171-176.
König, K; Wang, BG; Krauss, O; Riemann, I; Schubert, H; Kirste, S; Fischer, P. First in vivo
animal studies on intraocular nanosurgery and multiphoton tomography with low-energy
80 MHz near infrared femtosecond laser pulses. In: Manns F, Soderberg PG, Ho A
(editors). Title: Ophthalmic Technologies XIV. San Jose: SPIE; 2004; 5314:262-269.
König, K; Schenke-Layland, K; Riemann, I; Stock, UA. Multiphoton autofluorescence
imaging of intratissue elastic fibers. Biomaterials, 2005(a); 26(5):495–500.
König, K; Wang, BG; Riemann, I; Kobow, J. Cornea surgery with nanojoule femtosecond
laser pulses. In: Manns F, Soederberg PG, Ho A, Stuck BE, Belkin M (editors). Title:
Ophthalmic Technologies XV. San Jose: SPIE; 2005(b); 5688:288-293.
König, K; Riemann, I. High-resolution multiphoton tomography of human skin with
subcellular spatial resolution and picosecond time resolution. J. Biomed Opt, 2003;
8(3):432-439.
Levene, MJ; Dombeck, DA; Kasischke, KA; Molloy, RP; Webb, WW. In vivo multiphoton
microscopy of deep brain tissue. J. Neurophysiol, 2004; 91(4):1908-1912.
Li, Q; Timmers, AM; Hunter, K; Gonzalez-Pola, C; Lewin, AS; Reitze, DH; Hauswirth, WW.
Noninvasive imaging by optical coherence tomography to monitor retinal degeneration in
the mouse. Invest Ophthalmol Vis. Sci, 2001; 42:2981-2989.
Light, ND. Bovine type I collagen: A study of cross-linking in various mature tissues.
Biochim Biophys Acta, 1979; 581(1):96-105.
Lubatschowski, H; Maatz, G; Heisterkamp, A; Hetzel, U; Drommer, W; Welling, H; Ertmer,
W. Application of ultrashort laser pulses for intrastromal refractive surgery. Graefe’s
Arch. Clin. Exp. Ophthalmol., 2000; 238(1):33-39.
Lyubovitsky, JG; Spencer, JA; Krasieva, TB; Andersen, B; Tromberg, BJ. Imaging corneal
pathology in a transgenic mouse model using nonlinear microscopy. J. Biomed Opt,
2006; 11(1):014013.
Maltseva, O; Folger, P; Zekaria, D; Petridou, S; Masur, SK. Fibroblast growth factor reversal
of the corneal myofibroblast phenotype. Invest Ophthalmol Vis. Sci, 2001; 42(11):2490-
2495.
Malak, H; Castellano, FN; Gryczynski, I; Lakowicz, JR. Two-photon excitation of ethidium
bromide labeled DNA. Biophys Chem., 1997; 67:35-41.
Marmorstein, AD; Marmorstein, LY; Sakaguchi, H; Hollyfield, JG. Spectral profiling of
autofluorescence associated with lipofuscin, Bruch's Membrane, and sub-RPE deposits in
normal and AMD eyes. Invest Ophthalmol Vis. Sci, 2002; 43:2435–2441.
Masur, SK; Dewal, HS; Dinh, TT; Erenburg, I; Petridou, S. Myofibroblasts differentiate from
fibroblasts when plated at low density. PNAS, 1996; 93:4219-4223.
Maiti, S; Shear, JB; Williams, RM; Zipfel, WR; Webb, WW. Measuring serotonin
distribution in live cells with three-photon excitation. Science, 1997; 275 (5299):530-532.
Maurice, D.M. (1984). The eye: The cornea and sclera (Davson, H.) (ed.). New York,
London: Academic Press.
Maser, MD; Trimble, JJ. Rapid chemical dehydration of biological samples for scanning
electron microscopy using 2,2-dimethoxypropane. J. Histochem Cytochem, 1977;
25(4):247-251.
Masters, BR; So, PT; Gratton, E. Multiphoton excitation fluorescence microscopy and
spectroscopy of in vivo human skin. Biophys J., 1997; 72(6):2405-2412.
Merindano, MD; Costa, J; Canals, M; Potau, JM, Ruano, D. A comparative study of
Bowman's layer in some mammals: Relationships with other constituent corneal
structures. European Journal of Anatomy, 2002; 6(3):133-139.
Mohan, RR; Hutcheon, AE; Choi, R; Hong, J; Lee, J; Mohan, RR; Ambrósio, R Jr; Zieske,
JD; Wilson, SE. Apoptosis, necrosis, proliferation, and myofibroblast generation in the
stroma following LASIK and PRK. Exp Eye Res, 2003; 76(1):71-87.
Morishige, N; Wahlert, AJ; Kenney, MC; Brown, DJ; Kawamoto, K; Chikama, T; Nishida, T;
Jester, JV. Second-harmonic imaging microscopy of normal human and keratoconus
cornea. Invest Ophthalmol Vis Sci, 2007; 48(3):1087-1094.
Moreaux, L; Sandre, O; Mertz, J. Membrane imaging by second-harmonic generation
microscopy. J. Opt Soc Am. B, 2000; 17:1685-1694.
Nakayasu, K; Tanaka, M; Konomi, H; Hayashi, T. Distribution of types I, II, III, IV and V
collagen in normal and keratoconus corneas. Ophthalmic Res, 1986; 18(1):1-10.
Nishida, T; Yasumoto, K; Otori, T; Desaki, J. The network structure of corneal fibroblasts in
the rat as revealed by scanning electron microscopy. Invest Ophthalmol Vis. Sci, 1988;
29(12):1887-90.
Nitsch, R; Phol, EE; Smorodchenko, A; Infante-Durarte, C; Aktas, O; Zipp, F. Direct impact
of T cells on neurons revealed by two-photon microscopy in living brain tissue. J.
Neurosci, 2004; 24(10):2458-2464.
Patterson, GH; Knobel, SM; Arkhammar, P; Thastrup, O; Piston, DW. Separation of the
glucose stimulated cytoplasmic and mitochondrial NAD(P)H responses in pancreatic islet
cells. PNAS, 2000; 97(10):5203-5207.
Piston, DW. Imaging living cells and tissues by two-photon excitation microscopy. Trends
Cell Biol, 1999; 9(2):66-69.
Poole, CA; Brookes, NH; Clover, GM. Keratocyte networks visualised in the living cornea
using vital dyes. J Cell Sci, 1993; 106(Pt 2):685-691.
Ragan, T; Sylvan, JD; Kim, KH; Huang, H; Bahlmann, K; Lee, RT; So, PT. High-resolution
whole organ imaging using two-photon tissue cytometry. J. Biomed. Opt, 2007;
12(1):014015.
Roth, S; Freund, I. Second harmonic generation in collagen. J. Chem. Phys, 1979; 70:1637-
1643.
Rubart, M. Two-photon microscopy of cells and tissue. Circ. Res., 2004; 95(12):1154-1166.
So, PTC; Kim, H; Kochevar, I. Two-photon deep tissue ex vivo imaging of mouse dermal and
subcutaneous structures. Opt. Express, 1998; 3(9):339-350.
Scharenberg, K. The cells and nerves of the human cornea. Am J Ophthal, 1955; 40: 368-379.
Schenke-Layland, K; Riemann, I; Opitz, F; König, K; Halbhuber, KJ; Stock, UA.
Comparative study of cellular and extracellular matrix composition of native and tissue
engineered heart valves. Matrix Biol., 2004; 23(2):113-25.
Sekundo, W; Marshall, GE; Lee, WR; Kirkness, CM. Immuno-electron labelling of matrix
components in congenital hereditary endothelial dystrophy. Graefes Arch Clin. Exp.
Ophthalmol, 1994; 232(6):337-346.
Squirrell, JM; Wokosin, DL; White, JG; Bavister, BD. Long-term two-photon fluorescence
imaging of mammalian embryos without compromising viability. Nat. Biotechnol, 1999;
17:763–767.
Stern, D; Schoenlein, RW; Puliafito, CA; Dobi, ET; Birngruber, R; Fujimoto, JG. Ablation by
nanosecond, picosecond, and femtosecond lasers at 532 nm and 625 nm. Arch
Ophthalmol, 1989; 107(4):587-592.
Stoll, S; Delon, J; Brotz, TM; Germain, RN. Dynamic imaging of T cell-dendritic cell
interactions in lymph nodes. Science, 2002; 296(5574):1873-1876.
Stuart, BC; Feit, MD; Herman, S; Rubenchik, AM; Shore, BM; Perry, MD. Optical ablation
by high-power short-pulse lasers. J. Opt Soc Am (B), 1996; 13:459-468.
Terry, MA; Ousley, PJ; Will, B. A practical femtosecond laser procedure for DLEK
endothelial transplantation: cadaver eye histology and topography. Cornea, 2005;
24(4):453-459.
Teuchner, K; Ehlert, J; Freyer, W; Leupold, D; Altmeyer, P; Stücker, M; Hoffmann, K.
Fluorescence studies of melanin by stepwise two-photon femtosecond laser excitation. J.
Fluorescence, 2000; 10(3):275-281.
Tseng, SC; Smuckler, D; Stern, R. Comparison of collagen types in adult and fetal bovine
corneas. J Biol Chem, 1982; 257(5):2627-2633.
Vogel, A; Noack, J; Nahen, K; Theisen, D; Busch, S; Parlitz, U; Hammer, D; Noojin, G;
Rockwell, B; Birngruber, R. Energy balance of optical breakdown in water at nanosecond
to femtosecond time scales. Appl. Phys. B, 1999; 68(2):271-280.
Vogel, A; Venugopalan, V. Mechanisms of pulsed laser ablation of biological tissues. Chem
Rev, 2003; 103(2):577-644.
Williams, RM; Zipfel, WR; Webb, WW. Multiphoton microscopy in biological research.
Curr. Opin. Chem. Biol., 2001; 5(5): 603-608.
Williams, RM; Zipfel, WR; Webb, WW. Interpreting second-harmonic generation images of
collagen I fibrils. Biophy J, 2005; 88:1377-1386.
Wang, BG; König, K; Riemann, I; Krieg, R; Halbhuber, KJ. Intraocular Multiphoton
Microscopy with Subcellular Spatial Resolution by Infrared Femtosecond Lasers.
Histochem Cell Biol., 2006(a); 126(4):507-515.
Wang, BG; König, K; Riemann, I; Schubert, H; Halbhuber, KJ. Multiphoton Imaging of
Corneal Tissue with Near-Infrared Femtosecond Laser Pulses: Corneal Optical

Tomography and Its Use in Refractive Surgery. In: Periasamy A, So PTC (Editors). Title:
Multiphoton Microscopy in the Biomedical Sciences VI. San Jose: SPIE; 2006(b);
6089(59):1-12.
Wang, BG; Riemann, I; Schubert, H; Schweitzer, D; König, K; Halbhuber, KJ. Multiphoton
microscopy for monitoring intratissue femtosecond laser surgery effects. Lasers Surg
Med, 2007(a); 39(6):527-533.
Wang, BG; Riemann, I; Halbhuber, KJ; Schubert, H; Kirste, S; König, K. In-vivo animal
follow-up studies on intrastromal surgery with near-infrared nanojoule femtosecond laser
pulses. In: Jacques SL; Roach WP (Editors). Title: Optical Interactions with Tissue and
Cells XVI. San Jose: SPIE; 2005(a); 5695(47):292-302.
Wang, BG; Lohmann, CP; Riemann, I; Schubert, H; Halbhuber, KJ; König, K. Multiphoton-
mediated Corneal Flap Generation Using the 80 MHz Nanojoule Femtosecond Near-
Infrared Laser. J. Refract Surg, 2008(b); accepted.
Wang, BG; Riemann, I; Schubert, H; Halbhuber, KJ; König, K. In-vivo intratissue ablation by
nanojoule femtosecond near-infrared lasers. Cell Tissue Res., 2007(b); 28(3):515-520.
Wang, BG; Halbhuber, KJ; Riemann, I; König, K. In-vivo Corneal Nonlinear Optical
Tomography based on Second Harmonic and Multiphoton Autofluorescence Imaging
induced by Near-Infrared Femtosecond Lasers with Rabbits. In: Chatard JP, Dennis PNJ
(Editors). Title: Detectors and Associated Signal Processing II. Jena: SPIE; 2005(b);
5964(41):1-11.
Wang, BG; Halbhuber, KJ. Corneal Multiphoton Microscopy and Intratissue Optical
Nanosurgery by Nanojoule Femtosecond Near-Infrared Pulsed Lasers. Ann Anat,
2006(c); 188(5):395-409.
Wang, BG; Halbhuber, KJ; Riemann, I; Schubert, H; Kirste, S; König, K. Application of
multiphoton autofluorescence imaging (MAI) and second harmonic generation (SHG) to
intrastromal surgery with rabbits. abstract, FOM, 2005(c). www.focusonmicroscopy.org/
2005/PDF/073_Wang.pdf.
Wang, BG. Kornea Laserchirurgie und Multiphoton Mikroskopie mittels Naher-Infrarot
Nanojoule Femtosekunden Laser mit Kaninchen. Dissertation. Friedrich-Schiller-
Universität Jena, Medizinische Fakultät. 2006(d); urn:nbn:de:gbv:27-20061009-144159-
3.
Wang, BG; Eitner, A; Lindenau, J; Halbhuber, KJ. Cellular two-photon microscopy in ocular
tissues. abstract, Cytometry A, 2007(c); 71A (7):511-531.
Wang, BG; König, K; Halbhuber, KJ. Intraocular nonlinear optical tomography and corneal
flap generation using nanojoule femtosecond near-infrared lasers. abstract, Clinical and
Surgical Ophthalmology, 2007(d); 25(7):246.
Wang, BG; Eitner, A; Lindernau, J; Halbhuber, KJ. High-Resolution Two-Photon Excitation
Microscopy of Ocular Tissues in Porcine Eye. Lasers Surg. Med., 2008(a); 40(4):247-
256.
Watanabe, W; Arakawa, N; Matsunaga, S; Higashi, T; Fukui, K; Isobe, K; Itoh, K.
Femtosecond laser disruption of subcellular organelles in a living cell. Opt Express,
2004; 12:4203-4213.
Wilson, MJ. Structure of the corneal stroma. Vision Res, 1970; 10(6):519-520.
Wilson, SE. Analysis of the keratocyte apoptosis, keratocyte proliferation, and myofibroblast
transformation responses after photorefractive keratectomy and laser in situ
keratomileusis. Trans Am. Ophthalmol Soc., 2002; 100:411-433.

Xu, C; Williams, RM; Zipfel, WR; Webb, WW. Multiphoton Excitation Cross-Sections of
Molecular Fluorophores. Bioimaging, 1996; 4(3):198-207.
Zimmermann, DR; Trueb, B; Winterhalter, KH; Witmer, R; Fischer, RW. Type VI collagen is
a major component of the human cornea. FEBS Lett, 1986; 197(1-2):55-58.
Zipfel, WR; Williams, RM; Webb, WW. Nonlinear magic: multiphoton microscopy in the
biosciences. Nat. Biotechnol., 2003(a); 21(11):1369-1377.
Zipfel, WR; Williams, RM; Christie, R; Nikitin, AY; Hyman, BT; Webb, WW. Live tissue
intrinsic emission microscopy using multiphoton-excited native fluorescence and second
harmonic generation. PNAS, 2003(b); 100(12):7075-7080.
Zoumi, A; Yeh, A; Tromberg, BJ. Imaging cells and extracellular matrix in vivo by using
second-harmonic generation and two-photon excited fluorescence. PNAS, 2002;
99(17):11014-11019.
Zoumi, A; Lu, X; Kassab, GS; Tromberg, BJ. Imaging coronary artery microstructure using
second-harmonic and two-photon fluorescence microscopy. Biophys J., 2004;
87(4):2778-2786.
Chapter 8
OPTICAL PROPERTIES OF RARE EARTH IONS

INDUCED BY FEMTOSECOND LASER
Lixin Yu∗1 and Zhongxin Liu2

1
Department of Materials Science and Engineering, Nanchang University,
330031, Nanchang, P. R. China
2
College of Materials and Chemical Engineering, Hainan University
58 Renmin Road, 570228, Haikou, P. R. China
ABSTRACT
Ultrafast titanium-sapphire tunable femtosecond (fs) laser can produce fs and
provide high-density to be used as excitation resource, it is ideal tool to study the optical
properties of materials. Rare earth (RE) compounds have many potential and practical
applications in display device, upconverson laser, fluorescent probe, optical

communications and fiber optic amplifiers, etc. The upconversion luminescence (UCL)
and fluorescent changes of RE ions in transparent hosts and complex induced by fs laser
have aroused a large amount of interests. In this review, we systematically reported the
progresses of UCL of RE ions (such as Eu3+, Tb3+, Ce3+ and Dy3+, etc) in glass,
crystallized hosts, and complexes, including luminescent characteristic, upconversion
mechanisms and electronic transition processes. The changes of optical properties of RE
ions by fs laser irradiation can achieve valence manipulation of RE ions and
microfabrication. Thus in this review we also intensively reported the progresses of this
topic, including persistent spectral hole burning, structural changes, dynamic processes
etc.
∗
Email: yulixin72@yahoo.com.cn
280 Lixin Yu and Zhongxin Liu
INTRODUCTION
Because RE ions have a plenty of energy levels, the emissions can be modulated
according to practical applications. Table 1 is the electronic configuration of RE [1]. RE ions
doped glasses and crystals are extensively applied in optoelectonic devices.
Table 1. Electronic configuration and ground states of trivalent (R3+) RE ions
Atomic number element R2+ R3+

57 La Lanthanum 4f0, 1S0
58 Ce(a) Cerium 4f1, 2F5/2
59 Pr(a) Praseodymium 4f2, 3H4
60 Nd Neodymium 4f3, 4I9/2
61 Pm Promethium 4f4, 5I4
62 Sm Samarium 4f5, 6H5/2
63 Eu Europium 4f7, 8S7/2 4f6, 7F0
64 Gd Gadolinium 4f7, 8S7/2
65 Tb(a) Terbium 4f8, 7F6
66 Dy Dysprosium 4f9, 6H15/2
67 Ho Holmium 4f10, 5I8
68 Er Erbium 4f11, 4I15/2
69 Tm Thulium 4f12, 3H6
70 Yb Ytterbium 4f13, 2F7/2
71 Lu Lutecium 4f14, 1S0
(a) Ce, Pr and Tb may also be stabilized in tetravalent oxidation states. The electronic configuration and
ground state of RE in tetravalent state can be determined from this table.
The optical characteristics of RE ions have been widely investigated due to their potential
and practical applications in lighting device, display (such as cathode ray tubes, plasma
display panel, etc), optical communication, high-density optical storage, infrared detection,
and medical imaging, etc. As known, most of fluorescence emitters usually follow the well-
known principle of the Stokes law which simply states that excitation energy is higher than
emitted ones. This, in a sense, is an indirect statement that the efficiency cannot be larger than
1. On the contrary, anti-Stokes principle, as well known frequency UCL. The key feature of
UCL of RE ions is able to absorb and combine two or more photons of low near-infrared
(NIR) to emit a higher energy visible photon. Before the 1960s, all anti-Stokes emissions,
which were known to exist, involved emission energies in excess of excitation energies by
only a few kT. They were originated from the thermal population of energy states above
excitation states by such an energy amount. It was the well-known case of anti-Stokes
emission for the so-called thermal bands or in the Raman effect for the well-known anti-
Stokes sidebands [2]. Thermoluminescence, where traps are emptied by excitation energies of
the order of kT, also constituted a field of anti-Stokes emission of its own. Superexcitation,
i.e. raising an excited electron to an even higher level by excited-state absorption (ESA), was
also known. In previous, The studies of UCL were concentrated on the Er3+, Tm3+, Yb3+,
Ho3+, Pr3+ and Sm3+ due to many fascinating applications in diode laser, solid-state-laser [3-
22]. Because Eu3+, Tb3+, Dy3+, Ce3+ and some transition metal ions doped hosts are
Optical Properties of Rare Earth Ions Induced by Femtosecond Laser 281
extensively applied to commercial phosphors and display devices, the down-conversion

luminescence of these ions in different hosts has been extensively investigated [23-32]. But to
our knowledge, the resorts on UCL of these ions in crystallized and transparent hosts at NIR
excitation are very rare since the energy levels of them can not easily match with those of
other RE ions. In fact, UCL of above RE ions in different hosts can be achieved due to the
advent of ultrafast titanium-sapphire tunable NIR laser that can produce fs laser and provide
high-density to be used as excitation resource. As a new excitation source, the NIR fs laser
has more advantages than other ones. First, the fs laser can produce more intense pulses with
the higher repetition rates and shorter pulse duration. Second, the higher excitation density of
the fs pulse laser can achieve simultaneous multiphoton absorption more easily than the other
pulse lasers. Third, the infrared fs laser can penetrate the crystals and transparent hosts
without breaking the overall of the hosts including transparent crystallized hosts. Thus fs can
achieve the microfabrication and modulation of the optical properties. Therefore, the infrared
fs pulse laser has been applied to investigate the mechanisms of multiphoton absorption
frequency UCL and offers a possibility of near-infrared-pumped lasers. In this review, we
systematically and mainly discussed the recent advances in UCL properties of Eu3+, Tb3+,
Ce3+, Dy3+ and Sm3+ induced by fs laser.
In general, four mechanisms my be involved in UCL [1]: Besides (1) multistep excitation
due to ESA, there is (2) the very efficient process of UCL by sequential energy transfers due
to the fact that in many RE-doped solids ions may be connected energetically by energy
transfer diffusion. This last phenomenon has to be distinguished from (3) a third process,
namely cooperative up-conversion either between two ions or between a pair of ions and a
third one. Though some of their theoretical behaviors are rather analogous those of
upconversion by energy transfers, their efficiencies are usually much lower because they
involve virtual levels which have to be described in a higher order of perturbation. (4) A
fourth process, the photon avalanche effect. In ESA, another mechanism of UCL of RE ions
may work induced by fs laser because of high excitation energy: simultaneous multiphoton
absorption (MPA). In 1961, two-photon simultaneous absorption in CaF2:Eu2+ crystals was
observed [33]. From then on, the studies of simultaneous MPA were focused on the organic
materials. Although it is difficult to obtain simultaneous MPA in inorganic hosts, the fs laser
with high excitation density can achieve simultaneous MPA in inorganic solid-state materials.
Based on this point, simultaneous MPA of some RE ions in inorganic solid state hosts has
been observed.
To better understand and compare the mechanism of UCL of RE ions, emission intensity
of RE ions as a function of the excitation power is appreciate tool to study the UCL
mechanisms. The UCL intensity can be expressed as follow,
I∝pn
where I is the integrated intensity of UCL, p is the pump power of fs laser, and n is the
number of photons absorbed per visible photon emitted for any UCL mechanism. According
the above equation, n is the slope of log-log plot between the UCL intensity and pump power.
In this review, “n” represents “photon number”.
The considerable efforts have been made to obtain 3D selective microfabrication with
infrared fs laser to meet the requirements of high performance data storage. Materials
processing using fs laser has attracted many attentions in both scientific and technological
aspects. One advantage of focused ultrashort pulse laser microprocessing is its very large
electric field intensity, which is generated in the range of several terawatts per centimeter
squared [34]. More, ultrashort laser pulses can also induce nonlinear optical effects, so they
can be used with transparent materials in very small areas of the order of several micrometers
in size. For the microfabrication using NIR fs laser, previous reports include the metal
reduction and morphology control [35-37], the generation of submicrometer pores in glass
hosts [38] and optical defects in LiNO3 crystals [39] induced by fs laser, site-selecttive
formation of Si nanocrystals in SiO2 by fs laser irradiation [40], the phase changes of TiO2
[41], the optical changes of semiconductors by fs laser illumination [42,43], photonic crystals
[44,45], and waveguide [46,47] etc. The interaction between fs laser and transparent hosts
containing RE ions includes UCL, photoreduction of RE ions, the structural changes
surrounding RE ions or void creation etc. In this review, we mainly introduced the recent
advancements about the spectroscopic changes of RE ions, such as valence changes of RE
ions, persistent spectral hole burning (PSHB), structural changes and dynamic processes etc.
DISCUSSION
1. UCL Properties of RE Induced By Fs Laser
1.1. The UCL Properties of Eu3+ and Eu2+ Under Fs Laser Irradiation
Because Eu3+ ions are hypersensitive to the hosts and local microstructures and Eu3+
doped hosts are universally applied to commercial phosphors and fluorescent probe, the
down-conversion luminescence of Eu3+ in different hosts has been extensively investigated.
Because the energy levels of Eu3+ ions can not easily match with those of other RE ions, the
resorts on UCL of Eu3+ in different hosts at NIR laser excitation were very less. In order to
better analyze the UCL mechanisms of Eu3+ ions, the energy levels of Eu3+ are shown in
Figure 1. For fs laser at about 800 nm, the absorption of 7F0-5L6 transitions at about 396 nm is
imperfect matched by a two-photon absorption. The charge transfer (CT) absorption peaking
at about 260 nm is precisely matched by a three-photon absorption.
In 2001, Lakowicz and co-worker systematically reported the UCL spectra and properties
of Eu3+ and Tb3+ in D2O solution at fs laser excitation [48]. Figure 2 shows the emission
spectra of Eu3+ in D2O solution at 796-nm fs laser (a) and 394-nm light excitation (b), and the
emission intensity as a function of fs laser pump power (c). It is obvious that the spectral
shape of Eu3+ ions is similar for downconversion and UCL. The emission intensity at 592 nm
under 796-nm fs laser is found to depend quadratically on the incident power, indicating a
two-photon process.
This is the first reported emission of an unchelated lanthanide in solution with UCL at
two-photon excitation. After then on, You et al. reported the UCL characteristics of Eu3+ in
solid-state Al2O3-SiO2 sol-gel glasses [49]. Figure 3 shows the emission spectra of Eu3+ in
Al2O3-SiO2 glass (a), the log-log plot between the emission intensity and the incident power
of fs laser (b), and absorption spectra (c). The spectral configuration of Eu3+ ions
corresponding to CT and fs laser excitation is similar, implying that they have the same
origins. Moreover, the red emission from the 5D0-7F2 transitions is stronger than the orange
emissions from the 5D0-7F1 transitions. In Figure 3 (b), the slope is obtained to be about 2.91
by linear fitting, indicating the UCL is a three-photon process. In absorption spectra, the
broad band absorption ranging from 200 to 310 nm is from CT absorption and several weaker
absorption lines at low energy side are associated from the f-f transitions. Based on the
analysis of absorption of Al2O3-SiO2 containing Eu3+ ions and UCL processes, the
mechanisms of UCL of Eu3+ ions in this host are attributed to simultaneous three-photon
absorption. At fs laser excitation (800 nm), the three-photon simultaneous absorption leads to
the population of 4f levels via the excitation of the CT state of Eu3+ ions, then relaxes to the
lowest 5D0 level, as a consequence that the 5D0-7FJ transitions emissions are observed. Qiu et
al. also reported the similar UCL phenomena. They observed the highly efficient UCL of
Eu3+/Tb3+ codoped silicate glasses by fs laser excitation [50]. The origins of UCL for both
Eu3+ and Tb3+ in silicate glass can be attributed to the three-photon simultaneous absorption
that leads to the population of the 5d level of the Eu3+ and Tb3+ ions.
CT
three-photon absorption
5
L6
5
D3
5
D2
5
D1
5
D0
Two-photon absorption
5
4 7
3
2 FJ
1
Figure 1. Energy levels of Eu3+ ions and possible UCL processes.
Figure 2. Emission spectra of Eu3+ in D2O at 796 nm fs laser (a) and 394 nm (b) excitation, the
dependence of the lanthanide emission intensity on the laser power (c). (From Ref. 48).
Figure 3. Emission spectra of Eu3+ in Al2O3-SiO2 glass under fs laser irradiation (800 nm) and 267 nm
excitation (a), fluorescent intensity of the 5D0-7F2 transitions of Eu3+ as a function of pump power (b),
and the absorption spectra (c). (From Ref. 49).
The electrons in the upper 5d state then relaxes nonradiatively to the bottom 5d state of
5
D0 (Eu3+), 5D3 and 5D4 (Tb3+), and then radiatively to terminal state, creating characteristic
emissions of Eu3+ and Tb3+ ions. This result implies that the UCL processes of Eu3+ in silicate
melting glasses suggested by Qiu et al. is different from the result of Eu3+ in Al2O3-SiO2 sol-
gel glasses, in which the population of 4f level of Eu3+ is caused by the excitation of CT state
of Eu3+ through absorbing three photons simultaneously.
It is well known that nanosized materials exhibit some unique physical and chemical
properties in comparison with corresponding to the micrometer materials due to surface and
quantum confinement effect. The UCL of RE ions in nanomaterials induced by fs laser is an
interesting topic. Very recently, the UCL of Eu3+-doped Y2O3 nanoparticles and
Y2O3:Eu3+/SiO2 core/shell nanostructures was reported by Lu et al. [51]. Figure 4 is the UCL
spectra of Y2O3:Eu3+ and Y2O3:Eu3+/SiO2 nanoparticles under fs laser irradiation (a), double
logarithmic plots of UCL intensity of Y2O3:Eu3+ nanoparticles as a function of pump power
(b), and the UCL mechanisms (c).
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Optical Properties of Rare Earth Ions Induced by Femtosecond Laser 285
Figure 4. UCL spectra of Y2O3:Eu3+ and Y2O3:Eu3+/SiO2 nanoparticles under 800-fs laser (a), double
logarithmic plots of UCL intensity of Y2O3:Eu3+ nanoparticles as a function of pump power (b), and the
UCL mechanisms (c). (From Ref. 51).
It can be seen five emission bands at 581, 586-599, 611-631, 651 and 709 nm, which are
from the 5D0-7FJ (J=0-4) transitions of Eu3+ ions, respectively. It is clear that the emissions of
Eu3+ ions are enhanced after the nanoparticles are coated with SiO2. The longer the coating
time is, the stronger the intensity of Eu3+ ions is. As known, the ratio of surface to volume in
nanocrystals increases greatly than that in the bulk powder. In the surface, a large number of
surface defects exist, which act as the nonradiative transition channels, leading the
nonradiative transition rates to increase and causing that luminescence becomes weak. In
SiO2-coated Y2O3:Eu3+ nanocrystals, the surface defects are decreased because the
cooperative ligand fields between the Y2O3:Eu3+ core and non-crystalline SiO2 shell interface
active the ‘dormant’ Eu3+ ions near or on the surfaces of nanoparticles to participate in the
UCL process, leading the intensity of Eu3+ in coated samples is stronger than that of
noncoated samples. Authors prove that the core-shell structures can increase the radiative
quantum efficiency of Eu3+-doped nanoparticles according to J-O theory analysis. In Figure 4
(b), the slope of the double logarithmic fitted line is 2.04, indicating that the UCL of Eu3+ in
Y2O3 nanocrystals is a two-photon process, which is different from the above mentioned UCL
process of Eu3+ ions in silicate glasses [49, 50]. Authors proposed that the mechanism of UCL
of Eu3+ ions is a simultaneous two-photon excitation process. The two-photon simultaneous
absorption under 800-nm fs laser irradiation can be excited through a ‘virtual’ level to the
excited level 5L6 of Eu3+ with the aid of phonons of the host matrix. Detailed UCL process is
shown in Figure 4 (c). As known, glass-ceramic consists noncrystalline glassy phase and
nanocrystalline phase. Due to RE ions doped this material for their uses in telecommunication
systems, such as upconversion fibers, optical amplifiers, solid-state lasers and 3D displays
[52–54], efforts have been focus on finding materials with low-energy phonons, in order to
reduce the multiphonon non-radiative de-excitation and improve cross-sections of the RE
ions. In previous, the UCL of RE ions in glass-ceramic was focus on the Er3+, Tm3+ etc. In
2007, Qiu and coworker reported the UCL properties of Eu3+ in transparent SrO-TiO2-SiO2
glass-ceramics containing Sr2TiSi2O8 microcrystals with a very large second-order optical
nonlinearity for the first time [55]. Figure 5 show the emission spectra of Eu3+-doped glass (a)
and glass-ceramic (b) (left), the dependence of luminescent intensity of Eu3+-doped glass (a)
and glass-ceramic (b) on the fs laser pump power (right). The emission spectra of Eu3+ in both
samples under 790-nm fs and 394-nm light excitation are similar. It should be noted that an
emission at about 395 nm in the 790-nm fs laser irradiated glass-ceramic appeares, while it is
absent in the glass samples. This emission is assigned to the second-harmonic generation of
the fs laser when excited by fs laser. It is vital important that the UCL intensity of Eu3+ ions in
glass-ceramics is about 10 times stronger than that in the glass under the same condition
because the absorption of the second-harmonic generation as a result of precipitated
Sr2TiSi2O8 microcrystalline particles with second-order optical nonlinearities at 790 nm fs
laser excitation. The energy transfer from microcrystalites to Eu3+ ions takes place. The slope
(n) of double logarithmic fitted lines is determined to be 1.90 in glass-ceramic and 1.5 in
glass, indicating that the UCL process of Eu3+ ions in both glass-ceramics and glasses is two-
photon simultaneous excitation at fs laser excitation (790 nm). Mostly important, the damage
threshold of the glass-ceramics evidently increases in comparison with the glass. These
results implied the glass-ceramics containing RE ions will be a promising three-dimensional
display and data storage materials.
Figure 5. Emission spectra of Eu3+-doped glass (a) and glass-ceramic (b) under 794-nm fs laser
irradiation and 394-nm light excitation (left), the dependence of luminescent intensity of 5D0-7F2
transitions of Eu3+-doped glass (a) and glass-ceramic (b) on the fs laser pump power (right). (From Ref.
55).
We observed that different UCL process of Eu3+ ions in ZnO-SiO2 and SiO2 sol-gel
glasses [56]. Figure 6 is the emission spectra Eu3+ ions in ZnO-SiO2 and SiO2 glasses at fs
laser (800 nm) excitation (a), and excitation spectra monitoring the emission at 618 nm (b). It
can be seen that there is no evident difference of Eu3+ ions in both samples. The slope of
double logarithmic fitted lines is obtained to be 3.05 in ZnO-SiO2 glass and 1.79 in SiO2
glass, indicating that two-photon excitation of Eu3+ ions in SiO2 glasses dominates and three-
photon excitation dominates in ZnO-SiO2 glass. We carefully analyzed the excitation
processes (Figure 6 (b)). No excitation lines for ZnO-SiO2 glass with the absence of Eu3+ ions
were observed. In both glasses containing Eu3+ ions, the broad band extending from 220 to
310 nm was associated with CT transition from 2p orbital of O2- ions to the 4f orbital of Eu3+
ions, while the sharp lines at 317, 360, 380, 396, 412 and 464 nm with the direct excitation of
f-f shell transitions of Eu3+, which can be contributed to 7F0-5H5, 7F0-5D4, 7F0-5D4,3,2, 7F0-5L6,
7
F0-5D3 and 7F0-5D2 transition absorption, respectively. It is obvious that relative proportion of
f-f transitions to CT absorption in SiO2 glasses is much higher than that in ZnO-SiO2 glass.
According to the excitation spectra, the ratio of f-f transition to CT absorption in SiO2 glass
greatly increased in contrast with that in the ZnO-SiO2 glass. Thus the electrons at the ground
state 7F0 can be excited to 5L6 level through two-proton simultaneous absorption in SiO2
glass. The electrons at 5L6 level can nonradiatively relaxed to 5D3, 5D2, 5D1, and resultantly to
lowest excited state 5D0 level, and as a consequence 5D0-7FJ transition emissions occurre. In
ZnO-SiO2 glass, electrons are also excited from ground state 7F0 to 5L6 level of Eu3+ through
two-photon simultaneous absorption firstly. One the one hand, because the two-photon
absorption (400 nm) locates at the edge of 7F0-5L6 transition absorption (396 nm), the
absorption cross is low.
On the other hand, the ratio of f-f transitions to CT absorption remarkably decreases in
ZnO-SiO2 glasses Thus electrons at 5L6 level can absorb another photon to be resonantly
excited to CT band. The energy transfer from CT band to 5D3 level occurres. The followed
luminescent processes are same with that in SiO2 glasses.
Figure 6. Emission spectra of Eu3+ in SiO2 and ZnO-SiO2 glass at 800-fs laser excitation (a), and
excitation spectra monitoring the emission of Eu3+ ions at 618 nm (b). (From Ref. 56).
We introduced the recent advancement of UCL of Eu3+ in different solid-state hosts.

Different UCL processes are observed and different UCL mechanisms are presented. In order
to further investigate the UCL mechanisms, they need be studied in different hosts and more
spectral and dynamic data are obtained. These investigations of UCL of Eu3+ ions can offer a
new approach to obtain visible laser and enrich the understanding of UCL of Eu3+ in
transparent hosts and develop advanced phosphors.
There exist two valence-states of europium, Eu3+ and Eu2+. Above discussions of UCL
are focused on three valence Eu3+ ions. In fact, Eu2+ is also an important RE ions. Differently
from Eu3+, the transitions of Eu2+ are allowed f-d transitions. Although the two-photon
excitation of Eu2+ in CaF2 has been studied in 1961 [33], the little researches of UCL of Eu2+
were reported from then on. Recently, the UCL of Eu2+ in silica glasses was presented [57].
Figure 7 show the emission spectra of Eu2+doped silicate glasses (a), log-log plots of
integrated intensity vs pump power (b), and UCL process (c). The spectral profile excited by
fs laser is almost same as that by UV monochromatic light, indicating that the emission in
both cases is from an identical origin. The slope of double logarithmic fitted lines is about
2.0, implying that the blue UCL of Eu2+ is a two-photon process. Based on the absorption
analysis, the detailed UCL process is shown in Figure 7 (c). At 800 nm fs laser excitation, the
electron in the ground state of Eu2+ absorbs two photons to be excited to one of the higher 5d
levels with the help of a virtual energy level, then nonradiative relax to the lowest 5d level.
Subsequently, the transition from the lowest 5d levels to 8S7/2 and blue emissions of Eu2+ are
observed.
Above mentioned hosts are inorganic hosts. In fact, narrow emission bands and
millisecond luminescence lifetimes make lanthanide ions and their complexes with
photosensitizing organic ligands into useful photoluminescent species for biomolecular
research and medical diagnostics.
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Optical Properties of Rare Earth Ions Induced by Femtosecond Laser 289
Figure 7. Emission spectra of Eu2+ doped silicate glasses excited by 800-nm fs laser and 400-nm
monochromatic light (a), log-log plots of integrated intensity vs pump fs laser power (b), and UCL
process (c). (From Ref. 57).
Multiphoton excitation has the advantages of localized excitation at the optical focal
point and minimal photochemical and heating effects due to the absorption of water. Based on
this point, Werts and coworker reported the UCL of europium complexes solution at NIR fs
laser irradiation [58]. The UCL of europium complexes is a two-photon process. Fu et al.
suggested that the two-photon excitation of europium complexes will be promising in
biomedical probe and diagnoses [59].
1.2. The UCL of Tb3+ Ions Under Fs Laser Irradiation

As for UCL of Tb3+ ions, the low emission cross section is a disadvantage. But, this
disadvantage is directly related to the main advantages of Tb3+ ions, which make it attractive
for tunable-laser application. The advantages are the large number of luminescence range.
Another advantage of Tb3+ ion is low multiphoton losses from the 5D4 and 5D3 levels that are
good for any upper laser levels [60]. Rai and coworker reported pure Tb3+ in tellurite glass
gives no fluorescence on excitation with 980 nm diode laser, as there is no absorption at this
energy [61]. Even a small amount of Yb3+ added to the glass results in intense luminescence
from Tb3+ ions. Thus the UCL properties of in Yb3+ and Tb3+ codoped glasses and
crystallized hosts have been extensively investigated under the excitation of diode laser (980
nm), YAG:Nd laser (1064 nm), or fs laser, such as Tb-Yb codoped tellurite glasses on 980-
nm excitation [61, 62], Yb-Tb doped SrCl2 single crystals at fs laser excitation (942 nm)
[63],Yb-Tb codoped silica sol-gel glasses at 980-nm or 1064-nm laser excitation [64], Tb-Yb
codoped YSGG [61] and SrCl2 [65] at fs laser excitation, Tb-Yb doped fluoroindate glasses
[66-68] and crystals [69], and phosphate phosphors [70]. The mechanisms of UCL of Yb-Tb
codoped hosts are also arributed to the cooperative energy transfer from Yb3+ to Tb3+ ions, as
a sequence of fluorescence emissions of Tb3+ ions. The similar results were observed in
terbium protein complex [71]. Recently, under NIR fs laser irradiation, the UCL of pure Tb3+
ions was also reported. In 2001, Lakowicz et al reported weak green UCL of pure Tb3+ ions in
D2O at 768 nm fs laser excitation [48]. Figure 8 shows the emission spectra of Tb3+ in D2O at
fs laser excitation (a) and 256 nm light from Xe lamp (b), and the dependence of the emission
intensity at 543 nm on incident power (c).
Figure 8. Emission spectra of Tb3+ in D2O at 768 nm fs laser excitation (a) and 256 nm light from Xe
lamp (b), and the dependence of emission intensity at 543 nm on incident power (c). (From Ref. 48).
Figure 9. Emission spectra of Tb3+ in ZnO-B2O3-SiO2 glass at 800-nm fs laser excitation (a), 267 nm
excitation (b), and the excitation spectra monitoring 545nm (c). (From Ref. 72).
Figure 10. Dependence of UCL intensity of 5D4-7F5 transition of Tb3+ in ZnO-B2O3-SiO2 glass on pump
power (a), and absorption spectra (b). (From Ref. 72).
For multiphoton (Figure 8 (a)) and one photon (Figure 8 (b)) excitation of Tb3+, no
evident difference is observed. The results shown in Figure 8 (c) indicate that UCL of Tb3+ in
D2O is a three photon process. But authors do not discuss the detailed UCL mechanisms.
After then on, Qiu and coworker reported the multi-photon absorption frequency UCL of
Tb3+-doped ZnO-B2O3-SiO2 glass excited by a 800-nm fs laser [72]. Figure 9 shows the
emission spectra at 800-nm fs laser (a) and 267 nm light excitation (b), and the excitation
spectra monitoring 545nm (c). There exists four emission peaks at about 490, 545, 580 and
620 nm, which can be assigned to 5D4-7FJ (J=3-6) transitions of Tb3+ ions. The emission
configuration is similar for both excitation wavelengths, indicating the downconversion
luminescence and UCL of Tb3+ ions have the same origins. Figure 10 is the dependence of
UCL intensity of 5D4-7F5 transition of Tb3+ on pump power (a), and the absorption spectra (b).
According to Figure 10 (a), the UCL process of Tb3+ ions is three-photon excitation. The
broad band absorption centered at 267 nm is associated with the 4f-5d transitions of Tb3+ in
ZnO-B2O3-SiO2 glass. After the analysis of absorption of Tb3+ ions, the mechanisms of UCL
of Tb3+ in this hosts are attributed to the three-photon simultaneous absorption. Pumping the
Tb3+ by using 800 nm fs laser causes population of electrons in the excited state, the excited
electrons nonradiatively relax to the lowest 5D4 level, and radiatively return to the 7FJ levels,
leading to green emissions of Tb3+. The three-photon absorption coefficient of Tb3+ doped
ZnO-B2O3-SiO2 glass is determined by fitting, 8.24x10-24cm3W-2.
Zheng et al. reported different research results. They reported the UCL properties of
Tb -doped lithium niobate crystal at the room temperature for the first time [73] , as Tb3+
3+
luminescence thermal quenching is very obvious in LiNbO3 crystals at temperature about 150
K [74]. Figure 11 is the room-temperature UCL spectra of Tb3+ under 800-nm fs laser
irradiation (a), and logarithmic plot of the pump energy dependence of the UCL intensity
under 800-nm excitation. The UCL spectra of Tb3+ in LiNbO3 crystals are similar with that in
silicate glass. But the dependence of 5D4-7F5 transitions of Tb3+ on incident powder of fs laser
indicates that the relation of ln(I) versus ln(P) only holds for low excitation power and will
deviate from the n-power for high excitation power. The slope at the low excitation power is
about 2 which implies that the UCL process of Tb3+ in LiNbO3 crystls is a two-photon
excitation. The slope of about 1 for high pump power indicates that the saturation effect takes
place. The two-photon excitation of Tb3+ ions in LiNbO3 crystals at 800-nm NIR fs laser is
different from the three-photon simultaneous absorption of Tb3+-doped in ZnO-B2O3-SiO2
glass glasses reported by Qiu et al. [72]. This is due to that the valence electron transition of
LiNbO3 matrix from 2p orbits of O2- to 4d orbits of Nb5+ has large absorption coefficients
(~1.4 cm-1) at 400 nm which is precisely matched by two-photon absorption.
Figure 11. The room-temperature UCL spectra of Tb3+ in LiNbO3 crystals under 800-nm fs laser
irradiation (a), and dogarithmic plot of the pump energy dependence of the UC intensity under 800-nm
excitation (b). (From Ref. 73).
Our group also observed the strong UCL of Tb3+ in ZnO-SiO2 and SiO2 sol-gel glasses by
800-nm fs laser excitation. The saturation effect of Tb3+ in SiO2 glasses at high pump power
occurs, implying that the damage threshold of ZnO-SiO2 glasses increases in comparison with
SiO2 glasses. The UCL mechanisms of Tb3+ in SiO2 and ZnO-SiO2 glass are the three-photon
simultaneous excitation.
Above research results indicated the UCL properties of Tb3+ ions are strongly dependent
on the host matrixes. In order to further clarify the UCL properties of Tb3+ ions, UCL
characteristics of Tb3+ in different hosts (such as glass, glass-ceramics, crystals) should be
further studied, including emission spectra, dynamics processes etc.
1.3. UCL Properties of Ce3+ Under Fs Laser Irradiation

Differently from the Eu3+ and Tb3+ ions, both excitation and emission transitions of Ce3+
ions are from f-d transitions. Thus the excitation and emission spectra exhibit broad bands
due to allowed f-d transitions, not sharp lines of forbidden f-f transitions. Therefore, it is
suitable candidate for ultraviolet UCL applications.
Figure 12. Emission spectra of Ce3+ in Al2O3-SiO2 glass (a), the dependence of intensity Ce3+ on the
incident power (b), and the absorption spectra (c). (From Ref. 75).
.
However, the studies on UCL properties of Ce3+ ions are paid little attention. 2004, You
reported the UCL of Ce3+ in Al2O3-SiO2 sol-gel glasses at 800-nm fs laser excitation [75].
Figure [12] show the emission spectra of Ce3+ in SiO2-Al2O3 glasses under 800 nm fs laser
irradiation (a), the dependence of intensity of Ce3+ on the incident fs laser power (b) and the
absorption spectra (c). At fs laser excitation, the broad band emission centered at 392 nm is
from 5d→4f transition of Ce3+ ions. From Figure 12 (b), the slope of the logarithmic fitted
line is about 3, indicating that a three-photon excitation predominates in the UCL process of
Ce3+ in Al2O3-SiO2 glasses at 800-nm laser excitation. According to absorption properties
(Figure 12 (c)) and level analysis, UCL mechanisms of Ce3+ ions in this glass can be
suggested to be three-photon simultaneous excitation. The electron in the ground state of 2F5/2
level is first excited into the upper 5d excited state via a three-photon simultaneous absorption
that causes a population of the upper 5d level of the Ce3+ ions. The electron in the upper 5d
state relaxes nonradiatively to the lowest state of 5d, as a consequence the characteristic
emissions of Ce3+ ions follow.
Recently, the UCL properties of Ce3+ in Gd2SiO5 [76, 77] and Lu2Si2O7 [78] single
crystals, BK7 glasses [79], YAP crystals [80] were also investigated under 800-nm fs laser
excitation. Although the hosts are different, the UCL mechanisms of Ce3+ ions are same, as a
sense, three-photon simultaneous absorption.
1.4. UCL Properties of Other RE and Transition Metal Ions Under Fs Laser
Irradiation
Because of the dense energy levels of Dy3+ ions, favoring radiationless decay of excited
levels by multophinon relaxation, the reports on pure UCL of Dy3+ are virtually rare. In 1998,
Wermuth et al. systematically reported the UCL of pure Dy3+ ions in CsCdBr3 single crystals
under the fs laser excitation for the first time [14]. Figure 13 is the UCL spectra of Dy3+ in
CsCdBr3 crystals at 810-nm fs laser excitation (a), normalized uoconversion-excitation
spectra of CsCdBr3:Dy3+ at three different temperature monitoring the emission at 576 nm (b).
According to UCL spectra, the 4F9/2-6H13/2 transitions are the strongest emission, indicating
the yellow-green color of UCL. In addition, the transitions originating from 4M21/2 and 4P3/2
can be observed. They also studied the power dependence for 4F9/2 and 4M21/2 transitions
intensity. The results indicate that it is exactly quadratic, implying a two-photon process to
populate the emission level. The 4P3/2 luminescence show a cubic power dependence,
indicating a three-photon excitation. For upconversion excitation (Figure 13 (b)), with the
increasing temperature, the excitation lines became broader and there arise intense and
extraordinarily sharp lines at 12164 and 12177 cm-1, which is assigned to typical behavior of
ESA excitation. According the energy level of Dy3+ ions, upconversion-excitation spectra at
different temperature, dynamic results and analysis, the mechanisms of UCL of Dy3+ have
been assigned to the ESA and energy transfer upconversion (ETU). Figure 14 shows the
energy level of Dy3+ with two possible ETU mechanisms to populate 4F9/2 after 6F5/2
excitation. Figure 15 shows the energy level of Dy3+ with two possible ESA mechanisms after
excitation into 4F3/2 (a) and 6F5/2 (b), respectively.
Recently, Zheng and coworker reported the UCL of Dy3+ ions in LiNbO3 single crystals
codoped with ZnO [81]. Figure 16 show the UCL spectra of Dy3+ in LiNbO3 crystals under
806-nm fs laser excitation (a) , and the dependence of UCL intensity at 486 nm and 583 nm
emission on incident power (b).
Figure 13. UCL spectra of Dy3+ in CsCdBr3 crystals at 810-nm fs laser excitation (a), uoconversion-
exitation spectra of CsCdBr3:Dy3+ at three different temperature monitoring the emission at 576 nm (b).
(From Ref. 14).
Figure 14. Energy-level diagram of Dy3+ with two possible ETU mechanisms to populate 4F9/2 after
4
F5/2excitation. The cross-relaxation step is indicated with dashed arrows. (From Ref. 14).
Figure 15. Energy-level diagram of Dy3+ with two possible ESA mechanisms after excitation into 4F3/2
(a) and 4F5/2 (b), respectively. The cross-relaxation step preceding ESA is indicated with dashed arrows,
the multiphonon-relaxation step with a curly arrow. (From Ref. 14).
Figure 16. UCL spectra of Dy3+ in LiNbO3 crystals under 806-nm fs laser excitation (a) , and the
dependence of UCL intensity of Dy3+ at 486 nm and 583 nm emission on incident power (b). (From
Ref. 81).
In Figure 16 (a), at fs laser excitation, four UCL bands centered at 483, 581, 663 and 753
nm from 4F9/2 are assigned to characteristic emission of Dy3+ ions. I should be noted that
yellow emission from the transitions of 4F9/2-6H13/2 of Dy3+ in ZnO-codoped crystals is
stronger, whereas the blue emission associating with the 4F9/2-6H15/2 transitions of Dy3+ in
crystals with the absence of ZnO dominates, which can be attributed to that the local
environment around Dy3+ ions in ZnO-codoped LiNbO3 crystals is different from that in
nondoped samples. As known, there is no UCL of Dy3+ ion reported in LiNbO3 crystals
because the multiphonon relaxation causes the short lifetime of potential intermediate state.
Figure 17. Schematic energy-level diagram and UC excitation and visible emission schemes of Dy3+ in
LiNbO3 crystal. Full, doted and curly arrows indicate radiative energy transfer, nonradiative energy
transfer, and multiphonon relaxation processes, respectively. (From Ref. 81).
Authors suggested that the appearance of UCL of Dy3+ ions in LiNbO3 crystals is not
originated from the host with low phonon energy, but from high instantaneous power density
of fs laser. In Figure 16 (b), the n value for the two emission bands is about 1.8 at low pump
energy and about 1.0 at high pump energy. This result indicates that the UCL of Dy3+ is a
two-photon excitation at low pump energy, which is consistent with the reported by Wermuth
[14]. At high pump energy, saturation effect occurres. According to energy level shown in
Figure 17, authors presented the UCL mechanisms of ESA shown in Figure 17. Firstly, Dy3+
ion absorbs a laser photon from 6H15/2 to 6F5/2. The excited 6F5/2 level absorbs another photon
(ESA) to 4M21/2 level. Then through multiphonon relaxation from 4M21/2 to 4F9/2, Dy3+ ion
emits yellow and blue UCL. This group also reported the UCL of Er-Dy [82] and Er-Hf [83]
codoped LiNbO3 single crystals under fs laser excitation. The enhancement of UCL intensity
of Er3+ ions is caused by energy transfer from Dy3+ ions (4F9/2) to Er3+ (4F7/2). The energy
transfer efficiency from Dy3+ to Er3+ is calculated to be 33%. In Er-Hf codoped LiNbO3
single crystals, the HfO2 incorporation has affected on Er3+ radiative lifetime and fluorescent
branch ratio. The UCL mechanisms of Er-Hf codoped LiNbO3 crystals are discussed.
Considering the reported lifetime of 4I11/2 (200 μs) of Er3+, ESA of 4I11/2-4F3/2 of Er3+ ions can
not appear due to short pulse duration (120 fs) and low repetirion frequency. Whereas, they
observed the ESA of 4I9/2-2H9/2 transitions of Er3+ ions (emission at about 410 nm) because of
the fs laser with high instantaneous energy density in favor of sequential absorption
multiphoton. Another ESA, 4I13/2-2H11/2 of Er3+, should also exist. They calculated the Ω value
based on the modified Judd-Ofelt theory. For hf(4% mol):Er(1% mol):LiNbO3, Ω2=2.63x10-
20
cm2, Ω4=2.86x10-20 cm2, and Ω6=0.72x10-20 cm2. The relation of Ω2<Ω4 in this report is
contrary to the Er3+ general trend of Ω2>Ω>Ω6 as the Hf concentration is below its threshold.
YVO4 crystals is a very useful phosphor in use for a long time. In 2006, Qiu and
coworker reported the UCL of pure YVO4 single crystals excited by fs laser irradiation for the
first time [84].
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. 298 Lixin Yu and Zhongxin Liu
Figure 18. Emission spectra of YVO4 single crystals under fs laser and and 267-nm monochromatic
light excitation (a), absorption spectra (b), and the fluorescent intensity of the transition from 3T1, and
3
T2 to 1A1 of VO43- as function of the fs pump power (c). (From Ref. 84).
Figure 18 shows the emission spectra of YVO4 single crystals under fs laser and and 267-
nm monochromatic light excitation (a), absorption spectra (b), and the fluorescent intensity of
the transition from 3T1,and 3T2 to 1A1 of VO43- as function of the fs pump power (c). The
emission configuration of YVO4 at 267- and 800-nm excitation is similar. The spectra exhibit
a broad band centered at 438 nm and a shoulder peak at 464 nm, which can be assigned to the
characteristic emission of YVO4 single crystals. In absorption spectra, a strong absorption
band ranging from 200 to 350 nm is the superposition of intrinsic absorption of vanadate
group. The slope of double logarithmic fitted lines is about 2.93, indicating that the UCL
process of YVO4 is a three-photon absorption. According to absorption characteristics of
YVO4 crystal, they suggested that the UCL mechanisms of YVO4 is three-phonon
simultaneous absorption. The VO43- has two the strong absorption bands ranging from 240 to
350 nm. Under NIR fs laser (800 nm), the VO43- can be excited from ground states to excited
state, which nonradiatively relax to the lowest excited state. As a consequence the
characteristic broad band luminescence of VO43- occurs. Recently, this group also studied and
compared the UCL characteristics of YVO4:Nd3+ at a 800-nm fs laser and a 800-nm
continuous wave (CW) diode laser excitation, respectively [85]. Figure 19 shows the emission
spectra of the Nd3+:YVO4 crystals excited by fs laser and diode laser. Four emission peaks at
1338 (4F3/2→4I13/2), 1064 (4F3/2→4I11/2), 895 ( 4F3/2→4I9/2) and 754 nm (2G9/2+2K13/2→4I15/2)
excited by diode laser are observed, and only the emission at 754 nm belongs to UCL, while
seven group emission lines at 1328 (4F3/2→4I13/2), 1064 (4F3/2→4I11/2), 875 (4F3/2→4I9/2), 731
(2G9/2+2K13/2→4I15/2), 684 (4G7/2→4I13/2), 424 (4D3/2→4I13/2) and 393 nm (4D3/2→4I11/2) under
fs laser irradiation. Five emission peaks belong to UCL. In addition, the emission band at 393
and 424 nm is similar to the one peaking at 422 nm excited by the 267 nm monochromatic
light.
Figure 19. Emission spectra of the Nd3+:YVO4 crystals excited by 267 nm monochromatic light (a) and
800 nm fs laser(b). (From Ref. 85).
Figure 20. Energy level diagram for Nd:YVO4 upconversion process pumped at 800 nm. (From Ref.
85).
This result implies that the emission bands centered at both 393 and 424 nm are probably
MPA transitions of Nd3+. Based on the analysis of level of Nd3+ and UCL mechanisms,
author presented the possible mechanisms of UCL of Nd3+. The emission at 393 and 424 nm
excited by fs laser may be associated with three-photon absorption and the emissions at 684
and 731 nm are probably three-photon absorption or cooperative energy transfer process. As
for the emission at 754 nm excited by the diode laser, it may be due to ESA or cooperative
energy-transfer process. The energy level diagrams for these upconversion processes are
shown in Figure 20. Moreover, it is important that the upconversion fluorescence intensity of
Nd3+ excited by the fs laser is higher than the downconversion one, and that the UCL
efficiency pumped by the fs laser is higher than that by the diode laser. With the development
of the femtosecond laser technique, the frequency upconversion pumped by the fs laser will
be more widely applied. In fact, present group also reported the three-photon simultaneous
absorption of Pr3+ in Y2SiO5 crystals under fs laser irradiation [86]. At 800-fs laser excitation,
strong ultraviolet UCL band from 245 to 360 nm of Pr3+ in Y2SiO5 is observed. The result of
dependence of emission of Pr3+ on fs pumped power indicates three-photon excitation of Pr3+.
Authors suggested that the mechanisms of Pr3+ in Y2SiO5 are the three-photon simultaneous
absorption.
In previous, the UCL of Sm3+ under diode laser excitation has been extensively studied.
Recently, Qiu et al reported the UCL of Sm3+ under NIR fs laser excitation. This group
studied the UCL of Sm3+ doped Ba2TiSi2O8 glasses and glass-ceramics [87]. Figure 21 show
the emission spectra of the Sm3+-doped glass and glass-ceramics excited by 800-nm fs laser
and 400-nm Xe lamp (left), and fluorescent intensity as a function of fs laser pump power
(right). It can be seen that there exist three emission peaks at about 565, 602 and 648 nm
under fs laser excitation, which are similar with the spectra of Sm3+ by Xe lamp. Moreover,
an additional peak at 400 nm in glass-ceramics samples is observed under 800-nm fa laser
irradiation, which originates from the second harmonic generation of the fs laser.
Figure 21. Emission spectra of the Sm3+-doped glass and glass-ceramics excited by 800-nm fs laser and
400-nm Xe lamp (a), and fluorescent intensity as a function of fs laser pump power (b). (From Ref. 87).
A linear correlation between the logarithmic transformation of the excitation power and
the UCL intensity in the glass-ceramics (LnP=2.02), whereas the UCL becomes saturated in
the glass samples at high pump power. This result implies that the two-photon excitation
dominats in both glass and glass-ceramics, and the damage threshold is virtually improves in
the glass-ceramics because of the precipitation of Ba2TiSi2O8 microcrystallites. The
mechanisms of UCL of Sm3+ in both samples can be assigned the two-photon simultaneous
absorption. The enhancement of Sm3+-doped glass-ceramics is due to the effective energy
transfer from second harmonic generation of Ba2TiSi2O8 microcrystallites to the Sm3+ ions. In
fact, the similar phenomenon is observed in Eu3+-doped SrO-TiO2-SiO2 glass-ceramics
containing Sr2TiSi2O8 microcrystals [55]. These results implied that the multi-photon
absorption can be enhanced in glass-ceramic containing RE ions if the effective energy
transfer from microcrystallite to RE ions. This can help us to design advanced phosphors or
optical devices. In addition, the UCL of Sm3+ in Gd2O3 nanopowders excited by low-power
continuous-wave excitation at 514, 561, 594, and 633 nm and with a pulsed fs at 710 nm were
reported by Dosey et al [88]. They observed a broad, multicomponent blue emission band
centered at around 400 nm. This result is significant for the application of Gd2O3:Sm3+ in
fluorescent labeling. The reasons are: Firstly, The blue emission of Sm3+ ions can be excited
in the far end of the spectrum which can minimize the excitation of autofluorescence, a
conmon problem in bioimaging, and ensures deep penetration into tissues; Second, the
emission of Sm3+ falls beyond the range of common autofluorescence band, which are more
frequent than green, yellow and red. The blue emission is well separated spectrally with
respect to red excitation and that facilitates both imaging and the development for specialisted
systems such as that the readout of bioassays. The third, the excitation range where the
upconversion takes place is very broad. Thus authors suggested that Gd2O3:Sm3+
nanopowders systems have attractive potential for future application in fluorescent labeling
and probe.
Except for RE ions, the UCL properties of some transition metal ions also were studied
by NIR fs laser irradiation. Hirao group reported the UCL of Ta5+doped silicate (10CaO-
20Na2O2-70SiO2) glass under fs laser irradiation and observed a novel phenomenon [89].
Figure 22 is the emission spectra of Ta5+ in 10CaO-20Na2O2-70SiO2 glass at 780 nm fs laser
and 245 nm excitation (a), log-log plot between the UCL intensity and the incident power (b),
and UCL mechanism (c).
Figure 22. Emission spectra of Ta5+ in 10CaO-20Na2O2-70SiO2 glass at 780 nm fs laser and 245 nm
excitation (a), and log-log plot between the UCL intensity and the incident power (b). (From Ref. 89).
It can be seen that the broad band emission centered at 420 (blue emission) is observed
and the spectrum configuration is similar at 245 and 780 nm excitation. In Figure 22 (b), the
slope is equal to 2, indicating that the two-photon absorption is responsible for the blue
emission under 780 nm fs laser irradiation. In (c), the UCL process is proposed. An electron
can be promoted to 5d0 level of Ta5+ by absorbing two photons and then relaxes to the lowest
vibrational level of 5d0 of Ta5+. Finally, the recombination of the electron and hole causes the
UCL of Ta5+. This result facilitates understanding the mechanisms of UCL of Ta-doped
glasses and demonstrates the great potential of transition metal ions with the absence of d
electrons in the fields of optics.
For UCL of Cr3+, the previous reports were concentrated on the energy transfer from RE
ions to Cr3+. Recently, 2E2-4A2 luminescence of Cr3+ in Al2O3 crystals by direct pumping Cr3+
using NIR fs laser was reported by Yang et al. [90]. Figure 23 is the emission spectra of Cr3+
in Al2O3 crystals (a), the log-log relation between the fluorescent intensity of Cr3+ and pump
power (b), and the absorption spectra (c). It is obvious that the configuration of emission
spectra at 400 and 800 nm excitation is same, indicating that the UCL of Cr3+ under fs laser
irradiation can be also attributed to the 2E-4A2 transitions of Cr3+. According to Figure 23 (b),
the slope wis determined to be 2.05 by linear fitting, implying that UCL of Cr3+ ions is a two-
photon excitation. Based on the absorption properties (Figure 23 (c)) and UCL analysis,
authors presented that the mechanism of UCL of Cr3+ in Al2O3 crystals is a two-photon
simultaneous absorption.
Figure 23. Emission spectra of Cr3+-doped Al2O3 crystals under 800 nm fs laser and 400 nm light from
Xe lamp(a), the log-log relation between the fluorescent intensity of cr3+ and pump power (b), and the
absorption spectra (c). (From Ref. 90).
2. The Changes of Optical Properties of RE Ions Induced by Fs Laser
Active research has been carried out in the optical properties of RE ions induced by fs
laser. With fs laser, site-selective microfabrication can be achieved. Fs pulses irradiated
mixture of CeO2 and Tb4O7 under ambient conditions[91]. The results demonstrated that solid
solution CeTbO3+δ with cubic fluorite structure has been synthesized on the irradiated target
surface. SEM micrographs showed that the ultra-short laser irradiation resulted in the
formation of foamy structure and spherical particles with size varying from about 30 to
200 nm. In the vis VUV regions, the lanthanide and actinide ions have CT bands, in which
charge separation can be induced by excitation. Both Sm3+ and Eu3+ have CT excitation,
which caused them to be induced by fs irradiation. In 1994, Nakashima reported that two-
photon reduction of Eu3+ to Eu2+ via f’´←f transitions in methanol [92]. This is the first
reported the photoinduction of Eu3+ to Eu2+ in solution. The excitation spectra evidently
indicate that the photoinduction takes place via a 4f´←4f transiton. The second photon is
presumed to be absorbed in a transition band of an intermediate level to a CT level. The
reduction efficiency was about 3x10-5 for the excitation at 464.7 nm under 460 mJ/cm2 from
dye laser. In 1999, this group reported the three-photon reduction of Eu3+ to Eu2+ in methanol
using the second harmonic Ti:sapphire fs laser (394 nm, 2ps) [93]. The reaction scheme is
thought to be as follows:
hν hν hν
Eu 3+ X ⎯⎯→ ( Eu 3+ ) * X ⎯⎯→ ⎯⎯→ ( Eu 2+ X + ) * ⎯
⎯→
0
Eu 2+ + X +
hν
Eu 2+ Y ⎯⎯→ ( Eu 2+ Y ) * ⎯
⎯→
0
Eu 3+ + Y
The first equation is reduction and the second one is oxidation process. The first photon
at 394.3 nm pumps the f-electronic transition of 5L6←7F0, and the second and the third photon
pump into the CT states. X can be Cl-, methanol, or H2O. the final CT state efficiently
dissociate to Eu2+. It is interesting and important that the reduction efficiency of Eu3+ to Eu2+
is increased to be 0.3 under fs laser irradiation in comparison with the order of 10-5 for a
nanosecond dye laser pulse. The existence of back-reaction (oxidation) is introduced to
explain the saturation behavior of the photoproduct formation of Eu2+. Y can be methanol or
H2O. In 1999, Qiu et al. reported that permanent reduction of Eu3+ to Eu2+ in a
fluorozirconate glass using NIR fs laser [94]. In fs irradiated glass containing Eu3+, ESR
signals and the absorption of Eu2+ were observed, while no ESR signals and Eu2+ absorption
in non-irradiated glass. Because Eu3+-doped fluorozirconate glass has no absorption at 800
nm, photoreduction of Eu3+ to Eu2+ should be a nonlinear optical process. Authors suggested
the photoreduction mechanisms as follows: Active electrons and holes are generated in the
glass through a multiphoton ionization process [95]. Holes are trapped in the active sites in
the glass matrix, and some electrons are trapped by Eu3+, leading to the formation of V-type
hole-trapped defect centers and Eu2+. In addition, the trap levels of defect centers can be deep,
resulting in the stable Eu2+ in room temperature. The similar phenomenon of permanent
photoreduction of Eu3+ to Eu2+ in sodium borate glasses was also observed [96-98]. Except
for Eu3+ ions, trivalent Sm3+ ions also have CT transition, leading them to be reduced by laser
irradiation. In 1999, Qiu et al. reported the observation of permanent photo reduction of Sm3+
to Sm2+ inside a transparent and colorless Sm3+ -doped sodium aluminoborate glass [99].
After irradiation by an 800 nm focused fs pulsed laser, the focused part of the laser in the
glass became orange. Absorption and luminescence spectra show that a part of Sm3+ ions
were reduced to Sm2+ ions due to multiphoton reduction induced by fs laser. In irradiated
glass, ESR signals of Sm2+ were observed. The possible photoreduction mechanisms which
were similar with the mechanisms of photoreducton of Eu3+ to Eu2+ were presented. Active
electrons and holes ae created in the glass through multiphoton, Joule heating, and collisional
ionization processes due to fs laser irradiation. Holes are trapped by nonbridging oxygen ions
as well as by tetrahedral coordinated boron atoms, while some of electrons were trapped by
Sm3+, leading to the formation of Sm2+. The trap level of defect centers may be deep,
resulting in stable Sm2+ ions at room temperature. They thought that this phenomenon was
inferred to be useful for the fabrication of optical memory devices with altrahigh storage
density. After then on, several groups reported the photonreduction of Sm3+ to Sm2+ in
different transparent glass matrixes under fs laser irradiation, such as Al2O3-SiO2 sol-gel glass
[100], 95(0.85B2O3-0.15Na2O)-5Sm2O3 melting glass[101], ZnCl2 based glass [102] and
Al2O3-TiO2-SiO2 sol-gel glass[103]. It should be noted that the photoreduction of Sm3+ to
Sm2+ by fs laser is more stable than by X-ray [103].
Interaction between RE ions and fs laser can cause the structural and spectral changes
except for photoreduction of RE ions. Qiu et al reported the observation of memorized
polarization-dependent light scattering in a Eu2+-doped fluoroaluminate glass [104].
Anisotropic light scattering along the plane of light polarization was observed after the
excitation of 800-nm, 150 fs laser with a repetition rate of 200 KHz. When the direction of
the light polarization was changed and the same location was irradiated, an anisotropic light
scattering pattern identical to the original one at the beginning was observed, then a new
pattern along the new direction of the light polarization was observed while the original light-
scattering pattern disappeared gradually, which was attributed to the light scattering of the
polarization-dependent permanent microstructure induced by polarized ultrashort pulsed laser
itself. They also found that long-lasting phosphorescence in 40CaO-30Al2O3-30SiO2 doped
with Ce3+, Tb3+, Pr3+ glass induced by the irradiation of fs laser [105]. After irradiation by an
800 nm fs pulsed laser, the focused part of the laser in the glasses emits bright and long-
lasting phosphorescence able to be clearly seen with the naked eye in the dark even one hour
after removal of the activating laser. Moreover, by selecting appreciate glass compositions
and RE ions, optical three-dimensional image patterns emitting long-lasting phosphorescence
in various colors, including blue, green, and red. Only in the Ce3+-,Tb3+- and Pr3+- doped
glasses, a long-lasting phosphorescence was observed, while other RE ions doped glasses, no
long-lasting phosphorescence was observed. This origin can be discussed as follows: Ce3+,
Tb3+ and Pr3+ ions are easily oxidized to be Ln4+. Since all of glasses are prepared under an
ambient atmosphere, Ce, Tb and Pr may have already existed as mixed states of Ln3+
(Ln=RE) and Ln4+. Ln3+ can possibly act as hole trapping enters, while Ln4+ is known to act
as electron trapping centers. The long-lasting phosphorescence of above three valence RE
ions is considered to be due to the thermal simulated recombination of holes and electrons at
traps induced by fs laser irradiation, which leave holes or electrons in a metastable excited
state at room temperature. They also suggested that it may be possible to selectively induce
defects by using laser with different wavelengths, pulse widths, and repetition rates, leading
to photoluminescence in different colors. This will be vital important in the fabrication of
rewritable three-dimensional optical memory devices with an ultrahigh recording speed and
an ultrahigh storage density [106]. The present group also observed the similar phenomenon
in Eu2+-doped CaO-Al2O3-SiO2 melting glasses [107]. They found long lasting
phosphorescence in Eu2+-doped calcium aluminosilicate glasses induced by fs infrared laser.
It suggested that the long lasting phosphorescence arises from the thermal stimulated
recombination of holes and electrons at traps induced by the laser irradiation, which leave
holes or electrons in a metastable excited state at room temperature.
PSHB materials have been intensively studied due to their potential applications in high-
density frequency-domain optical storage and signal processing [108]. For practical usages,
materials exhibiting high temperature and stable PSHB are desirable. The room temperature
PSHB in Sm2+- and Eu3+-doped glasses has been reported [109-112]. Moreover, hole-burning
efficiency is strongly dependent on the local environments around RE ions. Therefore, in this
review, we mainly concentrat on the PSHB changes induced by fs laser. In 2001, Fujita
reported that as-prepared glasses with the composition of 35Na2O-65B2O3-1Eu2O3 melted in
air, do not show PSHB even at liquid nitrogen temperature (~77 K), but room temperature
PSHB occurs in the fs laser irradiated glasses [113]. The reduction of Eu3+ to Eu2+ and the
generation of intrinsic defects were observed in the glasses irradiated by fs laser pulses. They
suggested that the phenomenon of PSHB at room temperature is likely to arise from
photoinduced redistributing of electric charges between Eu3+ and Eu2+. After then on, Nogmi
observed the depth and width of burned holes in Eu3+ doped aluminosilicate sol-gel glasses
irradiated by fs laser reached to 15% and ~2.5 cm-1 fwhm at 7 K [114], respectively, while in
non-irradiated glasses, the values of depth and width were 6% and 3 cm-1 fwhm. This result
indicates that fs laser treatment can increase the hole-burning efficiency. Moreover, in the
glass without fs laser irradiation, the hole depth decreases when the temperature increases and
it disappears about 200 K. On the other hand, in the irradiated glass, the hole depth also
decreases and hole width increases as the temperature increases. It should be noted that the
hole still remains with 5% of its original depth even room temperature. The increase of
burning-hole efficiency in irradiated glasses is due to the rearrangement of OH bonds
surrounding the Eu3+ ions. In fact, the PSHB of Sm2+ induced by fs laser also reported.
Similar with Eu3+, the burning hole efficiency and temperature stability of Sm2+ in fs laser
irradiated glasses increase, such as Sm2+-doped borate [115], fluoride [116], aluminosilicate
glasses[117]. It is important that hole-burning rate is faster than the H2-treated glasses and X-
ray irradiated glasses. For the origins of enhancement of hole-burning efficiency of RE
induced by fs laser need be further studied.
Nogami reported the precipitation of SnO2 nanocrystals in in Eu3+-doped SnO2-SiO2 sol-
gel glasses [118, 119]. SnO2 nanoparticles with an average size of 5 nm were precipitated in
silicate sol-gel glasses by an irradiation with an 800 nm fs laser pulse. The mechanisms of
formation of SnO2 nanocrystals were discussed as follows. In glass network, the Sn ions are
coordinated by two oxygen ions with point defect of a molecularlike electronic structures.
Under the irradiation, the two fold-coordinated Sn atoms are activated to react with oxygen,
leading to the formation of SnO2 nanocrystals. It is important that the luminescence of Eu3+
ions are 100-time stronger than the glass without nanocrystals corresponding to the excitation
of the absorption edge of SnO2 nanocrystals. The enhancement of Eu3+ ions is caused by
energy transfer from SnO2 nanocrystals to Eu3+.
It can be seen that fs laser is a powerful tool to modulate the optical properties of RE
ions, This provides a bright prospect method to obtain high-densith memory devices and
materials.
ACKNOWLEDGEMENT
The authors gratefully thank the financial supports from Nation Natural Science
Foundation of China (Grant No. 10504030).
REFERENCES
[1] G. K. Liu, B. Jacquier, Spectroscopic properties of rare earths in optical materials.
Springer, p8.
[2] F. Auzel, Upconversion and Anti-Stokes Processes with f and d Ions in Solids. Chem.
Rev. 2004, 104, 139-173.
[3] Y. Wei, F. Q. Lu, X. R. Zhang, D. P. Chen, Polyol-mediated synthesis and
luminescence of lanthanide-doped NaYF4 nanocrystal upconversion phosphors. J.
Alloys Comp. 2008, 455, 376-384.
[4] G. Y. Chen, H. C. Liu, G. Somesfalean, L. H. Liang, Z. G. Zhang, Q. Sun, F. P. Wang.
Enhancement of the upconversion radiation in Y2O3:Er3+ nanocrystals by codoping with
Li+ ions. Appl. Phys. Lett. 2008, 92, 113114.
[5] G. Yi, G. Chow, Colloidal LaF3:Yb,Er, LaF3:Yb,Ho and LaF3:Yb,Tm nanocrystals
with multicolor upconversion fluorescence. J. mater. Chem. 2005, 15, 4460-4464.
[6] J. Shan, X. Qin, N. Yao, Y. Ju. Synthesis of monodisperse hexagonal NaYF4:Yb, Ln
(Ln = Er, Ho and Tm) upconversion nanocrystals in TOPO. Nanotech. 2007, 18,
445607.
[7] K. Hirao, M. Higuchi, N. Soga, Infrared to blue up-conversion in Pr3+ doped YAG and
LiYF4 crystals. J. Lumin. 1994, 60/61, 179-182.
[8] P. J. Deren, R. Mahiou, W. Strek, A. Bednarkiewicz, G. Bertrand, Up-conversion in
KYb(WO4)2:Pr3+ crystal. Opt. Mater. 2002, 19, 145-148.
[9] R. Balda, M. Sanz, J. Fernandez, J. M. Fdez-nvarro, Energy transfer and upconversion

