Physics of Interaction of Ultrashort Laser Pulses With Solids RPP2011

Physics Reports 508 (2011) 91243
Contents lists available at SciVerse ScienceDirect

Physics Reports
journal homepage: www.elsevier.com/locate/physrep
The physics of ultra-short laser interaction with solids at
non-relativistic intensities
E.G. Gamaly
Laser Physics Centre, Research School of Physics and Engineering, The Australian National University, Canberra, ACT 0200, Australia
a r t i c l e i n f o
Article history:
Accepted 20 July 2011
Available online 16 September 2011
editor: G.I. Stegeman
Keywords:
Ultra-short laser interaction with matter
Electronphonon coupling
Laser-induced melting
Pulsed laser ablation
Laser-induced micro-explosion
a b s t r a c t
The interaction of ultra-fast sub-picosecond laser pulses with solids is a very broad area
of research. The boundaries for research fields covered by this review are defined as
follows. A laser pulse in the context of the review is of ultra-short duration if the pulse is
shorter than all major relaxation times. Such pulses excite only electrons, leaving the lattice
cold for the time required for the transfer of the absorbed laser energy from the heated
electrons to the lattice. For this reason, any laser-induced phase transformations occur in
non-equilibrium conditions, making properties of the material drastically different from
their equilibrium counterparts. We study laser interaction with matter in a broad range
of intensities from those inducing subtle atomic excitations (10
10
W/cm
2
) up to high
intensity (10
16
W/cm
2
), when solid is swiftly transformed into hot and dense plasma.
The phenomena emerging in succession in response to increasing laser intensity, namely,
the excitations of coherent phonons, phase transitions, ablation, and transformation of
material into plasma, are described in consecutive chapters. Two interaction geometries
are investigated: the interaction of a laser pulse with a surface, and confined interaction
when a laser is focused inside a transparent solid. The highest intensity in all these studies
is well below the relativistic limit. Therefore, super-intense lasermatter interactions are
beyond the scope in this review.
All phenomena involved in lasermatter interaction are considered from the first
principles using explicit approximations, eventually aiming to establish the analytical
scaling relations, which link the parameters of the laser and the material and allow
comparison with experiments. We compare theory to experiments in all intensity ranges.
The applications of some studies are described in a separate chapter. The prospects of these
studies are indicated in the conclusion.
2011 Elsevier B.V. All rights reserved.
Contents
1. Introduction............................................................................................................................................................................................. 94
2. Fundamentals of ultra-short laser absorption in solids........................................................................................................................ 96
2.1. Introduction to interaction with strongly absorbing media .................................................................................................... 96
2.2. Laser field penetration into a solid ............................................................................................................................................ 96
2.3. Laser energy absorption and response of the material ............................................................................................................ 97
2.3.1. Skin layer approximation............................................................................................................................................ 97
2.3.2. Absorbed energy density............................................................................................................................................. 97
2.3.3. Electron temperature at the end of the pulse............................................................................................................ 98
2.3.4. Lattice temperature after equilibration...................................................................................................................... 99
Tel.: +61 2 6125 8659; fax: +61 2 6125 0029.

E-mail address: gam111@physics.anu.edu.au.
0370-1573/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.physrep.2011.07.002
92 E.G. Gamaly / Physics Reports 508 (2011) 91243
2.4. Modification of the material during the laser pulse................................................................................................................. 100
2.4.1. The Drude-like dielectric function.............................................................................................................................. 100
2.4.2. Electronic excitation and ionisation........................................................................................................................... 101
2.4.3. Modified dielectric function and heat capacity at high temperature ...................................................................... 102
2.4.4. Electron and ion temperature in the ionised solid.................................................................................................... 102
2.5. Relaxation processes in laser-excited solids............................................................................................................................. 103
2.5.1. Establishing the main parts of the distribution functions ........................................................................................ 103
2.5.2. Electronphonon interaction...................................................................................................................................... 105
2.5.3. Importance of the high-energy tail in the Maxwell distribution ............................................................................. 109
2.6. Electronion collisions in solid-density plasma ....................................................................................................................... 111
2.6.1. Electron-to-ion momentum exchange rate ............................................................................................................... 111
2.6.2. Electron-to-ion energy transfer rate .......................................................................................................................... 113
2.7. General picture for the electron-to-lattice (ion) exchange rate: from solid to plasma ......................................................... 113
2.8. Electronic heat conduction ........................................................................................................................................................ 114
2.9. Atomic motion in a solid induced by the ultra-short pulse action.......................................................................................... 115
2.9.1. Atomic response on the action of external high-frequency electric field................................................................ 115
2.9.2. Single electron and ion motion................................................................................................................................... 116
2.9.3. Macroscopic forces acting on a solid in the external laser field ............................................................................... 116
2.9.4. Heating and motion of atoms in laser-affected solid ................................................................................................ 117
2.10. Summary ..................................................................................................................................................................................... 118
3. The evolution of atomic motion: the origin, life, and death of coherent phonons ............................................................................. 118
3.1. Introduction ................................................................................................................................................................................ 118
3.2. Initial conditions in a laser-excited solid at the energy density comparable to enthalpy of melting................................... 119
3.2.1. Absorbed energy density, electron and lattice temperatures................................................................................... 119
3.2.2. Excitation of electrons................................................................................................................................................. 121
3.2.3. Stress tensor in laser-excited solid............................................................................................................................. 121
3.2.4. Time scales for the consecutive stages of atomic motion......................................................................................... 121
3.3. Forces driving atomic motion in solids ..................................................................................................................................... 122
3.3.1. Elastic force in an unperturbed solid.......................................................................................................................... 122
3.3.2. The forces in laser-excited solid ................................................................................................................................. 122
3.3.3. Equation of atomic motion in a laser-excited solid................................................................................................... 123
3.4. Consecutive stages of laser-induced atomic motion................................................................................................................ 124
3.4.1. Coherent atomic displacement ................................................................................................................................... 124
3.4.2. Harmonic vibrations under electronic force.............................................................................................................. 124
3.4.3. Non-linear phenomena during the process of lattice heating.................................................................................. 124
3.5. Effect of atomic motion on the transient properties of the laser-excited solid...................................................................... 126
3.5.1. Dielectric function ....................................................................................................................................................... 126
3.5.2. Time-dependent reflectivity ....................................................................................................................................... 127
3.6. Comparison with experiments .................................................................................................................................................. 128
3.6.1. Measurements of time-dependent reflectivity with fast optical probes ................................................................. 128
3.6.2. Atomic vibrations observed with fast X-ray and electronic probes ......................................................................... 131
3.7. Summary ..................................................................................................................................................................................... 135
4. Ultra-fast melting by fs-lasers................................................................................................................................................................ 135
4.1. Introduction ................................................................................................................................................................................ 135
4.2. Revisiting melting in equilibrium.............................................................................................................................................. 136
4.2.1. Criteria for melting ...................................................................................................................................................... 137
4.2.2. Critical entropy and critical temperature .................................................................................................................. 138
4.3. Transformations induced by ultra-fast heating ........................................................................................................................ 139
4.3.1. Distribution functions in a swiftly heated solid ........................................................................................................ 139
4.3.2. Generation of thermal defects driven by ultra-fast heating ..................................................................................... 140
4.3.3. Entropy changes produced by electron excitation.................................................................................................... 141
4.3.4. The onset of ultra-fast disordering: entropy catastrophe by superheating of the lattice....................................... 142
4.3.5. Summary: sequence of events before the onset of disordering ............................................................................... 143
4.4. Phase transformation: heterogeneous and homogeneous nucleation.................................................................................... 143
4.5. Transient state of matter created by the ultra-fast excitation................................................................................................. 144
4.6. Ultra-fast melting of metals and dielectrics.............................................................................................................................. 145
4.7. Comparison to experiments....................................................................................................................................................... 146
4.7.1. Superheating of ice ...................................................................................................................................................... 146
4.7.2. Superheating of Gallium ............................................................................................................................................. 146
4.7.3. Superheating of Aluminium........................................................................................................................................ 147
4.7.4. Ultra-fast excitation of Bismuth ................................................................................................................................. 148
4.7.5. Ultra-fast melting of dielectrics: experiments with superheated Indium Antimonide.......................................... 150
4.8. Conclusions ................................................................................................................................................................................. 151
4.9. Summary ..................................................................................................................................................................................... 151
5. Ablation of solids..................................................................................................................................................................................... 152
5.1. Introduction ................................................................................................................................................................................ 152
E.G. Gamaly / Physics Reports 508 (2011) 91243 93
5.2. Evaporation in equilibrium........................................................................................................................................................ 153
5.2.1. Cold pressure and energy ........................................................................................................................................ 154
5.2.2. Saturated pressure of vapours in equilibrium........................................................................................................... 154
5.2.3. Evaporation rate .......................................................................................................................................................... 155
5.2.4. Number of particles in the high energy tail of the Maxwell distribution ................................................................ 155
5.3. State of solid excited by ultra-short pulse at the energy density around the equilibrium enthalpy of vaporisation.......... 156
5.3.1. Swift ionisation of dielectrics...................................................................................................................................... 156
5.3.2. Relaxation processes in laser-excited solids.............................................................................................................. 156
5.3.3. Skin-layer approximation ........................................................................................................................................... 157
5.3.4. Electron-to-ion energy-transfer time......................................................................................................................... 157
5.3.5. Heat diffusion time...................................................................................................................................................... 157
5.3.6. Absorbed energy density in dielectrics below and at the ablation threshold ......................................................... 157
5.3.7. Maximum electron and ion temperatures ................................................................................................................. 159
5.4. Mechanisms of ablation by ultra-short laser pulses................................................................................................................. 159
5.4.1. Electrostatic ablation: ions pulled out by energetic electrons ................................................................................. 160
5.4.2. Non-equilibrium, semi-thermal ablation (k
B
T
ion

b
) ............................................................................................ 161
5.4.3. Thermal evaporation (k
B
T
ion
<
b
) ............................................................................................................................. 161
5.5. Single pulse ablation thresholds................................................................................................................................................ 161
5.5.1. Ablation thresholds in vacuum................................................................................................................................... 161
5.5.2. Long-lived non-equilibrium state in the outermost surface layer ........................................................................... 163
5.5.3. Ablation thresholds in ambient gas............................................................................................................................ 165
5.5.4. Comparison with the experimental data ................................................................................................................... 166
5.6. Ablation rate, mass, and depth by a single pulse...................................................................................................................... 168
5.7. Control over the ablation rate and phase state of laser-produced plume: spatial pulse shaping ......................................... 170
5.7.1. Local energy thresholds for phase transitions ........................................................................................................... 170
5.7.2. Criterion for total atomisation of ablated plume....................................................................................................... 171
5.7.3. Surface damage and evaporation ............................................................................................................................... 172
5.7.4. Optimum pulse profile for atomisation of the plume ............................................................................................... 172
5.7.5. Experiments: ablation and deposition of silicon films by spatially shaped pulses ................................................. 174
5.8. Accumulation effects in ablation of solids by high-repetition-rate short-pulse lasers.......................................................... 174
5.8.1. Dwell time and number of pulses per focal spot ....................................................................................................... 175
5.8.2. Smoothing of the evaporation conditions on the surface ......................................................................................... 175
5.8.3. Temperature accumulation in a multiple-pulse action on poor heat-conducting dielectrics................................ 175
5.8.4. Thermal ablation threshold in multiple-pulse action ............................................................................................... 176
5.8.5. The build-up of plume density near the target surface............................................................................................. 176
5.8.6. Change in the interaction regime ............................................................................................................................... 177
5.8.7. Concluding remarks on cumulative ablation............................................................................................................. 177
5.9. Summary ..................................................................................................................................................................................... 178
6. Confined interaction: laser beams tightly focused inside transparent solids ..................................................................................... 178
6.1. Introduction ................................................................................................................................................................................ 178
6.2. Focusing of the beam inside a transparent solid ...................................................................................................................... 179
6.2.1. Limitations imposed by the spherical aberrations .................................................................................................... 179
6.2.2. Limitations imposed by the laser beam self-focusing............................................................................................... 180
6.2.3. Laser intensity distribution in a focal domain ........................................................................................................... 181
6.3. Non-destructive interaction: formation of diffractive structures in photo-refractive materials.......................................... 182
6.3.1. Properties of lithium niobate...................................................................................................................................... 182
6.3.2. Electronic excitation by a low-intensity laser field................................................................................................... 182
6.3.3. Electrons excitation by the high-intensity ultra-short pulse.................................................................................... 184
6.3.4. Modification of the optical properties of fs-laser-excited crystal ............................................................................ 186
6.3.5. Quasi-stationary changes in optical properties of fs-laser-affected crystals........................................................... 187
6.3.6. Possible mechanisms for changes in the refractive index after the pulse ............................................................... 188
6.3.7. Comparison with experiments ................................................................................................................................... 188
6.4. Lasermatter interactions confined inside a solid at high intensity ....................................................................................... 189
6.4.1. Absorbed energy density............................................................................................................................................. 190
6.4.2. Ionisation...................................................................................................................................................................... 190
6.4.3. Increase in the absorbed energy density due to modification of optical properties............................................... 192
6.4.4. Energy transfer from electrons to ions: relaxation processes after the pulse ......................................................... 192
6.4.5. Shock-wave propagation, stopping, and void formation.......................................................................................... 193
6.4.6. Computer modelling of a confined micro-explosion ................................................................................................ 195
6.4.7. Density and temperature in the shock- and heat-affected solid.............................................................................. 196
6.4.8. Upper limit for the pressure achievable in confined interaction ............................................................................. 197
6.4.9. The propagation of an optical breakdown wave from the laser-absorption zone inside a solid ........................... 197
6.4.10. Similarity laws of hydrodynamics: can we model macroscopic explosions in tabletop conditions?.................... 198
6.4.11. Experimental observation of void formation in different materials ........................................................................ 199
6.4.12. Multiple-pulse interaction: energy accumulation .................................................................................................... 200
6.5. Laser-induced forward transfer ................................................................................................................................................. 201
6.5.1. Threshold conditions................................................................................................................................................... 201
6.5.2. Electron and ion temperature, relaxation time in the skin layer ............................................................................. 203
6.5.3. Propagation of a heat wave and a shock wave formation......................................................................................... 203
6.5.4. Expansion of a laser-affected solid into vacuum....................................................................................................... 204
6.5.5. Strong shock-wave created by a solid expanding in ambient gas............................................................................ 205
6.6. Summary ..................................................................................................................................................................................... 206
7. Applications of ultra-fast laser interaction with matter....................................................................................................................... 206
7.1. Introduction ................................................................................................................................................................................ 206
7.2. Experimental results: ablation and deposition of thin films ................................................................................................... 207
7.2.1. Ablation and deposition of carbon films with atomic surface quality ..................................................................... 208
7.2.2. Ablation and deposition of homogeneous chalcogenide glass films........................................................................ 208
7.2.3. Ablation by powerful short-pulse multi-MHz-repetition-rate lasers ...................................................................... 208
7.3. Precise micro-machining: removal of dental enamel by ultra-fast laser ablation................................................................. 209
7.4. Formation of nanoclusters by ultra-short pulses...................................................................................................................... 211
7.4.1. Kinetic model of cluster formation in a plume of laser-ablated atoms.................................................................... 212
7.4.2. Growth of nanoclusters in ambient gas and in vacuum............................................................................................ 217
7.5. Three-dimensional optical memory.......................................................................................................................................... 222
7.5.1. Recording in photorefractive media........................................................................................................................... 222
7.5.2. Reading, erasing, re-writing........................................................................................................................................ 223
7.6. Three-dimensional networks..................................................................................................................................................... 223
7.7. Creation of a super-dense material in the confined micro-explosion..................................................................................... 224
7.7.1. Progress in formation of new super-dense materials ............................................................................................... 224
7.7.2. Evidence of super-dense aluminium produced in confined micro-explosion......................................................... 225
7.8. Application of fs-lasers in art restoration.................................................................................................................................. 226
8. Conclusion remarks and future directions ............................................................................................................................................ 227
8.1. Ultra-short lasers ........................................................................................................................................................................ 227
8.1.1. Solid-state laser systems............................................................................................................................................. 227
8.1.2. Attosecond lasers......................................................................................................................................................... 227
8.1.3. Free electron lasers...................................................................................................................................................... 228
8.1.4. Sculptured laser beams............................................................................................................................................ 228
8.2. Theory and computer modelling ............................................................................................................................................... 228
8.3. Controlled excitation of phonons and coherent sound waves................................................................................................. 228
8.4. Control over chemical reaction-rates and transient phase states ........................................................................................... 229
8.5. Warm Dense Matter on the tabletop: mimicking conditions in stars and creating new materials ...................................... 229
Acknowledgements................................................................................................................................................................................. 230
Appendix A. Properties of some elements........................................................................................................................................ 230
Appendix B. Temperature-dependent bismuth properties from optical measurements in equilibrium..................................... 231
B.1. Electronphonon momentum exchange rate ........................................................................................................................... 231
B.2. Electronphonon energy exchange rate ................................................................................................................................... 232
B.3. Number of electrons in the conduction band from the optical measurements ..................................................................... 233
B.4. Electronic heat conduction and characteristic cooling time.................................................................................................... 233
Appendix C. Empirical pseudo-chemical inter-atomic potential.................................................................................................... 234
Appendix D. Ablation of metals......................................................................................................................................................... 235
Appendix E. Micro-explosion in transparent solids: material parameters and ionisation losses ................................................ 236
References................................................................................................................................................................................................ 236
1. Introduction
Several definitions are needed to make the field of ultra-fast lasermatter interaction and its boundaries clear. First, the
laser pulse is considered as of ultra-short duration if it is shorter than the duration of major relaxation processes, including
the electron-to-lattice energy transfer, heat diffusion and hydrodynamics. Therefore, an atomic motion is insignificant in the
laser-affected solid during the pulse and the atomic structure remains intact. The laser intensity to create such conditions
ranges fromthe lowlimit of 10
10
W/cm
2
up to 10
16
W/cm
2
, the upper intensity being well belowthe relativistic threshold.
A J laser pulse focused down to a micron-sized focal spot delivers energy density (fluence) from mJ/cm
2
up to 100s J/cm
2
to the surface. The associated absorbed energy density per unit volume delivered to a sample allows one to study the excited
solidstate fromsubtle excitations tophase transformations, suchas transitionfromone crystalline structure toanother, from
a solid to liquid, conversion to plasma, ablation, and, if the laser is focused in the bulk of a transparent dielectriccreation
of extreme pressure and temperature conditions confined inside a solid.
Second, the salient feature of the ultra-short pulse lasermatter interactionis that ultra-short pulses excite only electrons,
leaving the lattice cold for the time required for the transfer of the absorbed energy from the hot electrons to the lattice.
Therefore, any phase transformations in a laser-affected solid occur in non-equilibrium conditions, creating properties
of the material drastically different from their equilibrium counterparts. The statistical distributions in the electron and
lattice sub-systems are time-dependent, and the time required for attaining the state of equilibrium state depends heavily
on the parameters of the laser pulse and the characteristics of the material. The research field defined this way covers a
broad range of physical effects, as well as applications of the ultra-short laser interaction in material science, industry and
medicine.
Another important feature of the ultra-short lasermatter interaction is that the affected area is of only a few hundred
atomic layers, which relaxes within picoseconds after the excitation. Thus, the space constraints and surface phenomena
(surface states and quasi-equilibriumdistribution in the outermost surface layers) are essential for understanding the phase
transitions induced, such as laser ablation. Understanding the phenomena occurring on a time-scale of picoseconds and on
a space-scale of nanometres is essential for ultra-short lasermatter interaction, bringing these studies closer to research
into nanostructures.
Ultra-fast lasersolid interaction studies in the last twenty years have been naturally separated into the sub-fields in
accordance with the absorbed energy density. One can easily see that a high-intensity laser is required in order to induce
a particular transformation by an ultra-short pulse. Indeed, the absorption depth for metals is in the range of 1050 nm.
In order to melt a metal in equilibrium an energy density (the enthalpy of melting) of 0.51 kJ/cm
3
is required, while
for evaporation an energy density in a range 5060 kJ/cm
3
is needed. Thus the intensity ranges for a 100 fs pulse and
the relevant surface energy density over the focal spot, where particular phenomena may occur, can be estimated. Phase
transformations may be induced in a skin layer by short laser at fluence in excess of several mJ/cm
2
and intensity more than
10
10
W/cm
2
, while evaporation requires 0.21.0 J/cm
2
and intensity in the range TW/cm
2
(TW = 10
12
W). Consequently,
the sub-fields of research of the different states of excited solids include studies of fs-laser-induced subtle atomic motion
and coherent atomic vibrations, phase transitions, optical breakdown and ionisation, laser ablation and generation of dense
and hot plasma confined inside a cold solid.
Ultra-short lasermatter interaction studies are multidisciplinary by their nature. The areas of physics include the
interaction of electromagnetic field with solids and plasma, elements of atomic physics, plasma physics, optics, solid-state
physics, and statistical physics. This review is an attempt to present the major characteristics of lasermatter interaction,
such as the absorbed energy, electron and lattice temperature, relaxation times, ablation rates and other parameters by
simple analytical formulae, and scaling relations, which combine the laser irradiation conditions with properties of the
material properties and have clear physical meaning. In most cases the way from the first principles to the scaling relation
is indicated, allowing better understanding and control over the phenomenon described. However, chemistry and quantum
chemistry, electro-magnetic interactions with a single atom, laser-induced fusion, laser acceleration of particles, X-ray and
high harmonic generation, femtosecond X-ray diffractometry and attosecond physics are beyond of the scope of the review.
The studies in this review are restricted to the laser intensity well below the relativistic limit. There are excellent books
covering extreme energy-density physics such as Short Pulse Laser Interaction with Matter: An Introduction by Paul Gibbon,
and High Energy Density Physics by Paul Drake, where detailed accounts of super-intense lasermatter interactions can be
found.
The main frame of this review builds on the journal papers, reviews and book sections written by the author in close
cooperation with the colleagues. We use the approaches and experience gained from the classic works on long-pulse
interactions in laser-fusion studies. Kruers LaserPlasma Interaction (Kruer, 1988), the works of Afanasiev and Krokhin
(1971), Anisimov et al. (1971) and Anisimov and Lukyanchuk (2002) and the fundamental book by Ilinsky and Keldysh
Electromagnetic Response of Material Media (Ilinskii and Keldysh, 1994) are indispensable for a thorough understanding of
the physics of lasermatter interaction. We leave aside the descriptions of femtosecond lasers and the methods of generating
femtosecond pulses. These can be found in a number of excellent books, such as the first very good review of this subject
by Akhmanov et al. (1988), covering almost every issue of femtosecond laser-pulse phenomena, and a recent review of
ultrafast laser systems by Keller (2003) and the references therein. Nor do we cover the description of computer simulations
using complex computer codes such as Molecular Dynamics (MD) and Particle-in Cells (PIC) codes. The formulation of the
problemfor these computer simulations needs to be explained thoroughly and linked to physics, mathematics and relevant
computational problems. Such books as Laserplasma Interactions by William Kruer (Kruer, 1988) and Short pulse laser
interactions with Matter: An Introduction by Paul Gibbon (Gibbon, 2005) may be recommended for a first acquaintance with
PIC codes, while the references in these books allowone to develop an advanced knowledge. The MDcodes and simulations,
which treat different aspects of lasermatter interaction, can be found in the original papers, for example, (Car and Parinello,
1985; Gong et al., 1993; Ivanov and Zhigilei, 2003; Lin and Zhigilei, 2008).
The phenomena described in this review are arranged successively in accordance with increasing energy density
deposited in the laser-affected material, and therefore with increasing laser intensity. In Section 2 the basics of lasermatter
interactions are described. Section 3 is devoted to the excitation of subtle atomic motion, followed by coherent atomic
vibrations in various solids at various energy densities below and well above the threshold for phase transformation. The
non-equilibriumprocesses preceding phase transitions allowdeeper insight into the nature of well-known transformations
such as melting. Ultra-fast phase transitions induced by ultra-short lasers are the subject of Section 4. Ultra-fast evaporation
processes driven by short laser pulses, laser ablation, are the subjects of Section 5. Non-equilibrium and thermal ablation
mechanisms and ablation rates are considered in detail, along with the link to a long-pulse interaction mode. Processes
of lasermatter interaction when the laser focal zone is confined inside the bulk of a transparent solid are discussed in
Section 6. Reversible phase transitions at moderate intensity are considered first. Then the theoretical and experimental
studies of laser-induced micro-explosions confined in the bulk of a solid are described in detail. In Section 7 the numerous
applications of ultra-short lasersolid interaction are discussed, concentrating mainly on the applications of laser ablation
processes already implemented. In the conclusion, we address some unresolved fundamental and applied problems that
may be encountered in the near future.
The field of research covered in this review is very broad and embraces many thousands of references. I consider the
papers citedas guides inthe vast sea of different results, models andopinions. The list of references is by nomeans exhaustive
and reflects the authors views.
2. Fundamentals of ultra-short laser absorption in solids
2.1. Introduction to interaction with strongly absorbing media
The definition of an ultra-short pulse in this review relates to pulses with duration shorter than the major relaxation
time for electron-to-lattice energy transfer, heat conduction and hydrodynamic expansion. In practice, these are pulses
with duration of 100 fs or less. In order to create conditions for the ultra-short pulse lasermatter interaction defined
above, the laser pulse should comply with several strict conditions. First, the high-energy contrast should be achieved. That
means that there should be no pre-pulse or after-pulse accompanying the main pulse, as these may contain an amount of
energy sufficient to produce phase changes before or after the interaction of the main ultra-short pulse with material. In
practical terms the ratio of the average intensity during the main pulse to that in the pre-pulse (the contrast ratio) should
constitute 10
9
or higher. The intensity distributionover the focal spot also strongly affects the final results of a lasermatter
interaction experiment, as we demonstrate in the following sections. We assume in this section that the intensity is constant
inthe whole space over the focal spot anddrops tozero at the focal spot boundaries. The laser pulse inthis case has a so-called
top-hat spatial intensity distribution over the focal spot. The set of initial laser parameters sufficient for the description of
lasermatter interaction includes the laser wavelength , energy per pulse E
las
, pulse duration t
p
, and the focal spot area S
foc
,
assuming that the focal area is a circle. Useful characteristics are also the average intensity
I during the pulse time, defined

as,

I = E
las
/S
foc
t
p
and the fluence F as the surface energy density, F = E
las
/S
foc
. These laser characteristics allow one to
describe many details of a single pulse lasermatter interaction. In the multiple-pulse interaction many pulses hit the same
spot on the target surface, thus the laser repetition rate and laser beam scanning speed over the sample surface should also
be accounted for; this will be considered in detail in Section 5.
We are mainly considering interactions with strongly absorbing media. The absorption coefficient is defined as the ratio
of the absorbed fluence (or intensity) to the incident one. In most cases, the phase transformation of the material occurs in
a non-transparent medium, where the transmission may be ignored. The laser energy is absorbed in a thin layer, the skin
layer l
s
, with a typical thickness of several tens of nanometres, which is much smaller than the diameter of the focal spot.
For this reason the description of the interaction is presented in one-dimensional approximation in most cases.
The laser-affected material in the skin layer remains at a density close to the initial solid density, as the hydrodynamic
motion is negligible during the laser pulse. However, the material properties, and particularly the optical properties, rapidly
change fromthe very beginning of the laser pulse. Belowwe consider howthe ultra-short pulse affects the properties of the
material, whichremains at the initial soliddensity during the interactiontime. The absorbedenergy density, the temperature
of the electrons and the lattice, and the material properties are obtained in a form, which allows for comparison with the
experimental data.
2.2. Laser field penetration into a solid
As the electric field of the laser pulse penetrates into a solid, part of the incident energy is absorbed, swiftly changing
the state of the solid and its properties. We treat the problemof lasermatter interaction self-consistently, that is, the effect
of the laser electric field on matter and the inverse effect of matter on the electric field are considered simultaneously.
The field equations coupled to the material equations describe the medium response to the action of the electric field. The
term matter in the context of the problem considered stands for the ensemble of non-relativistic particles, electrons and
nuclei that compose the solid, liquid, gas or plasma, depending on the parameters of the laser beam, namely, the intensity,
wavelength, and pulse duration. Therefore, the coupled time-dependent field and material equations describe all continuous
transformations of the materials.
The electromagnetic field is treated classically, via the Maxwell equations coupled to the material equations:
j =
E;

D =
E (Landau et al., 1984). The electric field, E, current density, j, conductivity, , and induction, D, all are functions of three
space coordinates and time. The material equations are presented in linear in respect to the electric field and in the local
form (spatial dispersion neglected). For description of dielectrics and semi-metals, it is convenient to present the currents
and polarisation separately. The transient dielectric function is defined in general as follows:
=
0
() +i
4
; (2.1)
where = 2c/ is the laser field frequency. It is clear in this formulation that the current emerges due to the free carrier
excitation while the electric induction relates to the part of dielectric function,
0
(), which describes polarisation. For
poor conductors (such as semi-metals with small number of electrons in the conduction zone) at 4/
0
(), a solid
Fig. 2.1. Laser electric field E penetration into the solid in normal skin effect; the absorbed laser energy E
abs
T E
2
.
should be treated as dielectric with dielectric function
0
(). In the opposite case 4/
0
(), the excited free carriers
dominate the dielectric function. Therefore, it takes its conventional form as for good conductors:
= 1 +i
4
. (2.2)
The dielectric function and the conductivity are presented in the formof complex functions, =
+i
; =
+i
. The
optical properties of a solid under the action of powerful laser pulse are time-dependent functions changing in the course
of interaction.
2.3. Laser energy absorption and response of the material
In what follows, we will consider the processes of electric field penetration, energy absorption by electrons, and energy
gain by electrons under assumption that material properties are unchanged during the interaction. First, we remind the
general relations for the field penetration in a solid and derive a general formula for the absorbed energy density, which are
both valid for any absorbing medium without specifying particular absorption mechanisms.
2.3.1. Skin layer approximation
First, we approximate the laser beamat normal incidence by a plane-polarised electromagnetic wave. Let the s-polarised
wave,
E(E
x
, E
y
, E
z
) =

E(0, E
0
e
it
, 0), propagate along the x-axis that coincides withthe normal to the surface fromvacuumat
x < 0, while the matter fills a half-space at x > 0. The non-magnetic medium(magnetic permeability = 1) is characterised
by the local complex dielectric function, , or conductivity, .
For most of considered cases the electron mean free path, l
mfp
, and the distance travelled by electrons during the wave
period,
e
/, both are less than the skin depth l
s
so that the normal skin approximation is valid: l
mfp
l
s
;
e
/ l
s
; here
v
e
is the electron velocity. Then the laser electric field inside the medium is described by the well-known solution for the
normal skin effect (Landau et al., 1984):
E = E
0
exp
_
i
c
_
= E
0
exp
_
i
nx
c

x
l
s
_
; (2.3)
here =
+i
= n
2
k
2
;
= 2nk; and the electrostatic field is neglected (Fig. 2.1). The absorption l
abs
or, the skin
depth l
s
, is inversely proportional to the imaginary part of the refractive index, N
= n +ik:
l
abs
=
c
k
. (2.4)
Usually the absorption length is of the order of several tens of nanometres (16) 10
6
cm that comprises hundreds of
atomic layers.
2.3.2. Absorbed energy density
The absorbed energy density rate can be derived in a general form not specifying the mechanism of absorption. Indeed,
the absorbed laser energy per unit time and per unit volume, Q
abs
, is related to the gradient of the Poynting vector (Landau
et al., 1984) as follows:
Q
abs
= S =
_
c
4
E H
_
. (2.5)
Time averaging Eq. (2.5) over many laser periods and replacing the space derivatives with the time derivative from the
Maxwell equations results in the form:
Q
abs
=

8
|E
a
|
2
; (2.6)
where E
a
denotes the electric field inside the medium averaged over the laser period but it maintains, of course, the time
dependence on the field intensity of the incident laser pulse at t
1
. The spatial dependence of the field and the intensity
inside the solid is determined by (2.3). One can see that formula (2.6) expresses conventional Joule heating of a material by
laser-induced electron current, Q
abs
=
E
2
. Indeed, taking into account the relation of the real part of the conductivity
to the imaginary part of the dielectric function one gets
/4. The value of the evanescent electric field at the

solidvacuum interface, E
a
(0), relates to the amplitude of the incident laser field, E
0
, through the boundary conditions of
continuity for E
y
and H
z
as:
|E
a
(0)|
2
=
4
|1 +
1/2
|
2
|E
0
|
2
. (2.7)
Now (2.6) takes the following form:
Q
abs

1
2
Re Q
abs
=

8
4E
2
0
|1 +
1/2
|
2
. (2.8)
We introduce the incident laser intensity, I
0
= c|E
0
|
2
/8, and express the imaginary part of the dielectric function
through the real and imaginary parts of refractive index. The absorption coefficient A is defined by the Fresnel formula
(Landau et al., 1984):
A = 1 R =
4n
(n +1)
2
+k
2

4n
|1 +
1/2
|
2
. (2.9)
Then the absorbed energy density rate, the energy per unit volume per unit time, Eq. (2.8) reduces to the useful compact
form (Gamaly et al., 2002):
Q
abs
=
2A
l
abs
I(r, z, t) (2.10)
where l
abs
is the skin depth from Eq. (2.4). One could easily show that Eq. (2.10) is valid for the interaction of elliptically
polarised plane wave as well. It is instructive to note that the absorbed energy density rate in Eq. (2.10) is expressed through
the experimentally measured parameters without any ad hoc assumptions. One can also see from Eq. (2.10) that absorbed
energy density decrease e-fold in space inside the target at the half of the absorption length.
It was implicitly assumed in the above derivation that the optical parameters of the medium are space and time
independent and that they are not affected by lasermatter interaction. One should also note that the relations between
the Poynting vector, the intensity and the absorption presented above are rigorously valid only for a plane wave. However,
temporal intensity dependence during the pulse and spatial intensity distribution over the focal spot can be explicitly taken
into account in Eq. (2.10). When experimental results are compared with that of Eq. (2.10), it is apparent that the energy
deposition is predicted with sufficient accuracy.
Integration of Eq. (2.10) over the time immediately gives the energy density (in J/cm
3
) absorbed during the pulse time,
t
p
:
W(t
p
) =
2AF
p
l
abs
; F
p
=
t
p
0
I(t)dt (2.11)
F
p
is the fluence, the energy per unit area delivered to the surface by a laser pulse, and F
a
= AF
p
is the absorbed laser fluence.
Analysis of Eq. (2.11) allows one making several important conclusions. The enthalpy of melting in metals is of the order of
several kJ/cm
3
, while the absorption depth comprises 10
5
to 10
6
cm. Thus, the necessary absorbed laser fluence threshold
for the initiation of metal melting lies in a range of 110 mJ/cm
2
. One could estimate the ablation threshold in a similar
way. The enthalpy of vaporisation is conventionally 1520 times larger than that of fusion. Thus, the ablation threshold for
metals is in a range 0.10.2 J/cm
2
. These estimates are in a qualitative agreement with the experimental figures. Rigorous
thresholds for the transformations of the materials with material properties modified during the pulse are established later
in Sections 3 and 5.
2.3.3. Electron temperature at the end of the pulse
The absorbed laser energy is contained mostly in the electron sub-system at the end of the ultra-short pulse (Fig. 2.2).
Thus, one can calculate maximum electron temperature, T
e
, which is reached at the end of the laser pulse, from the energy
Fig. 2.2. Qualitative time dependence of electron and lattice temperature in the skin layer. The dotted line is a Gaussian shape of fs-laser pulse; t
eph
indicates the energy equilibration time.
conservation law. The electrons excited by laser to the temperature k
B
T
e
F
constitute the degenerated electron gas with
the number density n
e
, the Fermi energy
F
(n
e
), and with the heat capacity, C
e
=
2
k
2
B
T
e
/2
F
. The energy density of such
electron gas reads:
W
e
= C
e
n
e
T
e
= n
e
2
k
2
B
T
2
e
/2
F
. (2.12)
Equalising Eqs. (2.11), (2.12) and solving for the electron temperature one obtains:
k
2
B
T
2
e,m
(t
p
) =
4
F
AF
p
2
n
e
l
abs
. (2.13)
Let us consider, for example, the case when the absorbed energy density in Aluminium (n
e
= 1.8 10
23
cm
3
, the Fermi
energy 11.63 eV) is close to the enthalpy of melting in equilibriumof 1 kJ/cm
3
. That occurs at the absorbed laser fluence of
0.5 mJ/cm
2
delivered by laser with wavelength of 800 nmand at the absorption depth of l
abs
10
6
cm. Then the maximum
energy per electron reaches at the end of the pulse 4.8 10
20
J, which corresponds to the temperature 0.3 eV (3520 K).
The electron temperature is almost 9 times higher than the equilibrium melting temperature of Aluminium.
2.3.4. Lattice temperature after equilibration
The maximumlattice temperature canbe estimatedunder assumptionthat all energy losses are negligible. The maximum
lattice temperature is reached at the moment when the electron-to-lattice energy transfer is completed, T
e
= T
L
= T; it can
be calculated from the energy conservation:
C
e
n
e
T +C
L
n
a
T = W(t
p
). (2.14)
The electron heat capacity is taken as above, while the lattice heat capacity often has the value close to that of the
DulongPetit form, C
L
= 3k
B
as for the unperturbed solid. It is convenient to measure the lattice temperature in units
of the maximum electron temperature of Eq. (2.13). Then (2.14) reduces to the quadratic equation for the ratio of lattice
temperature to the maximum electron temperature T
e,max
:
2
+a 1 = 0; =
T
T
e,max
; a =
2C
L
n
a
2
n
e
k
2
B
T
e,max
. (2.15)
The estimate of the lattice temperature by Eq. (2.15) is legitimate when the losses for ionisation, and modifications of the
dielectric function, and electron and lattice heat capacities are all negligible. The initial lattice temperature is omitted in
Eq. (2.15) for brevity. One can see that in the condition C
L
C
e
and the electron and atomic densities are comparable, the
maximum lattice temperature, T
L,max
, reads:
T
L,max

2AF
C
L
n
a
l
s
. (2.16)
Comparing Eqs. (2.13) and (2.16) one can see that the maximum electron temperature is significantly larger than the
maximum lattice temperature:
T
e,max
T
L,max
C
L
n
a
2
n
e
k
2
B
T
e,max
.
A typical ultra-short laser pulse (100 fs, 100 nJ, wavelength 800 nm) focused onto a spot with a diameter of several microns
can easily deposit either in metal or in dielectric the surface energy density in a range sufficient for melting or ablation.
Using the above formula for estimating the electron temperature for the case when deposited energy density is around the
energy for vaporisation gives the maximum electron temperature in a range of several electron volts, which is sufficient
for ionisation of the material. The rise of the electron temperature to electron volts level during hundred of femtosecond
time definitely affects the material state and all its properties, thus indicating that the approximation of constant optical
properties fails to be true. The excitation of electrons from valence to conduction band in dielectrics, ionisation processes,
the temperature dependence of the collision rates of electrons and ions should be taken into account. Thus, the next step
is to introduce the model where the material properties are electron/lattice temperature and number density dependent
during the laser pulse. This is considered in the following section.
2.4. Modification of the material during the laser pulse
2.4.1. The Drude-like dielectric function
The laser excitation changes the number of conductivity electrons, the electron and lattice temperature, the electron
effective collision rates and thus affects the optical properties of a metal. The free electron can only scatter a photon but
cannot absorb the energy because in this case it is impossible to comply with the conservation laws. Thus, three body
interactions involving a photon, an electron and an atom (or ion) are responsible for the absorption of laser energy. In
the absence of the external field, this spontaneous process is well known as bremsstrahlung emission. In the presence of
external field the two processes occur: the induced emission and inverse process of photon absorption. The net result is the
photon absorption process called inverse bremsstrahlung or collisional absorption. All collisions of electrons with electrons,
ions and neutrals should be taken into account because the electronelectron collisions only do not change the total electron
momentum and the total electron current.
The free electron oscillates in the laser field,

E = E
0
e
it
with velocity v
osc
= e
E/m. The energy of oscillations averaged

over the many wave periods presents a convenient characteristic of the applied field strength:
osc
=
mv
2
osc
2
=
e
2
E
2
2m
2
. (2.17)
It is useful to present the oscillationenergy as functionof the incident intensity andthe laser wavelengthinthe form(Gamaly
et al., 2002):
osc
=
e
2
E
2
(1 +
2
)
2m
2
= 9.35 10
14
(1 +
2
)I
2
(2.18)
where = 1 stands for the circular, = 0 for the linear polarisation, m
1
is the effective electron mass introduced to
take into account the interaction of electrons with the crystal lattice, with phonons, and interaction with other electrons;
I is in (W/cm
2
) and in (m). For example, electrons quivering energy in the field of linear polarised 100 fs laser beam
at wavelength of 1 m at absorbed fluence of 10 mJ/cm
2
(average intensity 10
11
W/cm
2
) equals to 0.0094 eV, while at
fluence of J/cm
2
(at intensity 10
13
W/cm
2
) electron oscillates with the energy of 0.94 eV. In what follows, we consider
only plane-polarised light. It is known that the ionisation rate for the circularly polarised wave in the case of ionisation
by many photons is several orders of magnitude lower then that for the plane polarisation (Muller et al., 1992). The
selection rule for the angular momentum is very strict. The absorption of many photons creates big centrifugal barrier,
which decreases the overlap between the ground state and excited state resulting in small ionisation cross-section. It was
experimentally established (Temnov et al., 2006) that multi-photon absorption cross section for the powerful femtosecond
circular-polarised beam is much lower than that for the beam of the same intensity and pulse duration but being plane-
polarised.
The multiple electron collisions with the lattice (phonons) or with ions convert the oscillation energy into the electrons
kinetic energy. All collisions with the loss of momentumare accounted for through the effective collision frequency, v
eff
. The
conductivity and dielectric function of the free electron gas (or electron plasma) where ions are considered as a neutralising
background are presented in the Drude-like form. The conductivity of such plasma reads:
=
e
2
n
e
m
e
(
eff
i)
. (2.19)
The relation between the dielectric function and the conductivity stems fromthe Maxwell equations in the formof Eq. (2.2).
After the insertion (2.19) into (2.2) the dielectric function for the electronic plasma then attains well-known form of the
Drude-like function (Landau et al., 1984):
= 1 +i

2
pe
(
eff
i)
= 1

2
pe
(
2
eff
+
2
)
+i

2
pe
eff
(
2
eff
+
2
)
. (2.20)
Correction to the free electron gas model introduced through the electron effective mass. The electron effective mass and
electron number density enter into the dielectric function of (2.20) through the electron plasma frequency,
pe
:
2
pe
=
4e
2
n
e
m
e
. (2.21)
1
The optical electron effective mass is usually not known, it is often assumed to be equal to a free electron mass, m

= m
e
.
The collision frequency in general is a function of the electron and phonon (ion) temperature and electrons number
density, which will be discussed later in this section. Thus the optical properties of excited metal are time-dependent. This
purely classical derivation is rigorously valid at
h
e
. However, it appears to be approximately valid in many cases when
the quantum mechanical approach should apply at

h
e
.
2.4.2. Electronic excitation and ionisation
The heating rate of electrons by laser field defines the excitation and ionisation rates. The heating rate expresses
through the imaginary part of the dielectric function. The dielectric function in turn depends on laser frequency, effective
electronlattice (ion) collision rate, the electron number density and effective mass, which are changing in the course of
interaction. Indeed the imaginary part of dielectric function follows from Eq. (2.20):
=

2
pe
eff
(
2
eff
+
2
)
. (2.22)
In order to make clear the physical meaning of the heating process of electrons in laser field let us express the heating rate
from Eq. (2.10) through the imaginary part of the dielectric function and electrons oscillation energy as follows:
Q
abs
=

8
|E
a
|
2
= 2n
e
osc
eff
(
2
eff
+
2
)
. (2.23)
Then the average heating rate for a single electron in a medium assuming that number density of electrons is time-
independent, reads (Raizer, 1978):
d
dt
= 2
osc
eff
(
2
eff
+
2
)
. (2.24)
It follows from Eq. (2.24) that the energy gain rate of electron in high frequency laser field strongly depends on relation
between laser frequency and effective electronphonon (ion) collision rate. We show later in the section that collision rate
first grows up at the temperature that exceeds the Debye temperature but is well below the Fermi level, k
B
T
D
<
e

F
.
In this case >
eff
, and thus the electron-heating rate of Eq. (2.24) depends on the collision rate as the following, d
e
/
dt 2
osc
eff
. At this temperature level the collision rate is proportional to the phonon frequency of the order of THz.
Therefore the electron excitation time is of the order of
1
eff

1
ph
100 fs. The collision rate reaches its maximum at
high electron temperature k
B
T
e

F
when a solid converted into plasma near the ablation threshold for most of materials.
The maximum of collision rate is close to the electron plasma frequency,
eff

pe
. These conditions correspond to a
solid, which was heated by 100 fs laser with intensity around 10
13
W/cm
2
, and converted to plasma. The electrons heating
rate now expresses as d
e
/dt 2
osc
2
/
eff
. Typical time for the electrons heating at high temperature is around tens of
femtoseconds. Hence, the material properties may change during the pulse time. At even higher intensity >10
13
W/cm
2
the temperature increases further, k
B
T
e

F
I
0
(I
0
is the ionisation potential), and the solid is converted into a totally
ionised plasma. In plasma, the frequency of the electronion Coulomb collisions decreases with the increase in the electron
temperature. One can see from Eq. (2.24) that electron accelerated in applied laser field can gain the energy of several
electron volts depending on laser intensity and relation between laser and collision frequency.
Let us consider two major processes responsible for absorption in metals and dielectrics and therefore relevant to the
dielectric function modification during the absorption process. The intra-band transitions comprise the electrons excitation
and heating in metals. The inter-band transitions and electron excitation from valence to conduction band are important in
defining the dielectric function and absorption in dielectrics. These processes include single and multi-photon absorption,
avalanche acceleration of electrons in valence band to the energy exceeding band gap following by ionisation and conversion
of dielectric to the metal-like state.
Usually direct photon absorption by electrons in a valence band is negligibly small. A few (seed) electrons always exist
in the valence band. These electrons oscillate in the electromagnetic field of the laser and can gain net energy by collisions.
Electrons accelerated to energy in excess of the band-gap can collide with electrons in the valence band and transfers
sufficient energy to them for excitation into the conduction band. The process of successive accelerations and collisions
generates an avalanche of ionisation events (Yablonovitch and Bloembergen, 1972; Holway, 1972; Fradin et al., 1973;
Holway and Fradin, 1975; Raizer, 1978). The probability of such event per unit time can be estimated with the help of
single electron energy gain rate Eq. (2.24) as follows:
w
imp

1
gap
d
e
dt
. (2.25)
Let us define time for a single ionisation event as t
ion
= w
1
imp
. Then after k = t/t
ion
= w
imp
t events the seed number of
electrons, n
0
, increases in 2
k
times. The number density of electrons generated by such an avalanche process reads:
n
e
= n
0
2
w
imp
t
= n
0
exp[ln 2(w
imp
t)]. (2.26)
Another important ionisation process that contributes into the total ionisation rate is the multi-photon ionisation. We
consider both processes in more details in further sections where high intensity interaction is described. At this stage it is
important to note that ionisation time by both closely intertwined the avalanche and the multi-photon processes is shorter
than the pulse time. Therefore, material properties are modified during the pulse, affecting the interaction process.
One can see from Eq. (2.24) that in the condition >
eff
the real part of the dielectric function is zero when the
number density of charge carriers is such that the electron plasma frequency equals to the laser frequency, =
pe
. The so-
called critical electron density follows fromthis equality: n
c
= m
e
2
/4e
2
. The electromagnetic wave becomes evanescent
when the number density of free carriers is larger the critical value as happens in metals. It is generally accepted that
optical breakdown occurs when the number density of electrons reaches the critical density. The critical density for lasers in
wavelength domain for visible light of 4001000 nm varies in a range of (1.27.5) 10
21
cm
3
. The laser parameters (the
intensity, wavelength, and pulse duration) and material parameters (the band-gap width and the electronphonon effective
rate) at the breakdown threshold are combined together by the condition n
e
= n
c
. Some of the material parameters are
presented in Appendix.
2.4.3. Modified dielectric function and heat capacity at high temperature
The dielectric properties of metals and dielectrics change dramatically at high intensity near the ablation threshold. The
ionisation of dielectrics becomes significant to the extent that any dielectric converts to metal-like solid when laser intensity
approaches the ablation threshold. The number of electrons in the conduction band comes close to the atomic number
density and the crystalline band structure effectively collapses (Gruzdev, 2004, 2005). The electronion collision rate in
these conditions has a maximum value, which is close to the electron plasma frequency charge; it significantly exceeds the
laser frequency,
eff

pe
. The dielectric function (2.20) and the refractive index then transform to the follows:
2
pe
1;
= 2nk

pe
; n k =
_
pe
2
_
1/2
. (2.27)
The ratio of absorption coefficient defined by the Fresnel formula to the skin depth, A/l
s
, is a complete combination of
material parameters that determines the absorbed energy density. Dielectric function Eq. (2.28) describes well the solid
density plasma near the ablation threshold. The absorbed energy density for this case depends only on the laser wavelength
through A/l
s
4/ in accord with experiments (Gamaly et al., 2002). It was shown (Gamaly et al., 2005) that absorption
of the laser light with = 532 nm in the hot Aluminium plasma increases two times during the interaction time in
comparison to that in a cold metal.
Dielectrics under the action high intensity laser experience radical transformation from non-absorbing or weakly ab-
sorbing mediuminto highly absorbing plasma with metal-like properties (Kautek and Krausz, 1998). The number density of
electrons in the conduction band of dielectrics, which experienced the optical breakdown, exceeds the critical density; the
collision rate modifies accordingly. Therefore, the optical properties of the ionised dielectrics are also calculated through
relations of Eq. (2.27).
We show later in the paper that the phase state of a solid changes at much slower pace in comparison to the changes in
the optical and material properties. The conductivity electrons in metals are degenerate. At the electron temperature much
lower than the corresponding Fermi energy, k
B
T
e

F
, the electron heat capacity is derived by expansion of the Fermi
distribution into series in respect to the small parameter, k
B
T
e
/
F
(Landau et al., 1984). The electron heat capacity of the
degenerated electron gas reads:
C
e

2
k
2
B
T
e
/2
F
. (2.28)
The high temperature limit is also well known. At k
B
T
e

F
the electron gas can be treated as an ideal gas with the heat
capacity of 3k
B
/2. The electron distribution function and correspondingly the electronic specific heat both approach to
those of the ideal gas at k
B
T
e

b
. Therefore it is reasonable suggesting that the electron heat capacity weakly depends
on temperature near the ablation threshold and it is close to the ideal gas value, 3k
B
/2. Note that the swift increase in the
electron number density with the same rate as the temperature increase may lead to the transient decrease in the electron
heat capacity due to
F
n
2/3
e
.
The atomic motion has an oscillatory character at low temperature. Accordingly, the atomic specific heat C
L
is equal to
3k
B
per atom. At the gradually increasing temperature, the vibrational motion of atoms changes to a translation mode, as in
a mono-atomic gas. Correspondingly, the atom specific heat gradually decreases to 3k
B
/2 per atom. The effective boundary
dividing the temperature ranges, where the two limiting values of the atomic specific heat are valid, can be associated with a
potential barrier against the free motion of atoms through the solid. The temperature T
b
at the potential barrier is naturally
related to the binding energy: k
B
T
b
2
b
/3 (Zeldovich and Raizer, 2002). Thus, the increase in lattice temperature, as
the ablation threshold approaches, is accompanied by an increase in the specific heat of electrons and decrease in the heat
capacity of atoms. Both electron and ion heat capacities are converging to the same value as for ideal gas. The qualitative
dependence of electron and ion heat capacity is presented in Fig. 2.3.
2.4.4. Electron and ion temperature in the ionised solid
The maximal temperature of electrons expresses as (compare to Eq. (2.13)):
T
e,max
=
4AF
p
3k
B
n
e
l
abs
. (2.29)
Fig. 2.3. The dependence of the electron and lattice heat capacity on temperature; the C
e,L
are in units of k
B
, while T
e,L
are in units of binding energy
b
.
Similarly, the energy conservation, C
e
n
e
T +C
L
n
a
T = W(t
p
), gives the maximal ion temperature (all losses are neglected):
T
e,max
=
4AF
p
3k
B
(n
e
+n
a
)l
abs
. (2.30)
One shall note that the difference between the maximum electron and ion temperature in solid density plasma, T
e
/T
L
=
(n
e
+n
a
)/n
e
, is less thanthat ina neutral solidexcitedto lower temperature. However eveninmetals having one conductivity
electron per atom the maximum electron temperature is two times higher than that for ions.
2.5. Relaxation processes in laser-excited solids
The primary process of electron excitation by the ultra-short laser pulse occurs in totally non-equilibrium conditions.
The following processes take place in succession before a solid attains the full equilibrium state after the end of the laser
pulse. First, electronelectron and phononphonon collisions lead to establishment of local quasi-equilibrium distributions
in both sub-systems. The electrons establish the conventional FermiDirac distribution with T
e
< T
F
. However, the atomic
distribution at the end of the lasermatter interaction has the form of the incomplete Maxwell distribution. The high-
energy tail is absent because the creation of particles with the energy larger than the average one needs much more
collisions and therefore the longer time to be filled in. The main part of the Maxwell distribution that comprises
80%90% of the atoms establishes fast. The atomic energy averaged over this distribution corresponds to the lattice
temperature. The time for building up the high-energy tail appears to be the longest of all relaxation times. Electronphonon
momentumexchange rate plays animportant role during the interactionbeing responsible for light absorptionandtransport
processes. Electronphonon energy exchange leads to equilibration of electron and lattice temperatures before the complete
Maxwell distribution is established (Corkum et al., 1988). The electrons diffusion (the energy transport) depends on the
electronphonon momentum exchange rate, the electron velocity, and on the electrons heat capacity. Therefore the heat
conduction coefficient and characteristic cooling time both are the functions of electron and lattice temperature. Under the
action of powerful femtosecond laser, the matter undergoes swift transformation from initial cold solid state to the plasma
state. The time-dependent ionisation and recombination processes take place. When the absorbed energy per atomexceeds
the binding energy, an atom becomes free (ablated) and, having the excess of kinetic energy, leaves the surface of a laser-
affected solid. The characteristic expansion time (hydrodynamic time) defines the material removal rate.
It is instructive expressing the relaxation rates through the fundamental constants making a clear relation to the first
principles (Ilinskii and Keldysh, 1994). These constants are the following: the electron charge, e = 4.8 10
10
CGSE, the
speed of light in vacuum, c = 3 10
10
cm/s, the Planck constant,

h = 1.055 10
27
erg s; the free electron mass,
m
e
= 9.11 10
28
g; the masses of nuclei M, the fundamental unit of length, the Bohr radius, a
B
=

h
2
/m
e
e
2
, the unit of
energy is Rydberg: Ry = e
4
m
e
/2
h
2
= 13.6 eV, the unit of frequency is the atomic frequency
at
= I
0
/
h = 2.06 10
16
s
1
;
the unit of electronvelocity is v
e
= e
2
/
h = 2.1810
8
cm/s. We use mainly the CGSEunits because they are most appropriate
for description of interaction of electromagnetic field with matter. Whenever it is possible, we present scaling of relaxation
times through the fundamental units.
We begin with fast relaxation processes of electronelectron and phononphonon interaction leading to establishment
of the main parts of local distributions in both sub-systems.
2.5.1. Establishing the main parts of the distribution functions
(a) Electronelectron collisions
The collision rate v
ee
, or probability for collision per unit time, in the kinetic theory conventionally expresses through
the number density of colliding particles, n
e
, cross section for interaction,
ee
, and relative velocity of colliding particles,
ee
,
as
ee
= n
e
ee
v
e
. Let us first estimate the electronelectron collision rate taking the cross section as that for the Coulomb
collisions,
ee
e
4
/
2
e
, n
e
a
3
B
, and v
e
e
2
/
h. Then collision rate reads:
ee

e
4
a
3
B
m
2
e
v
3
e

at
. (2.31)
The estimate fromthe first principles gives the electronelectron collision rate in the order of
ee
10
16
s
1
that is compa-
rable to the electron plasma frequency for good metals,
ee

pe
. Correspondingly the mean free path travelled by electron
between collisions reads:
l
mfp
= v
e
/
ee
. (2.32)
Thus, the mean-free-path in fully degenerated electron gas comprises several angstroms.
The mean free path of the electrons slightly excited over the Fermi level in a strongly correlated degenerate electron
gas has been calculated by Quinn and Ferrell (1958) and Pines (1964) for low, in comparison to those for metals, electronic
densities. The rate of the single-particle electron scattering, taking into account screening effects (exchange neglected), can
be expressed as:
ee
=
v
F
l
ee
=

pe
1.88
F
; (2.33)
where v
F
is the Fermi velocity. This result holds for the electron energy being well belowthe Fermi level. The mean free path
of electron increases in 1.5 times in comparison with the results of Quinn and Ferrell (1958), when the exchange effect
along with the screening is taken into account (Penn, 1980).
The experimental measurements of the mean free path for 5 eV electrons (v
e
= 1.32 10
8
cm/s) in Al (Kanter, 1970)
give l
ee
= 50 to be compared with l
ee
= 62 given by Quinn and Ferrell (1958). That gives the electron collision time of
t
ee
4 fs.
The mean free path of electrons in Aluminium has been studied by examination of ultrasonic pulses generated by 0.2 ps,
632 nm, 0.5 nJ, 76 MHz laser (Tas and Maris, 1994). The authors estimated the temperature of the heated electron gas in
Aluminium in a range of 4001000 K (0.040.1 eV above the Fermi level). They estimated the electron thermalisation time
around 100 fs in qualitative agreement with the theory (Quinn and Ferrell, 1958; Pines, 1964) that predicts 12 fs (taking for
Al
pe
= 1.97 10
16
s
1
;
F
= 11.63 eV).
The electron-energy distribution in 300 thick gold film was measured with 700 fs time resolved photoemission
spectroscopy following laser heating by 400 fs, 674 nm laser pulse (Fann et al., 1992). They estimated lifetime of electrons
heated by laser to 0.4 eV above the Fermi level (v
e
= 0.37 10
8
cm/s) equal to the time resolution of 700 fs giving the
mean free path of l
ee
2600 . This result is by the order of magnitude larger that following fromthe theoretical predictions,
however time resolution is insufficient (for Gold
F
= 5.51 eV;
pe
= 1.37 10
16
s
1
).
It was observed by Bronson et al. (1987) that the fast electrons from the high-energy tail in Gold (with velocity of
v
e
2 10
8
cm/s, twice of the Fermi energy) propagate on the distance 100200 nm. This gives the collision time of
50100 fs in a qualitative agreement with the theory of Quinn and Ferrell (1958) and Pines (1964). It should be noted that
the energy range of these electrons is beyond the range of the validity of the above-mentioned theories.
As a final note to this section we shall point out the absence of comprehensive theoretical and experimental knowledge
of the exact electronelectron collision rate dependence on the electron energy in low energy range,
e

F
. However,
it looks well proved by the experiments and theory that electronelectron collision rate grows in direct proportion to the
increase of the electron energy and plasma frequency when it approaches to the Fermi energy,
F
, from the lower energy
range:
ee

pe
F
. (2.34)
It qualitatively complies with the estimate for the maximum collision rate when approaching to the Fermi energy from the
high-energy range as we show later. In the following section, we use Eq. (2.34) for calculation the time for establishing the
local energy distribution, the FermiDirac distribution, t
ee

1
ee
at the electron energy below and approaching the Fermi
energy.
(b) Phononphonon interactions
It is legitimate considering the harmonic atomic vibrations, as quasi-particles, phonons. The interaction of phonons with
phonons and with particles of different nature (photons, electrons) results in scattering, absorption and emission of phonons
with energy and quasi-momentum conserved (Frohlich, 1949).
A phonon gas can be also treated in the framework of the kinetic theory of gases. Therefore, such kinetic concepts as a
mean free path and effective collision frequency (the inverse of it is a phonon relaxation time) can be applied. Let us estimate
the phononphonon collision rate
phph
in accord to the kinetic theory as the follows:
phph
n
ph
phph
v
ph
. (2.35)
The number density of phonons at the temperature around the Debye temperature estimates as the atomic number
density n
a
n
ph
, and the characteristic phonons velocity equals to the speed of sound v
ph
v
sound
. The former expresses
through the phonons frequency,
ph
, and the inter-atomic distance, d, as v
sound

ph
d. The phononphonon scattering
cross section reasonably estimates as squared inter-atomic distance,
phph
d
2
, and atomic number density scales as
n
a
d
3
. Thus, the phononphonon collision rate is proportional to the phonons frequency:
phph

ph
. (2.36)
One could estimate the phononphonon collision rate through the fundamental constants. The atomic density estimates as
n
a
a
3
B
, the scattering cross section is
phph
a
2
B
, and the speed of sound expresses through the characteristic electron
velocity, v
e
e
2
/
h, with the help of the adiabaticity principle, v

sond

_
m
e
M
_
1/2
v
e
(Ashcroft and Mermin, 1976). Then one
arrives to: v
phph
(m
e
/M)
1/2
at

ph
, the same result as in Eq. (2.36).
The above estimates implicitly assume that the solid is at the temperature around the Debye temperature, T
D
. The
maximum phonon energy in all branches of the spectrum is of the order of the Debye energy,

h
D
. At the temperature
above the Debye temperature, T
L
> T
D
, the phonon spectrum broadens and the Debye frequency is no longer the upper
phonon frequency limit. The phonon energy averaged over the distribution function at T
L
> T
D
reads:
h
ph
3k
B
T
L
/2
(Kittel, 1996). Then the phononphonon collision rate becomes directly proportional to the lattice temperature:
phph

D
T
L
T
D
=
3k
B
T
L
2
h
. (2.37)
Thus the phonon sub-system attains the main part of equilibrium distribution function during the time,
phph

2
h/3k
B
T
L
. The Debye frequency for the majority of solids is around 10 THz 10
13
s
1
. Thus it takes around 100 fs for
establishing the main part of the Maxwell distribution at the Debye temperature. For example, for Bi the Debye temperature
is 119 K,
D
= 1.56 10
13
s
1
and equilibration time comprises 64 fs. However, at the room temperature the main part
of distribution in Bi establishes in less than 39 fs. In a solid excited by the 100 fs laser above the room temperature the
equilibration in the phonon subsystem establishes at the beginning of the laser pulse.
2.5.2. Electronphonon interaction
The optical response of any medium on the action of applied electric field relates to the relaxation of the electrons
perturbations induced by the field. The absorption occurs when electrons oscillating in the applied high frequency electric
field are scattered mainly by the lattice vibrations (phonons). The scattering results in the change of electron momentum
thus affecting the electric current inthe medium. The measure of efficiency of this scattering is a magnitude of the probability
of scattering per unit time or effective collision frequency that enters into the conductivity, heat diffusion coefficient,
dielectric function and into absorption coefficient as we introduced above.
The scattering of electron by lattice vibrations on quantum mechanical language can be considered as a process
of emission or absorption of a phonon. The multi-particle processes lead to conversion of the part of the oscillation
energy into the self-energy of electron and afterwards through electronphonon energy exchange into the lattice thermal
energy.
The electronphonon interaction (EPI) affects the lasersolid coupling process in several ways. Firstly, the EPI changes
the dielectric function thus affecting the absorption of the laser light. Secondly, the EPI leads to the energy transfer from
the electron sub system to the lattice, e.g. heating a lattice, and finally to the equilibration of the electron and lattice
temperature. Thirdly, the EPI affects the heat capacity of the electrons via the change in the electron effective mass. The
electron heat conduction is affected through the changes in the heat capacity and effective collision frequency. After the
laser pulse termination the energy transfer by means of the electron heat conduction dominates the processes in the target.
If the structural phase transition occurs due to laser action then the atoms in the lattice are shifted from the original
positions into the new ones corresponding to the other phase-state. The atomic vibrations during the transition period
change fromthe harmonic ones corresponding to phonons to strongly anharmonic oscillations following by instabilities and
then by the structural re-arrangement to a new phase. Below we account briefly for all this processes and present simple
formulae available for the following estimates. First, we remind the general properties of lattice vibrations (phonons) treated
as quasi-particles. Then, we introduce the characteristic time scales (the inverse collision rates): the electronphonon
time for the momentum exchange (the optical or transport relaxation time), and the electronphonon time for the energy
exchange (the energy relaxation time).
(a) Electronphonon momentum exchange
The electrons are travelling in a periodic self-consistent potential W(r) of an ideal crystalline lattice without losses.
Electrons scattering occurs on lattice imperfections: impurities, defects and on perturbations in periodicity caused by the
atomic vibrations (phonons). The electronphononscattering is a mainsource of changes inelectronmomentumandenergy.
Let us first consider a simple qualitative picture of electronphonon momentumexchange interaction. We assume here two-
particle interaction with the total quasi-momentum conserved.
Electromagnetic and kinetic approach: Let us first consider this interaction as a charge-and-field interaction. The atom
oscillating with amplitude
ph
creates a dipole moment, e
ph
, and corresponding electric field, e
ph
/d
3
, which interacts
with electron. The energy of such interaction is E
eph
e
2
2
ph
/d
3
, and corresponding frequency equal to:
eph
E
eph
/
h e
2
2
ph
/
hd
3
. (2.38)
One can easily see that the above estimate complies with the kinetic approach for the electronphonon scattering rate in the
form,
eph
n
e
eph
v
e
, taking the cross section for the electronphonon scattering as
eph