processes in Nd3+-doped GeO2-PbO-Nb2O5 glass. J. Opt. Soc. Am. B 2000, 17, 1671-77.
[10] J. Fernadez, R. Balda, A. Mendorioz, M. Sanz, J. L. Adam, Upconversion processes in
Nd3+-doped fluorochloride glass. J. Non-Cryst. Solids 2001, 287, 437-443.
[11] J. Fernadez, R. Balda, M. Sanz, L. M. Lacha, A. Oleaga, J. L. Adam, Upconversion
losses in Nd-doped fluoroarsenate glasses. J. Lumin. 2001, 94/95, 325-329.
[12] L. H. Acioli, J. T. Guo, C. B. de Araugo, T. Messaddeq, M. A. Aegerter, Violet-blue
emissions due to frequency up-conversion in Nd3+-doped fluoroindate glasses. J.
Lumin. 1997, 72/74, 68-70.
[13] R. Balda, M. Sanz, A. Mendorioz, J. Fernandez, L. Griscom, J. L. Adam, Infrared-to-
visible upconversion in Nd3+-doped chalcohalide glasses. Phys. Rev. B 2001, 64,
144101.
[14] M. Wermuth, T. Riedener, H. U. Gudel, Spectroscopy and upconversion mechanisms of
CsCdBr3:Dy3+. Phys. Rev. B 1998, 57, 4369-4376.
[15] M. Pollnau, E. Heumann, G. Huber, Time-resolved spectra of excited-state absorption
in Er3+ doped YAlO3. Appl. Phys. A 1992, 54, 404-410.
[16] S. R. Luthi, M. Pollnau, H. Gudel, M. P. Hehlen, Near-infrared to visible upconversion

in Er3+-doped Cs3Lu2Cl9, Cs3Lu2Br9, and Cs3Y2I9 excited at 1.54 μm. Phys. Rev. B
1999, 60, 162-178.
[17] M. P. Hehln, K. Kramer, H. U. Gudel, A. V. Butashin, A. Kaminskii, Upconversion in
Er3+-dimer systems: Trends within the series Cs3Er2X9 (X=Cl,Br,I). Phys. Rev. B 1994,
49, 12475-12484.
[18] O. Wenger, D. R. Gamelin, H. U. Gudel, A. V. Butashin, A. Kaminskii, Site-selective
optical spectroscopy and upconversion mechanisms of the laser material BaLu2F8:Er3+.
Phys. Rev. B 1999, 60, 5312-20.
[19] N. J. Cokcroft, K. M. Murdoch, Energy transfer between thulium and praseodymium
ions in solids. J. Lumin. 1994, 60/61, 891-894.
[20] X. B. Chen, G. Y. Zhang, Y. H. Mao, Y. B. Hou, Y. Feng, Z. Hao, Research on the up-
conversion luminescence of Tm3+ ion in crystal and amorphous pentaphosphate
materials. J. Lumin. 1996, 69, 151-160.
[21] H. W. Song, L. X. Yu, L. M. Yang, and S. Z. Lu, Luminescent properties of rare earth
ions in one-dimensional oxide nanowires. J. Nanosci. Nanotech. 2005, 5, 1519-1531.
[22] Y. Q. Lei, H. W. Song, L. M. Yang, L. X. Yu, Z. X. Liu, G. H. Pan, X. Bai, and L. B.
Fan, Upconversion luminescence, intensity saturation effect, and thermal effect in
Gd2O3:Er3+,Yb3+ nanowires. J. Chem. Phys. 2005, 123, 174710.
[23] F. Y. Konz, C. W. Sun, R. L. Thiel, R. W. Cone, R. L. Equall, Hutcheson, R. M.
Macfarlane, Temperature and concentration dependence of optical dephasing. spectral-
hole lifetime, and anisotropic absorption in Eu3+:Y2SiO5. Phys. Rev. B 2003, 68,
085109.
[24] R. Yano, M. Mtsunaga, N. Uesugi, Stimulated-photon-echo spectroscopy. I. Spectral
diffusion in Eu3+:YAlO3. Phys. Rev. B 1992, 45,12752-12759.
[25] J. Dexpert-Ghys, R. Mauricot, M.D. Faucher, Spectroscopy of Eu3+ ions in monazite
type lanthanide orthophosphates LnPO4, Ln = La or Eu. J. Lumin. 1996, 69, 203-215.
[26] J.Holsa, M. Leskela, L. Niinisto, Sensitization of Tb3+ luminescence with Ce3+ in
LaOBr: Tb3+, Ce3+. J. Solid State Chem. 1981, 37, 267-270.

[27] D. Jia, W. Jia, X. Wang, W. Yen, Quenching of thermo-stimulated photo-ionization by
energy transfer in CaAl4O7: Tb3+, Ce3+. Solid State Commu. 2004, 129, 1-4.
[28] M. Jose, Ce3+ to Tb3+ energy transfer in alkaline earth (Ba, Sr or Ca) sulphate
phosphors. Opt. Mater. 2004, 24, 651-659.
[29] L. X. Yu, H. W. Song, Z. X. Liu, L. M. Yang, S. Z. Lu, Z. H. Zheng, Electronic
transition and energy transfer processes in LaPO4: Ce3+/Tb3+ nanowires. J. Phys. Chem.
B 2005, 109, 11450-11455.
[30] L. X. Yu, H. W. Song, Z. X. Liu, L. M. Yang, S. Z. Lu, Fabrication and
photoluminescent characteristics of La2O3: Eu3+nanowires. Physical Chemistry
Chemical Physicas, 2006, 8, 303-308.
[31] L. X. Yu, H. W. Song, S. Z. Lu, Z. X. Liu, L. M. Yang, X. G. Kong, Luminescent
properties of LaPO4: Eu3+ nanoparticles and nanowires. J. Phys. Chem. B 2004, 108,
16697-16702
[32] N. Salaha, P. D. Saharea, S. P. Lochabb, P. Kumarc, TL and PL studies on CaSO4: Dy
nanoparticles. Radia. Meas. 2006, 41, 40-47.
[33] W. Kaiser, C. G. B. Garrett, Two-photon excitation in CaF2:Eu2+. Phys. Rev. Lett. 1961,
7, 229-231.
[34] N. Takeshima, Y. Narita, T. Nagata, S. Tanaka, Fabrication of photonic crystals in ZnS-

doped glass. Opt. Lett. 2005, 30, 537-539.
[35] J.P. Sylvestre, A. V. Kabashin, E. Sacher, M. meunier, J. H. T. Luong, Stabilization and
size control of gold nanoparticles during laser ablation in aqueous cyclodextrins. J. Am.
Chem. Soc. 2004, 126, 7176-7177.
[36] J. Qiu, X. Jiang, C. Zhu, M. Shirai, J. Si, N. Jiang and K. Hirao, Manipulation of gold
nanoparticles inside transparent materials. Angew. Chem. Inter. Ed. 2004, 116, 2280-
2284.
[37] J. Qiu, M. Shirai, T. Nakaya, J. Si, X. Jiang, C. Zhu, K. Hirao, Space-selective
precipitation of metal nanoparticles inside glasses. Appl. Phys. Lett. 2001, 81, 3040-
3042.
[38] R. An, J. D. Uram, E. C. Yusko, K. Ke, M. Mayer, A. J. Hunt, Ultrafast laser
fabrication of submicrometer pores in borosilicate glass. Opt. Lett. 2008, 33, 1153-
1155.
[39] P. Galinetto, D. Grando, M. C. Mozzati, D. Ballarini, E. Giulotto, Structural and optical
defects in LiNbO3 crystals induced by femtosecond laser irradiation. Phys. Stat. Sol.
2007, 4, 1329-1332.
[40] N. Uchida, Y. Mikami, H. Kintoh, K. Murakami, N. Fukata, M. Mitome, M. Hase, M.
Kitajima, Site-selective formation of Si nanocrystals in SiO2 by femtosecond laser
irradiation and Al deoxidization effects. Appl. Phys. Lett. 2008, 92, 153112.
[41] H. L. Ma, J. Y. Yang, Y. Dai, Y. B. Zhang, B. Lu, G. H. Ma, Raman study of phase
transformation of TiO2 rutile single crystals irradiated by infrared femtosecond laser.
Appl. Surf. Sci. 2007, 253, 7494-7500.
[42] M. Simurda, P. Nemec, F. Trojanek, P. Maly, Substantial enhancement of
photoluminescence in CdSe, nanocrystals by femtosecond pulse illumination. Solid
state Films, 2004, 453-454, 300-303.
[43] X. D. Guo, R. X. L, Y. Hang, Z. Z. Xu, B. K. Yu, H. L. Ma, X. W. Sun, Raman
spectroscopy and luminescent properties of ZnO nanostructures fabricated by
femtosecond laser pulses. Mater. Lett. 2007, 61, 4583-4586.

[44] H. Sun, Y. Xu, S. Juodkazis, K. Sun, M. Watanabe, S. Matsuo, H. Misawa, Arbitrary-
lattice photonic crystals created bu multiphoton microfabrication. Opt. Lett. 2001, 26,
325-327.
[45] M. Straub, M. Gu, Near-infrared photonic crystals with high-order bandges generated
by two-photon polymerization. Opt. Lett. 2002, 27, 1824-1826.
[46] R. R. Thomson, S. Campbell, I. J. Blewett, A. K. Kar, D. T. Reid, S. Shen, A. Jha,
Active waveguide fabrication in erbium-doped oxyfluoride silicate glass using
femtosecond pulses. Appl. Phys. Lett. 2005, 87, 121102.
[47] N. D. Psaila, R. R. Thomson, H. T. Bookey, A. K. Kar, N. Chiodo, R. Osellame, G.
Cerullo, A. Jha, S. Shen, Er:Yb-doped oxyfluoride silicate glass waveguide amplifier
fabricated using femtosecond laser inscription. Appl. Phys. Lett. 2007, 90, 131102
[48] J. R. Lakowicz, C. Piszczek, B. P. Maliwal, I. Gryczynski, Multiphoton excitation of
lanthanides, ChemPhysChem. 2001, 2, 247-252.
[49] H. P. You, M. Nogami, Three-photon-excited fluorescence of Al2O3-SiO2 glass
containing Eu3+ ions by femtosecond laser irradiation. Appl. Phys. Lett. 2004, 84, 2076-
2078.
[50] S. Q. Xu, W. wang, S. F. Zhu, B. Zhu, J. R. Qiu, Highly efficient red, green and blue
upconversion luminescence of Eu3+/Tb3+-codoped silicate by femtosecond laser
irradiation. Chem. Phys. Lett. 2007, 442, 492-495.
[51] Q. Lu, A. H. Li, F. Y. Guo, L. Sun and L. C. Zhao, The two-photon excitation of SiO2-
coated Y2O3:Eu3+ nanoparticles by a near-infrared femtosecond laser, Nnaotech. 2008,
19, 1-8.
[52] M. Mortier, A. Monteville, G. patriarche, G. Maze, F. Auzel, New progresses in
transparent rare-earth doped glass-ceramics. Opt. mater. 2001, 16, 255-267.
[53] M. Mortiery, Between glass and crystal: glass-ceramics, a new way for optical
materials. Philosophi. Mag. B 2002, 82, 745-753.
[54] F. Lahoz, I. R. Martin, U. R. Rodriguez-Mendoza, I. Iparraguirre, J. Azkargorta, A.
Mendioroz, R. Balda, J. Fernandez, V. Lavin, Rare earths in nanocrystalline glass–
ceramics. Opt. Mater. 2005, 27, 1762-1770.
[55] B. Zhu, S. M. Zhang, S. F. Zhou, N. Jiang, J. R. Qiu, Enhanced upconversion
luminescence of transparent Eu3+doped glass-ceramics containing nonlinear optical
microcrystals. Opt. Lett. 2007, 32, 653-655.
[56] L. X. Yu, M. Nogami, Upconversion luminescence properties of europium in ZnO–
SiO2 glasses by femtosecond laser excitation. Mater. Chem. Phys. 2008, 107, 186–188.
[57] Y. B. Qiao, D. P. Chen, J. J. Ren, B. T. Wu, J. R. Qiu, T. Akai, Blue emission from
Eu2+-doped high silica glass by near-infrared femtosecond laser irradiation. J. Appl.
Phys. 2008, 103, 023108.
[58] M. H. V. Werts, N. Nerambourg, D. Pelegry, Y. L. Grand, M. Blanchard-Desce, Action
cross section of two-photon excited luminescence of some Eu (Ш) and Tb (Ш)
complexes. Photochem. Photobiol. Sci. 2005, 4, 531-538.
[59] L. M. Fu, X. F. Wen, X. C. Ai, Y. Sun, Y. S. Wu, J. P. Zhang, Y. Wang, Efficient two-
photon-sensitized luminescence of a europium (Ш) complex. Angew. Chem. Int. Ed.
2005, 44, 747-750.
[60] M. A. Noginov, P. Venkateswarlu, M. Mahdi, Two-step upconversion luminescence in
Yb:Tb:YSGG crystals. J. Opt. Soc. Am. B 1996, 13, 735-741.

[61] N. K. Giri, D. K. Rai, S. B. Rai, UV-visible emission in Tb-Yb codoped tellurite glass
on 980-nm excitation. Appl. Phys. B 2007, 89, 345-348.
[62] M. V. D. Vermelho, P. V. dos santos, M. T. de Araujo, A. S. Gouveia-Neto, F. C.
Cassanjes, S. J. L. Ribeiro, Y. Messaddeq, Thermally enhanced cooperative energy-
transfer frequency upconversion in terbium and ytterbium doped tellurite glass. J.
Lumin. 2003, 102-103, 262-267.
[63] G. M.Salley, R. Valiente, H. U. Gudel, Photon-assisted cooperative sensitization of
Tb3+ in SrCl2:Yb, Tb. J. Phys. : Condens. Matter. 2002, 14, 5461-5475.
[64] I. R. Martin,A. C. Yanes, J. Mendez-Ramos, M. E. Torres, V. D. Rodriguez,
Cooperative energy transfer in Yb3+-Tb3+ codoped silica sol-gel glasses. J. Appl. Phys.
2001, 89, 2520-2524.
[65] G. M. Salley, R. Valiente, H. U. guedel, Luminescence upconversion mechanisms in
Yb3+-Tb3+ systems. J. lumin. 2001, 94-95, 305-309. M. A. Noginov, P. venkateswarlu,
M. Mahdi, Two-step upconversion luminescence in Yb:Tb:YSGG crystal, J. Opt. Soc.
Am. B 1996, 13, 735-741.
[66] E. Martins, C. B. de Araujo, C. E. Egalon, A. S. Gomes, B. J. Costa, Y. Messaddeg,

Cooperative frequency upconversion in Yb3+–Tb3+ codoped fluoroindate glass Opt.
Comm. 1994, 158, 61-64.
[67] L. de S. Menezes, G. S. Maciel, Cid B. de Araujo, Y. Messaddeq, Phonon-assisted
cccperative energy transfer and frequency upconversion in a Yb3+-Tb3+ codoped
fluoroindate glass. J. Appl. Phys. 2003, 94, 863-866.
[68] A. S. Gouveia-Neto, L. A. Bueno, A. C. M. Afonso, J. F. Nascimento, E. B. Costa, Y.
Messaddeq, S. J. L. Ribeiro, Upconversion luminescence in Ho3+/Yb3+- and Tb3+/Yb3+-
codoped fluorogermanate glass and glass ceramic. J. Non-Cryst. Solids 2008, 354, 509–
514.
[69] Y. N. Forrester, R. S. Pandher, B. R. Reddy, S. K. Nash-Stevenson, Stepwise two-
photon excitation studies in terbium-doped fluoride crystals. Opt. Mater. 2002, 19, 255-
258.
[70] C. H. Yang, Y. X. Pan, Q. Y. Zhang, Z. H. Jiang, Cooperative energy transfer and
frequency upconversion in Yb3+-Tb3+ and Nd3+- Yb3+-Tb3+ codoped GdAl3(BO3)4
phosphors. J. fluores. 2007, 17, 500-504
[71] G. F. white, K. L. Litvinenko, S. R. Meech, D. L. Andrews, A. J. Thomson,
Multiphoton-excited luminescence of a lanthanide ion in a protein complex:Tb3+bound
to transfer. Photochem. Photobio. Sci. 2004, 3, 47-55
[72] S. M. Zhang, B. Zhu, S. F. Zhou, S. Q. Xu, J. R. Qiu, Multi-photon absorption
upconversion luminescence of a Tb3+-doped glass excited by an infrared femtosecond
laser. Opt. Expre. 2007, 15, 6883-6888.
[73] A. H. Li, Z. R. Zheng, Q. Lu, L. Sun, W. Z. Wu, W. L. Liu, Y. Q. Yang, T. Q. Lu,
Two-photon-excited luminescence in a Tb3+-doped lithium niobate crystal pumped by a
near-infrared femtosecond laser. Opt. Lett. 2008, 33, 1014-1016.
[74] W. Ryba-Romanowski, S. Golab, G. Dominiak-Dzik, M. N. Palatnikov, N. V. Sidorov,
Influence of temperature on luminescence of terbium ions in LiNbO3. Appl. Phys. Lett.
2001, 78, 3610.
[75] H. P. You, T. Hayakawa, M. Nogami, Upconversion luminescence of Al2O3–SiO2:Ce3+

glass by femtosecond laser irradiation. Appl. Phys. Lett. 2004, 85, 3432-3434.
[76] [76] Q. Z. Zhao, J. R. Qiu, C. J. Zhao, X. W. Jiang, C. S. Zhu, G. J. Zhao, Investigation
of optical properties of Ce3+-doped Gd2SiO5 crystal irradiated by a femtosecond laser.
Opt. Commu. 2005, 255, 97-101.
[77] Y. J. Dong, J. Xu, G. Q. Zhou, G. J. Zhao, M. Y. Jie, L. Y. Yang, L. B. Su, J. R. Qiu,
W. W. Feng, L. H. Lin, Blue upconversion luminescence generation in Ce3+:Gd2SiO5
crystals by infrared femtosecond laser irradiation. Opt. Expre. 2006, 14, 1899-1904.
[78] Y. J. Dong, J. Xu, G. J. Zhao, C. F. Yan, G. Q. Zhou, L. B. Su, L. Y. Yang, J. R. Qiu, L.
H. Lin, X. Y. Liang, R. X. Li, Z. Z. Xu, Simultaneous three-photon-excited violet
upconversion luminescence of Ce3+:Lu2Si2O7 single crystals by femtosecond laser
irradiation. Opt. Lett. 2006, 31, 2175-2177.
[79] C. Wang, M. Y. Peng, L. Y. Yang, X. Hu, N. Da, D. P. Chen, C. S. Zhu, J. R. Qiu,
Upconversion luminescence of Ce3+ doped BK7 glass by femtosecond laser irradiation.
J. Rare Earths.2006, 24, 754-756.
[80] L. Y. Yang, Y. J. Dong, D. P. Chen, C. Wang, X. Hu, N. Da, G. J. Zhao, J. Xu, X. W.
Jiang, C. S. Zhu, J. R. Qiu, Three-photon-excited upconversion luminescence of
Ce3+:YAP crystal by femtosecond laser irradiation. Opt. Expre. 2006, 14, 243-247.
[81] A. H. Li, Q. Lu, Z. R. Zheng, L. Sun, W. Z. Wu, W. L. Liu, Y. Q. Yang, T. Q. Lu,

Yellow-green upconversion luminescence of Dy3+ ion in LiNbO3 crystal heavily
codoped with ZnO. J. Appl. Phys. 2007, 102, 113102.
[82] A. H. Li, Q. Lu, Z. R. Zheng, L. Sun, W. Z. Wu, W. L. Liu, H. Z. Chen, Y. Q. Yang, T.
Q. Lu, Enhanced green upconversion emission of Er3+ through energy transfer by Dy3+
under 800 nm femtosecond-laser excitation. Opt. Lett. 2008, 33, 693-695.
[83] A. H. Li, L. Sun, Z. R. Zheng, Q. Lu, W. Z. Wu, W. L. Liu, Y. Q. Yang, T. Q. Lu,
Spectroscopic properties of Er3+ ions in LiNbO3 crystals codoped with HfO2. Appl.
Phys. B 2008, 90, 29-34.
[84] L. Y. Yang, C. Wang, Y. J. Dong, N. Da, X. Hu, D. P. Chen, J. R. Qiu, Three-photon-
excited upconversion luminescence of YVO4 single crystal by infrared femtosecond
laser irradiation. Opt. Expre. 2005. 13, 10157-10162.
[85] X. S. Wang, J. Song, H. Y. Sun, Z. Z. Xu, J. R. Qiu, Multiphoton-excited upconversion
luminescence of Nd:YVO4. Opt. Expre. 2007, 15, 1384-1389.
[86] X. S. Wang, J. R. Qiu, J. Song, J. Xu, Y. Liao, H. Y. Sun, Y. Cheng, Z. Z Xu,
Simultaneous three-photon absorption induced ultraviolet upconversion in Pr3+:Y2SiO5
crystal by femtosecond laser irradiation. Opt. Commu. 2008, 281, 299-302.
[87] B. Zhu, S. M. Zhang, G. Lin, S. F. Zhou, J. R. Qiu, Enhanced multiphoton absorption
induced luminescence in transparent Sm3+-doped Ba2TiSi2O8 glass-ceramics. J. Phys.
Chem. C 2007, 111, 17118-17121.
[88] D. Dosev, I. M. Kennedy, M. Godlewski, I. Gryczynski, T. Tomsia, E. M. Golds,
Fluorescence upconversion in Sm-doped Gd2O3. Appl. Phys. Lett. 2006, 88, 011906.
[89] X. G. Meng, K. Tanaka, S. Murai, K. Fujita, K. Miura, K. Hirao, Two-photon-excited
fluorescence from silicate glass containing tantalum ions pumped by a near-infrared
femtosecond pulsed laser. Opt. Lett. 2006, 31, 2867-2869.
[90] L. Y. Yang, Y. J. Dong, D. P. Chen, C. Wang, N. dam, X. W. Jiang, C. H. Zhu, J. R.
Qiu, Upconversion luminescence from 2E state of Cr3+ in Al2O3 crystal by infrared
femtosecond laser irradiation. Opt. Expre. 2005, 13, 7893-7898.
[91] W. W. An, J. P. Miao, Z. G. Zhang, Z. Lü, W. H. Su, A. A. Sjögren, C. G. Wahlström,

S. Svanberg, Synthesis and surface characteristics of CeTbO3+δ induced by femtosecond
laser irradiation. Appl. Sur. Sci. 2007, 253, 3884-3887.
[92] M .Kusaba, N. Nakashima, Y. Izawa, C. Yamanaka, W. Kawamura, Two-photon
reduction of Eu3+ to Eu2+ via the f´←f transitions in methanol. Chem. Phys. Lett. 1994,
221, 407-411.
[93] N. Nakashima, S. Nakamura, S. Sakabe, H. Schillinger, Y. Hamanaka, C. Yamanaka,
M. Kusaba, N. Ishihara, Y. Izawa, Multiphoton reduction of Eu3+ to Eu2+ in methanol
using intense, short pulse from a Ti:sapphire laser. J. Phys. Chem. A 1999, 103, 3910-
3916.
[94] J. R. Qiu, K. Kojima, K. Miura, T. Mitsuyu, K. Hirao, Infrared femtosecond laser pulse-
induced permanent reduction of Eu3+ to Eu2+ in a fluorozirconate glass. Opt. Lett. 1999,
24, 786-788.
[95] B. C. Stuart, M. D. Feit, A. M. Rubenchik, B. W. Shore, M. D. Perry, Laser-induced
damage in dielectrics with nanosecond to subpicosecond pulses. Phys. Rev. Lett. 1995,
74, 2248.
[96] K. Lim, S. Lee, M. Trinh, S. Kim, M. Lee, D. S. Hamilton, G. N. Gibson, Femtosecond

laser-induced reduction in Eu3+-doped sodium borate glasses. J. Lumin. 2007, 122-123,
14-16.
[97] J. Qiu, K. Miura, K. Nouchi, T. Suzuki, Y. Kondo, T. Mitsuyu, K. Hirao, Valence
manipulation by lasers of samarium ion in micrometer-scale dimensions inside
transparent glass. Solid State Commu. 1999, 113, 341-344.
[98] H. P. You, M. Nogami, Optical properties and valence chage of europium ions in a sol-
gel Al2O3-B2O3-SiO2 glass by femtosecond laser pulses. J. Phys. Chem. B 2005, 109,
13980-13984.
[99] J. R. Qiu, K. Miura, T. Suzuki, T. Mitsuyu, Permanent photo reduction of Sm3+ to Sm2+
inside a sodium aluminoborate glass by an infrared femto second pulsed laser. Appl.
Phys. Lett. 1999, 74, 10-12.
[100] G. J. Park, T. Hayakawa, M. Nogami, Formation of Sm2+ ions in femtosecond laser
excited Ai2O3-SiO2 glasses. J. Phys.: Condens. Mater. 2003, 15, 1259-1265.
[101] S. Lee, M. Lee, K. Lim, Femtosecond laser induced PL change in Sm-doped sodium
borate glass and 3D optical memory. J. Lumin. 2007, 122-123, 990-992.
[102] J. Qiu, K. Kojima, A. Kubo, M.Yamashita, K. Hirao, Multiphoton reduction of Sm3+ to
Sm2+ in a ZnCl2 based glass. Phys. Chem. Glasses. 2000, 41, 150-152.
[103] M. Nogami, G. Kawamura, G. J. Park, H. P. You, T. Hayakawa, Effect of Al3+ and Ti4+
ions on the laser reduction of Sm3+ ion in glass. J. Lumin. 2005, 114, 178-186.
[104] J. R. Qiu, P. G. Kazanski, J. H. Si, K. Miura, T.Mitsuyu, K. Hirao, A. L. Gaeta,
Memorized polarization-dependent light scattering in rare-earth-ion-doped glass. Appl.
Phys. Lett. 2000, 77, 1940-1942.
[105] J. R. Qiu, K. Miura, H. Inouye, Y. Kondo, T. Mitsuyu, K. Hirao, Femtosecond laser-
induced three-dimensional and lng-lasting phosphorescence inside calcium
aluminosilicate glasses doped with rare earth ions. Appl. Phys. Lett. 1998, 73, 1763-
1765.
[106] P. E. Laibins, J. J. Hickman, M. S. Wrighton, G. M. Whitesides, Orthogonal self-
assembled monolayers: alkanethiols on gold and alkane carboxylic acids on alumina.

Science 1990, 245, 843.
[107] J. R. Qiu , K. Miura ,H. Inouye, S. Fujiwara, T. Mitsuyu, K. Hirao, Blue emission
induced in Eu2+ doped glasses by an infrared femtosecond laser. J. Non-Cryst. Solids
1999, 244, 185-188.
[108] W. E. Moerner, Persistent spectral hole-burning science and application. Topic in
Current Physics, V44, Springer, Berlin, 1988.
[109] K. Hirao, S. Todoroki, D. H. Cho, and N. Soga, Room-temperature persistent hole
burning of Sm2+ in oxide glasses. Opt. Lett. 1993, 18 1586-1588.
[110] K. Kurita, T. Kushida, T. Izumitani, Michimasa Matsukawa, Room-temperature
persistent spectral hole burning in Sm2+-doped fluoride glasses. Opt. Lett. 1994, 19,
314-316.
[111] M. Nogami, Y. Abe, K. Hirao, D. H. Cho, Room temperature persistent spectra hole
burning in Sm2 + -doped silicate glasses prepared by the sol-gel process. Appl. Phys.
Lett. 1995, 66, 2951-2951.
[112] M. Nogami, K. Suzuki, Fast Spectral Hole Burning in Sm2+-Doped Al2O3-SiO2 Glasses.
Adv. Mater. 2002, 14, 923-926.
[113] K. Fujita, M. Nishi, K. Hirao, Ultrashort-laser-pulse-induced persistent spectral hole

burning of Eu3+ in sodium borate glasses. Opt. Lett. 2001, 26, 1681-1683.
[114] G. park, T. Hayakawa, M. Nogami, Spectral hole-burning femtosecond laser-irradiated
Eu3+-doped aluminosilicate glasses. J. Sol-Gel Sci. Tech. 2005, 33, 47-50.
[115] K. Fujita, C. Yasumoto, K. Hirao, Photochemical reactions of samarium ions in sodium
borate glasses irradiated with near-infrared femtosecond laser pulses. J. Lumin. 2002,
98, 317-323.
[116] J. R. Qiu, K. Miura, K. Nouchi, T. Suzuki, Y. Kondo, T. Mitsuyu, K. Hirao, Valence
manipulation by lasers of samarium ion in micrometer-scale dimensions inside
transparent glass. Solid State Commu. 2000, 113, 341-344.
[117] G. Park, T. Hayakawa, M. Nogami, Spectral hole burning and fluorescence in
femtosecond laser induced Sm2+-doped glasses. J. Lumin. 2004, 103-108.
[118] M. Nogami, A. Ohno, H. P. You, Laser-induced SnO2 crystallization and fluorescence
properties in Eu3+-doped SnO2-SiO2 glasses. Phys. Rev. B 2002, 68, 104204.
[119] M. Nogami, A. Ohno, Laser precipitation of SnO2 nanocrystals in glass and energy
transferred-fluorescence of Eu3+ ions. J. Non-Cryst. Solids 2003, 330, 264-267.
Chapter 9
YTTERBIUM DOPED MATERIALS FOR FEMTOSECOND

LASERS
Pierre-Olivier Petit1, Philippe Goldner1, Bruno Viana1, Justine

Boudeile2, Frédéric Druon2, Dimitris N. Papadopoulos2,
Marc Hanna2 and Patrick Georges2
1
Laboratoire de Chimie de la Matière Condensée de Paris, CNRS
ENSCP, 11, rue Pierre and Marie Curie 75231 Paris cedex 05, France
2
Laboratoire Charles Fabry de l’Institut d’Optique, CNRS, Univ Paris-Sud,
Campus Polytechnique, RD128, 91127 Palaiseau Cedex, France
Keywords: solid state laser, structural characterization, ytterbium, near-infrared laser, optical
spectroscopy, femtosecond laser, thermo-mechanical properties.
1. INTRODUCTION
At the end of the last century, three important technological breakthroughs took place in
ultra-short laser pulse generation field. In 1985, Strickland and Mourou first demonstrated
Chirped Pulsed Amplification [1]. Then Moulton, in 1986, developed a new laser, based on
aluminium oxide doped by titanium ions (Ti3+:Al2O3), called Ti:Sapphire [2]. Finally, Spence,
Kean and Sibbert carried out the first phase mode-locking (“magic-modelocking”) in
Ti:Sapphire also called Kerr Lens Mode-locking [3]. Since this moment, all required elements
have been demonstrated to build reliable and efficient ultra-short pulses laser chain.
The following decade was the “golden age” of Ti:Sapphire. This hegemony resulted from
the combination of two extraordinary properties of this material. First, Ti3+ ion exhibits a very
broad emission band ranging from 680 nm up to 1100 nm. This spectral width could
theorically lead to 3.6 fs laser impulsions. Experimentally, 4.8 fs has been demonstrated and
this value represents the shortest pulse ever obtained in a laser material. Moreover, sapphire
thermal conductivity is very high (κundoped= 34 W.m-1.K-1). By comparison thermal
conductivity of common laser materials such as YAG (Y3Al5O12) is at least three times
316 Pierre-Olivier Petit, Philippe Goldner, Bruno Viana et al.
smaller. Such a thermal conductivity enables an efficient heat removal from the crystal under
high power pumping. For these two reasons, Ti:Sapphire became the most popular
amplification medium for tunable laser in the near infra-red and for ultra short pulses chains.
During the same period, another revolution took place in the solid state laser field with
the emergence of efficient laser diodes that became quickly available in the near infra-red.
Thanks to their high optic to electric efficiency and despites their bad spatial and spectral
quality, the idea of pumping active ions doped material became very attractive. It worth
noting that first demonstration goes back to 1964 with the diode pumping of U3+:CaF2 [4].
Rare earth ions and more particularly neodymium and later ytterbium have been widely used
as active ions in laser materials. So a new investigation field was opened for rare earth doped
laser crystals but it could not have been extended to Ti3+ ion because its absorption band is
centred in the visible domain where no high power laser diodes are yet available. Historically,
the first laser diodes emitted around 800 nm. Thus, neodymium which exhibits an efficient
absorption in this energy range was a good candidate. Efficient laser systems have been
developed with neodymium ions and picoseconds range lasers are available but its narrow
emission lines prevent it from being used as active ions for ultra short pulse generation in the
fs range. Later, 980 nm laser diodes became available and ytterbium ions seemed much more
interesting for high power laser and ultra short pulses thanks to a well-fitted electronic
structure. Ytterbium ion combines two interesting properties: first, it only has two electronic
manifolds and this simple energy scheme prevents any parasitic absorption such as ESA
(Absorption in the Excited State) under high pumping power [5]. Indeed 4f-4f transitions of
ytterbium are very broad by comparison to similar transitions in other rare earth ions which is
interesting for femto-second pulse generation.
The present chapter is divided into three parts. First, ytterbium ions optical properties and
origin of the broad emission band are presented. Then, spectroscopic and thermo-mechanical
criteria to choose the most appropriate laser material for a desired application are developed.
Finally, several families of material of relevant interest for femto-second pulse generation are
presented.
2. WHY IS YTTERBIUM ION SO INTERESTING?

a. General Presentation
Ytterbium has been discovered by Swiss chemist J.C. Galissard de Marignac in 1878.
This name comes from the Swedish town where he found the mineral containing ytterbium.
Chemical and physical properties of ytterbium have only been determined after 1953 when
pure metallic ytterbium became available.
Ytterbium possesses 67 electrons that locate it in 6th period at the last position but one of
the lanthanide series. Energy levels of electrons in Yb3+ ions results from two contributions:
free ion splitting and crystal field splitting.
First, to understand free ion splitting, it is necessary to consider orbital approximation in
a central field model. Hamiltonian for a Ze charge ion with N peripheral electrons has the
following expression:
Ytterbium Doped Materials for Femtosecond Lasers 317
− h2 N
Ze 2 N e 2 N r r
H free ion = − .∑i Δ i − ∑ + ∑ + ∑ ζ (ri ) si .li
2m i =1 ri i > j rij i =1
(Eq. 1)
H free ion = H kinetic + H potential + H coulombic repulsion + H spin −orbit
where i and j are for electrons, ri is the distance between an electron and the nucleus and rij is
the interelectronic distances. Terms are classified by decreasing order of magnitude. The first
two terms (kinetic and potential energies) represent the configurations energy and define
electronic orbitals. It splits the 67 electrons of Yb3+ ions in electronic shells and leads to the
electronic configuration of the ion:
1s22s22p63s23p64s23d104p65s25p64f13 = [Xe]4f13
4f shell is not complete and the missing electron is responsible for all optical properties of
trivalent ytterbium ions. 4f orbitals are less extended than 5s and 5p so lanthanides are
scarcely any sensitive to the environment (electronic shielding) and they are difficult to
separate.
Generally, electronic repulsion, coloumbic interaction between two electrons of a same
ion, can lead to partial splitting of electronic energy levels. But Yb3+ ions, 4f13, behave as 4f1
so only one single electron is to consider in the 4f shell and there is no splitting due to
columbic interaction. Energy levels determined at this point are called spectroscopic terms
2s+1
L. Ytterbium is in a term 2F (L=3, orbital momentum and S=1/2, spin momentum),
degenerated (2S+1) (2L+1) =14 times.
In the case of heavy ions like rare earth ions, spin-orbit coupling is high and its effects
cannot be treated as a perturbation of the kinetic and potential energies. So, in order to
describe ions energy states, total kinetic momentum J=Σli+Σsi has to be considered. Under
spin-orbit interaction, 2F term splits into to two manifolds 2FJ: 2F7/2 and 2F5/2 degenerated four
and three times respectively and separated by ΔESO ~ 104 cm-1.
Then interaction of electrons with the environment has to be taken into account by the
mean of the crystal field model. Basically, it consists in a Stark effect caused by the columbic
interaction of ytterbium electrons with surrounding atoms considered as point charges [6].
This contribution splits the two manifolds on an energy range ΔEcc between 102 cm-1 and 103
cm-1.
H = H free ion + H crystal field ( Eq. 2)
Effects of different splitting terms are depicted on a schematic emission profile on figure
1. First every electron has the same energy. Then Hspin-orbit generates two manifolds between
which optical transitions can occur (curve of left). Intensity of the transition is very high
(truncated on the picture) and has a narrow spectral width. After that, Hcrystal field splits
manifolds into stark sub-levels that change the single emission line into four lines with
equivalent width. This interaction is responsible for the changes in transitions energies from
one material to the other.
Figure 1. Energy levels splitting in Yb3+.
10000
1fs
1000
10 fs
Spectral bandwidth (nm)
Ti:Sapphire
100
Yb:crystals 100 fs
10
1 ps
Nd:crystals
1
0,1
0 500 1000 1500 2000 2500
Central wavelength (nm)
Figure 2. Required bandwidth for the production of fs pulses vs central wavelength of the laser
emission. Position of Ti:Sapphire , Nd and Yb doped crystals are depicted.
Finally, every level is broadened by the coupling of the electrons with host phonons
(vibronic coupling), so transitions are broadened. This interaction is developed into further
details in the following section.
Concerning laser operation, ytterbium ions possess several assets. First, its simple
electronic structure prevents from any absorption in the excited state or cross-relaxation like
that occurs in neodymium ions. So, heavily doped materials can be used and thin disk laser
configuration becomes very attractive. Moreover, ytterbium lifetime (~500 μs) is long
compared to neodymium that facilitates energy storage. In the case of doped YAG crystals,
Yb3+ lifetime is even 1.2ms while Nd3+ lifetime is only 230 μs. Another interesting point
concerns parasitic heat production. Ytterbium based lasers operate with very low quantum
defect values (ηq<5%). Quantum defect is the relative difference between laser and pump
energies.
λlaser − λ pump
η q = Quantum defect = ( Eq. 3)
λ pump
So, little heat is generated during laser action by comparison to operating conditions of
Ti3+ (ηq~50%) or even Nd3+ (ηq~25%) when half or quarter of the absorbed power is
converted into heat.
For further information on the lanthanide cations optical spectroscopy please refers to
dedicated works [6], [7].
b. Origin of the Broad Emission Lines
The 4f electrons give to lanthanides series their unique optical properties dominated by
relatively narrow lines due to intershell 4f-4f transitions. Some phenomena, such as
multiphonon relaxation, phonon assisted energy transfer and so on are only possible due to
interaction between the 4f electrons and the surrounding. The strength of the electron-phonon
interaction in the trivalent rare earth ion spectra depends (i) on the host matrix, (ii) on the
concerned ion and (iii) on the particular transition under consideration.
It was first expressed by Hellwege that vibronic phenomena are much stronger at both
ends of the trivalent rare earth ions (cerium, praseodymium, thulium and ytterbium) than in
the middle of the rare earth series (europium, gadolinium, and terbium). More recent work of
Ellens et al. [8] on the variation of the electron-phonon coupling in the lanthanide series in
different oxide and fluoride hosts indicate a specific trends : the coupling is indeed strong for
ions at the beginning and at the end of the lanthanide series and weak in the middle. In
addition, in covalent host lattice (such as oxides) the coupling is much stronger than in ionic
matrices (such as fluoride hosts).
The authors have performed line-broadening experiments on nine rare earth ions in LiYF4
(YLF) and the results are presented below.
After comparison to Raman broadening and inhomogeneous effect, Ellens et al. derived
an effective electron-phonon coupling parameter. The variation of the electron-coupling
parameter well follows the tendency first described by Hellwege. The origin of such variation
throughout the rare earth series is therefore an antagonistic effect of lanthanide contraction
and shielding of the 4f electrons:
• Due to lanthanide contraction, the 4f orbitals are contracted and the overlap between
the 4f orbitals and the surrounding decreases in the lanthanide serie.
• Shielding of the 4f electron by the 5s2 and 5p6 electrons varies in the serie from 1.1
for the trivalent cerium down to 0.6 in the trivalent ytterbium. This smaller screening
factor leads to an increase of the electron-phonon coupling in the lanthanide series.
The two effects explain the shape of the observed electron-phonon coupling curve. Then
Yb3+ presents the strongest electron-phonon coupling and therefore this ion will be very
favourable to obtain ultra short laser pulse as described in the following part of the paper.
3. REQUIRED PROPERTIES FOR FEMTO-SECOND GENERATION

MATERIAL
The primary spectroscopic requirement for ultra short pulse generation is related to the
conservation of the time-bandwidth product:
Δt.Δν ≥ 0.32 (Eq. 4)
where Δt and Δν are the full width at half maximum (FWHM) of pulse duration and the pulse
spectrum width respectively. This inequality corresponds to the transformed-limited pulses
(and the strict inequality corresponds to chirped pulses). Hence, ultra short pulses require gain
over a very broad spectral range and, for example, to generate 100 fs pulse at 1 μm, the
required spectral bandwidth is 10 nm.
As presented before, ytterbium ions can be pumped by laser diode and exhibit wide
emission bands. But, intrinsic bandwidth of Yb3+ could not be not sufficient to generate
femto-second pulses and a broadening by the host crystal is necessary. Hence, ytterbium is
the most interesting rare earth ions, but the choice of host material is more critical. In this
section, origin of important emission band broadening due to the matrix will be presented.
a. Multi Sites Materials
Femto second pulse generation requires broad gain range at laser wavelength. An
interesting way to get broad emission band is to consider multi-sites material and as ytterbium
concentrations should be high in the laser system, the occupancy of several sites is favoured,
that broaden optical features of Yb3+ ions. Multi-sites crystals consist in materials with
several insertion sites for ytterbium ions. Each site would produce a defined electric field and
split ytterbium levels according to a given energy level scheme. The superposition of several
energy schemes gives a broad emission band. Examples are numerous in litterature, in
particular with simple formula crystals such as Yb:CaF2 [9], or sesquioxides such as lutecia
(Lu2O3)[10].
350
300
250
200
αaverage
150
100
50
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Number of 4f electrons
Figure 3. Effective electron-phonon coupling parameter values of the 4f-4f transitions of RE3+ in LiYF4
extracted from Ellens[8]. Maximum electron/phonon coupling occurs in Yb3+ ion.
Figures 3a, 3b and 3c are extracted from [10].
Figure 3a. Low temperature (LT) spectrum of zero-line absorption of Yb:Lu2O3 exhibiting two lines
related to the two sites. Figure 3b and c: LT fluorescence spectra Yb:Lu2O3 excited around 974.4 nm
(b) and around 975.6 nm (c). Figure 3d: Schematic view of the two coordination sites in the lutecia
structure.
Lu2O3 material is a cubic phase derived from fluorite (CaF2) structure. But in order to
keep the electroneutrality, balance between positive and negative charges, oxygen vacancies
occupy a quarter of anionic sites. According to the position of these vacancies in regards to
Yb ions, two insertion sites are generated for ytterbium ions: on site1, vacancies are aligned
on the diagonal of a plane of the unit cell; on site2, vacancies are aligned along the diagonal
of the unit cell. Consequently related symmetries are different: C2 and C3i in proportion 3:1.
Fluorescence spectra recorded at low temperature [10] and presented on figures 3b and 3c
show the two sites emission spectra.
b. Disordered Materials
Another way to produce broad emission bands is to use disordered crystals. A material is
said to be a disordered crystal when its atomic positions can be described by long range
reproduction of a unit cell by its symmetry elements but parts of the constituting atoms are
randomly distributed on these well defined crystalline positions. The long range symmetry
differenciates a disordered material and a glass. Such atomic disorder leads to large range of
environments around the active rare earth with gradual variation. Each environment would
give one specific transition and the sum generates a broad and smooth band. It can even lead
to broader emission bands than in glass systems. For example, the Yb doped gehlenite,
Yb:Ca2Al2SiO7, (CAS) [11] [12] has a disordered structure that gives a very broad emission
ranging from 975 to 1100 nm. Yb:CAS structure can be described by the alternation of two
types of layers composed of tetrahedra connected by their corner. In the first layer, tetrahedra
T1 are exclusively occupied by Al3+ ions while in the following layer, tetrahedra T2 are
shared by Al3+ and Si4+ ions. Occupation of crystallographic positions in this layer is random.
Between these two layers, Ca2+ ions are on one unique crystallographic site, called Thomson
cube (Tc). This site is very distorted with Ca-O distances ranging from 2.4Ǻ to 2.9Å. Yb3+
ions substitute Ca2+ in the crystal structure and consequently are located on a distorted site.
Random occupancy of T2 tetrahedra by Si4+ and Al3+ and Tc by Yb3+ and Ca2+ gives rise in
this compound of a very disordered behaviour. Disorder is related to the local surrounding of
cations with different sizes and charges Ca2+/Yb3+ and Al3+/Si4+. Tungstates and molybdates
are also interesting families for ultra short pulse generation that exhibit structural disorder
[13][14].
4. THERMO-MECHANICAL PROPERTIES
a. Thermal Behaviour Under Pumping
A current problem occurring during laser operation in diode pumped solid state laser
system is heat generation. In order to get increase output power, it is necessary to increase
pumping power but it generates thermal constraints leading to a decrease of the efficiency. So
heat production has to be limited and thermal loading has to be efficiently removed.
Numerous works have been dedicated to identify the proportion of pump power
converted into heat [15] [16]. Experimental measurements of internal temperature in the
material with and without laser action are scarce in the literature [17] [18] even if theoretical
works are numerous [15]. Nowadays thanks to the rise of computer performance, FEA (Finite
Element Analysis) method enables the computation of thermal profile in the material from a
thermal cartography of a surface. [19]
Heat generation has two main origins: one is a direct consequence of laser action and the
other one is an indirect effect. In most cases, ytterbium ions absorb pump photons around 980
nm and produce laser emission between 1000 nm and 1100nm. Hence emitted photons are
less energetic than pump photons and the relative energy difference, called quantum defect, is
completely converted into heat. So intrinsically, as soon as laser action is occurring, heat is
produced.
The second source of heat production is far more complex because it can imply crystal
defects, multi-photons processes (absorption in the excited state, up-conversion [20] [21]) or
multi-centres processes (radiative trapping leading to defect). This production can be reduced
by improving quality of materials or by using low concentration of dopant in the crystal.
Since quantum defect is difficult to reduce in the case of femto-second pulse generation,
the main solution consists in having defect free crystals with high thermal properties.
Composite crystals are another way to obtain more efficient materials [22] [23]. The
underlying idea is to cool the hottest face of the crystal with a high thermal conductivity
material hence reducing thermal gradient and therefore thermal lensing effect.
b. Thermal Resistance Criteria
In order to compare the resistance of different materials to pump power, several criteria
have been developed. First, one quantifies the resistance of a crystalline material to thermal
shocks [24] by:
'
(1 − υ) K κ
RT = c
Eα
(Eq. 5)
where υ is Poisson’s constant, E the Young’s modulus of the material and Kc the material
toughness which represents the resistance to crack propagation. κ and α are the thermal
conductivity and the thermal expansion respectively. The higher this parameter, the higher the
resistance of the material to thermal stresses. The evaluation of this thermo-mechanical figure
of merit may however be simplified if one assumes that Poisson’s coefficient is close to 0.25
and the ratio Kc/E is almost constant for crystalline materials. Therefore thermal shock
resistance coefficient can be simplified to R'T≈κ/α [25]. This criterion is useful to compare
single crystals. Concerning polycrystalline materials, increase of material toughness is to
expect thanks to multi-grain structure and joint nature that can alter cracks propagation [26].
This feature of these materials explains the recent interest for the transparent laser ceramics
for high power applications.
Key parameter evaluated in this figure of merit is related to the maximum of the
temperature difference that a laser crystal can withstand. Assuming that all heat is generated
inside the material by the quantum defect ηc (relative difference between pump and laser
wavelengths) one can relate pumping power to heat flow φmax inside the crystal:
ϕ max η c .Ppump (Eq. 6)

R't ∝ ΔTmax ∝ ∝
κ κ
In the case of femto-second laser pulses, only central laser wavelength is considered. It is
therefore possible to define a laser power resistance parameter RP [27] taking into account
both heat production and thermal behaviour of the Yb-doped material:
κ2
RP ∝ (Eq. 7)
αη c
In order to choose an appropriate material for femto-second laser operation, it is
interesting to plot the ability of a material to resist to pump power versus its ability to produce
short pulses. Garnet and sesquioxides seem interesting for high power application but are
restricted to long pulses. On the contrary borates (GdCOB, BOYS) can produce shorter pulses
but with lower energy. Finally, silicates, tungstates and aluminates present both advantages of
short pulses with a rather high Rp factor and they offer the possibility to reach interesting
compromises.
5. INVESTIGATION OF MATERIALS WITH RELEVANT INTEREST

a. State of the Art
Femto-second laser pulses have been generated in numerous ytterbium doped crystals
and the following tables exhibit the main obtained results in terms of average power and pulse
durations:
Figure 5. Rp parameter vs theorical shortest pulse.
Figure 6. Average power Vs. Pulse duration for main Yb-doped crystals.
Three domains are emphasized in the figures 5 and 6: (i) materials such asYb:SYS and
Yb:BOYS presenting broad emission leading to short pulse generation but rather low thermo-
mechanical properties, (ii) materials like Yb:YSO and Yb:LSO presenting rather broad
emission leading to short pulse generation and high thermo-mechanical properties, (iii)
Yb:CALGO with ultra broad emission leading to the shortest pulse ever generated with a bulk
material (47 fs long [28]) and good thermo-mechanical properties. Structural and optical
properties of these families are developed below.
b. Yb:BOYS/Yb:SYS: Short Pulses But Low Thermal Properties
Yb:BOYS and Yb:SYS constitute the first family. They are both multisite materials and
present low thermal properties values. Yb:BOYS which stands for Yb:Sr3Y(BO3)3 is a
rhombohedral borate discovered in the nineties [29][30]. Its crystalline structure is composed
of one-dimensional chains bridged by Sr2+ ions. These chains made of [YO6] octahedra are
stacked along c axis and linked by [BO3] groups on their corner. Along the c axis, the
alternation of polyhedra location generates two different sites for Y3+ ions, 3a and 3b. Since
Y3+ and Yb3+ ions have very closed radii and same electrostatic charges, Y3+ is easily
substituted by Yb3+ in the structure, therefore Yb3+ ions are distributed on these two
crystallographic sites. Sum of these two contributions gives a 60 nm broad emission band.
Yb:SYS is the apatite silicate matrix SrY4(SiO4)3O in which [SiO4] tetrahedra create two
types of insertion sites. 4f sites form columns parallel to C axis linked by 6h sites. In this
structure, it worth noting that there is one oxygen which is not implied in any SiO4
tetrahedron but it is bonded to three cations located in 6h sites.
C axis
Figure 7. SYS structure.