2
ph
, the electron number
density as proportional to the atomic number density, n
e
d
3
, and the electron velocity as for the degenerated Fermi
gas, v
e
e
2
/
h. Formula Eq. (2.38) is an important relation, which establishes a link between the atomic vibrations and the
optical properties of solid, which we later use for the analysis of the experiments.
Electronphonon exchange rate from the solution of the transport Boltzmann equation: Solution of the Boltzmann equation
for electron transport with the electronphonon interaction taken into account gives the relaxation time (the inverse of
electronphonon momentum exchange rate) in the following form (Grimvall, 1981):
1
mom
eph
= 4

max
0
h
k
B
T
2
tr
F()d
[1 exp(
h/k
B
T)][exp(
h/k
B
T) 1]
;
where
2
tr
F() is the Eliashberg electronphonon coupling function. This function is a product of an effective
electronphonon coupling
2
involving phonons of energy

h, and the phonon density of states F(). At high temperature
the k
B
T

h relaxation time reads (Grimvall, 1981):
1
mom
eph
4k
B
T

max
0
2
tr
F()d
.
Taking the BlochGruneisen formula for the Eliashberg function
2
tr
F()
4
/
4
max
(Grimvall, 1981) one obtains the
electronphonon momentum relaxation time:
1
mom
eph
k
B
T
h
. (2.39)
Thus, the electronphonon momentumexchange rate at T
L
> T
D
estimated fromdifferent approaches is proportional to the
phonons frequency and, when averaged over the phonons spectrum, it is proportional to the lattice temperature.
Measurements of the temperature dependence of the optical parameters of Bi in equilibrium confirmed that the
electronphonon momentum exchange rate changes in linear proportion to the lattice temperature up to and above the
melting point (Comins, 1972; Garl, 2008). At room temperature this rate equals to 2 10
15
s
1
giving the electronphonon
momentum relaxation time in Bi less than femtosecond.
It was established theoretically and verified in experiments, that the rate of the electronphonon momentum exchange
averaged over the all phonon modes grows in proportion to the lattice temperature (Grimvall, 1981; Ilinskii and Keldysh,
1994; Kittel, 1996). The time of the electronphonon momentum exchange at T
L
= T
D
,
mom
eph
(
mom
eph
)
1
is of the order
of 10100 fs, it decreases with growth of the lattice temperature. The above scaling gives a reasonable semi-quantitative
estimate for the momentum relaxation time. However, the exact values can be extracted from the temperature dependent
optical properties of a particular solid.
(b) Electronphonon energy exchange and relaxation to equilibrium
The electronelectron and phononphonon collision times are shorter in comparison to the energy transfer time due
to the big difference in electron and atomic masses. Therefore, the FermiDirac distribution function for electrons with
the electron temperature, T
e
, and the BoseEinstein distribution for phonons with the lattice temperature, T
L
, T
L
T
e
,
are established in both sub-systems well before the temperature equilibration. Both temperatures are time dependent and
adiabatically follow the changes in the whole system. In order to consider solely the energy exchange and equilibration
between two sub-systems the other effects, such as charge separation field, laser electric field, spatial inhomogeneities, etc.,
on the electron motion are neglected. Under these assumptions, the problemof the energy transfer fromthe electrons to the
lattice is reduced to the kinetic equation for electrons where the only electronphonon collisions are retained in the collision
integral. The collision integral is obtained in an explicit form because the phonons and electrons distribution functions are
known. Under these assumptions, the problem has been first formulated and solved by Kaganov et al. (1957). Later on,
Allen (1987) re-formulated the problem in the modern notations, and derived the energy exchange rate in a compact and
explicit form using approach from the theory of superconductivity rigorously valid at the temperature range below the
Debye temperature. The exchange rate in Allens derivation does not depend on the lattice temperature because he uses the
Eliashberg phonon spectral function, discussed in the previous section, which is valid at T
L
< T
D
. In what follows the collision
integral in Eqs. (2.40)(2.43) is simplified along the path exploited by Landau (1937), Kaganov et al. (1957) and Allen (1987).
However, we replace the electronphonon scattering probability via the electronphonon momentum exchange rate at
T
L
> T
D
in a way similar to that done by Landau for the electronion collisions. In this case, the phonon spectral function is
the MaxwellBoltzmann distribution. Averaging over this distribution naturally results in the rate dependence on the lattice
temperature. The kinetic equation for the electron distribution function that takes into account solely the electronphonon
interaction takes a form:
f ()
t
= I
eph
(); (2.40)
where I
eph
() is the electronphonon collision integral. The kinetic equation (2.40) converts to the electron energy equation
by multiplication both sides by electron energy and integrating over the all electron states around the Fermi level under
assumption k
B
T
e

e
:
E
e
t
=

F
0
I
eph
()
1/2
d
2
3/2
F
. (2.41)
The energy per electron in the degenerate electron gas at temperature k
B
T
e

e
reads (Landau and Lifshitz, 1980b):
E
e
=

F
0
f
e
()
1/2
d
2
3/2
F
= E
0
+
1
2
C
e
(T
e
)T
e
; (2.42)
where E
0
= 0.3
F
is the energy at zero temperature and the second term is the total energy of electrons above the Fermi
level with the heat capacity C
e

2
k
2
B
T
e
/2
F
. Let us present now the explicit formula for the collision integral.
The electronphonon collision integral that describes an emission of phonon with quasi-momentum k (and energy
h) by electron with quasi-momentum p (energy ) and transition to the electron state p
, has familiar the

BlochPeierlsBoltzmann form(Frohlich, 1949; Kaganov et al., 1957; Ziman, 1960, 1964; Allen, 1987; Lifshitz and Pitaevski,
1981):
I
eph
() =
w(p, k; p
)
d
3
k
(2)
3
{f (
)[1 f ()]N
f ()[1 f (
)](1 +N
)}
p
2
dp
(2
h)
3
; (2.43)
where w(p, k; p
) is the probability of the process per unit time. The electrons distribution, f (), is taken in the FermiDirac
form and phonon distribution, N
, is that of the BoseEinstein with the different temperatures, T

e
, T
L
:
f () = {exp( )/k
B
T
e
+1}
1
N
= {exp
h/k
B
T
L
1}
1
.
(2.44)
We express following the Landaus recipe the probability of the scattering through the electronphonon momentum
exchange rate,
mom
eph
, and therefore through the cross section of the collision process and the electron velocity (Lifshitz
and Pitaevski, 1981):
w(p, k; p
)
d
3
k
(2)
3
= |v
e
|d
eph
=
d
mom
eph
n
e
mom
eph
= n
e
|v
e
|
eph
.
The collisionintegral is expandedinto the series inrespect of small ratios of k
B
T
e
/
F
and
h/k
B
T
L
, (k
B
T
e

F
andk
B
T
L

h)
as it was done by Kaganov et al. (1957) and Allen (1987). The collision integral of Eq. (2.43) takes the final form:
I
eph
() =

en
eph
f ()
T
e
(T
e
T
L
). (2.45)
By inserting Eq. (2.45) into the energy equation (2.41) and integrating over the electron initial states, the energy equation is
reduced to its familiar form (Kaganov et al., 1957; Allen, 1987):
C
e
(T
e
)T
e
t
=
en
eph
(T
L
)C
e
(T
e
)(T
e
T
L
). (2.46)
However, the major difference from the previous studies (Allen, 1987; Anisimov et al., 1974) is that the electronphonon
energy exchange rate is presented explicitly, averaged over the phonon distribution function, N
en
eph

1
en
eph
=
3(
h)
2
2
F
k
B
T
L
N
d
mom
eph
. (2.47)
As it is shown above, the momentum exchange rate is proportional to the phonons frequency:
mom
eph
C
. (2.48)
Here the dimensionless coefficient C
is the combination of the material parameters (see Section 2.5.2d). The momentum
exchange rate of Eq. (2.48) averaged over the equilibrium phonons distribution is proportional to the lattice temperature
(Kittel, 1996) at T
L
> T
D
:
mom
eph
= C
_
N
d
_
N
d
=
3C
2
k
B
T
L
h
. (2.49)
Now, the electronphonon energy exchange rate is calculated by inserting Eq. (2.48) into Eq. (2.47) and integrating over the
whole phonon spectrum to obtain:
en
eph
=
3C
(k
B
T
L
)
2
2
F
h

0
x
2
dx
e
x
1

3C
(k
B
T
L
)
2
F
h
. (2.50)
Thus, energy exchange rate governing approach the electronphonon system to equilibrium is strong function of lattice
temperature. The link with the momentum rate follows from the comparison of Eqs. (2.49) and (2.50):
en
eph

2k
B
T
L
mom
eph
. (2.51)
The removal of the angular brackets in Eq. (2.51) made for simplicity and could not be confusing because both functions are
spectrum-averaged.
It is worth noting that the longest time needed for the electronlattice equilibration can be estimated if the temperature
dependence of the collision rates is ignored. Indeed, the simple scaling stems from Eq. (2.51) through the upper frequency
in phonon spectrum,
D
:
en
eph

h
2
D
/
F
.
It follows fromthe adiabaticity principle that
D
= (m
e
/M)
1/2
at
, and
h
at

F
. Then, the energy exchange rate expresses
through the fundamental constants only (Ilinskii and Keldysh, 1994):
en
eph
m
e
at
/M =
m
e
M
_
m
e
e
4
2
h
3
_
.
Taking m
e
/M 10
5
as for metals and
at
2 10
16
s
1
one gets
en
eph
2 10
11
s
1
and equilibration time of 5 ps.
(c) Dependence of the electronphonon interaction rates on the electron temperature
This dependence is weak in conditions, k
B
T
e

F
that holds when the electronphonon interaction is important.
Therefore conventionally electron energy and velocity are taken as those at the Fermi level. Let us introduce explicitly the
term depending on T
e
into collision rates. The rates are proportional to the electronic density of states calculated to the
upper energy limit,
lim
,
mom
eph

en
eph

lim
= (T
e
). The maximumenergy in the degenerated electrons spectrumat zero
temperature conventionally introduced as the chemical potential at T
e
= 0,
F
= (T
e
= 0) (Landau and Lifshitz, 1980b).
Temperature correction for the chemical potential gives the correction to the electronic density of states and therefore to
the collision rates. This correction is of the same nature as that made for all thermodynamic functions at finite electron
temperature (electron energy and heat capacity). Calculating the density of electronic states and introducing it into the
collision rate (keeping the first term in expansion into series in respect to k
B
T
e
/
F
one obtains:
mom
eph
(T
e
)
mom
eph
(0)
_
1 +

2
k
2
B
T
2
e
12
2
F
_
. (2.52)
One can see that before the ionisation becomes significant (k
B
T
e
< 0.1
F
) the correction on the finite electron temperature
is negligibly small.
(d) Dependence of the electronphonon collision rate on electron density
In previous studies of the collision rate, it was implicitly assumed that the electron number density in the conduction
band is constant. However, at strong electronic excitation approaching the ablation threshold the electron number density
increases swiftly. Let us introduce explicitly the dependence of the metal parameters on the electron number density.
We estimate the electronphonon collision rate from the kinetic viewpoint,
mom
eph
n
e
eph
v
e
. The cross-section for the
electronphonon collision estimates as
eph

2
ph
, where
ph
is phonons amplitude, and electron velocity is taken
as the Fermi velocity, v
e
v
F
= 1.24
hn
1/3
e
/m
e
, because, k
B
T
e

F
. The squared amplitude expresses through the
phonons energy
2
ph
2
h/M
ph
, where M is the atomic mass while the phonons frequency relates to the electron
plasma frequency,
2
pe
= 4e
2
n
e
/m
e
, via the familiar relation,
2
ph
(m
e
/M)
2
pe
(Ashcroft and Mermin, 1976). Now, the
electronphonon momentumexchange rate reduces to the formof Eq. (2.48). The dimensionless proportionality coefficient
equals to C
= 0.62 a
B
n
1/3
e
; where a
B
=

h
2
/m
e
e
2
is the Bohr radius. Thus, the electronphonon momentum exchange
rate depends on the electron number density as:
mom
eph
a
B
n
1/3
e

ph
. (2.53)
Accordingly the energy exchange rate of Eq. (2.51) reads:
en
eph

h
2
2
ph
/e
2
n
1/3
e
. (2.54)
The dependence of the electronic heat capacity upon the electron density is straightforward:
C
e

2
k
B
T
e
2
F
n
2/3
e
. (2.55)
To the best of our knowledge, there are no experimental results published on the dependence of the collision rates on the
electron density Eqs. (2.53)(2.55).
(e) Electrons effective mass
It is known that for describing the heat capacity of ordinary metals by the free electrons model one needs introducing
the electron effective mass (Kittel, 1996). In description of the electronphonon interaction at low temperature the
renormalisation of the electron mass is always introduced (Allen, 1987; Maksimov et al., 1997; Grimvall, 1981). Allen
presented the energy exchange rate in the following form (Allen, 1987):
en

1
t
eph
=
3
h
2
F
2
.
The function (, T) is a characteristic renormalisation of the electron mass in the conditions of electronphonon
interaction; the angular brackets denote averaging by the Eliashberg electronphonon interaction function (see
Section 2.5.2.b above and also Maksimov et al., 1997; Grimvall, 1981).
Fig. 2.4. Temperature dependences of the electronphonon momentum relaxation time at elevated lattice temperature T
L
relative to that time at room
temperature of T
room
= 293 K, and the same for the energy relaxation times for beryllium (Be), aluminium (Al), bismuth (Bi), copper (Cu), and gold (Au).
There is no experimental or theoretical evidence on the necessity of electron mass renormalisation at T
L
> T
D
to the
best of our knowledge. Therefore the mass of electron in the review is generally taken as a free electron value. The
electronphonon spectral function is mainly relevant to the low temperature case. As pointed out by Grimvall, the electron
mass renormalisation and averaging by the Eliashberg function is unnecessary at T
L
> T
D
(see Section 2.5.2.b above and
Grimvall, 1981). Qualitatively, the value of the phonon frequency averaged over the phonon spectral function used by Allen
is of the order of magnitude of the Debye frequency (Allen, 1987). However the strong temperature dependence is missing
in the Allens approach.
(f) Extracting momentum exchange rates in metals from experimentally determined optical properties
The optical properties of metals as functions of the light frequency at room temperature are well documented (Palik,
1998). The electronphonon momentum exchange rate directly relates to the real,
re
, and imaginary,
im
, parts of the
dielectric permittivity if the former obeys the Drude-like form,
mom
eph
/
l
=
im
/(1
re
);
l
is the laser frequency. Therefore
the formulae of Eqs. (2.50) and (2.51) allow one to express the energy exchange time through the single experimentally
determined parameter, momentumrelaxation time at roomtemperature, t
mom
eph
(T
room
), in whole temperature range without
ad hoc assumptions. One can present these times in the dimensionless form applicable to any metal characterised by its
Fermi energy:
t
mom
eph
(T
L
)
t
mom
eph
(T
room
)
=
T
room
T
L
, and
t
en
eph
(T
L
)
t
en
eph
(T
room
)
=
_

F
k
B
T
room
__
T
room
T
L
_
2
.
These time ratios are plotted for Be (
F
= 14.3 eV), Al (
F
= 11.7 eV), Bi (
F
= 9.9 eV), Cu (
F
= 7.7 eV), and Au
(
F
= 5.53 eV), as an example, in Fig. 2.4.
The optical properties of Bismuth at 800 nm were measured at equilibrium from room temperature, 293 K, (Landolt-
Brnstein, 1983) up to 773 K well over the melting point of 544.7 K (Comins, 1972; Garl, 2008; Hodgson, 1962; Smith et al.,
1964). Under suggestion that electron has a free electron mass value (Comins, 1972) momentum transfer rate for Bi with
good accuracy follows the linear dependence on the temperature (Gamaly and Rode, 2009). The momentum transfer rate
extracted from equilibrium experiments at room temperature comprises
mom
eph
(T
room
) = 2 10
15
s
1
(Comins, 1972; Garl,
2008). The corresponding energy exchange rate equals to 1.1710
13
s
1
giving much shorter relaxation times than it would
be estimated without considering the temperature dependence.
2.5.3. Importance of the high-energy tail in the Maxwell distribution
It is known that atoms from the high-energy tail of the Maxwell distribution play crucial role in such equilibrium
processes as melting (Fecht, 1992) and evaporation (Landau and Lifshitz, 1980a). The proper concentration of thermal point
defects (7%10%) is crucial for the initiation of the equilibrium melting at the melting point, T
melt
. The energy for formation
of such defect is approximately 10 T
melt
. Therefore, in order to create such defects at the melting point the atoms with
higher than the average energy should be generated. Similarly, in order to evaporate an atom in equilibrium, i.e. to remove
it from a solid, the energy in excess of the binding energy
b
should be supplied to the atom. Conventionally, the binding
energy per atom is 1020 times larger than the vaporisation energy. Thus, melting and vaporisation at the melting point
and at the vaporisation temperature could proceed only if the high-energy tail in the atomic distribution is well established.
In order to describe these processes properly in the ultra-short pulse excited solid one needs to knowthe time-dependent
history of the setting up the full Maxwell distribution including its high-energy tail. In the collision between particles of the
same mass, the whole particle energy can be transferred in a single collision. Therefore, it is conventionally assumed that the
time for the establishment of the average energy per particle, the temperature, is the time for setting up the main part of the
Maxwell distribution involving 80%90% of atoms. This time is in inverse proportion to the electronphonon momentum
exchange rate, t
main
(
mom
eph
)
1
. However, while the main part follows the Maxwell distribution, it does not have the
high-energy tail.
The particle energy averaged over the mainpart of the distributionequals to the particle temperature. However, a particle
needs many successive collisions, and therefore longer time, to gain the energy above the average value, i.e. to build up
the high-energy tail of the distribution. This process can be considered as diffusion along the energy axis (Uhlenbeck and
Ornstein, 1930; Lifshitz and Pitaevski, 1981). The diffusion coefficient, D, conventionally expresses through the energy gain
rate, d/dt:
D

d
dt
. (2.56)
This general relation is valid for the collisions between particles of different nature in the different force fields. Time for
reaching the energy, , than reads:
t
=
2
/D

_
d ln
dt
_
1
. (2.57)
If the particle gains energy solely by collisions, d/dt

coll
, then time to reach the energy equals to inverse of the
collision rate at the energy :
t

1
coll
(). (2.58)
It was found (MacDonald et al., 1957) that the time needed for establishing the high-energy tail of the electron distribution
in plasma is much longer than that necessary for establishing a main part of distribution with temperature T
e
. The cross-
section for the Coulomb collision is a function of electron temperature, thus time for establishing the temperature, main
part of distribution, is t
main
T
3/2
e
. The time for electron to reach the energy
e
k
B
T
e
by the multiple Coulomb collisions
is sufficiently longer (MacDonald et al., 1957):
t
tail
t
main
_

e
k
B
T
e
_
3/2
.
One can apply the similar consideration to atomic collisions in laser-excited solid.
Let us consider the establishing of the distribution in a skin-layer of a solid rapidly overheated by the ultra-short laser
pulse at the deposited energy density below the ablation threshold. The excited solid remains intact. The atomic collisions
are responsible for establishing the Maxwells distribution in a neutral solid. The cross-section of atomic collisions in the
considered temperature range is practically independent on the temperature. It can be approximated by the cross-section
for collision of hard spheres. The collision rate for particles in the tail at
e
k
B
T then has a familiar form:
aa
= n()
aa
v(); (2.59)
where n() n
k
B
T
=
_
0
f (
)d
is the number of particles in the high-energy tail of the Maxwell distribution at k

B
T.
The ratio of the number of particles in a tail to the total number, n
a
, calculates in the following way:
n
k
B
T
n
a
= 1 erf{(/k
B
T)
1/2
} +
2
(/k
B
T)
1/2
exp(/k
B
T) (2.60)
where erf(x) is an error function: erf(x) = (2/
)
_
x
0
exp(u
2
)du. At = k
B
T the ratio Eq. (2.60) constitutes 0.645. In
application to melting and ablation problems the energy of atoms in the tail locates in a range = 10k
B
T. Then Eq. (2.60)
can be simplified (because 1 erf(2) 0 at > 2k
B
T) to the form:
n
2k
B
T
n
a
1.13(/k
B
T)
1/2
exp(/k
B
T). (2.61)
On the other hand the collision rate responsible for the setting up the main part of the distribution expresses as
main
=
n
a
(T)
aa
v(T). Now the ratio of two collision rates reads:
t
main
t
tail
=

tail
(T)
main
(T)
=
n
a
()
n
a
(T)
(/k
B
T)
1/2
.
Inserting Eq. (2.61) inthe above formula one obtains the time for the tail formationtime linkedto the time of the temperature
setting up in the explicit form:
t
tail
t
main
0.85(k
B
T/) exp(/k
B
T). (2.62)
It can be easily seen (Fig. 2.5), that building up the tail at the energy 10 k
B
T takes 10
3
times longer than that for
establishing the main part of the equilibrium distribution (temperature). This is the longest of the relaxation times; it
Fig. 2.5. Building up the high-energy tail in Maxwell distribution of energy with time: t
1
< t
2
< < t
5
.
comprises from tens to hundred picoseconds. It is necessary to note that in the processes of melting and ablation the role of
the tail in phase transformation is not only in reaching a particular energy threshold. It is necessary to fill up the tail to the
extent when the number of energetic atoms constitutes several percents of the total number of atoms. For example, in order
to initiate melting process the number of thermal point defects should exceed several percent. Therefore, time is required
to fill up the high-energy tail to create the necessary number of particles with the energy exceeding the threshold. This time
is much longer due to the exponential dependence on the energy. We will discuss this problem in detail in the Section 4
on ultra-fast melting. We discussed above the high-energy tail formation in the bulk of an excited material. However, the
lasermatter interaction starts at the surface where the highest temperature.
It appears that formation of the full Maxwells distribution in the surface layer differs from the process in the bulk. The
atoms in the outermost surface layer next to the vacuum are in fact in a different condition compared to the atoms in the
bulk. It is well known that the surface atoms are loosely bound to the bulk with the part of bonds dangling or saturated
with foreign atoms (Zangwill, 1988; Prutton, 1994). The effects of different bonding at the surface leads to decreases in
the Debye and melting temperatures; to changes in the bond length and inter-atomic distance as well as the crystalline
structure and nature and rate of any phase transition. Prutton noticed that: . . . many surface phases are actually metastable,
i.e. the surface is not in a true thermodynamic equilibrium (Prutton, 1994). The energy distribution in the outermost surface
layer is the important characteristic affecting the removal of atoms from the surface layer at the ablation threshold. This
energy distribution is responsible for the relative contribution of non-equilibrium ablation and thermal evaporation. We
will consider this problem in relation to the ablation thresholds in ambient gas and in vacuum in Section 5.
2.6. Electronion collisions in solid-density plasma
In the lasersolid interaction at intensity 10
13
W/cm
2
the energy of electron oscillations in a laser field comprises
several electron volts and ionisation becomes significant. The threshold intensity, the breakdown threshold, is defined as the
laser intensity level at which the density of electrons, pulled by the laser field fromthe valence band to the conduction band,
equals to the critical density determined by the laser frequency. Typically, the optical breakdown threshold for dielectrics
(the threshold for conversion a solid into plasma) lies in a range 10
12
10
13
W/cm
2
. The value of the breakdown threshold
depends strongly on the atomic composition of a solid, laser wavelength and intensity. Conversion to plasma significantly
changes optical properties of a laser-affected medium through the changes in the collision rates, and therefore alters the
lasermatter interaction mode. Collisions in plasma were thoroughly studied during the past decades. Below we briefly
summarise the results necessary for the studies of ultra-short lasermatter interaction at the high laser intensities.
2.6.1. Electron-to-ion momentum exchange rate
(a) Non-ideal plasma (
e

F
)
A solid transforms into the so-called non-ideal plasma when the contribution of the Coulomb interactions into the
electron energy is comparable to the thermal energy. The measure of the plasmas non-ideality is the ratio of the Coulomb
energy to the thermal energy of electron expressed as follows:
nl
=
Z
2
e
2
a
e
; a = (4n
e
/3)
1/3
. (2.63)
This ratio is large,
nl
1, in strongly correlated plasma where the electron energy is low compared to the Fermi energy.
The number of electrons in the Debye sphere is also a convenient physical parameter allowing one to distinguish between
the coupled and the collisionless (ideal) plasma, where the thermal energy exceeds the Coulomb energy:
N
D
= 4
3
D
/3 = 1.7 10
9
_
3
e
n
e
_
1/2
; (2.64)
here
e
is in (eV), and the Debye radius
D
, the charge screening distance, is conventionally defined as a ratio of the electron
velocity to the electron plasma frequency,
D
= v
e
/
pe
. One can see that both parameters are interconnected, N
D

3/2
nl
.
Thus, in a strongly coupled plasma the number of particles in the Debye sphere is small, N
D
1,
nl
1, while in the ideal
plasma it should be large, N
D
1,
nl
< 1. The relation between the collision rate, the plasma frequency and the coupling
parameter in two-component plasma (electrons and ions) was established by (Cauble and Rozmus, 1995):
mom
ei
pe
ln
10N
D
. (2.65)
Here, ln is the Coulomb logarithm, measure of contributions from far collisions (Kruer, 1988). In typical conditions of
femtosecond lasermatter interaction near the breakdown (ablation) threshold (
e
10 eV; n
e
10
23
cm
3
; ln 2)
one obtains N
D
= 0.17 and
mom
ei
/
pe
1.
The transport coefficients in strongly coupled solid density plasma have been calculated on the basis of the solutions
of coupled electron and ion kinetic equations by the moment expansion method (Cauble and Rozmus, 1995). Transient
electronion collision frequency has been found in non-equilibrium conditions as function of the different electron and
ion temperatures. In solid density plasma created by the femtosecond laser pulse plasma frequency is usually much larger
than the laser frequency
pe
. For example, in solid density aluminium irradiated by 800 nm light the ratio of these
frequencies comprises,
pe
/ 10. For coupling parameter
nl
1.5, in a solid-state aluminiumplasma (
e
10 eV; n
e

10
23
cm
3
; ln 2) the ratio predicted to be
mom
ei
/
pe
0.25 0.3 (Cauble and Rozmus, 1995). This demonstrates
that coupling decreases the collision rate in comparison to the ideal plasma in about 34 times. The significant difference
in electron and ion temperature (more correctly in the average energy,
e

i
) is typical for ultra-short lasermatter
interaction when electrons are heated to the level of several eV during the pulse while ions remain cold. The difference
between the electron and ion energy results in the increased screening that in turn decreases the collision rate. The decrease
in the collision rate in strongly coupled plasma may affect two important processes: the inverse bremsstrahlung absorption
process, which strongly depends on the collision rate, and the processes of establishing the equilibrium energy distribution
that in turn affects the laser absorption.
(b) Ideal plasma (
e

F
)
A solid becomes fully ionised at high temperature k
B
T
e
=
F
. The electron temperature dominates the electronion
interaction. The electronelectron, electronion, ionion interactions obey the model for ideal solid-state density plasma
when the thermal energy dominates over the energy of the Coulomb interactions. The transport cross section for scattering
of charged particle with mass m
1
and charge eZ
1
in the Coulomb field of the particle with mass m
2
and charge eZ
2
on small
angles expresses by the Rutherford formula (Lifshitz and Pitaevski, 1981):
12
=
4e
4
Z
2
1
Z
2
2
ln
2
(v
1
v
2
)
4
; (2.66)
where = m
1
m
2
/(m
1
+ m
2
) is the reduced mass. This approach is valid for the length scale much larger than the Debye
screening distance, l
D
= v
e
/
pe
, and for the time intervals longer than inverse plasma frequency, t
1
pe
. Obviously,
these conditions are fulfilled in short pulse lasersolid interaction. One can nowobtain the explicit formulae for all relevant
collision rates. Electronelectron collision rates reads:
ee
= n
e
ee
v
e

n
e
e
4
ln
ee
m
2
e
v
3
e
. (2.67)
The important difference from the low temperature case considered earlier is that the electron temperature dominates the
collision rate, which decreases with increase of temperature. It is opposite to the low temperature case where the collision
rate increases in direct proportion to the lattice temperature and is almost independent of the electron temperature. The
electronion momentum exchange rate now reads:
mom
ei
= n
i
ei
v
e

n
i
e
4
Z
2
ln
ei
m
2
e
v
3
e
. (2.68)
The above rate with the rigorously calculated numerical coefficient according to Kruer (1988) is:
mom
ei
3 10
6
ln
n
e
Z
2
T
3/2
eV
. (2.69)
For example, the electronion collision frequency in Copper at the electron temperature in the energy units coinciding with
the Fermi energy (n
e
= 0.845 10
23
cm
3
,
pe
= 1.64 10
16
s
1
, k
B
T 7.7 eV, ln 2) is
ei
= 2.38 10
16
s
1
,
that is almost two times larger than plasma frequency. However, it was demonstrated by Cauble and Rozmus (1995) that
corrections for the non-ideality effect in such plasma reduce the collision rate in more than two times. Therefore, suggestion
by Eidmann et al. (2000), that the maximum of the collision frequency for Aluminium is close to
pe
, qualitatively complies
with the theory.
The ionion collision rate, which is a characteristic for the establishment of the equilibrium distribution in the ion sub-
system, can be estimated as follows:
ii
n
i
ii
v
i

n
i
e
4
Z
2
ln
ii
M
2
i
v
2
i
. (2.70)
The ratio of the ionion collision rate to the electronion momentumexchange rate, assuming that the Coulomb logarithms
are the same for both cases, reads:
ii
ei
=
_
m
e
M
_
1/2
_
i
_
3/2
. (2.71)
Taking Copper as example (M
i
= 63.54 a.u.), when electrons are heated to the energy close to Fermi energy
e

F
, the
electronion collision frequency is
mom
ei
0.5
pe
10
16
s
1
. For the ionion collision rate one obtains from Eq. (2.70):
ii
3 10
13
s
1
. The time required for establishing an equilibrium energy distribution at
e

F
in the ion sub-system is
of the order of several tens of femtoseconds, while the local equilibrium in the electron sub-system is established in a tenth
of a femtosecond.
2.6.2. Electron-to-ion energy transfer rate
The electron and ion sub-systems form the Maxwell energy distribution faster than the electron-to-ion energy transfer
time. Therefore the kinetic equations for electrons and ions are reduced to the coupled equations for electron and ion
temperature where the electron-to-ion energy exchange termcan be explicitly calculated through the electronion collision
integral, while the electronelectron and the ionion collisions are insignificant. Landau has found the electronion collision
integral in 1937 in the approximation of a small change of momentum in a single collision that suffices the dominance of
the far collisions in the Coulomb interactions (Lifshitz and Pitaevski, 1981). The calculation of the Landau collision integral
with the Maxwell distributions for electrons and ions with different temperatures results in the explicit formfor the energy
exchange rate in coupled equations for electrons and ions temperatures, similar to that for the electronphonon energy
exchange. The equation for the electron energy density when the electronion energy transfer dominates and other losses
are neglected reads:
E
e
t
=
3n
e
2
T
e
t
=
en
ei
n
e
(T
e
T
i
). (2.72)
Accordingly, the electron energy sink in the Eq. (2.72) serves as the energy source for the ion energy density. The energy
transfer rate from electrons to ions with charge eZ has the following form (Landau, 1937; Lifshitz and Pitaevski, 1981):
en
ei
= 12(2)
1/2
n
i
Z
2
e
4
m
1/2
e
T
3/2
e
M
i
ln
ei
. (2.73)
The energetic electrons collision with much heavier ion results in a complete momentumtransfer while the energy transfer
constitutes only small part of the electron energy, m
e
e
/M
i
. Hence, the ratio of the electronion momentum transfer rate to
the energy exchange rate in a simple plasma equals to the ratio of ion to electron mass:
mom
ei
en
ei
=
M
m
e
. (2.74)
Therefore, the energy transfer time by the Coulomb collisions, in the conditions when the electronion momentumtransfer
is at its maximum
mom
ei

pe
, resides in a range of several picoseconds. Note that these conditions are close to the
optical breakdown and the ablation threshold. For example, in Copper (M
i
= 63.54 a.u.) at
e