1.6
1.4
Emission cross-section (x10 cm )
2
1.2 SYS
-20
1.0
0.8
0.6
BOYS
0.4
0.2
Glass
0.0
900 920 940 960 980 1000 1020 1040 1060 1080 1100
Longueur d'onde (nm)
Figure 8. Compared emission of Yb:BOYS, Yb:SYS and Yb:glass.
According to electrostatic arguments [31][32], Y3+ fully occupies 6h site and shares 4f
site with Sr2+ ions. Additional broadening of Yb emission lines is given by the random
occupancy of 4f sites in the crystal. Multisite and disordered structure of Yb:SYS lead to a
very broad emission band (FWHM=73 nm) as shown in figure 8.
Both crystals exhibit emission band broader than those of Yb:glass with higher cross
sections. This common feature is in both case related to the multisite character of Yb:BOYS
and Yb:SYS.
Yb:BOYS emission profile is rather similar to a Yb doped glass with a very smooth
profile. This crystal continuously tunable from 1051 nm to 1076 nm is suitable for sub 100-fs
pulses generation interesting to substitute neodymium glass on terawatts chains. The shortest
pulse generated with this crystal is 69 fs with 80 mW of average power [33]. In the case of
Yb:BOYS, the power is limited by the thermal behaviour of the material. Indeed, Yb:BOYS
has a low thermal conductivity of 1.8W.K-1.m-1 which prevents this material from being used
under high pumping power in cw laser operation [34].
Emission profile of ytterbium ions in SYS crystals is slightly different from Yb:BOYS.
Emission is composed of two main bands (figure 8). Emission cross-section are higher than
those of Yb:BOYS and pulses of 70 fs, 156 mW of average power, have been generated [35].
Yb:BOYS and Yb:SYS crystals present interesting spectroscopic features but their Rp
factors, defined in the previous part, are very low: 0.04 and 0.05 respectively. These low
values result from the combined effects of a low thermal conductivity, an important thermal
expansion and a high quantum defect. Every element taken separately is not very low, for
instance thermal conductivities are similar to double tungstates ones, but the combination of
these low parameters values lead to low Rp values.
Yb:SYS and Yb:BOYS have demonstrated similar performances in term of pulse
durations. But Yb:SYS, thanks to a better thermal conductivity and higher spectroscopic
cross-section gives a two times higher output power for same pulse duration [35]. Despite its
low thermal properties, this material may be very attractive for development of fs oscillators
at 1070 nm.
c. Yb:YSO/Yb:LSO: Longer Pulses and High Thermo-Mechanical Properties
The second family is a sub-family of orthosilicate materials whose general formula is

RE2SiO5 where RE can be yttrium (YSO), lutetium (LSO) and gadolinium (GSO). They
belong to the multisite family and exhibit good thermo-mechanical properties. In addition, as
these materials are widely used in scintillation when doped with Ce ions [36] crystal growth
is well handled and these crystals are industrially grown since about 10 years.
Two structures have been identified for orthosilicate crystals depending on ionic radius of
the rare earth. When rare earth is smaller than Dy3+, that is the case for Y3+ and Lu3+, the
structure is monoclinic with space group C2/c. In the structure, four oxygens are bonded to
silicon to form [SiO4] tetrahedra and one oxygen is only connected to RE ion and participates
in the formation of distorted polyhedra. This is the reason why these materials are sometimes
written as RE2(SiO4)O. Rare earth ions occupy two sites in the material (RE1 and RE2): RE2
is surrounded by 6 oxygen atoms while RE1 corresponds to a surrounding of seven oxygen
atoms. In the case of RE2 sites, RE2-O distances are rather close between 2.16 Ǻ and 2.26 Ǻ.
On the contrary, RE1-O distance variations are more important between 2.16 Ǻ and 2.61 Ǻ.
Coordination polyhedra RE1 and RE2 of low symmetry C1 form compact chains, bridged by
[SiO4] tetrahedra, along C axis. In LSO and YSO, crystal structure does not exhibit disorder
what is depicted, figure 10, by peak shapes on emission profiles.
The occurrences of two sites (50% of sites being RE1 and 50% RE2) in orthosilicate
crystals give broad emission bands respectively 40 nm and 35 nm for YSO and LSO. These
bands are clearly less extended than previously presented crystals (Yb:BOYS and Yb:SYS).
This makes these crystals not appropriated for sub 100 fs pulse generation. But, thanks to
their good thermal conductivities when doped with Yb3+ (κLSO= 5.3 W.K-1.m-1 and κYSO= 4.4
W.K-1.m-1) and their cw laser efficiency [37], they are appropriate for sub-250fs high power
(multiWatts) pulse generation.
Laser efficiency of orthosilicate crystals is related to its strong ground state manifold
splitting (760 cm-1 and 960 cm-1) [38] compared with other ytterbium doped materials.
Figure 9. Crystal structure of LSO. Inset: Focus on the two Lu sites of LESO.
Figure 10. Emission bands of Yb:LSO (on top) and Yb:YSO (bottom) along X and Z polarizations.
Energy level population is governed by Boltzmann statistics so an important energy gap

between ground level and final laser level facilitates the population inversion. Consequently
laser threshold is low and laser efficient is high.
High power fs pulses have been achieved with YSO and LSO [39]. They respectively
provided 2.6W of average output power with 198 and 260 fs pulse durations. Moreover,
Yb:LSO gave more than 1.1W of output power for a 233 fs pulse. Unfortunately,
orthosilicates crystals are not continuously tunable because their emissions are not smooth
enough. YSO tunability is limited to three spectral domains: 1034-1034nm, 1039-1043nm
and 1046-1046nm.
Orthosilicate crystals YSO and LSO have high Rp factors thanks to their good thermal
conductivities (4.4 W.m-1K-1 and 5.6 W.m-1K-1). In addition their thermal expansion factors
are lower than those of BOYS and SYS crystals that make them well suited materials to
produce high power fs pulses.
d. Yb:CALGO: Short Pulse Duration with High Thermo-Mechanical

Properties
Yb:CALGO is the third type of materials described in this chapter. It presents disordered
structure and high thermo-mechanical properties. CALGO is the acronym of the double
aluminate CaGdAlO4. It has a tetragonal structure derived from K2NiF4 and has only two
cationic sites: one for Al3+ ions that forms [AlO6] octahedra and a second one, of coordination
nine and C4v symmetry, randomly shared by Ca2+ and Gd3+. Since Yb3+ substitutes Gd3+, it
has only one insertion site and in Yb3+ doped CALGO, three cations occupy one
crystallographic site. Considering the second coordination sphere of an active ion, composed
of nine Ca2+ or Gd3+ ions, there are 29 different environments. It has been demonstrated that
cations at the apical position are closer (3.37 Ǻ) than other cations ( 3.64 Ǻ and 3.65 Ǻ) from
the second coordination sphere and whether it is Ca2+ or Gd3+, that occupy this position, one
can defines two environments families [40].

This disorder among site occupancy generates an important inhomogeneous broadening
of ytterbium emission. The resulting emission consists in the sum of a continuum of sites and
not only two sites like in Yb:YSO or Yb:LSO. Emission profile is hence very smooth and
broad. It exhibits a wide plateau from 1000 nm up to 1060 nm that is very interesting for the
generation of fs pulses. Homogeneous gain is expected on a broad wavelength range. Such
bandwidth could theorically lead to pulses as short as 17 fs (figure 12).
CALGO also possesses excellent thermo-mechanical properties. Rp factor of Yb:CALGO
is high thanks to good thermal conductivities and reach 6.9W.m-1.K-1 along a-axis, a value
even better than Yb:YAG of similar Yb3+ concentration.
Thermal lensing measurements have being carried out under high power pumping and
Yb:CALGO demonstrated very low sensitivity to this phenomenon, showing therefore a
scalable behaviour of the laser properties. With 86 W of pumping power corresponding to
absorption of 42 W and under laser action, measured thermal lens is f=25 cm what
corresponds to a dioptric power of 4m-1 (see figure 13) and wavefront is almost not distorted.
These characteristics make this material remarkably well appropriate for high power laser
applications and good scalability [41].
Temperature mapping of the surface of the crystal indicates a global temperature

elevation of 105°C for 83 W of incident pump power and a temperature gradient in the crystal
of 20°C. This relatively small increase of temperature clearly confirms the excellent
conductivity of this crystal and contributes to the small thermal lensing.
The broad emission band combined to the excellent thermal properties of CALGO
allowed the generation of 47 fs pulses at a low average power of 38 mW [28] and of 68 fs
pulses at an average power of 520 mW [42]. 47 fs, is to our knowledge the shortest pulse ever
obtained with a bulk Yb-doped material.
Figure 11. Crystal structure of Yb :CALGO and focus on Yb surrounding.

1,8
cm )
2
1,6
-20
Emission cross section (x10
1,4
1,2
π-polarization
1,0 σ-polarization
0,8
0,6
0,4
0,2
0,0
940 960 980 1000 1020 1040 1060 1080 1100 1120
Wavelength (nm)
Figure 12. Emission bands of Yb :CALGO in σ and π polarization.
5
4.5 No laser
Dioptric power (m-1)

4 Laser
3.5
3
2.5
2
1.5
1
0.5
0
0 20 40 60 80 100
Incident power (W)
Figure 13. Thermal lensing vs pump power in Yb:CaGdAlO4 under high power diode pumping.
CONCLUSIONS
In this chapter, we first present some singular aspects of the Yb3+ spectroscopy in solid
state laser hosts. The two manifolds scheme for energy levels, crystal field and electron
phonon coupling effects play an important role for this trivalent ion which is very attractive
for solid state lasers. Ytterbium ions are very favourable among rare earth ions thanks to the
intrinsic broadening of their optical properties caused by their strong electron-phonon
coupling. Ytterbium doped crystals have therefore become recently reference materials to
produce femto second laser pulses and could probably replace Ti:sapphire for this application.
The origin of the broad emission in materials of relevant interest has been related to the
disordered or the multisite nature of the crystals. Then materials used as femto-second laser
have been ranked according to their resistance to laser flux.
In the case of multi sites character, broadening due to disorder compensates intensity
variations related to different sites. This leads to shorter pulses, 70 fs, in Yb:SYS in
comparison to Yb:LSO or Yb:YSO where pulses are 198 fs and 260 fs respectively. However
borates (Yb:BOYS, Yb:GdCOB) and silicates (Yb:SYS) have low thermo-mechanical
coefficients partly due to disorder in their structure.
Yb:CALGO seems to be an interesting compromise as far as optical properties (broadest
emission band for an Yb doped crystal) and thermo-mechanical properties (thermal
conductivity similar to Yb:YAG) are concerned.
In order to get shorter pulse duration, new ways have been recently opened. For instance
composite material can be realized by combining two laser ceramics presenting different
emission spectra to obtain a gain medium with a very broad emission bandwidth. 60-fs pulses
have been generated in a Yb:Sc2O3/Yb:Y2O3 composite while shortest pulse that could
theorically be obtained with these material are respectively 84 fs and 98 fs for Yb:Y2O3 and
Yb:Sc2O3 [43]. Pulse duration can also be shortened down to 20 fs by non linear compression
in photonic crystal fiber [44] and power can be increased to 4 MW for 63fs by Optical
Parametric Chirped-Pulse Amplification using Yb doped fiber amplifier [45].
Therefore Yb doped materials for generation of ultra short pulses are in large expansion
and research in this field is very active.
REFERENCES
[1] D Strickland, G. Mourou, « Compression of Amplified Chirped Optical Pulses », Opt.
Commun. 55, 447-449 (1985).
[2] P.F. Moulton, “Spectroscopic and Laser Characteristic of Ti:Al2O3”, J. Opt. Soc. Am. B
3, 125-133, (1986).
[3] D.E. Spencer, P.N. Kean, W. Sibbett,”60-fsec pulse generation from a self-mode-
locked Ti: Sapphire”, Opt. Lett. 16, 42 (1991).
[4] R. J. Keyes, T.M. Quist, “Injection luminescent pumping of CaF2:U3+ with GaAs diode
lasers”, Appl. Phys. Lett. 4, 50 (1964).
[5] G. Boulon; “Why so deep research on Yb3+ doped optical inorganic materials?”; J.
Alloys and compounds 451, 1-11, (2008).
[6] P.H. Haumesser, R. Gaumé, B. Viana, E. Antic-Fidancev and D. Vivien,
“Spectroscopic and crystal-field analysis of new Yb-doped laser materials”, J. Phys.:
Condens. Matter 13 5427–5447 (2001).
[7] G. Liu and B. Jacquier, “Spectroscopic properties of rare earths in Optical Materials,
Springer, 2005.
[8] A. Ellens, H. Andre, M.L.H. ter Heerdt, R.T. Wegh, A. Meijerink and G. Blasse,”
Spectral line-broadening study of the trivalent lanthanide ion series. II The variation of
the electron-phonon coupling strength through the series”; Phys. Rev. 55 1, 180, (1997).
[9] Y. Voron’ko, V.V. Osiko and I.A. Shcherbakov,”Optical centers and the interaction of
Yb3+ ions in cubic fluorite crystals”, Soviet. Phys JETP 29, 1, 86-90, (1969).
[10] K. Petermann, G. Huber, S. Kuch, E. Mix, V. Peters, S.A. Basun, “Rare Earth Doped
Sesquioxides”; J.Lum. 87-89, 973 (2000).
[11] B. Viana, A. M. Lejus, D. Saber, N. Duxin, D. Vivien; “Optical properties and energy
transfer among Nd3+ in Nd:Ca2Al2SiO7 crystals for diode pumped lasers”; Opt.
Materials 3 (4), 307-316, 1994.
[12] A.M. Lejus, A. Kahn-Harari, J.M. Benitez, B. Viana; “Crystal growth, characterization
and structure refinement of neodymium doped gehlenite, a new laser material
[Ca2Al2SiO7]” Materials Research Bulletin 29 (7), 725-734, (1994).
[13] J. Liu, J.M. CanoTorres, C. Cascales, F. Esteban-Betegon, M.D. Serrano, V. Volkov, C.
Zaldo, M. Rico, U. Griebner, V. Petrov "Growth and CW laser operation of disordered
crystals of Yb:NaLa(WO4)2 and Yb:NaLa(MoO4)2"; Phys. Sta. Sol. 202, 4, R29 (2005).
[14] A. García-Cortés, J. M. Cano-Torres, X. Han, C. Cascales, C. Zaldo, X. Mateos, S.
Rivier, U. Griebner, V. Petrov, F. J. Valle; “Tunable continuous wave and femtosecond
mode-locked Yb3+ laser operation in NaLu(WO4)2” ; J. of Applied Physics, 101,
063110-1-7, (2007).
[15] T.Y. Fan: “Heat Generation in Nd:YAG and Yb:YAG”, IEEE Journal of Quantum
Electronics, 29 (6), 1457-1459 (1993).
[16] J.L. Blows, T. Omatsu, J. Dawes, H. Pask, M. Tateda, “Heat generation in Nd:YVO4
with and without laser action”, IEEE Photonics Technology Letters, 10 (12), 1727-1729
(1998).
[17] J.F. Bisson, D. Fournier, M. Poulain, O. Lavigne and R. Mévrel, “Thermal conductivity
of Yttria-Zirconia single crystals determined with spatially resolved infrared
thermography”, J. Am. Ceram. Soc., 83 (8), 1993-1998 (2000).
[18] W. Koechner, “Absorbed pump power, thermal profile and stresses in a cw pumped
Nd:YAG crystal”, Applied Optics, 9 (6), 1429-1434 (1970).
[19] J. Didierjean, E. Herault, F. Balembois, P. Georges, “Thermal conductivity
measurements of laser crystals by infrared thermography. Application to Nd doped
crystals”, Opt. Express, 16 (12), 8995-9010 (2008).
[20] F. Balembois, F. Falcoz, F. Kerboull, F. Druon, P. Georges, A. Brun : “Theoretical and
experimental investigations of small-signal gain for a diode-pumped Qswitched Cr-
LiSAF Laser”, IEEE J. of Quantum Electronics, 33 (2), 269-278 (1997)
[21] S. Guy, C.L. Bonner, D.P. Shepherd, D.C. Hanna, A.C. Tropper, B. Ferrand, “High-
Inversion Densities in Nd:YAG: Upconversion and Bleaching”, IEEE J. of Quantum
Electronics, 34(5), 900-909 (1998).
[22] F. Druon, S. Chénais, F. Balembois, P. Georges, R. Gaumé, B. Viana, ”Diode-pumped
continuous-wave and femto-second laser operation of a hetero-composite crystal
Yb3+ :SrY4(SiO4)3O||Y3Al5O12”, Opt. letters, 30, 8, 857-859 (2005).
[23] J. Didierjean, P.O. Petit, F. Balembois, F. Druon, P. Georges, J. Petit, C. Boissière, P.
Goldner, B. Viana, “Characterization of a new method of sol-gel bonding for high
power lasers”, under press.
[24] W.F. Krupke, M.D. Shinn, J.E Marion, J.A. Caird, S.E. Stokowski, "Spectroscopic,
optical and thermomechanical properties of neodymium and chromium doped
gadolinium scandium gallium garnet" J. Opt. Soc. Am. B 3(1) 102-113 (1986).
[25] H. Yagi, K. Takaichi, K. Ueda, Y. Yamasaki; T. Yanagitani, AA. Kaminskii; “The

physical properties of composite YAG ceramics”; Laser Physics, 15(9): 1338-44 (2005)
[26] R. Gaumé, B. Viana "Optical and thermo-mechanical propertiess of solid state laser
materials"; Ann. Chim. Sci. Mat. 28, 89-102 (2003).
[27] J. Petit, P. Goldner, B. Viana; “Laser emission with low quantum defect in
Yb:CaGdAlO4”, Opt. Letters; 30(11), 1345-7; (2005).
[28] Y. Zaouter, J. Didierjean, F. Balembois, G. Lucas-Leclin, F. Druon, P. Georges, J. Petit,
P. Goldner, B. Viana; “47-fs diode-pumped Yb3+:CaGdAlO4 laser”; Opt. Letters, 31(1),
119-211, (2006).
[29] P.H. Haumesser, R. Gaumé, B. Viana, J.-M. Benitez, B. Ferrand, G. Aka, D. Vivien, J.
Crsyt. Growth, 233 , 233 (2001).
[30] P.H. Haumesser, R. Gaumé, B. Viana, D. Vivien, « Yb3+ lasers based on Yb doped
M3RE(BO3)3 where M=Ba, S rand RE= Lu, Y, Sc », US patent n°09/782,787 14 feb
2001.
[31] G. Blasse, “Influence of Local Charge Compensation on Site Occupation and
luminescence of apatites”, Journal of solid state chemistry 14, 181 (1975).
[32] L.W. Schroeder and E. Wolf “Cation ordering in Ca2La8(SiO4)6O2” Journal of Solid
state chemistry 26, 383 (1978).
[33] F. Druon, S. Chenais, P. Raybaut, F. Balembois, P. Georges, R. Gaume, G. Aka, B.

Viana, S. Mohr, D. Kopf ; “Diode-pumped Yb:Sr3Y(BO3)3 femtosecond laser”, Optics
Letters, 27(3), 197-199 1Feb 2002.
[34] R. Gaumé, B. Viana, D. Vivien, J.P. Roger, D. Fournier, JP Souron, G. Wallez, S.
Chenais, F. Balembois, “Mechanical, thermal and laser properties of Yb:(Sr1-
xCax)3Y(BO3)3 (Yb:CaBOYS) for 1 μm laser applications”, Opt. Materials, 24(1-2):
385-92, (2003).
[35] F. Druon, F. Balembois, P. Georges “Ultra-short-pulsed and highly-efficient pumped
Yb:SYS mode-locked oscillators”, Opt. Express, 12(20), 5005-5012, 4 Oct. 2004.
[36] C. Van Eijk, “Perspectives on the future development of new scintillators”, Nuclear
Instruments and methods in Physics Research Section A: Accelerators, Spectrometers
Detectors and Associated Equipment, 537, 1-2, 6, 2005.
[37] R. Gaumé, P.H. Haumesser, B. Viana, D. Vivien, B. Ferrand, G. Aka; “Optical and
laser properties of Yb:Y2SiO5 single crystals and discussion of the figure of merit
relevant to compare ytterbium-doped laser materials” Opt. Materials 19(1), 81-8.
(2002).
[38] S Campos, A Denoyer, S Jandl, B Viana, D Vivien, P Loiseau and B Ferrand,
“Spectroscopic studies of Yb3+-doped rare earth orthosilicate crystals ”, J. Phys.:
Condens. Matter 16 4579–4590 (2004).
[39] F. Thibault, D. Pelenc, F. Druon, Y. Zaouter, M. Jacquemet, P. Georges; “Efficient
diode-pumped Yb3+:Y2SiO5 and Yb3+:Lu2SiO5 high-power femtosecond laser
operation”; Opt. Letters, 31(10), 1555-7, (2006).
[40] P.O. Petit, J. Petit, P. Goldner, B. Viana; “Inhomogeneous broadening of optical
transitions in Yb:CaYAlO4”; Opt. Materials 30(7), 1093-1097, (2008).
[41] P.O. Petit, P. Goldner, B. Viana, J. Boudeile, J. Didierjean, F. Balembois, F. Druon, P.
Georges; “Diode pumping of Yb3+:CaGdAlO4”, Proceedings-of-the-SPIE-The-
International-Society-for-Optical-Engineering. 25 April 2008; 6998: 69980Z-1-9.
[42] J. Boudeile, F. Druon, M. Hanna, P. Georges, Y. Zaouter, E. Cormier, J. Petit, P.
Goldner, B. Viana, “Continuous-wave and femtosecond laser operation of

Yb:CaGdAlO4 under high-power diode pumping”; Opt. Letters, 32(14): 1962-4. (2007).
[43] M. Tokurakawa, A. Shirakawa, K. Ueda, H. Yagi, T. Yanagitani, AA. Kaminskii;
“Diode-pumped sub 60-fs Kerr-lens mode-locked Yb-doped sesquioxide combined
ceramic laser”; Conference on Lasers and Electro Optics, (2008).
[44] F. Druon, P. Georges; “Pulse-compression down to 20 fs using a photonic crystal fiber
seeded by a diode-pumped Yb:SYS laser at 1070 nm”; Opt. Express 12, 3383-3396
(2004).
[45] N. Papadopoulos, Y. Zaouter, M. Hanna, F. Druon, E. Mottay, E. Cormier and P.
Georges, “Generation of 63 fs 4MW peak power pulses from a parabolic fiber amplifier
operated beyond the gain bandwidth limit”; Opt. Express 32, 2520-2522, (2007).
Chapter 10
ULTRAFAST DYNAMICS OF PORPHYRINS IN HIGHER

EXCITED STATE
Dae Won Cho1, Mamoru Fujitsuka2 and Tetsuro Majima∗2

1
Department of Chemistry, Chosun University, Gwangju, 590-711, Korea
2
The Institute of Scientific and Industrial Research (SANKEN),
Osaka University, Mihogaoka 8-1, Ibaraki, Osaka 567-0047, Japan
1. INTRODUCTION
It is well known that in the natural photosynthesis system excitation energy captured by
the light-harvesting complex is transferred to the reaction center after an energy migration
process [1]. In the natural system, higher excitation energy generating the higher excited
states of dyes such as chlorophylls and carotenes is also utilized, despite its very short lifetime
[2]. Inspired by the natural system, artificial energy transfer systems, in which energy transfer
of higher excited-state energy is realized, have been demonstrated by various researchers [3].
On the other hand, for most molecules, no excited singlet states (S2, S3…states) have been
observed to emit light upon excitation, the same being true for triplet states (T2, T3…states).
This is summarized in Kasha’s rule: The emitting electronic level of a given multiplicity is the
lowest excited level of that multiplicity [4].
In most photochemical reactions, once higher excited states are produced by excitation of
the second or higher absorption bands, the generated excited states usually deactivate very
rapidly to the lowest excited state, which deactivates to the ground state by emission or
radiationless processes or undergoes reactions competing with the deactivation processes.
These behaviors of higher excited states are often said to obey Kasha’s rule. The courses of
most photochemical reactions do not depend on the wavelength of excitation. However, we
can easily recognize that some photochemical reactions show wavelength-dependent features
which do not obey Kasha’s rule. [5,6]
∗
Email: majima@sanken.osaka-u.ac.jp
336 Dae Won Cho, Mamoru Fujitsuka and Tetsuro Majima
Ar Ar
N N
N N
N M N N M N
N N
N N
Ar Ar
Azulene porphyrin phthalocyanine

Scheme 1.
In contravention of Kasha’s rule, one of the well-known groups of compounds is azulene

(Scheme 1) and its derivatives. Azulene shows two absorption bands in the visible and UV
light regions, but the second singlet excited state (S2) emits ‘anomalous’ fluorescence in the
UV region competing with internal conversion to the lowest singlet excited state (S1) as
shown in Figure 1 [7]. Beer and Longuet-Higgins discovered that azulene displays the
anomalous (S2 →S0) fluorescence from the second excited state upon excitation in 1955 [8],
whereas that normal S1→S0 fluorescence and T1→T0 phosphorescence were too weak to be
observed. The absence of S1→S0 fluorescence suggests that S1 to S0 internal conversion is a
very rapid process in azulene, this supposition was confirmed by the lifetime of the S1 state
being determined as 1.9 ps [9]. Azulene has much wider spacing between its zero–point levels
of the first two excited singlet states (S1, S2) compared with that of naphthalene, the isomer of
azulene [10].
Figure 1. The normalized absorption (—) and emission (---) spectra for Azulene in CH3CN. Abs is the
absorption band, while em is the emission band. Reprint with permission form [7], Yeow. et al., J.
Phys. Chem. A 2004, 108, 10980-10988. © 2004, American Chemical Society.
Ultrafast Dynamics of Porphyrins in Higher Excited State 337
The anomalous fluorescence of azulene subsequently occurs because internal conversion

fails to take place between the S2 and S1 level, on account of wide energy separation.
Porphyrins are well known compounds as the other authentic example of a molecule
which violates Kasha’s rule. Porphyrins also have two major absorption bands, namely the Q
band in the visible light region and the Soret band (B band) in the violet or ultraviolet region.
Usually, photochemical reactions of porphyrins occur from the S1 or T1 states, of which the
latter is generated from the S1 state by the intersystem crossing process, although the
reactions should compete with the deactivation processes such as internal conversion and
emission processes including the fluorescence and phosphorescence, Figure 2.
Fluorescence (S2-emission) from the Soret band of porphyrins was first observed by
Bejema et al. [11] for Zn(II)-5,10,15,20-tetrabenzprophyrin (ZnTBP) in octane containing
0.5% pyridine.
The reported quantum yields of S2 emission were of the order of 10-3 - 10-4, much lower
than those from the S1 state. Furthermore, the lifetime on S2 state of porphyrins was of the
order of picoseconds.
In spite of a lot of difficulty in studying on the photochemistry of porphyrins in the S2
state, the S2 state of porphyrins attracts much attention, particularly in the standpoint of solar
energy conversion mimicking biological systems in nature.
Phthalocyanine compounds also exhibit two well-separated absorption bands, namely an
intense Q band at 600-800 nm and B band at 300-400 nm as shown in Figure 3 [12]. Some
phthalocyanines showed fluorescence not only at the longer wavelength side of the Q band
but also in the violet region at the longer-wavelength side of the B band [13-15].
Interestingly, the violet emission from the S2 state of phthalocyanines shows the lifetime of
several nanoseconds. This is roughly comparable to, but a little longer than, the S2 lifetime of
azulene. However, the S2 lifetime of porphyrins is much shorter than those of azulenes or
phthalocyanines.
Figure 2. Schematic energy diagram of porphyrins in the excited states: Fl, fluorescence; phosp,
phosphorescence.
Figure 3. Electronic absorption and fluorescence emission spectra (λex = 355 nm) of argon saturated
solution of chloroaluminum-phthalocyanine with tetra-α-butoxy chains in THF. Reprint with
permission form [12], Gan. et al., Optics Express 2005, 13, 5424. © 2005, Optical Society of America.
Therefore the photochemistry in the S2 state of porphyrins is fertile subject for

investigation by approaches, including ultrafast time-resolved spectroscopy using a
femtosecond pulse laser.
The higher excited states generally dissipate very rapidly, but competing with such rapid
deactivation, often show unique behavior different from the lowest excited states, such as
characteristic emission and enhanced reactivity in electron transfer, energy transfer and so on.
This chapter aims to describe the photophysical and photochemical behaviors of the higher
excited states of porphyrins populated mostly by irradiation of their second or generally
higher absorption bands with violet or ultraviolet light, particularly utilizing femtosecond
time-resolved spectroscopy.
2. FEMTOSECOND TIME-RESOLVED MEASUREMENTS FOR

ULTRAFAST S2-STATE DYNAMICS OF PORPHYRINS
Because of quite short lifetime of S2-excited state of porphyrin derivatives, spectroscopic
methods employing femtosecond laser are essential to evaluate reaction dynamics.
Nowadays, the sub-picosecond transient absorption spectra are measured by the pump and
probe method using a regenerative amplified Ti:sapphire laser. Also the time-resolved
fluorescence measurements are quite useful to get information about the excited S2 state of
porphyrins.
Especially, the fast fluorescence dynamics of the excited molecules shorter than about 10
ps can be measured by using fluorescence up-conversion method.
2.1. Transient Absorption Measurement
Sub-picosecond time-resolved transient absorption spectra can be measured by the pump

and probe method using femtosecond laser.
Usually the second harmonic oscillation (~400 nm) of the output of the Ti:sapphire laser
can be used as the excitation pulse to excite the porphyrin samples to the S2 state. For
excitation to the S1 state, visible pulse (~550 nm) generated by OPA can be used.
A white light continuum pulse, which is generated by focusing the residual of the
fundamental light to the continuum generator such as a sapphire glass or flowing water, is
used as the probe light.
Temporal delay of the probe pulse from the pump pulse can be controlled by a computer-
controlled optical delay. For spectrum measurements, a CCD detector equipped with a
polychromator is often employed. Time resolution of the transient absorption measurements
is limited by pulse duration of the laser.
2.2. Fluorescence Up-Conversion Technique
Usually, fluorescence time profile has been measured by time-correlated single photon
counting (TCSPC) method using a time-to-amplitude converter or a streak camera. The time-
resolution of TCSPC is usually several ten picoseconds due to various electric factors such as
jitter of the trigger diode even when the excitation pulse duration is shorter than 1 ps. For
measurement of fluorescence decay with shorter lifetime than ps, a gating technique should
be implements, one of which is fluorescence up-conversion. In this technique, a short laser
pulse of ~100 fs (the gate) is mixed with the emission from the sample, which is excited by
another ~100 fs pulse, producing a signal at the sum optical frequency (ωgate + ωem). Since the
mixing takes place with co-existence of the emission and the gate pulse, this provides time-
resolution comparable to the laser pulse-width (~100 fs).
The power of the sum-frequency generated (P3) is given by P3 = αP1P2I2(ωsum/ωem)

(where P1 and P2 are the powers of the gating and fluorescence pulses, respectively, α is a
coefficient related to the crystal nonlinearity, and I is the length of the crystal) and therefore is
directly proportional to the fluorescence intensity.
Time-dependent data is obtained by optically delaying the gate pulse with respect to the
excitation pulse.
Usually, the up-conversion measurement can work with either a mode-locked Ti-sapphire
oscillator or Ti:sapphire amplifier. The second harmonic oscillation (400-420 nm) of the
output of the femtosecond laser (800-840 nm) was used to excite metalloporphyrins in a cell
with short optical path length.
After ultrafast excitation of the sample, the emitted light is collected and focused onto the
nonlinear crystal (BBO type I) where it is mixed with the residual fundamental gate-pulse.
The change of temporal delay between the pump and gate pulse is realized with a variable
delay-stage. The sum frequency signal is collected by an arrangement comprising reflectors
or lens, a PMT and the photon counting after passing through the monochromator as shown in
Figure 4.
Fundamental
BBO
Delay Stage WP
BBO ωem Sample pump
RO
ωsum
up-conversion signal gate ωg
Monochromator
PC
PMT
Figure 4. Fluorescence up-conversion set-up for sub-picosecond time-resolved emission measurements:

WP, wave plate; RO, reflective microscope objective.
3. PHOTOPHYSICAL PROPERTIES OF PORPHYRINS IN THE S2 STATE

Porphyrins have been the focus of a series of time-resolved spectroscopic investigations
over the past years in order to gain a better insight into the excited-state dynamics of natural
photosynthesis systems. The radiationless transitions are important factors affecting
photophysical properties of porphyrins. The energy gap law has often been used successfully
to interpret the variation in radiationless decay rates of the electronic excited states of families
of structurally related molecules. The energy gap law is an extension of the weak coupling
case of radiationless transition theory. For the families of molecules in which the electronic
coupling energies are approximately constant, the rate of radiationless decay varies with the
Franck–Condon factor for the two coupled states – approximately as the inverse exponent of
their electronic energy spacing. Early successes in the application of the energy gap law are to
be found in correlations of the rates of radiationless decay of the lowest excited singlet and
triplet states of families of aromatic hydrocarbons with their S1–S0 and T1–S0 electronic
energy spacing, respectively. The rates of radiationless decay of some higher electronic states
of porphyrins have been shown to exhibit energy gap law correlations. This section
introduces notable the S2–S1 internal conversion rates of porphyrin derivatives and porphyrins
array systems.
3.1. Steady-State Spectroscopic Properties of Porphyrins in the S2 State
The characteristic absorption and fluorescence spectra of Zn tetraphenylporphyrin (TPP)

in benzene are shown in Figure 5 [16]. The ground state absorption is very typical for Zn
porphyrins with an intense sharp Soret (B) band at 418 nm, which has a large absorption
coefficient of the order of 105 M-1cm-1, and the S0-S1 transition (Q-band) in the red spectral
region at 547 nm.
Figure 5. The ground-state absorption and static fluorescence spectra of ZnTPP in benzene. For the
fluorescence spectrum, the excitation wavelength was 397 nm and the sample concentration was 1.2 ×
10-7 M in a 5 mm cell. Reprint with permission form [16], Yu, H. Z. et al., J. Phys. Chem. A 2002, 106,
9845. © 2002, American Chemical Society.
In the case of Soret-band excitation, the emission consists of the S1 fluorescence from Q
band (655 and 602 nm) and the S2-fluorescence from Soret band in the region 430 nm. In
deviation from Kasha’s rule, the observed anomalous luminescence is a radiative emission
from a higher excited state (S2).
The first report on S2 fluorescence of ZnTBP was followed by observations of several
MTPPs (M = Zn, Cd and Al) [17] as well as a series of rare earth MTPP (M = Mg, Lu, Gd,
Tb, Sm, Yb and Eu) in solutions [18], although the quantum yields of S2 emission were much
lower than those from the S1 state. Martarno et al. [19] found also S2 fluorescence for
YIIITPP, LuIIITPP , and ThIVTPP complexes. Recently, the S2 fluorescence properties of SbV
[20] and PV [21] complex with TPP are reported by Fujitsuka et al., and SnIVTPP [22] by Jang
et al. Among various porphyrins, especially the spectral properties of ZnTPP are well
established [17,23,24].
3.2. Ultrafast Deactivation of the S2 State in Porphyrins
3.2.1. Energy Gap Law

The main deactivation process of the S2 state of zinc porphyrin should be the internal
conversion from the S2 to the S1 state, since the quantum yield of the S2 fluorescence is quite
small. According to the energy gap law [25,26], the internal conversion rate is given as kIC ∝
exp(−βΔE), where β and ΔE are the proportionality constant and the energy gap between the
zero vibrational states in two electronic states, respectively. One of the reasons for the small
S2 fluorescence quantum yield has been attributed to the modest (ca. 1 eV) energy gap
between their S2 and S1 electronic states [27]. A larger energy gap between the lowest and
higher excited state tends to reduce the deactivation and thus enhance the fluorescence from
the higher electronic states. ZnTBP has a much larger S2–S1 gap, ΔE(S2–S1) = 8953 cm-1 in
the bare molecule [28]. The relatively large energy gap (ΔE(S2–S1) = 7752 cm-1) between S1
and S2 in ZnTPP [29] might be the reason for the measurable S2 lifetime and also the reason
for the unusual fluorescence. ZnTPP show the good energy gap law correlation in a wide
variety of solvent. However, when other diamagnetic zinc porphyrins are excited in the Soret
region, the vast difference in radiationless decay rates was found. The energy gap law of
radiationless transition theory does not apply to other Zn porphyrins [30]. For example, Zn
octaethylporphyrin (ZnOEP) is almost non-fluorescent on excitation in the Soret band. The
upper state lifetime of ZnOEP is about two orders of magnitude shorter than that of ZnTPP
even though the S2–S1 electronic energy gaps of the two molecules are similar [31]
3.2.2. Deactivation of the S2 State by Internal Conversion

Regarding deactivation of the S2 state of Zn porphyrins, many groups reported that the
decay of the S2 emission is accompanied by a simultaneous rise of the S1 emission with the
same time constant and showed that the S2 state undergoes internal conversion to the S1 state
in competition with fluorescence emission [32]. For ZnTPP, the variety of S2 lifetimes has
been reported for measurements in a wide range of concentrations in a variety of solvents. It
should be pointed out that the measurement made at solute concentrations over ca. 20 μM
could be influenced by the excitation of solute aggregates. Furthermore, estimation using high
fluent laser will lead to wrong results because of degradation of the sample by multi-photon
process.
In the 1980s, S2 fluorescence quantum yields of ZnTPP were measured in various
solvents [17,18,33,34], and used to give estimations of the S2 fluorescence lifetime ranging
from 240 fs [17] to 3.6 ps [18]. Line widths measured in a detailed study of the fluorescence
excitation spectrum of ZnTPP seeded in a pulsed supersonic expansion of helium [29] set a
lower limit of 1 ps for S2 relaxation in the isolated molecule.
In the 1990s, Chosrowjan et al. [35] extended ultrafast laser techniques to the study of
ZnTPP dynamics, reporting S2 fluorescence lifetimes of 3.5 and 0.75 ps for ZnTPP in CH3CN
and CH2Cl2, respectively. On the basis of those measurements and a dependence of ZnTPP
fluorescence yield on CH2Cl2 concentration when CH3CN /CH2Cl2 mixtures were used as
solvent, an effective intermolecular electron transfer from the S2 state of ZnTPP to CH2Cl2
was suggested. The same year, separate spectral hole-burning experiments on ZnTPP in
polymer matrixes doped with electron acceptors at cryogenic temperatures were reported by
two groups, [36,37] and opposite conclusions were reached on the question of electron
transfer from the S2 state. S2 lifetimes ranging from 0.85 to 1.6 ps were also inferred from the
hole line widths [37]. Additional data supporting charge transfer from S2 to CH2Cl2 is
reported in detail in a review by Tokumaru [38].
In 1998, Gurzadyan et al. [23] monitored the decay of the S2→S0 fluorescence and the
rise of S1→S0 fluorescence for ZnTPP in ethanol using the fluorescence up-conversion
technique and fit both to a single lifetime of 2.35 ps. In addition to this data, and
measurements in other solvents by the same group [39], including a 2 ps S2 lifetime for
ZnTPP in CH2Cl2, supported a simple picture of relaxation of an equilibrated S2 population
by internal conversion to S1. Reports of ZnTPP S2 relaxation lifetimes of 1.2 ps in CHCl3 [40]
and 1.6 ps in dimethylformamide [41] have been obtained from the application of transient
absorption methods in conjunction with electron transfer or energy-transfer studies in zinc
porphyrins and derivatives. Furthermore, a weak 18 ps component was also observed and
attributed to vibrational cooling in the S1 state following internal conversion from S2. This is
similar to the ~10 ps time scale of vibrational cooling inferred from evolution of the transient
absorption of a variety of different porphyrin systems under various excitation conditions
[42,43].
Mataga et al. [24] used fluorescence up-conversion to explore the fluorescence from
ZnTPP in ethanol over a wide wavelength range. Their measurements showed the same 2.3 ps
S2 decay and S1 rise as reported previously [23]. On the other hand, Yu et al. [16] reported
that excitation of the low-lying singlet excited state (S1) results in nanosecond relaxation,
while excitation to S2, the Soret band, leads to multiple electronic and vibrational relaxation
time scales of S2 and S1 populations, from hundreds femtoseconds to tens picoseconds
The vibrational relaxation process have been studied other zinc porphyrins. For example,
for zinc(II) 2,8,12,18-tetraethyl-3,7,13,17-tetramethyl-5,15-di(3,5-di-t-butylphenyl)porphyrin
and its free base behavior of the fluorescence. In the zinc derivative the S2 fluorescence
decays in 150 fs to vibrationally excited states of the S1 state, which rises with the same time
constant and subsequently undergoes vibrational relaxation within the S1 state in 600 fs [44].
For zinc(II) 5,10,15,20-tetra(sulfonatophenyl) porphyrin (ZnTSPP), the internal conversion
decays with a lifetime of 1.3 ps in aqueous phosphate buffer (pH=7) to give transient
absorption of the S1 state [45].
3.2.3. Deactivation of the S2 State of Porphyrin Array Systems
3.2.3.1. Zinc Porphyrin Polypeptides

Solladié et al. [46] have already synthesized L-lysine-based polypeptides bearing
porphyrin pendants (porphyrin polypeptides), in which the central metals can be varied. The
porphyrin polypeptides are expected to form a helical structure, in which the porphyrin units
are aligned in a favorable conformation for the energy migration. The singlet energy
migration process in the free-base porphyrin (H2TPP) polypeptides is reported from the
results of transient absorption methods [47] while the free-base porphyrin polypeptides does
not show S2 emission.
Fujitsuka et al. reported the S2-fluorescence properties of the ZnTPP polypeptides (ZP1,
ZP2, ZP4, and ZP8, Scheme 2) based on fluorescence up-conversion techniques [48]. The S2
state properties dependent on the polymerization degree and the singlet energy migration
were observed.
The absorption spectrum of ZP8 showed peaks at 429.3 nm (B band, λabs(B)) and 556.2
and 600.0 nm (Q bands, λabs(Q)) as shown in Figure 6. It was revealed that the peak position
of the B band shifted to the longer wavelength side with the number of the porphyrin unit in
the polypeptides. In addition, B band showed broadening. The distance between the porphyrin
units is expected to become shorter with the degree of the polymerization. Thus, the red shift
and broadening of the B band can be attributed to the exciton coupling of the transition dipole
moment of the porphyrin units. From the observed shift and broadening, the dipole moments
of the porphyrin rings are considered to be in the oblique configuration. The oblique
configuration of porphyrins due to α-helix structure of the L-lysine backbone is indicated [47]
supporting by the induced CD pattern.
O H
N
O BOC
n
(CH2)4
HN
O
(H2C)3
O
t t
Bu Bu
N N
Zn
t
N N t
Bu Bu
n = 1, 2, 4, 8
t t
Bu Bu
Scheme 2. Molecular structure of zinc porphyrin polypeptides (ZPn, n = 1, 2, 4, 8).
1.0
Fluorescence Intensity
0.8
Fl. Int.
Absorbance
0.6
0.4 0.0 2.0 4.0 6.0

-12
Time / 10 s
0.2
0.0
300 400 500 600 700 800
Wavelength / nm
Figure 6. Steady state absorption (solid line) and S2-fluorescence spectra (dashed line) of ZP8 in
toluene. Inset: Decay profile of S2-fluorescence (465 nm) of ZP8. Solid and dashed lines are fitted
curve and irf. Reprint with permission form [48], Fujitsuka, M. et al. Photochem. Photobiol. A: Chem.
2007, 188, 346. © 2007, Elsevier B. V.
S2 fluorescence spectrum of ZP8 is shown in Figure 6. Both S2 and S1 fluorescence bands

show shift to the longer wavelength side in accordance with the absorption peak shift
compared with others. The lifetime of the S2 fluorescence (τ(S2)) of ZP8 was estimated to be
0.84 ps (Inset of Figure 6), which is much shorter than that of ZP1 (1.49 ps). The τ(S2)
became shorter as the polymerization degree. Furthermore, the S2 fluorescence quantum
yields of longer polypeptides are smaller than that of ZP1. This can be attributed to the
smaller oscillator strength of oligomers due to the exciton coupling.
S2
Δλf(S2)
ΔE(S2-S1) ΔE(S2-S1)
S1
Δλf(S1)
ZP1 ZPn (n = 2, 4, 8)
Figure 7. Schematic energy diagram for S2 and S1 states of ZP1 and ZPn (n = 2, 4, 8). Reprint with
permission form [48], Fujitsuka, M. et al. Photochem. Photobiol. A: Chem. 2007, 188, 346. © 2007,
Elsevier B. V.
28.0
27.6
ln(kIC)
27.2
26.8
0.70 0.72 0.74 0.76 0.78 0.80
ΔE(S2-S1) / eV
Figure 8. Relation between the S2 decay rate and S2-S1 energy gap. Reprint with permission form [48],
Fujitsuka, M. et al. Photochem. Photobiol. A: Chem. 2007, 188, 346. © 2007, Elsevier B. V.
In the presence of the exciton coupling, the internal conversion process can be attributed
to the deactivation from the lower exciton state of the S2 to the upper exciton state of the S1 as
indicated in Figure 7. The rate of the internal conversion (kIC) is usually governed by the
energy gap law, kIC ∝ exp(-βΔE) [25,26]. In the case of ZP1, the energy gap between the S2
and S1 states (ΔE(S2-S1)) can be evaluated from the state energies estimated from the
corresponding fluorescence peaks. For ZPn (n = 2, 4, 8), the energy gap between the lower
exciton state of the S2 and the upper exciton state of the S1 can be evaluated by taking the
fluorescence peak shift into the consideration as indicated in Figure 7.
The linear relation between the ln(kIC) values and the ΔE(S2-S1) values was confirmed as
shown in Figure 8. Thus, it is suggested that the present S2-S1 internal conversion of zinc
porphyrin polypeptides is governed by the energy gap law.
3.2.3.2. Directly Linked Porphyrin Arrays

Osuka and co-workers have reported synthesis of various meso, meso-linked zinc
porphyrin arrays [49]. The S2 state lifetimes of meso, meso-linked zinc porphyrin dimer,
trimer, and tetramer were largely reduced because of participation of exciton split B band as a
ladder type deactivation channel. Therefore, the exciton coupling of zinc porphyrin
polypeptides is rather smaller than the meso, meso-linked zinc porphyrin arrays. Song et al.
[50] measured the fluorescence decay profiles of zinc 5,15-di-(3,5-ditert-butylphenyl)
porphyrin monomer (Z1) and its meso, meso-linked dimer (Z2), trimer (Z3) and tetramer (Z4)
(Scheme 3) by using the fluorescence up-conversion technique. Fluorescence from the S2
state in Z1 was observed to decay with the time constant of 1.2 ps at 460 nm.
Scheme 3. Molecular structures of synthesized porphyrins monomer and arrays. Reprint with
permission form [50], Song, N. W. et al. Bull. Chem. Soc. Jpn. 2002, 75, 1023. © 2002 The Chemical
Society of Japan.
(a) (b)
Figure 9. Temporal profiles of (a) Z1 and (b) Z2 fluorescence at various probe wavelengths. Reprint
with permission form [50], Song, N. W. et al. Bull. Chem. Soc. Jpn. 2002, 75, 1023. © 2002 The
Chemical Society of Japan.
Corresponding fluorescence rise due to the internal conversion from S2 to S1 states in Z1

was also observed at 635 nm with the time constant of 1.5 ps (Figure 9). The S2 state lifetime
of Z2 was largely reduced (90 fs) due to the participation of exciton-split Soret band as a
ladder type deactivation channel [51-53]. The internal conversion rate is given as kIC ∝
exp(−βΔE) according to the energy gap law [25]. Thus the internal conversion rate in Z2
might be much increased along with the large decrease in ΔE due to the addition of ladder
state between the Soret and Q bands. This behavior was also observed in the fluorescence
decay profiles of Z3 and Z4.
Kim et al. [53] reported that the internal conversion dynamics from the upper excited
state to the lowest excited state of longer porphyrin arrays. The Soret band of Zn(II)
porphyrin has two perpendicular components of Bx and By as depicted in Scheme 4.
In a simple monomer, they are degenerate, but in a porphyrin dimer they couple
differently. In the case of Z2, only Bx transitions are parallel, and other dipole interactions
should be zero for an averaged perpendicular conformation of Z2 (Scheme 4).
Unperturbed Soret transitions observed at ca. 413 nm for all the arrays (Z2-Z128) suggest
an orthogonal conformation and vice versa. And these bands become broader as the porphyrin
arrays become longer, indicating the increasing conformational heterogeneities caused by the
dihedral angle distribution, aggregate formation, and partial insolubility, if any, of longer
arrays. Transitions are allowed to the lower energy of the two Bx states and the two
unperturbed transitions By and Bz. Thus, the Soret band of Z2 is split into a red-shifted Bx
component and unperturbed By, Bz components (Scheme 4).
On the other hand, formation of the emitting state occurs on a faster time scale for the
porphyrin arrays than the constituent monomeric Z1. In the porphyrin arrays, the splitting of
the component levels of the S2 manifold establishes a band of intermediate levels between the
state accessed by 400 nm excitation and the lowest emitting state.
This provides a “ladder” for sequential relaxations between successive pairs of levels that
are separated by energies much smaller than the S2-S1 gap of conventional porphyrins
monomers and less strongly coupled porphyrin arrays (typically 7000 - 10 000 cm-1).
Scheme 4. Reprint with permission form [53], Kim, Y. H. et al. J. Am. Chem. Soc. 2001, 123, 76. ©
2001 American Chemical Society.
This facile internal conversion process for the porphyrin arrays as compared with Z1
monomer may be facilitated by improved Franck-Condon factors, with respect to the

monomeric building blocks and/or from the fact that the highly split S2 levels for the
porphyrin arrays provide a large number of closely spaced intermediate states that enhance
this process. For the quantitative analysis of the energy gap dependence on the internal
conversion rates for the porphyrins arrays, they plotted the S2→S1 internal conversion rates
(~1.6 ps for Z1, ~0.3 ps for Z2, ~0.5 ps for Z3, and ~0.7 ps for Z4 and the limiting values of
~0.7 ps for longer oligomers) against the energy gap between the high-energy monomeric
Soret band (S2) and the low-energy exciton split Soret band (S2’), because the S2’→S1
internal conversion is believed to be much faster than the S2→S2’ process (Figure 10). The
plot gives rise to an excellent linear correlation between the S2-S2’ energy gap and the
formation time of the lowest emissive S1 state. The rate-determining step in the overall
internal conversion processes upon photoexcitation to the S2 states in the porphyrin arrays is
likely to be the S2→S2’ internal conversion process. This argument can also be rationalized
on the basis of the fluorescence polarization anisotropy measurement of the porphyrins arrays
which shows that the S2’ states are the excitonic states being similar to the S1 states in the
transition dipole orientation in contrast with the monomeric localized electronic feature of the
S2 states. Hence the electronic coupling between S2’ and S1 states is suggested to be much
stronger as compared with that between S2 and S2’ states.
Figure 10. Plot of log(internal conversion rate) vs the energy gap between the high-energy Soret band
and the low-energy exciton split Soret band (S2 - S2’) for porphyrin arrays. As for monomer Z1, the
energy gap between the Soret and Q-bands (S2 - S1) was used. Reprint with permission form [53], Kim,
Y. H. et al. J. Am. Chem. Soc. 2001, 123, 76. © 2001 American Chemical Society.
As indicated in this section, for porphyrin polypeptides, the lifetime of S2 fluorescence

became shorter with the polymerization degree. This phenomenon can be explained by
applying the energy gap law to the S2–S1 internal conversion. The annihilation rate became
faster with the polymerization degree because of the close distance between the zinc
porphyrin units.
Therefore, the exciton coupling among the porphyrin units plays an important role in the
deactivation process of the S2 and S1 states. On the other hand, the molecular design of
directly meso-meso-linked porphyrin arrays as a new model of light harvesting antenna as
well as a molecular photonic wire was envisaged to bring the porphyrin units closer for rapid
energy transfer.
The directly linked porphyrin arrays up to 128 porphyrin units connected together
linearly were successfully isolated, and their photophysical properties depending upon the
length of the arrays were investigated by ultrafast fluorescence decay measurements. The rate
of S2→S1 internal conversion process seems to follow the energy gap law, which is mainly
determined by the energy gap between the two Soret bands of the porphyrin arrays. Overall,
the regularly arranged porphyrin arrays with ample electronic interactions will be promising
as a light-harvesting photonic wire by transmitting singlet excitation energy rapidly over the
array.
4. ENERGY MIGRATION ON PORPHYRIN POLYMERS IN THE S2 STATE

Inspired by the natural light-harvesting system, various kinds of porphyrin arrays, in
which fast energy migration for light harvesting is realized, have been demonstrated. For S1
energy migration in these natural and artificial porphyrin array systems have been well
investigated. On the other hand, contribution of the S2 state of chlorophyll in the natural
photosynthesis system is also well known [54,55]. Here, we show an example revealing the
S2 energy migration in the artificial porphyrin array system.
4.1. Cyclic Porphyrin Arrays
Cho et. al. [56] have measured energy-transfer and ultrafast internal conversion processes
in the cyclic porphyrin arrays mimicking the natural photosynthetic light-harvesting antennas
by utilizing anisotropy decay of S1 and S2 fluorescence. In the porphyrin arrays, a pair of
porphyrin moieties is joined together via triphenyl linkage to ensure cyclic and rigid
structures such as shown in scheme 5.
Time-resolved fluorescence anisotropy measurement can give detailed information on the
dynamics of intermolecular and intramolecular excitation energy transfer processes. When an
initial excitation is transferred to a state whose transition dipole moment is different in
direction from that of the initial state, the anisotropy decays according to the transfer rate.
Especially, the time-resolved anisotropy measurement in a multichromophore system is
useful in studying the excitation energy transfer processes. For fluorescence from S1 state,
energy migration process was confirmed and evaluated hopping rates are listed in Table 1. It
was revealed that the rate of energy hopping of S1 energy is on the order of several ten
picoseconds.
The anisotropy decay of S2-fluorescence is shown in Figure 11. For the arrays, the
anisotropy decay arises from the interporphyrin excitation energy transfer as well as the
electronic dephasing between Bx and By dipoles of the monomer unit.
Scheme 5. Structures of the o-Dimer, m-Trimer, and Hexamer Optimized by Semiempirical PM3
Hamiltonian along with the Excitation Energy Hopping Process. Reprint with permission form [56],
Cho, H. S. et al. J. Am. Chem. Soc. 2003, 125, 5851. © 2003 American Chemical Society.
Figure 11. Anisotropy decay profiles of the B-band emission following 405-nm photoexcitation of the
monomer, o-dimer, m-trimer, and hexamer from top to bottom. See Table 1 for the details.
The anisotropy decays of the m-trimer and hexamer can be expressed by three
components: τmono(1), τmono(2), and τhop/3. The estimated hopping rates of the porphyrins arrays
are listed in Table 1. The hopping times of S2 energy between the porphyrin units in the o-
dimer and hexamer were estimated to be ~100 and ~220 fs and that in the m-trimer was ~450
fs. The anisotropy decay rates of the arrays can be qualitatively well described by the sum of
the energy hopping rates between porphyrin moieties and a considerable portion of
dephasing/depopulation rates within the monomer, which indicates that the electronic
excitation energy is localized initially on one porphyrin unit.
Anisotropy decay times of the porphyrin arrays can be well described by the Förster
incoherent excitation hopping process between the porphyrin units. Exciton coupling
strengths of 74 and 264 cm-1 for the m-trimer and hexamer estimated from the observed
excitation energy hopping rates are close to those of B800 and B850, respectively, in the LH2
bacterial light-harvesting antenna. Thus, these cyclic porphyrin array systems have proven to
be useful in understanding energy migration processes in a relatively weak interaction regime
in light of the similarity in overall structures and constituent chromophores to natural light-
harvesting arrays.
Table 1. Fitting Results for the Anisotropy Decay, r(t), and Estimated Hopping Time
sample Time r( ∞ ) τETa τET

constant Förster b
monomer 75, 380 fsc 0.11
o-dimer B-band 100 fsd 0.08 100 fs 172 fs
Q-band 60 ps (580 60 ps 78 ps
m-trimer B-band ps)e 150 fsd 0.04 450 fs 6.7 ps
Q-band 90 ps (800 270 ps 3.1 ns
hexamer B-band ps)e 0.04 220 fs (125 246 fs
Q-band 73 fsd fs) 111 ps
60 ps (2.8 180 ps (130
ns)e ps)
a
The energy-transfer times in m-trimer and hexamer were obtained by 3 times the anisotropy decay
times. b The time constants were calculated based on the Förster equation. See the text for the
details. c The time constants from a biexponential fit. d The time constants in the B states of o-
dimer, m-trimer, and hexamer were obtained by the function, in the form of A1rm(t) + A2 exp(-t/τ).
Only the τ values excluding monomer anisotropy decay are shown here. e The values in
parentheses are second-exponential time constants, which are assigned to the rotational diffusion
processes.
In summary, from the analysis of anisotropy decay of S2 fluorescence, Cho et al.

successfully revealed the energy migration in cyclic porphyrin arrays. The energy hopping
rate of S2 energy is quite fast compared to the S1 energy hopping because of strong coupling
among chromophores due to large B band of porphyrins. Interestingly estimated coupling
between S2 excited porphyrins is equivalent to those of S1 excited bacteriochlorophylls, which
are components of natural light harvesting complex and exhibit large Q band. An example of
S2 energy migration is limited, but based on the strong B band of porphyrins, which allows
strong coupling between porphyrins, ultrafast energy transport systems will be realized.
5. ELECTRON TRANSFER IN THE S2 STATE OF PORPHYRIN DYADS