F
, the electronion
momentum transfer rate is
mom
ei
0.5
pe
10
16
s
1
, while the energy exchange rate in accordance to Eq. (2.73) equals
to 8.6 10
10
s
1
. Thus, electron-to-ion energy transfer occurs in 11.6 ps.
The fact that electrons and ions energy distribution functions have Maxwell distributions with separate time-dependent
temperatures and with well-defined energy exchange rate is the basis of two-temperature (two-fluid) description (Landau,
1937) of the laser-created plasma. It is similar to the two-temperature approach for electronlattice energy exchange while
the nature of the temperature dependence of the exchange rate is different (Landau, 1937; Kaganov et al., 1957; Anisimov
et al., 1974).
2.7. General picture for the electron-to-lattice (ion) exchange rate: from solid to plasma
The general formula for the dependence of the electron collision rate on the electron energy ranging from a small
excitation over the Fermi level (lowtemperature) up to the hot plasma state does not exist yet, to the best of our knowledge.
However, the processes of electronphonon interaction are well understood qualitatively and quantitatively in a lowenergy
range k
B
T
e

e
as well as in the high-temperature ideal plasma state at k
B
T
e
>
e
. Moreover, the qualitative understanding
also exists for transition from solid to ideal plasma through the non-ideal plasma state.
Fig. 2.6. Temperature dependence of the electronphonon collision rate at low temperature (left solid line) and electronion rate at high temperature
(right solid line) for Aluminium. Solid lines are the known functions for electronphonon interaction from this paper and electronion collision rate from
Kruer (1988). The dashed line is conjectured qualitative behaviour during solid-to-plasma transition.
As it follows from Eq. (2.50), the electronphonon momentum and the energy exchange rates coincide at k
B
T
e
=
F
/2
when a metal is near the ablation threshold, and solid is converted into non-ideal plasma. The approximation used for
description of electronphonon interaction fails at these conditions. However, in a non-ideal plasma the electronelectron
collision rate increases in direct proportion with the increase in the electron energy,
ee

pe
e
/
F
, when approaching the
Fermi level from the low energy range, qualitatively supporting the approach to its maximum value. The growth rate slows
down with increase of temperature when the solid becomes partly ionised and the Coulomb collisions become significant.
The dependence of the collision rates on temperature becomes weaker due to the effects in non-ideal plasma discussed
above. When the energy per atom approaches the ablation limit at k
B
T
L

b
, the electronphonon collision rate becomes
close to the atomic frequency,
mom
ei

b
/
h
at

pe
apparently approaching the maximum value. In plasma, where
the Coulomb collisions dominate, the electronion momentum exchange rate decreases with the increase of temperature,
mom
ei
n
e
T
3/2
e
(
F
/T
e
)
3/2
see Eqs. (2.73) and (2.74). Thus, the existence of a maximum around the electron plasma
frequency seems qualitatively obvious. Hence, Eidmann et al. (2000) suggested that the collision rate reaches its maximum
value close to the electron plasma frequency,
mom
ei

pe
10
16
s
1
. However the exact value of the maximum is still
unresolved, to the best of our knowledge. The electronion collision rates in solid density Al plasma can be presented in
the following form:
mom
ei
(T) 1.4 10
16
(
F
/T
e
)
3/2
s
1
(Kruer, 1988). The collision rate dependence on temperature in
the whole range from electronphonon interaction to electronion interaction in plasma is presented in Fig. 2.6 for
Aluminium.
2.8. Electronic heat conduction
The dominant heat conduction mechanismin the ultra-short-laser heated solid is the electronic transport. The electronic
heat flux reads:
Q
eth
= k
e
T
e
. (2.75)
The electron heat conduction coefficient, k
e
, expresses through the electrons mean free path, l
e
, number density, n
e
, heat
capacity and electron velocity (Lifshitz and Pitaevski, 1981):
k
e
= C
e
(T
e
)n
e
D
diff
_
W
cm K
_
. (2.76)
The diffusion coefficient reads:
D
diff
=
l
e
v
e
3
=
v
2
e
3
mom
eph
_
cm
2
s
_
. (2.77)
The electrons mean free path directly relates to the electron momentumexchange rate through the familiar kinetic relation,
l
e
= v
e
/
mom
eph
. The heat diffusion behaves differently at low excitation level
F
> k
B
T
e
k
B
T
L
> k
B
T
D
and near the ablation
threshold, where solid is converted to plasma (Kanavin et al., 1998). The difference is due to the different dependence of the
momentum exchange rate and heat capacity on temperature.
Let us consider, for example, the heat conduction in metal at low excitation, T
e

F
. The electron velocity is close
to the Fermi velocity, v
2
F
= 2
F
/m
e
, heat capacity is as that for the degenerated electron gas, C
e

2
k
2
B
T
e
/2
F
, and the
electronphonon collision rate expresses as,
mom
eph
= 3C
k
B
T
L
/2
h (T
L
> T
D
), C
being the numerical coefficient that must

be extracted from experiments. Then the heat conduction coefficient at low excitation is the following:
e
=
2
2
hk
B
n
e
9C
e
T
e
T
L
. (2.78)
The heat conduction coefficient is temperature-independent after the electronlattice temperature equilibration (Blatt,
1968). The characteristic time for cooling the absorption layer swiftly heated by laser follows directly from the heat
conduction equation:
t
cool
l
2
abs
/D
diff
. (2.79)
The electronphonon momentumexchange rate defines the optical properties through the dielectric function. The same
rate dominates the heat diffusion process. Therefore the momentum exchange rate can be in principle recovered from the
optical measurements of real and imaginary parts of the dielectric function, if it is described by the Drude model, or from
the measurements of static conductivity. However, the momentum exchange rate measured by these two methods differs
significantly. We demonstrate that taking Aluminium as an example. The results of measurement of static conductivity in
equilibrium allow one to recover the momentum collision rate (in fact, some effective collision frequency responsible for
resistance) and therefore the heat conduction coefficient of metals (see Kittel, 1996, p. 168). The static conductivity reads:
stat
=
2
pe
/4
eff
= e
2
n
e
/m
eff
.
The static conductivity of Al (m
e
= 1.48m
e
; n
e
= 1.86 10
23
cm
3
) comprises 3.28 10
15
s
1
(3.63 10
5
cm) at
295 K. One gets for Al (
eff
)
stat
= 9.710
15
s
1
that complies with the heat conduction measurements and the electron heat
diffusion coefficient of D
diff
= 0.978 cm
2
/s. It follows that time for cooling the aluminium skin layer of 3 10
6
cm occurs
in 10 ps well after the end of the ultra-short laser pulse. The dielectric function of Aluminium at room temperature as
function of light frequency presented in Palik (1998). The electronphonon collision rate can be recovered from the optical
measurements through the real and imaginary parts of the dielectric function in the Drude form,
mom
eph
/
l
=
im
/(1
re
);
where
l
is the light frequency. The momentum rate recovered from the optical measurements in the wavelength range
from500 to 900 nmcomprises (
eff
)
opt
= (12) 10
15
s
1
giving the relaxation times 510 times larger. It should be noted
that ambiguity with the value of the electron mass remains unresolved (see Section 2.5.2f and Figs. 2.4 and 2.6).
The heat conductionindiffusionapproximationis a validapproachwhenthe electron-mean-free pathis muchlower than
the scale length of the heated zone, l
mpf
= v
e
/
eff
< l
abs
. The electron mean-free-path in Al resides in a range of 1.313.0 ,
taking into account the uncertainty in the momentum rate for Al,
eff
= 10
16
10
15
s
1
. Thus, the diffusion approximation
for description the heat conduction in metals in a skin layer with thickness in a range 100300 is valid even with this
uncertainty. The diffusion approximation is a legitimate description of the transport processes at high and at low excitation
levels.
2.9. Atomic motion in a solid induced by the ultra-short pulse action
2.9.1. Atomic response on the action of external high-frequency electric field
Bounded and conductivity electrons start oscillating with the frequency of the imposed laser field immediately after
field penetration into a solid. The time for the electrons response to the external perturbation equals to the inverse of
the electron plasma frequency, t
response

1
pe
10
16
s. The electrons can accumulate part of the oscillation energy in
collisions with the lattice during the time of a few femtoseconds that is comparable to the several periods of the laser light.
Indeed, the frequency of the laser light in visible region is in a range of (25 10
15
s
1
). Conventionally, the fast electronic
motion during the time
1
pe
can be neglected and the laser field oscillations are also averaged over many field periods. This
is legitimate approximation until the laser pulse duration complies with condition t
p
/2 1. This condition holds for
the laser pulses down to a few femtoseconds.
The main effect of excited electrons on atomic cores occurs through the action of the electrostatic field of charge
separation well before the energy transfer by the electronphonon collisions becomes important. Indeed, the effect of the
electrostatic field of excited electron on the parent ion is almost immediate after the electrons started moving and gaining
energy fromthe applied field. It is well known that the electrons oscillations and atomic vibrations in metals are intrinsically
coupled via the electrostatic field of charge separation. This is the reason why the frequency of atomic vibrations (phonons)
in metals,
ph
, is directly related to the electron plasma frequency,
ph
(m
e
/M)
1/2
pe
(Ashcroft and Mermin, 1976). The
force exerted by the excited electrons on the lattice during the laser pulse can be considered as a rapid blow because the
pulse duration is much shorter than the phonon period t
ph

1
ph
t
p
. Under the action of this force atoms experience fast
coherent displacement, which in turn affects the harmonic atomic vibrations with the cold phonon frequency. We consider
these delicate atomic motions in detail in the next Section 3.
The harmonic vibrations start during the period shorter than electronphonon energy transfer time. The atomic motion
looses its harmonic character and becomes chaotic when the lattice temperature gradually increases to the level when phase
transformation (melting) becomes possible. The phononphonon interactions enter into a non-linear, multi-phonon regime.
Finally, atoms are shifted into new positions corresponding to the melt or other crystalline phase. The duration of a period,
when harmonic motion is discernible, sharply decreases when the deposited energy density and intensity during the pulse
are increasing. At the energy density around and over the ablation threshold the atoms gain the energy of the order of the
binding energy in a solid, and with the kinetic energy obtained by an atomafter the bonds breaking the macroscopic motion
and expansion (hydrodynamics motion) becomes significant. We consider ablation in detail in Section 5.
2.9.2. Single electron and ion motion
Let us consider in detail the primary action of excited electrons on the atomic cores. It is instructive to consider first the
free electronoscillations inthe absence of external fieldandwiththe presence of neutralising ionbackground. The oscillation
of electrons in an unperturbed metal is the intrinsic property of any quasi-neutral system of charges (the Earnshaws
theorem). The Newton equation describes the oscillatory motion of electron with the restoring electro-static field between
electrons and ions E
elst
4e(n
e
n
i
)x
e
included. The electron plasma frequency, the electrons oscillation frequency, is of
the order of characteristic electronic frequency in atom. The electrostatic fields magnitude compares to that of the atomic
field E
elst
E
at
= e/a
2
B
= m
2
e
e
5
/
h
4
. The same electro-static field affects the atomic core motion. In the accounting for
electrons effect on the atomic motion one can neglect the electrons inertia because an ion mass is much larger than that of
electron, M m
e
(adiabatic principle). Then, it follows fromthe Newton equation for atomic motion that the atomvibrates
in the electrostatic field with the phonons frequency,
ph
(m
e
/M)
1/2
pe
(Ashcroft and Mermin, 1976).
Now let us consider the case when electrons are swiftly and non-homogeneously excited to temperature T
e
while atoms
remain cold. The electron temperature rises during the laser pulse; the gradient of the electron temperature in the skin
layer follows the spatial intensity distribution with characteristic scale length of l
s
/2. Thus, the new force, the gradient
of the electronic pressure, starts affecting the electrons motion, along with the increase of electrons temperature during
the laser pulse action. In the equation of a single electron motion this force enters as the electron temperature gradient.
Therefore, this force modifies the electrostatic field to the form, eE
elst
m
e
2
pe
x
i
T
e
. Hence, the equation of atomic
motion also modifies as follows:
M
i
dv
i
dt
m
e
2
pe
x
i
+T
e
. (2.80)
Therefore, excited electrons affect atomic motion through the changes in the force imposed by the electrostatic field. The
build up of the electron temperature gradient is completed to the end of the pulse. Hence, the electrons start affecting the
motion of cold atoms well before the energy transfer by the electronphonon collisions becomes significant. The atoms in a
lattice, still remaining at the initial temperature, can be swiftly displaced by the action of the electrostatic field.
2.9.3. Macroscopic forces acting on a solid in the external laser field
Local equilibrium in the electron and lattice sub-systems establishes early in the pulse time. Therefore, one can describe
the medium response to the laser excitation by the stress tensor in the equilibrium form. Forces acting on a solid placed in
the high frequency electric field can be expressed through the stress tensor,
ik
, which includes both the internal energy of
a medium and the energy of the field inside the medium. For an initially isotropic medium the stress tensor reads (Landau
et al., 1984):
ik
= P
ik
E
2

ik
8
+
E
2
8
_
n
a
_
jk
n
a
_
T
_
ij
+
E
i
D
j
4
;
where the electric field displacement vector has a form D
k
=
kj
E
j
. We assume that the dielectric tensor modified by the
laser effect consists of two terms, the Drude-like term and the polarisation term. The force per unit volume reads:
f
i
=

ik
x
k
. (2.81)
The laser field produces atomic displacements (deformations), which modify initially homogeneous dielectric function,
(0)
ik
,
ik
=
0
ik
+ ()
ik
. The medium is non-magnetic, and all magnetic field effects are neglected. The force per unit
volume expresses through the stress tensor,
ik
, in the form (Landau et al., 1984):
f
i

P
x
i
()
il
x
k
E
l
E
k
8
+
()
il
8
x
k
E
l
E
k
. (2.82)
The first term represents the thermal force, where P = P
e
+ P
Lattice
is the pressure inside the interaction zone,
which includes electron and lattice contributions. The second term accounts for the changes in the dielectric function. The
polarisation force and the ponderomotive force in the third termare effective only during the pulse, while the thermal force
of the pressure gradient works long after the pulse end during the cooling stage.
It is instructive to compare the magnitude of all laser-exerted forces to the end of the pulse under the assumption that
the intensity is constant during the pulse time (flat-top-hat pulse shape). The space scale of the intensity gradient is of the
order of magnitude of the skin-depth. It is convenient comparing the forces acting per single atom, F
a
= f /n
a
. We also take
for simplicity that n
a
= n
e
. Then the scalings for all forces are:
F
th
T
e
/l
abs
F
pol
I
abs
/n
e
cl
pol
F
pond
I
abs
/n
e
cl
abs
.
(2.83)
Consider, for example, the case of Aluminium excited by 800 nm, 100 fs pulse at fluence of 5 mJ/cm
2
and intensity
5 10
10
W/cm
2
(n
e
= 1.8 10
23
cm
3
, skin depth l
abs
= 4.8 10
6
cm). Then comparison of the above forces shows that
the thermal force dominates:
F
th
= F
pol
F
pond
.
The thermal electronic force acts as a swift weak blowon the atomic chains inducing the atomic vibrations with a frequency
of cold phonons. The contribution of the thermal force increases with laser intensity and finally the thermal force dominates.
2.9.4. Heating and motion of atoms in laser-affected solid
We consider lasermatter interaction in the broad range of the absorbed energy density, starting from the very weak
excitations and slight changes in material properties up to the high intensity when the solid converts to plasma early in the
interaction time. There are several levels for description of the mediumresponse to the action of the external electric field: a
quantummechanical or a classical (as opposed to quantum) approach. The rigorous treatment of electron excitation, energy
absorption and ionisation are usually performed by quantum mechanical methods, which we are not considering here. We
describe mainly the macroscopic response that reveals itself in the enhanced atomic vibrations, in phase transformation
(such as melting), in conversion to plasma and expansion. Therefore the kinetic and hydrodynamic approaches are most
appropriate. Nevertheless we consider the processes in the volumes of several cubic nanometres containing thousands of
atoms where the macroscopic properties are drastically different from those in bulk. The collective effects and effects of
inter-atomic potential on the atomic response can be taken into account by force terms or exchange terms in equation for
electrons andatoms (ions). Classical approachproduces results withsufficient accuracy for the interpretationof experiments
and for reasonable predictions for many practical applications.
(a) Truncated moment equations
Let us consider a solid as a system containing different sets (sub-systems) of particles, electrons and atoms (phonons,
ions). The external and internal energy sources, fields and forces, the energy and momentum exchange between
subsystemsall are included in frame of the model.
The kinetic theory is an appropriate tool for description such a systemin the external field. Each sub-systemis described
by the distribution functions, for constituent particles of type i in six-dimensional phase space and time coupled with
the set of the Maxwell equations for the field. Solution for such set of equations is a formidable task even for modern
supercomputers. However, as we shown in the preceding sections, the majority of processes can be described by the
quasi-equilibrium distributions in each sub-system. The non-equilibrium effects of building up the high-energy tail of the
distribution can be treated separately for the description of such phenomena as melting and ablation. Therefore, one can
righteously simplify the above approach as follows.
The first step of simplification is reducing the set of kinetic equations to the set of coupled equations for the velocity
moments for the distribution function. Rigorously the kinetic equation approximates by the infinite set of coupled
momentum equations, where the equation for n-order moment includes (n + 1) moment. In many practically important
cases the response of the material can be described by the set of momentum equations truncated at the second velocity
moment (that is pressure) (Kruer, 1988). The first three moments represent the number density n
k
, average velocity u
k
, and
pressure P
k
(subscript k denotes electrons, phononsatoms or ions). The additional relation between the pressure, density,
and temperature the Equation of State (EOS) should be added to make the equation set complete. The properly chosen
EOS is crucial for correct description of phase transitions, melting and especially ablation. We discuss that in detail in the
following sections.
The truncated moment equations for the first three moments express respectively the conservation laws of mass (zero
moment), momentum (first moment) and energy (second moment). The zero moment of the kinetic equation gives the
continuity equation for the particle density. We take into account transient changes in the electron number density in
the conduction band when the ionisationrecombination processes are important. The first moment gives the momentum
conservation for electrons and ions as follows:
n
e
m
e
_
u
e
t
+u
e
u
e
_
= en
e
E
elst
P
e
n
a
m
a
_
u
a
t
+u
a
u
a
_
= en
a
ZE
elst
P
a
(2.84)
here E
elst
is the electrostatic field of electronion interaction. One can ignore fast electron motion on a time-scale of the
inverse electron plasma frequency and electron inertia due to the fact that ions are much heavier than electrons. By this
reason the left-hand-side in the first equation of Eq. (2.84) can be nullified, and the electrostatic field force expresses through
the electron pressure gradient. Now Eq. (2.84) reduces to the momentum conservation equation for ions:
n
a
m
a
_
u
a
t
+u
a
u
a
_
= (P
e
+P
a
). (2.85)
The charge conservation, n
e
= Z
n
a
, is taken into account in derivation of Eq. (2.85). The energy conservation for electrons
(ignoring fast electron motion) and ions reads as:
(n
e
e
)
t
= Q
abs
Q
eth
Q
eph
1
2
t
(n
a
m
a
u
2
a
+P
a
) = Zn
a
u
a
E +Q
eph
Q
ph-heat
_
n
a
u
a
_
m
a
u
2
a
2
+
3P
a
2n
a
__
(2.86)
where Q
abs
is the absorbed laser energy density rate, Q
eth
is the electronic heat conduction flux, Q
eph
=
en
eph
n
e
(T
e
T
L
) is
the energy transfer rate from electrons to lattice (ions), and Q
ph-heat
is the phonon heat conduction flux which we ignore in
the following calculations. In addition, the energy expenses for particular phase transition should be included in Eq. (2.86)
if the transformation of the material occurs.
(b) Two-temperature approximation
In many cases of ultra-short lasermatter interaction a laser-excited solid remains intact, i.e. the macroscopic atomic
motion could be neglected. Let us find howlong and at what absorbed energy level this assumption holds. An atomcould be
removed froma solid (ablated) if it receives the energy in excess of the binding energy. In the conditions close to the ablation
threshold, which we consider in detail in Section 5, the kinetic velocity of a freed atom is of the same order of magnitude
as the sound velocity v
sound
10
5
cm/s. Thus the expansion time could be estimated as the time when a surface atom
moves on a distance comparable to the absorption depth, t
hydro
l
abs
/v
sound
(2 3) 10
11
s. This means that a solid
remains practically intact during several tens of picoseconds after 100 fs pulse even when the deposited energy density
exceeds the ablation threshold. Thus, the hydrodynamic motion is not important during the laser pulse and long after of the
electron-to-lattice energy transfer time. Therefore one can ignore the hydrodynamics in Eq. (2.86) and assume u = 0. Then,
the set Eq. (2.86) reduces to so-called two-temperature approximation (Kaganov et al., 1957; Anisimov et al., 1974; Allen,
1987) that is a legitimate description of the state of laser-excited medium during the time when the mass and momentum
are unchanged, and two coupled equations express the energy conservation law:
(C
e
n
e
T
e
)
t
= Q
abs
Q
eth
Q
eph
(C
L
n
a
T
L
)
t
= Q
eph
(2.87)
where C
e
and C
L
are correspondingly heat capacities for electrons and the lattice, which, in general, are functions of electron
and lattice temperature. The energy exchange terms in the set of Eq. (2.87) are also temperature-dependent through the
electronphonon energy exchange rate in the form of Eq. (2.50). The heat conduction coefficient is non-linear in respect to
temperature. These equations are broadly used during the past decades for description a state of laser-excited solid with
temperature-independent energy exchange rates.
2.10. Summary
Absorbed energy density at the end of the laser pulse W(t
p
) =
2AF
p
l
abs
_
J
cm
3
_
; F
p
=
_
t
p
0
I(t)dt
_
J
cm
2
_
Maximum electron temperature k
B
T
max
e
=
_
4
F
AF
p
2
n
e
l
abs
_
1/2
; k
B
T
e

F
Maximum lattice temperature k
B
T
max
L

2AF
p
C
L
n
a
l
abs
; C
L
n
a
C
e
n
e
Major relaxation time scales:
Electronelectron collision time t
ee

1
pe
_

F
1.88
e
_
Phononphonon equilibration time t
main
t
eq

_
h
k
B
T
L
_

1
at
_
M
m
e
_
1/2
Electronphonon momentum exchange rate and time
mom
eph
=
3C
2
k
B
T
L
h
; t
mom
eph
(
mom
eph
)
1
Electronphonon energy exchange rate and time
en
eph
=
2k
B
T
L
F

mom
eph
; t
en
eph
(
en
eph
)
1
The time for building up the high-energy tail in the
Maxwell distribution
t
tail
0.85 t
main
_

k
B
T
_
1
exp
_

k
B
T
_
3. The evolution of atomic motion: the origin, life, and death of coherent phonons
3.1. Introduction
The fast development of femtosecond lasers from the early 1980s was accompanied by the advances of a diagnostic
technique for fast probing of swiftly excited solids by X-ray, optical and electronic probes with time resolution in the
range 10100 fs. These achievements generated a broad variety of experiments, which for the first time made possible
the observation of new phenomena in swiftly excited solids on the femtosecond time scale and on a space-scale of tens
of nanometres. The most impressive observations are the oscillations in the optical probe reflection from a laser-excited
solid with a frequency close to that of cold phonons in a solid (Cheng et al., 1991; Zeiger et al., 1992). It was also found
that the intensity of an X-ray probe beam diffracted from the laser-excited solid oscillates with the cold-phonon frequency
(Sokolowski-Tinten et al., 2003).
The excitation and detection of coherent lattice vibrations has been produced in many transparent and opaque materials,
such as semi-metals, (Cheng et al., 1991; Zeiger et al., 1992; Garrett et al., 1996; Hase et al., 1996, 1998; Misochko et al., 2007;
Murray et al., 2005; DeCamp et al., 2001; Johnson et al., 2008) transition metals (Hase et al., 2002), cuprates (Albrecht et al.,
1992), insulators (Ishioka et al., 2006), and semi-conductors (Rousse et al., 2001). It was then realised that the generation
of a phonon is preceded and followed by a subtle laser-induced atomic motion that is imprinted into transient material
properties of the excited solid (Boschetto et al., 2008a,b). The various stages of laser-induced atomic motion were identified
on the basis of analysis of the experimental data (Garl et al., 2008). First, the electronic excitation during the pulse, which
is shorter than the electronphonon energy transfer time, produces fast coherent atomic displacement followed by the
harmonic vibrations of the lattice with cold phonon frequency because the lattice remains cold. Heating of the lattice leads
to decay of harmonic modes due to non-linear phononphonon interaction, and the coherent vibrations cease to exist. The
onset of thermal expansion and instability manifests the onset of the transformation of the materials, eventually leading to
the disorder and melting of the lattice. Studies of all stages of the laser-induced atomic motion allow us to obtain a deep
insight into the microscopic nature of the material transformations produced by laser.
The ability to drive and control transient atomic motion via an external photon flux opens a number of interesting
applications, such as the possibility to induce phase transitions: transient phase state (Boschetto et al., 2008a,b) paraelectric-
to-ferroelectric (Cavalleri et al., 2001) or insulator-to-metal transitions Chollet et al. (2005), the selective opening of the
caps of nano-tubes in non-equilibrium conditions (Dumitrica et al., 2004). They also provide a basis for SASER (Sound
Amplification by Stimulated Emission of Radiation) experiments.
In this section we consider in succession the interconnected processes of laser interaction with matter at a level of
absorbed energy close to and exceeding the equilibrium enthalpy of melting. The laser-exerted forces start displacing
atoms early in the pulse time. The atomic motion in turn affects the optical and material properties. Thus, observing the
time-dependent history of the reflected probe beam allows one to retrieve the atomic motion on the sub-picosecond and
picosecond time scale. These processes gradually transform the material phase state during the time period up to several
tens of picoseconds. The studies of atomic motion on a femtosecond time scale and on a nanometre spatial scale at the pre-
melting stage allow a better understanding of the microscopic processes preceding the transformation and disordering of
the material.
3.2. Initial conditions in a laser-excited solid at the energy density comparable to enthalpy of melting
The succession of the processes in the short-pulse laser-excited solid is the same at the different levels of the absorbed
energy density. In this section we consider the processes at a deposited energy density close to the equilibrium enthalpy
of melting (or enthalpy of fusion). Enthalpy of fusion, H
f
, for the majority of solids, including metals and dielectrics, is of
the order of (0.31.0) kJ/cm
3
(see Table 4.1 in Section 4). The fluence of several mJ/cm
2
delivered by a 100 fs laser pulse
and absorbed in a skin layer 2030 nm thick easily creates such energy density in metals. Therefore the average during
the 100 fs pulse laser intensity is around 10
10
W/cm
2
, that is, about two orders of magnitude lower than the breakdown
threshold for dielectrics (Arnold and Cartier, 1992). Hence the processes of electron excitation and effects on the transient
properties of metals and dielectrics affected by an ultra-short laser pulse near the melting point are different. Below we
consider the laser interaction mainly with metals and semi-metals where the electron excitation, optical properties and
relaxation processes can be treated explicitly. In order to deposit the melting energy density into dielectrics the significant
increase in absorption should be achieved first. This is only possible if optical breakdown conditions are created (average
per-pulse intensity increases up to 10
12
W/cm
2
) and significant numbers of electrons from the valence band (more than
10%) are promoted to the conduction band. Then bremsstrahlung absorption on conduction electrons increases to the level
of several tens of percent, the absorption depth becomes comparable to the skin depth in metals, and then melting similar
to that in metals proceeds.
3.2.1. Absorbed energy density, electron and lattice temperatures
The electron temperature reaches its maximum value T
e,max
at the end of the pulse, along with the total absorbed
energy, which is all contained in the electron component E
el
= C
e
n
e
T
e
= 2AF(t
p
)/l
s
; where l
abs
is the absorption depth,
C
e
is the electron heat capacity, n
e
is the electron density, Aabsorption coefficient, and F(t
p
) is the incident laser fluence.
The relaxation times decrease in the inverse proportion to increasing temperature increase. Therefore, the duration of all-
important transient processes decreases with the growth in the absorbed energy density.
We discuss below the experimental and theoretical results for Bismuth, one of the most studied solids in respect to
the ultra-fast excitation, as an example for the numerical characterisation of ultra-fast processes in laser-excited solid.
We believe that the data and discussion below describe quite generally the parameters of a metal excited close to the
melting conditions. Enthalpy of fusion in Bi is 0.5 kJ/cm
3
(Landolt-Brnstein, 1983; Lide, 2008), just in the middle of the
Table 4.1
Melting parameters for metals and dielectrics in equilibrium.
b
binding energy; T
m
melting point; H
f
specific heat; S
f
enthalpy of fusion; n
l
, n
s
atomic number density in liquid and solid.
b
(eV/at) T
m
(eV) k
B
T
m
/
b
H
f
(kJ/cm
3
) S
f
Comments
Al 3.39 0.08 0.0236 1.07 1.378k
B
n
l
< n
s
Cu 3.49 0.117 0.0335 1.865 1.17k
B

Ag 2.95 0.1065 0.0361 1.097 1.098k
B

Fe 4.28 0.156 0.0364 1.95 1.296k
B

Au 3.81 0.1154 0.0303 1.23 1.13k
B

Bi 2.16 0.047 0.0217 0.5 2.4k
B
n
l
> n
s
Si 4.63 0.1454 0.0314 4.17 3.725k
B

Ga 2.81 0.026 0.0093 0.474 2.267k
B

InSb 2.65 0.0685 0.0258 0.975 3.0k
B

Ice (frozen water) 0.0235 0.335 2.638
whole range of the melting enthalpy values for solids. It is essential that the temperature-dependent optical properties
of Bi introduced below were retrieved from the experiments in equilibrium (see Gamaly and Rode, 2009, and references
therein) in agreement with the theory presented in Section 2. The experimental data indicate that Bismuth can be treated
as a good metal obeying the Drude-like dielectric function at elevated temperature range from room temperature and
up to 200 K over the melting point, see Appendix. Absorption coefficient and the skin depth for 800 nm laser light in
Bismuth are respectively A = 0.26; ls = 2.98 10
6
cm (Garl et al., 2008). Thus, 800 nm laser pulses with the fluence,
F(t
p
) = 2.76.7 mJ/cm
2
, deposit the energy density in Bi from the enthalpy of melting in equilibrium to twice higher
value (0.481.19) 10
3
J/cm
3
(Garl et al., 2008). The maximum electron temperature for the above range of fluences is
found to be T
e,max
= [4
F
AF(t
p
)/
2
n
e
l
s
]
1/2
, T
e,max
= 28254450 K (0.240.38 eV). The maximum lattice temperature is,
correspondingly, 701.5 and 1273 K (0.060.11 eV). Thus, the first distinctive feature of the ultra-fast excitation is apparent:
the maximum electron temperature is three times larger than the lattice temperature at the deposited energy density
comparable to the equilibrium enthalpy of melting. Therefore, one may expect a strong impact of the electron excitation
on the transient state and on the atomic displacements in the excited material. It is worth noting that the melting point for
Bismuth is T
melt
= 544.7 K. Hence, another question arises: when and how the solid transforms into a different phase state
in non-equilibrium conditions under fs-laser excitation.
Let us consider the relaxation times. The Fermi energy,
F
= 5.17 eV, and plasma frequency,
pe
= 1.310
16
s
1
for solid
Bi at roomtemperature are extracted fromthe optical data (Comins, 1972; Garl, 2008; Garl et al., 2008), see also Appendix B.
The maximumelectron temperature (that is the energy in excess over the Fermi level,
e
) in excited Bi,
e
= 0.240.38 eV =
(0.040.065)
F
. Thus equilibriumdistribution for electrons is established in a period, t
ee

1
ee
=
F
/
p
e
(12) fs (see
Section 2). Similarly, the phononphonon interactions lead to the establishment of an equilibriumin the phonon subsystem
(the lattice temperature). The phononphonon relaxation time lies in the range t
phph