Reports on the charge separation via the higher excited states are rather limited, because
of quite short lifetime of these intermediates. It is well known that the electron donating and
accepting character of porphyrins changes according to the central metal. In the case of Zn-
porphyrin, S2-excited Zn-porphyrin acts as an electron donor. As discussed several times in
the previous sections in this chapter, intermolecular electron transfer from Zn-
tetraphenylporphyrin (ZnTPP) to solvent has been reported by Chosrowjan et al. [35] Some
intramolecular CS systems using ZnTPP have been also reported [41,45]. Mataga et al. have
reported the systematic studies on the intramolecular CS process of S2-excited Zn-porphyrin
derivatives, in which an electron acceptor is substituted at the meso-position of the porphyrin
unit [57-60]. Hayes et al. revealed that S2 electron transfer rate in ZnTPP-pyromellitimide
(PI) dyads enhanced when PI is connected at meso position compared to the dyad connected
at β-pyrrole [61]. On the other hand, very recently, Fujitsuka at el. found the electron transfer
between axial ligand and excited porphyrins in S2 state. This section summarizes the recent
works on some electron transfer between the axial-ligand and porphyrin macrocycle in S2
state. In addition, salient earlier results will be also discussed to provide necessary
information on the S2-state.
5.1. Directly Linked Porphyrin-Imide Dyads
Mataga et al. [57-60] employed a series of directly linked zinc (II)-porphyrin (ZP)-imide
(I) dyads as shown in scheme 6. The direct linkage ensures the electron donor-acceptor (D-A)
electronic interaction strong enough to realize the rapid charge separation (CS) reaction in
competition with the S2→S1 internal conversion. It should be noted here that the Soret bands
of these ZP-I dyads and the excitation energies ΔE(S2) ≈ 2.9 eV are practically the same
throughout the series, being close to that of ZP itself. Somewhat reduced D-A electronic
coupling may be ascribed to the nearly perpendicular conformation of the I moiety as
revealed by X-ray crystallography and MO calculation. The LUMOs of the attached I
moieties commonly have a node at the connecting meso-positions, thereby minimizing the
electronic interactions.
Mataga et al. have examined in detail the solvent polarity effect on the kCS ~ -ΔGCS
energy gap law in various solvents of different polarities for the CS reaction from the S2 state
of the ZP-I series. In accordance with the interpretations associated with the reaction
coordinate diagram of Figure 12, the energy gap law in nonpolar solvent shows clearly the
characteristics of the less prominent normal region and a more prominent inverted region with
gentle slope extending over a wider range of the -ΔGCS values (Figure 12(a)) compared with
the more typical bell-shaped energy gap law in strongly polar CH3CN solution with more
prominent inverted region with more steep slopes within a little narrower range of -ΔGCS
values (Figure 12(b)).
Scheme 6.
Figure 12. Schematic reaction coordinate diagrams for the crossings between free energy surfaces for
ZP*(S2)-I, ZP*(S1)-I and ZP+-I- states: (a) in nonpolar solvents and (b) in polar solvents. Reprint with
permission form [59], Mataga. et al. Chem. Phys. 2003, 295, 215. © 2005 Elsevier B.V.
This result indicates that the energy gap law in nonpolar solvent somewhat approaches
that of the intramolecular radiationless transition in the weak coupling limit [62].
The kCS ~ -ΔGCS energy gap law of ZP-I series shows less prominent normal region and
more prominent inverted region with more gentle slope extending over wider range of the -
ΔGCS values in nonpolar or less polar solvents like MCH and toluene compared with more
prominent normal region and less prominent inverted region with a little narrower range of -
ΔGCS values in strongly polar or a little polar solvents like CH3CN and THF, as indicated in
Figure 13.
Figure 13. Observed kCS vs. -ΔGCS relations in various solvents: methylcyclohexane (○), toluene ( ),
tetrahydorfuran (□), ethanol (●) and CH3CN ( ). Reprint with permission form [59], Mataga. et al.
Chem. Phys. 2003, 295, 215. © 2005 Elsevier B.V.
It should be noted here that the εS value of ethanol is only a little smaller than that of
CH3CN but much larger than that of THF. Nevertheless, the slope in normal region in ethanol
solution seems to be much more gentle than those in THF and CH3CN solutions and further,
more gentle than those in MCH and toluene solutions. We should invoke here another kind of
mechanisms to give a consistent interpretation for the observed kCS ~ -ΔGCS energy gap law in
ethanol solution. In the ethanol solutions of ZP-I, solvent molecules surrounding the solute
can make hydrogen bonding interactions with imide group, which seems to increase the
electron affinity of the imide group enhancing the -ΔGCS values, shifting the energy gap law
in the normal region near to the top part and also the increase of the electron affinity of the
imide part probably a little enhances the charge transfer interaction between ZP and I group,
contributing to the increase of the CS rate constant.
5.2. Electron Transfer from Axial Ligand of Porphyrins in the S2 State
5.2.1. Electron Transfer from Axial Ligand to Excited Sb-Porphyrins in the S2 State
Porphyrin including SbV as a central metal exhibits an electron acceptor-ability.
Shiragami et al. [63] have showed the intermolecular electron transfer to the S1- and T1-
excited Sb-tetraphenylporphyrin (SbTPP) derivative from the electron donor. Fujitsuka et al.
investigated the S2-state properties of SbTPP derivatives (Scheme 7) and reported SbTPP
derivatives exhibits electron acceptor-ability in the higher excited state. The S2-fluorescence
from SbTPP was observed for the first time [20]. Furthermore, CS from the ligand to the S2-
excited SbTPP was confirmed.
Figure 14 shows an absorption spectrum of 5 in acetonitrile. The peak attributable to the
B band was observed at 420 nm, while the Q bands were at 553 and 594 nm. Upon excitation
of the B band at 405 nm, fluorescence peak attributable to the S2 → S0 transition appeared at
425 nm with a shoulder at 446 nm. This is the first observation of the S2-fluorescence of
SbTPP [20]. The S2-fluorescence lifetime was estimated to be 1.8 ps (Inset of Figure 14). The
S2-fluorescence was observed for all the compounds investigated (1 - 6). The lifetime is
slightly shorter than that reported for ZnTPP (3.5 ps in acetonitrile) [35]. As seen in Table 2,
the S2-fluorescence lifetime (τS2) largely depends on the ligand of SbTPP, indicating the
additional deactivation pathway to the radiative and non-radiative processes. It has been
shown that the excited SbTPP works as an electron acceptor [51]. The lifetimes are in the
order of 1 ≈ 2 ≈ 3 < 4 < 5.
R - 1: R = NC6H4OCH3, R'=OCH3
Br
2: R = NC6H4CH3, R'=OCH3
N
N Sb N 3: R = NC6H5, R'=OCH3
N
4: R = NC6H4CF3, R'=OCH3
R' 5: R = NC6H4CN, R'=OCH3
6: R = R' = OH
Scheme 7. Molecular structures of SbTPP derivatives.
1.5
Fluor. Int.
Absorbance
Fluor. Int.
1.0
0.5 0.0 4.0

-12
8.0
Time / 10 s
*
0.0
300 400 500 600 700 800
Wavelength / nm
Figure 14. Absorption (solid line) and fluorescence spectra (dashed line) of 5 in acetonitrile. The
asterisk indicates a Raman peak of the solvent. Inset: fluorescence decay profile at 425 nm upon
excitation at 400 nm (fwhm 80 fs). Dashed line is IRF. Reprint with permission form [20], Fujitsuka,
M. et al. J. Phys. Chem. B 2006, 110, 9368. © 2006, American Chemical Society.
Table 2. S2-State Properties (λS2 and τS2), Oxidation and Reduction Potentials, Free
Energy Change for Charge Separation, Charge Separation Rates, Quantum Yields for
Charge Separation, and Charge Recombination Rates of SbTPP derivatives
λS2 / τS2 / ps Eox / V a Ered / V a −ΔGCS / eV kCS / 1012 s-1 b ΦCS b kCR / 1012 s-1
nm
1 429 0.45 0.76 - 0.57 1.56 1.7 0.77 6.8
2 427 0.37 0.78 - 0.56 1.56 2.2 0.81 5.8

3 425 0.40 1.03 - 0.55 1.34 2.0 0.80 6.1
4 427 0.89 1.23 -0.42 1.25 0.62 0.55 -c
5 425 1.8 1.23 - 0.48 1.21 0.06 0.11 -c

a
vs. SCE. b The values were estimated based on the S2 fluorescence lifetime. The lifetime of 6 in
acetonitrile (2.0 ps, λS2 = 425 nm) was used as a standard. c Charge separated state was not
observed. Reprint with permission form [17], Fujitsuka, M. et al. J. Phys. Chem. B 2006, 110,
9368. © 2006, American Chemical Society.
The free energy change of CS (ΔGCS) was estimated by assuming CS from the axial
ligand to SbTPP via the S2-state using following relation, GCS = Eox - Ered - ES2, where ES2 is
the S2-state energy estimated from the peak position of the S2-fluorescence. Therefore, the
measured lifetimes are in the good relation with the ΔGCS values. That is, the fluorescence
lifetime became shorter with an increase in the -ΔGCS values. The presence of the CS process
during the deactivation of the S2-state was examined by the sub-picosecond transient
absorption measurements. In order to generate the S2-state of SbTPP, the sample in
acetonitrile was excited with the femtosecond laser of 400 nm. As shown in Figure 15, an
absorption band appeared around 710 nm immediately after the laser excitation of 2 as well as
the negative absorption bands at 553 and 594 nm due to the Q bands of 2. The absorption
band at 710 nm decayed within 1 ps (Figure 15).
(a)
0.015
0.010
Δ O.D. 0.3
0.005 0.5
0.7
0.000 0.9
1.1
-0.005 1.3 ps
-0.010
550 600 650 700 750
Wavelength / nm
(b)
0.015
Δ O.D.710
0.010
0.005
0.000
0 5 10 15
-12
Time / 10 s
Figure 15. Transient absorption spectra (a) and kinetic trace at 710 nm (b) during the laser flash
photolysis of 2 in acetonitrile upon excitation with 400 nm femtosecond laser.
The absorption spectrum after 1 ps time-delay can be attributed to the S1-state of SbTPP.
On the other hand, the absorption band at 710 nm can be attributed to the reduced SbTPP, i.e.,
SbIVTPP from the comparison with the previous intermolecular electron transfer study on
SbTPP and amine [63]. Therefore, the appearance of SbIVTPP indicates intramolecular CS
from the ligand to SbTPP via the S2-excited state according to following equation,
[SbVTPP(OCH3)]+(S2)-NC6H5CH3 → [SbIVTPP(OCH3)]•-N•+C6H5CH3.
The estimated kCS and ΦCS values are listed in Table 2. The formation of the CS state
was also confirmed for 1 and 3 in acetonitrile. The large ΦCS values (~ 0.8 for 1 - 3) indicate
the efficient CS in spite of short lifetime of the S2 state. On the other hand, the absorption
band due to the CS state was not observed for 4 and 5 because of the smaller ΦCS values
resulted from their smaller -ΔGCS values. The estimated kCS values are in the order of 1 ≈ 2 ≈
3 > 4 > 5, while the -ΔGCS values are 1 ≈ 2 > 3 > 4 > 5. From the -ΔGCS dependence of the kCS
values, it seems that the CS processes in 1 - 5 are in the Marcus “normal” and “top region”
[64].
As shown in Figure 15 (b), the CS state disappeared within 1 ps after the formation of the
CS state. Because the CS states of 1 - 5 (1.33 - 1.71 eV relative to the ground state) are
located at lower energies compared with the S1-state of SbTPP (2.08 eV), the charge
recombination (CR) does not generate the S1-state but generates the ground state as indicated
in Figure 16. Based on this energetic consideration, kinetic trace of 2 was fitted with 5.8 ×
1012 and 2.2 × 1012 s-1 of rate constants for the appearance and disappearance of the CS state,
respectively.
S2 state
S2
kr knrS2 kCS
S1 state
hν [SbIVTPP(OCH3)] -N +C6H5CH3
kCR
[SbVTPP(OCH3)]+-NC6H5CH3
Figure 16. Schematic energy diagram for the charge separation and recombination processes of 2.
The agreement between the kCS and the disappearance rate of the CS state indicates that
the CR rate (kCR) is larger than the kCS [65]. Consequently, this is the first example indicating
that the S2-excited Sb-porphyrin acts as an electron acceptor.
5.2.2. Electron Transfer from Axial Ligand to Excited P-Porphyrins in the S2 State
Fujitsuka et al. [21] also examined the CS and CR processes in P-tetraphenylporphyrin
(PTPP) derivatives ([PTPP-(NHC6H4X)2]+Cl-, X = OCH3, CH3, H, Cl, CF3, and CN) (Scheme
8). Soret band and Q bands of 2 appeared at 423 and 549 and 590 nm, respectively. On the
other hand, those of 1b were red-shifted to 429 and 568 nm, respectively. These shifts are
attributable to the interaction with N atom of the axial ligand of PTPP. Upon excitation at 420
nm, which corresponds to the Soret band, fluorescence peak was observed at 434 nm
indicating the fluorescence from the S2 state. The lifetime of the S2 fluorescence of 2 in
CH3CN was estimated to be 1.5 ps. However, the S2 fuorescence lifetime of 1b was estimated
to be 0.16 ps. The substantial decrease of the S2 fluorescence lifetime indicates that a new
process other than internal conversion is included by introducing the axial ligands including
N-phenyl groups, which have electron donating nature. For other compounds, i.e., 1a, 1c-1f,
both S1 and S2 fluorescence bands were not observed. As the new deactivation pathway,
electron transfer from the axial ligand to porphyrin ring is expected, because the axial ligand
has an electron donating nature.
+ 1a: R = NHC6H4OCH3
R - 1b: R = NHC6H4CH3
Cl
1c: R = NHC6H5
N
N P N 1d: R = NHC6H4Cl
N 1e: R = NHC6H4CF3
R 1f: R = NHC6H4CN
2: R = OH
Scheme 8. Molecular structure of PTPP derivatives in this study.
From the energy minimized structure of 1b, PTPP has a distorted structure, because of
small radius of the included P atom. LUMO is localized on the TPP ring. HOMO is on the N-
phenyl group. HOMO-1 is on another N-phenyl group of 1b, which is energetically
degenerated with HOMO. HOMO-2 of 1b corresponds to HOMO of TPP ring. Thus, upon
excitation of TPP ring, electron transfer from HOMO of 1b, i.e. N-phenyl group, is expected.
That is, PTPP and N-phenyl ring are expected to act as an electron acceptor and a donor,
respectively, in the electron transfer.
Photoexcitation process upon excitation to the S2 state of PTPP was examined by
measuring the transient absorption spectra during the laser flash photolysis using femtosecond
laser at 400 nm. Figure 17 shows transient absorption spectra of 1b in CH3CN during the laser
flash photolysis. Upon excitation, transient absorption peak appeared at 700 nm with (0.15
ps)-1 of rate constant, with the ground state bleaching at 565 nm. The absorption band around
700 nm indicates the generation of radical anion of PTPP. Furthermore, generation rate
agreed well with the S2 fluorescence decay rate indicated above. These results indicate that
CS occurred from the S2 state (eq. 1),
(A) 0.008
0.4 ps
Δ O.D.
-0.1 ps
0.004
0.000
550 600 650 700 750

Wavelength / nm
(B) 0.008
Δ O.D.700
0.004
0.000
-0.5 0.0 0.5 1.0 1.5 2.0

-12
Time / 10 s
Figure 17. (A) Transient absorption spectra at -0.1, 0, 0.1, 0.2, 0.3, and 0.4 ps of 1b in acetonitrile
during the laser flash photolysis using 400 nm laser pulse (fwhm 100 fs). (B) Kinetic trace of ΔO.D. at
700 nm during the laser flash photolysis. Reprint with permission form [21], Fujitsuka, M. et al. J.
Phys. Chem. B 2007, 111, 10574. © 2007, American Chemical Society.
[PTPP(NHC6H5CH3)]+(S2)-NHC6H5CH3 → [PTPP(NHC6H5CH3)]•- (NHC6H5CH3)•+.

(1)
The observed electron transfer rates, i.e., kCS and kCR, are discussed on the basis of the
Marcus theory [64]. In Figure 18, the observed electron transfer rates are plotted against the
free energy changes. Usually, electron transfer rate depends on free energy change according
to eqs. (2) - (4),
π ⎛ (λ S + ΔG + mh ω ) 2 ⎞
∑m (e (S / m!)) exp⎜⎜ − ⎟
2 −S
k ET = V m
(2)
h λS k BT
2
4 λ k T ⎟
⎝ S B ⎠
⎛ 1 1 1 ⎞⎛ 1 1 ⎞ (3)
λ S = e 2 ⎜⎜ + − ⎟⎟⎜⎜ 2 − ⎟⎟
⎝ 2rD 2rA r ⎠⎝ n εS ⎠
λV . (4)
S=
hω
In eq (2), λS is the solvent reorganization energy given by eq. (3), V is the electronic
coupling, S is the electron-vibration coupling constant given by eq. (4), <ω> is the averaged
angular frequency. In eq (3), rD, rA, r, and n are donor radius, acceptor radius, center-to-center
distance, and refractive index, respectively. In eq. (4), λV is the internal reorganization
energy. From the theoretical calculation, the rD, rA, and r values were estimated to be 2.1, 4.2,
and 3.0 Å, respectively. Using these values, λS value was estimated to be 0.18 eV. In Figure
18, eq. (2) was calculated as a red line by assuming λV, V, and η<ω> are to be 0.57, 0.035,
and 0.15 eV, respectively. The calculated curve well reproduced the kCS1, kCR1, and kCR2
values.
13
10
-1
12
kET / s
10
11
10
10
10
0.0 0.5 1.0 1.5 2.0
-ΔG / eV
Figure 18. Free energy change (-ΔG) dependence of electron transfer rate (kET, i.e., kCS (filled mark) and
kCR (opened mark)) of PTPP derivatives upon excitation to S1 (circle) and S2 (triangle) states. Solid
lines were calculated using eq. (4) in text by assuming λs = 0.18 eV, V = 0.035 eV, h<ω> = 0.15 eV,
and λv = 0.57 eV for S1 excitation (red line) and 1.65 eV for S2 excitation (blue line). Reprint with
permission form [21], Fujitsuka, M. et al. J. Phys. Chem. B 2007, 111, 10574. © 2007, American
Chemical Society.
That is, CS from the S1 state is in the top region, while CR in the inverted region. On the
other hand, although the kCS2 values are located at close position to the red line, kCS2 values
become larger as -ΔGCS values, indicating that the kCS2 values are in the normal region of
another parabola. This fact suggests that the total reorganization energy for CS from the S2
state is larger than that from the S1 state, probably because of difference in molecular
structure in the S1 and S2 states, which causes difference in the λV value. The large difference
in λS value cannot be expected. In Figure 18, the blue curve was obtained by just changing the
λV value to be 1.65 eV, while other parameters are the same as the red line. The good fit
supports above consideration.
As indicated in above, the reports on the electron transfer from the S2 state of porphyrins
are limited. Mataga et al. investigated free energy change dependence of electron transfer rate
from the S2-excited Zn porphyrin derivatives, in which Zn porphyrin acts as an electron donor
[57-59]. They showed that the observed electron transfer rates were well reproduced by the
Marcus theory assuming 0.3 eV of λV, which is much smaller than the value employed in this
study. Although they have not examined CS from the S1 state of the same compounds, the λV
values for electron transfer from the various S1-excited Zn porphyrin derivatives have been
reported to be 0.3 - 0.6 eV [66], which is similar to the λV for electron transfer from the S2
state reported by Mataga et al.. Thus, the Zn porphyrin derivatives, which have a planar
structure, do not have large structural difference between the S1 and S2 states. On the other
hand, present PTPP derivatives have a distorted structure in the ground state [21]. The excited
PTPP possibly take a different structure depending on the excited state from the ground state,
in order to reduce instability in the excited state. Such structural change causes the difference
in the λV values.
5.2.3. Electron Transfer from Porphyrins to Axial Ligand in the S2 State

Fujitsuka et al. [67] have been investigated that the electron transfer reactions from the S1
and S2 states of some Zn porphyrins (ZnTPP, ZnOEP), Zn phthalocyanine (ZnPc) and Zn
naphthalocyanine (ZnNc) to the axial ligand. An axial ligand was synthesized as an acceptor,
which an asymmetric pyromellitic diimide (PI) compound has an alkyl chain and a pyridine
ring on N and N’ atoms, respectively. The pyridine ring of PI can coordinate to Zn of
tetrapyrrole macrocycles as shown in scheme 9.
When the ZnTPP-PI was excited to the S2 state by 400 nm laser pulse, the peak of PI•-
appeared obviously faster than the S1-excitation (Figure 19). This implies that the ET
occurred from the S2 state in this complex. The apparent rate of ET from the S2 state was
estimated to be (4.8 ps)-1. The CS rate constant from the S2 state was 5.5 times faster than ET
from the S1 state. Because the S2-fluorescence lifetime of ZnTPP is reported to be 2.4-3 ps
[68], the CS is not an efficient process due to the other deactivation processes such as internal
conversion. Thus the quantum yield of ET from the S2 state is as low as 0.63.
It should be stressed that CS in ZnPc-PI was also confirmed when ZnPc was excited to
the S2 state. On the other hand, ZnOEP and ZnNc did not show CS even when they were
excited to the S2 state, probably because of short S2 lifetime of ZnOEP and smaller driving
force of ZnNc-PI in S2 state. In the case of ZnOEP and ZnNc, CS was observed only from
their S1 state.
Scheme 9. Structure of ZnTPP-PI complex.

Figure 19. Time profiles of ZnTPP-PI complex in short time scale. Excitation wavelength: 400 nm
(circle) and 570 nm (square). The black and red lines were fitted curves. Reprint with permission form
[67], Harada, K. et al. J. Phys. Chem. A 2007, 111, 11430. © 2007, American Chemical Society.
In Figure 20, the ET rate from the S2 state was also plotted against driving force. It is
clear that the ET rate from the S2 state was not on the Marcus parabola for ET from the S1
state. This is a quite unique feature for ET from the S2 state. As a reason for this behavior, the
following two possibilities can be pointed out. First, ET from the S2 state was not on the
Marcus parabola for ET from the S1 state.
Figure 20. (A) The rate constants, kET, were plotted against -ΔG under the assumption that ET from the
S2 state was on another Marcus parabola from ET from the S1 state. The red curve was drawn with
larger reorganization energy λ= 2.0 eV. (B) The energy level diagram of the complex including the hot
CS state. (C) The rate constants, kET, were plotted against -ΔG under the assumption that ET from the S2
state has smaller -ΔG. Furthermore, ET from the S2 state of ZnTPP and ZnPc to PI was confirmed. ET
from the S2 state of ZnPc was observed for the first time. The ET rate from the S2 state was faster than
that from the S1 state. In the case of ZnOEP-PI and ZnNc-PI complexes, ET from the S2 state was not
observed. Reprint with permission form [67], Harada, K. et al. J. Phys. Chem. A 2007, 111, 11430. ©
2007, American Chemical Society.
To explain the ET rate from the S2 state, Marcus parabola with larger reorganization
energy should be considered. Because solvent reorganization energy may not depend on
excited state, internal reorganization energy varies to much extent. That is, when the
complexes were excited to the S2 state, their structures might change larger than S1-
excitation. By assuming λ =2.0 eV, Marcus parabola can be drawn as the red line in Figure
20A. Second, ET from the S2 state to the hot CS state will be another possible ET process
(Figure 20B). In this case, -ΔGCS value should be smaller as indicated in Figure 20C [69].
Thus, -ΔG(S2) may be overestimated. But the CR rates do not have a difference between S1-
excitation and S2-excitation. Thus, relaxation rate from the hot CS state to the CS state should
be quite fast. Th rate will be <100 fs because the transient absorption spectra attributable to
the hot CS was not observed [68]. When the plots of ET from the S2 state were shifted to the
Marcus parabola of ET from the S1 state, -ΔG(S2) was decreased to 0.6-0.9 eV and the energy
difference between the hot CS state and the CS state can be estimated 0.8-1 eV. From the
absence of corresponding electronic transition both in ZnTPP•+ and PI•-, the higher vibration
state of the complexes should be included in the CR process. To distinguish these
possibilities, detailed study on free energy dependence of the electron-transfer rates is needed.
As summarized in this section, CS processes from the S2 excited ZnTPP have been
investigated by some groups. Since ZnTPP is an electron donor in the excited state, electron
acceptor has been attached at meso, β-pyrrole, or axial ligand. In each case, CS from the S2
state is confirmed. For efficient CS from the higher excited state, relatively longer S2 lifetime
of ZnTPP seems to be essential, otherwise CS can not compete with the internal conversion,
as in the case of ZnOEP. On the other hand, SbTPP and PTPP act as electron acceptors in the
excited state. Efficient CS is also attributed to relatively long S2 state lifetimes. It should be
noted that electron transfer rates via S2 excited state could not be explained on the basis of
Marcus parabola for the electron transfer via S1 state. This observation is important
characteristics of the electron transfer via higher excited state and detailed mechanism has to
be clarified. In addition ZnPc also showed CS from the S2 state. Thus, the CS via higher
excited state will be a common phenomenon in electron transfer study.
CONCLUDING REMARKS
Photophysical and photochemical properties of porphyrin derivatives have attracted
scientists in the fields of photosynthesis, photodynamic therapy, solar energy conversion and
so on. Therefore, the photophysical and photochemical works on porphyrins have been
reviewed in detail by a lot of scientists. The most of them deal with the photophysics and
photochemistry of porphyrins in the lowest excited singlet- and triplet states because they are
main processes in many cases according to Kasha’s rule. On the other hand, the S2 state is
quite reactive state in spite of its very short lifetime. Thus, under the condition, where
ultrafast reaction dominates, new and high efficient reactions, which is not possible for the
lowest excited states, become possible. Understanding of the photophysical and
photochemical properties of S2 state in porphyrins becomes possible by using ultrafast laser
spectroscopy. Actually, recent improvements in instrumentation and related techniques such
as the fluorescence up-conversion and the transient absorption spectroscopy using the
femtosecond pulsed laser provide wide variety of information on the properties of the S2-state
of porphyrins and detailed understanding to the S2-state photophysics and photochemistry in
porphyrins is now becoming possible. As reviewed in this chapter, photoinduced processes
characteristic to the S2 state are now reported, such as CS and energy transfer. It should be
stressed that the higher excited state chemistry is also applicable to various molecules, not
only porphyrins. Furthermore, other higher excited states will also be a target for the ultrafast
laser spectroscopy. These information on the reactivities of the higher excited states will be
useful to achieve efficient photoinduced process.
REFERENCES
[1] (a) McDermott, G.; Prince, S. M.; Freer, A. A.; Hawthornthwaite-Lawless, A. M.;
Papiz, M. Z.; Cogdell, R. J.; Isaacs, N. W. Nature 1995, 374, 517. (b) Koepke, J.; Hu,
X.; Muenke, C.; Schulten, K.; Michel, H. Structure 1996, 4, 582. (c) McLuskey, K.;
Prince, S. M.; Cogdell, R. J.; Isaacs, N. W. Biochemistry 2001, 40, 8783.
[2] (a) Shrev, A. P.; Trautman, J. K.; Frank, H. A.; Owens, T. G.; Albrecht, A. C. Biochim.
Biophys. Acta 1991, 1058, 280. (b) Nagae, H.; Kakitani, T.; Katoh, T.; Mimuro, M. J.
Chem. Phys. 1993, 98, 8012.
[3] (a) Nakano, A.; Yasuda, Y.; Yamazaki, T.; Akimoto, S.; Yamazaki, I.; Miyasaka, H.;
Itaya, A.; Murakami, M.; Osuka, A. J. Phys. Chem. A 2001, 105, 4822. (b) Yeow, E. K.
L.; Steer, R. P. Phys. Chem. Chem. Phys. 2003, 5, 97.
[4] Kasha, M. Discuss. Faraday Soc. 1950, 9, 14.
[5] Turro, N.; Ramamurthy, V.; Cherry, W.; Farneth, W. Chem. Rev. 1978, 78, 125.
[6] Steer, R. P.; Ramamurthy, V. Acc. Chem. Res. 1988, 21, 380.
[7] Yeow, E. K. L.; Ziolek, M.; Karolczak, J.; Shevyakov, S. V.; Asato, A. E.;
Maciejewski, A.; Steer, R. P. J. Phys. Chem. A 2004, 108, 10980-10988.
[8] Beer, M.; Longuet-higgins, H. C. J. Chem. Phys. 1955, 23, 1390.

[9] Ippen, E. P.; Shank, C.V.; Woerner, R. L. Chem. Phys. Lett. 1977, 46, 20.
[10] Gillispie, G.D.; Lim, E.C. J. Chem. Phys. 1978, 68, 4578.
[11] Bajema, L.; Gourterman, M.; Rose, C. B. J. Mol. Spectrosc. 1971, 39, 421.
[12] Gan, Q.; Li, S.; Morlet-Savary, F.; Wang, S.; Shen, S.; Xu, H.; Yang, G. Optics Express
2005, 13, 5424.
[13] Kaneko, Y.; Nishimura, Y.; Takane, N.; Arai, T.; Sakuragi, H.; Kobayashi, N.;
Matsunaga, D.; Pac, C.; Tokumaru, K. Photochem. Photobiol. A: Chem. 1997, 106,
177.
[14] Kaneko, Y.; Nishimura, Y.; Arai, T.; Sakuragi, H.; Tokumaru, K.; Matsunaga, D.; J.
Photochem. Photobiol. A: Chem. 1995, 89, 37.
[15] Kaneko, Y.; Arai, T.; Sakuragi, H.; Tokumaru, K.; Pac. C. J. Photochem. Photobiol. A:
Chem. 1996, 97, 155.
[16] Yu, H. Z.; Baskin, J. S.; Zewail, A. H. J. Phys. Chem. A 2002, 106, 9845.
[17] Kurabayashi, Y.; Kikuchi, K.; Kokubun, H.; Kaizu, Y.; Kobayashi, H. J. Phys. Chem.
1984, 88, 1308.
[18] Tsvirko, M. P.; Stelmakh, G. F.; Pyatosin, V. E.; Solovyov, K. N.; Kachura, T. F.
Chem. Phys. Lett. 1980, 73, 80.
[19] Martarno, L. A.; Wong, C. P.; Dew-Horrocks, W. Jr.; Goncalves, A. M. P. J. Phys.
Chem. 1976, 80, 2389.
[20] Fujitsuka, M.; Cho, D. W.; Shiragami, T.; Yasuda, M.; Majima, T. J. Phys. Chem. B
2006, 110, 9368.
[21] Fujitsuka, M.; Cho, D. W.; Tojo, S.; Inoue, A.; Shiragami, T.; Yasuda, M.; Majima, T.
J. Phys. Chem. A 2007, 111, 10574.
[22] Jang, J. H.; Kim, H. J.; Kim, H.-J; Kim, C. H.; Joo, T.; Cho. D. W.; Yoon, M. Bull.
Korean Chem. Soc. 2007, 28, 1967.
[23] Gurzadyan, G. G.; Tran-Thi, T. –H.; Gustavsson, T. J. Phys. Chem.1998, 108, 385.
[24] Mataga, N.; Shibata, Y.; Chosrowjan, H.; Yoshida, N.; Osuka, A. J. Phys. Chem. B
2000, 104, 4001.

[25] (a) Englman, R.; Jortner, J. Mol. Phys. 1970, 18, 145. (b) Siebrand, W. J. Chem. Phys.,
1967, 46, 440.
[26] (c) Siebrand, W. J. Chem. Phys. 1966, 44, 4055. (d) Siebrand, W.; Williams, D. F. J.
Chem. Phys. 1968, 49, 1860.
[27] Turro, N. J. Modern Molecular Photochemistry, University Science Books, California,
1991.
[28] Kobayashi, H.; Kaizu, Y. in Porphyrins: Excited States and Dyanmics, eds. Gouterman.
M.; Rentzepis, P. M.; Straub, K. D. American Chemical Society, Washington DC,
1986, p 105.
[29] Even, U.; Magen, J.; Jortner, J. J. Chem. Phys. 1982, 77, 4384.
[30] Even, U.; Magen, J.; Jortner, J.; Friedman, J.; Levanon, H. J. Chem. Phys. 1982, 77,
4374.
[31] Velate, S.; Liu, X.; Steer, R. P. Chem. Phys. Lett. 2006, 427, 295.
[32] Lie, A.; Yeow, E. K. L.; Velate, S.; Steer, R. P Phys. Chem. Chem. Phys. 2006, 8, 1298.
[33] Kotlo, V. N.; Solovyev, K. N.; Shkirman, S. F. Izv. Akad. Nauk. SSSR Ser. Fiz. 1975,
39, 1972.
[34] Ohno, O.; Kaizu, Y.; Kobayashi, H. J. Chem. Phys. 1985, 82, 1779.
[35] Stelmakh, G. F.; Tsvirko, M. P. Opt. Spectrosc. (USSR) 1983, 55, 516.
[36] Chosrowjan, H.; Tanigichi, S.; Okada, T.; Takagi, S.; Arai, T.; Tokumaru, K. Chem.
Phys. Lett. 1995, 242, 644.
[37] Kishii, N.; Shirai, K.; Tamura, S.; Seto, J.; Tokumaru, K.; Takagi, S.; Arai, T.;
Sakuragi, H. J. Lumin. 1995, 64, 125.
[38] Vacha, M.; Machida, S.; Horie, K. J. Phys. Chem. 1995, 99, 13163.
[39] Tokumaru, K. J. Porphyrins Phthalocyanines 2001, 5, 77.
[40] Gurzadyan, G. G.; Tran-Thi, T.-H.; Gustavsson, T. Proc. SPIE Int Soc. Opt. Eng. 2000,
4060, 96.
[41] Kumble, R.; Palese, S.; Lin, V. S.-Y.; Therien, M. J.; Hochstrasser, R. M. J. Am. Chem.
Soc. 1998, 120, 11489.
[42] LeGourriérec, D.; Andersson, M.; Davidsson, J.; Mukhtar, E.; Sun, L. C.;
Hammarström, L. J. Phys. Chem. A 1999, 103, 557.
[43] Rodriguez, J.; Kirmaier, C.; Holten, D. J. Am. Chem. Soc. 1989, 111, 6500.
[44] Rodriguez, J.; Kirmaier, C.; Holten, D. J. Chem. Phys. 1991, 94, 6020.
[45] Akimoto, S.; Yamazaki, T.; Yamazaki, I.; Osuka, A. Chem. Phys. Lett. 1999, 309, 117.
[46] Anderson, M.; Davidsson, J.; Hammerstrom, L.; Korppi-Tommola, J.; Peltola, T. J.
Phys. Chem. B 1999, 103, 3258.
[47] (a) Solladié, N.: Hamel, A.; Gross, M. Tetrahedron Lett. 2000, 41, 6075, (b) Solladié,
N.: Hamel, A.; Gross, M. Chirality 2002, 13, 736, (c) Aubert, N.; Troiani, V.; Gross,
M. ; Solladié, N. Tetrahedron Lett. 2002, 43, 8405.
[48] Fujitsuka, M.; Hara, M.; Tojo, S.; Okada, A.; Troiani, V.; Solladié, N.; Majima, T. J.
Phys. Chem. B 2005, 109, 33.
[49] Fujitsuka, M.; Cho, D. W.; Solladié, N.; Troiani, V.; Qiu, H.; Majima, T. Photochem.
Photobiol. A: Chem. 2007, 188, 346.
[50] (a) Osuka, A.; Shimidzu, H. Angew. Chem., Int. Ed. Engl., 1997, 36, 135. (b) Khoury,
R. G.; Jaquinod, L.; Smith, K. M. Chem. Commun., 1997, 1057. (c) Ogawa, T.;
Nishimoto, Y.; Yoshida, N.; Ono, N.; Osuka, A. Chem. Commun., 1998, 337.
[51] Song, N. W.; Cho, H. S.; Yoon, M.-C.; Jeoung, S. C.; Yoshida, N.; Osuka, A.; Kim, D.
Bull. Chem. Soc. Jpn. 2002, 75, 1023.
[52] Cho, H. S.; Song, N. W.; Kim, Y. H.; Jeoung, S. C.; Hahn, S. J.;Kim, D.; Kim, S. K.;
Yoshida, N.; Osuka, A. J. Phys. Chem. A 2000, 104, 3287.
[53] Aratani, N.; Osuka, A.; Kim, D.; Kim, Y. H.; Jeong, D. H. Angew. Chem. 2000, 39,
1458.
[54] Kim, Y. H.; Jeong, D. H.; Kim, D.; Jeong, S. C.; Cho, H. S.; Kim, S. K.; Aratani, N.;
Osuka, A. J. Am. Chem. Soc. 2001, 123, 76.
[55] Shrev, A. P.; Trautman, J. K.; Frank, H. A.; Owens, T. G.; Albrecht, A. C. Biochim.
Biophys. Acta 1991, 1058, 280.
[56] Nagae, H.; Kakitani, T.; Katoh, T.; Mimuro, M. J. Chem. Phys. 1993, 98, 8012.
[57] Cho, H. S.; Rhee, H.; Song, J. K.; Min, C-. K.; Takase, M.; Aratani, N.; Cho, S.; Osuka,
A.; Joo, T.; Kim, D. J. Am. Chem. Soc. 2003, 125, 5851.
[58] Mataga, N.; Chosrowjan, H.; Shibata, Y.; Yoshida, N.; Osuka, A.; Kikuzawa, T.;
Okada, T. J. Am. Chem. Soc. 2001, 123, 12422.
[59] Mataga, N.; Chosrowjan, H.; Taniguchi, S.; Shibata, Y.; Yoshida, N.; Osuka, A.;
Kikuzawa, T.; Okada, T. J. Phys. Chem. A 2002, 106, 12191.
[60] Mataga, N.; Taniguchi, S.; Chosrowjan, H.; Osuka, A.; Yoshida, N. Chem. Phys. 2003,
295, 215.
[61] Mataga, N.; Chosrowjan, H.; Taniguchi, S. J. Photochem. Photobiol. C: Rew. 2005, 6,
37.
[62] Hayes, R. T.; Walsh, C. J.; Wasielewski, M. R. J. Phys. Chem. A 2004, 108, 2375.
[63] Englman, R.; Jortner, J. Mol. Phys. 1970, 18, 145.
[64] Shiragami, T.; Tanaka, K.; Andou, Y.; Tsunami, S.; Matsumoto, J.; Luo, H.; Araki, Y.;
Ito, O.; Inoue, H.; Yasuda, M. J. Photochem. Photobio. A: Chem. 2005, 170, 287.
[65] (a) Marcus, R. A. Annu. ReV. Phys. Chem. 1964, 15, 144. (b) Marcus, R. A.; Sutin, N.
Biochim. Biophys. Acta 1985, 811, 265. (c) Marcus, R. A. Angew. Chem., Int. Ed. Engl.
1993, 32, 1111.
[66] Ozeki, H.; Nomoto, A.; Ogawa, K.; Kobuke, Y.; Murakami, M.; Hosoda, K.; Ohtani,
M.; Nakashima, S.; Miyasaka, H.; Okada, T. Chem. Eur. J. 2004, 10, 6393.
[67] (a) Gaines, G. L., III; O’Neil, M. P.; Svec, W. A.; Niemczyk, M. P.; Wasielewski, M.
R. J. Am. Chem. Soc. 1991, 113, 719. (b) Asahi, T.; Ohkohchi, M.; Matsusaka, R.;
Mataga, N.; Zhang, R. P.; Osuka, A.; Maruyama, K. J. Am. Chem. Soc. 1993, 115,
5665. (c) Heitele, H.; Pöllinger, F.; Häberle, T.; Michel-Beyerle, M. E.; Staab, H. A. J.
Phys. Chem. 1994, 98, 7402. (d) DeGraziano, J. M.; Liddell, P. A.; Legett, L.; Moore,
A. L.; Moore, T. A.; Gust, D. J. Phys. Chem. 1994, 98, 1758. (e) Ha¨berle, T.; Hirsh, J.;
Pöllinger, F.; Heitele, H.; Michel-Beyerle, M. E.; Anders, C.; Döhling, A.; Krieger, C.;
Rückemann, A.; Staab, H. A. J. Phys. Chem.1996, 100, 18269. (f) Osuka, A.; Noya, G.;
Taniguchi, S.; Okada, T.; Nishimura, Y.; Yamazaki, I.; Mataga, N. Chem. Eur. J. 2000,
6, 33.
[68] Harada, K.; Fujitsuka, M.; Sugimoto, A.; Majima, T. J. Phys. Chem. A 2007, 111,
11430-11436.
[69] El-Khouly, M. E.; Rogers, L. M.; Zandler, M. E.; Suresh, G.; Fujitsuka, M.; Ito, O.;
D’Souza, F. Chem. Phys. Chem. 2003, 4, 474.
[70] Beckers, E. H. A.; Meskers, S. C. J.; Schenning, A. P. H. J.; Chen, Z.; Wurthner, F.;
Janssen, R. A. J. J. Phys. Chem. A 2004, 108, 6933.
Chapter 11
FEMTOSECOND EXCITED-STATE ULTRAFAST

DYNAMICS OF COMPLEX MOLECULAR SYSTEMS:
SEMICLASSICAL DYNAMICS SIMULATIONS
∗
Guang-Jiu Zhao , Yu-Hui Li and Ke-Li Han*
State Key Laboratory of Molecular Reaction Dynamics, Dalian Institute of Chemical
Physics, Chinese Academy of Sciences, Dalian 116023, Liaoning, P. R. China
ABSTRACT
In past years, the femtosecond lasers have been widely used to study the time-
dependent phenomena in the fields of physics, chemistry, biology, and material sciences.
At the same time, more and more experimental techniques and theoretical methods are
developed to investigate the ultrafast dynamics in the electronic excited states of complex
molecular systems. Nowadays, the time-resolved ultrafast spectroscopy, excited-state

quantum chemical calculations, and excited-state dynamics simulations have been
versatile tools for the study of the electronic excited-state ultrafast dynamics of complex
molecular systems. In this chapter, we have reviewed our recent semiclassical excited-
state dynamics simulations results of some complex molecular systems. The detailed
changes for molecular conformations and electronic structures in the femtosecond
excited-state ultrafast dynamics of Tetraphenylethylene (TPE) and 9,9’-bianthryl (BA)
molecules are dynamically simulated. Moreover, the dynamics simulations results are
compared with the femtosecond time-resolved spectroscopic results.
Keywords: Femtosecond lasers; Semiclassical dynamics simulation; Ultrafast processes;

Excited-state dynamics; Hydrogen bonding; Charge transfer; Photoisomerization.
∗
Corresponding author. Tel: +86-411-84379293; Fax: +86-411-84675584; E-mail: gjzhao@dicp.ac.cn (G.-J.
Zhao); klhan@dicp.ac.cn (K.-L. Han); Group Homepage: http://beam.dicp.ac.cn
370 Guang-Jiu Zhao, Yu-Hui Li and Ke-Li Han
1. INTRODUCTION
Photochemistry of complex molecular systems has been widely studied using modern
experimental and theoretical techniques [1]. The time-resolved ultrafast spectroscopy,
quantum chemical calculations for excited states, and excited-state dynamics simulations
have been versatile tools for the study on the electronic excited-state ultrafast dynamics of
complex molecular systems. It is very valuable to combine time-resolved spectroscopic
experiments with excited-state quantum chemistry calculations and dynamics simulations.
However, one can note that it is difficult to do electronic excited-state dynamics simulations
on complex molecular systems.
The electronically excited molecular systems, which are prepared by absorbing UV/vis
light, can be deactivated to the ground state through radiative and/or nonradiative channels
[2]. The complex molecular systems in the singlet excited state can directely emit
fluorescence [3,4], or emit phosphorescence from the triplet state after intersystem crossing
(ISC) [5-7]. At the same time, the radiative deactivations are competed with several important
nonradiative proceses, such as internal conversion (IC) [8-12], vibrational relaxation (VR)
[13-16], photoisomerization [17-20], excited-state electron transfer (ET) [21-29] or proton
transfer (PT) [30-33], twisted intramolecular charge transfer (TICT) [34,35], etc. All these
nonradiative processes significantly affect the fluorescence or phosphorescence emission and
are paid more and more attention in recent photochemistry research. Moreover, the dynamic
response of the solvent molecules to photoinduced changes in the charge distribution of
complex molecular systems is also of the great interest to photochemistry. The dynamic
response of the solvent molecules is crucial in determining the excited-state dynamics of the
complex molecular systems. The processes involved are generally quite complex, and their
dynamics are controlled by both the intramolecular and intermolecular interactions.
The nonradiative deactivation processes of the electronically excited state always take
palce in the ultrafast timescale (from tens of femtoseconds to hundreds of picoseconds) [36-
61]. Therefore, the femtosecond and picosecond time-resolved spectroscopies become the
good candidates to investigate the ultrafast excited-state dynamical processes. Until now,
transient absorption [36], time-resolved fluorescence [37], fluorescence up-conversion [9,38],
time-resolved infrared (IR) [39] and Raman [11,40] spectra, time-resolved photon echos
[41,42], have been the conventional spectroscopic methods to study the ultrafast excited-state
dynamics in solution. Moreover, some other techniques, such as time-resolved
multidimensional spectroscopy [43-45], time-resolved X-ray diffraction [46,47], ultrafast
electron diffraction, crystallography, and microscopy [48], have been developed and applied
widely to study the structural dynamics in real time.
Excited-state dynamics simulation has been a very useful method for understanding the
details of the femtosecond ultrafast excited-state dynamics. Moreover, several excited-state
dynamics simulation methods have been developed. Herein, a semiclassical excited-state
dynamics simulation method will be introduced.
The outline of the present review is as follows. In Section 2, we present the semiclassical
excited-state dynamics simulation methods employed in our studies. In Section 3, some recent
theoretical results about the femtosecond ultrafast excited-state dynamics of complex
molecular systems are discussed. The main conclusions are summarized in Section 4.
Femtosecond Excited-State Ultrafast Dynamics of Complex Molecular Systems 371
2. SEMICLASSICAL EXCITTED-STATE DYNAMICS SIMULATIONS

In our semiclassical dynamics simulation approach, the state of the valence electrons is
calculated by the time-dependent Schrödinger equation, but the radiation field and the motion
of the nuclei are treated classically.
The detailed description of the method has been published elsewhere [62-81] and only a
brief explanation is presented here. The one-electron wave function Ψj is expressed by a set
of localized basis functions φlα(x) [62,68]:
Ψ j (x, t ) = ∑ Ψ j (lα , t )φlα (x ), (1)

lα
with ∫d
3
xφl+α (x )H eφl 'α ' (x ) = H e (lα , l 'α '), (2)
∫d
3
xφl+α (x )φl 'α ' (x ) = S (lα , l 'α '). (3)
Here, j labels a valence electron; l labels a nucleus; α = x, y, z. Ψj is the vector with

components Ψj(lα) and H is the matrix with elements He(lα, lα). φlα(x) are the adiabatic basis
functions which move with the nuclei, so the He(lα, lα) are functions of nuclear positions Xlα.
The electron wave function is updated at each time step by solving the time-dependent
Schrödinger equation [62-70]:
∂Ψ j
ih = S −1 ⋅ H ⋅ Ψ j . (4)
∂t
The time-dependent Schrödinger equation (4) is solved using a unitary algorithm which is
based on iterative solution of the equation for the time evolution operator:
Ψ j (t + Δt ) = U(t + Δt , t ) ⋅ Ψ j (t ) (5)
U(t + Δt , t ) = (1 + (i 2h )ΔI ) (1 + (i 2h )ΔI )

−1
(6)
t + Δt
ΔI = ∫ dt ' H (t ')
t
(7)
Nonorthogonal basis functions are employed in order to properly model molecules

moving far from their equilibrium geometries [70]. This aspect is obviously important during
the breaking and forming of chemical bonds after the application of an intense laser field.
The motion of the nuclei is described by Ehrenfest’s theorem [62,70]:
1 ⎛ ∂H 1 ∂S −1 1 ∂S ⎞
F =− ∑ Ψ +j ⋅⎜⎜ − S ⋅ H − H ⋅ S −1 ⋅ ⎟⎟ ⋅ Ψ j − ∂U rep ∂X lα ,
2 j ⎝ ∂X lα 2 ∂X lα 2 ∂X lα ⎠ (8)
In these formulas, X is any nuclear coordinate, Urep is the nucleus-nucleus repulsive

potential, Uii is the repulsive interaction between nuclei and is a function of their coordinates
Xlα. S is the overlap matrix for the atomic orbitals.
The electrons are coupled to the vector potential A of the radiation field through the time-
dependent Peierls substitution [70]:
H ab (X − X') = H ab
0
(X − X') exp⎛⎜ iq A ⋅ (X − X')⎞⎟ (9)
⎝ hc ⎠
Here Hab0( X – X’) and Hab( X – X’) are the matrix elements between basis functions a
and b on atoms labeled by their instantaneous position vectors X and X’, before and after
modification by the Peierls substitution, and q = -e is the charge of the electron. The coupling
of the vector potential has been found to provide a proper treatment of the laser excitation
process, with the laser pulse having its specific properties, including wavelength, fluence, and
duration.
The Hamiltonian matrix, overlap matrix, and effective nuclear-nuclear repulsion are
calculated using density functional based tight binding (DFTB) method [79]. The DFTB
model is a semiempirical two-center approximation to the density functional theory (DFT).
Due to the limitation inherent in the DFT, typically the energy positions of the excited states
are not accurately presented in the current calculation. Specifically with the photon energy
appropriately scaled to the highest occupied molecular orbital (HOMO)–lowest unoccupied
molecular orbital (LUMO) gap, the basic conclusions, regarding the electronic and nuclear
response to a laser pulse, is quite reliable [79].
3. RESULTS AND DISCUSSIONS

3.1. Excited-State Dynamics Simulation of Tetraphenylethylene (TPE)
Photoisomerization of conjugated polyene molecules is a process of fundamental

importance in both chemistry and biology [82-113]. A representative example of this reaction
is cis-trans photoisomerization of stilbene, where a quantum yield of 0.35 was found for the
trans product [82-101]. Tetraphenylethylene (TPE), another extensively studied molecule,
preferentially reverts back to the initial cis conformation along the twisting coordinate after
decaying to the electronic ground state at the twisted geometry [102-113].
The photoisomerization of TPE has been investigated by a variety of nonlinear optical
techniques [102-105] and the most notable observations include multicomponent exponential
and nonexponential fluorescence decay kinetics in nonpolar solvents. Based on their
picosecond pump-probe study on TPE in 3-methylpentane solvent [106], Barbara et al.
proposed a three-state model, to account for the large Stokes shift in emission, the
multicomponent fluorescence, the viscosity dependence of the fluorescence lifetime and the
radiationless decay from the last state along the isomerization coordinate back to the ground
state. Their experiments provide a clear evidence of formation of at least two distinct
emissive species prior to the twisted excited singlet state [106]. These species were identified
as the nascent conformation, generated by vertical excitation of the ground state, and a
conformationally relaxed state. In his picosecond transient absorption study, Greene found
that TPE in hexane has an excited-state absorption band at about 620 nm, which decays on a
picosecond time scale [107]. Another transient absorption band at 420 nm observed in the
same measurement was assigned to a transition from the charge-resonance state, due to the
twisting of the ethylenic bond. Schilling and Hilinski studied the effect of solvent polarity on
the decay of this band and suggested that a zwitterionic rather than a charge-resonance state
was involved [108]. Picosecond optical calorimetry measurements by Morais et al. strongly
suggest the existence of a zwitterionic state along the twist coordinate [104]. Ma et al.
determined the excited state potential energy surface for the photoisomerization of TPE using
fluorescence and picosecond optical calorimetry and assigned the three states to the vertically
excited state, the conformationally relaxed state, and the twisted excited state, which they
interpreted as a zwitterionic state [105]. In a recent femtosecond pump-probe study, Wiersma
et al. found that the ethylenic bond twisting occurs at around 250 fs rather than the 5-15 ps
reported previously [109,110]. They further proposed a four-state model to explain the
excited-state dynamics of TPE. The four states that play an important role in the excited-state
dynamics include the Franck-Condon (vertically) state, the relaxed state, the charge-
resonance state, and the zwitterionic state [110]. However, later investigation by the same
group gave conflicting results regarding time scale of the ethylenic bond twisting process
[111].
In our work, we have reported the results of the first semiclassical simulation study of the
excited-state dynamics of photoisomerization of TPE. Geometry optimization and infrared
spectra calculation are performed using the B3-LYP functional and SV(P) basis set. No
constraints to bonds/angles/dihedral angles were applied and all atoms were free to optimize.
Vibrational frequencies were determined by numerical differentiation of analytical gradients

using central differences and default displacements of 0.02 Bohr. For dynamical simulations,
all nuclei are explicitly included in the calculations. The laser pulse was taken to have a full-
width at half-maximum (FWHM) duration of 60 fs, a Gaussian profile, a fluence of 0.20
kJ/m2, and a wavelength corresponding to a photon energy of 3.8 eV. The wavelength
matches the density functional energy gap between the HOMO and LUMO levels of TPE.
The fluence was chosen such that the forces on ion cores are large enough to produce a
change of geometry, but not a dissociative reaction. Simulation was performed at 300 K.
The optimized geometric structure of TPE is shown in Figure 1, where all the carbon
atoms in TPE molecule are labelled and all the hydrogen atoms are omitted. The calculated
length of ethylenic bond (C2-C3) is 1.37 Å, which is typical for a double bond. The four vinyl-
phenyl bonds, each with a length of 1.50 Å, are nearly planar. The dihedral angle of C1-C2-
C3-C4 is 11.6˚ and that of C15-C2-C3-C16 is 11.3˚. Both C1-C2-C15 and C4-C3-C16 angles are
115˚. On the other hand, all four C2-C3-C bond angles are 123˚. The dihedral angles between
each phenyl ring and the plane formed by the C2-C3 and the vinyl-phenyl bond are all about
50˚.
Figure 1. Optimized geometry of TPE (only the carbon atoms are shown here). The ethylenic bond
lengthening and twisting modes as well as the phenyl ring torsional mode are labeled. Taken from Ref
[19].
3182
1.0
IR Intensity (norm.)
0.8 725 C=C

0.6 Stretching Mode
0.4
1528
0.2 1062 1654
0.0
0 1000 2000
3000 4000
Wavenumber ( cm-1)
Figure 2. Calculated IR spectrum of TPE in ground state. Taken from Ref [19].
Some calculated infrared absorption peaks for the optimized TPE in the ground state are
shown in Figure 2. It is found that the stretching vibrational frequency is 1528 cm-1 with a
stretching vibrational period of 22 fs. In the dynamical simulation, the stretching vibrational
mode of ethylenic bond in the excited state can be distinctly studied by monitoring the change
in the length of the ethylenic bond with time. While the other strong IR vibrational absorption
peaks are mainly corresponding to the breathing mode of the benzene rings and C-H
stretching mode in benzene rings. Therefore, these vibrational modes will play less important
role on the twisting motion around the ethylenic bond compared to the stretching vibrational
mode of the ethylenic bond.
Details the excited-state dynamics of the photoisomerization in the first 1000 fs after
flashing the laser pulse are given here. The reaction from 1000 fs until the decay of the TPE
molecule to the electronic ground state is discussed in the next section. The variations with
time of the energies for the HOMO and LUMO are presented in Figure 3. It is seen from
Figure 3, following the photon excitation of the laser pulse, the energy gap between HOMO
and LUMO levels begins to decrease drastically from the initial value of around 3.8 eV and
the two orbitals are coupled at around 240 fs. At around 300 fs, the energy gap transiently
rises up to approximately 2 eV and then drops off gradually. Starting from about 600 fs until
1000 fs, two orbitals are coupled again.
The time-dependent electron population of the LUMO orbital is shown in Figure 4. The
laser pulse promotes about one electron from the HOMO to the LUMO within about 120 fs.
Then the electron population fluctuates until 300 fs due to the nonadiabatic transitions
induced by the coupling of the orbitals. A more detailed examination of the electron
populations of different molecular orbitals indicates that the decrease in the population of the
LUMO results from the nonadiabatic downward transition to the HOMO and the rise of the
electron population in the LUMO is mostly caused by the transitions from orbitals lower than
the HOMO in energy to the LUMO.
HOMO
-2
LUMO
240 fs
Energy ( eV )
-3
-4
-5
-6
-7
0 200 400 600 800 1000
Time ( fs )
Figure 3. Variation with time of HOMO and LUMO energy levels. Taken from Ref [19].
1.2
1.0
Electron Population
0.8
0.6
0.4
0.2
0.0
0 200 400 600 800 1000
Time ( fs )
Figure 4. Time-dependent population of LUMO orbital. Taken from Ref [19].
Shortly after 300 fs, the population of the LUMO is nearly unchanged until 600 fs. After
600 fs, the electron population of the LUMO fluctuated more strongly because of the
vigorous coupling of the two orbitals for this period of time.
The variation of the ethylenic double bond length with time is shown in Figure 5. By
monitoring the change of the ethylenic bond length with time, the stretching vibrational mode
of ethylenic bond in the excited state can be examined. In the early stage (< 120 fs), the
stretching vibration period of the ethylenic bond is about 20 fs. This value is close to one
calculated for the period of stretching vibration of the ethylenic bond in ground state for the
equilibrium structure, suggesting that the stretching mode of ethylenic bond has not been
strongly excited. After the laser pulse is turned off at 120 fs, the bond length increases
instantly from the initial 1.37 Å to around 2.20 Å and this length lasts until about 150 fs. It
clearly indicates that the stretching mode of ethylenic double bond is strongly excited at this
time by the electronic energy released during the nonadiabatic transitions. The calculated
period of the stretching vibration of the lengthened ethylenic bond is about 50 fs or its
frequency is about 667 cm-1. By 300 fs, the period of the stretching vibration of the ethylenic
bond is reduced to about 30 fs and the energy of the stretching mode is dissipated. The length
of ethylenic bond is only slightly fluctuated after 300 fs.
In Figure 6, we show the variation with time of some of the dihedral angles involved in
the ethylenic bond twisting process. One can find that all the dihedral angles are nearly
unchanged in the first 300 fs following the laser pulse excitation. From 300 fs to 600 fs, the
dihedral angles of C1-C2-C3-C4 and C15-C2-C3-C16 increase from 0˚ to 90˚ and the dihedral
angles of C1-C2-C3-C16 and C15-C2-C3-C4 decrease from the initial value of 180˚ to 90˚.
Consequently, the four C-C bonds which are nearly in one plane initially are twisted to be
perpendicular with each other by about 600 fs. After 600 fs, all four dihedral angles stay
about 90˚. These results of the simulation clearly show that the ethylenic bond twisting takes
place in the subpicosecond scale.
2.8
Length (Angstrom)
2.4
2.0
1.6
1.2
0 200 400 600 800 1000

Time ( fs )
Figure 5. Variation of the ethylenic bond length with time. Taken from Ref [19].
240
C1-C2-C3-C4
Dihedral Angle (degree)
210
C15-C2-C3-C16
180 C1-C2-C3-C16
150 C15-C2-C3-C4
120 C2-C3
90
60
30
0
0 200 400 600 800 1000
Time ( fs )
Figure 6. Variation with time of dihedral angles associated with ethylenic bond twisting. The variation
with time of the ethylenic bond length is also shown for comprison. Taken from Ref [19].
A large decrease in the dihedral angles of C1-C2-C3-C4 and C15-C2-C3-C16 and increase in
the dihedral angles of C1-C2-C3-C16 and C15-C2-C3-C4 start at about 300 fs when the
amplitude of the ethylenic bond drops sharply, indicating that energy transfers from the
stretching mode to the twist mode. An increase in the amplitude of the dihedral angle of C15-
C2-C3-C4 at about 300 fs further supports this conclusion.
The time scale of the twisting dynamics of the ethylenic bond, as presented above, is
consistent with the recent experimental measurements and contradicts the conclusion about
the twisting occurring in the time scale of 5-15 ps made by several groups [106-110].
Moreover, it can be concluded from above discussions that the significant lengthening of the
ethylenic bond from 150 to 300 fs is associated with the avoided crossing between the
HOMO and the LUMO, as the bond gap shrinks between the electronically excited and
ground states. Therefore the two observations reported by Wiersma et al. -- a large change in
the excited-state absorption spectrum and a rise in the stimulated emission, shifted as much as
12 000 cm-1 to the red of the excitation frequency -- probably take place during the first 250 fs
and is a direct consequence of laser excitation of the stretching mode of the ethylenic bond.
The experimental findings were explained initially in terms of an excited-state 90˚ twist,
taking place on a time scale of about 250 fs. The dynamics simulation results presented above
show that the stretching mode of the ethylenic bond is first excited to a high vibrational level
evidenced by a large increase in the length of ethylenic bond.
HOMO
-2 (a)
LUMO
-3
Energy (eV)
-4
-5
-6
-7
1.2
Electron Population
(b) LUMO
1.0
0.8
0.6
0.4
0.2
0.0
0 1000 2000 3000 4000 5000
Time ( fs )
Figure 7. (a) Variation with time of HOMO and LUMO energy levels. (b) Time-dependent population
of LUMO orbital. Taken from Ref [19].
The excited stretching mode is then relaxed to a lower energy level. The twisting motion
of the ethylenic bond is then activated by the excess energy of the stretching motion via
intramolecular vibrational energy redistribution (IVR) [106-110].
The variation with time (up to 5 ps) of the energies of the HOMO and LUMO orbitals is
shown in Figure 7(a), and the time-dependent population of the LUMO level for this period of
time is shown in Figure 7(b). It is seen from Figure 7(a) that after strong couplings from
about 600 fs and 1300 fs, the energy gap between the LUMO and HOMO levels slowly
increases before 2000 fs and remains roughly constant through 4500 fs. At about 4800 fs, the
two orbitals are coupled. This coupling leads to a remarkable transition of electrons from the
LUMO to the HOMO, as shown in Figure 7(b). This downward transition eventually brings
TPE to the electronic ground state.
Figure 8 displays the variations of the four dihedral angles involved in the twisting of the
ethylenic bond until 5 ps. It is seen that after twisting to about 90˚ before 1000 fs, the dihedral
angles strongly fluctuate about this value until about 4800 fs. At about 4800 fs, all four
dihedral angles reach about 90˚. Thereafter, the dihedral angles C1-C2-C3-C16 and C15-C2-C3-
C4 start increasing toward 180˚ while the C1-C2-C3-C4 and C15-C2-C3-C16 angles start
decreasing toward 0˚. These processes ultimately lead to the formation of cis-TPE isomer.
The extraordinary fluctuations of the dihedral angles from about 2000 fs to about 4800 fs
must be associated with the full geometric relaxation of the TPE molecule: the ethylenic bond
twisting is ultrafast so that the phenyl rings are temporally lagging behind. The large strain
between the phenyl rings and ethylenic bond forms a major driving-force for the torsional
dynamics of phenyl rings to achieve new equilibrium positions. This suggests that the
experimentally observed viscosity dependence taking place in a picosecond time scale must
be associated with the torsional dynamics of the phenyl rings. The phenyl rings’ torsional
dynamics has significant viscosity dependence as a consequence of the relatively large
conformational changes involved and the significantly lower rate compared to dynamics of
the ethylenic bond twisting.
270
C1-C2-C3-C4
Dihedral Angle (degree)
240
C15-C2-C3-C16
210
C1-C2-C3-C16
180
C15-C2-C3-C4
150
120
90
60
30
0
-30
0 1000 2000 3000 4000 5000
Time ( fs )
Figure 8. Variation of dihedral angles with time in the first 5 ps. Taken from Ref [19].
The predominantly nonadiabatic transition to electronic ground state for the

photoisomerization reaction of TPE molecule at about 4800 fs when all four dihedral angles
achieve about 90˚ suggests that the torsional motion of the phenyl rings in the electronically
excited state also plays a role in the radiationless decay of the photoexcited TPE [106]. In this
sense, the simulation results presented provide a fundamental understanding for the viscosity
dependence of the flureoscence lifetime of TPE observed by several groups.
3.2. Simulation on Excited-State Torsional Dynamics of 9,9’-bianthryl (BA)
9,9’-bianthryl (BA) in ground state with two identical anthracene moieties has D2d
symmetry (see Figure 9) and therefore possesses a vanishing dipole moment [114–125]. In
solutions, the fluctuations of solvent dipoles surrounding the excited molecule cause
symmetry breaking and initiate the formation of the charge-transfer (CT) excited-state with
unequal anthracene moieties [126–128]. Therefore, the photophysical properties of BA are
strongly dependent on the solvents.
The ground-state absorption spectrum of BA in a non-polar solvent is very similar to that
of anthracene although it is a little red-shifted as a whole. However the fluorescence spectrum
in the same solvent is much broader and red-shifted in comparison with that of anthracene
[129–137]. A one-dimensional model involving only the torsional relaxation around the bond
linking the two anthracene moieties (see Figure 10) has been proposed to understand the
spectral relaxation of BA in the non-polar solvents.
Figure 9. Ground-state structure of 9,9’-bianthryl (BA) molecule.
Figure 10. Schematic S0 and S1 torsional potentials of BA molecule as well as the perpendicular
geometric conformation in the ground state. The vertical arrow represents the transition corresponding
to excitation at the maximum absorption band. The bent arrows denote the excited-state twisting around
the central C–C bond. Taken from Ref [20].
The ground-state torsional potential has a single minimum at the perpendicular

conformation [119]. According to the Franck-Condon and band-shape analysis, the angle
between the anthracene planes in BA at the minimum of the torsional potential in the S1 state
is 62˚ or 118˚ in non-polar 2-methylbutane solution and 58˚ or 122˚ in slightly polar benzene
solution [122–126]. On the other hand, the formation of the CT state and the spectral
relaxation of BA in polar solvents can be understood by a one-dimensional model involving
only the solvent coordinate [126–128].
The torsional dynamics of BA in both electronically excited and ground states is a centre
to understand many features of photophysical properties in the non-polar solvents. Various
spectroscopic techniques have been applied to study this process [121–137]. These
investigations have revealed the complicated nature of the torsional dynamics of BA in the
electronically excited state and torsional relaxation to the tilted configuration in the S1 state.
Herein, the results of the first semiclassical simulation study of the excited-state relaxation
dynamics for BA in gas phase are reported.
In our dynamics simulations, all nuclei are explicitly included in the calculations. The
laser pulse was taken to have a full-width at half-maximum (FWHM) duration of 100 fs, a
Gaussian profile, a fluence of 0.20 kJ/m2, and a wavelength corresponding to a photon energy
of 3.2 eV. The wavelength matches the density functional energy gap between the HOMO
and LUMO levels of BA. The fluence was chosen such that the forces on ion cores are large
enough to produce a change of geometry, but not a dissociative reaction. The time-step of our
dynamic simulation is set as 20 attoseconds. Temperature is included in the simulations
through initial random velocities that are scaled so that the total kinetic energy is equal to
(3N–3)kT⁄2, where N is the number of atoms, k the Boltzmann constant, and T the
temperature. In the present work, the temperature is set as 300 K [20].
Following the photoexcitation of the femtosecond laser pulse, the excited-state dynamics
of the gas-phase BA molecule has been simulated for around 5000 fs. The variations with
time of the orbital energy levels for the HOMOs and LUMOs are presented in Figure 11. It is
seen that the energies of the HOMO and LUMO orbitals are only slightly fluctuated, while no
couplings between HOMO and LUMO orbitals occur in this time scale. Thus, no
nonadiabatic downward transition from the LUMO to the HOMO orbital can take place in the
relaxation process of the excited-state BA [20]. This is in accordance with that the S1 state of
BA in non-polar solvents or gas phase is a locally excited (LE) state and no radiationless
charge transfer (CT) process can occur in this state [119–126]. So the electronic structure of
BA molecule calculated by DFTB method is consistent with the experimental results.
It should also be noted that the molecule has D2d symmetry when BA takes the
perpendicular configuration. At this symmetry, the HOMO and HOMO–1 as well as the
LUMO and LUMO+1 are degenerated. The energy of each of two degenerated orbitals splits
when the molecule rotates about its central C–C bond away from the perpendicular structure.
When the geometry is near to the perpendicular structure, the HOMO and HOMO–1, and the
LUMO and LUMO+1 are close each other. This allows electronic transitions occurring
between two orbitals of each pair. This feature is observed three times after the laser pulse is
applied. The first one takes place at around 1000 fs and lasts nearly to 2000 fs. The other two
take place at about 3200 fs and 4500 fs, respectively. The importance of these results
indicates that BA molecule approaches the perpendicular conformation in these times and
takes the twisted conformational in other times [20].
In Figure 12, the time-dependent electron populations of the four orbitals discussed above
are shown. One can find that electron populations of these orbitals in the time regions of
orbital couplings are significantly different from that in other regions without orbital
couplings.
Figure 11. Variations of HOMO–1, HOMO, LUMO, and LUMO+1 orbital energy levels with the time.
Taken from Ref [20].
Figure 12. Time-dependent populations of the HOMO–1, HOMO, LUMO and LUMO+1 orbitals.
Taken from Ref [20].
In the time regions without orbital couplings, around one electron is located in both the
HOMO and LUMO orbitals; while the HOMO–1 orbital has about two electrons and no
electrons are transitioned into the LUMO+1 orbital. In the time regions of orbital couplings,
the electron populations of the HOMO–1 and LUMO orbitals are decreased. At the same
time, the electron populations of the HOMO and LUMO+1 orbitals are correspondingly
increased. It is noted that the total populations of the HOMO–1 and HOMO orbitals as well as
that of the LUMO and LUMO+1 orbitals remain unchanged, since no nonadiabatic downward
transition from the LUMO to the HOMO orbital can take place as discussed above [20].
Figure 13. Variations of dihedral angles of C3–C7–C15–C20, C8–C7–C15–C16, C3–C7–C15–C16 and C8–C7–
C15–C20 with time. Taken from Ref [20].
To delineate the ultrafast excited-state torsional dynamics, the variations of some dihedral
angles around the central C–C bond with time are shown in Figure 13. The real-time twisting
motion of BA molecule in the S1 state can be distinctly monitored by observing the changes
of dihedral angles formed by the two anthracene moieties. One can find that the dihedral
angles of C3–C7–C15–C20 and C8–C7–C15–C16 are nearly the same in all the times and twisted
in the range from 60˚ to 120˚; while the dihedral angles of C3–C7–C15–C16 and C8–C7–C15–
C20 are also nearly the same and twisted in the range from 240˚ to 300˚. Thus, it is distinctly
confirmed that the perpendicular BA molecule in ground state can be twisted to tilted
conformation in the S1 state within ultrafast time scale. Moreover, the twisted angles in our
simulation for the excited-state torsional dynamics are in good agreement with the
experimental values in the previous studies [122–126]. So the reliability of the torsional
dynamics simulated by our semiclassical method used in this work can be confirmed. At the
same time, it can be observed that the twisted conformation of BA molecule in the S1 state is
oscillated along the torsional coordinate between 60˚ and 120˚ (or 240˚ and 300˚). This
oscillation behavior confirms the double-well potential in the S1 state with the torsional
coordinate. Furthermore, the minima of the double-well torsional potential obtained by our
dynamic simulation are in good agreement with the values in the previous works [20].
From the Figure 13, one can see that the conformational change of BA molecule is
consistent with the change of electronic structure as discussed above. At about 500 fs, the
dihedral angles of C3–C7–C15–C20 and C8–C7–C15–C16 are twisted from 90˚ to 120˚. At the
same time, the dihedral angles of C3–C7–C15–C16 and C8–C7–C15–C20 are also twisted from
the initial value of 270˚ to 300˚. One can find that the twisted BA returns to its initial
perpendicular conformation at the end of 1000 fs. Then the two anthracene moieties remain
perpendicular with each other for about 1000 fs. As a consequence, all the dihedral angles are
nearly unchanged in the time range of 1000–2000 fs. This is the time region for the coupling
of HOMO and HOMO–1 as well as LUMO and LUMO+1 orbitals. Thus, the electronic
structures are tightly associated with the molecular conformation of BA. So it is confirmed
that the orbital couplings are induced by the molecular symmetry of the perpendicular
conformation for BA molecule.

It should be noted that the two perpendicular anthracene moieties start to twist along
inverse direction after 2000 fs. The dihedral angles of C3–C7–C15–C20 and C8–C7–C15–C16 are
decreased to the minimum value of 60˚ at about 2600 fs. At the same time, the dihedral angles
of C3–C7–C15–C16 and C8–C7–C15–C20 are decreased to 240˚. In the following around 600 fs
time scale, the BA molecule returns to the perpendicular conformation. At this time, the
HOMO and HOMO–1 orbitals as well as the LUMO and LUMO+1 orbitals are coupled
together again.
After that, the dihedral angles of C3–C7–C15–C20 and C8–C7–C15–C16 are twisted into the
range from 90˚ to 120˚. At the same time, the dihedral angles of C3–C7–C15–C16 and C8–C7–
C15–C20 return into the range from 270˚ to 300˚. The twisting motion is repeated as that in the
first 1000 fs, but with more time (around 1300 fs) to return to the perpendicular conformation.
This time delay indicates that the energy of the photoexcited BA molecule has being
significantly dissipated to other low frequency vibrations and therefore the BA molecule can
no longer have enough torsional energy to across the energy barrier into the other potential
minimum of the S1 state. Consequently, the dihedral angles of C3–C7–C15–C20 and C8–C7–
C15–C16 are fluctuated only in the range from 90˚ to 120˚ and C3–C7–C15–C16 and C8–C7–C15–
C20 are only in the range from 270˚ to 300˚. At the same time, the torsional motion of BA
molecule localized in one potential well after energy dissipation indicates the presence of a
torsional barrier between the two potential minima. Thus, the double-well trosional potential
in the S1 state for BA can be confirmed again. Moreover, it could be expected that the
twisting period will become longer following the energy dissipation of the photoexcited BA
molecule. Experimentally, Fujiwara et al. observed that this energy dissipation process can be
completed in 8 ps through the intramolecular vibrational energy redistribution (IVR) for the
BA-Ar gas complex [124]. Our simulation time up to 5 ps is not sufficient to discuss the
exacted time scale of this dynamical process as by the end of the simulation these torsional
vibration modes still show significant large amplitudes.
The variant of the C7–C15 bond length with time is shown in Figure 14. The change
tendency of the bond length is labeled by the dotted line, which is obtained by monitoring the
central positions of the vibrational bond length. It should be noted that the fluctuation of the
bond length of C7–C15 is explicitly consistent with that of dihedral angles formed by the two
anthracene moieties shown in Figure 13. When the two anthracene moieties of BA are
perpendicular with each other, the bond length of C7–C15 reaches its maximum. However,
when the conformation of BA molecule is twisted to a planar conformation, the bond length
of C7–C15 becomes shorter. This can be explained in term of the strengthened electronic
coupling between the two anthracene moieties at the planar conformation compared to that at
the perpendicular conformation [135–137]. The two anthracene moieties are electronically
decoupled when they are perpendicular with each other [137]. Thus, the bond length of C7–
C15 gets its maximum value at the perpendicular conformation. One can also find that the
period becomes longer following the energy dissipation, which is consistent with the twisting
motion.
Figure 14. Variation of the bond length of C7–C15 with time. The dotted line labels the change tendency
of the bond length. The solid arrow indicates that the amplitude of the C7–C15 stretching vibrational
mode becomes weaker and weaker with time due to the energy dissipation. Taken from Ref [20].
In addition, the amplitude of the C7–C15 stretching vibrational mode becomes weaker and
weaker with time. Therefore, it is confirmed that the energy of photoexcited BA molecule is
significantly dissipated.
SUMMARY AND CONCLUSION

In this chapter, our recent studies on the femtosecond excited-state ultrafast dynamics of
some complex molecular systems have been reviewed. The femtosecond excited-state
ultrafast dynamics of TPE and BA molecules have been investigated by the semiclassical
dynamics simulations. The detailed dynamics simulations for the excited-state process of TPE
show that following an ultrashort laser pulse, a substantial lengthening of the ethylenic bond
takes place within first 300 fs and the energy gap between the LUMO and HOMO orbitals is
drastically reduced. An avoided crossing of the LUMO and HOMO levels at about 240 fs
leads an electronic transition from the LUMO to HOMO levels, which may be associated
with the experimentally observed ultrafast dynamic Stokes shift of 12 000 cm-1 within about
250 fs. The ethylenic bond twists about 90˚ from roughly 300 fs until 600 fs and this twisted
structure remains approximately until about 4800 fs. At 4800 fs, a nonadiabatic transition to
the electronic ground state occurs. The simulations reveal that both the twisting of the
ethylenic bond and the torsional motion of the phenyl rings play important roles in the
radiationless decay of the photoexcited TPE and that the reaction proceeds to the radiationless
decay through four processes: vertically excitation, ethylenic bond lengthening, ethylenic
twisting, and torsional motion of the phenyl rings.
Moreover, the detailed simulation for the excited-state torsional dynamics of gas-phase
BA molecule in the S1 state has also been performed. The ground-state BA molecule is of a
geometric conformation with the two anthracene moieties perpendicular each other.
Following the photoexcitation of the femtosecond laser pulse, the ultrafast twisting motion
around the central C–C bond is initiated.
By real-time monitoring the dihedral angles formed by the two anthracene moieties, the
novel oscillation behavior between the two symmetric potential minima with the torsional
coordinate has been observed. The twisted angles in our simulations are between 60˚ and
120˚, which is in good agreement with the previous values. At the same time, the central C–C
bond length is shortened when the conformation of BA molecule is twisted to get close to the
planar conformation.
Furthermore, the electronic structure of BA molecule is associated with its geometric
conformation. When the twisted BA molecule returns its perpendicular conformation, the
HOMO and the lower HOMO–1 orbitals as well as the LUMO and the higher LUMO+1
orbitals are coupled with each other due to the good symmetry of the perpendicular
conformation. It has also been demonstrated that no charge transfer occurs in the S1 state of
the gas-phase BA molecule, since there has no nonadiabatic coupling for the HOMO and
LUMO orbitals in our simulation. Following the energy dissipation, the photoexcited BA
molecule can only be fluctuated between the perpendicular conformation and one of the
potential minima.
This semiclassical dynamics simulation method for the excited-state ultrafast processes is
developing in our group and will be applied to study more excited-state dynamical processes,
such as hydrogen bonding dynamics, electron transfer, vibrational relaxation, internal

conversion, twisted intramolecular charge transfer, intersystem crossing, proton transfer, etc.
for many important complex molecular systems in the future [50-81, 138-140].
ACKNOWLEDGEMENT
This work was supported by NSFC (No. 20833008) and NKBRSF (Nos. 2007CB815202
and 2009CB220010). Authors wish to thank Y. Dou for the helpful discussions.
REFERENCES
[1] C. H. J. Wells, Introduction to Molecular Photochemistry, Chapman and Hall, London
(1972).
[2] N. J. Turro, Modern Molecular Photochemistry. University Science Books: Sausalito,
CA (1991).
[3] F. Bernardi, M. Olivucci, M. A. Robb, Chem. Soc. Rev. 25 (1990) 321.
[4] J. Mack, J. R. Bolton, J. Photochem. Photobiol. A: Chem. 128 (1999) 1.
[5] T. E. Bitterwolf, J. Organomet. Chem. 689 (2004) 3939.
[6] V. Vaida, Int. J. Photoenergy 7 (2005) 61.
[7] M. Garavelli, Theo. Chem. Acc. 116 (2006) 87.
[8] A. N. Macpherson, T. Gillbro, J. Phys. Chem. A 102 (1998) 5049.
[9] A. H. Zewail, J. Phys. Chem. A 104 (2000) 5660.
[10] J. M. L. Pecourt, J. Peon, B. Kohler, J. Am. Chem. Soc. 123 (2001) 10370.
[11] F. S. Rondonuwu, Y. Watanabe, J. P. Zhang, K. Furuichi, Y. Koyama, Chem. Phys.
Lett. 357 (2002) 376.
[12] Y. Shi, J.-Y. Liu, K.-L. Han, Chem. Phys. Lett. 410 (2005) 260.
[13] J.-Y. Liu, W.-H. Fan, K.-L. Han, W.-Q. Deng, D.-L. Xu and N.-Q. Lou, J. Phys. Chem.
A 107 (2003) 10857.
[14] J. C. Owrutsky, D. Raftery, R. M. Hochstrasser, Annu. Rev. Phys. Chem. 45 (1994) 519.
[15] D. J. Nesbitt, R. W. Field, J. Phys. Chem. 100 (1996) 12735.
[16] S. K. Pal, J. Peon, B. Bagchi, A. H. Zewail, J. Phys. Chem. B 106 (2002) 12376.
[17] D. H. Waldeck, Chem. Rev. 91 (1991) 415.
[18] C. Dugave, L. Demange, Chem. Rev. 103 (2003) 2475.
[19] G.-J. Zhao, K.-L. Han, Y.-B. Lei, Y. Dou, J. Chem. Phys. 127 (2007) 094307.
[20] G.-J. Zhao, Y.-H. Liu, K.-L. Han, Y. Dou, Chem. Phys. Lett. 453 (2008) 29.
[21] M. R. Wasielewski, Chem. Rev. 92 (1992) 435.
[22] M. J. Weaver, Chem. Rev. 92 (1992) 463.
[23] R. A. Marcus, Rev. Mod. Phys. 65 (1993) 599.
[24] R. A. Bissell, A. P. de Silva, H. Q. N. Gunaratne, P. L. M. Lynch, G. E. M. Maguire, C.
P. McCoy, K. R. A. S. Sundanayake, Top. Curr. Chem. 168 (1993) 223.
[25] P. F. Barbara, T. J. Meyer, M. A. Ratner, J. Phys. Chem. 100 (1996) 13148.
[26] H. B. Gray, J. R. Winkler, Annu. Rev. Biochem. 65 (1996) 537.
[27] P. Piotrowiak, Chem. Soc. Rev. 28 (1999) 143.
[28] T.-S. Chu, Y. Zhang, K.-L. Han, Int. Rev. Phys. Chem. 25 (2006) 201.
[29] K.-L. Han, G.-Z. He, J. Photochem. Photobiol. C: Photochem. Rev. 8 (2007) 55.
[30] A. C. Legon, Chem. Soc. Rev. 22 (1993) 153.
[31] R. I. Cukier, D. G. Nocera, Annu. Rev. Biochem. 49 (1998) 337.
[32] L. M. Tolbert, K. M. Solntsev, Acc. Chem. Rev. 35 (2002) 19.
[33] K. Bhattacharyya, Acc. Chem. Rev. 36 (2003) 95.
[34] Z. R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 103 (2003) 3899.
[35] Y. Lei, S. Yuan, Y. Dou, Y. Wang, Z. Wen, J. Phys. Chem. A (2008)
10.1021/jp802483b.
[36] Y. Kimura, J. C. Alfano, P. K. Walhout, P. F. Barbara, J. Phys. Chem. 98 (1994) 3450.
[37] M. H. Gehlen, F. C. Deschryver, Chem. Rev. 93 (1993) 199.
[38] N. Mataga, Y. Shibata, H. Chosrowjan, N. Yoshida, A. Osuka, J. Phys. Chem. B 104
(2000) 4001.
[39] E. T. J. Nibbering, H. Fidder, E. Pines, Annu. Rev. Phys. Chem. 56 (2005) 337.
[40] P. Kukura, D. W. McCamant, R. A. Mathies, Annu. Rev. Phys. Chem. 58 (2007) 461.
[41] M. S. Pshenichnikov, K. Duppen, D. A. Wiersma, Phys. Rev. Lett. 74 (1995) 674.
[42] X. J. Jordanides, M. J. Lang, X. Y. Song, G. R. Fleming, J. Phys. Chem. B 103 (1999)
7995.
[43] J. C. Kirkwood, A. C. Albrecht, J. Raman Spectrosc. 31 (2000) 107.
[44] J. Bredenbeck, J. Helbing, R. Behrendt, C. Renner, L. Moroder, J. Wachtveitl, P.
Hamm, J. Phys. Chem. B 107 (2003) 8654.
[45] R. M. Hochstrasser, Proc. Natl. Acad. Sci. USA 104 (2007) 14189.
[46] C. Rischel, A. Rousse, I. Uschmann, P. A. Albouy, J. P. Geindre, P. Audebert, J. C.
Gauthier, E. Forster, J. L. Martin, A. Antonetti, Nature 390 (1997) 490.
[47] F. Schotte, M. H. Lim, T. A. Jackson, A. V. Smirnov, J. Soman, J. S. Olson, G. N.
Phillips, M. Wulff, P. A. Anfinrud, Science 300 (2003) 1944.
[48] A. H. Zewail, Annu. Rev. Phys. Chem. 57 (2006) 65.
[49] J. B. Foresman, M. Head-Gordon, J. A. Pople, J. Phys. Chem. 96 (1992) 135.
[50] G.-J. Zhao and K.-L. Han, ChemPhysChem 9 (2008) 1842.

[51] G.-J. Zhao, R.-K. Chen, M.-T. Sun, G.-Y. Li, J.-Y. Liu, Y.-L. Gao, K.-L. Han, X.-C.
Yang, and L. Sun, Chem. Eur. J. 14 (2008) 6935.
[52] R.-K. Chen, G.-J. Zhao, X.-C. Yang, X. Jiang, J.-F. Liu, H.-N. Tian, Y. Gao, K.-L. Han,
M.-T. Sun, and L. Sun, J. Mol. Struct. 876 (2008) 102.
[53] G.-J. Zhao and K.-L. Han, Biophys. J. 94 (2008) 38.
[54] G.-J. Zhao and K.-L. Han, J. Chem. Phys. 127 (2007) 024306.
[55] S.-B. Wan, S.-S. Liu, G.-J. Zhao, M.-D. Chen, K.-L. Han M.-T. Sun, Biophys. Chem.
129 (2007) 218.
[56] L.-C. Zhou, G.-J. Zhao, J.-F. Liu, K.-L. Han, Y.-K. Wu, X.-J. Peng, and M.-T. Sun, J.
Photochem. Photobiol. A: Chem. 187 (2007) 305.
[57] L.-C. Zhou, J.-Y. Liu, G.-J. Zhao, Y. Shi, X.-J. Peng and K.-L. Han, Chem. Phys. 333
(2007)179.
[58] A. Bernhardsson, R. Lindh, J. Olsen, M. Fulscher, Mol. Phys. 96 (1999) 617.
[59] J. Stalring, A. Bernhardsson, R. Lindh, Mol. Phys. 99 (2001) 103.
[60] G.-J. Zhao and K.-L. Han, J. Phys. Chem. A (2009) DOI: 10.1021/jp8083666.
[61] G.-J. Zhao, and K.-L. Han, J. Comput. Chem. 29 (2008) 2010.
[62] Y. Dou, B. R. Torralva, R. E. Allen, Chem. Phys. Lett. 378 (2003) 323.
[63] Y. Dou, B. R. Torralva, R. E. Allen, J. Phys. Chem. A 107 (2003) 8817.

[64] Y. Dou, R. E. Allen, J. Chem. Phys. 119 (2003) 10658.
[65] Y. Dou, B. R. Torralva, R. E. Allen, Chem. Phys. Lett. 392 (2004) 352.
[66] Y. Dou, R. E. Allen, J. Mod. Opt. 51 (2004) 2485.
[67] T. Dumitrica, M. E. Garcia, H. O. Jeschke, B. I . Yakobson, Phys. Rev. B 74 (2006)
193406.
[68] Y. Dou, Y. B. Lei, A. Y. Li, Z. Y. Wen, B. R. Torralva, G. V. Lo, R. E. Allen, J. Phys.
Chem. A 111 (2007) 1133.
[69] B. R. Torralva, R. E. Allen, J. Mod. Opt. 49 (2002) 593.
[70] Y. Dou, B. R. Torralva, R. E. Allen, J. Mod. Opt. 50 (2003)2615.
[71] Y. Dou, Recent Res. Devel. Physical Chem. 7 (2004) 185.
[72] T. Dumitrica, A. Burzo, Y. Dou, R. E. Allen, phys. Stat. sol. (b) 241 (2004) 2331.
[73] M. O. Scully, R. E. Allen, Y. Dou, K. T. Kapale, M. Kim, G. Chen, A. Svidzinsky,
Chem. Phys. Lett. 389 (2004) 385.
[74] B. R. Torralva, T. A. Niehaus, M. Elstner, S. Suhai, Th. Frauenheim, R. E. Allen, Phys.
Rev. B 64 (2001)153105.
[75] M. Elstner, D. Porezag, G. Jungnickel, J. Lsner, M. Haugk, Th. Frauenheim, S. Suhai,
G. Seifert, Phys. Rev. B 58 (1998) 7260.
[76] P. Sauer, R. E. Allen, J. Mod. Opt. 53 (2006) 2619.
[77] P. Sauer, Y. Rostovtsev, R. E. Allen, J. Chem. Phys. 126 (2007) 024502.
[78] P. Sauer, R. E. Allen, Chem. Phys. Lett. 434 (2007) 260.
[79] B.R. Torralva, T.A. Niehaus, M. Elstner, S. Suhai, Th. Frauenheim, R.E. Allen, Phys.
Rev. B 64 (2001) 153105.
[80] G.-J. Zhao, K.-L. Han, J. Phys. Chem. A 111 (2007) 2469.
[81] G.-J. Zhao, K.-L. Han, J. Phys. Chem. A 111 (2007) 9218.
[82] D. H. Waldeck, Chem. Rev. 91 (1991) 415.
[83] G. Orlandi, F. Zerbetto, M. Z. Zgierski, Chem. Rev. 91 (1991) 867.
[84] R. J. Sension, S. T. Repinec, A. Z. Szarka, R. M. Hochstrasser, J. Chem. Phys. 98
(1993) 6291.
[85] B. E. Kohler, Chem. Rev. 93 (1993) 41.
[86] C. Dugave, L. Demange, Chem. Rev. 103 (2003) 2475.
[87] A. A. Heikal, J. S. Baskin, L. Banares, A. H. Zewail, J. Phys. Chem. A 101 (1997) 572.
[88] J. Saltiel, J. D’Agostino, D. Megarity, L. Metts, K. R. Neuberger, M. Wrighton, O. C.
Zafiriou, Org. Photochem. 3 (1973) 1.
[89] R. J. Buenker, V. Bonanic-Koutecky, L. Pogliani, J. Chem. Phys. 73 (1980) 1836.
[90] R. W. J. Zijlstra, A. H. De Vries, P. T. Van Duijnen, Chem. Phys. 204 (1996) 439.
[91] L. V. Zhigilei, E. Leveugle, B. J. Garrison, Y. G. Yingling, M. I. Zeifman, Chem. Rev.
103 (2003) 321.
[92] C. Szakal, J. Kozole, M. F. Russo, B. J. Garrison, N. Winograd, Phys. Rev. Lett. 96
(2006) 216104.
[93] W. L. Ryan, D. J. Gordon, D. H. Levy, J. Am. Chem. Soc. 124 (2002) 6194.
[94] W. L. Ryan, D. H. Levy, J. Am. Chem. Soc. 123 (2001) 961.
[95] H. S. Lee, J. C. Light, J. Chem. Phys. 118 (2003) 3458.
[96] H. Chen, J. C. Light, J. Chem. Phys. 112 (2000) 5070.
[97] R. W. Schoenlein, L .A. Peteanula, R. A. Mathies, C. V. Shank, Science 254 (1991)
412.
[98] R. S. Judson, H. Rabitz, Phys. Rev. Lett. 68 (1992) 1500.

[99] V. Ramakrishna, M. V. Salapaka, M. Dahleh, H. Rabitz, A. Peirce, Phys. Rev. A 51
(1995) 960.
[100] M. Artamonov, T.-S. Ho, H. Rabitz, Chem. Phys. 328 (2006)147.
[101] M. Artamonov, T.-S. Ho, H. Rabitz, J. Chem. Phys. 124 (2006) 064306.
[102] D. B. Toublanc, R. W. Fessenden, A. Hitachi, J. Phys. Chem. 93 (1989) 2893.
[103] W. Schuddeboom, S. A. Jonker, J. M. Warman, W. F. Jager, B. De Lange, B. L.
Feringa, R. W. Fessenden, J. Am. Chem. Soc. 115 (1993) 3286.
[104] J. Morais, J. Ma, M. B. Zimmt, J. Phys. Chem. 95 (1991) 3885.
[105] J. Ma, G. B. Dutt, D. H. Waldeck, M. B. Zimmt, J. Am. Chem. Soc. 116 (1994) 10619.
[106] P. F. Barbara, S. D. Rand, P. M. Rentzepis, J. Am. Chem. Soc. 103 (1981) 2156.
[107] B. I. Greene, Chem. Phys. Lett. 79 (1981) 51.
[108] C. L. Schilling, E. F. Hilinski, J. Am. Chem. Soc. 110 (1988) 2296.
[109] E. Lenderink, K. Duppen, D. A. Wiersma, J. Phys. Chem. 99 (1995) 8972.
[110] R. W. J. Zijlstra, P. Th. Van Duijnen, B. L. Feringa, T. Steffen, K. Duppen, D. A.
Wiersma, J. Phys. Chem. A 101 (1997) 9828.
[111] Z. R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 103 (2003) 3899.
[112] X. M. Guo, A. D. Xia, J. Lumin. 122 (2007) 532.
[113] Z. X. Huang, D. M. Ji, A. D. Xia, F. Koberling, M. Patting, R. Erdmann, J. J. Am.
Chem. Soc. 127 (2005) 8064.
[114] A.H. Zewail, Discuss. Faraday Soc. 75 (1983) 315.
[115] J.L. Fry, S.A. Nizkorodov, M. Okumura, C.M. Roehl, J.S. Francisco, P.O. Wennberg, J.
Chem. Phys. 121 (2004) 1432.
[116] [116] L.E. Christensen, M. Okumura, J.C. Hansen, S.P. Sander, J. Phys. Chem. A 110
(2006) 6948.
[117] L.R. Khundkar, A.H. Zewail, J. Chem. Phys. 84 (1986) 1302.
[118] L.V. Zhigilei, E. Leveugle, B.J. Garrison, Y.G. Yingling, M.I. Zeifman, Chem. Rev.
103 (2003) 321.
[119] K. Elich, M. Kitazawa, T. Okada, R. Wortmann, J. Phys. Chem. A 101 (1997) 2010.
[120] K. Yamasaki, K. Arita, O. Kajimoto, K. Hara, Chem. Phys. Lett. 123 (1986) 277.
[121] A. Subaric-Leitis, Ch. Monte, A. Roggan, W. Rettig, P. Zimmermann, J. Heinze, J.
Chem. Phys. 93 (1990) 4543.
[122] N. Mataga, S. Nishikawa, T. Okada, Chem. Phys. Lett. 257 (1996) 327.
[123] T. Ishida, Y. Fujimura, T. Fujiwara, O. Kajimoto, Chem. Phys. Lett. 288 (1998) 433.
[124] T. Fujiwara, Y. Fujimura, O. Kajimoto, J. Chem. Phys. 113 (2000) 11109.
[125] J.J. Piet, W. Schuddeboom, B.R. Wegewijs, F.C. Grozema, J.M. Warman, J. Am. Chem.
Soc. 123 (2001) 5337.
[126] M. Jurczok, P. Plaza, M.M. Martin, Y.H. Meyer, W. Rettig, Chem. Phys. 253 (2000)
339.
[127] T. Fujiwara, K. Egashira, Y. Ohshima, O. Kajimoto, Phys. Chem. Chem. Phys. 2 (2000)
1365.
[128] M. Jurczok, T. Gustavsson, J.-C. Mialocq, W. Rettig, Chem. Phys. Lett. 344 (2001)
357.
[129] K. Tanaka, K. Honma, J. Phys. Chem. A 106 (2002) 1926.
[130] M. Mac, P. Kwiatkowski, U. Pischel, Chem. Phys. Lett. 357 (2002) 440.
[131] E.J.A. De Bekker, A. Pugzlys, C.A.G.O. Varma, J. Phys. Chem. A 106 (2002) 5974.
[132] K. Sakata, K. Hara, Chem. Phys. Lett. 371 (2003) 164.

[133] T. Takaya, H. Hamaguchi, H. Kuroda, K. Iwata, Chem. Phys. Lett. 399 (2004) 210.
[134] K. Sakata, Chem. Phys. 312 (2005) 69.
[135] F.C. Grozema, M. Swart, R.W.J. Zijlstra, J.J. Piet, L.D.A. Siebbeles, P.Th. van Duijnen,
J. Am. Chem. Soc. 127 (2005) 11019.
[136] T. Takaya, S. Saha, H. Hamaguchi, M. Sarkar, A. Samanta, K. Iwata, J. Phys. Chem. A
110 (2006) 4291.
[137] Z.R. Grabowski, K. Rotkiewicz, W. Rettig, Chem. Rev. 103 (2003) 3899.
[138] G.-J. Zhao, J.-Y. Liu, L.-C. Zhou, K.-L. Han, J. Phys. Chem. B 111 (2007) 8940.
[139] S. Chai, G.-J. Zhao, P. Song, S.-Q. Yang, J.-Y. Liu, K.-L. Han, Phys. Chem. Chem.
Phys. (2009) DOI:10.1039/B816589K.
[140] G.-J. Zhao and K.-L. Han, “Chapter 5: Hydrogen Bonding Effects on the
Photochemistry of Chromophores in Solution” In: Photochemistry Research Progress,
Editor: Alejandro Sánchez and Sergio J. Gutierrez, ISBN: 978-1-60456-568-3, Nova
Science Publishers, Inc. New York (2008).
Chapter 12
HIGH INTENSITY LASER-MATTER INTERACTIONS:

MEASUREMENTS OF ION AND NEUTRAL EMISSION
AND APPLICATIONS
A. Daskalova1 and W. Husinsky2

1
Institute of Electronics, Bulgarian Academy of Sciences,
Boul.Tzarigradsko Chausee 72, 1784 Sofia, Bulgaria
2
IAP, Vienna University of Technology,Wiedner Hauptstrasse 8-10,
1040 Vienna, Austria
ABSTRACT
This chapter represents an overview of the current state of the art in the field of
femtosecond technology with a special emphasis put on the research on high-intensity

laser-matter interactions. Ultrashort pulse laser ablation provides a promising approach
for processing of almost any kind of material. However, this method has been
investigated mainly by research on solid materials such as metals, semiconductors, etc.
Studies on large biological molecules are still only slightly covered. Following a brief
description of the fundamentals of femtosecond laser technology, an introduction to the
physics of the high intensity laser interaction will be provided. Thorough understanding
of heat transfer mechanism and energy conversion routes is of crucial importance for
controlling the resulting modification of the target material in many practical
applications. In the current studies femtosecond laser ablation was applied to complex
biological molecules. Measurements of TOF mass spectra obtained from the laser–tissue
interaction are presented. TOF mass spectroscopy has been applied to characterize the
species produced in the femtosecond (fs) and nanosecond (ns) regimes. These studies aim
to detect and give information on the photodissociation of these biological molecules and
to compare the two ablation regimes. The morphological changes of the material were
investigated using Environmental Scanning Electron Microscopy. The results will
demonstrate the feasibility of the USLA-TOF method as a high sensitivity analytical
technique for elemental analysis and for detection of fragile biological molecules. Some
comments on the potential novel applications of ultra-short laser radiation are presented,
namely efficient implementation in ophthalmology, dentistry, nanosurgery and innovative
trends in femtosecond technology.
394 A. Daskalova and W. Husinsky
INTRODUCTION
The progress in ultra-fast laser technology since the 1990’s (demonstration of Kerr-lens
mode-locking, CPA concept) has led to significant new insights into ultra-fast phenomena in
solid-state, chemical and biological materials as well as fundamental studies of high intensity
laser-matter interactions.
For example if a pulse with energy of 1mJ and duration 25fs is focused down to a tiny
spot with diameter of 20μm, the intensity I will be in the order of ≈1.2 × 1016W/cm2, such
high powers achievable with ultrashort pulses are suitable for precise ablation of many
materials and successful applications for micromachining, drilling, cutting and welding
[1].Ultrashort laser pulses nowadays have entered many areas of research and development.
Besides all other applications described in this chapter, femtosecond laser-tissue
interaction with high peak intensity favors a rapid plasma formation and, consequently, the
transformation of the deposited energy into material ionization and ejection rather than heat
transfer to the environment, so the heat conduction to the surrounding tissue is greatly
reduced and the sample remains essentially cold during the interaction process– “Cold
Ablation”.
This nonthermal interaction is what makes ultrafast laser systems so valuable in most
ultrafast applications. The result is a clean ablation which gives a possibility for more
efficient and precise structuring of the tissue.
1. FEMTOSECOND TECHNOLOGY
Femtosecond technology is based on the advantages of the femtosecond light pulses:
−15
energy can be concentrated in a temporal interval as short as several 10 s, which
corresponds to only a few optical cycles in the visible range, the pulse peak power can be
extremely large even at modulate pulse energies, the geometrical length of a fs pulse amounts
only to several micrometers [2].
One of the necessary conditions for generating short pulses is the presence of a
broadband gain medium in the laser cavity. In practice, solid state materials such as Titanium
sapphire are used in ultra-short laser sources. The development of Ti: sapphire systems are
accompanied by the discovery of the Kerr lens mode-locking technique [3]. At high
intensities, it is possible to introduce an intensity-dependent modification to the refractive
index n.
For a Gaussian beam profile this intensity-dependent modification of n causes self-
focussing of the beam as it passes through the cavity. This allows pulses as short as 10fs to be
produced directly from a laser. Having generated a short seed pulse, it is indispensable to
address the design consideration for the amplifier.
During the amplification, the self-focussing effect can cause big damage to the gain
medium. The goal of the CPA concept [4] is mainly to reduce the peak pulse intensity during
the amplification Figure 1.
High Intensity Laser-Matter Interactions 395
Figure 1. Representative scheme for chirped pulse amplification.
This can be done in several stages. The first stage is to stretch the seed pulse duration by
several orders of magnitude. Usually stretching is possible to be achieved with a combination
of gratings and telescope. The second stage is to amplify the pulse to the required level. And
in the final, the stretched, amplified pulse has to be recompressed again to its original
duration. In conclusion after propagating through the oscillator, amplifier, stretcher and
compressor, the pulse reaches its final state. The amplification process introduces extra
dispersion and is one of the major considerations in system design. After the pulse is formed,
two mechanisms are affecting the fs pulse length and structure—the self-phase modulation
(SPM) and the group-velocity dispersion (GVD) [5]. Group velocity dispersion causes
different wavelengths to travel through an optic at different velocities. In a broadband pulse,
this has the effect of lengthening the overall pulse envelope. Pulse lengthening due to GVD
can be compensated, by using a pulse compressor, which is an optical setup or component
that possesses GVD opposite in sign.
There are many modifications of ultrafast laser systems. Organic dyes have proven to
possess excellent properties in the generation of ultrashort pulses, however their application
in spectroscopic investigations is restricted. For measurement with femtosecond time
resolution, titanium sapphire material had acquired popularity in the late 1980’s. In
comparison with dye lasers, the titanium sapphire laser has about four times as broad a tuning
range which covers the entire fluorescence band from 680nm to 1100nm, it has also a much
smaller pump absorption cross section than organic dyes. In the structure of the sapphire
crystal the Ti3+ ion substitutes the Al3+ ion. The absorption band in the blue-green part of the
spectrum is very wide. The emission bands are shifted towards low energies (750nm). The
Ti:Al2O3 can be efficiently pumped by frequency doubled Nd lasers and Ar lasers.These
characteristics greatly affect the resonator designs for such kind of lasers. Optical parametric
oscillators were first used as an alternative to dye lasers for generating coherent radiation
tunable over a wide wavelength region. The setup of our experimental system is one of the
typical representatives of this generation of oscillators. It offers the possibility for generation
of optical pulses with high quality, stability and reproducibility. Optical pulses of sub-12 fs in
duration can be achieved directly from the mirror dispersion controlled (MDC) oscillator
Figure 2.
Figure 2. Schematic diagram of the MDC Ti:sapphire oscillator.
The MDC oscillator contains a set of chirped mirrors [6] for broadband intra-cavity
group-delay-dispersion (GDD) control in the Ti:sapphire oscillator.The presented oscillator
utilizes the Kerr lens mode-locking (KLM) technique. Self-focusing introduced by the optical
Kerr effect in the gain medium changes the confocal parameter of the cavity mode. The
mode-locking in the experimental set up is started by realizing a small perturbation of the
cavity length. After the oscillator the optimized mode-locked seed beam is directed onto
Ti:sapphire multipass amplifier. The typical pulse energy emitted from a KLM oscillator is of
the order of a few nJ. With a suitable amplifier design this energy can be improved. The
current laser system consists of two stages: amplifier and compressor Figure 3.
Figure 3. Schematic of the experimental arrangement for KLM Ti:Sapphire laser.
The Ti:sapphire crystal is pumped by frequency doubled Nd:Ylf laser. In order to ensure
a homogeneous, Gaussian pump-beam profile the output of the Nd:Ylf laser is imaged onto
the crystal with a lens. The focussed spot has about 1.5mm diameter. The last part of the
system is the compressor. It is a prism compensator. The Prisms are mounted on a translation
stage. By moving the translation stage the path can be changed and thus the overall
compressor dispersion is adjusted. This can prolong the pulses to about 100fs. The beam exits
the compressor via a periscope consisting of dispersive dielectric mirrors with a high positive
third order dispersion. With several reflections of these mirrors the final pulse duration can be
fine-tuned to below 30 fs.
2. OVERVIEW OF LASER - TISSUE INTERACTIONS

THE NATURE OF LASER ABLATION MECHANISM DIFFERS FROM UV TO IR RANGE [7]. UV
ABLATION IS believed to be predominantly a photochemical process. It is stated that the
photochemical process develops when the UV photon has an energy which covers the bond
energy of the material and is resulting in bond breaking (due to electronic excitation). Energy
of one UV photon is high enough (≈ 6.43 eV) to dissociate C=C bonds. In the case of IR
ablation energy of one photon (≈1eV) is not enough to dissociate the chemical bonds, by
absorption of several photons it triggers molecular vibrations and rotations. There can be
distinguished two cases in which the molecule can decay its excitation: radiative decay, the
molecule discards its excitation trough a photon and nonradiative decay, the excess energy is
transferred to vibration, rotation or translation of the neighboring molecules. Thus the optical
energy is transferred to heat energy.
The variety of medical applications of lasers is based on the absorption of laser light in
tissue. The absorbed amount of light by the media at certain wavelength is described by the
Beer-Lambert law equation (1):
I=I0exp(-ad) (1)
where I0 is the incident intensity, I is the transmitted intensity, a - the absorption coefficient of
the tissue and d- length of the light path. This indicates that with increasing depth the
absorbed radiation decreases in correspondence with equation (1).
The thermal effects of laser – tissue interaction are related with two other optical
properties: optical scattering and optical reflection. Optical scattering occurs when the
direction of the incident rays is changed by the molecules present in the tissue. The
consequence of the scattering is the effect which it imposes on the properties of the laser
beam, after propagating a certain length (few mm in tissue) the laser beam changes its
characteristics: the coherence and the collimation, it becomes incoherent and isotropic.
Scattering is wavelength dependent. The optical reflection, determines in what proportion the
input beam will penetrate in the tissue. This knowledge of the reflectivity is necessary
because it can reach high values, for example the argon laser beam is reflected 40%-50%
from the skin.
The classification of the different types of laser-tissue interactions can be summarized as
follows [8]:
Photochemical interaction: results when the light is absorbed by molecules present or added
to the tissue
Thermal interactions: it is an effect of deposition of energy into the tissue. They are
characterized with increase in local temperature.
Photoablation: it is described as photo-dissociation process of the molecular bonds of the
tissue.
Plasma-induced ablation: arises when applying very high power densities more than
1011W/cm2. Ablation occurred due to ionizing plasma formation.
Mechanical: usually is associated with high laser fluence rates and is related to formation of
optical breakdown of the tissue, generation of plasma and shock waves.
Typical example of photochemical interaction is the Photodynamic therapy (PDT). It is

based on injection of spectrally adapted chromophores (photosentisizer) into the body. After
irradiation with monochromatic light the photochemical reaction is triggered. The
photosentisizer performs several simultaneous decays, which result in intramolecular transfer
reaction. After being excited by absorption of photon, the drug returns to its ground state and
transfers energy to ambient oxygen, which is transformed to singlet oxygen. The singlet
oxygen is very toxic substance and oxidizes all the tissue constituents. This method is used
mainly for treating cancer cells. Another type of photochemical interaction is biostimulation.
It takes place at very low laser powers (1-5mW).The thermal interactions results from
conversion of light to heat. For thermal decomposition of tissue it is necessary to adjust the
duration of the laser pulse in order to minimize thermal damage to surrounding tissue. For
this reason thermal relaxation time τ is defined, which measures thermal susceptibility of the
tissue. When τ<1µs for nano- and picosecond pulses the heat diffusion during the laser pulse
is negligible When τ>1µs the thermally damaged zone is broader. At microscopic level,
thermal effects can be outlined as a two-step process:
Absorption: A+hν → A∗ (2)
Deactivation: A∗+M(Ekin) → A+M(Ekin+ΔEkin) (3)
Depending on the degree and duration of the heating of the tissue can be distinguished
several phases.
Coagulation – usually the coagulated tissue appears darker than the other tissue, and during
this process the temperatures reach 60°C.
Vaporization – leads to pressure increase, water tries to expand in volume as it vaporizes and
thus leads to microexplosions. The result is thermal decomposition.
Carbonization – the carbonization of tissue develops due to drastically increase in the
temperature. At temperatures above 150°C the tissue starts to carbonize leading to

blackening in color.
Melting – it is characterized with a significant increase in the temperature (few hundred
degrees Celsius).
The photoablative effect can be described as an ablation of the material without thermal
lesions. The energy of single UV photon is enough for dissociation of the former bound, the
molecular bonds are broken and the tissue components are vaporized without heat generation.
This effect is associated with lasers operating in the UV range (ArF, KrF, XeCL). The
decomposing of the material happens when exposing to high intense laser irradiation (107-
108W/cm2) in the nanosecond range pulse duration. The photoablation is a two-step process:
Step1: Excitation
AB+ hν → (AB)* (4)
Step2: Dissociation
(AB)* → A+B+Ekin (5)
To trigger the photoablation it is necessary to apply certain threshold intensity. Above

this threshold a well defined depth is ablated.
Plasma-induced ablation proceeds with well defined removal of the tissue, without
thermal or mechanical damage. The characteristic parameter for plasma-induced ablation is
the local electric field E.
1
I(r,z,t)= ε0cE2 (6)
2
If E exceeds a certain threshold (the applied electric field forces the ionization of
molecules and atoms) than the breakdown occurs. It is established that Q-switch and mode-
locked pulses in pico- and femtosecond time domain could initiate localized microplasma.
The difference is in the mechanism of generation of free electrons. By the Q-switched pulses
it is a thermoionic emission, but in the case of mode-locked pulses is multi-photon ionization
(several photons provide the energy for ionization).
The Mechanical effect leads to fragmentation and cutting of tissue by mechanical forces
(cavitation, jet formation and e.t.c). These effects take place at different time scale.
Photodisruption starts with optical breakdown. Shock wave generation starts during the
expansion of plasma then it propagates into adjacent tissue and leaves the focal volume.
Cavitation starts 50-150ns after the laser pulse and occurs when focusing the laser beam into
the tissue.
3. MASS SPECTROSCOPY OF BIOLOGICAL MOLECULES

3.1. The LSNMS Method
Previously successful efforts have been attained to examine many organic biological
molecules with matrix-assisted laser desorption ionization (MALDI) which produces intact
molecular ions. In MALDI analysis the analyte is crystallized with large molar excess of
matrix compound. The matrix serves to efficiently absorb the laser light and thus causing the
vaporization of the analyte. The method of Laser Secondary Neutral Mass Spectroscopy
(LSNMS) is established since early 1980s [9]. With the development of new class, high
power, ultra-short lasers the approach of photoionization of neutral species with ultra-short
pulses is a topic of considerable interest due to the unique ionization mechanism. In our
research we have used the post-ionization method because of the high sensitivity of the
detection and analysis of molecules and atoms. LSNMS is a technique which provides
qualitative elemental and chemical species analysis from localised areas of the surface of a
solid sample. It uses a fs laser to ablate the sample surface and time-of-flight (ToF) mass
spectrometer to detect the ionized species [10].
Three types of photon excitation can be distinguished, which lead to ionization of neutral
atom or molecule: Single Photon Ionization (SPI), Resonant Multiphoton Ionization (RMPI)
and Nonresonant Multiphoton Ionization (NRMPI) [11].
The Single Photon Ionization (SPI) means that the absorption of one photon will be
enough for promoting the electron into the ionization continuum. When the atomic system is
brought into the ionic continuum through simultaneous absorption of several photons the
ionization process is called Multiphoton Ionization. Depending of the spectral parameters of
the radiation and of the atomic states two cases of MPI can be distinguished. The Resonant
Multiphoton Ionization (RMPI) is a multi-step process where an electron is promoted to
intermediate electronic state trough absorption of a photon and the subsequent absorption of a
photon of the same energy initiates the ionization. The typical power density required for
RMPI is in the order of 106W/cm2. The neutrals are intersected by a pulsed laser beam of high
flux density of photons. The other case of MPI is the NRMPI, and can be encountered when
the multiphoton absorption occurs even when the electronic intermediate state is not excited.
Power densities exceeding 109 W/cm2are required for achieving saturation of the ion yield.
High power pulsed ultra violet lasers are used for this type of ionization. The majority of the
emitted particles (sputtered due to energetic ion bombardment or ablated, desorbed due to
interaction with laser beam) from a solid surface are neutral in charge but it also exists a small
proportion of positively and negatively species. The crucial parameter for the post-ionization
experiments is the energy necessary for ionization of the neutral species. The ionization
efficiency depends strongly from the precise timing of the laser pulses and from the
geometrical overlap of the focus of the post-ionizing laser beam with the emitted particle
cloud.
Femtosecond laser-tissue ablation [12] can be described as a multiphoton process. The
essence of the process is expressed in ionization of bond and free electrons which develops
with plasma-induced ablation and finally introduces an optical breakdown. The extremely
short duration of femtosecond pulses is the main advantage in tissue processing. With ultra-
short pulses the thermal component can be neglected because the pulse duration is smaller
than the thermal relaxation time, which for biological material is in the order of 1µs, thus the
time for heat distribution is limited and the sample remains relatively cold.
In the present experiment the technique incorporates two steps: ablation of neutral
molecules, atoms from the sample surface and post-ionization of the ablated particles with fs
laser pulses. This separate mode of exploring the method allows the ablation and the post-
ionization step to be optimized separately thus is achieved maximum sensitivity. In this study
the method is applied for analysis of polypeptides and proteins.
3.2. Experimental Setup
The idea of the method of ultra-short laser ablation time-of-flight mass spectroscopy is
expressed in the possibility for simultaneous registration of the whole mass spectrum of the
ablated species. The recorded masses appeared according the rule: ions with the same kinetic
energy will move with different velocities depending of their m/z value. After the ablation the
particles are accelerated, mass separated into the drift tube of the ToF spectrometer and
detected with a micro channel plate detector according to their mass-to-charge ratio Figure 4.
In our research we have examined the ultra-short laser ablation mass spectra of tooth
sample at two wavelengths (λ=193nm, λ=800nm) and with ion beam sputtering of 9KV
accelerated Xe ions. The experimental system is already described [13]. In brief it includes a
femtosecond laser system (Femtopower-Compact pro) Figure 5 and UHV chamber with a
base pressure of 10-8Torr equipped with xyz rotating target holder.
Figure 4. A schematic view of an experimental setup: UHV-Ultra high vacuum chamber, MCP-Micro
channel plate detector, TOF-Time-of-Flight spectrometer.
Figure 5. The table-top 1KHz Ti:sapphire laser at the Institute of General physics, Vienna University of
Technology.
The properties of this laser system are 1mJ pulse energy, 25 fs pulse duration at 800 nm
central wavelength at 1KHz repetition rate. The laser-beam has a Gaussian intensity
distribution. The laser beam is incident at an angle of 45◦ on the target and focused with a lens
(focal length of 35cm) onto the sample inside the ablation chamber to a spot of 50µm
diameter. The mass spectra are detected by means of reflectron ToF mass spectrometer. Our
aim was to gain maximal signal detection with suppressed molecular fragmentation.
3.3. Analysis and Discussion of ToF Mass Spectra
3.3.1. Composition of Biological Tissue

The aim of this work is to investigate how femtosecond laser ablation would influence
the mass removal and elemental fragmentation. Our experiments were performed on
examining biological targets such as tooth, bones. Tissue consists of hundreds of different
types of molecules (DNA, protein, water fats and etc.) distributed non-homogenously. More
specifically the main composites of hard tissue (bone and tooth) are represented by organic
and inorganic matter respectively collagen fibers and hydroxyappatite crystals. The organic
matter (collagen fibers) in the tooth and bone samples represents a considerable part of their
composition. Collagen occurs in all multicellular animals and it belongs to the family of
proteins. It is organized into insoluble fibers and is the main stress-bearing component of
connective tissues such as bone, tooth, cartilage and the fibrous matrices of skin and blood
vessels. The molecular mass of one molecule of collagen is ≈ 285kD*. The molecule length is
≈ 3000Å and it consists of three polypeptide chains wound together in a tight triple helix
Figure 6. Collagen has an amino acid composition, almost one third of it‘s residues are Gly;
another 15 to30% of its residues are Pro and 4-hydroxyproline (Hyp) (Figure 23). There is
very little Cys or Trp present. The primary structure of collagen has a frequent repetitive
pattern, Gly-Pro-Hyp (or Lys or Hyl). The proteins consist of one or several peptide chains.
The amino acid units in the protein molecule are connected trough a peptide bond. The amino
acids contain an amino group and a carboxyl group. The resulting CO-NH linkage is called
peptide bond. Proteins, which are composed of several amino acids, are called di-, tri-
…polypeptides Figure 7.
Figure 6. Example of collagen triple helix.
Figure 7. Example of polypeptide backbone.
All proteins are composed of 20 amino acids, which can be distinguished by the R -
functional group.
It is very important to have knowledge of the composition of the tissue material, in order
to make chemical and physical analysis. The acquired laser ablation ToF mass spectra
strongly depend on the target composition. It was difficult to examine the behavior of more
complex molecules in the presence of intensive laser field due to their involatility and thermal
labiality.
3.3.2. Laser Ablation Experiments

The results will be presented in two parts; the first one is laser ablation of biological
samples with femtosecond and nanosecond pulses (the primary objective will be to study the
differences in the mass spectra for femtosecond and nanosecond laser-induced ablation from
hard biological tissue) and SEM measurements and the second one is the post-ionization
results. The irradiations made with fs laser pulses on both samples prove that the interaction
of ultra-short laser pulses with bio-tissue can be possible to proceed with reduced molecular
fragmentation and limited peripheral thermal damage [14]. The time-of-flight mass spectra
for bone sample at λ=800nm, pulse duration 30fs are shown below Figure 8.
Figure 8. (Continued).
Figure 8. Mass spectrum of ions produced from Ti:sapphire laser-ablated bone sample with a 30fs
duration at 800nm.
Since the main composite material of bone is hydroxyapatite and collagen, we are
expecting in the obtained spectra to detect Ca peak at m/z = 40, resulting from the interaction
of the laser light with the present in the hydroxyapatite structural OH- group Figure 8c.This
interaction breaks the OH- bond releasing H+ ions and Ca ions.
The peaks at m/z =19 and m/z =23 can be attributed to F-and Na+. Fragmentation is
observed from the breakage of the polypeptide chain of collagen which is confirmed by the
presence of CmHn residues from the polypeptide chain of collagen Figure 8b, c, d, e, f, and g.
The mass spectra from tooth sample, acquired for ns, fs ablation and SIMS are shown on
Figure 9.
The laser ablation (800nm) ion spectra of tooth sample show a series of hydrocarbon
peaks as in the mass spectra acquired for bone sample. The fragmentation pattern, in the laser
ablation at 800nm and in SIMS mass spectra for tooth sample, exhibit the Ca+ ion, this is not
observable for the mass spectra obtained at 193nm. The Ca+ ion peak is surrounded in all
spectra by a moiety of low mass ion peaks of H, C, N, O, Na, P, K, and low mass ion
fragments such as COOH, CnHm. Mass fragments that are unique to the peptide chain are
observed in the higher mass range.
By comparing Figure 9 a, b and c we observed that ion signals with high masses do not
exist in the nanosecond ablation spectra and SIMS. As it is shown in Figure 9 the degree of
fragmentation in the fs mass spectra appeared to be less [15], which is confirmed by the
detection of higher mass fragments with m/z up to 177.
Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. High Intensity Laser-Matter Interactions 405
Figure 9. Comparison between laser ablation mass spectra of tooth sample and SIMS from tooth
sample: a) SIMS; b) femtosecond laser ablation mass spectra at λ=800nm, pulse duration 30fs and
I=5.1*1013W/cm2; c) Nanosecond laser ablation mass spectra at λ=193nm, pulse duration 30ns and
I=2.5*108W/cm2.
Comparing the experimental results obtained for femtosecond and nanosecond laser
ablation and SIMS can be concluded that there is a great difference in the effectiveness of
producing of ionized species of organic samples. It is valuable to examine the effect of laser
power density on fragmentation efficiency. Figure 10 shows a comparison between the mass
spectra of ablated ions from bone sample by increasing laser power density.
For lower laser intensities (8*1013-2.5*1014W/cm2) the overall fragmentation pattern
comprises only H, Na and Ca ion peaks (Figure 10 c, d). In addition, the yield of atomic
fragments was higher at low ablation intensity. As the ablation intensity increases (3.4 *1014-
3.7*1014W/cm2) a shift in ion abundance toward lower mass was typical, multiply charged
species of C, N, and O ions (C2+,C3+, N2+, N3+, O2+, O3+) have been observed (Figure 10 a, b).
At higher intensities the individual profiles of the CmHn groups are registered typical
representatives of the polypeptide chain. As a general observation it was concluded that at
low laser power densities the signals increase slowly. The signals increase drastically when
power densities more than 3¯1014W/cm2 are reached.
Figure 10. Comparison of ToF mass spectra of laser ablated species of bone sample over range of
ablation intensities (1013-1014W/cm2) at λ=800nm.
3.3.3. Fs Laser Post-Ionization Experiments

From the analysis of the post-ionization mass spectra the coupling mechanism between
the intense laser pulse and the molecules can be deduced. The ionization mechanism of
molecular species interacting with intense laser pulses was examined in order to gain more
inside into the fragmentation pathways. Particle desorption was achieved using 9keV
accelerated Xe+ ions. Neutral species intersect the laser beam at a distance of 2mm above the
target surface. The spectra were recorded at λ=800nm, pulse duration of 30fs and with typical
ionization laser intensities of I=2.2*1014W/cm2. Figure 11a shows a typical mass spectrum of
sputtered ions from bone sample under 9keV Xe+ion bombardment. The spectrum exhibits
atomic ions at m/z = 23(Na), 40(Ca) and 44(44Ca) and also some molecular fragments at m/z
=56 (CaO).
Figure 11. A comparison of ToF mass spectra of SIMS and post-ionization mass spectra: a) ToF mass
spectrum of desorbed particles of bone sample bombarded with 9keV Xe+ ions; b) ToF mass spectrum
of post-ionized species with femtosecond laser (λ=800nm, 30fs, I=2.2*1014W/cm2).
Figure 11b shows the 800nm LSNMS mass spectrum of bone samples. The spectrum
yields structurally specific and non-specific fragments. Besides the masses at m/z = 23(Na),
40(Ca), 44(44Ca) and 56 (CaO) which appear on Figure 11a , the mass spectrum on Figure
11b shows signals at m/z = 20, 47, 49, 60, 82, 96, 103, 113, 120.
A comparison between laser-SNMS and ToF SIMS showed that laser-US LSNMS is
particularly well-suited for identifying specific high mass fragments of biological molecules
[16]. With laser intensity of > 1.1*1014W/cm2 we measured increase in the amount of
characteristic ions produced with post-ionization over SIMS under identical ion sputtering
conditions. The emission of charged species was detected in the case of fs post-ionization
mass spectrum. Moreover the fs post-ionization experiments are in good agreement with fs
laser ablation data, more specifically both approaches proved that ultra-short laser tissue
interaction is a nondestructive mechanism for obtaining tooth and bone tissue mass spectra.
3.3.4. ESEM of Biological Samples