=
h/T
L
(625) fs for the considered
Bi excitation energy span. Thus establishing the statistical distributions (electron and lattice temperature, T
e
= T
L
) early in
the pulse time is the rationale for using the 2-temperature approximation (T
e
(t), T
L
(t)) to describe the non-equilibrium
lasermatter interaction during the pulse and after the end of the laser pulse (Kaganov et al., 1957).
The temperature dependence of the electronphonon momentum transfer rate
mom
eph
has been retrieved directly from
the equilibrium optical experiments in agreement with the theory in the form (see Appendix B):
mom
eph
= 2 10
15
T
T
room
(s
1
).
Note that this dependence and numerical values are close to those for aluminium retrieved in a similar way and presented
in Section 2. The link between the electronphonon energy exchange rate
en
eph
and the momentum transfer rates has been
established in Section 2,
en
eph
= (2k
B
T
L
/
F
)
mom
eph
. Taking the value of the Fermi energy for bismuth,
P
F = 5.17 eV from
optical measurements at 800 nm (Garl et al., 2008), one obtains the energy exchange rate in the form:
en
eph

= 2 10
13
_
T
T
room
_
2
(s
1
).
The Fermi energy is 9.9 eV if all five valence electrons in Bi are in the conduction band (Ashcroft and Mermin, 1976).
The lattice heating time therefore appears to be shorter than the pulse duration, t
en
eph
= (
en
eph
)
1
= (1550) fs in the
temperature range, T
room
T
L
T
melt
. Heat diffusion coefficient (diffusivity) D is directly linked to the Fermi energy and
the electronphonon momentumexchange rate retrieved fromthe optical experiments under assumption that the electron
effective mass coincides with that for the free carrier: D = v
2
F
/3
mom
eph
. The diffusivity values recovered fromthe temperature
dependences in equilibrium are D = (2.02.89) cm
2
/s for the temperature range 293793 K (see Appendix B). These
results are in good agreement with the recent non-equilibrium measurements from the X-ray reflectivity data of fs-laser
excited bismuth giving D = 2.3 cm
2
/s (Johnson et al., 2008). However, these results are in sharp contrast to what would
be expected from the equilibrium data at low temperature. Indeed, the diffusivity from the textbook, D = 0.067 cm
2
/s
(Landolt-Brnstein, 1983) is 30 times lower. There is no a convincing explanation for this discrepancy to the best of our
knowledge.
Nevertheless, the time for cooling the skin depth of 29.8 nm Bi with the diffusivity D = 2.3 cm
2
/s recovered from
the optical measurements is: t
cool
l
2
s
/D = 2.3 ps, much longer than the 45 fs pulse duration. This result shows the
importance of the heat conduction for defining the state of excited solid after the pulse (Boschetto et al., 2008a,b). Indeed,
the temperature at the target surface decreases due to the linear electronic heat conduction:
T
max
(x = 0; t) T
max
(x = 0; t
p
)
_
t
cool
t
cool
+t
_
1/2
.
Thus, the temperature decreases 2.7 times in 25 ps after the pulse, diminishing significantly the effects of excitation.
3.2.2. Excitation of electrons
The number density of electrons excited to the conduction band in dielectrics and semiconductors by an avalanche-like
process is proportional to the electron energy density. A simple estimate suggests that the whole absorbed energy goes to
the electron transfer from the valence to the conduction band separated by the energy gap of E as:
n
e
(t
p
)
=
2AF(t
p
)
El
s
. (3.1)
In this section we are considering metals and semi-metals where the conduction and valence bands are either partly
overlapped or the band gap is narrow. For example, in such semi-metal as equilibrium liquid Bismuth all valence electrons
are transferredinto conductionbandas it was evidencedby observationof the optical properties of liquidBismuthby Comins
(1972). The band gap in Bi has a complicated structure in the energy space. The bands are either slightly overlapping or
separated by a narrow gap less then tenth of electron Volt. The above estimate at the absorbed energy density equal to the
equilibrium enthalpy of melting 2AF(t
p
)/l
s
H
melt
and at E 0.023 eV gives the number of excited electrons equal to
that for the liquid Bismuth. On the other hand during the ultra-fast excitation of metals the whole absorbed energy goes
directly to the electrons heating. In ultra-fast excitation of dielectrics the number of excited electrons decreases due to
recombination after the end of the pulse. We consider ionisation and recombination processes in dielectrics in Section 6.
3.2.3. Stress tensor in laser-excited solid
The equilibrium concept of the stress tensor in a laser-affected solid after electronelectron and phononphonon
equilibration can be used in the same form as in the equilibrium conditions. The stress tensor for initially isotropic medium
reads (Landau et al., 1984):
ik
= P
ik
_
E
2
ij
8
+
E
i
E
j
4
_

jk
+
E
2
8
_
n
a
_
jk
n
a
_
T
_

ij
+
E
i
D
k
4
; (3.2)
where the displacement vector has a form D
k
=
kj
E
j
. We assume that the dielectric tensor modified by the laser effect
consists of two terms, the Drude-like term,
D
, and polarisation term,
(p)
jk
:
jk
=
D

jk
+
(p)
jk
. (3.3)
The difference between the equilibrium state and the transient state is that in equilibrium n
e
is constant and the electron
temperature is equal to the lattice temperature, while in the transient state n
e
is changing along with time-dependent
electron and lattice temperature. The pressure in the transient state is the sum of the contributions from the laser-affected
electron and lattice sub-systems, P = P
e
+ P
L
. During the pulse and up to the temperature equilibration the electron
temperature is much higher than that of the lattice and therefore the electrons create the pressure. The dielectric function
is also modified by the laser action.
3.2.4. Time scales for the consecutive stages of atomic motion
The general scenario for the different stages of atomic motion under the swift excitation is the following. In unperturbed
solid atoms performharmonic oscillations around the equilibriumpositions. The fast excitation heats electrons while atoms
continue their unperturbed vibrations (see Section 2.9.2). The electrons reach the maximum temperature and create the
electron temperature gradient to the end of the laser pulse. The force, proportional to the electronic pressure (temperature)
gradient, acts on atoms inducing the atomic motion. This non-homogeneous excitation of electrons leads to increase in the
electro-static field of electronion-core separation that is expressed through the gradient of electronic temperature (see
Section 2.9.2). This force acts during the time shorter than phonons period (t
ph
= 2/
ph
100s fs) producing the fast
atomic displacement. The magnitude of this force is proportional to the absorbed energy density as it is described in detail
later. At the energy density of the order of the enthalpy of melting the elastic force responsible for the atomic vibrations
in unperturbed solid is larger than the electronic force. Therefore, the electronic force acts as a short blow inducing the
atomic vibrations with the cold phonon frequency. The next stage of atomic motion in time corresponds to harmonic
vibrations with cold phonon frequency, which lasts during the period for energy transfer from the electrons to the lattice.
The period of harmonic vibrations continues until the lattice acquires temperature close to the melting point. During this
period atomic vibrations gradually lose the harmonic character. Non-linear interaction between different phonon modes
becomes dominant and that eventually transforms a solid into a different phase or into a disordered state.
Fig. 3.1. Qualitative picture for the change of potential energy due to electronic excitation (scales are exaggerated).
One can also describe qualitatively the fast atomic motion in a swiftly excited solid using the language of the inter-atomic
interaction (Fig. 3.1). Indeed, the inter-atomic potential consists of the attractive electronic part and that of repulsion of ionic
cores. Fast electronic excitation decreases attraction and therefore reduces the binding energy. As a consequence, the inter-
atomic distance increases (atoms are displaced) as a result of electronic excitation. Taking the inter-atomic potential in the
Morse-like form allows one expressing the atomic displacement as a function of the electron temperature, binding energy,
equilibrium inter-atomic distance and the gradients of attractive and repulsive parts of the potential; we present this later
in this section, Section 3.4.
3.3. Forces driving atomic motion in solids
Let us describe quantitatively the atomic motion on the consequent time stages in laser-excited elemental solid with
identical atoms of mass M. First, we consider the swift excitation of atomic vibrations by the force acting on an atom
during the time much shorter than the phonon period. As the first approximation the spatial dispersion can be neglected
and the excited phonon is considered as a standing wave. It is instructive first to re-visit the atomic vibrations in a cold,
unperturbed, solid.
3.3.1. Elastic force in an unperturbed solid
The elastic force driving harmonic vibrations in a solid expresses through the first (quadratic) perturbation term in the
inter-atomic potential, which has the form:
U
el
=
1
2
_
2
U
q
2
_
0
q
2
1
2
M
2
0
q
2
. (3.4)
One estimates (
2
U/q
2
)
0

b
/d
2
; here q is the cold phonon amplitude,
b
is the binding (cohesive) energy, and d is inter-
atomic distance in equilibrium. The cold phonon frequency in Eq. (3.4) is
2
0

b
/Md
2
. The elastic force driving harmonic
vibrations immediately follows from Eq. (3.4):
F
el
=
U
el
q
M
2
0
q. (3.5)
The cold phonon amplitude (at a temperature lower than the Debye temperature) estimates as (Ilinskii and Keldysh,
1994, see also Section 2):
q
0

_

h
M
0
_
1/2
. (3.6)
Now the elastic force explicitly expresses through the basic characteristics of a solid:
F
el
(M
3
0
h)
1/2
. (3.7)
For example, in the unexcited Bi (atomic mass, M
Bi
= 3.47 10
22
g, inter-atomic distance in c-direction d = 3.3 , and
b
= 2.16 eV) the longitudinal A
1g
optical mode (vibration in c-direction) has a frequency of 310
12
s
1
, which qualitatively
complies with estimate by the above formula for the cold phonon frequency. The amplitude of the cold phonon fromEq. (3.6)
equals to q
0
= 1.4 10
9
cm. Now the elastic force driving vibrations in cold Bi is F
el
= M
Bi
2
o
q
0
= 4.4 10
6
erg/cm.
3.3.2. The forces in laser-excited solid
The laser electric field at the moderate intensity induces internal atomic motion (deformations) in solids and liquids,
while the total volume can be considered being constant. The atomic motion brings changes to initially homogeneous
dielectric function. The volume forces induced by the laser field express through the stress tensor modified by the field
action,
ik
, taken from Eq. (3.2) as follows (Landau et al., 1984):
f
i
=

ik
x
k
=
P
x
k
+

(p)
ik
x
k
E
2
8
+
(
D
1)
8
E
2
x
i
= f
th
i
+f
(p)
i
+f
pond
i
. (3.8)
Here we took into account that for the Drude-like part of the dielectric function the following relation holds, n
a
_
D
n
a
_
T
=
D
1. The first term in Eq. (3.8) is the thermal force of electronic pressure introduced earlier. During the pulse, when
only electrons are excited and the lattice is cold, the gradient of electronic pressure is a manifestation of the electrostatic
interaction between electrons and ions. The second term is the polarisation force when the polarisability depends on the
atomic displacement (the Placzek effect, see Ilinskii and Keldysh, 1994), and third term is the ponderomotive force. Note
that polarisation and ponderomotive forces are effective only during the pulse, while thermal force drives atomic motion
until spatial smoothing of temperature gradients. It is also worth noting that polarisation force in Eq. (3.8) is similar but not
identical to the force driving phonons excitation in the Raman effect (Shen and Bloembergen, 1965). The difference relates to
the fact that laser pulse duration is much shorter than the phonons period. Therefore, there is no interaction between laser
electric field and vibrational field of atomic motion; the phonon frequency does not enter into polarisation force explicitly.
Let us compare the laser-exerted forces to each other. We take the polarisability and dielectric tensor in the Placzek form
(Shen and Bloembergen, 1965; Ilinskii and Keldysh, 1994), and estimate the changes as, (
ik
/x
k
)
0

0
/d, d being the
inter-atomic distance. The unperturbed polarisability estimates with the LorentzLorenz formula,
0
= 3(1)/4(+2),
(Kittel, 1996). Then the polarisation force in Eq. (3.8) reduces to the following:
f
(p)
i
=

(p)
ik
x
k
E
2
8
=
_
ik
x
k
_
0
E
2
8

0
d
E
2
8
. (3.9)
Thermal force reaches its maximumat the end of the pulse. The total pressure is proportional to the absorbed energy density.
Therefore the thermal force expressed through the fluence, F(t) It, reads:
f
th
P
e
+P
L
l
abs
2AIt
l
2
abs
; (3.10)
where I = cE
2
/8 is the average laser intensity during the pulse. Now one can compare the magnitudes of all forces during
the laser pulse:
f
th
2AIt
l
2
abs
; f
(p)

0
d
I
c
; f
pond
(
D
1)
l
abs
I
c
. (3.11)
At the beginning of the pulse the polarisation and ponderomotive forces dominate the thermal force. After several
femtoseconds the thermal force compares with the polarisation force in the heated layer of tens of nanometres thickness.
At the end of the pulse the thermal force is significantly larger of the intensity dependent forces.
Let us consider, for example, Bismuth in conditions of the experiments of Garl et al. (2008). In these experiments 40 fs,
800 nm, laser pulses excited Bi layer at fluence of 7 mJ/cm
2
(l
abs
= 28.3 nm; = 22.39; d = 3.3 10
8
cm; n
a
=
2.82 10
22
cm
3
;
0
= 0.2; 6.34 10
6
cm
1
). The magnitudes of the forces acting on Bismuth atoms to the end of the
pulse are the following:
f
(p)
3.67 10
14
erg
cm
4
f
pond
4.38 10
14
erg
cm
4
f
th
4 10
15
erg
cm
4
.
The forces above are volume forces (force per unit volume). Let us convert the thermal force of Eq. (3.10) into the force acting
on a single atom, F
th
2AIt
n
a
l
2
abs
, and compare it to the elastic force driving atomic vibrations of Eq. (3.7). One can easily see that
the thermal force in Bismuth, excited to the energy density twice of the equilibrium enthalpy of melting, is by the order of
magnitude lower then the force driving cold harmonic vibrations. Another feature of the thermal force is the short period
of action equal to the pulse duration, which is shorter than the phonons period.
3.3.3. Equation of atomic motion in a laser-excited solid
The laser-exerted force can be considered as a perturbation imposed on the harmonic vibrations during the period when
lattice remains cold. One can consider the oscillations of an individual atom under the combined action of elastic and laser-
imposed forces as follows:
M
d
2
q
k
dt
2
= F
las
k
+F
el
k
;
where F
las
k
= F
th
+ F
(p)
+ F
pond
is the sum of the laser-exerted forces; q
k
is atomic displacement in k-direction. The
above equation holds for the description of initial purely harmonic motion. However, gradually the lattice is heated by
the energy transferred fromelectrons, and interaction between different phonon modes becomes significant. Therefore, the
phenomenological damping with coefficient should be introduced. We discuss the physical meaning of this damping later
in this section. Now the equation for atomic vibrations in a solid where electrons are excited but lattice remains cold reads:
d
2
q
k
dt
2
+2
dq
k
dt
+
2
0
q
k
=
F
las
k
M
. (3.12)
The elastic force is taken in the form Eq. (3.5).
3.4. Consecutive stages of laser-induced atomic motion
3.4.1. Coherent atomic displacement
The relative effect of the elastic force and laser-imposed forces is different on a short-time scale t <
1
0
and on a
long-time scale t
1
0
. The elastic force is a slow one; it is not effective during the period, which is much shorter than
the period of the atomic vibration. Correspondingly, the first term in the left-hand-side in Eq. (3.12) dominates, thus this
equation reduces at t <
1
0
to the simplest form of a Newton equation:
d
2
q
dt
2

F
las
k
(t)
M
. (3.13)
Solution of this equation is straightforward:
q
k
(t) = M
1
t
t
dt
0
F
las
k
(t
)dt

F
las
max
t
2
p
2M
. (3.14)
Thus, on a short-time scale, the laser-imposed forces produce a coherent displacement of atoms. It is instructive to compare
fast atomic displacement to the amplitude of cold phonon q
0
(2
h/M
0
)
1/2
of Eq. (3.6). The electronic force, F
th
2AIt
p
/n
a
l
2
abs
, is a dominant contribution into the sum of laser-imposed forces to the end of the pulse. Let us consider the
case when the absorbed energy density compares to the enthalpy of melting, 2AIt
p
/l
abs
H
fusion
10
3
J/cm
3
and calculate
the coherent displacement of Bismuth atom to the moment when harmonic vibrations commence at about 300 fs after the
pulse. One obtains from Eq. (3.14) that in these conditions Bismuth atoms are displaced on the distance less than 10
10
cm
that is approximately one tenth of the cold phonons amplitude.
3.4.2. Harmonic vibrations under electronic force
One can see that the maximum value of the sum of the laser imposed forces at the absorbed energy density in excess of
enthalpy of melting equals to approximately one tenth of the elastic force driving cold phonons,
2
0
q
0
F
las
k
/M. Thus on a
long-time scale t
1
0
the laser-imposed forces act as a small perturbation. This perturbation results in a small change in
the amplitude of the cold atomic vibration, q
k
(t). Therefore, one can search for solution of Eq. (3.12) in a form:
q
k
(t) = q
k0
(t) +q
k
(t) (3.15)
where q
0
is the unperturbed solution:
q
k0
(t) = q
k0
exp{i(
2
0

2
)
1/2
t}. (3.16)
Now Eq. (3.12) transforms into equation for the perturbation is as follows:
d
2
q
k
dt
2
+
2
0
q
k
+2
dq
k
dt
=
F
las
k
M
. (3.17)
We assume as the first order approximation that the perturbation oscillates with the same frequency as unperturbed
vibration and therefore that the perturbation affects only the vibration amplitude. Later on we account for the change in the
vibration frequency. The change in the vibrational amplitude at t
1
0
takes the form:
q
k
=
F
th,k
2
0
M
exp{i(
2
0

2
)
1/2
t }. (3.18)
Note that only the thermal force is effective after the pulse end. The phase and some pre-exponential constant should be
introduced in order to stitch the above solution to the initial atomic displacement.
3.4.3. Non-linear phenomena during the process of lattice heating
(a) Thermal expansion
The approximation of damped harmonic oscillations of atoms under the laser excitation is a valid description of atomic
motion only during the period when the lattice remains cold and electronic force drives the atomic vibrations. After the end
of the pulse the lattice temperature increases due to the transfer of energy from the electrons. The Boltzmann distribution
with temperature T,
ph
(q) = exp{U
el
/T} establishes earlier in the pulse time. Phonons energy is also distributed in
accord with the Boltzmann function where the perturbation in the inter-atomic potential has only harmonic (quadratic)
term and satisfies to condition, U
el
M
2
ph
q
2
T. Phonons do not interact in harmonic approximation; respectively
the average position of vibrating atoms does not depend on temperature; moreover, the displacement of oscillating atom
from the equilibrium position averaged over the Boltzmann distribution is zero q = 0 (Kittel, 1996). However, the mean
square displacement (or average energy of phonon treated as one-dimensional oscillator) is proportional to temperature,
q
2
= T/M
2
ph
. This derivation qualitatively complies with a more rigorous calculation for average amplitude of atomic
vibrations around the lattice vertex at the temperature that exceeds the Debye temperature, T
D
(Pines, 1964; Ziman, 1960),
M
2
D
q
2
T; here
D
= T
D
/
h is the Debye frequency.

The phononphonon interactions become important when increasing lattice temperature approaches the equilibrium
melting point. The atomic vibrations loose their harmonic character; the non-linearity in the interaction potential should
be taken into account. The perturbation in the inter-atomic potential with a 3-order term included reads (Kittel, 1996):
U
nl
= Cq
2
gq
3
; C

b
2d
2
; g =
1
6
_
3
U
x
3
_
0

b
6d
3
. (3.19)
Now the atomic displacement averaged over the distribution function with potential of Eq. (3.19) increases as the lattice
temperature grows up (Kittel, 1996):
q
nl
=
3gT
4C
2

T
2
b
d. (3.20)
This is the thermal expansion of heated solid. The atomic motion looses its oscillatory character and becomes chaotic
(randomised) when the change in the average atomic position by Eq. (3.20) constitutes significant part of the cold
oscillation amplitude. The displacement of equilibrium atomic position due to non-linear interaction of phonons calculated
by Eq. (3.20) at temperature equal to that of the equilibrium melting point coincides with 10% accuracy with the average
displacement following from the Lindemann criterion of melting (Lindemann, 1910; Pines, 1964)). This amazing from the
first glance coincidence is simply the evidence of failure of harmonic approximation for description of atomic motion during
the phase transition stage.
The non-linear processes of multi-phonon interaction at temperature close to the equilibrium melting point gradually
result firstly in the mode softening (that is a decrease of the oscillation frequency) and then later the instability develops,
when squared frequency may turn negative. Now we define time scale when non-linear interaction becomes dominant.
(b) Three-phonon interaction: phonons lifetime
The probability of multi-phonon processes per unit time defines the characteristic phonon decay time. Probability of
decay of an optical phonon in two acoustic phonons per unit time can be calculated with the help of quantum perturbation
theory as third order term(Ilinskii and Keldysh, 1994). We present here a simplified version of similar derivation estimating
the perturbation Hamiltonian, H
, as the third term in the series of potential expansion in powers of atomic displacement
similar to that in Eq. (3.19):
U U
0
+
1
2
_
2
U
x
2
_
0
q
2
1
6
_
3
U
x
3
_
0
q
3
+ = U
0
+H
+H
+ .
Then the scaling for the probability for the phonon decay is the following:
w
2|H
|
2
hk
B
T
; (3.21)
here we take the average phonon energy as, M
2
D
q
2

h
ph
k
B
T. Respectively, the perturbation Hamiltonian expresses
in the following form:
H

1
6
_
3
U
q
3
_
0
q
2
3/2

b
6
_
k
B
T
b
_
3/2
. (3.22)
For the temperature in excess of the Debye temperature, kT

h
D
the single phonon decay rate using Eq. (3.21) and
Eq. (3.22) reads:
w
decay

18
h
b
(k
B
T)
2
. (3.23)
For Bi at melting point of 544.7 K (
b
= 2.16 eV) the decay rate equals to several picoseconds in qualitative agreement with
the experimental observations. Non-linear phononphonon interactions also result in a dependence of phonon frequency
on the lattice temperature that is considered in the next section.
(c) Decrease in the phonon frequency and increase in the inter-atomic separation due to laser excitation
The excitation of electrons and lattice heating lead to the decrease in the binding energy, in the phonons frequency,
while the inter-atomic distance grows compared to that in a cold state. One can describe the property of slightly excited
solid using a simplified form for the empirical chemical pseudo-potential (Abell, 1985; Tersoff, 1986, see Appendix III):
V(r) = V
R
exp(r) V
A
exp(r). (3.24)
Here V
R
, V
A
, , (or s = / and ) are respectively the repulsive and attractive potentials along with their gradients. It
follows from Eq. (3.24) that the inter-atomic distance in equilibrium (where the potential is a minimum, V(d
0
) =
b
) can
be expressed via the binding energy:
d =
1
ln
_
(s 1)V
A
s
b
_
. (3.25)
The excitation of electrons reduces the attraction part of the inter-atomic potential and the binding energy. Thus in the
excited state at |
b,exc
| |
b
| k
B
T
e
; k
B
T |
b0
| the inter-atomic separation Eq. (3.25) can be presented as an expansion in
series in respect to powers of a small parameter, k
B
T/
b0
. The increase in the inter-atomic spacing reads, d d
0
+k
B
T/
b0
.
This increase depends on the gradient of the attraction part of the potential:
dq
1
k
B
T
e
b,0
. (3.26)
One can see that the above relation qualitatively complies with the atomic displacement calculated as the thermal
expansionfromEq. (3.20). However, inEq. (3.26) the asymmetry inthe inter-atomic potential (the difference inthe gradients
of attractive and repulsive parts) is taken into account. Similarly, the phonon frequency is calculated through the second
derivative of the potential:
2
ph
=
1
M
_
2
V
r
2
_
r=d
=

2
s|
b
|
M
. (3.27)
Thus, the frequency of phonons inexcitedsolidlinearly decreases withthe temperature growthas the first approximation
when the inter-atomic distance change with the temperature is ignored:
2
ph

2
0
_
1
k
B
T
e
|
b0
|
_
. (3.28)
Therefore the above estimate implicitly assumes that electrons are excited while the lattice remains cold. For example
in Bi excited to the maximum electron temperature of 0.3 eV (binding energy 2.16 eV) the phonon frequency decrease
predicted by Eq. (3.28) is 7%. Thus for A
1g
mode the frequency of 3 THz in accord to Eq. (3.28) should decrease to 2.78 THz.
The experimental result from Garl et al. (2008) is 2.82 THz. One can see that when the temperature approaches to the
equilibrium melting point (that is transition to the intermediate potential minimum corresponding to the liquid state) the
second derivative of the potential passes through zero value. That is a manifestation of the onset of the vibrational instability
when phonons frequency temporarily may turn negative. In fact at the stage when the non-linear inter-atomic interactions
became essential the transformations of the material should be considered with the statistical (thermodynamic) methods
as it is presented in Section 4.
3.5. Effect of atomic motion on the transient properties of the laser-excited solid
3.5.1. Dielectric function
In order to trace experimentally the subtle atomic motion following laser excitation one should determine how the
atomic motion affects transient properties of the excited solid. Apparently obvious way is to use the ultra-short X-ray and
electronic beam probes diffracted from the excited sample. Analysis of time-dependent diffracted intensity in principle
would allow directly trace the changing atomic positions. However, as we discuss later the interpretation of diffraction
experiments should be carried on with caution and leaves some experimental features unclear. Another way is to use optical
single and double simultaneous probes in order to measure the transient optical properties of excited solid. Both methods
are complementary and allow obtaining a general picture of the ultra-fast transformation of the material. Our goal in this
sectionis totrace the atomic motionandsubsequent phase changes by time-dependent dielectric properties of laser-affected
material. We assumed that the dielectric tensor modified by the laser effect consists of two terms, the Drude-like term,
D
,
and polarisation term,
(p)
jk
,
jk
=
D

jk
+
(p)
jk
. The polarisation termis a real number. Thus, the total dielectric function can
be presented as a sum:
jk
=
D

jk
+
(p)
jk

r
+i
i
r
=
(p)
jk
+
r,D
;
i

i,D
.
(3.29)
The Drude-like term has its conventional form:
D
= 1

2
p
2
+
2
eph
+i

2
p
2
+
2
eph
eph

r,D
+i
i,D
. (3.30)
This function depends on the number density of the conductivity electrons, n
e
, and on the electron effective mass through
the plasma frequency,
2
p
= 4e
2
n
e
(t)/m
e
. Number density of conductivity electrons may increase when a solid is heated
up to the melting point. For example it is known that dielectric properties of semi-metal Bismuth at the melting point and
above can be well fitted by the Drude-like dielectric function under assumption that all five valence electrons are transferred
into the conduction band. It is unknown to the best of our knowledge how the electrons effective mass changes (if any)
during a solid heating and melting. For example, it is established that in liquid Bismuth (Comins, 1972) the effective mass
is equal to that of free electron, supposedly it is also the case for solid Bi at room temperature. However, in simple metals
such as Aluminium, Iron, Copper, Lead, the electron effective masses at room temperature are well above the free electron
mass (Kittel, 1996, see Appendix II). There is no data on how the ultra-fast excitation might change the effective electron
mass. Therefore in the future analysis we ignore the changes of the electron effective mass. The dielectric function depends
on the temperature-dependent phonon amplitude through the electronphonon momentum exchange rate in the explicit
form:
eph

(0)
eph
q
2
q
2
0
. (3.31)
Eq. (3.31) was obtained under assumption that electronphonon scattering cross-section is proportional to the squared
phonons amplitude while the electron velocity equals to the Fermi velocity and it is unchanged during the interaction. At
the initial stage of atomic motion the electrons are excited but lattice still remains at the initial temperature. Therefore
one can neglect change in the phonon frequency and account for the changes in the phonons amplitude only thus taking
Eq. (3.31) at unperturbed frequency. The initial phonon amplitude has a conventional form:
q
0

_
2
h
M
0
_
1/2
. (3.32)
Nowthe small perturbation in the electronphonon rate expresses through the change in the phonon amplitude as follows:
eph

(0)
eph
2q(t)
q
0
. (3.33)
The polarisability part from Eq. (3.29) expresses directly through the laser-induced atomic displacement:
(p)
ik
= 4
_
ik
x
l
_
0
q
l
(t). (3.34)
The coherent atomic displacement q(t) during the pulse is expressed by Eq. (3.14); laser-perturbed atomic vibrations on
the later time presented by Eq. (3.18).
The real and imaginary parts of slightly perturbed dielectric function are the following:
r
=
(p)
jk
+
r,D
=
(p)
jk

_

D,r
ln n
e
_
0
n
e
n
e
+
_

D,r
eph
_
0
eph
0,eph
i

i,D
=
_

i,D
ln n
e
_
0
n
e
n
e
+
_

i,D
eph
_
0
eph
0,eph
.
(3.35)
Subscript 0 denotes that a derivative is taken fromthe unperturbed function (Garl et al., 2008). Thus, the dielectric function
in a swiftly excited metal-like solid is changing due to increase in the number density of conductivity electrons and in
variation of the electronphonon momentumexchange rate. Both are explicit functions of electron and lattice temperature.
In Bi the real part of the dielectric function in equilibriumliquid is slightly higher than that in a solid while the imaginary
part increases more than two times. The number density of electrons, and electronphonon rate in liquid are almost 3 times
higher of those in a solid (see Comins, 1972, and Appendix II). On the other hand the logarithmic derivatives of real and
imaginary parts of the dielectric function on electron number density and derivatives of real parts on the electronphonon
rate are of the same sign in both states. However, the logarithmic derivative of the imaginary part on the electronphonon
rate in a solid is slightly positive (0.86) while in a liquid it is strongly negative (11.95). Thus, the sign of this derivative
changes somewhere during the solidliquid transition time. One can make two conclusions from this analysis. First, one
cannot present the transient dielectric function in Bi as expansion into series during solidliquid transition. Second, the
transient phase state may have peculiar optical properties due to strong changes in the electronphonon collision rate
directly related to the laser-excited atomic motion.
3.5.2. Time-dependent reflectivity
Now it is instructive to present the time-dependent reflectivity of the probe laser beam from laser-excited solid through
the solid internal properties. Such reflectivity was directly measured in numerous experiments from different solids. First,
we express the Fresnel reflection coefficient R through the real and imaginary parts of the dielectric function =
r
+ i
i
as follows:
R =
+1
2
=
|| +1
2(|| +
r
)
|| +1 +
2(|| +
r
)
(3.36)
where || =
_
2
r
+
2
i
. Small variation of reflectivity then reads:
R =
_
R
r
_
0
(
r,D
+
(p)
jk
) +
_
R
i
_
0
i,D
. (3.37)
Substituting variations in the dielectric function fromEq. (3.35) one can present the small first order reflectivity variation
expressed through the changes in polarisation, in the number of the conductivity electrons and in the electronphonon rate
with the coefficients, which are combinations of the unperturbed solid parameters:
R =
_
R
r
_
0