The surface modification of Ti:sapphire laser ablated tooth tissue surfaces were examined
with ESEM microscope in order to give qualitative estimation of the ablation efficiency under
irradiation with different number of laser pulses. Teeth samples were cut transversally in
1mm thick disks using diamond saw and afterwards polished with 1200-grit SiC paper. After
the cut the dentine disks were stored in distilled water until examination. Laser surface
exposure was performed in air. The laser beam was focused on to the dentine disks using 8cm
focal length lens to a focal spot of 30µm diameter. All the specimens were evaluated with an
environmental scanning electron microscopy ESEM (XL30 ESEM-Philips) operated in wet
mode at 1Torr pressure and 20KV electron beam.
The acquired ESEM micrographs show the morphological changes that occurred in ultra-
short laser exposed dentine disks under different irradiation conditions. The results of ESEM
evaluation show the difference in surface morphology which was clearly identifiable. In
Figure 12a is presented a surface modification of tooth dentine surface exposed to modest
laser energy (1.36J/cm2) with N= 500 laser pulses. At this sub-ablative conditions can be
possible to achieve chemical changes without morphological damages in dentine surfaces. As
can be seen from the micrograph it is present a well defined threshold for laser ablation. In the
next Figure 12b, there is no evidence of thermal damages, cracks, or melting.
Figure 12. Representative ESEM micrographs of ultra-short laser irradiated dentine samples with 25fs pulse
duration, (a) F=1.36J/cm2 and N=500 pulses; (b) crater rim, N=100pulses, F=1.55J/cm2 there is no collateral
damage around the fs-laser ablated cavity; (c)-(d) N=100, 5000 pulses, F=1.55J/cm2 with increasing number
of pulses there is a localized thermally modified area around the crater and occasionally some cracks.
The result is pure ablation without any modification and collateral damage of the
surrounding crater areas [17]. In Figure 12c-d, dentine surfaces were irradiated with 500 and
5000 number of pulses respectively. An evidence of thermal load and formation of cracks at
the bottom of the craters was observed with increasing number of pulses. If we make a
juxtaposing between the Mass spectrometry results and ESEM examinations we can observe a
good agreement between both types of results. The detection of the highest mass fragments
under fs laser ablation conditions and the observation of very smooth and clean ablation
craters and surroundings correspond to the expected (non-thermal) mechanism for energy
transfer in ultrashort-pulsed laser ablation. More specifically can be concluded that the ultra-
short laser ablation is a process which develops with limited heat diffusion into the adjacent
areas of the processed surface, minimally invasive with efficient, fast and localized energy
deposition; almost absent mechanical damage, molten zones and liquid phase[18].
An understanding of the physical, mechanical and chemical properties of prepared dental
tissue after laser treatment may help to find the optimized conditions to enhance the clinical
applications of ultra-short pulses in dentistry. The focus of modern dentistry recently was
directed toward prevention of caries decay. The combination of the two different techniques
(LASNMS and ESEM) for analysis of tooth surface can be used as well for characterization
of protective coatings of tooth enamel.
4. FEMTOSCIENCE IN MEDICINE
Along with the latest development in the femtosecond laser technology the number of
medical ultra-fast laser applications has increased immensely in the recent years. In the
intensity range 1012W/cm2-1014W/cm2 a very wide fields of applications of femtosecond
lasers are the precision microstructuring (ranging from metals, dielectrics to biological
tissues), nanotechnology and microsurgery. This work covers a broad range of different
applications in femtoscience [19].
The field of ultrafast science and technology have progressively influenced different
applications in medicine [20]. For example, not direct application of femtosecond pulses in
medicine was found in the structuring of medical implants – stents. The excellent cut quality
conditions which ultrafast pulses provide makes possible to manufacture new stent designs
with dimensions of 1.5-3mm in diameter. They are used to replace blocked by the plaque
arteries for making a bypass operation and treatment of strokes. Fs pulses are finding uses in
biomedicine for brain surgery (treatment of tumors and and Parkinson’s disease), angioplasty
(ablation of plaque from blocked artery walls). Femtosecond laser now makes possible the
reconstruction of defect auditory ossicles. About 1.5 percent of all newborn babies are
partially or totally deaf because their ossicles are stiffened. With femtosecond laser is possible
to dissolve such stiffening without destroying the ossicles which are of a size below one
millimetre. In modern dentistry researchers have found that the treatment of dental hard tissue
is supported with reduced side effects such as cracks and heat conduction to the nerves [21]. It
is expected processing times comparable to that of turbine drilling, as well as an almost pain-
free treatment.
DENTAL HARD TISSUE IS CHARACTERIZED WITH STRONG

ABSORPTION IN THE IR REGION. THIS PLACE LIMITS TO THE TYPE OF
RADIATION USED FOR DENTAL APPLICATIONS.
When comparing fs laser dentistry to previous laser applications (Er:Yag, CO2 lasers)
[22] the quality is improved due to the reduced thermal and mechanical side-effects. It is
possible to avoid collateral injuries to the healthy tissue by minimizing the heat load. The
deposition of energy in ultrashort pulse laser machining is in an optical skin depth. Moreover
the removal of thin layers develops in time scales faster than that required to transfer the
energy into the bulk material. These are the primary advantages for employing ultra-short
pulses in dentistry nowadays because it is possible to achieve a better control over the
ablation depth per pulse and to reduce the limit for the ablation threshold which drastically
influences the ablation process. The carious tissue is ablated about ten times faster than sound
tissue, this leads to slowdown of ablation speed when sound tissue is reached. Recently an
innovative femtosecond laser shaping method for preparation of ceramic restorations by
ablation has gained promising results [23]. The system consists of three components: three
dimensional evaluation of the tooth surface by coherence radar, CAD system for construction
of the restoration, CAM system for production of the tooth model. With the tiny spot of the fs
laser it is possible to form the fine occlusal surfaces of a crown with less collateral damage. A
decade ago for the needs of eye surgery was used UV laser ablation (LASIK method) which
requires surgical intervention. With a mechanical blade a flap of cornea is cut and than an
excimer laser reshapes a portion of the cornea, afterwards the flap is repositioned. Today with
the advent of fs pulses can be provided the required intensities while delivering sufficiently
low average power in order to avoid collateral damage. Thus surgeons can very precisely
create the initial flap and reshape the human eye cornea by ablation of thin layers of corneal
tissue using ultra-short pulses without surgical intervention [24] Figure 13.
Figure 13. Femtec Laser system, commercially available, for treatment of transparent corneal tissue.
Perforating keratoplasty using the femtosecond laser can be performed safely and accurately with
precise cut quality.
Figure 14. Femtosecond laser flap preparation with FSLASIK.
The LASIK software directs the tiny 3 micron spot of a focused fs laser at the central
layer of the cornea. Then the laser starts to create a dissection plane by forming series of
bubbles. The effect of photodisruption occurs only in the focal point. When the pattern of
bubbles is made the surgeon folds the tissue back to expose the underlying corneal layer
Figure 14.
The advantages of this method is expressed in higher precision tissue cuts, smaller
heating effects in the tissue, very weak shock waves, reduced collateral damage, improved
clinical safety. Another potential application in ophthalmology which is still under
investigation is the creation of a new drainage system in the eye and also for glaucoma
treatment [25]. Successful innovative applications of femtosecond lasers were shown in
nanosurgery. With pulses of intense laser light researchers have succeed to vaporize tiny
structures inside living cells without destroying them [26]. This technique could be valuable
to test how the cell work and to perform diagnosis and therapy of individual cells -
superprecise surgery. For example with Fs laser can be removed a single mitochondrion (the
cell’s powerhouse) without damaging surrounding structures. Researchers have found a way
to create the smallest incision into living tissue - with a width of 70 nanometers. Thus a
foreign genetic material can be inserted into living cells using ultra-short laser pulses, without
destroying the cells.
The implementation of ultrashort pulses in clinical use still meet several requirements for
improvement of the treatment procedures. Some of the main difficulties which have to be
overcome are the increase of the ablation speed, miniaturization of the systems by integrating
the laser beam into glass fiber in order to use the laser as a conventional drill and high costs
of such laser systems. Integration of laboratory Fs systems in “One box” user friendly
ultrafast laser with a possibility for implementing of “On board” laser diagnostics and
overcoming the high costs of such laser systems will help to adapt this technology for
commercial applications. Nowadays the improvements to the laser technology by means of
tweaking the design of the optical cavities and precision in the pumping mechanisms give
possibilities for the laser producers to provide lasers with five to six femtoseconds of pulse
durations.
CONCLUSION
The obtained results clearly show the sensitivity of US LA – SNMS mass spectrometry
TOF technique for elemental analysis of biologically related samples. Further studies are
necessary in order to expand and supplement the knowledge of the interaction mechanisms of
ultra-short pulses by investigating soft tissue as well. For example, one possible direction of
examination which can give real influence of medical treatment in ophthalmology will be to
study the processes which cause inflammation due to insertion of eye implants. The
knowledge of reaction products, which are formed after the implant insertion and cause
contamination, can be the main source of eye inflammation. By using the described
techniques in this chapter, physical and chemical analysis can be performed to prevent the eye
inflammation after surgical intervention and input of eye implants.
In conclusion, Ti-sapphire laser used for laser ablation in this study was found very
effective for processing of hard bio-tissue. Furthermore, we demonstrated the great potential
of ultra-short laser technology for successful implementation in medicine.
REFERENCES
[1] Krueger, J.; Kautek, W. Adv. Polym. Sci. 2004, vol 168, 247-289.
[2] Dausinger, F.; Nolte, S. Topics in Appl. Phys. 2004, vol 96, 1-6.
[3] Xu, L.; Tempea, G.; Poppe, A.; Lenzner, M.; Spielmann, Ch.; Krausz, F.; Stingl, A.;
Ferencz, K. Appl. Phys. B 1997, vol 65, 151-159.
[4] Le Blank, C.; Grillon, G.; Chambaret, J. P.; Migus, A.; Antonetti, A.; Optics Lett 1993,
vol 18, 140-142.
[5] Szipoes, R.; Ferencz, K. Optics. Lett. 1994, vol 19, 201-203.
[6] Tempea, G.; Krausz, F.; Spielmann, Ch.; Ferencz, K. IEEE J Selected Topics in Quant
Electronics 1998, vol 4, 1- 4.
[7] Dayem, R. N. Lasers Med. Sci. 2007, DOI 10.1007/s10103-007-0499-x.
[8] Sorokina, I.T.; Vodopyanov, K.L. Solid-State Mid-Infrared Laser Sources, Springer-
Verlag Berlin Heidelberg, 2003; Topics Appl. Phys., vol 89, 511–546.
[9] Adams, F.; Vertes, A. Fresenius J. Anal. Chem. 1990, vol 337, 638-647.
[10] Chew, A.; Sykes, D. E.; Waddilove, A. E. J.Anal. Atomic. Spectr. 1997, vol 12, 1101–
1103.
[11] Varoucha, e.; Papadogiannis, N. A.; Charalambidis, D.; Saenz, A.; Schroeder, H.;
Witzel, B. Phys. Rev. A 2001, vol 65, 012901-7.
[12] Strassl, M.; Wieger, V.; Brodoceanu, D.; Beer, F.; Moritz, A.; Wintner, E. JLMN 2008,
vol 3, 30-40.
[13] Daskalova, A.; Husinsky, W. Plasma Process Polym 2006, vol 3, 257-261.
[14] Ozono, K.; Obara, M.; Appl. Phys. A 2003, vol 77, 303–306.
[15] Nakashima, N.; Shimizu, S.; Yatsuhashi, T.; Sakabe, S.; Izawa, Y. Journ of
Photochemistry and Photobiology C 2000, 131–143.
[16] Kato, T.; Kobayashi, T.; Kurata-Nishimura, M.; Okamura-Oho, Y.; Sano, T.; Oyama,
R.; Matsumura, Y,; Yamamoto, H.; Hayashizaki, Y.; Matsuo, Y.; Kawai, J. NIM B
2008, vol 266, 992-997.
[17] Wieger, V.; Strassl, M.; Wintner, E. Laser and Particle Beams 2006, vol 24, 41–45.
[18] Serbin, J.; Bauer, T.; Fallnich, C.; Kasenbacher, A.; Arnold, W. H. Appl. Surface
Science 2002, vol 197-198, 737-740.
[19] Steiner, R. Medical Laser Applications 2006, vol 21, 131-140.
[20] Lubatschowski, H.; Heisterkamp, A.; Will, F; Singh, A.I.; Serbin J.; Ostendorf,A.;
Kermani, O.; Heermann, R.; Welling, H.; Ertmer, W. RIKEN Review 2003, vol 50, 113-
118.
[21] Rode, A. V.; Madsen, N. R.; Kolev, V. Z.; Gamaly, E. G.; Luther-Davies, B. Proc. of
SPIE 2004, vol 5340, 76-86.
[22] Lizarelli, R. de F. Z.; Moriyama, L. T.; Jorge, J. R. P.; Bagnato, V. S. Laser Physics
2006, vol 16, 849–858.
[23] Deppe, H.; Horch, H.; Lasers Med. Sci. 2007, vol 22, 217-221.
[24] Dupps W.J.; Wilson, S. E.; Experimental Eye Research 2006, vol 83, 709-720.
[25] Ngoi, B. K. A.; Hou, D. X.; Koh, L. H. K.; Hoh, S. T. Lasers in Medical Science 2005,
vol 19, 218–222.
[26] Watanabe, W.; Matsunaga, S.; Shimada, T.; Higashi, T.; Fukui, K.; Itoh, K. Medical
Laser Application 2005, vol 20, 185–191.
Chapter 13
IMPROVED THERMAL MODEL AND ITS APPLICATION

IN HIGH-POWER LASER ABLATION OF TARGET
Ranran Fang*1,2, Duanming Zhang2, Hua Wei3 and Zhihua Li2

1
College of Mathematics and Physics, Chongqing University of Posts and
Telecommunications, Chongqing 400065, China
2
School of Physics, Huazhong University of Science and Technology,
Wuhan 430074, China
3
Center for Modern Physics and Department of Physics, Chongqing University,
Chongqing 400044, China
ABSTRACT
Among various candidates considered to grow high-quality thin film materials,
pulsed laser deposition (PLD) technique has attracted much attention over past years due
to its special advantage that the target component is similar to the thin film. During PLD
process, pulsed laser ablation (PLA) is an important stage which determines the quality
and the efficiency in making thin film. Therefore, the investigation of the corresponding
thermodynamic process of PLA is very significant and the relevant researches are widely
developed. In this chapter, we review recent works of PLA including the improved
thermal model. The concrete works include an improved thermal model describing UV
high-power nanosecond laser ablation of metal target, and a tailed thermal model
describing IR high-power nanosecond laser ablation of multi-elemental oxide
superconductor target. Under the same parameters, the comparison between the
numerical results and the experimental data is also presented.
*
ranranfang2@gmail.com
416 Ranran Fang,, Duanming Zhang, Hua Wei et al.
1. INTRODUCTION
Since the high-quality YBa2Cu3O7 superconducting thin film was successfully prepared
in the American Bell laboratory for the first time using long pulsed short-wave excimer laser
in 1987, pulsed laser deposition (PLD) technique has been developing into an important
technology for the preparation of thin films [1]. Although many kinds of functional thin films
have been prepared and lots of progress in experiments have been made, the understanding of
PLD mechanism is lagging behind the experimental study, especially in the high-power
pulsed laser ablation(PLA) processes. During laser ablation, a large amount of electrons, ions
and excited neutrals coexisting in the vaporized material will absorb the laser light and form
plasma around the target surface. The interactions in laser, solid and plasma become much
8 −2
more complicated, especially when the incident laser intensities greater than 10 Wcm [2,
3, 4, 5, 6].
The chapter focuses on PLA based on the improved thermal model. The features of the
ablation process are influenced by the quantity of laser energy absorbed by the target. The
absorption coefficient introduced by the Lambert-Beer's law accounts for the attenuation of
the absorbed energy in the target [7], and the absorptivity refers to the magnitude of the ratio
of absorbed energy to incident energy [8]. They both describe the ability of absorbing laser
energy for target. For the thermal mechanism dominates almost all the physical effects in the
process of nanosecond pulsed laser ablation [9], the absorption coefficient and the
absorptivity are the important optical parameters in the pulsed laser ablation. Recently, there
are some experimental studies about the absorption coefficient and the absorptivity [10, 11].
However, few literatures investigate the characters of the absorption coefficient and the
absorptivity theoretically in detail in PLA [9, 12, 13]. The temperature dependence of the
absorption coefficient and the absorptivity are widely studied without considering the effects
of vaporization and plasma shielding in the laser ablation process (before melting) [14]. In
order to describe the whole process (before and after melting) of the UV high-power
nanosecond laser ablation of metal target, a tailored theoretical model including the
vaporization effect, the plasma shielding effect, and the temperature dependence of
absorption coefficient and absorptivity is reviewed in this chapter. The modified model is
used to simulate UV high-power laser ablation of iron target. We obtain the space- and time-
dependence of temperature of the target, and the time-dependence of temperature at two
different position for a certain laser fluence. The satisfactory agreement between our
numerical results and experimental data of ablation depth indicates that the vaporization
effect, the plasma shielding effect, and the temperature dependence of absorption coefficient
and absorptivity in laser ablation of metal target should not be neglected under UV high-
power nanosecond laser ablation situation.
We also present a tailed theoretical model for the high-power laser ablation of multi-
elemental oxide superconductor which includes the effects of vaporization and the plasma
shielding. The numerical results for YBa2Cu3O7 target are obtained by the finite different
method. We give the time-dependence of temperature of target at a certain laser fluence, the
transmitted intensity for different laser fluences, and the ablation rate per pulse as a function
of the laser fluence. The satisfactory agreement between our numerical results and
Improved Thermal Model and its Application in High-Power Laser Ablation … 417
experimental data indicates that the influences of the vaporization effect and the plasma
shielding effect in PLA of multi-elemental oxide superconductor are unambiguous.
In section 2, we will concentrate on the improved thermal model and its application in
UV high-power pulsed laser ablation of metal target. In contrast, section 3 is for laser-target
interaction during high-power pulsed laser deposition of superconducting thin films, which,
different from the implementation in section 2, is robust to high-power laser ablation of multi-
elemental oxide target. We will conclude in section 4.
2. IMPROVED THERMAL MODEL AND ITS APPLICATION IN UV

HIGH-POWER PULSED LASER ABLATION OF METAL TARGET
In this section, we will consider the whole process (before and after melting) of the UV
high-power nanosecond laser ablation of metal target, which takes into account the
vaporization effect, the plasma shielding effect, and the temperature dependence of
absorption coefficient and absorptivity. To describe such a complicated case, we may use the
improved thermal model.
2.1. Temperature Dependences of the Absorption Coefficient and the

Absorptivity
This section is based on the optical characters of material and classical electromagnetism
theory. It is assumed that a pulsed laser beam irradiates on a target surface vertically. The
electric field intensity in the target can be described as follows,
x
E = E0 exp[iω ( − t )], (1)
E0 is the electric field intensity at x=0, ω is the incident laser angle frequency, υ is the
laser propagation velocity in target satisfying the following relation [15],
υ = c/nc , (2)
where c is the laser propagation velocity in vacuum. nc is the complex refractive index,
which can describe the optical characters of material.
nc = n + iκ , (3)
where n is the refractive index, κ is the extinction coefficient.

Substituting Eqs. (2) and (3) into Eq. (1) yields
xn xκ
E = E0 exp(−iωt ) exp(−iω ) exp(−ω ). (4)
c c
By the operation of module to Eq. (4), we have
xκ
| E |2 = EE * =| E0 |2 exp(−2ω ). (5)
c
According to Lambert-Beer-Bouguer Law, when a laser irradiates the target surface, the
laser intensity attenuates in exponent in the target, which can be expressed by [16, 17]
I ( x, t ) = βI (0, t ) exp(−bx), (6)
I (0, t ) is the intensity of the incident laser. β is the absorptivity of the target, b is the
absorption coefficient of the target.
In the target, the intensity equals to the square of the module of the electric filed intensity,
2
namely | E | = I ( x, t ) . Based on this relation, comparison of Eq. (5) with Eq. (6) gives the
following relation,
2ωκ 4πυκ 4πκ

b= = = , (7)
c c λ0
where λ0 is the laser wavelength in vacuum.

Substituting Eq. (4) into Maxwell equation, we obtain the refractive index n and the
extinction coefficient κ as the functions of the conductance σ and the permittivity ε as
follows,
1 σ
n2 = ε{ 1 + ( ) + 1},
2 ωεε 02
(8)
1 σ
κ 2 = ε{ 1+ ( ) − 1},
2 ωεε 02
where ε 0 is the permittivity of vacuum.

−1 −1
The conductance of general metal in the range ∼ 10 − 10 Ω m
8 12
is big enough, so
we can regard that
σ 2
( ) 1. (9)
εε 0ω
So Eq. (14) can be simplified by
σ
n κ , (10)
2ωε 0
σ can be presented as linear increasing function of electron temperature T [18],
σ0
σ (T ) = , (11)
1 + α (T − T0 )
where σ 0 is the target conductance at initial temperature T0 , α is the target temperature

coefficient of resistance. For general metal, α is in the range
∼ 400 ×10 −5 − 700 × 10 −5 K −1 [19].
Substituting Eqs. (10) and (11) into Eq. (7) yields
4πσ 4πσ 4πσ 0 4πσ 0

b(T ) = = = = . (12)
λ0 ε 0 λ0 c ε 0 λ0 c ε 0 λ0 c[1 + α (T − T0 )]
This is an important relationship that we derived in this section. From this formula, it can
be found that the absorption coefficient monotonously decreases with target temperature
when the resistance temperature coefficient and irradiated laser wavelength are all constants.
With the same point of view, we give a discuss about the absorptivity, i.e., the ratio of the
absorbed intensity to incident intensity. When a laser irradiates a target surface vertically, the
target absorptivity is generally described as follows [18],
(n − 1) 2 − κ 2 2κ 2
β = 1− R = 1− = , (13)
(n − 1) 2 + κ 2 (n − 1) 2 + κ 2
where R is the target reflectivity.

Substituting Eqs. (10) and (11) into Eq. (13) gives
4πυε 0 4πcε 0 (1 + α (T − T0 ))
β (T ) = 2 = . (14)
σ λ0σ 0
This is another important relationship that we derived indicating the relation between the
resistance temperature coefficient, the laser wavelength, the target electron temperature and
the absorptivity.
2.2. Absorption and Ionization Mechanisms
The process of high-intensity laser pulse interacts with a solid sample involves the laser-
solid interactions at the target surface and the formation of plasma. For high laser fluences the
vapor becomes significantly ionized, absorbing part of the incident laser beam and leading to
vapor breakdown and plasma formation [3]. When the density and temperature of the laser-
produced plasma become high enough then an efficient target shielding occurs during the
laser pulse. The features of the laser ablation process are strongly influenced by laser-plasma
coupling and plasma kinetics. Before describing the laser-solid-plasma interaction model, the
main absorption and ionization mechanisms active in the laser-ablation process at UV
wavelengths should be introduced at the very start.
Two different photon-absorption processes are considered in the model, i.e., the inverse
bremsstrahlung (IB) and the photoionization (PI) of excited species.
We will firstly consider IB absorption processes for UV laser photons ( λ =248nm).
During the IB process, the probability of photon absorption in an individual electron-ion
collision is much greater than the one of the scattering by neutrals. However, at the early
stages of the laser evaporation process, when the ionization degree of the iron vapour cloud is
very low, the large density of neutrals in the initial vapour significantly enhance this last
contribution. The IB process is usually described by the inverse absorption length
α IB ( cm −1 ) [20],
N0
α IB ≈ [1.25 ×10 − 46 λ3Te−0.5 ( ) + N e ]N e = σ IB N e , (15)
200
where λ is the laser wavelength in cm , Te is the electron temperature in eV , N 0 and N e

−3
are densities of the neutral atom and electron in cm , respectively. For the excimer laser
−3
ablation of metals, such as iron, the plasma particle density is 5 ×10 cm , and the electron
18
temperature is 3 eV [26]. In Eq. (15), σ IB represents the total cross section of the IB process,
and the contributions of both neutral and ion IB processes are included [2, 3],
σ IB = σ IB,neu + σ IB,ion . (16)
The cross section of neutral σ IB,neu equals approximately to 2.75 ×10 −23 cm 2 . Since a
−2 −2
laser fluence of 1 − 6Jcm is equivalent to 0.125 ×10 − 0.75 ×10 photonscm , the
19 19
product of the cross section and photon density clearly indicates a probability of
∼ 0.0034% − 0.02% for the incoming photons to be absorbed by IB due to neutrals,
which is too small to significantly attenuate the laser or heat plasma.
Another absorption process is the direct photoionization of excited atoms. In the UV laser
ablation process, direct PI of excited atoms in the vapor plays a significant important role.
The photoionization cross-section can be expressed as [21],
( EI − E ∗ ) 2.5 I H0.5
σ PI ≈ 7.9 × 10 −18 cm 2 , (17)
( hν ) 3
where I H is the ionization potential of hydrogen, E I is the ionization potential of the atom
∗ ∗
and E is the energy of the excited level which can be photoionized, and here E I − E is
taken to be equal to the photon energy hν . The photon energy of the 248nm excimer laser is
5eV, making σ PI ≈ 1.3 ×10 −17 cm 2 , which are several orders of magnitude higher than the
IB cross section. Since a laser fluence of 1 − 6Jcm −2 is equivalent to
−2
0.125 ×10 − 0.75 ×10 photonscm , the proability for an excited atom of absorbing an
19 19
UV photon equals approximately to 1 , i.e., the ionized condition is formed. As a

consequence, the vapour ionization degree is readily enhanced, PI is the most important
absorption mechanism. Therefore, it can be pointed out that only considering PI is reasonable
in this section.
2.3. The Improved Thermal Model for Two Different Stages
2.3.1. The Stage Before Target Melting

We give the heat diffusion equation for target before melting as follows,
∂T ( x, t ) ∂ ∂T ( x, t )
cs ρ s = (k s ) + β (T )b(T ) I 0 (t ) exp(−b(T ) x)(0 < t ≤ τ m ), (18)
∂t ∂x ∂x
where cs , ρ s and k s are the thermal capacity, the mass density and the thermal conductivity
of the solid target, respectively.

In order to describe the incident laser intensity more accurately, we adopt the laser
incident intensity which can be expressed by a Gauss function [14],
( t −τ ) 2
−
2σ 2
I 0 (t ) = I 0 e , (19)
where I 0 is the maximal laser power density.

According to Ref. [18], the time for the surface to reach the melting point is
πcs 2 ρ s 2Tm 2
τm = , (20)
4β (T ) 2 I 0 (t ) 2
where Tm is the melting temperature.
The initial temperature of the sample is 300 K. Based on the energy conservation law
and the adiabatic condition, the boundary conditions can be expressed as
∂T ( x, t )
− ks | x = 0 = β (T ) I 0 (t )(0 < t ≤ τ m ), (21)
∂x
∂T ( x, t )
− ks | x = d = 0(0 < t ≤ τ m ), (22)
∂x
where d is the target thickness.
2.3.2. The Equation for Liquid Target During the Laser Irradiation
Once the absorbed energy is high enough, the target surface will be melt. When the
temperature of target surface is higher than the melting temperature, i.e., T ≥ Tm , the
vaporization effect will become noticeable. Then the effects of vaporization and plasma
shielding will be important and Eq. (18) is no longer valid. After considering the vaporization
effect and the plasma shielding effect, the temperature of the target in the period from t = τ m
to t = τ is determined by
∂T ( x, t ) ∂T ( x, t ) ∂ ∂T ( x, t )
cl ρ l + cl ρ l u (t ) = (kl ) + β (T )b(T ) I 0 (t ) exp(−σ PI nx H ) exp( −b(T ) x )
∂t ∂x ∂x ∂x
(τ m < t < τ ),
(23)
where cl , ρ l and kl are the thermal capacity, the mass density and the thermal conductivity
of the liquid target, respectively. nx is the number density of the excited neutrals and H is
the vapor height.

The vaporization flux can be obtained according to Hertz-Knudsen equation [22]. The
velocity of vaporization (or say, the velocity of surface recession) can be expressed as [41],
1
p⎛ m ⎞ 2
u (t ) = 0.82 ⎜⎜ ⎟ , (24)
ρ ⎝ 2πkbT (0, t ) ⎟⎠
where p is the gas pressure, ρ is the material density, m is the particle average mass, and
kb is the Boltzmann constant. T (0,t ) is the surface temperature.
Under the limit Vlip Vgas , Vlip and Vgas being the molar volume of liquid and gas,
respectively, the vapor pressure above the vaporized surface can be estimated by the
Clausius-Clapeyron equation
⎡L⎛ 1 1 ⎞⎤
p = pb exp ⎢ ⎜⎜ − ⎟⎟⎥, (25)
k T
⎣ b⎝ b T (0, t ) ⎠⎦
where pb is 1 atm, L is the latent heat of the target material, and Tb is the boiling
temperature under the reference pressure pb .
It is noted that because the cross section is very small, our model only consideres the first
ionization from the ground state to the first excited state while the multiple ionization is
neglected, which is reasonable in general conditions. The excited particles density nx as a
function of plasma temperature is expressed as a Boltzmann distribution
nx (T p ) = n0 exp(−ε/kbTp ), (26)
where n0 is the total particle density, ε corresponds to the energy from the ground state to
the first excited state, and T p is the plasma temperature.
During the laser pulse, the temperature in the plasma is assumed as isothermal. The
plasma length H varies with time when plasma expands outward. The dynamic equation of
plasma expansion is given by [23, 24]
⎡ 1 dX (t ) d 2 X (t ) ⎤ ⎡ 1 dY (t ) d 2Y (t ) ⎤ ⎡ 1 dZ (t ) d 2 Z (t ) ⎤ kbTp
X (t ) ⎢ + ⎥ = Y (t ) ⎢ + ⎥ = Z (t ) ⎢ + ⎥= ,
⎣τ dt dt 2 ⎦ ⎣τ dt dt 2 ⎦ ⎣τ dt dt 2 ⎦ m
(27)
where dX (t )/dt , dY (t )/dt and dZ (t )/dt refer to the expansion velocities of the plasma
edges X (t ) , Y (t ) and Z (t ) , respectively. We fix X (t ) in the direction perpendicular to

target of the expanding plasma. The X (t ) direction of plasma expansion is considered in this
section.
The dynamic equation (15) is numerically solved by a finite difference scheme in order to
evaluate the plasma length. In our calculations, the initial velocity υ is the thermal particle
velocity defined by [25]
υ = 8kbT (0, tm )/πm . (28)
Based on continuous temperature condition we get the initial condition,
T ( x, t ) |t =τ = T ' ( x, t ) |t =τ , (29)
m m
'
where T ( x, t ) |t =τ is the temperature distribution on the threshold of melting.
m
In order for energy to be conserved, the boundary conditions can be expressed as
∂T ( x, t )
− kl | x = 0 = − Lρ l u (t ) + β (T ) I 0 (t ) exp(−σ PI nx H )
∂x (30)
(τ m < t < τ ),
∂T ( x, t )
− kl | x = d = 0(τ m < t < τ ). (31)
∂x
2.4. Results and Discussion

2
Iron target is irradiated by the fluences ranged in 1-6J/cm Gaussian profile laser [26].
Table 1 is the thermal and optical properties of iron from Refs. [18,26-28].
Table 1. Thermal and optical properties of iron
k s ( Wm −1K −1 ) kl ( Wm −1K −1 ) ρ s (kgm −3 ) ρ l (kgm −3 )

80 41.5 7870 7860
cs (Jkg −1K −1 ) cl (Jkg −1K −1 ) L(Jkg −1 ) α
4.48× 10 2 6.93×10 2 6288 4.8 × 10 −3
2.4.1. Profile and Evolution of Temperature

2
Figure 1. The time and depth dependence of the target temperature for the laser fluence of 4J/cm
Firstly, we investigate the time and depth dependence target temperature obtained from
our improved thermal model. Fig. 1 displays the three-dimensional profile of temperature for
2
the incident laser fluence of 4J/cm . From the figure we can see that in time period
0 ≤ t ≤ 40 ns, at a given time, the temperature decreases along with the depth. In time period
40ns ≤ t ≤ 60 ns ( 60ns is the pulse duration), for the certain time, the temperature increases
along with the ablation depth, and after it reaches a peak it begins to decrease. This
phenomenon is called subsurface heating which is studied in detail in Ref. [41]. At a given
depth, the target temperature evolves like Gauss profile, which is concerned with the
tendency of Gauss profile of pulsed laser. The evolvement of target temperature in this
section is consistent with the fact.
2.4.2. Comparison of the Temperature Distribution During UV High-Power Pulsed

Laser Radiation
Figure 2. Comparison of the evolution of temperature at d=0.00 μ m and d=0.05 μ m obtained from
three different models , (1) neglecting the effects of vaporization and plasma shielding, (2) only taking
the vaporization effect into account, and (3) considering the two effects of vaporization and plasma
shielding. The temperature dependence of absorption coefficient and absorptivity are taken into account
in all curves.
In order to investigate the two effects of vaporization and plasma shielding on the UV
high-power pulsed laser ablation of metal target, we plot the target temperature evolution
obtained from three heat conduction models, as shown in Fig. 2. The variations of the target
temperature T ( x, t ) with time t at x=0.0 μ m and x=0.05 μ m under the laser fluence of
2
4J/cm are respectively obtained.
We can see from Fig.2, at the beginning of the ablation ( t ≤ τ m ), the target temperature
profiles obtained from the three models are almost the same as each other. The reason is that
there is no vaporization around the target in this period, i.e., the energy loss of vaporization
can be neglected. However, during the time period of τ m ≤ t ≤ τ , the vapor emerges around
the surface, thereafter, the plasma comes into being. It can be seen clearly that the
temperature of the model considering the vaporization effect increases slower than that of
neglecting the vaporization effect. The phenomenon can be analyzed as follows, with
temperature gradually rising, vaporization becomes more and more remarkable, and the
energy loss at the surface becomes larger and larger, which leads to the temperature decrease
accordingly. Moreover, the temperature curve involving the effects of vaporization and
plasma shielding rises slower than that of only taking the vaporization effect into account.
The reason is that with the time the more laser energy is absorbed by the plasma. From the
foregoing analysis, we can find that the effects of vaporization and plasma shielding are
prominent in UV high-power pulsed laser ablation of metal target.
2.4.3. Ablation Depth as a Function of Laser Fluence

Figure 3. Fluence dependence of the calculated and measured ablation depths.
How the calculated and measured ablation depths depend on laser fluence is shown in
Fig.3 where the experimental data comes from Ref. [26]. Only considering the plasma
shielding, Lunney et al. simulated the first curve. The second curve is simulated from our
present model which describes the thermal phenomena before and after melting. The
satisfactory well agreement between model predictions and experimental data confirms that
the improved thermal model is correct and reasonable.
In this section, we have analytically demonstrated the temperature dependence of
absorption coefficient and absorptivity, the vaporization effect, the plasma shielding effect.
As the vaporization effect, the plasma shielding effect, and the temperature dependence of
absorption coefficient and absorptivity are included in the improved model, our present model
would be helpful for the experimental investigation of metal thin films prepared by pulsed
laser deposition.
3. LASER-TARGET INTERACTION DURING HIGH-POWER PULSED

LASER DEPOSITION OF SUPERCONDUCTING THIN FILMS
As discussed in section 2, much attention was focused on fundamental studies of the
laser-ablation process of simple one-component targets (such as metals) to gain a better
knowledge of the basic mechanisms of the laser-solid-plasma interaction. Among the
fundamental studies of the PLA process of one-component target [28-32], many effective
methods and models are developed. Amoruso et al [29] have investigated the absorption and
saturation mechanisms in aluminium laser ablated plasmas through experimental and
semiquantitative analysis methods. Particular attention has been devoted to a comparison
between experimental and theoretical time of flight profiles in Refs. [30]. A. Bogaerts et al
[31] have developed a comprehensive numerical model which studies the effect of laser
parameters on the target and plume characteristics. Influence of a finite initial ion density
gradient on plasma expansion into a vacuum has been studied in Ref. [32]. A simple
analytical model is presented for hydrodynamic expansion of laser-produced plasma with a
limited mass has been created in Ref. [33]. A great effort has been done in pulsed laser
ablation of one-component target. However, there are few literatures considering the
influence of the vaporization effect and the plasma shielding effect during the high-power
nanosecond pulsed laser ablation of multi-elemental oxide superconductor.
In this section, we will develop a tailed theoretical model for the high-power laser
ablation of multi-elemental oxide superconductor which includes the effects of vaporization
and the plasma shielding. The numerical results for YBa2Cu3O7 target are obtained by the
finite different method. We give the time-dependence of temperature of target at a certain
laser fluence, the transmitted intensity for different laser fluences, and the ablation rate per
pulse as a function of the laser fluence. The satisfactory agreement between our numerical
results and experimental data indicates that the influences of the vaporization effect and the
plasma shielding effect in PLA of superconductor can not be neglected under high-power
nanosecond laser ablation situation.

The section will be separated into several parts. We first study the formation of vapor and
the vaporization mechanism. Then the plasma shielding effect of multi-elemental oxide
superconductor is investigated in detail. The modified theoretical model for the high-power
laser ablation of multi-elemental oxide superconductor are presented in a separate part. As an
example, The numerical results for high-power laser ablation of YBa2Cu3O7 target are
obtained by the finite different method.
3.1. The Formation of Vapor and the Vaporization Mechanism
When a high-power laser irradiates a target surface, the target will absorb laser energy
and be heated. Once the absorbed energy is high enough, the target surface will be melt.
When the temperature of target surface is higher than the melting temperature, i.e., T ≥ Tm ,
the vaporization effect will become noticeable.
The vaporization flux can be obtained according to Hertz-Knudsen equation [22], and the
velocity of vaporization (say, the velocity of surface recession) can be expressed as,
1
p⎛ m ⎞2
u (t ) = 0.82 ⎜⎜ ⎟ , (32)
ρ ⎝ 2πkbT (0, t ) ⎟⎠
where p is the gas pressure, ρ is the material density, m is the particle average mass, kb
is the Boltzmann constant, and T (0,t ) is the surface temperature of the target. The target
temperature is a function of the target depth x and the interacting time t , T = T ( x, t ) ,
where the direction of the depth is perpendicular to the target surface.
Under the limit condition that the molar volume of liquid Vlip is far smaller than the one
of the gas Vgas , i.e., Vlip Vgas , the vapor pressure above the vaporized surface can be
estimated in the term of Clausius-Clapeyron equation
⎡L⎛ 1 1 ⎞⎤
p = pb exp ⎢ ⎜⎜ − ⎟⎟⎥, (33)
⎣ kb ⎝ Tb T (0, t ) ⎠⎦
where pb is the reference pressure 1 atm, L is the latent heat of the target material, and Tb
is the boiling temperature under pb .
3.2. Absorption And Ionization Mechanisms
During nanosecond laser ablation process, target evaporation begins after the impact of
the leading part of the laser pulse on the surface. The interaction of the following part of the
laser beam with the vapor in the vicinity of the target surface leads to strong heating and
ionization of the vapor and plasma formation [29]. There are also two dominant photon
absorption processes in the vapor. The first one is the inverse IB absorption process, free
electrons gain kinetic energy from the laser beam and promote plume ionization and
excitation through collisions with excited and ground-state neutrals. The second one is PI of
excited species [29, 34].
It is worth discussing the main absorption and ionization mechanisms active in the laser-
ablation process at the infrared (IR) wavelengths. We firstly consider IB absorption processes
using IR laser photons ( λ =1064nm). During the IB process, the probability of photon
absorption in individual electron-ion collisions is much greater than that of the scattering by
neutrals. However, at the early stage of the laser evaporation process, when the ionization
degree of the YBa2Cu3O7 vapor cloud is very low, the large density of neutrals in the initial
vapor will significantly enhance the photon-neutral scattering rate. The IB process is usually
described by the inverse absorption length α IB ( cm −1 ), which is given as [20],
N0
α IB ≈ [1.25 × 10 − 46 λ3Te−0.5 ( ) + N e ]N e = σ IB N e , (34)
200
where λ is the laser wavelength in nm, Te is the electron temperature in eV, and N 0 and
N e are the neutral atom and electron densities in cm −3 , respectively. The plasma particle
−3
density is in the range of 10 − 10 cm , the plasma's temperature is of the order 10 K
19 21 4
19 −3
[23]. It is supposed that particle density value is 10 cm , and the plasma's temperature is
Tp = 30000 K. In Eq. (34) σ IB represents the total cross section of the IB process, and the
contributions of both neutral and ion IB processes are also included [35, 36],
σ IB = σ IB,neu + σ IB,ion . (35)
The cross section of neutral σ IB,neu here equals approximately to 5 ×10 −21 cm 2 . A laser
−2 −2
fluence of 1-15 Jcm is equivalent to 0.5 ×10 − 8.1× 10 photons cm , the product of
19 19
the cross section and photon density clearly indicates a probability of ∼ 2.5% − 40% for
the incoming photons to be absorbed by IB due to neutrals, thus heating and exciting the
vapor species and, then, enhancing the free electron density by electron impact (EI) ionization
from excited states.
The mean ionization of crystal is defined to be the weighted mean of the first ionization
energy of each atom [37, 38],
〈U 0 〉 = ∑X iU 0i / ∑X i , (36)
i i
where the subscript i refers to the four kinds of atom in YBa2Cu3O7 , U 0i is the first
ionization energy of the i th type of atom, X i is the number of the i th type atom in
YBa2Cu3O7 . From Eq. (4), we can obtain the mean ionization 〈U 0 〉 of YBa2Cu3O7 is
10.398 eV.
If each evaporated atom in local thermal equilibrium has only one electron which can be
excited, the free electron density N e is given by Saha Equation [18],
2
N e = 3 × 10 21 (kbT p ) 3/2 exp(−〈U 0 〉 )/kbT p . (37)
From Eq. (37), we can see that N e = 1.44 × 10 , at which the IB by electron-ion
20
collisions become the primary mechanism of photon absorption ( σ IB,ion ∼ 1.25 × 10

−19
cm 2 ).
Therefore, PI process can be neglected for photon absorption processes under these
conditions.
3.3. Ablation Model
The target's temperature can be described by the heat conduction equation, depending on
the incident laser parameters, the thermodynamic parameters of the target, the boundary
conditions and the initial conditions. In order to describe the temperature fields more
accurately, we give the heat conduction equations with the initial and boundary conditions,
taking both the effects of vaporization and plasma shielding into account.
Before the target's surface begins to vaporize, the target conforms to the equation
∂T ∂ ∂T
cρ = k + α b βI 0 (t ) exp(−α b x)(0 < t ≤ τ m ). (38)
∂t ∂x ∂x
After considering the vaporization effect and the plasma shielding effect, the temperature
of the target in the period from t = τ m to t = τ is determined by
⎛ ∂T ∂T ⎞ ∂ ∂T
cρ ⎜ − u (t ) ⎟= k + α b β I 0 (t ) exp(−α IB H ) exp(−α b x)
⎝ ∂t ∂x ⎠ ∂x ∂x (39)
(τ m ≤ t ≤ τ ),
where c is specific heat, k is the thermal conductivity, τ is the width of laser pulse, α b and
β are the absorption coefficient and the absorptance of the target, respectively. In order to
describe the incident laser intensity more accurately, we also adopt the laser incident intensity
expressed by a Gauss function, i.e.,
( t −τ/2) 2
−
2σ 2
I 0 (t ) = I 0 e , (40)
where I 0 is the maximal laser power density.

During the laser pulse, the temperature in the plasma is assumed as isothermal. The
plasma length H varies with time when plasma expands outward. The dynamic equation of
plasma expansion is given by [23, 24]
⎡ 1 dX (t ) d 2 X (t ) ⎤ ⎡ 1 dY (t ) d 2Y (t ) ⎤ ⎡ 1 dZ (t ) d 2 Z (t ) ⎤ kTp
X (t ) ⎢ + ⎥ = Y (t ) ⎢ + ⎥ = Z (t ) ⎢ + ⎥= ,
⎣τ dt dt 2 ⎦ ⎣τ dt dt 2 ⎦ ⎣τ dt dt 2 ⎦ m
(41)
where dX (t )/dt , dY (t )/dt and dZ (t )/dt refer to the expansion velocities of the plasma
edges X (t ) , Y (t ) and Z (t ) , respectively. We fix X (t ) in the direction perpendicular to
target of the expanding plasma. The initial dimension of the direction X (t ) is shorter three
magnitude than the initial dimension of the direction Y (t ) and Z (t ) . However, the initial
velocity of the direction X (t ) is higher three magnitude than the initial dimension of the
direction Y (t ) and Z (t ) . So the expansion of the plasma results in a similar elongated
ellipsoid shape [39]. The X (t ) direction of plasma expansion is considered in this section.
The dynamic equation (41) is numerically solved by a finite difference scheme in order to
evaluate the plasma length. In our calculations, the initial velocity υ is the thermal particle
velocity defined by [25]
υ = 8kbT (0,τ m )/πm . (42)

For the equations (38) and (39), the initial condition is
T ( x,0) = T0 , (43)
and the boundary conditions are
∂T
−k | x = 0 = I ( x, t )(0 < t ≤ τ m ), (44)
∂x
∂T
−k | x = 0 = − Lu (t ) + I ( x, t )(τ m ≤ t ≤ τ ), (45)
∂x
where T0 =300K is the initial temperature which is uniform across the target.
4. RESULTS AND DISCUSSION

Under vacuum conditions, YBa2Cu3O7 target is irradiated by the Gaussian profile laser,
2
the fluences of which ranged in 1-15J/cm . The laser wavelength and width are 1064nm and
13ns, respectively [40]. Under the initial condition (43), and the boundary conditions (44) and
(45), the equations (38) and (39) are numerically solved by a finite difference scheme. Table
2 is the thermal and optical properties of YBa2Cu3O7 (see Refs. [40, 41]).
Table 2. Thermal and optical properties of YBa2Cu3O7
Average mass of vaporized species ( m ) 89g

Density (20C ) ( ρ ) 4.55 g/cm
3
Sunlimation temperature (at10 Pa )

5 2173K
Latent heat ( L ) 430.3 kl/mol

Absorptance ( A0 ) 0.85
Specific heat ( c ) 2.46 J/cm K
3
Average mass of vaporized species ( m ) 89g

Thermal conductivity ( k ) 3.2 W/mK
Absorption coefficient ( b )
α 5 ×10 4 cm −1
temperature conductivity ( α ) 0.011cm /s
2
4.1. Vaporization At The Surface

Figure 4. Profiles of vaporization front velocity versus irradiated time for three different laser fluence.
The vaporization front velocity is determined by the pressure and the temperature on the
target surface, the pressure is also a function about the temperature, and the temperature is a
function of laser fluence. Therefore, the vaporization front velocity at the surface for different
laser fluence can be obtained.
Fig. 4 is the vaporization front velocities versus radiated time for three different laser
fluence. We note that the vaporization front velocity evolution has a nearly Gaussian shape.
In time period t < τ m , three curves nearly overlap on the verge of zero. We can find that it's a
stage of the vaporization relaxation. We can also see that the laser intensity evidently affects
the vaporization front velocity. However, the rise in vaporization front velocity is not a linear
variety with that of laser intensity. Along with the intensity becomes bigger, the increase ratio
of vaporization front velocity turns to be slower.
4.2. Subsurface Heating
We can learn from Fig. 4, in the period 8ns ≤ t ≤ 13ns, for the certain time, the
temperature increases then it decreases along with the ablation depth, namely, the subsurface
heating phenomenon is the most remarkable when the pulsed laser finished.
Fig. 5 shows the temperature profiles within the target at the time moments when the
laser ends. The values of the subsurface heating, defined as ΔT = Tmax − Ts ( Tmax being the
maximum temperature in spatial distribution). We can see from Fig. 5, the values are 10000K,
2 2 2
4000K and 1000K for different laser fluences of 15J/cm , 8J/cm and 2J/cm , respectively.
In other words, the higher the laser fluence, the higher the value of the subsurface heating is.
Figure 5. The calculated temperature profiles within the target at the time moments when the laser ends.
The temperature within the target is determined by three factors, the heating by absorbing
the laser energy, the surface cooling due to the vaporization, and the inner cooling by the heat
diffusion. The absorption of the laser energy can be governed by the Beer-Lambert law
−α b x
I ( x, t ) = β (t ) I (t )e . The smaller the absorption coefficient α b , the greater depth the
laser energy penetrates is. The absorption coefficient α b of YBa2Cu3O7 is very small (Table
2), so the depth penetrated by the laser energy is very large. Generally speaking, the rate of
temperature equalization along the distance with time is controlled by the temperature
conductivity α = k/ρc . The α value for YBa2Cu3O7 is 0.011, so the heat is conducted to
the bulk depth slowly. As for the vaporization process, the rate of the vaporization is so high,
and much more energy is took away. Therefore, the three factors result in a collective
tendency for YBa2Cu3O7 , although the target surface absorbing the energy of the laser, the
ascend trendy of the surface temperature raises slowly due to the forceful vaporization. No
vaporization phenomena happens in the subsurface, and the rate of the temperature
equalization along the distance with time is very slow, then the subsurface temperature
ascends quickly and its temperature is higher than the one on the surface.
The higher the laser fluence, the more remarkable the vaporization and the more energy
accumulated in the subsurface, the higher the value of the subsurface heating is. Certainly, the
higher the laser fluence, the more energy is absorbed by the target, the temperature within the
target ascends quickly. We can see from Fig. 5, the higher the laser fluence, the higher the
curve locates.
4.3. The Formation of Plasma snd the Plasma Shielding
Figure 6. The relative laser energy reaching the target surface for three different laser fluences.
More noticeable plasma shielding come into being following the incident laser pulse
become higher. We represent the plasma shielding by the relative laser energy that reaches
the target surface, namely, the ratio of the actual laser energy reaching the target surface to
the peak value of the laser energy.
We can see from Fig.6, the solid curve expresses the relative laser energy reaching the
target surface without regarding for the plasma shielding. The other curves consider the
2 2 2
plasma shielding corresponding to the laser fluences of 2J/cm , 8J/cm and 15J/cm ,
respectively. In the same period, the higher the incident laser fluence, the less the relative
laser energy reach the target surface. In other words, the higher the incident laser fluence, the
more noticeable the plasma shielding is.
Furthermore, we can find that at the beginning of the ablation, four curves nearly overlap.
The reason is that during this period, the relaxation time of vaporization is considered. There
is no plasma existing at the surface during this stage, so the transmitted intensity is a constant
and the plasma shielding hardly emerges. However, during the time period of τ m ≤ t ≤ τ ,
the four curves detach. We can analyze the phenomenon as follows, in this time period, there
is vapor around the surface, thereafter, the plasma forms. The IB photo-absorption processes
shields part of laser energy and forms the plasma shielding.
The higher the laser fluence and the shorter the relaxation time τ m , the more remarkable
the plasma shielding is. For the high laser fluence, the plasma shielding happens in the prior
time, the slope coefficient of the curve diminishes (the increment rate of relative laser energy
that reaches the target surface is decreasing). To sum up, the higher the laser fluence, the
more remarkable the plasma shielding, so the less the transmitted relative laser energy is.
4.4. Plasma Shielding

2 2
In Fig. 7, the transmitted intensity for three different laser fluences of 5J/cm , 8J/cm
2
and 13J/cm are presented, respectively. For the fixed time, the higher the incident laser
fluence, the less the transmitted intensity is. In other words, the higher the incident laser
fluence, the more noticeable the plasma shielding effect is. The plasma shielding effect is
described by the transmitted intensity, i.e., the ratio of the actual laser energy reaching the
target surface to the incident laser energy.
Figure 7. Evolution of the transmitted intensity for three different laser fluences.
Furthermore, we can find that the value of the transmitted intensity for different laser
fluences is 1.0 at the beginning of ablation for the relaxation time of vaporization is
considered. The plasma around the surface is absent at this stage, so the transmitted intensity
is a constant and the plasma hardly emerges. However, during τ m ≤ t ≤ τ , the values of the
transmitted intensity for different laser fluences are different each other. It can be explained
that during this period, vapor begins to appear around the surface and interact with following
part of the laser beam, therefore, the plasma appears. The IB photo-absorption processes
shields part of laser energy and produces the plasma shielding effect. The higher the laser
fluence, the shorter the relaxation time τ m and the more remarkable plasma shielding effects
are.
4.5. MASS REMOVAL PER PULSE AS A FUNCTION OF LASER

FLUENCE
Figure 8. Mass removal per pulse as a function of laser fluence
Fig. 8 presents the numerical results of two theoretical models, considering both the
effects of vaporization and plasma shielding, and only taking the vaporization effect into
account.
We can learn from Fig. 8 that, the two curves nearly overlap in the range of the laser
2 2
fluence from 1J/cm to 3J/cm and begin to detach gradually when the laser fluence exceeds
2
3J/cm . This phenomenon can be explained that with the laser fluence rising gradually, the
plasma shielding effect becomes more and more remarkable. The numerical results of
considering the effects of vaporization and plasma shielding produce a good agreement with
the experimental data. It is clear that the modified model considering the vaporization effect
and the plasma shielding effect can describe the ablation process more accurately.
4.6. CONCLUSION
We have reviewed our recent research work on improved thermal model and its
application in high-power laser ablation of target, including metal and multi-elemental oxide
superconductor. By using the finite different method, we have analytically presented the
satisfactorily good agreement between our numerical results and experimental results. What
we have discussed in this chapter shows that the vaporization effect, the plasma shielding
effect are detrimental factors in PLA.
5. ACKOWLEDGEMENTS
This work was supported by the National Science Foundation of China through Grant
No.10675048, No.10604017 and No.50872038.
REFERENCES
[1] D. Dijkksmp, T. Venkatesan, X. D. Wu, App. Phys. Lett. 51(8), 619 (1987).
[2] S. Amoruso, B. Toftmann and J. Schou, Phys. Rev. E 69, 056403(2004).
[3] S. Amoruso, Appl. Phys. A 69, 323(1999).
[4] B. Toftmann, J. Schou, T. N. Hansen and J. G. Lunney, Phys. Rev. Lett. 84, 3998
(2000) .
[5] R.R. Fang, D.M. Zhang, Z.H. Li, F.X. Yang, L Li, X.Y. Tan and M. Sun, Solid State
Commun. 145, 556 (2008).
[6] R.R. Fang, D.M. Zhang, Z.H. Li, L Li, X.Y. Tan and F.X. Yang, Phys. Status Solidi A
204, 4241 (2007).
[7] A. Inam, X.D. Wu, T. Venkatesan, S.B. Ogale, C.C. Chang, D. Dijkkamp, Appl. Phys.
Lett. 51, 1112 (1987) .
[8] D.M. Zhang, L. Li, et al., Acta Phys. Sin. 54, 1283 (2005) .
[9] E.G. Gamaly, A.V. Rode, B. Lutter-Davies, Phys. Plasmas 9, 949 (2002).
[10] Y.L. Huang, F.H. Yang, G.Y. Liang, J.Y. Su, Chin. J. Lasers 30, 449 (2003).
[11] C.H. Huang, H.Y. Shen, Z.D. Zeng, Opt. Laser Technol. 22, 345 (1990).
[12] C.H. Huang, G. Zhang, et al., Opt. Laser Technol. 34, 209 (2002).
[13] T. Kunii, J. Kitao, K. Mori, N. Yoshida and S. Nonomura, Sol. Energy Mater. Sol. Cells
74, 415 (2002) .
[14] L. Li, D.M. Zhang, Z.H. Li, Phys. B 383, 194 (2006).
[15] S. Guo, Electric Dynamic, (Higher Education Publishing Company, 1997).
[16] T. Venkatesan, S.B. Ogale, C.C. Chang, D. Dijkkamp, Appl. Phys. Lett. 51, 1112
(1987).
[17] P. Andrea, M. Antonio, Phys. Rev. Rev. E 50, 4716 (1994).
[18] C. W. Sun, Q. S. Lu, Z. X. Fan, Y. Z. Chen, The effect of Laser Irradiation (National
Defence Industry Press, Beijing, 2002).
[19] B.H. Billings, H.P.R. Frederikse, et al., American Institute of Physics Handbook. 3rd
ed.(McGraw-Hill, New York, 1972).
[20] C. Knight, AIAA J. 17, 519 (1979).

[21] Ya.B. Zel¡¯dovich, Yu.P. Raizer, Physics of Shock Waves and High-Temperature
Hydrodynamics Phenomena, Vol. 1 (Academic Press, New York, 201 1966).
[22] Q.M. Lu, S. Mao, X.L. Mao, and R.E. Russo, Appl. Phys. Lett.80, 3072 (2002).
[23] R.K. Singh, J. Narayan, Phys. Rev. B 41, 8843 (1990).
[24] Duanming Zhang, Zhihua Li et al,Science In China (Series A) 44, 1485 (2001).
[25] [F. Garrelie, J. Aubreton, A. Catherinot, J. Appl. Phys. 83, 5075 (1998).
[26] J.G. Lunney, R. Jordan, Appl. Surf. Sci. 127, 941 (1998).
[27] Y.S. Touloukian, R.W. Powell, C.Y. Ho, and P.G. Klemens, Thermophysical
Properties of High Temperature Solid Materials (Macmillan, New York, 1976).
[28] CRC Handbook of Materials Science, Vol.1 General Properties, edited by C.T. Lynch
(CRC Press, Cleveland, 1974).
[29] S. Amoruso, M. Armenante, V. Berardi, R. Bruzzese, N. Spinelli, Appl. Phys. A 65,
265(1997).
[30] M. Capitelli, A. Casavola, G. Colonna, A. De Giacomo, Spectrochimica Acta Part B
59, 271(2004).
[31] A. Bogaerts, Z. Chen, Spectrochimica Acta Part B,60, 1280 (2005).
[32] T. Grismayer, P. Mora, Physics of Plasmas 13, 032103 (2006).
[33] M. Murakami, Y.-G. Kang, K. Nishihara, S. Fujioka, and H. Nishimura, Physics of
Plasmas 12, 062706 (2005).
[34] S. Amoruso, Appl. Phys. A 69, 323 (1999).
[35] S. Amoruso, B. Toftmann and J. Schou, Phys. Rev. B 69, 056403(2004).
[36] S. Amoruso, G. Ausanio, R. Bruzzese, M. Vitiello, and X. Wang, Phys. Rev. B 71,
033406 (2005).
[37] P. Villars, J. C. Phillips, Phys. Rev. B 37, 2345(1988).
[38] W. Chen, G. H. Cao, R. Z. Ma, Chinese journal of chemical physics 7 263(1994).
[39] D. M. Zhang, L. Guan, Z. H. Li, Chin. Phys. Soc. 52, 242(2003).
[40] N. M. Bulgakova, A. V. Bulgakov, Appl. Phys. A. 73, 199 (2001).
[41] D. Bhattacharya, R. K. Singh and P. H. Holloway, J.Appl. Phys. 70, 5433 (1991).
SHORT COMMUNICATIONS
Short Communication A
DYNAMICS OF EXCITONS CONFINED IN

SEMICONDUCTOR THIN FILMS
Osamu Kojima1 and Toshiro Isu2

1
Department of Electrical and Electronics Engineering, Graduate School of Engineering,
Kobe University, 1-1 Rokkodai, Nada, Kobe 657-8501, Japan
2
Department of Nano-Technology, Institute of Technology and Science,
The University of Tokushima, 2-1 Minamijosanjima-cho,Tokushima 770-8506, Japan
ABSTRACT
We report the dynamics of confined exctions when several exciton states are excited.
The dynamics are measured by using a transient grating (TG) technique and a degenerate
four-wave-mixing (DFWM) technique. In the system of the quantization of exciton
center-of-mass motion, the excitons have several levels in the narrow energy region. The
transient signals generated by the excitation of several exciton levels by using an
ultrashort-pulse laser with a broadband spectrum are different from those generated by
the excitation of single exciton level. The excitons confined in GaAs films with a specific
thickness show enhanced excitonic optical nonlinearity, which enhancement is described
by the nonlocal response theory, and an ultrafast response that is comparable to the pulse
width of the DFWM signals. This result is an important factor in the development of
ultrafast optical devices that exhibit large optical nonlinearity and ultrafast response.
Moreover, the TG signals show a long-lived component, which is longer than that of the
single exciton level. The changes in the response profile can be attributed to the
excitation by a spectrally broadband pulse. Our results will provide impetus to the
development of ultrafast optical devices based on the excitonic optical nonlinearity.
I. INTRODUCTION
The development of epitaxial growth techniques has made it possible to fabricate various
semiconductor nanostructures such as quantum wells and quantum dots for applications to
novel optical devices. Ultrafast processes in nanostructured semiconductors are currently the
subject of intense research. A large number of studies have been carried out to investigate
442 Osamu Kojima and Toshiro Isu
their application to optical devices, including all-optical switching devices, used in the field
of information and communication technology [1, 2] and to elucidate their excitonic
properties in the ultrashort time scale [3]. In these structures, the optical properties
remarkably depend on the confinement size [4]. In the strong confinement regime, the
decrease in the electron–hole relative motion causes the enhancement of the oscillator
strength [5], resulting in a large nonlinear susceptibility. On the other hand, as the
confinement of the excitons becomes weaker, the coherent extension of the center-of-mass
motion causes the size-linear enhancement of the oscillator strength, and the nonlinear
susceptibility is enhanced in proportion to confinement size [6, 7]. Moreover, when the
excitons are confined in thin films with the thicknesses considerably larger than the Bohr
radius and close to the light wavelength, the excitons are characterized by a nonlocal response
[8−10]. The excitonic properties based on the nonlocal response are different from those
obtained from the long-wavelength approximation in the usual exciton confinement regime.
For instance, the transition selection rule changes from odd parity [11, 12] to even parity as a
result of the phase difference between the exciton wave functions and the light wave
functions.
In the case of a nonlocal response, the enhancement of the optical nonlinearity of
spatially extended excitons, due to the size-resonant enhancement of the internal field using a
nano-scale spatial structure, has been an attractive topic in theoretical and experimental work
[8−10, 13−15]. This effect is called nonlocality induced double resonance in energy and size
(NIDORES) [9]. The strong nonlinear optical response [13, 14] and the fast population decay
induced by NIDORES effect [15] of excitons confined in the thin films with a specific
thickness indicate the possibility of overcoming the incompatibility between the optical
nonlinearity and response time.
In the case of a nonlocal response system, several exciton states exist in a narrow energy
region because of the small quantum confinement effect [16]. The excitons have to be excited
by using an ultrashort-pulse laser with a broadband spectrum for the realization and/or
observation of their ultrafast response. When several exciton levels are simultaneously
excited by an ultrashort pulse, interference phenomena such as excitonic quantum beats or

polarization beats will be expected [3]. These phenomena are never observed when single-
exciton levels are excited by using picosecond pulses with a narrow-band spectrum.
Therefore, new dynamical characteristics can be expected to be exhibited when several
exciton levels are excited by using ultrashort pulses with a broad spectral width.
In this study, we demonstrate the dynamical behavior of excitons confined in the GaAs
thin films. This review is arranged as follows. In section II, we briefly describe the
characteristics of our sample and the experimental measurement system for the measurement
of the exciton dynamics. In sections III and IV, the dynamics of confined excitons when
several exciton levels are simultaneously excited are presented. Finally, in section V, our
results are summarized and we comment the future works. Our results provide some
guidelines for the development of ultrafast optical devices operating approximately at the rate
of terabits per second.
II. EXPERIMENT
II-1. Sample Structure
We have used a double-heterostructure (DH) sample with three periods of GaAs(110

nm)/Al0.3Ga0.7As(5 nm) on a (100) GaAs substrate grown by molecular beam epitaxy, as
shown in Figure 2.1. The Al0.3Ga0.7As barrier layer has sufficient thickness to confine the
excitons in the GaAs thin films [17, 18]. The enhancement of the nonlinear optical response
of the n = 2 exciton has been reported for a film thickness of approximately 110 nm by
coupling with the radiative field [13, 14].
In Figure 2.2, the photoluminescence (PL) spectrum of the sample is depicted. The
origins of the peaks are determined by comparing the spectrum with the results calculated by
the quantization of the center-of-mass motion. In order to estimate the exciton energies
numerically, we have solved the Schrödinger equation for the excitons confined in a finite
potential well using the effective mass approximation [19]. Actually, we should consider the
energy shift owing to coupling effect with the light fields in the calculation [10]. However, in
the case of the GaAs thin film with the thickness of 110 nm, the calculation results using the
simple effective mass approximation almost agrees with that considering the coupling effect
with light. Then, we do not cite the energy shifts owing to the coupling effect through this
review. In the calculation, we used the effective film thickness of 95.6 nm, considering the
exciton dead layer. Moreover, we used the bulk exciton energy of 1.5154 eV. The estimated
confined exciton energies are indicated by the dashed lines in Figure 2.2. The notations HHn
and LHn indicate the n-th quantized heavy-hole (HH) and light-hole (LH) exciton states,
respectively.
Figure 2.1. Sample structure.
Figure 2.2. PL spectrum of sample.
Here, the even parity is interpreted on the basis of the nonlocal response theory
mentioned above. In the case of nonlocal response regime, the phase difference between the
light waves and exciton waves should be considered, and the transition probability is
determined by the integration of the wave function of excitons multiplied by the resonant
electromagnetic field. Therefore, for a thickness of 110 nm, only the excitons with even
quantum numbers are allowed, which is the one of the unique characteristics of the sample.
II-2. Measurement of Exciton Dynamics

The excition dynamics were measured by using a transient-grating (TG) technique and a
degenerate four-wave-mixing (DFWM) technique at 5 K. The plots of TG and DFWM signals
as a function of time delay provide the information on the population decay and dephasing of
resonantly excited excitons, respectively [3]. The schematic of the measurement system is
shown in Figure 2.3. A mode-locked Ti:sapphire pulse laser with a pulse width of 160 fs was
used as the light source. The central laser photon energy was tuned at 1.5158 eV. The spectral
width of the laser pulse was 20 meV, which is sufficient to excite the several exciton states.
The pulse train is split into three parts that are focused to a single 100-μm-diameter spot on
the sample. The excitation power was kept at 100 μW, which corresponds to an excitation
density of 12 nJ/cm2. Pulse 1 and pulse 2 were copolarized, and pulse 3 was cross polarized.
The TG signals diffracted in the direction of k1+k2−k3 were recorded as a function of the time
delay between pulse 2 and pulse 3 when the time delay between pulse 1 and pulse 2 is 0 fs.
The DFWM signals diffracted in the direction of 2k1−k2 were also recorded. In order to
demonstrate the effects of the excitation spectral width on the exciton dynamics, we
introduced a spectral-width control system. The system consists of the two gratings and a slit.
By placing the slit between the grating pair, the spectral width of the pulses (ΔE) can be
controlled in the range from 0.8 to 18.1 meV without changing the central energy of 1.5158
eV.
Figure 2.3. Schematics of measurement system.
III. APPEARANCE OF LONG-LIVED COMPONENTS INDUCED BY

ULTRAFAST-PULSE EXCITATION
Figure 3.1 shows the TG signal of the sample in the time domain [17]. The rise time
(τRise) of the signal is considerably longer than the pulse width, and the decay time (τDecay)
exceeds 30 ps. τRise and τDecay evaluated by fitting the curve with a single-exponential function
are 4 and 120 ps, respectively. τDecay observed in this measurement is considerably longer
than that expected by the nonlocal response theory and that measured by exciting single-
exciton level [15].
In order to elucidate the origin of the long-lived component, TG spectra were measured at
various time delays. Figure 3.2 depicts the time-resolved TG spectra of the sample. The
dashed lines indicate the exciton energies estimated from the PL spectrum. All the spectra
show excitonic peaks, and the HH2 exciton exhibits the maximum intensity. These facts
indicate that the TG signals in all the time regions are generated due to the nonlinear optical
response of the confined excitons, and the nonlocal response is valid when several exciton
levels are excited, except for the measured τDecay. Moreover, even at −20 ps, the excitons
show a nonlinear response. Thus, we have deduced that these phenomena are caused by the
excitation of several states.
Next, we examined the relation between ΔE and the exciton decay profile in order to
determine the origin of the long-lived component. Figure 3.3 shows the ΔE dependence of the
TG signals [20]. The excitation power was kept at 100 μW for all ΔE values. Initially, we
focus on the TG signal generated by pulses of ΔE = 0.9 meV. In this case, the laser pulse
almost excites a single exciton state.
Figure 3.1. TG signal of sample at 5 K.

Figure 3.2. Time-resolved TG spectra. The dashed lines indicate the quantized exciton energies.
The TG signal has two decay components. The fast component of the TG signal is caused
by the radiative decay process arising from the coherent processes based on the nonlocal
response [15]. This two step decay process has also been reported by Schultheis et al. [21].
Figure 3.3. TG signals observed at various ΔE. All signals are normalized at the maximum intensity.
With an increase in ΔE, the fast component becomes smaller and disappears at ΔE = 15.3
meV. We have already stated that the confined excitons contribute to the TG signals in all
time regions. Hence, the disappearance of the fast component may be attributed to that of the
coherent radiative processes due to the nonlocal response induced by the excitation of several
exciton states, and not due to the reduction in the excitonic optical nonlinearity.
In order to reveal the origins of the change in the decay profile, we measured the TG
spectra at various ΔE values. Figure 3.4 shows the TG spectra at zero delay for various ΔE
values; the laser spectra are depicted by dotted curves. Every TG spectrum shows a sharp
peak at the exciton energy of 1.5158 eV. The full-width at half maximum of the excitonic
peaks is almost the same as that at ΔE=15.3 meV. Hence, the disappearance of the fast
component in the time-domain signal cannot be explained by the high-density excitation
effects. If an increase in the exciton density results in the exciton screening and the exciton–
exciton scattering effects, the excitonic peaks will show an energy shift and/or broadening of
the line width. However, the results show the same spectral shape. Thus, we concluded that
the multiexciton effects caused by an increase in the excitation density with a decrease in the
pulse width hardly contribute to the disappearance of the fast component.
In the case of ΔE = 3.3 meV, the TG spectrum displays higher-order exciton peaks, which
indicates the simultaneous excitation of several exciton levels. This simultaneous excitation
induces a modulation of the population dynamics of fundamental excitons. Moreover, Kanno
et al. have reported the rise of PL under nonresonant excitation conditions in a confined
exciton system [22]. The slow rise of PL is caused by the exciton formation process in states
with large wave vectors which do not couple to light waves directly when the excitation
energy is higher than the exciton energies [23, 24]. In addition, recent measurement by using
a pump-probe technique has clarified the contribution of the bulk-like exciton-polariton
propagation to the transient dynamics of confined excitons [25]. The slow signal rise in the
negative time-delay region, as shown in Figures. 3.1 and 3.3, originates from the exciton-
polariton propagation. The propagation time estimated from the signal rise time and film
thickness almost corresponds to the group velocity of the bulk exciton-polariton, which
indicates that the bulk-like exciton-polariton propagation is a significant phenomenon
affecting the confined exciton dynamics even in the thin film. Therefore, our results
demonstrate that the simultaneous excitation of several exciton states, the exciton formation
process, and the contribution of nonresonant excitation components lead to the disappearance
of the coherent processes and changes in the exciton decay profiles.
Figure 3.4. ΔE dependence of TG spectrum (solid curves) observed at zero delay. The dotted curves are
laser profiles.
IV. ULTRAFAST NONLINEAR RESPONSE INDUCED BY EXCITATION OF

SEVERAL EXCITON STATES
When the several exciton states are simultaneously excited by an ultrashort pulse,
interference phenomena such as excitonic quantum beats or polarization beats will be
observed [3]. If the excitons under the NIDORES conditions demonstrate an ultrafast
response as a result of these interference effects, ultrafast response based on these interference
phenomena will be expected. In this section, we demonstrate ultrafast nonlinear optical
responses observed in the DFWM signals when the confined excitons in GaAs thin films are
excited by using an ultrafast pulse under the NIDORES condition.
Figure 4.1 shows the ΔE dependence of the DFWM signals at 5 K [18]. The intensity of
each DFWM signal was normalized by the maximum intensity. The decay profile at ΔE = 0.8
meV agrees closely with the result obtained using a picosecond-pulse laser with ΔE = 0.7
meV [26]. The signal decay time (τ) at ΔE = 0.8 meV is estimated to be 1.72 ps by fitting a
single exponential function to curve for the signal after 7 ps. Since the dephasing time of the
excitons under inhomogeneous broadening is given by T2 = 4τ [27], that of the single
confined exciton state under the NIDORES condition is evaluated to be 6.88 ps.
Figure 4.1. Δ E dependence of the DFWM signals.
The signals excited using laser pulses with larger ΔE indicate an oscillatory structure and
an ultrafast response comparable to the pulse width at around zero delay. The ultrafast
population decay due to elastic or inelastic scattering of excitons formed by the excitation of
several exciton states and the pure dephasing of the excitons are assumed to be the possible
factors behind the ultrafast response; hence, these factors should be considered first.
However, the response time, which is comparable to the pulse width, is considerably shorter
than the radiative decay time expected from the nonlocal response theory. Moreover, the long
exciton lifetime exceeding 100 ps, determined by a TG technique using the ultrashort pulse as
described previously, clearly denies contribution of the population decay to the ultrafast
response. Furthermore, since there is a component with a decay time of 1.69 ps around the
signal tail of ΔE = 18.1 meV agreeing with the signal decay time of the single state of the
confined excitons, as mentioned above, the origin of the ultrafast response is not pure exciton
dephasing. Therefore, the population decay and pure dephasing of the excitons are
insufficient to explain the ultrafast response. In order to determine the origin of the ultrafast
response, the DFWM spectra obtained by the excitation of exciton states using pulses of ΔE =
18.1 meV were measured at various time delays.
Figure 4.2. Image map of the DFWM spectra at various time delays with a step of 100 fs. The
brightness is proportional to the DFWM intensity. The solid curve indicates the numerically summed
spectrum in from –2.0 to 2.0 ps. The dashed lines indicate the confined exciton energies.
Figure 4.2 shows the image map of the DFWM spectra at various time delays with a time
step of 100 fs, where the brightness is proportional to the DFWM intensity. The solid curve
indicates the numerically summed spectrum for time delay from –2.0 to 2.0 ps. The DFWM
spectra clearly exhibit several peaks around zero delay in addition to the main peak at the
energy of the HH2 exciton. Figure 4.3 shows the temporal evolution at each peak shown in
Figure 4.2 with a spectral resolution of 0.2 meV. The top profile is the spectrally integrated
signal, which corresponds to the time-domain signal shown in Figure 4.1. All excitonic peaks
rapidly disappear within 200 fs, which is consistent with the characteristic of time-domain
signals. This result indicates that the excitons confined in several states simultaneously
contribute to the ultrafast response, whereas exciton dephasing is a negligible factor. Then,
we focused on the oscillatory structure in the time-domain signals, which causes the ultrafast
response.
For the clarification of the origins of the oscillatory structures, the excitation energy
dependence of the DFWM signals was measured using pulses of ΔE = 5.4 meV, as shown in
the linear (a) and logarithmic (b) plots in Figure 4.4. The intensity of each signal was
normalized by the maximum intensity. In the linear plot in Figure 4.3(a), while the pulse
width is constant, the response time around zero delay decreases with an increase in the
excitation energy, and the oscillatory period varies with the excitation energy. These
characteristics of the excitation energy dependence resemble those observed in the ΔE
dependence. In order to evaluate the oscillatory period, we performed fitting the curve with
the following equation describing the excitonic quantum beats in the DFWM signals [3]
⎛ ω2 ω2 2ω1ω 2 ⎞ 2 −2 Γ1t
I (t ) ∝ ⎜⎜ 1 + 2 + (
⎟ ω1 e
⎟ )
+ ω 22 e − 2 Γ2t + 2ω1ω 2 cos(ΔE ES t )e −(Γ1 + Γ2 )t ,
⎝ 2Γ1 2Γ2 (Γ1 + Γ2 ) + ΔE ES
2 2
⎠
(1)
where Γ1 and Γ2 are the dephasing rates of two transitions; ω1 and ω2 are the spectral weights
of two transitions; ΔEES is the energy separation of two exciton states; and t is the time delay.
The values of these parameters evaluated by fitting the curve with eq. (1) are listed in Table I.
Figure 4.3. Temporal evolution at each exciton energy in Figure 4.2 the spectral resolution of 0.2 meV.
Table I. Parameter values obtained by fitting curves with eq. (1)

Excitation energy (eV) ω1 ω2 Γ1 Γ2 ΔEES (meV)

1.5151 0.11 0.78 0.65 0.32 3.05
1.5159 0.19 0.77 0.65 0.37 3.08
1.5166 0.17 0.85 0.94 0.55 4.53
1.5174 0.12 0.77 0.65 0.34 5.36
The values of ΔEES corresponding to the oscillatory periods agree closely with the energy
separations of the HH2–LH4, HH2–HH6, HH2–HH8, and HH2–HH10 excitons at the
excitation energies of 1.5151, 1.5159, 1.5166, and 1.5174 eV, respectively. Thus, the
oscillatory structure is attributed to interference effects of the confined exciton states. In the
case of an interference with excitons from the higher states, such as the HH10 exciton, the
values of the exciton energy calculated using the confinement model of the center-of-mass
motion are slightly different from the experimental values due to low barrier height.
However, we observed the peak of the HH10 exciton in the DFWM spectra of a 120-nm-
thickness sample [28]; therefore, it concluded that the excitons from the higher states
contribute to the temporal oscillation. In addition, it should be noted that the values of Γ1 are
almost constant and larger than Γ2 at all excitation energies. This Γ1 represents the dephasing
rate of HH2 excitons having the maximum value in all exciton states and determines the
signal decay. The values of signal decay times estimated by fitting are approximately close to
those signal decay times at 1.5136 and 1.5143 eV, shown by the dotted lines, estimated by
fitting with the single-exponential function in Figure 4.4 (b). Therefore, the dephasing of the
n = 2 exciton gives the decay time of the DFWM signals. These results indicate that the
excitation of several exciton states by a pulse with a broadband spectrum induces an ultrafast
response as a result of the interference effect of the exciton states.
Figure 4.4. Excitation energy dependence of DFWM signals generated using pulse with Δ E = 5.4 meV:
(a) linear and (b) logarithmic plots. The dotted lines indicate the decay time fitted with a single-
exponential function.
SUMMARY
We have investigated the dynamics of confined excitons when several exciton states are
excited by using an ultrashort-pulse laser with a broadband spectrum. The excitonic optical
nonlinear response confined in the thin films with the thickness close to the light wavelength
is characterized by the nonlocal response under the excitation of the single exciton state. On
the other hand, we observed a variation in the exciton dynamics induced by the excitation of
several exciton states due to excitation of the several exciton states. In the exciton population
dynamics, the longlived excitons appear, which exciton lifetime is considerably longer than
that observed by single-exciton-state excitation. This indicates that the excitons from the
higher states, the nonresonant components, and the exciton formation process modify the
dynamics of the fundamental exciton state, and the longlived component appears. Moreover,
a slow rise in the signal due to the exciton-polariton propagation was also observed. On the
other hand, in the DFWM signals, an ultrafast response due to the interference effects of the
exciton states was observed. In the analysis of the oscillatory structures, the temporal
oscillation is attributed to the quantum beats of the confined excitons. While this ultrafast
response due to the interference implies the new possibility for the ultrafast operation (e.g.
Kerr rotation [29]), we have also observed the decrease in the signal intensity due to the
residual exciton under the multiple pulse incident condition [30]. Therefore, the ultrafast erase
of the residual excitons, e.g. by using Rabi oscillation [31], will be important. However, the
present demonstration of the modification in the exciton dynamics induced by the excitation
of several exciton states may aid the development of ultrafast optical devices operating at the
rate of terabits per second.
ACKNOWLEDGEMENT
We would like to thank M. Tsuchiya, J. Ishi-Hayase, M. Sasaki, A. Kanno, and R. Katouf
from the National Institute of Information and Communications Technology (NICT); O.
Wada, T. Kita, S. Watanabe, and A. Miyagawa from Kobe University; and H. Ishihara from
Osaka Prefecture University for their fruitful discussions. These works were partially
supported by CREST in Japan Science Technology Agency and the Grant-in-Aid for the
Scientific Research and from the Ministry of Education, Culture, Sports, Science, and
Technology of Japan.
REFERENCES
[1] J. M. Hvam, in Ultrafast Photonics, eds. A. Miller, D. T. Reid, and D. M. Finlayson
(IOP Publishing, Bristol, U.K., 2004) p. 265.
[2] Wada, O. New J. Phys. 2004, 6, 183-1-35.

[3] J. Shah, in Ultarafast Spectroscopy of Semiconductors and Semiconductor
Nanostructures, ed. M. Cardona, Springer Series in Solid-State Sciences, Vol. 115
(Springer, Berlin, 1996).
[4] M. J. Kelly, in Low-Dimensional Semiconductors: Materials, Physics, Technology,
Devices, Series on Semiconductor Science and Technology Vol. 3 (Clarendon, Oxford,
1995).
[5] Takagahara, T. Phys. Rev. B 1987, 36, 9293-9296.
[6] Hanamura, E. Phys. Rev. B 1988, 37, 1273-1279.
[7] Belleguie, L.; Mukamel, S. J. Chem. Phys. 1994, 101, 9719-9735.
[8] Ishihara, H.; Cho, K. Phys. Rev. B 1993, 48, 7960-7974.
[9] Ishihara, H.; Cho, K. Phys. Rev. B 1996, 53, 15823-15833.
[10] Ishihara, H. J. Phys. Condens. Matter 2004, 16, R247-R273.
[11] Tang, Z. K.; Yanase, A.; Yasui, T.; Segawa, Y.; Cho K. Phys. Rev. Lett. 1993, 71,
1431-1434.
[12] Silva-Castillo, A.; Perez-Rodriguez, F. J. Appl. Phys. 2001, 90, 3662-3664.
[13] Akiyama, K.; Tomita, N.; Nomura, Y.; Isu, T. Appl. Phys. Lett. 1999, 75, 475-477.
[14] Ishihara, H.; Cho, K.; Akiyama, K.; Tomita, N.; Nomura, Y.; Isu, T. Phys. Rev. Lett.
2002, 89, 17402-17405.
[15] Akiyama, K.; Tomita, N.; Nishimura, T.; Nomura, Y.; Isu, T.; Ishihara, H.;Cho, K.
Physica B 2002, 314, 293-296.
[16] Tredicucchi, A.; Chen, Y.; Bassani, F; Massies, J.; Deparis, C.; Neu, G. Phys. Rev. B
1993, 47, 10348-10357.
[17] Kojima, O.; Isu, T.; Ishi-Hayase, J; Tsuchiya, M. phys. stat. sol. (c) 2006, 3, 675-678.
[18] Kojima, O.; Isu, T.; Ishi-Hayase, J; Sasaki, M.; Kanno, A.; Katouf, R.; Tsuchiya, M. J.
Phys. Soc. Jpn 2008, 77, 044701-1-4.
[19] M. Fox, in Optical Properties of Solids, Oxford Master Series in Condensed Matter
Physics (Oxford University Press, 2001) p119.
[20] Kojima, O.; Isu, T.; Ishi-Hayase, J; Sasaki, M.; Kanno, A.; Katouf, R.; Tsuchiya, M. J.
Lumin. 2008, 128, 966-968.
[21] Schultheis, L.; Kuhl, J.; Honold, A.; Tu, C. W. Phys. Rev. Lett. 1986, 57, 1797-1800.
[22] Kanno, A.; Katouf, R.; Kojima, O.; Ishi-Hayase, J.; Sasaki, M.; Tsuchiya, M.; Isu, T. J.
Lumin. 2008, 128, 1069-1071.
[23] Damen, T. C.; Shah, J. ; Oberli, D. Y.; Chemla, D. S.; Cunningham, J. E.; Kuo, J. M.
Phys. Rev. B 1990, 42, 7434.
[24] J. Kusano, Y. Segawa, Y. Aoyagi, and S. Namba, Phys. Rev. B 1989, 40, 1685-1691.
[25] S. Watanabe, O. Kojima, T. Kita, O. Wada, T. Isu, phys. stat. sol. (c), in press.
[26] K. Akiyama, N. Tomita, Y. Nomura, and T. Isu, Physica B 1999, 272, 505-508.
[27] T. Yajima and Y. Taira, J. Phys. Soc. Jpn. 1979, 47, 1620-1626.
[28] O. Kojima, T. Isu, J. Ishi-Hayase, M. Sasaki, and M. Tsuchiya: Physics of
Semiconductors: 28th International Conference on the Physics of Semiconductors, AIP
Conf. Proc. 2007, 893, 457-458.
[29] Kanno, A.; Katouf, R.; Kojima, O. Ishi-Hayase, J.; Tsuchiya, M.; Isu, T. phys. stat. sol.
(c) 2008, 5, 360-363.
[30] Kojima, O.; Isu, T.; Ishi-Hayase, J; Sasaki, M.; Kanno, A.; Katouf, R.; Tsuchiya, M.
phys. stat. sol. (c) 2008, 5, 2858-2860.

[31] Kojima, O.; Miyagawa, A.; Kita, T; Wada, O.; Isu, T. Appl. Phys. Express 2008, 1,
112401-1-3.
Short Communication B
HIGH POWER LONG PULSE WIDTH QCW LASER

DIODE BARS FOR OPTICAL PUMPING OF YB-ER
GLASS SOLID STATE LASERS
N. I. Katsavets1, V. A. Buchenkov2 and A. L. Ter-Martirosyan1

1
ATC-Semiconductor Devices Joint-Stock Company, St. Petersburg, Russia
2
Jenoptik Components Company, St. Petersburg, Russia
ABSTRACT
This publication is devoted to development of 100 W quasi-continues wave 950 nm
long pulse width (5 ms) laser diode bars. Results on optical output power and
degradations characteristics within wide temperature range (from -400C to +850C) are
presented. Developed LDBs are based on advanced single quantum well separate
confinement heterostructures with index-graded waveguide design optimized for
generation wavelength 950 nm at +250C, low threshold current density (70 A/sm2), high
characteristic threshold current temperature (T0=150 K) and good slope efficiency
(1,1 W/A).
Estimated LDB lifetime in long pulse width regime is more than 109 and 108 pulses
at heat-sink temperature +250C and +850C, respectively.
These laser diode bars are developed for effective pumping of Yb-Er glass solid state
lasers (SSLs) with generation within the “eye-safe” spectral range (1.54 µm). Design of
Yb-Er glass SSLs using for optical pumping two 950 nm laser diode bars was developed.
Such SSLs have the pulse energy up to 8mJ at multimode regime, the pulse width 20 ns
and repetition rate up to 20 Hz.
Solid state lasers (SSLs) are widely used for science and technical applications.
Significant interest is attracted to SSLs which generate radiation in the “eye-safe” optical
range (1.5 µm ÷ 2.0 µm) [1-3]. Radiation of Yb3+- Er3+ (Yb-Er) doped glass SSLs is in this
range. Optical radiation in 940÷970 nm spectral range effectively pumps of ions Yb.
Absorption energy is then non-radiatively transferred to the “pump” level of ions Er with
generation spectral line at 1.54 µm [1]. Due to durable lifetime of the metastable lasing
manifold of ions Er (about 5 ms) Yb-Er glass SSLs have possibility to generate high-energy
optical pulses at relatively low optical pumping density.
456 N. I. Katsavets, V. A. Buchenkov and A. L. Ter-Martirosyan
High power laser diode bars (LDBs) with generation wavelength 940÷970 nm can be
successfully used for effective optical pumping of Yb-Er doped glass SSLs. However, long
pulse width regime results in significant thermo-cycle stress in the LDB crystal [4] and makes
additional demands to laser heterostructure post-growth processing and crystal bonding
technologies.
In this publicity results of development of high power quasi-continues wave (QCW)
950 nm LDBs operated in long pulse width (up to 5 ms) regime are presented. They have high
efficiency and can reliably operate in the wide temperature range (from -400C to +850C). The
publication also presents SSLs based on Yb-Er glass with optical pumping by developed
LDBs.
Original design of laser wafers based on single quantum well AlxGa1-xAs/InyGa1-yAs
/GaAs heterostructures with separate electronic and optical confinement having the index-
graded waveguide [5] was developed. This heterostructure design was optimized for
generation wavelength 950 nm at +250C, low threshold current density, high characteristic
threshold current temperature and good slope efficiency [6].
Figure 1 shows the composition diagram X(d), Y(d) of grown heterostructures as well as
calculated distribution of optical intensity I(d) in the waveguide (d – distance from the active
region). The calculation was made using the SiLENSe program [7] adapted to cubic III-V
compounds. The calculation gives the optical confinement factor Г = 0,034 and a full width
of half maximum (FWHM) of fast axis far-field laser radiation distribution (in the direction
perpendicular to the p-n junction) Θ⊥= 400 which agrees well with calculated value. Such
laser wafers were grown by MBE technique. Technology installation features were presented
in [8]. Measuring threshold current as well as quantum efficiency of laser diodes made from
the grown laser wafer versus cavity length gives transparency threshold current density 70
A/sm2, intrerior quantum efficiency about 100% and interior optical losses 1,5 sm-1.
Figure 1.
High Power Long Pulse Width QCW Laser Diode Bars for Optical Pumping … 457
LDB crystal (chip) design was the array (50 units) of optically isolated broad-area laser
diodes (LDs) with stripe-width 160 µm, period of 200 µm and cavity length 1000 µm. High
reflection (95%) and antireflection (5%) coating were deposited on the laser chip mirrors with
electron beam vacuum technique after surface passivation treatment.
Special equipment was used for bonding of laser chips on the heat-sink with placement
accuracy plus/minus one micron. Chips were bonded p-side down with the In-containing
solder on the Cu-W submounts placed onto the Cu heatsinks. Solder thickness and
temperature regime were optimized for minimum mechanical stress appearing after bonding.
For output optical power measurement the laser power meter “LASERMATE” was used.
Spectral parameters are measured with the spectrometer “MDR-23”.
Typical output optical power dependencies versus pumping current of LDB operated at
different pulse width (300 µs and 5 ms) at duty factor (10%) are presented in the figure 2.
Figure 2.
Figure 3.
Slope efficiency is slightly (from 1.1 W/A to 1.05 W/A) decreased at long pulse width
operating regime. The inset in the figure 2 shows the typical oscilloscope image of the output
optical pulse measured at a pumping current of 110 A. As can be seen, the decay of the output
optical power during the 5 ms pulse does not exceed 5%. This fact says about low LDB active
region overheating at long pulse width regime. On the other hand, low active region
overheating is confirmed also by not-significant red shift and small broadening of generation
spectrum (figure 3).

Figure 4 shows typical output optical power as well as power conversion efficiency
versus pumping current of LDB at different heat-sink temperature (T). As can be seen, from
the figure, output optical power riches to 100W (80W at +850C) without any bend with power
conversion efficiency from 60% (T= -400C) to 40% (T= +850C). Analysis of LDB threshold
current Ith(T) versus heat-sink temperature using expression
Ith (T) = Ith (0) * exp (T/T0)
gives the characteristic temperature [4] T0=150 K.

Figure 5 shows typical spectral characteristics of LDBs at different temperature. FWHM
spectral generation line increases from 3.5 nm to 5.6 nm at heat-sink temperature rise from
-400C to +850C. Shift temperature coefficient of spectrum line maximum is 0.35 nm/degree.
Developed LDBs were exposed to long-term test at +250C and +850C of heat-sink
temperature in the constant pumping current (100A) regime and pulse width 5 ms for 5*107
pulses. Estimation of LDB lifetime defined as pulse quantity for 20% decreasing of initial
output optical power gives the value 109 pulses at +250C and more than 108 pulses at +850C.
Figure 4.
Figure 5.
a) the exterior view of the LDB;

b) the simple schema of Yb-Er glass SSL (1 – the resonator with theYb-Er active crystal, 2 - two LDBs,
3 - mirrors, 4 – the active optical shutter, 5 – the telescope);
c) the exterior view of Yb-Er SSL.
Figure 6.
Reliable operation of developed LDBs at wide temperature range (from -40 to +850C)
and a broad shoulder in the Yb absorption spectrum [3] gives opportunity to create Yb-Er
SSL without thermo-stabilization of pumping system. Such SSL design results in low energy
consumption, construction simplicity and high reliability.
Yb-Er glass SSLs with optical pumping based on developed LDBs were created. Such
SSLs use two LDBs and operate in “eye-safe” range (1,54µm) with energy 8 mJ at multimode
operation, pulse width 20ns and frequency up to 20Hz. The exterior views of developed
LDBs and SSLs with simple optical pumping schema are presented in the figure 6.
In conclusion, high power long pulse width QCW 950nm LDBs operating in wide
temperature range were developed. They are successfully used as optical pumping in SSLs
based on Yb-Er glass.
REFERENCES
[1] Hutchinson, J.A. and Allik,T.H. Diode array-pumped Er, Yb: phosphate glass laser
Applied Physics Letters, 1992, Vol.60, 12, pp. 1424-1426.
[2] Buchenkov, V.A. and Nikitichev, A.A. Eye-safe Diode Pumping Solid State Lasers,
Laser-Inform, 2003, Vol.13, pp.268-269.
[3] Hamlin, S. J.; Hays, A. D.; Trussell, C. W.; King, V. Eyesafe Erbium Glass Microlaser,
Proc. of SPIE, 2004, Vol. 5332, pp.97-101.
[4] Amzajerdian, F.; Meadows, B.; Baker, N.; Sudesh, V.; Kavaya, M.; Sudesh,V. Risk
Reduction and Advancement of High Power Quasi-CW Laser Diode Pump Arrays,
Solid State and Diode Laser Technology Review, 2004, P-1.
[5] Casey, H.C. and Panish, M.B. Heterostructure lasers, Academic Press, NY, 1978.
[6] Alexandrov, S.B.; Alekseev, A.N.; Demidov, D.M.; .Dudin, A.L.; Katsavets, N.I.;
Kogan, I.V.; Pogorel’skii, Y.V.; Ter-Marirosyan, A.L.; Sokolov, E.G.; Chaly, V.P.and
Shkurko, A.P. High-Power Low-Threshold Laser Diodes (λ = 0,94 μm) Based on MBE-
Grown In0.1Ga0.9As/AlGaAs/GaAs Heterostructures, Technical Physics Letters, 2002,
Vol.28, 8, pp.696-698.
[7] Bulashevich, K.A.; Evstratov, I.Yu.; Karpov, S.Yu. Hybrid ZnO / III-nitride light-
emitting diodes: modelling analysis of operation, Phys. stat. solidi (a), 2007, 204,
pp.241-245.
[8] Katsavets, N.I.; Buchenkov, V.A.; Demidov, D.M.;.Leus, R.V.; Iskandarov, M.O.;
Nikitichev, A.A.; Ter-Martirosyan A. L. Highly efficient High-Power Quasi-
Continuous Diode Laser Bars for Pumping Solid-State Lasers Based on Yb-Containing
Active Media, Technical Physics Letters, 2004, Vol.30, 12, pp.1039-1041.
INDEX
air, ix, 35, 49, 131, 132, 138, 140, 207, 208, 209,
A 213, 218, 223, 224, 225, 306, 408
Air Force, 19
Aβ, 398
Albino, 238
absorption coefficient, 2, 18, 42, 52, 220, 292, 340, algorithm, 127, 148, 149, 371
397, 416, 417, 418, 419, 425, 426, 430, 433 alkaline, 308
absorption spectra, 282, 284, 291, 292, 293, 294, alkane, 313
298, 302, 303, 338, 339, 357, 359, 363 alloys, 29
absorption spectroscopy, 364 alternative, 46, 139, 395
acceptor, 352, 353, 355, 358, 359, 360, 361, 363 alters, 176
acceptors, 342, 363 aluminates, 324
accounting, 16, 127 aluminium, xi, 315, 427
accuracy, 5, 102, 127, 129, 131, 155, 170, 213, 215, aluminosilicate, 31, 47, 306, 313, 314
216, 220, 222, 457 aluminum, 101, 139, 140
acetate, 143 aluminum oxide, 139, 140
acetonitrile, 355, 356, 357, 359 Alzheimer’s disease, 269, 271
acid, 143, 402 ambient air, 131, 132
acoustic, 3, 11, 74, 78, 86, 111, 113 ambient pressure, 223, 224
acoustic waves, 74 amine, 357
acoustical, 110 amino acids, 402, 403

acquisitions, 234, 235 amorphization, 36
actin, 268 amorphous, 30, 190, 241, 259, 308
actinide, 304 amphibians, 273
activation, 268 amplitude, 15, 16, 77, 79, 80, 81, 82, 83, 84, 85, 87,
active site, 304 88, 89, 91, 94, 95, 97, 98, 99, 100, 101, 103, 105,
actuators, 72 106, 107, 114, 115, 126, 148, 151, 212, 217, 243,
adiabatic, 371, 422 256, 339, 377, 385, 386
adjustment, 89, 92, 94, 101, 102, 106, 135, 145, 228 Amsterdam, 66
adult, 244, 248, 276 analog, 74, 76, 100, 111
aerospace, 190 anatomy, 248
AFM, 40, 41, 44, 45, 56, 59, 132 angioplasty, 409
Ag, 32, 33, 34, 37, 39, 40, 41, 42, 44, 45, 46, 47, 48, animal studies, 267, 274
49, 50, 51, 52, 53, 54, 57, 58, 63, 64, 65 animals, x, 232, 236, 237, 238, 239, 249, 270, 402
age, 238, 263, 272, 315 anisotropy, 134, 348, 350, 351, 352
agent, 140 annealing, viii, 27, 30, 31, 35, 38, 45, 46, 49, 50, 51,
age-related macular degeneration (AMD), 263 52, 53, 55, 58, 59, 60, 61, 62, 64, 65, 131
aggregates, 38, 266, 268, 342 annihilation, 349
aid, 63, 153, 154, 286, 453 anode, 134, 135
AIP, 454 anomalous, 336, 337, 341
antagonistic, 319
464 Index
antenna, 349, 351 band gap, 211

anthracene, 380, 381, 384, 385, 386 bandgap, 42, 59, 162, 204
antibody, 259 bandwidth, 9, 16, 18, 84, 318, 320, 329, 331, 334
antifibrotic, 270 barrier, 241, 384, 443, 451
apatite, 325 basal layer, 241
apatites, 333 basement membrane, 240, 241, 243, 247, 261, 262
apoptosis, 277 basic research, x, 232, 239, 270
appendix, 73, 75, 76, 77, 85, 86, 89, 91, 92, 93, 109, basis set, 373
110, 111, 113, 116 beams, vii, viii, 36, 102, 125, 127, 209, 235, 236,
application, ix, x, 5, 15, 27, 58, 61, 64, 74, 82, 86, 238, 263, 266, 267, 268
106, 107, 128, 134, 146, 149, 180, 190, 207, 209, behavior, 15, 16, 17, 18, 57, 64, 65, 83, 84, 85, 87,
231, 267, 270, 289, 301, 313, 316, 324, 331, 340, 101, 103, 110, 129, 162, 167, 197, 294, 304, 338,
343, 371, 395, 409, 411, 417, 437, 442 343, 347, 362, 384, 386, 403, 442
aqueous humor, 15 Beijing, 207, 437
aqueous solution, 47 bell, 353
aqueous solutions, 47 bell-shaped, 353
argon, 134, 140, 338, 397 bending, viii, 125, 144
argument, 135, 348 benefits, 208, 234, 269
arid, 270 benzene, 340, 341, 375, 381
aromatic hydrocarbons, 340 Bessel, 107, 151
arrest, 271 bias, 140
arteries, 257, 409 binding, 372
arterioles, 257 bioassays, 301
artery, 277, 409 bioenergetics, xi, 232, 243, 249, 270
aspect ratio, 227 bioengineering, 190
assessment, x, 232, 234, 257, 258, 270 biological systems, 272, 337
assets, 319 biomechanics, 253
assumptions, 75, 76, 86, 222 biomolecular, 234, 247, 271, 288
asymmetry, 44 biopsies, 259
atmosphere, 305 biopsy, 238
atomic force, 40 biosciences, 277
atomic force microscope, 40 biotechnology, 272
atomic orbitals, 372 bipolar, x, 231, 232, 261, 263

atomic positions, 322 bipolar cells, x, 231, 232, 261, 263
atoms, viii, 27, 36, 37, 39, 44, 46, 48, 49, 52, 53, 54, birefringence, 2, 4, 72, 74, 76, 92, 94, 95, 107, 113,
60, 61, 64, 65, 134, 187, 192, 197, 305, 306, 317, 117
322, 327, 361, 372, 373, 374, 381, 399, 400, 420 bismuth, 44, 46
Austria, 393 bleaching, 333, 359
autofluorescence, vii, x, 231, 232, 235, 236, 237, blocks, 87, 102
241, 243, 244, 245, 247, 248, 249, 250, 251, 252, blood, x, 231, 236, 238, 257, 258, 402
254, 255, 257, 258, 259, 260, 261, 263, 265, 268, blood flow, 257
270, 273, 274, 276, 277, 301 blood vessels, 402
autolysis, 237 Bohr, 373, 442
availability, 232 boiling, 423, 428
averaging, 52, 173, 222 Boltzmann constant, 160, 168, 173, 187, 197, 381,
axon, 261 422, 428
Boltzmann distribution, 423
bonding, 54, 333, 355, 369, 387, 456, 457
B bonds, 54, 210, 306, 373, 376, 397, 398
borate, 304, 306, 313, 314, 325
babies, 409
borosilicate glass, 309
background noise, 9
boundary conditions, 79, 81, 177, 180, 182, 194,
backscattering, 38, 257, 261
199, 200, 422, 423, 430, 431
bacterial, 351
Index 465
Boussinesq, 197 ceramics, 286, 293, 300, 301, 310, 312, 323, 331,
bovine, 271, 276 333
brain, 15, 234, 269, 272, 273, 274, 275, 409 cerebral cortex, 261
breakdown, ix, 64, 198, 207, 218, 223, 224, 270, cerium, 140, 141, 272, 319, 320
276, 397, 399, 400, 420 channels, 226, 227, 228, 235, 286, 370
breathing, 375 chemical bonds, 54, 371, 397
breeding, 149 chemical etching, 128, 134, 136, 137, 140, 143, 146
bremsstrahlung, 210, 420 chemical properties, 40, 284, 409
broadband, xii, 18, 394, 395, 396, 441, 442, 452 chemical reactor, 129
bubble, 195 chemical stability, 140
bubbles, 195, 265, 266, 267, 273, 410, 411 chemokines, 268
buffer, 343 chemotaxis, 268
building blocks, 348 China, 207, 279, 307, 369, 415, 437, 438
Bulgaria, 393 chloride, 203, 253
bulk materials, 190, 208 chlorophyll, 350
burning, xi, 279, 282, 306, 313, 314, 342 choroid, 263
bypass, 409 chromium, 140, 141, 333
chromosomes, 264, 274
circular polarized light, 91, 95, 97, 99
C circulation, 257
cis, 372, 379
Ca2+, 322, 329
classes, 110
CAD, 410
classical, ix, 2, 18, 128, 159, 162, 163, 164, 165,
cadaver, 276
167, 170, 197, 201, 202, 417
calcium, 29, 306, 313
classification, 397
calorimetry, 373
cleaning, 135
CAM, 410
CLSM, 233, 263
Canada, 207
clustering, 16
cancer, 398
clusters, 60, 64, 65
cancer cells, 398
Co, 32, 38, 39, 154, 217, 248, 255, 269
candidates, 370, 415
CO2, viii, 105, 125, 126, 127, 128, 132, 133, 144,
capillary, 232, 257, 258, 260
156, 209, 410
carbon, 157, 373, 374
coatings, 12, 13, 146, 147

carbon atoms, 373, 374
cobalt, 38
carbonization, 398
co-existence, 339
carboxyl, 271, 402
coherence, 14, 92, 261, 272, 273, 274, 397, 410
carboxylic, 313
collagen, 232, 237, 240, 241, 243, 244, 245, 246,
carboxylic acids, 313
247, 248, 249, 251, 252, 253, 254, 255, 256, 257,
cardiac myocytes, 269
258, 259, 260, 265, 266, 268, 269, 271, 272, 273,
cardiomyocytes, 270
274, 275, 276, 277, 402, 404
carrier, 11, 12, 162, 171
collateral, 162, 195, 267, 408, 409, 410, 411
cartilage, 402
collateral damage, 162, 195, 408, 409, 410, 411
CAS, 322
collisions, 36, 42, 53, 210, 211, 428, 429
cathode, 134, 280
colors, 305
cathode ray tube, 280
Columbia, 159
cavitation, 273, 399
combined effect, 327
cavities, 412
communication, 244, 280, 442
C-C, 376
competition, 16, 45, 342, 353
cell, 86, 87, 234, 240, 241, 242, 244, 245, 246, 248,
complement, 153
249, 253, 254, 255, 260, 261, 263, 266, 267, 268,
complexity, 17, 18, 94, 222
269, 273, 277, 322, 339, 341, 411
compliance, 109, 110, 114
cell body, 245
components, x, 28, 108, 109, 113, 221, 231, 234,
cell membranes, 261
240, 241, 245, 248, 249, 251, 253, 255, 258, 260,
ceramic, 286, 287, 301, 311, 334, 410
466 Index
267, 268, 269, 270, 275, 347, 351, 352, 371, 398, core-shell, 286
410, 446, 448, 452 cornea, x, 15, 231, 232, 237, 238, 239, 240, 241,
composites, 29, 30, 38, 42, 47, 51, 402 242, 243, 247, 249, 250, 251, 252, 253, 256, 257,
composition, 29, 149, 247, 254, 275, 306, 402, 403, 260, 263, 264, 265, 266, 267, 268, 271, 272, 273,
456 274, 275, 277, 410, 411
compounds, xi, 29, 134, 279, 332, 336, 337, 355, corneal epithelium, 241
358, 361, 456 correlation, 155, 249, 253, 256, 270, 301, 342, 348
computation, 180, 323 correlations, 340
computer simulations, 132 costs, 140, 411
computing, 127, 148 Coulomb, 212
concentrates, 46 couples, 41, 59
concentration, 2, 29, 34, 37, 38, 39, 44, 45, 47, 48, coupling, 9, 18, 82, 134, 144, 160, 161, 162, 163,
49, 54, 62, 143, 190, 194, 236, 238, 297, 308, 168, 174, 177, 180, 187, 190, 192, 195, 196, 198,
323, 329, 341, 342 200, 202, 210, 317, 319, 320, 321, 331, 332, 340,
concrete, 415 343, 345, 346, 348, 349, 351, 352, 353, 354, 360,
conductance, 418, 419 372, 375, 376, 379, 384, 385, 386, 406, 420, 443
conduction, ix, 42, 59, 159, 161, 162, 163, 164, 165, covalent, 319
166, 167, 168, 169, 170, 171, 179, 181, 191, 201, covering, 195, 228
202, 208, 210, 211, 219, 222, 394, 409, 425, 430 CPA, 216, 394
conductivity, 5, 128, 138, 140, 144, 160, 164, 168, CPU, 18
169, 171, 177, 181, 186, 191, 198, 315, 323, 327, crack, 323
330, 331, 333, 421, 422, 430, 432, 433 CRC, 158, 438
conductor, 171, 198 critical density, 209, 212
cone photoreceptors, 261 critical value, 15, 209
configuration, 4, 9, 17, 19, 72, 73, 80, 86, 87, 90, 91, crossbreeding, 149
94, 101, 102, 103, 104, 105, 106, 107, 116, 280, cross-linking, 256, 257, 274
282, 284, 291, 298, 302, 317, 319, 343, 381, 382 cross-sectional, 60
confinement, xii, 284, 442, 451, 455, 456 cryogenic, 164, 342
confocal laser scanning microscope, 273 crystal growth, 139, 327
conformity, x, 207, 219 crystal structure, 31, 36, 322, 327
connective tissue, 248, 255, 258, 259, 260, 402 crystalline, 190, 286, 322, 323, 325
Connective tissue, 272 crystallization, 36, 314
conservation, 150, 172, 173, 202, 320, 422 crystals, viii, 2, 3, 14, 29, 72, 74, 84, 92, 97, 100,
constraints, viii, 125, 198, 322, 373 101, 102, 105, 107, 109, 114, 139, 140, 187, 192,
construction, 113, 148, 410, 460 232, 280, 281, 282, 290, 292, 293, 294, 295, 296,
consumption, 28, 134, 196, 202, 460 297, 298, 299, 300, 302, 303, 307, 309, 310, 311,
contamination, 412 312, 316, 318, 319, 320, 322, 323, 324, 325, 326,
continuity, 172, 173 327, 329, 331, 332, 333, 334, 402
contracts, 88 CVD, viii, ix, 125, 127, 128, 129, 130, 140, 155
control, 17, 29, 50, 53, 126, 143, 144, 156, 162, 182, cycles, 116, 394
183, 190, 228, 282, 309, 396, 410, 444 cyclodextrins, 309
convection, 3, 180 cytometry, 269, 275
convective, 29 cytoplasm, 241, 244, 246, 248, 249, 250, 254, 255,
convergence, 148 257, 258, 261, 263
conversion, vii, viii, xii, 1, 2, 13, 14, 16, 18, 19, 28,
60, 281, 282, 307, 308, 323, 336, 337, 338, 339,
340, 341, 342, 343, 346, 347, 348, 349, 350, 353, D
358, 361, 363, 364, 370, 387, 393, 398, 458
damping, 72, 79, 80, 81, 82, 83, 84, 87, 88, 101
conversion rate, 340, 341, 347, 348, 349
data set, 117
convex, 214
decay, 216, 294, 340, 342, 343, 345, 346, 350, 351,
cooling, 3, 4, 45, 49, 58, 101, 343, 433
352, 373, 375, 380, 386, 397, 409, 442, 444, 445,
copper, viii, 27, 29, 53, 54, 55, 56, 57, 58, 59, 60, 61,
446, 447, 448, 449, 452, 458
64, 65, 140, 222, 223, 224, 227, 272
decay times, 351, 352, 452
Index 467
decomposition, 149, 398 diffusion, 3, 44, 45, 47, 48, 49, 58, 64, 65, 164, 167,
defects, 6, 36, 58, 60, 62, 64, 271, 282, 286, 305, 171, 186, 197, 208, 210, 211, 219, 222, 281, 308,
306, 309, 323 352, 398, 409, 421, 433
deficiency, 245, 246, 247, 255, 266 diffusion process, 352
definition, 89, 96, 265 diffusivity, 45, 161, 164, 165, 198, 200
deformability, 259 dihedral angles, 373, 376, 377, 379, 380, 383, 384,
deformation, 73, 79, 80, 83, 85, 88, 89, 94, 99, 100, 385, 386
108, 163, 194, 195, 198, 200, 202, 208 dimer, 308, 346, 347, 351, 352
degenerate, xii, 17, 347, 441, 444 dimethylformamide, 343
degradation, 51, 101, 342 diode laser, 19, 99, 102, 122, 280, 290, 299, 300, 332
dehydration, 241, 251, 274 diodes, 99, 100, 101, 102, 122, 316, 456, 457, 461
delivery, 222 dipole, 347, 348
dendritic cell, 276 dipole moment, 343, 350, 380
density, xi, xii, 3, 11, 30, 33, 34, 36, 42, 46, 55, 56, dipole moments, 343
57, 58, 59, 61, 78, 81, 84, 140, 160, 161, 168, direct observation, 261
172, 174, 182, 186, 187, 197, 202, 208, 211, 212, diseases, 247, 263
213, 223, 224, 225, 226, 227, 228, 233, 246, 258, disorder, 322, 327, 329, 331
260, 264, 274, 279, 280, 281, 297, 305, 306, 372, dispersion, 2, 13, 16, 18, 238, 395, 396
373, 381, 400, 405, 420, 421, 422, 423, 427, 428, displacement, 2, 75, 78, 79, 80, 81, 82, 83, 84, 95,
429, 430, 444, 447, 455, 456 108, 110, 113, 115, 118, 119, 120, 121, 145, 160,
density functional theory, 372 197, 200
dentistry, xii, 393, 409, 410 dissociation, 51, 65, 397, 398
dependent populations, 383 distilled water, 143, 408
dephasing time, 449 distortions, ix, 207, 209, 250
deposition, 29, 135, 210, 397, 409, 410, 415, 416, distribution, vii, viii, 1, 2, 4, 6, 8, 9, 11, 20, 27, 29,
417, 426 30, 38, 39, 43, 44, 45, 47, 48, 49, 50, 52, 53, 54,
deposits, 274 65, 126, 127, 132, 133, 134, 144, 150, 151, 152,
depressed, 202 153, 154, 155, 160, 165, 172, 176, 180, 191, 219,
derivatives, 197, 336, 338, 340, 343, 352, 355, 356, 232, 247, 248, 249, 253, 261, 263, 274, 347, 370,
358, 360, 361, 364 400, 401, 423, 433, 456
desorption, 44, 399, 406 distribution function, 2, 52, 160, 172
destruction, 65 divergence, 6, 18, 19, 91, 93, 94, 102, 132, 153
detection, viii, x, xii, 2, 19, 231, 234, 245, 251, 252, DNA, viii, 2, 234, 273, 274, 402
255, 260, 261, 269, 270, 280, 393, 399, 402, 404, dogs, 271
409 donor, 251, 352, 353, 355, 359, 360, 361, 363
detergents, 272 dopaminergic, 261
deviation, 89, 91, 133, 152, 153, 186, 196, 214, 341 dopant, 323
DFT, 372 doped, xi, 2, 11, 28, 54, 280, 281, 282, 284, 286,
diabetes, 257 287, 288, 289, 290, 291, 292, 300, 301, 302, 303,
diabetes mellitus, 257 304, 305, 306, 307, 308, 309, 310, 311, 312, 313,
diamond, viii, ix, 125, 127, 128, 129, 130, 131, 132, 314, 315, 316, 318, 319, 322, 324, 325, 326, 327,
134, 136, 138, 140, 155, 157, 408 329, 330, 331, 332, 333, 334, 342, 455, 456
diamond films, 127, 128, 138 doping, 29
diamonds, 138 drainage, 411
dielectric materials, vii, viii, 27, 29 dumping, 105, 165
dielectrics, x, 27, 29, 30, 36, 37, 38, 39, 42, 58, 207, duration, ix, 28, 31, 42, 45, 60, 61, 101, 105, 106,
209, 210, 211, 212, 228, 312, 409 129, 141, 162, 168, 176, 182, 186, 188, 190, 194,
differential equations, 18, 196 207, 208, 211, 219, 222, 233, 234, 281, 297, 320,
differentiation, 264, 268, 373 325, 327, 331, 339, 372, 373, 381, 394, 395, 396,
diffraction, x, 2, 17, 18, 127, 132, 143, 144, 149, 398, 400, 401, 403, 404, 405, 406, 408, 425
151, 153, 155, 208, 231, 234, 241, 263 dyes, xi, 244, 269, 275, 335, 395
dysplasia, 269
468 Index
300, 301, 302, 310, 312, 313, 315, 316, 317, 318,
E 320, 322, 323, 325, 326, 327, 329, 330, 331, 333,
335, 336, 337, 338, 339, 340, 341, 342, 343, 351,
earth, xi, 279, 308, 310, 313, 316, 317, 319, 320,
370, 372, 378, 395, 399, 407
322, 327, 331, 334, 341
emitters, 28, 280
ECM, 268, 269
endothelial cell, 240, 241, 250, 251, 253, 257, 258,
ecological, 127
260
edema, 247, 249
endothelial cells, 240, 241, 250, 251, 253, 257, 258,
education, 437
260
elastic deformation, 108
endothelium, 232, 239, 240, 241, 251
elasticity, 109, 110, 117
energy consumption, 28, 202, 460
elastin, 233, 259, 272
energy density, 31, 32, 33, 34, 35, 42, 46, 49, 62, 66,
electric charge, 306
129, 161, 173, 297
electric current, 72
energy efficiency, 126, 134, 152, 153, 154
electric field, 14, 81, 111, 113, 115, 121, 171, 172,
energy transfer, vii, xi, 1, 45, 60, 163, 211, 219, 222,
173, 282, 320, 399, 417
270, 281, 286, 287, 290, 294, 297, 300, 301, 302,
electric potential, 78
306, 308, 310, 311, 312, 314, 319, 332, 335, 338,
electrical resistance, 87, 88
349, 350, 364, 377, 409
electricity, 109, 113, 117
enlargement, 52, 64
electrodes, 75, 79, 80, 81, 87, 88, 94, 95, 100, 101,
enthalpy, 205
105, 110, 115, 190
entropy, 110
electromagnetic, 29, 51, 113, 208, 222, 444
environment, vii, ix, x, 51, 138, 144, 207, 208, 209,
electromagnetic wave, 113
231, 234, 263, 296, 317, 322, 394
electromagnetic waves, 51
epitaxial growth, 441
electromagnetism, 417
epitaxy, 443
electron beam, 408, 457
epithelial cell, 240, 241, 243, 253, 273
electron charge, 213
epithelial cells, 240, 241, 243, 253, 273
electron density, 213, 429
epithelium, 232, 239, 240, 241, 247, 251, 257, 258,
electron diffraction, 370
269
electron gas, 162, 163, 169, 174, 195, 197
epoxy, 37, 38
electron microscopy, 39, 244, 267, 274, 275, 408
epoxy resins, 38
electronic structure, xi, 306, 316, 319, 369, 382, 384,
equality, 131
386
equilibrium, vii, ix, 2, 29, 159, 162, 163, 168, 171,

electron-phonon, 42, 160, 162, 168, 173, 174, 177,
173, 202, 203, 205, 207, 208, 210, 211, 225, 371,
187, 192, 196, 198, 200, 210, 319, 320, 321, 331,
376, 379, 429
332
erbium, 309
electron-phonon coupling, 160, 162, 168, 174, 177,
ESR, 304
187, 192, 198, 200, 210, 319, 320, 321, 331, 332
estimating, 64
electrons, ix, 28, 42, 53, 55, 57, 59, 60, 65, 159, 160,
etching, viii, 14, 125, 128, 129, 131, 132, 134, 135,
162, 163, 164, 168, 169, 171, 172, 173, 174, 175,
136, 137, 140, 143, 144, 156
176, 177, 180, 181, 186, 187, 191, 195, 196, 197,
ethanol, 342, 343, 354, 355
198, 199, 200, 201, 202, 208, 210, 211, 212, 219,
ethyl acetate, 143
222, 284, 287, 292, 302, 304, 305, 316, 317, 319,
Europe, 122
320, 371, 372, 379, 383, 399, 400, 416, 428
europium, 288, 289, 310, 313, 319
electron-transfer, 363
evaporation, viii, 32, 34, 35, 44, 46, 125, 135, 210,
elliptic polarization, 77
211, 221, 420, 428
e-mail, 27
evolution, 15, 76, 96, 215, 221, 343, 371, 425, 432,
embossing, 144
450, 451
embryos, 234, 275
examinations, x, 59, 231, 236, 238, 260, 267, 269,
EMG, 129
409
emission, viii, x, 2, 3, 51, 100, 103, 232, 235, 237,
exciton, vii, xii, 343, 345, 346, 347, 348, 349, 441,
238, 241, 242, 244, 247, 256, 257, 258, 260, 263,
442, 443, 444, 445, 446, 447, 448, 449, 450, 451,
269, 270, 277, 280, 281, 282, 283, 284, 285, 287,
452
288, 289, 290, 291, 293, 294, 295, 296, 297, 298,
Index 469
experimental condition, 210 filters, 100, 102, 144, 217

exploitation, 210 financial support, 307
exposure, viii, 15, 27, 64, 408 FISH, 273
extinction, 50, 51, 52, 57, 417, 418 fitness, 149
extracellular matrix, x, 232, 240, 241, 254, 260, 263, fixation, 239, 259, 269
269, 275, 277 flatness, 4
extrusion, 143 flight, 399, 400, 403, 427
eyeball, 235 float, 30, 44
eyes, 15, 237, 253, 274 flow, ix, 4, 111, 159, 202, 208, 210, 323
flow rate, 4
fluctuations, 15, 16, 18, 60, 131, 247, 249, 379, 380
F fluid, 208, 219, 241
fluorescence, x, 15, 232, 233, 235, 242, 243, 247,
fabricate, 28, 29, 53, 128, 129, 219, 441
259, 261, 263, 268, 269, 270, 271, 272, 273, 275,
fabrication, viii, ix, 14, 27, 29, 127, 128, 139, 143,
277, 280, 290, 300, 307, 309, 312, 314, 321, 336,
144, 147, 148, 156, 159, 162, 208, 209, 219, 224,
337, 338, 339, 340, 341, 342, 343, 344, 345, 346,
225, 227, 228, 305, 309
347, 348, 349, 350, 352, 355, 356, 358, 359, 361,
fabrication errors, 127, 144
364, 370, 372, 380, 395
failure, 101, 163
fluorescence decay, 339, 343, 346, 347, 349, 356,
family, 325, 327, 402
359, 372
fats, 402
fluoride, 306, 311, 313, 319
fatty acid, 263
fluorophores, 232, 233, 239, 241, 243, 244, 270, 277
fatty acids, 263
focusing, vii, viii, ix, 1, 6, 8, 17, 20, 125, 126, 127,
fax, 231
207, 208, 212, 218, 223, 233, 264, 339, 396, 399
feedback, 17, 19, 98, 99
Fourier, 18, 167, 213, 215, 216
feet, 261, 262
fovea, 263
Fermi, 161, 162, 168, 169, 176, 186, 187
fracture, 2
Fermi energy, 168
fragmentation, 33, 60, 64, 65, 399, 402, 403, 404,
Fermi level, 162
405, 406
Fermi-Dirac, 176
France, 29, 156, 315
fetal, 276
free energy, 54, 354, 356, 360, 361, 363
Feynman, 230
freedom, 95
FFT, 17, 18
friction, 79, 81, 101

fiber, ix, xi, 4, 9, 85, 105, 122, 126, 143, 144, 145,
fuel, 190
146, 147, 156, 216, 262, 279, 332, 334, 411
fuel cell, 190
fibers, ix, x, 28, 126, 142, 146, 231, 232, 253, 254,
fulfillment, viii, 125
256, 257, 258, 259, 260, 261, 262, 274, 286, 402
Full Width at Half Maximum (FWHM), 83, 84, 105,
fibrils, 245, 247, 248, 251, 256, 257, 259, 260, 273,
160, 181, 191, 264, 320, 326, 373, 381, 456, 458
276
fusion, 160, 179
fibroblast, 268
fibroblast growth factor, 268
fibroblasts, x, 231, 232, 244, 246, 254, 255, 258, G
268, 274, 275
fibrocytes, 254, 258, 260 GaAs, vii, xii, 1, 11, 18, 105, 332, 441, 442, 443,
fibronectin, 272 448, 456, 461
fibrosis, 248, 249, 270, 272 gadolinium, 319, 327, 333
filament, 209, 219 gallium, 333
film, ix, 13, 46, 126, 128, 134, 135, 140, 159, 163, ganglion, x, 231, 261, 262
170, 171, 177, 178, 179, 180, 181, 182, 185, 186, gas, ix, 28, 29, 35, 36, 128, 134, 140, 160, 169, 182,
187, 190, 194, 198, 199, 200, 201, 202, 415, 443, 187, 197, 199, 207, 208, 213, 381, 382, 385, 386,
448 422, 428
film thickness, 181, 187, 199, 443, 448 gas phase, 381, 382
films, ix, xii, 38, 127, 128, 129, 135, 138, 156, 159, gases, 134, 208, 209, 217, 218
163, 203, 208, 416, 441
470 Index
Gaussian, 4, 6, 8, 11, 15, 17, 18, 38, 126, 132, 150,