(p)
jk
+C
n
e
n
e
n
e,0
+C
eph
eph
eph,0
(3.38)
where the coefficients are the combinations of the derivatives taken from the unperturbed functions:
C
n
e
=
_
R
r
_
0
_

r
ln n
e
_
0
+
_
R
i
_
0
_

i
ln n
e
_
0
C
eph
=
_
R
r
_
0
_

r
ln
eph
_
0
+
_
R
i
_
0
_

i
ln
eph
_
0
.
(3.39)
It is known from the numerous experiments that the relative amplitude of the reflectivity changes in observed oscillations
is of the order of 10
3
10
4
. One can use the expansion Eq. (3.38) for description of experiments if both coefficients in
expansion Eq. (3.38) have the same signs in solid and liquid states. This condition holds for experiments with Bismuth
(see Boschetto et al., 2008a,b; Garl et al., 2008) where C
n
e
is positive and C
eph
is negative and they are of the same order of
magnitude in solid and liquid states.
Each term in the formula Eq. (3.38) describing the reflectivity oscillations has clear physical meaning. Indeed, the first
term describes the reflectivity decrease due to polarisation changes in the dielectric function during the pulse, which
is positive, while the reflectivity derivative is negative (for excited Bismuth). The second term describes increase in the
reflectivity during the laser pulse due to the increase in the number of conductivity electrons. This contribution after the end
of the pulse gradually decreases (remaining positive) due to the recombination of excited carriers. The third contribution
into the reflectivity changes is the result of the effect of atomic vibrations on the optical properties of excited solid. This
contribution is of the same physical nature as the first termand it is negative. All variations in optical properties are explicit
functions of electron and lattice temperature, which are functions of laser and solid parameters.
3.6. Comparison with experiments
The state of a solid excited by femtosecond lasers at the absorbed energy density around and exceeding the equilibrium
enthalpy of melting has been studied using ultra-short optical, electron and X-ray probes in more then two decades (Cho
et al., 1990; Cheng et al., 1991; Zeiger et al., 1992; Garrett et al., 1996; Merlin, 1997; Rousse et al., 2001; Sokolowski-Tinten
et al., 2003; Fritz et al., 2007; Johnson et al., 2008; Sciaini et al., 2009). The most salient features of observations are the
oscillations in the optical probe beam reflection from the laser-excited solid with the frequency close to that of the cold
phonons in the solid (Zeiger et al., 1992). It was also found later that the intensity of X-ray probe beam diffracted from
the laser-excited solid oscillates with the down shifted phonon frequency at the absorbed energy density exceeding the
equilibrium enthalpy of melting (Sokolowski-Tinten et al., 2003). Below we discuss the experimental results for several
solids and compare them to the scenario and quantitative description of the processes during and after the femtosecond
laser excitation presented earlier in this section.
3.6.1. Measurements of time-dependent reflectivity with fast optical probes
The reflectivity of a probe beam was recorded with time resolution of 100 fs and less in many experiments (Zeiger et al.,
1992; Boschetto et al., 2008a,b) where the absorbed energy density has been below and up to several times larger the
equilibrium enthalpy of melting. Solid in these conditions may experience transition from solid to liquid state.
In order to understand what kind of information about transient phase state of laser-excited solid can be extracted from
the time-dependent reflectivity let us recollect briefly the scenario of processes occurring during the pulse and after the
end of the pulse. In the beginning of the pulse the polarisation force dominates over the electronic pressure force that for
a very short time pushes the reflectivity to drop below the initial (solid) level. Then the reflectivity increases over the level
for a cold solid and after the end of the pulse reaches its maximum value during the time comparable to the cold phonon
period. During the next period, the duration of which depends on the lattice temperature, the reflectivity oscillates with the
cold phonon frequency and with the gradually decreasing amplitude. At all stages the comparison of the average reflectivity
values to those in both liquid and solid states allows to relate the transient state to the known equilibrium values.
We discuss below experiments with laser-excited Bismuth, one of the most studied solid in relation to the excitation
of coherent lattice oscillations through optical probe reflectivity change. Reflectivity of a single (111)-oriented crystal of
bismuth excited by Ti: Sapphire 35 fs laser pulses at 800 nm was studied in a standard pumpprobe geometry, where pulse
fluence up to 20 mJ/cm
2
was used as a pump, and a small part of the main pulse was used as a probe (Boschetto et al.,
2008a,b). The reflectivity changes R were measured with the accuracy R/R 10
5
and temporal resolution of 40 fs.
Fig. 3.2. Time-dependent reflectivity changes, R/R, experimentally measured (solid lines) and calculated (dotted lines) in accord with the presented
above theory, for two pump fluencies: 2.7 mJ/cm
2
(a) and 6.7 mJ/cm
2
(b); the zero point on a time-scale indicates the arrival of the laser excitation pulse.
The insets show the first 3 ps after the pulse in detail.
Source: Taken from Boschetto et al. (2008a,b).
The results presented in Fig. 3.2 contain all characteristic features of the processes mentioned in the scenario above and
observed at the pump fluences in the range (1.515) mJ/cm
2
. There are several distinguishable features in the reflectivity
behaviour. First, the sharp initial drop in reflectivity was observed experimentally with 35 fs pumpprobe, while with
lower temporal resolution above 50 fs it was not detected. That qualitatively agrees with the theory, which attributes the
negative change in reflectivity to the polarisation effective during several femtoseconds in the beginning of the pulse. The
reflectivity reaches its maximum during 300 fs, that equals to phonons period, and then reflectivity starts oscillating with
phonon frequency, 2.9 THz, corresponding to the A
1g
mode in cold Bismuth at laser fluence 2.7 mJ/cm
2
. The oscillation
frequency decreases to 2.86 THz at 6.7 mJ/cm
2
that qualitatively agrees with the theory suggesting that phonon mode
softens when electron temperature grows up along with increasing fluence. Amplitude of the reflectivity oscillations decays
nearly exponentially with time. The atomic motion becomes anharmonic at the stage when lattice temperature approaches
to that for equilibrium melting and the reflectivity oscillations disappear. The duration of the period, where harmonic
oscillations are discernible, decreases with the increase in the pump fluence in qualitative agreement with the theory
(Boschetto et al., 2008a,b; Garl et al., 2008).
About 10 ps after the excitation pulse the reflectivity drops below the level corresponding to unperturbed solid and
remains constant and below this level for 20 ps, before slowly returning to the unperturbed value 4 ns after excitation.
The material excitation was fully reversible under excitation (2.76.7) mJ/cm
2
, which corresponds to the absorbed energy
density 12 times of the equilibrium enthalpy of melting. Thus, it was demonstrated by (Boschetto et al., 2008a,b), that
the meticulous analysis of the time-dependent reflectivity measured with high temporal resolution allows recovery of the
subtleties of atomic motion in a fs-laser excited solid during the period from several femtoseconds to tens of picoseconds
in accord with the theory.
The theory presented in this section suggests that the swift action of a femtosecond pulse excites phonons with the
frequency equal to that in the unperturbed solid. At the elevated initial temperature the frequency of atomic vibrations
decreases in accord with the increase in temperature. The time-dependent reflectivity change of bismuth crystal excited
at a pump fluence of 6.9 mJ/cm
2
at five different initial temperatures from 50 to 510 K presented in Fig. 3.3 (Garl et al.,
2008). One can clearly see that the increase in temperature results in the decrease of the time period when reflectivity
oscillations are discernible. Careful analysis also reveals that the initial frequency of oscillations decreases with the increase
of temperature. At the same time, the rate of damping is higher for the higher temperature.
The reflectivity time history at the different excitation levels and at different initial temperatures has common features.
The reflectivity increases first over the level for aninitial solidstate after the endof the pulse to its maximumvalue during the
time comparable to the cold phonon period. Then, reflectivity starts to oscillate with the phonon frequency that corresponds
to the initial lattice temperature and with the gradually decreasing amplitude. In several picoseconds time reflectivity
decreases below solid level thus indicating the onset of the phase transition to a liquid state.
It follows from Fig. 3.4 that the frequency of the reflectivity oscillations that corresponds to the initial phonon frequency
decreases (red-shifted) with the increase of the initial lattice temperature. We discussed above that the phonon frequency
decreases (red-shifted) compared to the cold one with the increase of the pumping fluence, i.e. with the growth of the
maximum electron and lattice temperature in excited solid at the constant initial temperature of a sample. However, the
magnitude of the shifts in these two cases, namely, with the increase of the initial temperature and with the increase of the
laser fluence, is different. Let us discuss this in detail.
In a solid heated in the equilibrium conditions, where the electron and lattice temperatures are equilibrated, the
binding energy decreases along with the increase of the inter-atomic distance due to thermal expansion; that is shift in
the equilibrium positions of vibrating atoms. The phonon frequency scales with the binding energy and the inter-atomic
distance as:
2
ph

b
/Md
2
. Following Eqs. (3.27) and (3.28), the changes in the inter-atomic distance d d
0
+k
B
T
b0
and
Fig. 3.3. Time-dependent reflectivity change R/R
0
of the probe beam for initial crystal temperatures from 50 to 510 K at a constant pump fluence of
6.9 mJ/cm
2
. The horizontal lines indicate the zero change for each measurement.
Source: Taken from Garl et al. (2008).
Fig. 3.4. Frequency of the A
1g
-phonon mode as a function of the initial crystal temperature at a constant pump fluence of 6.9 mJ/cm
2
, the line corresponds
to a fit performed with Eq. (3.43).
in the binding energy,
b

b0
k
B
T, both contribute into the frequency change as follows:
2
ph

2
ph,0
_
1
k
B
T
b,0
__
1 +
k
B
T
d
0
b,0
_
2
.
Keeping only the first order terms in expansion of above squared frequency in series of a small parameter, k
B
T/
b0
, one gets
the red shift of the phonon frequency as the function of the temperature in the form:
2
ph

2
ph,0
_
1
_
1 +
2
d
0
_
k
B
T
b,0
_
. (3.40)
One can see that the dependence from Eq. (3.40) fits well to the experimental data at Fig. 3.4.
Another distinctive feature of transient reflectivity presentedinFig. 3.3is a periodof time inwhichreflectivity oscillations
cease to exist. This time is associated with the lifetime of phonons. It can be directly recovered from the time-dependent
reflectivity as an inverse of the damping coefficient for the amplitude of oscillations. The theory suggests that the phonons
lifetime is inverse proportional to the probability per unit time for an optical phonon to decay into two acoustic phonons.
Optical phonondecay might be alsoconsideredas the onset of the vibrationinstability, whichinturnmanifests the beginning
of the disordering of a solid crystal. The initial temperature in Garl et al. (2008) was varied from50 K that is belowthe Debye
temperature of 119 K for bismuth, up to 510 K, that was close to the melting point of 544.6 K. The single phonon decay rate
from Eq. (3.23) reads:
w
decay

18
h
b
(k
B
T)
2
= C
(k
B
T)
2
h
b
. (3.41)
Fig. 3.5. Damping constant of the A
1g
-mode as a function of crystal temperature at a constant fluence of 6.9 mJ/cm
2
. The dashed line corresponds to a fit
with Eq. (3.41) with power 1.4; the solid line is the damping constant obeying to the temperature-squared law.
Here C is a dimensionless proportionality parameter in the scaling law. Comparison to the experimental data in Fig. 3.3 and
Fig. 3.4 with Eq. (3.41) shows that the decay time follows a temperature power-law of 1.4 (dashed line) at kT kT
D
while
the temperature-squared law from Eq. (3.41) overestimates the decay rate (solid line)see Fig. 3.5. At low temperature
range Eq. (3.41) predicts the constant decay rate in a qualitative agreement with the observations. Most probably, the two-
phonon decay process was mixed with some other involved dissipation processes. Thus the moment when the reflectivity
oscillation amplitude ceases to exist can be considered as the onset of the vibration instability, which in turn instigates the
beginning of the disordering of a solid crystal.
The time-dependent reflectivity changes in Bi excited by various pump fluence was also measured at the initial constant
temperature, the results are presented in Fig. 3.6. The down-shift in the oscillation frequency occurs due to the rise in the
electron temperature because the coherent displacement of atoms during the short pulse does not change significantly the
inter-atomic distance in a cold lattice. Therefore, the frequency change by the action of the ultra-short pulse in accord to
Eq. (3.28) expresses as:
2
ph

2
ph,0
_
1
k
B
(T
e
+T
0
)
b,0
_
. (3.42)
By comparing Eq. (3.40) and Eq. (3.42) one can see that the frequency shift in equilibrium is almost three times larger
than that during the sub-picosecond excitation at the same temperature due to big contribution of thermal expansion.
The dependence of the frequency of reflectivity oscillations in Bismuth sample at the initial room temperature (T
0
=
290 K, 0.025 eV) as function of exciting laser fluence is presented at Fig. 3.7. The dashed line with
ph,0
= 3.04 THz and
binding energy for bismuth of 2.16 eV in Eq. (3.42) fits well to the experimental data up to fluence of 13 mJ/cm
2
when the
maximum electron temperature reaches 0.573 eV. However, this linear approximation is only indication of a trend. When
temperature approaches to the melting point non-linear effects may become dominant and the above expansions will be
an underestimate.
The damping rate for the reflectivity oscillations excited exclusively by the electron excitation at the initial room
temperature was extracted from the results of Fig. 3.6 and plotted in Fig. 3.8 (Garl et al., 2008). It follows from Fig. 3.8
that the damping rate grows in linear proportion to the increase of the laser fluence in accord with the theory predictions.
3.6.2. Atomic vibrations observed with fast X-ray and electronic probes
It was demonstrated above that the fast optical probes with time resolution of 4050 fs allow uncover many details of
subtle atomic motion, whichare imprinted into transient optical properties. However, inoptical measurements the transient
changes in atomic positions are traced indirectly. During the last decade the ultra-fast X-ray sources and sources of energetic
electrons of 100200 fs duration became available. With such probe beams it seems possible to follow directly the atomic
displacements in excited solid by analysing the time-dependent behaviour of the diffracted probe beam. It is instructive to
compare studies of Bismuth excited by the ultra-fast laser in almost identical conditions (fluences in a range 620 mJ/cm
2
)
providedby the optical probes (Hase et al., 2002; Misochko et al., 2004; WuandXu, 2007; Boschetto et al., 2008a,b; Garl et al.,
2008) to those with fast X-ray and electrons probes (Sokolowski-Tinten et al., 2003; Fritz et al., 2007; Johnson et al., 2008;
Sciaini et al., 2009; Zhou et al., 2009). The main idea of probing the state of excited material by time-dependent behaviour
of the diffracted X-ray or energetic electron beamis to relate the changes in diffracted beamintensity to the transient phase
state of evolving material. Change in the diffracted intensity occurs when the Bragg conditions for reflection fromthe initial
crystalline lattice are violated. There are at least two reasons for such changes. First, it might be the non-homogeneity inside
the laser-excited sample due the spatial distribution of the incident laser intensity and spatial dependence of the absorbed
Fig. 3.6. Transient reflectivity signals R/R
0
for various pump fluences in the experiments at room temperature. The horizontal lines correspond to the
zero level for each measurement.
Fig. 3.7. Initial frequency of the A
1g
-mode as a function of the pump fluence at room temperature, the dashed line corresponds to frequencies calculated
with (42).
laser energy. Second, the phase transformation in excited material (for example, melting) could be a reason for the atomic
re-arrangement. Both processes may induce fast decrease of the diffraction efficiency. Is it possible to distinguish between
them? Let us discuss the spatial anisotropy induced by the laser beaminthe skinlayer that might be a reasonfor the violation
of the Braggs conditions.
(a) Laser-induced spatial anisotropy inside the laser-excited layer
The spatial distribution of the absorbed laser energy across the skin layer might be the primary source of the observed fast
drop of the diffracted beam intensity. Indeed, the absorbed laser energy decreases exponentially in the skin layer, E
abs
(x) =
E
abs,max
exp(2x/l
s
). Both the absorbed energy density and temperature have maximum values at the samplevacuum
interface. Atomic layers close to the interface start to expand with the speed of sound 2 10
5
cm/s after the energy
transfer fromthe laser-excited electrons. Thus in a picosecond time the atoms in several outer atomic layers move on several
angstroms outside the initial interface position. The gradient force F = (T
e
+T
L
) = (T
e
+T
L
)/l
skin
acts on atoms inside
the skin layer slightly displacing them in direction of the laser beam (into the sample).
The spatial anisotropy of heating exists until the temperature is smoothed by the heat conduction. The smoothing time
due to heat conduction across the skin layer is conventionally longer than the time for the diffracted intensity fall-off (see
Figs. 3.93.11). For example, the cooling time for 30 nm thick skin layer in Bi is 3.9 ps. Gaussian distribution of the laser
Fig. 3.8. Damping constant of the A
1g
-mode as a function of pump fluence at room temperature, the line corresponds to a linear fit to the data points.
Fig. 3.9. X-ray probe reflectivity in [222] direction in Bi excited by 6 mJ/cm
2
laser pump fluence as function of pumpprobe time delay.
Source: Adopted from Sokolowski-Tinten et al. (2003).
2
laser pump fluence as function of pumpprobe time delay.
intensity of both the pump and the probe beams over the surface induce lateral inhomogeneity, which further complicates
the interpretation of the diffracted beam intensity behaviour.
Undoubtedly, all these displacements affect the probe beam diffraction. It is worth noting that the sharp decrease in the
diffracted intensity was observed by Sciaini et al. (2009) even at the absorbed energy density half of the equilibriummelting
enthalpy when the disordering due to melting is energetically impossible. Obviously, the reason for the observed intensity
decrease calls for a different explanation than disordering due to non-thermal melting given by the authors. However, there
is no obvious way of distinguishing the decrease in the diffraction beam intensity caused by the above-mentioned non-
homogeneities from that produced by phase transformation.
2
laser pulse as function of pumpprobe time delay.
(b) The diffracted intensity of probe beam from the ultra-fast excited solid and its relation to the classic DebyeWaller factor
Many authors (Sokolowski-Tinten et al., 2003; Sciaini et al., 2009; Zhou et al., 2009) use the classic DebyeWaller factor
(Debye, 1913; Kittel, 1996) for determining the lattice temperature in a solid swiftly excited by the ultra-fast laser. The
DebyeWaller factor was introduced in order to account for the random thermal motion effect on the intensity decrease of
the diffracted X-ray beamfroma crystal in equilibriumconditions. Let us discuss howthe non-equilibriumconditions at the
temperature approaching the melting point may affect the validity of classical (equilibrium) approach. The decrease in the
beamintensity diffracted froma mediumwhere the atoms are displaced at a small distances u (u G
1
; G is the reciprocal
lattice vector) expresses as the following (Kittel, 1996):
I
I
0
= exp(iG u) 1 iG u
1
2
(G u)
2
. (3.43)
The angular brackets means averaging over the distribution function. The classical DebyeWaller formula relates the
scattered intensity decrease to the sample temperature in equilibrium conditions, I(h, k, l)/I
0
= exp{k
B
TG
2
/M
2
}
(Kittel, 1996). Equilibriumimplies that the randomthermal displacements are uncorrelated with G. Therefore the condition
G u = 0 holds. It also means that the harmonic vibrations are distributed in accord with the Boltzmann function
f
B
exp{M
2
ph
u
2
/k
B
T}, here
2
ph
is the phononfrequency. The average atomic positions remainunchanged inequilibrium,
u
_
uf
B
du = 0. At the temperature approaching the melting point the atomic vibrations lose their harmonic character,
and the third order correction to the interaction potential becomes significant. Now the first term in RHS of Eq. (3.43)
becomes non-zero, G u = 0, along with the average displacement from the equilibrium position u k
B
Td/
b
(Kittel,
1996). Therefore at the temperature approaching the melting point the classic DebyeWaller expression, ignoring the first
term in Eq. (3.43), becomes invalid. The first and the second order terms of the diffracted intensity decay are both non-zero
in respect to the averaged displacement and are proportional to the lattice temperature. The first order term is also non-
zero in the case of coherent atomic displacement induced by the electronic pressure gradient. For these reasons the classic
DebyeWaller expression cannot be used to determine the lattice temperature approaching or even exceeding the melting
point.
(c) A comparison between the X-ray and the optical probing of Bismuth in identical excitation conditions
It was observed in the experiments of Sokolowski-Tinten et al. (2003) that intensity of the probe X-ray beam diffracted
fromBismuth crystal excited at 6 mJ/cm
2
laser fluence (800 nm, 120 fs) oscillates during the observation period of up to 4 ps.
These observations are in qualitative agreement with the results from optical probes (Garl et al., 2008) where reflectivity
oscillations ceased to exist at 8 ps after excitation by 6.7 mJ/cm
2
. Indeed, the X-ray diffraction intensity oscillates with
2.12 THz that might be compared to 2.9 THz frequency for A
1g
phonon mode in cold Bismuth and to 2.75 THz measured
by Garl et al. (2008) from the optical reflectivity oscillations. The absorbed energy density at such fluence is twice of the
melting enthalpy. The theory predicts the frequency of phonons at these conditions to be 2.67 THz (T
e,max
= 0.35 eV and
T
L,max
= 0.077 eV) close but between the optical and the X-ray results.
Time-dependent features of the X-ray diffraction efficiency from Bismuth excited by 20 mJ/cm
2
laser fluence (absorbed
energy density about 7 times of the equilibriumenthalpy of melting) does not oscillate (see Fig. 3.11). However, (Sokolowski-
Tinten et al., 2003) attributed the decrease of the mean value of the X-ray signal (dotted line in Fig. 3.9) to the DebyeWaller
effect, suggesting that it reflects the increasing random component of the atomic motion, and extracted the atomic
displacement using the classic DebyeWaller factor. Previous analysis suggests that the observed diffraction fall-off might
be result of non-homogeneities in the laser-affected layer. Therefore application of the classic DebyeWaller factor for
estimation of atomic displacement is inappropriate.
Summing up, one can conclude that both optical and X-ray probing of bismuth excited in almost identical conditions
give similar qualitative picture in accord to the theoretical scenario. First, the coherent atomic displacement is produced
by the polarisation force and electron pressure force during the laser pulse that is recorded by the negative drop in the
reflectivity. Then oscillations with the frequency corresponding to that of phonon in a sample at the initial temperature and
red-shifted due to electron excitation were observed. The phonons amplitude gradually decreases while electrons transfer
the energy to the lattice. Heating of the lattice and thermal expansion transformed initially harmonic vibrations of atoms
into strongly non-linear motion that manifests the onset of solid-to-liquid phase transformation. This process is identified
by the measurement and interpretation of the damping rate of the reflectivity oscillations. The observed dependence of the
damping rate on temperature is close to the dependence of the rate of the optical phonon decay into two acoustic phonons
that confirms interpretation of the inverse damping rate as the phonon lifetime. The measurements and analysis of the
reflectivity after the decay of oscillations gave the evidence that the solid experiences transition to some transient state; its
reflectivity being the intermediate between those for solid and liquid. There are no indications that the phase transition
is completed during several tens of picoseconds. Both optical and X-ray probe measurements clearly demonstrate that
bismuth is not melted at the energy density twice of the equilibrium enthalpy of melting. However, it is difficult to say
at this stage when and at what conditions the transition occurs and what might be the transient state of a material during
the transformation. It is clear that the phase transitional stage of disordering and transformation of the material should be
described by statistical and thermodynamic methods. That is the subject of the next section.
3.7. Summary
Amplitude and frequency of phonons in an
unperturbed solid
q
0

_
2 h
M
0
_
1/2
;
2
0

b
Md
2
Volume density of forces exerted by laser pulse
in a solid
f
i
=

ik
x
k
=
P
x
k
+

(p)
ik
x
k
E
2
8
+
(
D
1)
8
E
2
x
i
= f
th
i
+f
(p)
i
+f
pond
i
Electronic pressure force driving atomic
vibrations
F
th
= P
e
/n
a

2AIt
n
a
l
2
abs
Quasi-harmonic vibrations (phonons) driven by
electronic force
q
k
=
F
th,k
2
0
M
exp{i(
2
0

2
)
1/2
t }
Average displacement from the cold equilibrium
position (thermal expansion)
q
nl

k
b
T
2
b
d
Single phonon decay rate into two acoustic
phonons
w
decay

18 h
b
(k
B
T)
2
Red shift in the cold phonon frequency due to
electronic excitation (atomic positions are
unaffected)
2
ph

2
0
_
1
k
B
T
e
|
b0
|
_
Atomic motion imprinted into the changes in
reflectivity of the optical probe beam through
the electronphonon coupling rate
R =
_
R
r
_
0

(p)
jk
+C
n
e
n
e
n
e,0
+C
eph
eph
eph,0
Red shift of the phonon frequency as the function
of the lattice temperature in equilibrium, T
e
= T
L
(thermal expansion included)
2
ph

2
ph,0
_
1
_
1 +
2
d
0
_
k
B
T
b,0
_
4. Ultra-fast melting by fs-lasers
4.1. Introduction
Numerous experiments have demonstrated that all transformations induced by the ultra-short laser action need more
energy than similar transformations in equilibrium. We start from the study of the well known and seemingly the simplest
transformation of the material, namely, melting. In this section we consider the processes occurring in a solid where the laser
energy, well in excess of the equilibrium enthalpy of melting, is deposited during a time shorter than the major material
relaxation times.
Melting is one of the most common crystals transformations in equilibrium. Melting in thermodynamic equilibrium has
been intensively studied for more than 100 years. Nevertheless this phenomenon is considered as a mystery (Cahn, 1986),
because it is still impossible to establish unequivocally the common properties of melting for different solids. However,
important steps in developing a general theory for crystalliquid transition were made by clarifying many common features
of melting in different crystals. It was established that a succession of instabilities (catastrophes) precedes the disordering
of the crystal, and the ultimate stability limit of a superheated crystal is determined by the equality of the crystal and
liquid entropies (entropy catastrophe) (Fecht and Johnson, 1988; Tallon, 1989). Moreover, it was found by Fecht (1992) that
conditions for disordering (entropy catastrophe) depend heavily on the concentration of non-equilibrium point defects. In
turn, the formation of defects depends on the presence of atoms from the high-energy tail of equilibrium distribution.
The advent of fs-lasers and novel diagnostic techniques (ultra-short X-ray, optical and electronic probes) during recent
decades has provided an opportunity to study the transient states of a material excited by ultra-short powerful lasers. The
observation of the time history of solidliquid transition at fs and ps time scale allows one to uncover the effects of the
non-equilibrium processes on the ultra-fast transformation of the material, and better understanding of the much slower
process of equilibrium melting, thus elucidating 100 years old mystery.
The ultra-short pulse swiftly excites electrons to an average energy-per-electron several times larger than the lattice
temperature. The lattice temperature is established later, after the electrons have shared their energy with the lattice, and
electronlattice temperatures have beenequilibrated. Therefore, it is quite reasonable to ask whether the energetic electrons
disorder (melt) the lattice well before the energy is transferred to the lattice by the electron-to-phonon collisions? In order
to answer this question one needs to understand first the processes of interaction between the electrons and the lattice on
a very short time-scale, and also to find what would be the signatures of the disordered lattice that can be diagnosed with
time resolution higher than the electron-to-lattice energy-transfer time.
The primary effects of the excited electrons on a cold lattice are twofold. The first effect relates to the spatially non-
homogeneous heating of electrons in the skin-layer. It is directly connected to the action of the electrostatic field of charge
separation on ion cores. The second effect appears in conditions of homogeneous excitation. It causes a deformation of the
inter-atomic potential, a decrease in the attraction force due to swift electron excitation, and a corresponding increase in
the inter-atomic separation. However, these effects produce atomic displacements similar to thermal expansion, which
should not be confused with disordering. In order to account for disordering, statistical methods for the description of
transformation of the material should be employed. The change from the long-range order of the crystal to the short-range
order of a liquid is characterised by the correlation functions that could be extracted from electron and X-ray diffraction
experiments. The main parts of local equilibriumdistribution functions in the electronic and lattice subsystems (the electron
and lattice temperatures) are established very fast by the ultra-short pulse action, in a few fs. The main non-equilibrium
processes, which require much longer time, are electron-to-lattice energy transfer and the building up of the high-energy
tail in the Maxwell distribution function (Gamaly et al., 2005). It appears that statistical thermodynamics can be applied to
describe what happens in both sub-systems. Therefore, the time-dependent entropy, being the most fundamental measure
of disorder, should be applied for the description of the degree of ultra-fast disordering.
The superheating of ideal defect-free crystals in equilibriumto a temperature beyond melting point without melting is a
well-known phenomenon. The crystal melts due to an intrinsic instability (entropy catastrophe), which has been predicted
as the ultimate stability limit of a superheated crystal (Fecht and Johnson, 1988; Tallon, 1989). The major contributions to the
critical entropy come from the heating (thermal disordering) of the lattice and from the thermal point defects, or vacancies.
The point defects, the inherent feature of equilibrium, are generated by the energetic particles from the high-energy tail
of equilibrium distribution function (Fecht, 1992). If the vacancy contributions were neglected, the critical entropy (and
the onset of disordering) would be achieved solely by superheating a crystal to 3T
m
(Fecht and Johnson, 1988). Thus the
transient processes of relaxation and establishing the distribution functions in electron and lattice sub-systems are crucial
for understanding the reconstruction of crystals following the swift excitation.
The section is organised as follows. First, the melting process and equilibrium criteria for melting are revisited. Then, the
processes of excitation, the redistribution of energy, and the directed, oscillatory and randomatomic motions are considered
in sequence. The critical entropy criterion for melting in equilibrium (Fecht and Johnson, 1988; Tallon, 1989) implies that
catastrophic disordering commences when the entropy of a solid equals to that of its liquid state. The temperature point of
the lattice when the overheated crystal catastrophically melts is defined by the corresponding value for the critical entropy.
Then it will be demonstrated that ultra-fast melting may only occur as a consequence of the strong overheating of the lattice
due to the high density of energy deposition. The disordering of the solid induced solely by strong electronic excitation, non-
thermal melting, is proved to be thermodynamically impossible.
Finally, the critical entropy value is connected to the time-dependent parameters of the material. In the final section the
experimental results on ultra-fast excitation of various solids probed by femtosecond optical, electronic and X-ray beams
with high temporal resolution are described and compared with the theoretical predictions.
4.2. Revisiting melting in equilibrium
The major characteristics of melting in equilibrium are the melting point, enthalpy and entropy of melting. The melting
point in equilibrium at a pressure P
m
is conventionally defined as the temperature at which the Gibbs free energies of the
two phases are equal, and beyond which thermodynamic instability occurs indicating commencement of the transformation
process. The temperature increase during the melting also indicates that the cohesive (binding) energy between atoms is
decreased. Thus the melting process leads to changes in the inter-atomic potential. The standard enthalpy of fusion (specific
melting heat), H
f
, is the amount of thermal energy, which must be absorbed in unit volume to change states from a solid
phase to a liquid at the melting point. The corresponding entropy of fusion is the increase in entropy entirely due to the
heating of a solid to the melting point, S
m
= H
f
/T
m
.
There are several major differences between solid and liquid states. First, the long-range order characteristics for the
crystalline state change to those of the short-range order, or disorder, in a liquid state. Second, a remarkable feature is
that the shear modulus is zero in liquids. This feature directly relates to the change in the inter-atomic potential during the
melting process. Third, the number of conductivity electrons in the conduction band may significantly increase in the molten
phase, as happens in Bismuth (see Comins, 1972). The entropy difference between the crystal and the liquid (entropy of
fusion) phases reaches its maximumat the melting point. Therefore, an increase of the temperature in a solid to the melting
point signifies the commencement of the melting process. We briefly recollect below the numerous attempts to establish
theoretically the criteria for melting in equilibrium.
4.2.1. Criteria for melting
The commencement of the solid-to-liquid transformation is preceded by a succession of instabilities (or catastrophes),
which delimit the range of stability for the crystalline state. These are the vibrational instability of Lindemann (1910), the
elastic shear instability (rigidity catastrophe) of Born (1939); the instability generated by the presence of defects (vacancies)
of Gorecki (1977). Lennard-Jones and Devonshire (1939) were the first who treated melting as transition froman ordered to
a disordered state. The entropy catastrophe of Fecht and Johnson (1988) at T > T
m
sets the ultimate limit for the crystalline
stability and allows connecting the value of entropy to the critical temperature at which the catastrophic melting occurs.
In fact, the theories and criteria of melting mentioned above allow identifying the different time stages during the
transformation of the material from solid to liquid state when a particular process dominates. The duration of these stages
in equilibrium is long when compared to the time for establishing the full distribution function in a material experiencing
transition. On the contrary, in the condition of strong excitation by laser pulse shorter than the major relaxation times, the
processes similar to those inequilibriumdevelopina short time while the distributionfunctionis still changing. By analysing
the time history of transformations induced by the ultra-fast excitation with the help of the equilibrium theories of melting
allows one to associate the particular time stage with the specific physical process in non-equilibrium conditions.
(a) The Lindemann criterion for melting
Atoms experience harmonic vibrations in solid in equilibrium. The energy of harmonic vibrations averaged by the
Boltzmann distribution equals to the temperature in the energy units. Thus, the squared amplitude of vibrations grows
in proportion to the temperature under the assumption that the frequency of vibrations remains unchanged and equal to
the Debye frequency,
D
. Lindemann (1910) assumed (see also, Pines (1964)) that any solid melts when the amplitude of the
vibrations of the atoms about their equilibrium positions exceeds some threshold value, q
m
. He found this threshold value
from the condition that the average phonon energy at the melting point equals to the melting temperature in the energy
units while the average phonon frequency is close to the Debye frequency, M
2
D
q
2
/2 3k
B
T
m
/2. The squared ratio of the
average amplitude of atomic vibrations q
2
m
defined from this conditions to the inter-atomic spacing, d
0
, reads:
q
2
m
d
2
0
=
T
m
M
2
D
d
2
0
= const. (4.1)
It is remarkable that this ratio is fairly constant for the alkali metals, and it changes only slightly (15%) fromLi to Au (Pines,
1965). The above formula can be re-written in the simple form expressing the Debye frequency via the binding energy as
M
2
D

b
/d
2
0
:
q
2
m
d
2
0
T
m
b
= const. (4.2)
One can see that the melting temperature constitutes approximately 3% fromthe binding energy over the whole periodic
chart of elements with rare deviations to 15% for Gallium, 2% for Uranium and Lithium, and 4% for Beryllium. Data for some
metals and dielectrics are presented in Table 4.1.
The non-linear effects are strong close to the melting point and equilibrium positions of atoms are shifted due to
the thermal expansion q
nl
melt
d
0
T
m
/
b
; this shift is larger than the amplitude of vibrations calculated in harmonic
approximationby Eq. (4.1). Therefore the phonons are transformedinto non-linear vibrations whentemperature approaches
the melting point. Hence the Lindemanncriterionis a qualitative indicationonthe onset of the vibrationinstability. Phonons
frequency decreases withincrease of temperature incomparisonwiththe Debye frequency. The moment whenthe phonons
frequency squared becomes negative,
2
ph
< 0, manifests the onset of an instability at which the lattice starts to reconstruct.
Summing up, one can see that the Lindemann criterion does not establish the melting thresholdit indicates qualitatively
on the onset of the vibration instability
(b) The Born melting criterion
Born (1939) suggested that the difference between solid and liquid is that solids have elastic resistance against shearing
stress which is absent in liquids. Therefore the limit for the pressuretemperature range for which a crystal is stable is
determined by the condition of zero shear stress. According to Born (1939) the temperature point at which the shear stress
turns to be zero signifies the transition to a liquid state. The condition of zero elastic stiffness, c
44
= 0, allows one to find
the ratio between the melting temperature and the binding k
B
T
m
/
b
energy that equals to 0.35, ten times overestimate
in comparison to real value (see Table 4.1). Born also derived the Lindemann formula with numerical coefficient slightly
different from that of Lindemann. Qualitatively the condition of zero shear modulus coincides with the onset of vibration
instability when
2
ph
= 0, because both the modulus and the squared phonon frequency are proportional to each other (and
to the second space derivative of the potential: c
44

2
ph

2
U/x
2
).
(c) Entropy and enthalpy catastrophe as a stability limit for crystalline material
It should be noted that none of the above criteria predicts melting temperature correctly or allow one characterising
the melting process quantitatively. The entropy catastrophe of Fecht and Johnson (1988) sets the ultimate limit for the
crystalline stability at T > T
m
. Fecht and Johnson (1988) and Tallon (1989) established that the solid to liquid transition
occurs in a succession of steps with progressively elevated temperature, which are characterised by the corresponding
change in entropy. The critical temperature corresponds to the state where the entropy of the solid equals to that for the
liquid. Beyond this point a superheated solid cannot exist because the entropy of solid at T > T
cr
becomes larger than in
liquid, which is impossible from the thermodynamic viewpoint. The catastrophic solid-to-liquid transition commences in
a state characterised by the critical values, the values of critical entropy and critical temperature, which are interrelated.
Therefore, the advance to the critical temperature indicates the commencement of disordering distinctive to a liquid state.
These critical values can be considered as the ultimate criteria defining the disordering.
4.2.2. Critical entropy and critical temperature
It was suggestedandsupportedby superheating experiments (Cahn, 1986andreferences therein) that melting is initiated
by continuous vibrational instability at the free solidsurfaces andgrainboundaries, whichserve as heterogeneous nucleation
sites. The kinetics of melting depends strongly onthe development of hetero-phase fluctuations, representedby small liquid-
like clusters (melting seeds) evolving in the crystalline phase and breaking the symmetry of the crystalline lattice (Landau
and Lifshitz, 1980a). Fecht (1992) pointed out that the destabilisation of a crystal can be enhanced by formation of non-
equilibrium point defects, such as vacancies, anti-site defects and interstitials, which are intrinsic features of equilibrium
characterising static disorder. He also indicated that such defects are typical for a transition externally driven by laser
irradiation. The number of vacancies is a strong function of temperature and this number is proportional to the number of
broken bonds in a lattice. Vacancies could condense to free surfaces and in the bulk forming small clusters (melting seeds)
with the surface energy being proportional to the heat of formation of a vacancy. For a closed system near the melting
point the probability of a fluctuation that leads to a seed formation depends on the entropy of fusion. For example, taking
S
fusion
= 1.1k
B
as for silver, one gets w exp(S
f
/k
B
) 0.33. This probability is strongly enhanced by the increase in
the vacancy concentration because S
f
0.
In what follows we neglect, as a first approximation, the heterogeneous nucleation at the surface while considering
the stability limit of a superheated crystal (critical entropy and temperature) assuming that instability would result in
homogeneous disordering and catastrophic melting in the bulk (Fecht and Johnson, 1988; Tallon, 1989, 1984). The specific
implications relevant to the heterogeneous nucleation, which are connected to the ultra-fast heating of the sample surface
by laser to be discussed later in this section.
(a) Entropy difference between crystal and liquid at the melting point
The entropy and enthalpy differences between crystal and liquid have its maximum at the melting point. The entropy
of fusion defined as, S
fusion
(T
melt
) = H
fusion
/T
melt
, and measured in the units of the Boltzmann constant. For the binary
inter-metallic compounds it equals to S
fusion
= 1.5k
B
; for the good metals it is 1.11.4k
B
; while for the semi-metals and
dielectrics it equals to 24k
B
(see Table 4.1).
The total change in entropy is a sumof contributions fromthe electrons, the lattice and the thermal defects. The electron
and lattice entropy changes are volume and temperature dependent. We take into account only temperature dependence
for the concentration and entropy of defects assuming the volume dependence to be negligible. It is also suggested that the
thermal defects implicitly include the effects of mixing (static disorder). Therefore the contribution of communal entropy
(Tallon, 1989) or entropy of mixing (Kittel, 1996) is not included explicitly.
The entropy of an overheated metal can be calculated in a general form as a sum of contributions from the isothermal
volume changes due to expansion, S
exp
, the isochoric (constant volume) heating of electrons, S
e
, heating the lattice, S
th
,
and that from the thermal defects (vacancies) S
vac
(Fecht and Johnson, 1988; Fecht, 1992; Tallon, 1989):
S
crit
= S
exp
+S
e
+S
th
+S
vac
. (4.3)
We shall measure the entropy in units of the Boltzmann constant.
(b) Entropy increase from the isothermal volume change
The first term in Eq. (4.3) comes from the isothermal volume changes:
S
exp
=
_
S
V
_
T
V. (4.4)
It is shown (Tallon, 1989) that for the majority of simple solids the entropy change due to isothermal volume change
expresses as:
S
exp
k
B
V
V
0
. (4.5)
(c) Isochoric contribution from electrons and lattice
The isochoric (constant volume) contribution comes from the heat capacity of the degenerated electrons, C
e
=
2
k
2
B
T
e
/2
F
:
S
e
(T) =
T
T
0
C
e
T
dT =

2
k
2
B
(T T
0
)
2
F
. (4.6)
Here
F
is the Fermi energy. For example, in aluminium at the melting point S
exp
0.2k
B
; and S
e
0.023k
B
. Both
contributions are small in comparison to entropy of fusion because the melting temperature of Al is small compared to the
Fermi and binding energies.
(d) Isochoric contribution from lattice heating
A major contribution to entropy comes from the thermal disordering due to lattice heating:
S
th
(T) =
T
T
0
C
p
T
dT C
p
ln T
T
0
. (4.7)
Thus, for good metals the entropy change due to heating from room temperature up to the melting point, taking the
DulongPetit value for the lattice heat capacity, is in the range S
th
(T
m
) = (3.55.5)k
B
see Table 4.1.
(e) Contribution from the thermal point defects
Thermal point defects are intrinsic feature of the equilibrium solid. Destabilisation of a crystal when its temperature
continuously increases is enhanced by formation of non-equilibrium point defects, such as vacancies, anti-site defects
and interstitials (Fecht, 1992). It is known that -irradiation lowers the melting point of pure metals by an amount that
is proportional to the dose, and thus to the number of point defects generated (Cahn, 1986). The number of vacancies
corresponds to the number of broken bonds in a lattice and it is proportional to the number of atoms in high-energy tail
of the Maxwell distribution, n
vac
exp(H
v
/k
B
T); H
v
is the heat for the vacancy formation. The increase in vacancy
concentration c
v
from 10
3
(crystal at T
m
) to 10% leads to solidliquid transformation. This transition could be suppressed
(or delayed) in the absence of vacancies.
Vacancy formation directly connected to the presence of the energetic atoms from the high-energy tail of the Maxwell
distribution. It is known that heat of formation for a vacancy is related to the equilibrium melting point T
m
by the universal
relationship, H
vac
= 9.28k
B
T
m
= 8 10
4
T
m
(eV) established for a wide range of metals (Fecht, 1992). The total energy
density of vacancies estimates as follows:
E
v
H
v
n
vac
AH
v
exp
_
H
v
k
B
T
_
.
Then the entropy density change due to presence of vacancies can be calculated by standard formula, S
vac
=
_
T
m
T
dE
v
/T.
However, the numerical coefficient should be extracted from experiments. Fecht (1992) calculated the change in the
entropy per atomarising fromthe presence of vacancies with the proportionality coefficient found fromthe measurements,
S
vac
2k
B
. It is clear, then, that lattice heating and thermal defects provide key contributions to the critical entropy.
The critical entropy for Al is S
crit
= 6.41k
B
and this value (Fecht and Johnson, 1988; Tallon, 1989) is reached at the
temperature T
vs
m
1.38T
m
. If the vacancy contribution is neglected then the critical temperature corresponding to the
entropy catastrophe in Al dramatically increases to T
s
m
3T
m
(Fecht and Johnson, 1988). The critical entropy for the majority
of metals is in a range S
crit
57k
B
. This value can be considered as a criterion for establishing the threshold of disordering
in non-equilibrium conditions.
4.3. Transformations induced by ultra-fast heating
We would like to follow the temporal evolution of the solid-to-liquid transformation on microscopic time and space
scales fromthe onset of transitionuntil its completion. We shall consider the contributionof all processes insuccessionat the
consequent time stages: the laser energy absorption, electrons heating, transfer of energy to the lattice and establishment of
the full distribution function. The fast, in a fewfemtoseconds, establishment of local quasi-equilibriumenergy distributions
in electron and lattice sub-systems is a key event, which allows one to apply statistical thermodynamics for quantitative
characterisation of the disordering process. The measure of orderdisorder is time-dependent entropy.
In what follows, we first define the material conditions in the fs-laser excited solid at the absorbed energy density
comparable to and higher than the equilibrium enthalpy of melting. Then the contributions from the different processes
into the entropy changes are considered.
4.3.1. Distribution functions in a swiftly heated solid
(a) Maximum electron and lattice temperature
Direct estimate of the maximum electron and lattice temperature can be made on the basis of energy conservation in
accord with the calculations presented in Section 2. The electron temperature reaches its maximum at the end of the pulse
when all of the absorbed energy is confined solely in electrons E
el
= C
e
n
e
T
e
= 2AF(t
p
)/l
s
. Here F (J/cm
2
), A, l
s
, C
e
, n
e
, T
e
are
the laser fluence, the absorption coefficient, absorption depth, electrons heat capacity, number density and temperature
respectively. We consider metals, thus the electron temperature is calculated as that for a degenerated electron gas,
T
e,max
= [4
F
AF(t
p
)/
2
n
e
l
s
]
1/2
. The maximumlattice temperature is reached at the moment of electronlattice temperature
equilibration and, taking into account the fact that T
e,max

F
is expressed in the form, T
Lm
2AF(t
p
)/C
L
n
a
l
s
. Let us
consider, for example, the ultra-fast heating of Aluminium. At the deposited energy density equal to the enthalpy of
melting the maximum electron temperature in Al (
F
= 11.63 eV; n
e
= 1.806 10
23
cm
3
; T
m
= 933.47 K) reaches
T
e,max
= [2
F
H
f
/
2
n
e
]
1/2
= 0.23 eV = 2.9 T
melt
. Note that the electrons overheating is typical for all metals where the
enthalpy of melting is in the order of kJ/cm
3
and higher.
(b) Applicability of statistical thermodynamics for description of ultra-fast transformations
In order to apply the notion of statistical thermodynamics for description of fast transformations in the material,
two conditions should be fulfilled. First, the distribution function that defines the thermodynamic parameters such as
temperature, pressure, entropy etc. should be set up. Second, temperature, entropy and time scale of the processes should
satisfy conditions when quantum fluctuations are negligible.
Let us recollect now the times for setting up the main parts of electron and lattice distribution (see for details Sections 2
and 2.5.1). Time for establishing the equilibrium temperature in the electron sub-system is, t
ee

1
ee

1
pe

F
/
e
, (see
Eq. (2.22), Sections 2.4.1 and 2). In Aluminium heated by a fs-pulse up to the enthalpy of melting (
F
/
e

= 50.56,
pe
=
1.97 10
16
s
1
), the electron temperature establishes in 2.6 fs.
Likewise the time for setting up the main part of the Boltzmann distribution in Al swiftly heated fromroomtemperature
to the melting temperature establishes in the time range of several femtoseconds. One can see that the time scales for
establishing statistical distributions in other fast-excited metals are also in a few femtosecond ranges.
The second condition for application of thermodynamics is that the entropy fluctuation during the process in question
should exceed the quantum limit (Landau and Lifshitz, 1980b): S

h/(tT). One can see that this condition holds for
melting solids by 100 fs pulses. Therefore statistical thermodynamics is a legitimate tool for description the time evolution
for the state of a solid excited by ultra-short laser for time periods longer than 100 fs. We emphasise that the setting up of the
local equilibrium means that the majority of particles (80%90%) in each sub-system reach the average energy equal to the
sub-system temperature. However, the major part of the absorbed energy is confined in the electrons while lattice remains
cold. During much longer time of the electron-to-lattice energy transfer the distribution functions adjust adiabatically to
the changing temperatures.
The time for the energy transfer from the electrons to the lattice is determined as the inverse electronphonon energy
transfer rate (see Section 2). The energy and momentumexchange rates for Al recovered fromthe optical experiments were
established in Section 2. Using these data one finds that electrons transfer their energy to the lattice in Al laser-excited to
the melting temperature, T
m
= 0.08 eV, during 25 fs. The thermal point defects play essential role in a solid disordering as it
is known from the equilibrium studies (McCarty et al., 2001). Therefore, the next step is to determine the time required for
generation of the sufficient number of the thermal defects and for triggering a phase transformation following the ultra-fast
excitation.
4.3.2. Generation of thermal defects driven by ultra-fast heating
The process of equilibration takes a few femtoseconds. Many more consecutive collisions of atoms with average energy
are needed to create a particle with an energy significantly exceeding the average. Therefore, the time for building up
the high-energy tail of the Maxwell distribution is much longer than that for establishing the central, main part of the
distribution function (see Section 2, Section 2.5.3). In equilibrium the concentration of point defects at the melting point
should reach 7%10% (Fecht, 1992) in order to trigger the catastrophic disordering. The energy necessary for formation of
a single defect in a metal is H
vac
= 9.28k
B
T
m
= 8 10
4
T
m
(eV). In equilibrium there are no time or energy constraints
imposed on the melting solid.
However, one can easily see that the energy constraints impose the definite conditions on the ultra-fast melting.
Indeed, in order to trigger the instantaneous melting 10% of defects, with the energy 9.28k
B
T
m
per atom spent for the
defect formation, and the rest 90% atoms at the melting temperature should be instantaneously created. That means the
average energy per atom of
abs
= 0.928k
B
T
m
+ 0.9k
B
T
m
= 1.828k
B
T
m
should be instantaneously deposited into a solid.
But depositing such energy means that the average temperature should be 1.8T
m
. Thus the energy considerations only
show that the instantaneous melting needs the deposited energy 1.8 bigger the enthalpy of melting. The vacancy (defects)
concentration is proportional to the concentration of high-energy atoms in the tail at
at
H
v
(Fecht and Johnson, 1988),
c
vac
= Ac
H
v
T
m
.
In equilibrium the proportionality coefficient was recovered from the experiments (Fecht, 1992). One can see that
concentration of atoms in the high-energy tail, H
v
, is a strong function of temperature:
c
H
v
T
m
=
n( H
v
)
n
a
1.13
_
H
v
T
_
1/2
exp
_
H
v
T
_
. (4.8)
The vacancy concentration reaches 9% at the superheating 3 times over the melting point in a qualitative agreement
with the energy considerations (see Table 6.2). However, at such level of superheating the entropy change equals to the
catastrophe value. Therefore, in non-equilibrium conditions one may expect that the fast melting should occur in strongly
overheated solid without the defect contribution if the electron-to-lattice energy transfer occurs faster than the defect
generation (Table 4.2).
Table 4.2
Concentration of energetic atoms in the tail as function of
the temperature.
c
v
(%) 0.032 0.58 2.35 5.3 9 16.92
T/T
m
1 1.5 2 2.5 3 4
Time-dependent generation of the thermal point defect proceeds in two interconnected steps with the different time
scales. At the first step the atoms in the process of diffusion along the energy axis reach the energy necessary for defect
formation,
at
H
vac
. This is the time for the high-energy tail of the distribution to be generated (see (2.63)):
t
tail
(H
vac
) 0.85t
main
(T)
T
H
vac
exp
_
H
vac
T
_
. (4.9)
However, inorder to fill the high-energy tail particles shouldproceedalong the energy axis further to the point
at
mH
vac
where the number of particles in the tail H
vac

at
mH
vac
will be at least 90% from the total number. For example at
T = 3T
m
the number of atoms at energy in a range H
vac

at
2H
vac
constitutes more than 90% of the total number in
the tail. Thus, we define the tail filling time t
tft
as the time when a particle with the energy of 2H
vac
was generated:
t
tft
0.85t
main
(T)
T
2H
vac
exp
_
2H
vac
T
_
. (4.10)
Time for establishing the main part of the distribution function in a lattice can be presented with sufficient accuracy as
t
main
(T)

h/T. Now the defects (vacancies) formation time of Eq. (4.10) for metals can be presented as function of the
overheating = T/T
m
in the simple form:
t
def
t
tft
4.58 10
2

h
T
m
exp
_
18.56
_
. (4.11)
Generation of the thermal point defects at the outermost atomic surface layer might be more efficient than in the bulk
because the outermost atoms are more loosely bounded. This puts the additional argument to the reasoning that the melting
initiates at the surface.
4.3.3. Entropy changes produced by electron excitation
The main parts of the distribution functions (temperatures) in electrons and lattice sub-systems are established early in
the pulse time (1020 fs). Therefore, one can apply statistical thermodynamics at the stage when the electron temperature
significantly exceeds the lattice temperature and the energy exchange between species proceeds with slower pace than
setting up the distribution in each sub-system. Thus disordering can be quantified by the entropy changes, which is the sum
of contribution from the different processes during the adiabatically slow adjustment of the distribution functions to the
temperature equilibration. The electron contribution into the entropy change is threefold. First contribution comes fromthe
electron temperature rise. Two other contributions are due to non-homogeneous electron heating; they are specific to the
swift and strong electronic excitation.
(a) Entropy rise due to strong electrons heating
The rapid electron temperature rise during the ultrafast laser pulse results in an isochoric (constant volume) increase
in the entropy of degenerate electrons whilst the lattice remains cold, S
e
(T
e
) =
2
(T
e
T
0
)/2
F
. The maximum entropy
increase coincides with the maximum electron temperature. Let us, for example, consider a swift overheating of electrons
in Al to a temperature ten times of melting point to T
e
= 10T
m
= 9934 K (0.804 eV). In accord with the above formula
it results in an entropy increase of 0.34k
B
, about 20 times less than the critical entropy value. The physical reason behind
this effect is clear: even at strong overheating the electron temperature remains 14.5 times less than the Fermi energy for
conductivity electrons in Al ensuring low electron heat capacity and respective low entropy change.
(b) Modification of inter-atomic potential and atomic displacement induced by excited electrons
Swift electronic excitation modifies the inter-atomic potential while atoms remain cold (see Section 3.4.3(c)). The
increased atomic repulsion (reduced attraction) effectively reduces the binding energy and increases the inter-atomic
distance. The binding energy is in the range of 25 eV for the majority of simple solids. The electron temperature is much
lower than the binding energy, k
B
T
e

b
, even for strong superheating over the melting point. Hence, the ratio of electronic
temperature to the binding energy can be used as a small parameter for estimations of the modification in the inter-atomic
potential. Changes in the binding energy and the increase in the inter-atomic distance are calculated by expanding the
interatomic potential in the Taylor series in respect to this small parameter. The atomic displacement caused by the rapid
electron excitation to temperature T
e
, can be expressed through the change in the binding energy,
b
k
B
T
e
, and the
unperturbed gradient, , of the attractive part of the potential in accord with Eq. (3.26), d k
B
T
e
/
b,0
. For example,
the energy density of kJ/cm
3
(the enthalpy of melting, H
f
) swiftly deposited in Aluminium (n
e
= 1.86 10
23
cm
3
;
F
= 11.63 eV) elevates the electron temperature to T
e
(2H
f
F
/
2
n
e
)
1/2
= 0.28 eV. The corresponding change in the
Fig. 4.1. The dependence of the electron-to-lattice energy transfer time (dashed lines) and the high-energy tail filling time on the absorbed energy density
(lattice temperature) relative to the equilibrium enthalpy of melting, for aluminium and bismuth.
inter-atomic spacing is d = 0.01 . The relative change in the lattice volume is V/V 3d/d = 0.03. Parameters of
the simplified inter-atomic potential for Al are the following: d
0
= 2.55 ;
b
= 3.065 eV; / = 3.29; = 0.93
1
,
(Abell, 1985, see Appendix III).
The swift laser heating of electrons is essentially a non-homogeneous effect: the space scale of electron
temperature/pressure gradient equals to half of the skin depth, l
s
/2. Therefore the displacement caused by the electronic
gradient force, F
el
= T
e
2T
e
/l
s
, is added to the local effect of the increase in the inter-atomic spacing due to modification
of the inter-atomic potential discussed above. It is easy to calculate that in the conditions near the melting point the non-
homogeneous displacement approximately doubles the homogeneous displacement as a result of potential modification
giving the relative expansion of V/V 3d/d 0.1. The characteristic time scale of the process is the time required for
the electron temperature gradient smoothing across the skin depth by heat conduction, which is around a picoseconds. The
changes inthe lattice entropy due to the isothermal volume changes are calculated inaccord withEq. (4.5) S
exp
k
B
V/V.
Then the increase in the entropy due to atomic displacements induced by excited electrons is S
exp
0.20.4k
B
. The
increase of total entropy induced solely by the excited electrons is the sum of the contributions from the electrons heating
and electrons induced atomic displacement, it is in a range 0.70.8k
B
for strong overheating. This constitutes one tenth of the
critical entropy value required for melting at the deposited energy density seven times exceeding the enthalpy of melting.
Thus, the entropy approach clearly indicates that electronic excitation alone does not produce a specific disordering; it leads
to the isothermal volume change similar to that as the thermal expansion does.
Therefore, ultra-fast disordering of the lattice (ultra-fast melting) can only occur either due to a strong overheating, or
within a time period much longer than electronlattice equilibration time. Ultra-fast melting always occur at the thermal stage
when the statistical distribution and thermodynamic parameters, namely temperature, entropy, and pressure, are all set up
and fully characterise the transient state of matter undergoing the swiftly excited phase transition.
4.3.4. The onset of ultra-fast disordering: entropy catastrophe by superheating of the lattice
As it follows fromthe previous section, the sole electron contribution into the total entropy rise is only a small fraction of
the catastrophic level. That holds even for a very large energy deposition several times higher than the enthalpy of melting.
One may expect that contributions from the lattice heating and thermal point defects generation could be as essential for
ultra-fast disordering as it is in equilibrium.
A comparison of tail growing time of Eq. (4.11) with the electron-to-lattice energy transfer time, t
en
eph
[
en
eph
(T)]
1
=
[
en
eph
(T
room
)]
1
(T
room
/T
L
)
2
, shows that lattice heating occurs much faster than the filling up of the high-energy tail of
the distribution (see Eqs. (2.55)(2.57), Section 2; Table 4.3 and Fig. 4.1). Therefore, the thermal defects and associated
disordering are absent over the period of time much longer than that for electronphonon temperature equilibration.
Thermal defect formation may be even more suppressed if the cooling of a laser-excited layer by the electron heat diffusion
is significant. By this reason one may expect in non-equilibrium conditions that fast melting should commence in strongly
overheated solid before the defect contribution becomes significant. The superheating level values follows from the energy
considerations and from the comparison of the tail filling time and energy transfer time (Table 4.3 and Fig. 4.1).
The superheating of simple solids to T
L
3T
m
ensures that the entropy change solely from the lattice heating S
lattice
=
C
p
ln{3T
m
/T
room
} reaches the catastrophe value. For example, for Aluminium swiftly superheated to T
L
3T
m
the entropy
increase due to the lattice heating constitutes 6.98k
B
comparing to the catastrophe value of 6.3k
B
(Fecht and Johnson, 1988).
Table 4.3
Defects formation time (tail filling time Eq. (4.11)) and electron-to-lattice energy transfer time as functions of the
overheating for Al (I
0
= 5.86 eV;
b
= 3.065 eV;
F
= 11.63 eV; T
m
= k
B
933.5 K = 0.08 eV), c
v
is the concentration of
energetic atoms in the high-energy tail.
= T
L
/T
m
1 1.5 2 2.5 3 3.5 4
t
def
(fs) = 37.7 exp
_
18.56
_
3.8 s 8.9 ns 400 ps 63 ps 18 ps 7.6 ps 3.9 ps
t
en
eph
(fs) = 25.2
_
Tm
T
L
_
2
; (fs) 25.2 11.2 6.28 4.03 2.8 2.1 1.6
c
v
(%) 0.032 0.58 2.35 5.3 9 13 16.92
4.3.5. Summary: sequence of events before the onset of disordering
Let us now summarise the main processes leading to the onset of transformation. Electrons are heated to the maximum
temperature at the end of the laser pulse. Statistical distributions within the electron and lattice sub-systems are established
in 1020 fs, this allows the use of averaged thermodynamics parameters such as temperature, pressure and entropy, the
ultimate characteristic of disorder in a system, for description the state of each sub-system. Nevertheless, the crystal as a
whole remains inthe non-equilibriumstate after setting upthe local statistical distributions withthe separate temperatures:
the electron temperature still remains much higher than that for the lattice. Thus disordering and temperature equilibration
are going along simultaneously. It was demonstrated in the previous sections that the entropy changes produced solely
by strongly excited electrons are insufficient for the lattice disordering. The excited electrons produce significant atomic
displacement comparable to that of thermal expansion, and, as the thermal expansion does, the entropy changes are small
in comparison to the catastrophic limit.
The salient feature of the ultra-fast transformation is that the time for setting up the high-energy tail of the atomic
distribution, containing approximately 10% of the atoms, appears to be longest of all relaxation times. The high-energy
tail is responsible for the generation of thermal point defects, which contribute half of the entropy change necessary for
disordering. In the absence of the point defects the main contribution to the disordering comes fromthe lattice heating. The
critical (catastrophic) value of the entropy is achieved if a lattice is superheated to the temperature approximately tree times
of the equilibrium melting point. At the same time the number density of thermal point defects reaches several percents,
necessary for triggering a catastrophic disordering (Gamaly, 2010).
The onset of disordering by the lattice superheating could be realised if the energy losses due to heat conduction are
negligible during the lattice heating by the energy transfer from electrons. Therefore the energy-lattice transfer time,
t
en
eph

F
/(2
mom
eph
k
B
T), should be shorter than the cooling time, t
cool
= 3l
2
skin
mom
eph
/v
2
F
allowing completion of the phase
transformation.
4.4. Phase transformation: heterogeneous and homogeneous nucleation
The transformations of the material induced by a laser pulse in less than 100 fs time takes place in a surface layer
with a thickness that comprises approximately hundred of atomic layers. Therefore, the implications of the space and time
constraints on the kinetics of the phase transformation in such conditions are essential.
A crystal lattice heated above its melting point up to the entropy catastrophe limit passes into explosively unstable state.
Thermal point defects, or vacancies, are microscopic seeds for formation of the macroscopic nuclei of new phase in the bulk
of a skin layer and at the skinvacuuminterface. Vacancies could condense at free surfaces, interfaces, and grain boundaries
forming small clusters with the surface energy being proportional to the heat of formation of a vacancy (Fecht, 1992). The
kinetics of melting depends strongly on the development of these small liquid-like clusters evolving in the crystalline phase
and breaking the symmetry of the crystalline lattice. There are two processes that can drive the thermal phase transition in
a skin layer having a free surface. The temperature has a maximum near the free surface; therefore the transformation into
the new phase there is energetically favoured driving then the process of heterogeneous nucleation.
Alternatively, as shown by Landau and Lifshitz (1980b), the formation of liquid seeds by condensation of the thermal
defects within an internally heated bulk crystal could be also a driving force for the transformation, if the inside temperature
exceeds the melting point. This is the process of the homogeneous nucleation. Small seeds of the new molten phase are
created in the overheated layer due to the lattice fluctuations generated by the thermal defects. For a closed system near
the melting point the probability of such a fluctuation is (Landau and Lifshitz, 1980b):
w exp
_
S
fusion
k
B
_
.
Here S
fusion
denotes the entropy difference between solid and liquid. This probability is strongly enhanced by increase in
the vacancy concentration because while S
fusion
> 0, the vacancy concentration increases to the critical value of 7.7% at
the temperature close to the melting point (Fecht, 1992).
These seeds are, however, unstable structures because the formation of an interface between the two phases requires
extra energy to overcome the surface tension at that interface. Therefore the homogeneous nucleation needs additional
energy expenses to proceed in comparison to the heterogeneous nucleation. There is a critical radius for a seed of the
molten phase defined from the condition that a surface energy is balanced by the internal thermal energy of a seed. The
critical radius of a seed is related to the temperature of the overheated layer, T, as (Landau and Lifshitz, 1980b):
r
cr
=
2
P
P