153, 181, 191, 214, 217, 218, 373, 381, 394, 396,
H
401, 424, 431, 432
H1, 100
GCS, 353, 354, 355, 356, 357, 361, 363
H2, 35, 306
gel, 29, 305, 306
Hamiltonian, 15, 316, 350, 372
generation, viii, xi, xii, xiii, 1, 2, 4, 5, 6, 17, 18, 29,
handling, 28, 74
126, 128, 209, 212, 232, 267, 270, 271, 272, 273,
hardness, 140
275, 276, 277, 282, 286, 300, 301, 306, 311, 315,
harmonics, 99, 104, 105, 107
316, 320, 322, 323, 325, 326, 327, 329, 330, 332,
harvesting, xi, 335, 349, 350, 351, 352
333, 359, 395, 397, 398, 399, 455, 456, 458
healing, 268, 273
geometrical optics, viii, 125, 156
heart, 260, 269, 272, 275
Germany, 27, 136, 231, 234, 235, 237, 238, 270
heart valves, 260, 275
Gibbs, 54
heat, vii, ix, xii, 2, 3, 4, 18, 28, 42, 45, 53, 58, 60, 64,
Gibbs free energy, 54
101, 140, 144, 159, 160, 161, 162, 163, 164, 165,
glass, xi, xiii, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,
166, 167, 168, 169, 170, 171, 174, 175, 176, 177,
39, 40, 41, 42, 43, 44, 45, 47, 48, 49, 51, 52, 53,
179, 180, 181, 182, 183, 186, 187, 190, 191, 194,
58, 61, 62, 64, 65, 71, 72, 105, 109, 135, 138,
195, 196, 197, 198, 200, 201, 202, 205, 208, 210,
143, 208, 235, 238, 279, 282, 284, 286, 287, 288,
211, 222, 225, 264, 316, 319, 322, 323, 324, 393,
289, 291, 292, 293, 294, 300, 301, 302, 304, 305,
394, 397, 398, 400, 409, 410, 420, 421, 423, 425,
306, 307, 309, 310, 311, 312, 313, 314, 322, 326,
428, 430, 431, 433, 455, 457, 458
339, 411, 455, 456, 460
heat capacity, 160, 163, 168, 174, 180, 181, 187,
glasses, viii, 27, 29, 31, 32, 33, 35, 36, 38, 40, 42,
190, 191, 198, 202
46, 49, 51, 53, 58, 280, 282, 284, 286, 287, 288,
heat conductivity, 140, 144
289, 290, 292, 293, 294, 300, 302, 304, 305, 306,
heat loss, 182, 194, 222
307, 309, 310, 313, 314
heat removal, 316
glaucoma, 411
heat transfer, vii, ix, xii, 159, 161, 163, 164, 168,
glial, x, 231, 261, 262
180, 201, 205, 393, 394
glial cells, x, 231, 261, 262
heating, ix, 3, 42, 44, 49, 51, 60, 64, 65, 135, 159,
glucose, 275
162, 163, 168, 171, 172, 175, 180, 187, 194, 195,
glycoprotein, 271
196, 197, 198, 200, 202, 203, 210, 211, 289, 305,
gold, ix, 159, 170, 171, 174, 175, 176, 177, 178, 179,
398, 411, 425, 428, 429, 433, 434
180, 182, 185, 186, 190, 194, 200, 201, 202, 205,
hegemony, 315
246, 253, 309, 313

height, 86, 104, 105, 129, 136, 139, 141, 143, 238,
gold nanoparticles, 309
422, 451
grafting, 251
helium, 342
grain, 323
helix, 247, 344, 402
grains, 191
hematoxylin-eosin, 251
granules, 241, 242, 243, 254, 255, 261, 263, 268
heterostructures, xii, 455, 456
graph, 85, 101
hexane, 373
graphite, 132
high pressure, 195
gratings, 395, 444
high resolution, 223, 247, 248, 255, 261
gravitational waves, 19
high tech, 260
gravity, viii, 2
high temperature, 49, 167, 186, 306
Great Britain, 156, 204
high-frequency, 134, 140
Green’s function, 213
high-speed, 28
grids, 182
histochemistry, 272
groups, 63, 180, 195, 305, 325, 336, 342, 358, 363,
histological, 233, 239, 240, 244, 247, 250, 259, 260,
377, 380, 405
263, 267, 269
growth, viii, 27, 30, 37, 39, 58, 61, 64, 139, 140,
histology, 239, 244, 246, 250, 255, 267, 269, 276
268, 272, 274, 327, 332, 441, 456
Hm, 179
growth arrest, 271
homeostasis, 244, 245, 246, 248
growth factor, 268, 272, 274
HOMO, 359, 372, 373, 375, 378, 379, 381, 382, 383,
guidelines, 238, 442
384, 386
Index 471
homogenous, 250, 251, 262, 263 indices, 73, 76, 92, 93, 96, 101, 108, 113, 117, 118,
host, 36, 255, 283, 284, 286, 293, 297, 319, 320 120
human, 15, 116, 237, 250, 263, 264, 271, 273, 274, indirect effect, 323
275, 277, 410 industrial, 29, 127, 208
humans, 251 inelastic, 449
hydration, 241, 251 inequality, 320
hydro, 340 inert, 36
hydrocarbon, 404 inertness, 138, 139
hydrodynamic, 427 infinite, 4, 79, 80, 110, 164, 165, 201
hydrofluoric acid, 143 inflammation, 412
hydrogen, 132, 355, 373, 387, 421 inflammatory, 268
hydrogen atoms, 373 information and communication technology, 442
hydroxyapatite, 404 infrared, viii, x, 13, 34, 71, 107, 125, 141, 156, 231,
hydroxyproline, 402 232, 233, 260, 273, 274, 276, 277, 280, 281, 306,
hyperbolic, ix, 159, 163, 165, 166, 169, 170, 171 308, 309, 310, 311, 312, 313, 314, 315, 333, 370,
hypersensitive, 282 373, 375, 428
infrared light, 13, 71, 233
initial state, 350
I initiation, 179
injection, 15, 17, 18, 398
ICE, 134, 135
injuries, 410
id, 41, 45
inorganic, 281, 288, 332, 402
illumination, 15, 41, 45, 269, 282, 309
insertion, 320, 322, 325, 329, 412
images, x, 40, 56, 223, 231, 233, 235, 237, 239, 243,
insight, 115, 340
244, 245, 248, 249, 253, 254, 256, 257, 260, 261,
inspection, 265
263, 265, 269, 271, 276
instability, 9, 15, 16, 17, 129, 131, 361
imaging, vii, x, 231, 232, 233, 234, 235, 236, 237,
instruments, 239
238, 239, 241, 242, 245, 246, 247, 248, 249, 250,
insulators, 47, 65
251, 252, 253, 254, 255, 256, 257, 258, 260, 261,
integration, 18, 116, 154, 215, 216, 444
262, 263, 264, 265, 266, 267, 268, 269, 270, 271,
integrity, 241, 267
272, 273, 274, 275, 276, 277, 280, 301
interaction, vii, viii, ix, xii, 27, 30, 40, 42, 45, 53, 58,
immersion, 142
61, 87, 132, 144, 159, 162, 170, 171, 180, 181,
immunohistochemistry, 271
183, 186, 191, 202, 213, 228, 282, 317, 319, 332,
implants, 29, 58, 409, 412
353, 355, 358, 372, 393, 394, 397, 398, 400, 403,
implementation, xii, 139, 232, 393, 411, 412, 417
404, 407, 412, 417, 420, 427, 428
impregnation, 246
interaction process, 394
impurities, 36, 139, 211
interactions, vii, xii, 30, 40, 171, 195, 234, 270, 276,
in situ, x, 231, 233, 238, 239, 241, 244, 245, 246,
347, 349, 353, 355, 370, 393, 394, 397, 398, 416,
248, 249, 250, 251, 253, 255, 257, 260, 264, 269,
420
277
interface, 160, 163, 180, 181, 182, 183, 286
in transition, 317
interference, 442, 448, 451, 453
in vitro, 246, 250, 259
intermolecular, 342, 350, 352, 355, 357, 370
in vivo, x, 231, 233, 238, 239, 244, 245, 247, 248,
intermolecular interactions, 370
250, 251, 253, 255, 267, 268, 269, 271, 272, 274,
interval, 64, 139, 215, 216, 248, 249, 256, 266, 394
275, 277
intervention, 410, 412
inactive, 267
intraocular, 260, 274
incidence, 57, 90, 191
intravenous, 236, 238
inclusion, 171, 180
intrinsic, x, 164, 176, 195, 196, 232, 233, 239, 241,
income, 132
242, 243, 244, 247, 249, 250, 260, 261, 263, 266,
incompatibility, 442
268, 277, 298, 306, 320, 331
incubation, 271
invasive, 409
independent variable, 75
inversion, vii, 1, 2, 3, 11, 103, 104, 329
indication, 145
inversions, 17
472 Index
investigations, 260
ion beam, 29, 36, 54, 56, 57, 58, 61, 400
L
ion bombardment, 147, 156, 400, 406
labeling, 301
ion implantation, viii, 27, 29, 30, 36, 37, 38, 39, 42,
lambda, 235
46, 49, 50, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61
lamella, 247
ion-exchange, 32, 33, 35
lamellar, 251, 257
ionic, 29, 319, 327, 400
lamina, 251
ionization, ix, 36, 196, 207, 208, 209, 211, 212, 213,
lanthanide, 282, 283, 284, 288, 304, 307, 308, 311,
217, 218, 223, 304, 308, 394, 399, 400, 403, 406,
316, 319, 320, 332
407, 420, 421, 423, 428, 429
laser ablation, viii, ix, xii, 125, 126, 128, 129, 130,
ionization energy, 429
134, 137, 155, 210, 223, 276, 309, 393, 397, 400,
ionization potentials, 208, 217
402, 403, 404, 405, 407, 408, 409, 410, 412, 415,
ions, xi, 28, 29, 30, 36, 37, 38, 40, 46, 49, 53, 56, 58,
416, 417, 420, 425, 426, 427, 428, 437
59, 61, 62, 65, 279, 280, 281, 282, 283, 284, 285,
laser radiation, xii, 3, 15, 58, 59, 60, 61, 64, 65, 126,
287, 288, 289, 291, 292, 293, 294, 296, 297, 301,
127, 133, 191, 393, 456
302, 304, 305, 306, 307, 308, 309, 311, 312, 313,
laser range finders, 105
314, 315, 316, 317, 319, 320, 322, 323, 325, 326,
lasers, vii, viii, ix, x, xi, xiii, 1, 3, 9, 11, 32, 65, 105,
327, 329, 331, 332, 399, 400, 404, 405, 406, 407,
125, 127, 128, 138, 161, 162, 163, 190, 194, 195,
416, 455
202, 205, 207, 208, 209, 212, 221, 231, 232, 233,
IOP, 453
264, 266, 267, 276, 277, 281, 286, 313, 314, 316,
iron, 30, 37, 416, 420, 424
319, 331, 332, 333, 369, 395, 397, 398, 399, 400,
irradiation, viii, xi, 27, 29, 30, 33, 36, 37, 38, 44, 45,
409, 410, 411, 412, 455, 461
49, 50, 52, 53, 54, 59, 62, 63, 64, 65, 125, 131,
lattice, ix, 2, 3, 45, 159, 160, 161, 162, 168, 169,
170, 185, 198, 279, 282, 284, 286, 287, 289, 290,
170, 171, 172, 175, 178, 179, 181, 182, 183, 184,
292, 294, 297, 299, 300, 301, 302, 303, 304, 305,
185, 187, 188, 190, 191, 192, 194, 195, 196, 197,
306, 309, 310, 311, 312, 338, 398, 408
198, 201, 202, 207, 208, 210, 211, 212, 309, 319
irradiations, 403
lattices, vii, ix, 159, 162, 163, 168, 169, 170, 180,
ISC, 370
195, 196, 197, 198, 202
isolation, 273
law, ix, 91, 93, 150, 159, 161, 164, 165, 168, 169,
isomerization, 373
170, 176, 197, 201, 280, 340, 341, 346, 347, 349,
isothermal, 423, 430
353, 354, 355, 397, 416, 422, 433
isotropic, 18, 75, 105, 109, 117, 173, 174, 397
laws, ix, 159, 161
iteration, 148, 149, 150

lens, 3, 4, 6, 9, 11, 15, 24, 128, 130, 131, 132, 133,
iterative solution, 202, 371
141, 145, 146, 150, 151, 152, 153, 154, 155, 213,
214, 217, 224, 225, 227, 228, 261, 329, 334, 339,
J 394, 396, 401, 408
lenses, 128, 132, 133, 144
Japan, 335, 346, 347, 441, 453 lesions, 266, 268, 398
Joule heating, 210, 211, 305 lie group, 93
Jung, 1 life sciences, 273
lifetime, xi, xii, 135, 271, 296, 297, 308, 319, 335,
336, 337, 338, 339, 342, 343, 345, 347, 349, 352,
K 355, 356, 357, 358, 361, 363, 364, 373, 380, 449,
452, 455, 458
keratoplasty, 251, 410 ligand, 234, 272, 286, 352, 355, 356, 357, 358, 361,
ketamine, 236, 238 363
kinetic energy, 173, 187, 381, 400, 428 ligands, 288, 358
kinetics, 180, 181, 202, 372, 420 light beam, 14, 71, 106, 154
Kirchhoff, 78, 132 light scattering, 132, 305, 313
KOH, 140 light-emitting diodes, 461
Korea, 335 limbus, 240, 241, 253, 259
Korean, 365 limitation, 2, 62, 94, 260, 372
krypton, 32 limitations, 57, 139, 148
Index 473
linear, vii, viii, x, 2, 4, 9, 13, 16, 29, 33, 42, 53, 73, manipulation, xi, 279, 313, 314
77, 78, 85, 95, 97, 101, 107, 108, 113, 115, 128, manufacturing, 128, 131, 134, 156, 161, 190, 208
129, 139, 166, 167, 174, 196, 231, 232, 233, 234, mapping, 234, 272, 330
236, 246, 248, 250, 253, 254, 256, 257, 261, 262, mask, 37, 129, 135, 140, 141
263, 265, 267, 269, 270, 283, 284, 301, 303, 332, masking, 135
346, 348, 419, 432, 442, 450, 452 mass spectrometry, 412
linear dependence, 73 material sciences, xi, 369
linear function, 4 material surface, 223
linkage, 350, 353, 402 matrices, 146
lipofuscin, 233, 243, 244, 263, 274 matrix, viii, x, 5, 27, 29, 30, 31, 32, 33, 34, 35, 36,
liquid nitrogen, 140, 306 37, 51, 53, 54, 55, 57, 58, 59, 60, 61, 64, 65, 76,
liquid phase, ix, 159, 163, 179, 180, 181, 182, 185, 86, 93, 96, 109, 111, 114, 117, 118, 119, 120,
187, 192, 194, 202, 409 121, 143, 144, 145, 146, 147, 232, 240, 241, 254,
lithium, 74, 122, 123, 292, 311 260, 263, 269, 275, 277, 286, 292, 304, 319, 320,
lithography, 29, 134, 135 325, 371, 372, 399
localised, 43, 46, 399 MBE, 456, 461
localization, x, 232, 245, 248, 255, 260, 263, 269 MCP, 401
London, 66, 69, 122, 157, 204, 207, 274, 387 MDR, 457
long distance, 5 mean-square deviation, 152
long work, 223 measure of value, 149
losses, ix, 17, 52, 125, 128, 132, 146, 147, 150, 221, measurement, 84, 244, 245, 248, 250, 255, 339, 342,
222, 289, 307, 456 348, 350, 373, 395, 442, 444, 445, 447, 457
low power, 208 measures, 107, 398
low temperatures, 175 mechanical properties, 315, 325, 327, 329, 331, 333
low-density, 186, 264 mechanical stress, 457
low-power, 301 media, viii, ix, 2, 30, 46, 50, 207, 209, 239, 270, 397
LSM, 235, 236, 237 medical diagnostics, 288
luminescence, xi, 266, 279, 281, 282, 286, 288, 289, medicine, 272, 409, 412
292, 294, 299, 302, 305, 306, 307, 308, 310, 311, melanin, 263, 276
312, 333, 341 melanoma, 248
LUMO, 359, 372, 373, 375, 376, 378, 379, 381, 382, melt, ix, 45, 49, 60, 139, 140, 195, 207, 208, 210,
383, 384, 386 211, 212, 219, 220, 222, 225, 422, 427
lutetium, 327 melting, viii, ix, 27, 42, 44, 45, 51, 58, 60, 61, 64,
lying, 238, 343 65, 140, 159, 160, 162, 163, 179, 180, 181, 182,
lymph, 276 183, 185, 186, 187, 190, 191, 194, 202, 210, 212,
lymph node, 276 221, 284, 305, 306, 408, 416, 417, 421, 422, 423,
lysine, 343, 344 426, 427
lysosomes, 263 melting temperature, 45, 140, 202, 421, 422, 427
melts, 58, 210, 222
membranes, 263
M memory, 219, 222, 227, 305, 306, 313
metabolic, 244, 245, 246, 248
M.O., 461
metal ions, 28, 29, 30, 36, 38, 46, 280, 301, 302
macrophage, 271
metal nanoparticles, vii, viii, 27, 38, 42, 47, 51, 309
macular degeneration, 263
metal oxide, 65
magnetic, 37, 134, 140, 172, 210
metalloporphyrins, 339
magnetic field, 134, 140, 172
metals, vii, ix, x, xii, 29, 42, 159, 162, 163, 167, 186,
magnetron, 29, 135
201, 202, 208, 209, 210, 211, 212, 228, 343, 393,
magnetron sputtering, 29
409, 420, 427
MAI, 277
methanol, 304, 312
maintenance, 244, 246, 259
methylcyclohexane, 354
mammals, 273, 275
mica, 57
manifold, 316, 317, 327, 331, 347, 455
microcavity, 18
manifolds, 316, 317, 331
474 Index
microelectronics, 30 monochromatic light, 73, 288, 289, 298, 299, 398

microfabrication, xi, 279, 281, 304, 309 monochromator, 339
micro-lens, 128 monolayer, 240, 241, 250, 257, 258, 261, 263
micrometer, 263, 284, 313, 314 monolayers, 313
microparticles, ix, 159 monomer, 346, 347, 348, 349, 350, 351, 352
microscope, 37, 136, 141, 142, 144, 145, 146, 223, monomeric, 347, 348
233, 234, 235, 236, 238, 241, 244, 246, 247, 250, monomers, 347
258, 259, 260, 263, 273, 340, 408 morphological, xii, 244, 251, 256, 257, 261, 266,
microscopy, x, 36, 39, 136, 231, 232, 233, 234, 235, 393, 408
236, 239, 241, 244, 248, 250, 251, 253, 258, 260, morphology, 57, 241, 244, 249, 254, 255, 260, 266,
261, 263, 265, 267, 268, 269, 270, 271, 272, 273, 268, 269, 282, 408
274, 275, 276, 277, 370 morphometric, 251
microstructure, 129, 141, 240, 268, 277, 305 mosaic, 250, 261
microstructures, vii, x, 231, 234, 244, 255, 260, 261, Moscow, 125, 157
265, 268, 270, 282 motion, xii, 17, 88, 89, 95, 101, 110, 111, 162, 176,
microsurgery, 264, 267, 409 177, 186, 187, 197, 198, 202, 371, 375, 379, 380,
microtubule, 272 384, 385, 386, 441, 442, 443, 451
microtubules, 234 mouse, 274, 275
microvasculature, 269 movement, 43, 47, 51, 219
microwave, 129 MPA, 281, 300
migration, xi, 49, 60, 335, 343, 350, 351, 352 MPI, 400
mimicking, 337, 350 multidimensional, 370
minerals, 29 multiplexing, 73, 99, 100, 101, 102, 111, 113, 122
miniaturization, 411 multiplication, 96, 215, 216
Ministry of Education, 453 multiplicity, xi, 335
mirror, 13, 18, 395 muscle, 268
Missouri, 159 mutation, 149
MIT, 19 mutations, 149
mitochondria, 233, 239, 241, 242, 243, 245, 249, myofibroblasts, x, 231, 232, 266, 268, 273
255, 261, 263, 270, 272
mitochondrial, xi, 232, 261, 263, 270, 275
mixing, xii, 339, 441, 444 N
mobility, 49, 52, 58, 173

NA, 217
modeling, 18, 50, 51, 65, 194, 203, 210
Na+, 404
models, vii, ix, 3, 4, 11, 18, 159, 162, 163, 167, 168,
NAD, xi, 232, 233, 234, 239, 241, 242, 243, 244,
170, 171, 172, 175, 176, 178, 179, 180, 186, 202,
245, 249, 255, 261, 263, 268, 270, 273, 275
222, 425, 427, 436
nanocomposites, 66
modulation, 14, 17, 72, 99, 123, 271, 281, 447
nanocrystal, 307
modulus, 75, 78, 109, 160, 200, 323
nanocrystalline, 286, 310
moieties, 350, 351, 353, 380, 384, 385, 386
nanocrystals, 282, 286, 306, 307, 309, 314
moisture, 238
nanomaterials, 284
molar volume, 422, 428
nanometer, 60, 208
molecular beam, 443
nanometer scale, 60, 208
molecular beam epitaxy, 443
nanometers, 40, 55, 56, 57, 59, 162, 196, 411
molecular mass, 402
nanoparticle synthesis, 29
molecular orbitals, 375
nanoparticles, vii, viii, 27, 29, 30, 31, 32, 33, 34, 35,
molecular structure, 361
37, 38, 39, 40, 42, 43, 44, 45, 47, 49, 50, 51, 52,
molecules, x, xi, xii, 213, 231, 233, 270, 335, 338,
53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65,
340, 342, 355, 364, 369, 370, 371, 372, 386, 393,
191, 284, 285, 306, 308, 309, 310
397, 399, 400, 402, 403, 406, 407
nanostructured materials, 50
molybdenum, 140
nanostructures, 239, 284, 309, 441
momentum, 172, 173, 174, 175, 202, 220, 317
nanostructures, 453
monazite, 308
nanotechnology, 409
Index 475
nanowires, 308 nuclear, 36, 53, 371, 372

naphthalene, 336 nucleation, 37, 38, 39, 45, 49, 58, 61, 163, 180, 182,
narcotics, 236, 238 195, 202
Nash, 311 nuclei, 52, 241, 242, 243, 244, 246, 248, 250, 254,
National Research Council, 207 255, 257, 258, 261, 263, 266, 268, 371, 372, 373,
National Science Foundation, 202, 437 381
natural, xi, 74, 107, 138, 139, 180, 335, 340, 350, nucleus, 65, 317, 371, 372
351, 352 numerical analysis, 177
Nb, 135 numerical aperture, x, 231, 233, 237, 263, 264
Nd, 2, 3, 4, 11, 13, 15, 19, 30, 31, 32, 33, 34, 35, 51, nutrition, 237, 238
63, 65, 85, 280, 290, 300, 307, 312, 318, 332,
333, 395, 396
necrosis, 275 O
neglect, 84
observations, 51, 209, 217, 239, 241, 251, 255, 260,
neodymium, 316, 319, 326, 332, 333
266, 273, 341, 372, 378
neoplastic, 269
occipital lobe, 261
neoplastic tissue, 269
OCT, 261
nerve, 240, 252, 253, 261, 262
octane, 337
nerve fibers, 253, 261, 262
oil, 234, 263
nerves, 261, 275, 409
oligomers, 345, 348
network, 32, 33, 34, 35, 244, 246, 275, 306
on-line, 222, 241, 245, 250, 251, 260, 267
neurons, 261, 271, 272, 275
operator, 371
New Jersey, 238
ophthalmic, 239
New York, 24, 69, 70, 122, 203, 204, 230, 271, 274,
optic nerve, 15, 261
391, 437, 438
optical absorption coefficient, 128
New Zealand, 238, 241, 249
optical activity, 72
Ni, 314
optical communications, xi, 15, 279
niobium, 135
optical density, 143
NIR, x, 71, 72, 102, 107, 231, 232, 233, 234, 237,
optical fiber, ix, 9, 126
264, 267, 269, 280, 282, 289, 290, 292, 299, 300,
optical imaging, x, 232, 234, 236, 238, 239, 245,
301, 302, 304
246, 247, 250, 251, 252, 253, 255, 260, 261, 265,
NiTi shape memory, 227
267, 269, 270, 272
nitride, 461
optical parameters, 117, 416
nitrogen, 140, 306
optical properties, xi, 29, 30, 46, 59, 200, 279, 281,
noble metals, 29
304, 306, 311, 316, 317, 319, 325, 331, 397, 424,
noise, 9, 15, 18, 239, 242, 269
431, 442
non-crystalline, 286
optical pulses, 395, 455
non-destructive, 253
optical storage, 280, 306
nonequilibrium, vii, ix, 65, 159, 162, 168, 176, 195,
optical systems, 142
196, 198, 201
optical transmission, 52, 58, 61, 132
non-invasive, x, 232, 255, 260, 269, 270, 272
optics, vii, viii, 1, 4, 19, 24, 53, 122, 125, 126, 127,
nonlinear optical response, viii, 2, 442, 443, 445, 448
156, 264, 302
nonlinear optics, 19, 53
optimization, viii, ix, 2, 3, 125, 126, 131, 144, 148,
nonlinearities, 286
149, 150, 152, 156, 373
nonlocality, 442
optimization method, ix, 126
non-metals, 42
optoelectronic, 27, 28, 29
non-thermal, 40, 409
optoelectronic devices, 29
non-uniform, vii, 1, 2, 3, 4, 17, 30, 38, 39, 46, 65,
optoelectronics, 27
153, 163, 171, 172, 182
orbit, 237, 317
non-uniformity, 153
ordinary differential equations, 96, 97
normal, 191, 209, 220, 240, 241, 247, 266, 268, 269,
organ, 269, 275
271, 272, 274, 275, 336, 353, 354, 355, 357, 361
organelle, xi, 232, 267, 270
normal conditions, 271
organelles, 234, 263, 264, 277
476 Index
organic, 281, 288, 395, 399, 402, 405 permittivity, 139, 418
organic matter, 402 peroxidation, 263
orientation, 73, 92, 139, 247, 249, 255, 256, 257, 348 perturbation, 4, 213, 281, 317, 396
orthophosphates, 308 perturbations, 214
oscillation, viii, 2, 16, 17, 18, 19, 71, 72, 74, 75, 76, pH, 343
77, 78, 80, 81, 84, 86, 98, 99, 100, 101, 102, 104, phagocytosis, 263, 267, 268
105, 106, 111, 113, 116, 180, 339, 384, 386, 451, phase space, 172
453 phase transformation, 309
oscillations, 55, 75 phenotype, 271, 274
oscillator, 15, 16, 17, 216, 264, 339, 345, 395, 396, Philadelphia, 203
442 phonon, 164, 165, 170, 171, 172, 195, 196, 203, 205,
ossicles, 409 212, 297, 298, 311, 319, 320, 321, 331
Ostwald ripening, 38 phonons, 163, 164, 168, 170, 171, 174, 195, 196,
oxidation, 31, 51, 53, 132, 280, 304 197, 198, 199, 200, 201, 202, 210, 211, 212, 286,
oxide, xi, 28, 65, 135, 139, 140, 308, 313, 315, 319, 319
415, 416, 417, 427, 437 phosphatases, 272
oxides, 319 phosphate, 238, 290, 343, 460
oxygen, 54, 135, 236, 238, 239, 305, 306, 321, 325, phosphor, 297
327, 398 phosphorescence, 305, 313, 336, 337, 370
phosphors, 281, 282, 288, 290, 301, 307, 308, 311
photobleaching, 233, 246, 271
P photochemical, 40, 272, 289, 335, 337, 338, 364,
397, 398
pain, 409
photodynamic therapy, 204, 398
pancreatic, 275
photoexcitation, 40, 348, 351, 382, 386
pancreatic islet, 275
photoionization, 32, 399, 420
parabolic, ix, 159, 164, 166, 169, 170, 171, 181, 191,
photolithography, 135, 143
334
photoluminescence, 305, 309, 443
parameter, 3, 6, 8, 19, 20, 36, 74, 101, 128, 200, 218,
photolysis, 357, 359
219, 319, 321, 323, 324, 396, 399, 400
photon, vii, x, 11, 15, 20, 28, 36, 37, 60, 103, 104,
Paris, 123, 315
163, 209, 211, 228, 231, 232, 233, 234, 241, 242,
Parkinson’s disease, 409
247, 248, 254, 255, 257, 258, 259, 262, 263, 264,
partial differential equations, 196
269, 271, 272, 273, 274, 275, 276, 277, 280, 281,
particle density, 420, 423, 429
282, 284, 286, 287, 288, 289, 291, 292, 293, 294,
particle nucleation, 39, 49
297, 298, 300, 301, 302, 303, 304, 308, 309, 310,
particle shape, 33, 34, 35
311, 312, 339, 342, 370, 372, 373, 375, 381, 397,
particles, viii, 27, 29, 30, 31, 32, 33, 34, 35, 38, 39,
398, 399, 420, 421, 428, 429, 444
44, 45, 46, 47, 49, 51, 52, 53, 54, 55, 57, 58, 59,
photonic, 147, 282, 309, 332, 334, 349
60, 62, 64, 65, 163, 190, 191, 209, 286, 304, 400,
photonic crystal fiber, 332, 334
407, 423
photonic crystals, 282, 309
partition, 198
photons, x, 42, 59, 171, 211, 231, 232, 233, 261,
passivation, 457
280, 281, 284, 288, 302, 323, 397, 399, 400, 420,
passive, vii, 1, 5, 11
428, 429
pathogens, 241
photoreceptor, 261, 263
pathology, 248, 274
photoreceptors, x, 231, 261, 263
pathophysiology, 257, 263
photosynthesis, xi, 335, 340, 350, 364
pathways, 406
photosynthetic, 350
patterning, 129
phthalocyanines, 337
PDEs, 196
physical mechanisms, 195
peptide, 402, 404
physical properties, 129, 316, 333
peptide chain, 402, 404
physics, xi, xii, 369, 393, 401, 438
perfusion, 271
physiological, 247, 260, 265, 267
periodic, 9, 14, 132
piezoelectric, 72
permit, 239
Index 477
pig, 251, 271 praseodymium, 308, 319

pigments, 272 praxis, 87, 102
pinhole, 233 precipitation, 36, 301, 306, 309, 314
PL spectrum, 444, 445 prediction, 40, 98, 99
planar, 361, 373, 385, 386 pressure, 140, 144, 160, 172, 173, 174, 195, 197,
plaque, 409 198, 210, 223, 224, 398, 400, 408, 422, 423, 428,
plasma, ix, 32, 34, 35, 64, 128, 129, 132, 135, 186, 432
196, 207, 208, 209, 210, 212, 218, 219, 222, 223, prevention, 409
224, 225, 227, 228, 264, 280, 394, 397, 399, 400, probability, 58, 149, 208, 211, 420, 428, 429, 444
416, 417, 420, 422, 423, 425, 426, 427, 428, 429, probe, xi, 136, 142, 146, 270, 279, 282, 289, 301,
430, 431, 434, 435, 436, 437 338, 339, 347, 372, 447
plastic, 90, 238 process duration, 141
platelet, 268 producers, 412
platinum, 227 production, ix, 139, 140, 143, 207, 208, 209, 318,
play, 13, 145, 194, 250, 331, 373, 375, 386 319, 322, 323, 324, 410
PLD, 415, 416 productivity, 129
plexus, 252, 253 profit, 105
PM3, 350 program, 18, 19, 95, 97, 235, 456
point defects, 36 programming, 145
Poisson, 109 proliferation, 275, 277
polarity, 272, 353, 373 propagation, ix, 28, 42, 44, 45, 74, 160, 163, 164,
polarization, 14, 33, 71, 72, 77, 86, 90, 92, 94, 95, 165, 167, 170, 195, 201, 202, 207, 208, 209, 213,
96, 97, 99, 100, 101, 102, 107, 111, 114, 115, 216, 323, 417, 447, 453
116, 305, 313, 330, 348, 442, 448 property, 18, 82, 168, 208
polarized light, 73, 76, 77, 90, 92, 94, 95, 101, 107, prophylaxis, 261
115 proportionality, 341
polycarbonate, 190 protective coating, 409
polycrystalline, 128, 140, 143, 156, 323 protein, 240, 247, 259, 270, 273, 290, 311, 402
polycrystalline diamond, 156 protein structure, 240, 247
polyene, 372 proteins, 233, 271, 400, 402, 403
polyimide, 37 Proteins, 402
polymer, 259, 342 prototype, 27, 102
polymer matrix, 342 PSI, 134

polymerization, 32, 34, 35, 309, 343, 345, 349 pulsed laser, viii, 27, 49, 52, 102, 105, 161, 162, 163,
polymers, 29, 36, 208 180, 182, 264, 276, 305, 312, 313, 364, 400, 409,
polymorphism, 272 415, 416, 417, 425, 426, 427, 433
polynomial, 5, 218 pulsed laser deposition, 415, 416, 417, 426
polypeptide, 402, 403, 404, 405 pumping, xiii, 2, 3, 4, 6, 13, 16, 18, 19, 103, 302,
polypeptides, 343, 344, 345, 346, 349, 400, 402 316, 322, 323, 327, 329, 331, 332, 334, 412, 455,
poor, 233 456, 457, 458, 460
population, vii, 1, 11, 17, 148, 149, 280, 283, 284, pumps, 304, 455
292, 294, 329, 342, 375, 376, 378, 379, 442, 444, pyridine ring, 361
447, 449, 452 pyrrole, 352, 363
pores, 190, 282, 309
porosity, 190
porous, 190 Q
porphyrins, 233, 243, 244, 337, 338, 340, 341, 342,
quantization, xii, 130, 131, 148, 149, 150, 152, 153,
343, 344, 346, 347, 348, 351, 352, 361, 364
156, 441, 443
postoperative, 266, 268
quantum, xi, xii, 3, 15, 18, 197, 213, 284, 286, 319,
potential energy, 373
323, 327, 333, 337, 341, 342, 345, 361, 369, 370,
powder, 190, 191, 205, 286, 292
372, 441, 442, 444, 448, 450, 453, 455, 456
powders, 180, 190, 205
quantum chemical calculations, xi, 369, 370
powers, 105, 208, 237, 339, 394, 398
quantum chemistry, 370
478 Index
quantum confinement, 284, 442 red shift, 343, 458

quantum dot, 441 redistribution, 176, 379, 385
quantum dots, 441 reflectance spectra, 39, 42, 43, 46, 48, 52, 55, 63
quantum mechanics, 213 reflection, ix, 12, 13, 18, 49, 50, 52, 55, 57, 59, 62,
quantum theory, 15 63, 64, 65, 90, 91, 100, 107, 111, 116, 125, 126,
quantum well, xii, 441, 455, 456 127, 128, 146, 147, 252, 266, 397, 457
quantum yields, 337, 341, 342, 345 reflectivity, 17, 18, 19, 39, 160, 176, 179, 182, 397,
quartz, 71, 72, 134 419
refraction index, 14
refractive index, ix, 4, 29, 76, 111, 114, 125, 128,
R 136, 139, 140, 146, 147, 148, 156, 208, 250, 360,
394, 417, 418
radar, 410
refractive indices, 73, 76, 92, 93, 96, 101, 113, 117,
radial distance, 222
118, 120
radial glia, 261, 262
regeneration, 241, 247, 267
radiation, viii, xii, 2, 3, 15, 28, 36, 58, 59, 60, 62, 61,
regrowth, 33, 45, 49, 65
64, 65, 126, 127, 132, 133, 138, 142, 143, 144,
regular, 261, 263
153, 156, 191, 208, 209, 213, 233, 234, 255, 257,
regulation, 244, 246
258, 259, 260, 307, 371, 372, 393, 395, 397, 400,
relationship, 9, 249, 419
410, 425, 455, 456
relationships, 175
radius, 4, 76, 144, 149, 150, 152, 153, 181, 191, 213,
relaxation, 40, 45, 59, 60, 160, 161, 164, 165, 166,
214, 217, 218, 219, 223, 224, 225, 227, 228, 327,
173, 187, 190, 196, 198, 200, 202, 210, 294, 295,
359, 360, 442
296, 297, 319, 342, 343, 363, 370, 379, 380, 381,
Raman, 280, 309, 319, 356, 370, 388
382, 387, 398, 400, 432, 434, 435
Raman spectroscopy, 309
relaxation process, 297, 343, 382
random, 149, 222, 322, 326, 381
relaxation processes, 297
range, viii, ix, xii, xiii, 15, 17, 19, 27, 29, 30, 36, 37,
relaxation rate, 363
42, 46, 55, 57, 58, 60, 62, 64, 84, 97, 98, 105,
relaxation time, 40, 160, 161, 164, 165, 166, 173,
107, 125, 126, 127, 128, 129, 131, 135, 138, 140,
187, 190, 196, 198, 200, 210, 343, 398, 400, 434,
141, 144, 145, 147, 150, 156, 161, 176, 180, 186,
435
187, 207, 208, 209, 210, 212, 219, 221, 223, 228,
relaxation times, 40, 165, 190
241, 267, 282, 289, 301, 316, 317, 320, 322, 329,
reliability, 384, 460
342, 343, 353, 354, 384, 394, 395, 397, 398, 404,
remodelling, 269
406, 409, 418, 419, 429, 436, 445, 455, 456, 460
remote sensing, viii, 2, 5, 15
rapid prototyping, 105, 205
repair, x, 231, 253, 268
rare earth, 307, 308, 313, 316, 317, 319, 320, 322,
reproduction, 322
327, 331, 332, 334, 341
research and development, 394
rare earths, 307, 332
residues, 402, 404
RAS, 125, 134
resilience, 259
rat, 275
resin, 37, 38
rats, 272
resistance, 87, 88, 127, 323, 324, 331, 419
reaction center, xi, 335
resistive, 135, 140
reactivity, 338
resolution, x, 128, 132, 198, 223, 231, 233, 234, 239,
reading, 145
242, 244, 245, 246, 247, 248, 249, 251, 255, 260,
real time, 265, 267, 370
261, 262, 263, 269, 270, 271, 272, 274, 275, 339,
reality, 30, 195
395, 450, 451
reasoning, 213
resonator, 3, 4, 5, 6, 9, 18, 111, 113, 395, 460
recession, 422, 427
response time, 442, 449, 450
recombination, 302, 305, 357, 358
restorations, 410
recombination processes, 358
retardation, 51, 73, 85, 87, 90, 91, 92, 94, 96, 97,
reconstruction, 409
100, 101, 103, 104, 105, 106, 107, 114, 115
rectification, 100
retina, x, 15, 231, 232, 233, 238, 261, 262, 263
recursion, 51
retinal disease, 261, 263
red light, 13
Index 479
returns, 384, 386, 398 sclera, 232, 238, 247, 249, 254, 255, 256, 257, 259,
rhombohedral, 325 263, 273, 274
rifting, ix, 159 search, 149
rings, 343, 375, 379, 380, 386 secret, 268
risk, 263 seed, 17, 139, 211, 394, 395, 396
rods, 261 seeding, 15, 17, 18, 19
room temperature, 3, 14, 173, 180, 187, 292, 304, selecting, 149, 305
305, 306 selectivity, 134, 246, 260
room-temperature, 292 self-assembling, 272
root-mean-square, 152 self-focusing, ix, 207, 208, 212, 218, 223
rotations, 93, 397 self-phase modulation, 209, 395
roughness, 40, 132, 211 SEM, 226, 271, 304, 403, 408
roughness measurements, 132 SEM micrographs, 304
routines, 213 semiconductor, vii, 1, 11, 13, 16, 18, 28, 29, 441
RPE, x, 231, 232, 233, 261, 263, 271, 274 semiconductors, xii, 36, 211, 212, 228, 282, 393, 441
RTT, 119 semimetals, 228
Russia, 125, 147, 455 senile, 271
Russian, 27, 68, 134, 158 senile plaques, 271
Russian Academy of Sciences, 27, 134 sensations, 15
Rutherford, 38 sensing, viii, 2, 5, 15, 19, 234
rutile, 309 sensitivity, xii, 223, 234, 269, 329, 393, 399, 400,
412
sensitization, 310
S separation, 32, 33, 34, 129, 143, 269, 304, 337, 352,
353, 358, 451
safety, 411
serotonin, 239, 274
saline, 238
serum, 271
samarium, 313, 314
shape, 12, 15, 30, 32, 33, 34, 35, 39, 42, 63, 76, 102,
sample, 29, 31, 35, 38, 39, 40, 42, 43, 44, 45, 47, 48,
210, 217, 222, 224, 227, 244, 246, 250, 261, 263,
49, 50, 52, 107, 129, 131, 135, 147, 233, 235,
266, 282, 284, 320, 381, 431, 432, 447
236, 238, 244, 250, 260, 263, 269, 339, 341, 342,
shaping, 410
352, 356, 394, 399, 400, 401, 403, 404, 405, 406,
shares, 326
407, 420, 422, 442, 443, 444, 445, 446, 451
shear, 77, 95, 98, 99, 101, 105, 109, 111, 113
sapphire, viii, x, xi, 27, 28, 33, 35, 37, 53, 54, 55, 57,
shock, 209, 273, 323, 397, 411
58, 59, 60, 125, 127, 138, 140, 141, 156, 231,
shock waves, 209, 273, 397, 411
233, 234, 236, 271, 279, 281, 304, 312, 315, 331,
shocks, 323
338, 339, 394, 395, 396, 401, 404, 408, 412, 444
shoulder, 39, 298, 355, 460
satellite, 19, 244, 246, 268
Si3N4, 57
satellite cells, 268
side effects, 409
saturation, 4, 13, 292, 293, 297, 304, 308, 400, 427
sign, 14, 99, 100, 395
scaffolding, 240, 241, 243
signals, xii, 9, 15, 18, 28, 233, 235, 236, 257, 261,
scalability, 329
266, 270, 304, 404, 406, 407, 441, 444, 445, 447,
scalable, 329
448, 449, 450, 452, 453
scalar, 127, 220
signal-to-noise ratio, 242, 269
scaling, 18
silica, ix, 30, 39, 44, 46, 54, 58, 61, 65, 126, 138,
scandium, 333
143, 146, 288, 290, 310
scanning electron microscopy, 271, 274, 275, 393,
silica glass, 44, 46, 54, 58, 288, 310
408
silicate, viii, 27, 31, 38, 40, 47, 48, 49, 51, 283, 284,
scatter, 131, 176
286, 288, 289, 292, 301, 306, 309, 310, 312, 313,
scattering, ix, 47, 132, 173, 191, 207, 209, 228, 234,
325
239, 269, 272, 305, 313, 397, 420, 428, 447, 449
silicates, 324, 331
schema, 460
silicon, 28, 134, 135, 139, 203, 327
Schrödinger equation, 371, 443
scintillators, 334
480 Index
silver, viii, ix, 27, 29, 30, 37, 38, 39, 40, 43, 44, 45, specificity, 248, 255, 269
46, 47, 48, 49, 50, 51, 52, 53, 54, 56, 58, 59, 64, spectroscopic methods, 338, 370
65, 126, 127, 143, 146, 147, 156 spectroscopy, xi, xii, 49, 53, 62, 234, 270, 273, 275,
similarity, 52, 351 308, 315, 319, 331, 338, 364, 369, 370, 393, 400
simulation, vii, 9, 19, 95, 101, 133, 152, 155, 179, spectrum, xii, 15, 18, 30, 39, 44, 48, 50, 58, 61, 62,
190, 369, 370, 371, 373, 375, 376, 378, 380, 381, 64, 65, 98, 99, 234, 235, 242, 243, 244, 256, 263,
384, 385, 386 301, 302, 320, 321, 339, 341, 342, 343, 345, 355,
simulations, xi, 87, 99, 132, 209, 369, 370, 373, 381, 357, 374, 378, 380, 395, 400, 404, 406, 407, 441,
386 442, 443, 444, 445, 447, 448, 450, 452, 458, 460
Singapore, 205 speed, 28, 45, 160, 162, 164, 167, 170, 201, 202,
single cap, 87 213, 217, 226, 227, 305, 410, 411
single crystals, 140, 290, 294, 297, 298, 309, 311, speed of light, 213, 217
323, 333, 334 spin, 317
singular, 331 SPR, 29, 31, 32, 33, 34, 35, 46, 47, 51, 52, 53, 55,
sintering, 180, 190, 205 57, 60, 61, 62, 63
SiO2, 28, 31, 32, 33, 34, 35, 36, 37, 56, 57, 61, 62, sputtering, 33, 36, 38, 39, 40, 54, 56, 400, 407
63, 64, 65, 113, 282, 284, 285, 287, 288, 291, squamous cell, 239, 240, 241, 242
292, 293, 294, 301, 305, 306, 309, 310, 311, 313, St. Petersburg, 455
314 stability, 12, 140, 143, 180, 306, 395
sites, 268, 304, 320, 321, 325, 326, 327, 328, 329, stabilization, 460
331 stages, 60, 134, 140, 143, 212, 395, 396, 420
skeleton, 190 stainless steel, 194, 195
skin, 179, 190, 234, 274, 275, 397, 402, 410 Stark effect, 317
slit lamp, 236, 239 stars, 28
Sm, 280, 312, 313, 341 statistical mechanics, 2
smooth muscle, 268 statistics, 15, 329
sodium, 29, 304, 313, 314 steady state, 17
software, 19, 223, 235, 410, 411 steel, 162, 194, 195, 225, 227, 238
solar, 337, 364 steel plate, 225
solar energy, 337, 364 stent, 409
sol-gel, 282, 284, 287, 290, 293, 294, 305, 306, 310, stiffness, 75, 78, 81, 111, 114
313, 333 stochastic, viii, 125, 148, 149, 152, 153, 156
solid phase, 180, 183, 194 Stokes shift, 372, 386

solid state, vii, xiii, 1, 2, 4, 11, 13, 18, 19, 30, 281, storage, 280, 281, 286, 305, 306, 319
315, 316, 322, 331, 333, 394, 455 strain, 72, 75, 76, 77, 78, 79, 80, 81, 82, 84, 85, 86,
solidification, 44, 45, 163, 180, 181, 182, 186, 202, 91, 94, 108, 109, 110, 111, 114, 119, 121, 161,
208, 227 196, 198, 199, 200, 253, 379
solidification processes, 182 strains, 4, 75, 95, 108, 199, 200
solid-state, x, 3, 231, 280, 281, 282, 284, 286, 288, streams, 28
394 strength, 3, 115, 140, 259, 319, 332, 345, 442
soliton, 214 stress, 2, 4, 18, 78, 85, 86, 197, 198, 199, 268, 402,
solubility, 29, 37, 39, 45, 54, 61 456, 457
solvent, 342, 352, 353, 354, 355, 356, 360, 363, 370, stress fracture, 2
372, 380, 381 stretching, 375, 376, 377, 378, 379, 385, 386, 395
solvent molecules, 355, 370 strokes, 409
solvents, 342, 353, 354, 372, 380, 381, 382 stroma, 239, 240, 241, 243, 244, 245, 247, 248, 249,
SPA, 11 251, 252, 253, 275, 277
spatial, vii, x, 1, 4, 6, 11, 15, 17, 18, 19, 20, 111, stromal, x, 232, 240, 241, 244, 245, 246, 247, 248,
143, 163, 164, 176, 198, 200, 232, 234, 239, 242, 249, 251, 252, 266, 268, 271, 273
244, 246, 251, 261, 263, 269, 274, 316, 433, 442 stromal cells, x, 232, 241, 245
species, xii, 250, 288, 373, 393, 399, 400, 405, 406, structural changes, xi, 279, 282
407, 420, 428, 429, 431, 432 structural protein, 271
specific heat, 160, 221, 430 structuring, 128, 155, 394, 409
Index 481
subcellular, 239, 261, 264, 276 therapy, 204, 270, 364, 398, 411
substances, 244 thermal conduction, 208
substitutes, 329, 395 thermal conductivity, 204
substitution, 372 thermal decomposition, 398
substrates, viii, 28, 29, 46, 125, 129, 134, 135, 138, thermal energy, 168, 196
139, 140, 141, 146, 156 thermal equilibrium, 2, 162, 168, 202, 208, 211, 225,
sulfate, 140, 141 429
sulphate, 308 thermal evaporation, 46
superconducting, 416, 417 thermal expansion, 2, 5, 161, 196, 197, 323, 327, 329
superconductor, 415, 416, 427, 437 thermal lens, 2, 3, 9, 11, 13, 323, 329, 330
superposition, 3, 30, 47, 62, 77, 114, 115, 298, 320 thermal load, vii, 1, 2, 4, 5, 322, 409
surface layer, 31, 36, 42, 51, 52, 57, 58, 62, 132 thermal properties, 128, 195, 323, 325, 327, 330
surface modification, 36, 408 thermal quenching, 292
surface structure, 208 thermal relaxation, 202, 398, 400
surface water, 3 thermalization, 36, 42, 162, 176, 177, 205
surface wave, 75 thermodynamic, 415, 430
surgeons, 410 thermodynamic parameters, 430
surgery, x, 231, 232, 247, 250, 251, 264, 265, 266, thermo-mechanical, 315, 316, 323, 325, 327, 329,
267, 268, 270, 273, 274, 276, 277, 409, 410, 411 331, 333
surgical, x, 232, 265, 266, 267, 410, 412 thin film, 161, 163, 181, 182, 415, 416, 417, 426,
surgical intervention, 410, 412 442, 443, 448, 452
susceptibility, 14, 29, 53, 398, 442 thin films, 416, 417, 426, 442, 443, 448, 452
swelling, 36 third order, 29, 53, 167, 396
switching, viii, 2, 3, 11, 73, 101, 102, 103, 105, 106, Thomson, 322
122, 123, 442 three-dimensional, x, 232, 253, 272, 286, 305, 313,
symbols, 255, 258 424
symmetry, 6, 18, 72, 109, 139, 191, 222, 322, 327, threshold, viii, x, xii, 1, 6, 8, 12, 14, 16, 17, 18, 19,
329, 380, 382, 384, 386 20, 35, 37, 38, 42, 61, 129, 132, 179, 196, 208,
syntax, 76 212, 219, 221, 222, 225, 232, 237, 264, 286, 293,
synthesis, vii, viii, 29, 52, 61, 125, 126, 127, 128, 297, 301, 329, 399, 408, 410, 423, 455, 456, 458
150, 307, 346 thresholds, 179
thulium, 308, 319
time consuming, 129

T time resolution, 274, 395
timing, 100, 400
T cell, 275, 276
TiO2, 282, 286, 301, 305, 309
T cells, 275
tissue, x, xii, 231, 233, 234, 235, 239, 244, 247, 248,
Taiwan, 1, 19
249, 250, 251, 252, 253, 254, 256, 257, 258, 260,
tantalum, 312
261, 262, 263, 266, 267, 269, 270, 272, 273, 274,
targets, 402, 427
275, 277, 393, 394, 397, 398, 399, 400, 402, 403,
telecommunication, 286
407, 408, 409, 410, 411, 412
TEM, 7, 245, 249, 267
titanium, xi, 30, 279, 281, 315, 394, 395
temperature dependence, 128, 416, 417, 425, 426
toluene, 344, 354, 355
temperature gradient, 161, 163, 164, 165, 166, 171,
topology, 129
195, 197, 201, 330
total energy, 154
temporal, 15, 18, 107, 115, 163, 181, 191, 200, 238,
toughness, 323
339, 394, 450, 451, 453
toxic, 398
tendon, 272
TPA, 11
tensile, 200, 259
tracking, 180, 181, 191, 202
tensile strength, 259
trade, 101
tension, 75, 78, 79, 80, 82, 108, 109, 110, 114, 119,
trade-off, 101
121, 259
trans, 58, 372
terawatt, 264
transducer, 111
terbium, 290, 310, 311, 319
482 Index
transfer, vii, ix, xi, xii, 1, 45, 64, 143, 159, 161, 163, two-dimensional (2D), 126
164, 168, 180, 201, 205, 210, 211, 212, 219, 220,
222, 237, 270, 281, 282, 286, 287, 290, 294, 297,
300, 301, 302, 306, 307, 308, 310, 311, 312, 319, U
332, 335, 338, 342, 343, 349, 350, 352, 355, 357,
ultraviolet, 30, 58, 59, 62, 141, 156, 204, 255, 259,
358, 359, 360, 361, 363, 364, 369, 370, 380, 382,
260, 293, 300, 312, 337, 338
386, 387, 393, 394, 398, 409, 410
ultraviolet light, 59, 338
transformation, 60, 64, 108, 180, 266, 268, 277, 301,
uniform, vii, 1, 2, 3, 4, 17, 30, 38, 39, 45, 50, 52, 53,
309, 394
65, 163, 172, 176, 182, 431
transformations, 213
United States, 156
transforming growth factor, 268, 272
USSR, 366
transforming growth factor-β, 268
UV, viii, ix, 2, 42, 50, 62, 125, 126, 128, 132, 134,
transgenic, 274
233, 234, 259, 288, 310, 336, 370, 397, 398, 410,
transgenic mouse, 274
415, 416, 417, 420, 421, 425, 426
transglutaminase, 271
UV light, 234, 336
transition, xi, 18, 98, 99, 279, 280, 286, 287, 288,
UV radiation, 132, 233
291, 292, 294, 298, 301, 302, 304, 308, 317, 319,
322, 340, 342, 343, 348, 350, 354, 355, 363, 373,
375, 379, 380, 381, 382, 383, 386, 442, 444 V
transition metal, 280, 294, 301, 302
transition metal ions, 280, 301, 302 vacancies, 321
transition rate, 286 vacuum, ix, 18, 90, 106, 135, 140, 207, 209, 213,
transitions, 29, 59, 282, 284, 285, 287, 288, 291, 218, 401, 417, 418, 427, 431, 457
292, 293, 294, 296, 297, 300, 303, 304, 312, 316, valence, xi, 42, 59, 174, 279, 282, 288, 292, 305,
317, 319, 321, 334, 340, 347, 375, 376, 382, 451 313, 371
translation, 396, 397 values, 9, 14, 19, 29, 40, 45, 51, 54, 56, 58, 59, 84,
transmission, 5, 9, 39, 52, 53, 58, 61, 62, 72, 73, 85, 86, 89, 90, 91, 101, 102, 107, 109, 110, 113, 131,
92, 93, 94, 97, 98, 99, 101, 102, 103, 104, 106, 139, 148, 150, 154, 166, 167, 176, 177, 195, 200,
107, 115, 116, 126, 127, 132, 147, 163, 245, 249, 201, 210, 215, 217, 218, 219, 222, 250, 257, 264,
272 306, 319, 321, 325, 327, 346, 348, 352, 353, 354,
transmits, 14, 115 355, 356, 357, 360, 361, 384, 386, 397, 433, 435,
transmittance spectra, 42 445, 447, 451
transparency, 10, 30, 61, 62, 139, 140, 239, 241, 244, vapor, 210, 420, 422, 425, 427, 428, 429, 434, 435
246, 456 variables, 75, 96, 107, 164, 199, 214
transparent, xi, 14, 58, 60, 64, 65, 72, 208, 209, 251, variation, 2, 4, 16, 129, 135, 224, 247, 250, 257, 261,
270, 279, 281, 282, 286, 288, 305, 309, 310, 312, 263, 319, 322, 332, 340, 376, 377, 379, 452
313, 314, 323, 410 vector, 17, 19, 76, 92, 95, 96, 108, 110, 111, 113,
transplantation, 276 114, 160, 161, 163, 165, 166, 172, 196, 197, 201,
transport, ix, 159, 162, 171, 172, 196, 197, 198, 202, 220, 221, 371, 372
204, 352 velocity, 13, 18, 74, 78, 86, 90, 103, 106, 111, 160,
traps, 280, 305 162, 163, 169, 172, 173, 175, 176, 179, 180, 182,
travel, 395 183, 185, 186, 187, 190, 194, 203, 395, 417, 422,
triggers, 397 423, 427, 431, 432, 448
trimer, 346, 351, 352 vibration, 360, 363, 376, 385, 397
Trp, 402 vibrational modes, 375
tsunami, 367 viscosity, 373, 379, 380
TTM, 180 visible, x, 12, 13, 15, 30, 55, 63, 71, 72, 107, 116,
tumor, 204, 272 133, 223, 232, 239, 253, 269, 280, 281, 288, 297,
tumor cells, 204 307, 308, 310, 316, 336, 337, 339, 394
tumor progression, 272 visual processing, 261
tumors, 271, 409 visualization, 239, 244, 253, 263, 270
tungstates, 324, 327
two-dimensional, 63, 126, 127, 215
Index 483
W X
water, 143, 209, 234, 249, 250, 269, 273, 276, 289, X-ray crystallography, 353
339, 398, 402, 408 X-ray diffraction, 57, 370
wave number, 212
wave power, 18
wave propagation, 80, 167 Y
wave vector, 113, 447
yield, 39, 79, 82, 170, 171, 245, 341, 342, 361, 372,
waveguide, ix, xii, 28, 126, 142, 143, 282, 309, 455,
400, 405
456
ytterbium, 310, 315, 316, 317, 319, 320, 322, 323,
wavelengths, x, 15, 19, 29, 30, 39, 42, 43, 47, 50, 51,
324, 327, 329, 334
55, 57, 60, 62, 63, 64, 65, 147, 232, 233, 235,
yttrium, 327
237, 245, 247, 248, 254, 260, 263, 269, 291, 305,
323, 347, 395, 400, 420, 428
weak coupling limit, 354 Z
weak interaction, 351
welding, 394 zinc, 341, 343, 344, 346, 349, 353
windows, 128 ZnO, 287, 288, 291, 292, 293, 294, 296, 309, 310,
wires, 27, 87, 88 312, 461
workers, 18, 346 ZnO nanostructures, 309
wound healing, 247, 266, 267, 268, 270, 273
wound repair, x, 231, 253, 268

High-Power and Femtosecond Lasers: Properties, Materials and Applications: Properties, Materials and Applications, Nova Science Publishers

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

High-Power and Femtosecond Lasers: Properties, Materials and Applications: Properties, Materials and Applications, Nova Science Publishers

Uploaded by

Copyright:

Available Formats

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

HIGH-POWER AND FEMTOSECOND

HIGH-POWER AND FEMTOSECOND

Nova Science Publishers, Inc.

NOTICE TO THE READER

AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS.

LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA

Published by Nova Science Publishers, Inc. 

Processing of Metals 159

Chapter 10 Ultrafast Dynamics of Porphyrins in Higher Excited State 335

scleral tissues in their native environment. Finally, included is a review of recent

of the hot-electron blast in multi-layered thin metal films is also presented.

controlled experimentally thereby yielding a cleaner fabricated surface. Due to significant

transfer of higher excited-state energy is realized, have been demonstrated by various

thickness show enhanced excitonic optical nonlinearity, which enhancement is described by

RECENT PROGRESS OF HIGH PEAK POWER SOLID

Yung-Sheng Huang∗1, Jung-Sheng Huang1 and Fang-Ling Chang2

2. THERMAL FOCAL LENGTH MEASUREMENT

Figure 3. TEM 00 laser mode in the x-y plane.

Figure 4. TEM 5, 0 laser mode in the x-y plane.

Figure 5. LG 1,5 laser mode in the polar coordinate.

4. RECENT DEVELOPMENT OF HIGH POWER SIMULTANEOUSLY

Figure 8. The same conditions as in Figure 7 , except for T 0 =0.5.

Figure 9. The same conditions as in Figure 7, except for T 0 =0.35.

5. COATING TECHNIQUE AFFECTING THE LASER OUTPUT BEAM

determining the beam quality of the laser output.

6. QPM TECHNIQUE AND QUASI PHASE MATCHING NONLINEAR

7. THEORY OF PARAMETRIC OSCILLATION (OPO)

8. PRACTICAL OPO PERFORMANCE AND RECENT DEVELOPMENT

× exp ⎢− 2⎜⎜ P 2 2 1 ⎟⎟⎜⎜ x − 2 Δx ⎟ ⎥

(1 − exp[− αL]) exp 2δ

[37] D. C. Brown, IEEE J. Quantum Electronics, 33, 861-873, 1997.

Am. B 20, 2319-2327, 2003.

LASER ANNEALING OF COMPOSITE MATERIALS

Table 1. Laser annealing of MNPs in solid dielectric matrix

ION SYNTHESIS OF METAL NANOPARTICLES IN DIELECTRIC MATRIX

Figure 3. Three-dimensional photon tunnel microscope image of sapphire implanted with

INTERACTION OF HIGH-POWER EXIMER LASER PULSES WITH

pulse length of 25 ns full-width at half-maximum at a wavelength of 248 nm with the total

MULTIPULSE LASER ANNEALING OF SODA-LIME GLASS WITH

penetration depth of in the sample is assumed to be approximately 60 nm. Inward silver

COMBINED LASER AND THERMAL ANNEALING OF SODA-LIME GLASS

pulse length of 25 ns full-width at half-maximum at a wavelength of 248 nm with the total

In Figure 17 modeled Mie extinction spectra of silver nanoparticles in the glass

samples can be furnace annealed.

LASER ANNEALING OF SIPPHIRE WITH ION-SYNTHESISED

LASER ANNEALING OF SILICA WITH ION-SYNTHESISED

Composite Materials (Elsevier Science, Amsterdam, 1991).

SINGLE CRYSTAL PHOTO-ELASTIC MODULATORS

1. BASICS OF SINGLE CRYSTAL PHOTO-ELASTIC MODULATORS

Conventional photo-elastic modulators (PEMs, Figure 1) modulate the polarization of a

Figure 1. A conventional photo-elastic modulator (PEM) [1].

1.2. Definition and Description of a SCPEM

A Single Crystal Photo-Elastic Modulator (SCPEM) is a piezoelectric optical transparent

T(δ) = cos2(δ/2) = (1 + cos δ)/2 . (2)

This is deduced in appendix 5.4.

T(δ) = (1 + cos (δ ± π/2))/2 (3)

1.3. Two Candidates for a SCPEM: LiTaO3 and LiNbO3

2.1. Reduction to One Dimension: Y-Excited X-Oscillation

We consider a y-excited oscillation in x-direction (Figure 2). For this purpose it is

tension: σ1 = C ε1 – e E2 [N/m²] (4)

electric displacement: D2 = e ε1 + χ E2 [C/m²]

C = 202.813 GPa, e = 1.59482 C/m², χ = 4.62033 10 −10 As/Vm.

Published by Nova Science Publishers, Inc.