2
n
a
(T T
m
)
; (4.12)
where is the surface tension between the crystal and liquid, P
is the transient pressure in the skin layer, and P is the

pressure corresponding to the melting temperature. The difference in the densities of liquid and solid is neglected above
for simplicity. Hence seeds with a size less than a critical value will decay back into the initial phase, whilst seeds with
a size exceeding this critical radius will grow rapidly driving the transformation of the bulk into the molten phase. A
clear consequence of Eq. (4.12) is that the critical radius increases rapidly as the temperature drops towards the melting
temperature due to cooling by heat conduction. As a result, when the critical seed radius compares to the thickness
of the heated layer the phase transition may be strongly inhibited or rather terminated. Let us consider, for example,
formation of critical seeds in Aluminium (the surface tension at the solidliquid interface is 1 J/m
2
, melting point
T
m
= 933.5 K; n
a
= 6.02 10
22
cm
3
. Then, one obtains that the critical seed with a diameter equal to the skin depth of
Aluminium of 13.1 nm for 800 nm can be formed at T = 1.4 T
m
. The probability of such a seed formation for the above
conditions is extremely low, w exp{4r
2
cr
/T} 1. One can conclude that the homogeneous nucleation in Aluminium
layer terminates at the temperature 2T
m
.
Superheating experiments indicate that melting is usually a process driven by heterogeneous nucleation that is
energetically more favourable. The phase transition starts fromthe targetvacuuminterface due to the energy reason: there
is no energetic barrier for conversion solid to liquid phase (no energy should be spent for the solidliquid surface tension
compensation) at the solidvacuum interface. Thus, the melting wave starts at the outer surface at the moment when the
entropy catastrophe conditions for the onset of melting were achieved. It is reasonable to suggest that the thermal defects
condensation occurs with a rate proportional to the local speed of sound. Then the phase transition wave propagates inside
the crystal with the speed of sound. Hence, the heterogeneous melting time for the layer with thickness of l
s
estimates as:
t
hetero
l
skin
/v
sound
. (4.13)
However, the temperature in the skin layer in condition when the entropy catastrophe limit is already achieved exceeds
the melting point by 3 times. Therefore the processes of heterogeneous melting wave propagating inside the layer
from the outer surface, the homogeneous melting inside the crystal, and the cooling by the heat conduction are going
simultaneously (Rethfeld et al., 2002). Interplay between these processes defines the time for the transition to be completed
and life-time of new phase, and then reverse transformation before final cooling to the ambient temperature.
Let us consider for example the ultra-fast melting of Aluminium(l
s
= 13.1 nm; heat diffusioncoefficient, D = 0.979 cm
2
/
s; v
sound
= 5 10
5
cm/sall data are at room temperature). The wave of the heterogeneous melting propagates through
the skin layer in 2.6 ps (under the assumption that the catastrophe condition holds everywhere in the skin during the
propagation) while the cooling time for the layer is 1.75 ps; during this time temperature drops to 0.7 of the maximum.
Thus it is unlikely at the superheating to T
max
= 3T
melt
the fast Aluminium melting would be completed during the couple
of picoseconds.
It is clear that competition between the homogeneous and heterogeneous nucleation and cooling is a common feature
of the ultra-fast melting in different materials defining the time and degree of the transformation of the material.
4.5. Transient state of matter created by the ultra-fast excitation
The ultra-short laser-excited solid can be temporarily transformed into the new phase state under the action of laser
fluence well below the ablation threshold. The heated material can exist in this new state for some time; then cooling
brings the laser-affected solid back to the ambient conditions. A transformation can bring short-lived changes in optical and
material properties. These changes include the modification of the phonon spectrum, variation in the number of free carriers
in the conduction band, vanishing of the shear modulus, coherent and chaotic atomic displacements, and finally the loss of
the long-range order. We are not discussing here the permanent (meta-stable) modifications of the dielectric function in
the photo-refractive materials. These effects are discussed below in Section 6.
The probing of the excited layer by single and double simultaneous optical beams allows measuring the time-dependent
reflectivity of the probe beam with high temporal resolution comparable to the pulse duration of the probe beam
(40100 fs). Changes in the free carriers number, average phonons frequency and phonons lifetime along with the rate
of transformation into liquid are all imprinted into the dielectric function and, in principle, can be recovered from the
reflectivity measurements as it was discussed in Section 3.
In principle the time-dependent diffraction pattern and diffracted intensity of short X-ray and electron beam probes
allows directly observe and identify the transient structural modifications, change in the inter-atomic spacing and
disordering in the laser-excited layer. However, powerful lasermatter interaction induces phase transformation in non-
equilibrium conditions and with significant spatial non-homogeneity as we discussed in Section 3 (see Section 3.6.2(a)
and (b)). Therefore, the interpretation of the experiments with the ultra-fast X-ray and electron beam probes should
Fig. 4.2. Dependence of the absorbedenergy density onthe absorbedlaser fluence for aluminiumandfusedsilica. The ablationthresholdpoint 0.35J/cm
2
is marked on the Al curve. The solid line part of the fused silica dependence compiled.
Source: Taken from Temnov et al. (2006).
take into account non-equilibrium conditions and spatial non-homogeneity. Also application the classical (equilibrium)
DebyeWaller factor for the interpretationof fall-off the electrons (X-ray) diffractionintensity innon-equilibriumconditions
is questionable (see 3.6.2(b)) for details).
There is no obvious way of distinguishing the decrease in the diffraction beamintensity caused by the non-homogeneity
from the disordering during melting. Due to above-discussed non-homogeneity the Bragg conditions should be different at
the different spots across the focus and inside the skin layer. One may expect the increase of the diffusive background in the
diffracted beam intensity similar to that of the inelastic scattering.
4.6. Ultra-fast melting of metals and dielectrics
First, let us recollect the similarities and differences of melting characteristics for the different materials in equilibrium.
For the majority of simple solids (for example, for good metals, see Table 4.1) the density in a liquid state is lower than
that for the solid. For these group the melting entropy lies in a range 1.11.4k
B
while for the materials, which have a liquid
density higher than that for a solid state (H
2
O, Gallium, Bismuth, InSb), the melting entropy ranges in 2.33k
B
. It is worth
noting that the difference is small in comparison to the catastrophic value for disordering of 67k
B
. On the other hand the
energy characteristics of melting, the ratio of melting temperature (in energy units) to the binding energy for the majority of
elements fromthe periodic chart constitutes k
B
T
m
/
b
3%. The enthalpy of fusion for broad range of metals and dielectrics
varies in a range from0.335 kJ/cm
3
for ice and 0.5 kJ/cm
3
for Bi at lowend, and up to 0.975 kJ/cm
3
for InSb and 1.86 kJ/cm
3
for Copper at the high enthalpy limit. Therefore in equilibrium, when the heating conditions are comparable, the process
of melting of the majority of solids proceeds in a similar way. Consequently, one cannot expect dramatic differences in the
ultra-fast melting of metals and dielectrics by the swift laser action if the same absorbed energy density deposited at the
same time.
The laser interaction with metals and dielectrics however is drastically different, and this difference results in the
dissimilarities of the ultra-fast melting of diverse solids. Indeed, the metals are well absorbing even at the low laser
intensity while most of the dielectrics (especially the wide band gap dielectrics) are transparent up to the optical breakdown
threshold. Optical breakdown for dielectrics lies in the intensity range of a few TW/cm
2
(Arnold and Cartier, 1992) that
corresponds to fluence of 0.1 J/cm
2
for 100 fs laser pulse. At the optical breakdown up to 10% of valence electrons are
transferred to the conduction band that results in a strong increase of intra-band absorption with minor effect on other
parameters. This difference can be clearly seen from comparison of the absorbed energy density as function of laser fluence
for metal (Al) and for silica (Temnov et al., 2006) (see Fig. 4.2).
The absorbed energy density for Al in Fig. 4.2 plotted in accord to formula E
Al
abs
(kJ/cm
3
) 91.4 F (J/cm
2
), which
gives correct value for the experimentally observed ablation threshold, 0.35 J/cm
2
(Gamaly et al., 2005). The corresponding
formula for silica is chosen in the form E
abs,SiO
2
(kJ/cm
3
) 17.5 (F (J/cm
2
))
7
closely reproduces the ablation threshold
predictedfor silica at 800 nm, 1.84 J/cm
2
(Gamaly et al., 2002) and1.5 J/cm
2
damage thresholdas well as the experimentally
observed dependence of the electron number density on fluence from (Temnov et al., 2006). One can easily see that the
ablation thresholds for Al and for silica differ by 5.3 times, while the absorbed energy density at that threshold in silica is
ten times larger than in Al, thus signifying the dramatically different interaction modes for metals and dielectrics.
4.7. Comparison to experiments
Let us now discuss the experiments where the fast laser-excited solids were probed by the short optical, X-ray and
electron beam probes with the time resolutions better than the time for electron and lattice temperature equilibration.
We begin the analysis of the experiments defining first the conditions in laser-excited solid depending on the main
lasermatter interaction parameters: taking the experimentally determined absorbed energy density allows directly
obtaining the maximum electron and lattice temperature from the energy conservation and comparing to the melting
point find the level of superheating. Then times for the establishment of the statistical distributions, for the electronlattice
temperature equilibration and the point defects formation can be found in a frame of general approach described in the
previous sections. Superheating level can be related to the value of the critical entropy and therefore one can judge on the
commencement of disordering. Then comparing the measured time of transition from solid to melt, where it is available,
to the above characteristic times one could comment on the nature of the observed transformation and the phase state
achieved.
4.7.1. Superheating of ice
High purity defect-free HDO:D
2
O ice was heated by 0.9 ps, 1.85 m, 3 J pulse to the temperature 2932 K by exciting
OH-stretching modes (Iglev et al., 2006). The chemical purity of the ice was 10
4
, and the initial defect concentration
was very low. The OH and OD stretching vibrational modes are known to be temperature and structure sensitive probes
for hydrogen bonding. The strength of hydrogen bonds is significantly smaller than that of covalent and metallic bonds.
Therefore formation and migration of structural defects in the hydrogen-bonded network of ice is enhanced. Large
topological defects have a lifetime of 0.5 ns.
A comparison of time-dependent infrared spectra at different temperatures allows one to distinguish between the local
ice and the water structures with time resolution of a fewps. It was found by Iglev et al. (2006), that HDO:D
2
Oice maintained
its crystalline structure during 250 ps at 293 2 K, being superheated 20 K over the melting point.
Superheating in these experiments is low in a sense that the lattice thermal contribution into the critical entropy is
small. Thus one may expect that the observed delay with the melting relate to the time for the thermal point defect
formation. There are no excited electrons at such low overheating. The statistical distribution in a lattice establishes in
t
distr

1
ph
30100 fs during the pulse time (phonon frequency for ice,
ph
= (13) 10
13
s
1
). Therefore superheated
ice is in a state of thermal equilibriumwhen the phase transformation commences and proceeds. The cooling time of micron
thick ice sample is longer than a microsecond (thermal diffusion coefficient D = (1.3358.43)10
3
cm
2
/s (James, 1968));
therefore the heat losses during the observation and transformation time were negligible.
The energy for formation of point defect in ice is unknown to the best of our knowledge. It is reasonable to suggest that
the energy for the vacancy formation is close to the strength of hydrogen bonds that is in a range, H
d
0.10.22 eV/atom.
The vacancy formation time calculated with the help of Eq. (4.11) is within 120500 ps range (taking 2H
d
/k
B
T = 11 and
t
distr
t
main
). Thus the theory prediction that the delay in melting of the overheated ice relates to the time necessary for
formation of point defects is in qualitative agreement with observations, taking into account the uncertainties with the
energy of defect formation. The catastrophic disordering commences when the number of defects reaches 7%9% level.
It is reasonable to suggest that the melting starts from the outer free boundary. The heterogeneous melting wave moves
with the speed of sound fromthe outer surface. The melting is completed when the wave reaches the deep end of the micron
thick layer of ice. The melting completion time comprises 320 ps (taking the speed of sound in ice of 3.125 10
5
cm/s)
that is in qualitative agreement with observation and with the above considerations of defects formation. Summing up,
the melting of the slightly superheated ice occurs in the conditions when the main parts of statistical distributions were
established during the laser pulse. The time delay of the observed disordering relates to the accumulation of the necessary
amount of the defects and to the completion of the transformation in a micron thick layer.
4.7.2. Superheating of Gallium
Transient transformations of Gallium excited by 150 fs laser pulse (775 nm, 1.43.85 mJ/cm
2
) have been studied
experimentally using two simultaneous fs-optical probes with time resolution of 200 fs (Uteza et al., 2004). Two
simultaneous probes allowrecovering the real andimaginary parts of the transient dielectric function. The dielectric function
of solid and liquid Gallium is well described by the Drude-like form. Therefore double simultaneous probes allow recovery
time-dependent electronphonon momentum exchange rate and plasma frequency thus giving the deeper insight into the
phase transformation history. In Gallium the differences in optical properties of solid and liquid are significant: reflectivity
in solid state equals to 0.633 comparing to 0.8 in melt. Hence the time-dependent changes in optical properties allow
recovering of the phase transition pace. The deposited energy density exceeded more than two times the equilibrium
enthalpy of melting. The maximum lattice temperature reached 809 K, which is 2.67 times the equilibrium melting
temperature (303 K) during 400 fs. The entropy catastrophe value for Ga is unknown to the best of our knowledge. However,
one may expect that the critical entropy value for the majority of materials is around (67)k
B
and the critical overheating is
around 3T
melt
. The maximum temperature in Ga is close to the critical value, so one can expect the commencement of the
phase transition.
Indeed, it follows from the measured dielectric function (Fig. 4.3) of the excited Ga that the onset of a change towards
the liquid state occurred after the electronlattice temperature equilibration at the end of the pulse. The phase transition
Fig. 4.3. The time-dependent real (open circles) and imaginary (closed circles) parts of the dielectric function of Gallium experiencing the phase
transformation recovered from the reflectivity. In solid state
1
= 7.05;
2
= 12.79; in liquid
1
= 66.00;
2
= 34.93.
Source: After Uteza et al. (2004).
commenced when electrons transferred their energy to the lattice, which was strongly overheated. This is a clear evidence
of thermal nature of the ultra-fast phase transition. The defects formation time in Gallium overheated to = T/T
m
= 2.67
in accord to Eq. (4.11) is 1.3 ps. Thus theory predicts the phase transition commencement in qualitative agreement with
observations. However, the experiments demonstrate that the transient dielectric function of so strongly excited Ga did
not reach the liquid state values 20 ps after excitation, and only 60% of material can be attributed to the liquid state (Uteza
et al., 2004). Uteza et al. indicated, that the cooling time for 34 nm skin layer in Gallium equals to 3.75 ps. Thus, phase
transition begins in competition with the cooling, which slows down the defects formation rate and the transition pace.
Indeed, temperature inthe skinlayer decreases 2.5times during 20ps of observation. The soundwave (that is associatedwith
the heterogeneous melting wave) reaches the far end of the skin layer in 12.4 ps (v
sound
= 2.74 10
5
cm/s) when the heat
conduction already reduces the average temperature about 2 times. The theory, therefore, suggests that the transformation
in the material cannot be completed during 20 ps of observations.
It is important to note that ice and Galliumare materials for which the density of the liquid exceeds that of the crystalline
solid (including also Bi, Sb, and KI). Unusual physical properties of these materials have been attributed to their very open
crystalline structures. To the best of the authors knowledge the critical entropy values for Ga and ice are unknown. However,
as described above, the entropy contributions due to alterations of volume solely due to the electron excitation are minor
in comparison with those due to lattice heating and defects, thereby legitimising the critical entropy criterion for a physical
description of melting in these materials.
4.7.3. Superheating of Aluminium
Aluminium is representative of the majority of materials whose liquid (melt) density is lower than that for the solid
state. The transformation Al to the liquid state has been widely studied theoretically and experimentally under equilibrium
conditions and with ultra-fast excitation. First experiments where the transient state of laser-excited aluminiumwas probed
by time-resolved diffraction of the electron beam were performed more than 20 years ago (Williamson et al., 1984) but the
resolution was insufficient for definite conclusions.
The structural evolution of 20 nm thick Al layer excited by ultra-short laser (120 fs, 70 mJ/cm
2
, 775 nm) was studied
by time-resolved electron diffraction of 600 fs electron probe pulses in the experiments of Siwick et al. (2003). The
authors recorded the time sequence of diffraction patterns. Then information contained in these patterns was analysed
in terms of average atomic pair correlation function, which contains information about the nearest-neighbour distances
and coordination numbers at each instant of observation. The analysis of diffraction patterns revealed both the loss of long-
range order and the emergence of short-range atomic correlations in a time of 3.5 ps after excitation, thereby providing
direct evidence of transition to liquid phase.
It can be readily shown that Al was strongly overheated in these conditions. Indeed, the equilibrium enthalpy of melting
for Al of 1.07 kJ/cm
3
corresponds to the absorbed laser energy density at the incident fluence of 7 mJ/cm
2
(the absorption
coefficient is 0.13 and the skin depth is 11.3 nm). Thus in the experiments of Siwick et al. (2003) the maximum lattice
temperature to the end of the laser pump pulse was ten times of the equilibrium melting point, 10 T
melt
. At such high
temperature the statistical distributions inelectronandlattice subsystems andelectron-to-lattice temperature equilibration
were established early in the laser-interaction time. Superheating also results in the fast formation of the necessary number
of the thermal defects. Therefore the lattice entropy reached the catastrophic level due to thermal disordering near the end
of the pulse, the lattice transformation commenced at that time and proceeds in a thermal mode (see Fig. 4.1).
The melting started at the vacuumsample interface where the temperature is at its maximumas Siwick et al. suggested.
The wave of heterogeneous nucleation (speed of sound in Al, v
sound
= 5 10
5
cm/s) passes through the skin depth of
11.3 nm in 2.26 ps time completing the transformation in good agreement with the experimental observation. The cooling
time for the skin layer is 1.3 ps (diffusivity in Al D = 0.979 cm
2
/s). Therefore, the temperature in the skin layer decreases in
1.8 times to the end of the observation period of 3.5 ps still being 5 times above the melting point. The transformation pace
should be only slightly affected by the cooling.
More recently the phase transformation of Aluminium excited by extremely powerful ultra-short laser pulse
(40 fs, 800 nm, 0.756 J/cm
2
) was followed by measuring the time-dependent reflectivity of two probe optical beams
with 65 fs resolution (Kandyla et al., 2007). The laser fluence in these experiments is almost twice above the ablation
threshold for Aluminium (see Section 5). Therefore the absorbed energy density is 120 times of the equilibrium enthalpy
of melting. Aluminium experiences the melting first, swiftly followed by ablation. At such extreme energy density electron
temperature rises up to 3.26 eV that is higher than the binding energy in Aluminium of 3.065 eV. Therefore the statistical
distributions, electronlattice equilibration and formation of vacancies all occur during the pulse time giving start to the
phase transformation. However, even at this extreme electron excitation the entropy change solely due to electrons is two
times less than the critical value for melting. Therefore the material disordering occurs due to lattice overheating. The
heterogeneous melting wave propagates through the skin depth in 2 ps. It is worth noting that the optical response of a
material builds up on the skin length. As a result, the experimental observation of Kandyla et al. (2007) that superheated Al
attains the optical properties of a liquid in 1.9 ps is the evidence that the phase transformation was completed during this
time in the whole skin layer.
The absorbed energy density in these experiments constitutes, E
abs

= 2AF/l
s
= 1.7410
5
J/cm
3
. Therefore the absorbed
energy per particle after electronlattice equilibration comprises 4.5 eVper atomand electron that is in excess of the binding
energy in Aluminiumof 3.065 eV. Thus the atomic bonds are broken in several atomic layers next to the sampleair interface.
The kinetic energy per ablating atom after the bond breaking equals to 1.45 eV; atomic velocity is 3 10
5
cm/s. It means
that the atoms from the outmost surface layer move out at the distance of 3 nm during 10 ps of observation. Remarkable
result of these experiments is that the reflectivity during the 1.910 ps period after the excitation coincides with that for
liquid Al. Therefore during the measurement time of 10 ps the target remains intact, in a quasi-solid state with already
broken bonds and being disordered. The results of these experiments are another unequivocal demonstration that the phase
transformations of extremely superheated Al are of thermal nature.
4.7.4. Ultra-fast excitation of Bismuth
The transient state of the femtosecond laser excited Bi crystal has been studied with time-resolved fast optical probes
(Zeiger et al., 1992; Misochko et al., 2004; Boschetto et al., 2008a,b; Garl et al., 2008), with X-ray probes (Fritz et al., 2007;
Sokolowski-Tinten et al., 2003; Johnson et al., 2008) and with energetic electron beams (Sciaini et al., 2009; Zhou et al., 2009)
at the deposited energy density below and up to eight times above the equilibrium enthalpy of melting. That corresponds
to the wide range of the pump lasers fluences (0.523) mJ/cm
2
.
Bismuth is an element with quite unusual solid properties. At lowtemperature it is a semi-metal, it acquires the metallic
properties at elevated temperature. The liquid Bismuth density in equilibrium is higher than that for solid. Temperature
dependence of optical properties of bismuth in equilibrium conditions has been measured in the wide temperature range
from the room temperature up to 773 K (200 K over the melting point) (Hodgson, 1962; Smith et al., 1964; Comins, 1972;
Garl, 2008). These data allow one to recover such temperature-dependent material properties as the electronphonon
momentumand the energy transfer rates, the heat diffusion coefficient, the number of electrons in the conduction band and
the Fermi energy. The set of these parameters form a strong basis for the quantitative interpretation of the experimental
results with the ultra-fast excitation of bismuth (see Appendix B).
The electronphonon momentum exchange rate in Bi directly obtained from the optical measurements linearly growth
with temperature in the temperature range, 293773 K, in accord with the theory (see Section 2),
mom
eph
= 210
15
T
L
/T
room
.
Melting point for Bi is 544.6 K; liquid range in equilibrium extends up to 1292.6 K. We adopt, as a first approximation, the
linear temperature dependence andfor higher temperature range innon-equilibriumexperiments. Theoretically established
link between the momentum and energy exchange rates allows obtaining the temperature dependent electronphonon
energy exchange rate (see Appendix B). It is convenient to present the electronphonon energy exchange time as function
of the ratio of lattice temperature to the melting point (overheating ratio) = T/T
m
:
t
en
eph
= (
en
eph
)
1

=
24.1
2
(fs).
Thus, the equilibrium experiments along with the theory suggest that thermalisation following the fast heating occurs
in a few femtoseconds when lattice temperature is around and higher the equilibrium melting point. Therefore, all
transformations induced by the femtosecond laser pulses proceed in the thermal mode when electrons and lattice
temperature are equilibrated.
The electron number density in the conduction band (and the Fermi energy), retrieved from the optical data in
equilibrium, grows from 40% of the total number of the valence electrons to 100%, in direct proportion to temperature.
Thermal diffusivity, D = v
2
F
/3
mom
eph
, recovered from the equilibrium experiments is D = (2.02.89) cm
2
/s. These results
are in good agreement with the recent non-equilibrium measurements from the X-ray reflectivity data of fs-laser excited
bismuth giving D = 2.3 cm
2
/s (Johnson et al., 2008).
Another important process of the ultra-fast transformationis generationof thermal defects. The optical properties of solid
Bismuth above the room temperature and in the melt are well described by the Drude function. Therefore, it is legitimate
to consider the thermal defects formation in Bismuth in the same way as in good metals. Then the defects formation time
in Bismuth as a function of superheating is described by the formula (see (4.11)):
t
def
t
tft
4.58 10
2

h
T
m
exp
_
18.56
_
.
One can see from the above formula and from Fig. 4.1 that in Bi superheated 3 times over the melting point, the number
of defects, sufficient for initiation of phase transformation, can be formed in 300 fs after the pulse. Therefore, the lattice su-
perheating and number of point defects are ensuring the onset of the phase transformation shortly after the end of the pulse.
However, the cooling due to heat conduction could slow down the transformation pace. The cooling time of 30 nm skin
layer is in a range, t
cool
= l
2
skin
/D = 3.13.9 ps. It means that propagation of the phase transition wave is accompanied
by strong cooling that decreases the temperature in at least (t
cool
+ t
complete
/t
cool
)
1/2
2.3 times. Summing up the results
from the equilibrium studies and the theory presented in this section one can see that the completion of melting may occur
during more than 17 ps under assumption that the initial maximumtemperature would be 710 times of the melting point.
Let us now analyse the experiments.
The phase transition can be detected by the change in reflectivity of a probe optical beam if the skin layer for the
wavelength of the beamis transformed into a newphase. The transformation is completed when the heterogeneous melting
wave passes throughthe entire 30nmof excitedskinlayer. The speedof this wave is associatedwitha speedfor condensation
of the thermal defects into the liquid seeds of new phase. The upper limit for the melting wave velocity is a sound speed
in solid Bi of 1.79 10
5
cm/s. The lower limit is the local thermal speed of heated atoms, which equals to 3.6 10
4
cm/s
for Bismuth heated 3 times over the melting point. This gives the span for the time for transformation to be completed in a
range, t
complete
= l
skin
/v
th
1783 ps.
Sokolowski-Tinten et al. (2003) probed the state of 50 nm Bi film excited by 800 nm laser at fluences 6 mJ/cm
2
and
20 mJ/cm
2
by short X-ray beam. The absorbed energy density was respectively 2 and 6.7 times larger than the equilibrium
enthalpy of melting. The measurements were performed at two different Bragg angles 20 (111) and 44 (222) in order to
compare diffraction from different lattice planes. The authors made their conclusions on the transient state of excited Bi on
the basis of the analysis of temporal behaviour of the diffracted beam intensity. At low fluence of 6 mJ/cm
2
the intensity
of the probes from both planes oscillates with 2.12 THz frequency and with gradually decreasing amplitude during 4 ps
of observations. The oscillations were attributed to the excitation of the red-shifted A
1g
phonon mode (2.92 THz in cold
Bismuth). The probe beam intensity diffracted from Bi layer excited at fluence of 20 mJ/cm
2
smoothly decreased without
oscillations (see Figs. 3.9 and 3.10).
The authors attributed the diffracted intensity decrease to the disordering due to ultra-fast melting. They used the
classical (equilibrium) DebyeWaller factor to extract the maximum atomic displacement of 0.04 nm from the diffracted
intensity decrease. However the entropy changes calculated in accord to theory presented earlier in this section with
these displacements, S
exp
= (0.150.24)k
B
, are 25 times less than the critical value indicating the commencement of
disordering. Thus, displacements of this magnitude cannot be an indication of disordering due to non-thermal melting. Most
probably the laser-induced inhomogeneity in the layer causes the probe beamintensity fall-off. The temperature-dependent
relaxation times for Bi clearly show that the electron and lattice temperatures equilibrate in less than picosecond at this
level of superheating thus making any processes after the end of the pulse to proceed in the thermal conditions. Another
important feature of these experiments not allowing the detection of the phase transformation is short observation time of
4 ps in comparison to the completion of transformation time that is longer than 17 ps. Thus conclusions of our analysis are
twofold: first, the transformation is unequivocally of the thermal nature; and, second, the transformation was not completed
and could not be observed in the above experimental conditions due to a rather short observation time.
The state of Bi excited by 800 nm, 150 fs laser with fluences in a range 6.720 mJ/cm
2
, similar to conditions of the work
of Sokolowski-Tinten et al. (2003), was probed with one probe (Boschetto et al., 2008a,b; Garl et al., 2008) and with two
simultaneous optical probes (Rode et al., 2008). The measured time-dependent real and imaginary parts of the dielectric
function of excited Bismuth remained in a state different from that for solid and liquid during 25 ps of observation. The
above analysis indicates that even at highest superheating at 20 mJ/cm
2
(6.7 times of the melting enthalpy) the phase
transformation was not completed because the cooling reduced the average temperature in 2.7 times from the maximum
value thus significantly slowing down the defect formation rate and therefore transformation rate. The transient state of
excited Bi then might be a mixture of solid and liquid phases. However, the existence of some unknown transitional state
cannot be excluded. The optical measurements clearly indicate that the phase transition proceeds in the thermal mode, and
even at conditions of strong superheating it is not completed during 25 ps time due strong cooling.
Johnson et al. (2008) employed grazing-incidence X-ray diffraction to study the structural changes in laser-excited Bi
crystal as function of depth from the surface. The authors recovered the structural changes as a function of depth from the
measurements of the time-resolved evolution of the diffracted X-ray intensity for different angles of incidence. The pump
laser excited Bi with fluences in a range 0.562.24 mJ/cm
3
that is belowthe enthalpy of melting. The author retrieved from
the measurements the carrier diffusion coefficient of 2.3 0.3 cm
2
/s, which is in a good agreement with 2.12.89 cm
2
/s
obtained from the optical measurements of solid and liquid Bi in equilibrium (Hodgson, 1962; Smith et al., 1964; Comins,
1972; Garl et al., 2008; Gamaly and Rode, 2009).
50 nm thick Bi, excited by 70 fs, 800 nm laser with the fluences in a range of 0.33 mJ/cm
2
below and in the vicinity
of the equilibrium melting enthalpy, has been probed with 100 fs, 9 keV X-ray pulses (Fritz et al., 2007). Time-dependent
signal of diffracted X-ray probe was the primary source of the experimental information on the temporal changes in atomic
positions following the excitation. It shouldbe pointedout that the authors didnot use intheir analysis the knownabsorption
coefficient of 0.26 for 800 nm light in Bi. Instead it was assumed that the single electronhole pair was produced per
absorbed photon (that means 1.55 eV energy expense per this act). On this basis the authors calculated, using Density
Functional Theory (DFT) methods, the number of excited electrons in the conduction band, the modification of the inter-
atomic potential and the induced coherent displacement of atoms. Their calculations suggest that upon excitation of 2.5% of
valence electrons bismuth undergoes a structural phase transition into a higher symmetry state, whereas at approximately
2% excitation the barrier between the wells (the Peierls distortion) is lowered sufficiently for the atoms to move in between
both wells. However there is no comparison of their results to the known properties of Bi at room temperature and in
the liquid state. Nevertheless the experimental results and calculations lead the authors to the conclusion that there is no
experimental or DFT evidence of non-thermal melting transition in Bi at this excitation level.
Sciaini et al. (2009) have studied the transient state of the freestanding 30 nm thick bismuth films excited by 200 fs
at the absorbed energy density ranging from two to eight times higher the equilibrium enthalpy of melting. The rate of
structural changes was recovered from analysis of the time-dependent diffracted intensity of 350 fs electron beam. At the
absorbed energy 8 times higher of the equilibrium enthalpy of melting the diffracted intensity reached a stationary level
at 190 fs, which was interpreted by the authors as a moment of non-thermal melting solely by the electron excitation. It
is clear from the preceding analysis that this interpretation is inconsistent with the relaxation times for Bi and previous
studies of Bi at the same level of excitation by different methods. The laser interaction with a freestanding layer in this work
imposes further implications for interpretation of the electron diffraction pattern. The pressure at the front surface of the
film, which is equal to the absorbed energy density, is 35 kBar at 23 mJ/cm
2
, while at the rear surface it is 4 kBar. The
sound velocity in solid Bi is 1.79 10
5
cm/s; thus the front surface of the free standing film shifts towards the laser beam
during the 200 fs pulse by 3.6 while the rear surface expands in the opposite direction. This creates additional source for
the intensity decrease in the diffracted electron beam irrelevant to melting. Note that the temperature smoothing time due
to heat conduction across the skin layer is 3.9 ps, which is much longer than the time for the diffracted intensity fall-off.
Possible inhomogeneous laser intensity distribution of both the pump and the probe beams over the surface induce further
complication into the interpretation of the diffracted beam intensity behaviour.
Let us draw attention to several important issues followed from the above analysis of the experiments with Bismuth
swiftly excited by ultra-fast laser pulses.
Firstly, the recovery of the temperature dependence for the relaxation times of Bi allows performing the quantitative
interpretation of the experiments with the ultra-fast excitation. The number of electrons in the conduction zone of Bi at
the temperature in excess the room temperature ranges from 40% to 100% of the valence electrons. Modification of the
potential surface by the electron excitation occurs in Bi in a way similar to that in other metals such as aluminium (Siwick
et al., 2003) or Gallium (Uteza et al., 2004), and can be related to the increase in the inter-atomic spacing due to directed,
coherent displacement.
Secondly, the temperature dependence of the electronphonon energy exchange time allows unequivocally establish
purely thermal nature of the ultra-fast laser-induced transformations in Bismuth: the electronphonon temperature
setting up and energy equilibration occurs faster than the transformation takes place, that is in accord with the observed
transformations in aluminium (Siwick et al., 2003) and gallium (Uteza et al., 2004).
Thirdly, the laser-induced inhomogeneity in the absorbed energy density and temperature results in the significant
atomic displacements, comparable to the inter-atomic distance, might be a major source for the violation of the Braggs
condition and the diffraction intensity decrease when the electron and X-ray probes were used.
The additional concern in the interpretation of laser-induced melting in Bi relates to the fact that liquid bismuth in
equilibrium conditions is denser than solid. In a typical experimental set up for the ultra-fast excitation the temperature is
highest at free vacuumsample boundary, therefore the importance of expansion is obvious. A special diagnostics is needed
to distinguish the effect of expansion on the diffracted probe from the phase transition changes in a sample.
The presented analysis allows concluding that the fast fs-laser-induced excitation of Bi in all experiments published so
far is of the entirely thermal nature. Another conclusion strongly supported by the experiments with the optical probes is
that the excited Bi was not transformed to the liquid phase during the observation time of 25 ps due to strong cooling and
not sufficient time for the transformation to be completed in the whole laser affected layer.
4.7.5. Ultra-fast melting of dielectrics: experiments with superheated Indium Antimonide
Rousse et al. (2001) were the first to claim the observation of the non-thermal melting in semi-conductors, namely in
InSb, excited by the ultra-short laser (see also Guo et al., 2000; Zijlstra et al., 2008). The layer of InSb was excited by the
powerful laser pulse 120 fs, at 800 nm, and at the fluence of 120 mJ/cm
2
. The conclusions on the transient state of excited
layer were derived from the analysis of the measured time-dependent diffracted X-ray beam probe intensity.
First, the analysis of the lasermatter interaction in these experiments shows that the powerful laser swiftly superheated
the sample. Indeed, the absorbed energy density at the fluence of 120 mJ/cm
2
equals to 7.55 kJ/cm
3
, 7.6 times higher
than the equilibrium enthalpy of melting (0.975 kJ/cm
3
for InSb). The average laser intensity during the pulse constitutes
10
12
W/cm
2
that is close to the breakdown threshold for many dielectrics. The estimate for the number density of excited
electrons in the conduction band performed by the authors, 7 10
21
cm
3
, is 4 times over the critical density for 800 nm,
which comprises 1.7 10
21
cm
3
. Therefore, early in the laser pulse (at 120 mJ/cm
2
) the large number of free carriers was
createdandthe laser absorptionoccurreddue tointra-bandas well as inter-bandtransitions similar tothat inmetals. It is also
documented by the crater formation observed by Rousse et al. (2001). The observation of the crater formation is evidence
of ablation. With such large number of free carriers the previous analysis of the relaxation times applies. The statistical
distributions in electron and lattice sub-systems, electron-to-lattice energy transfer and temperature equilibration occur
during the pulse time due to strong overheating over the melting point. All the processes involved in transformation of the
material occur then in a completely thermal mode.
The diffracted X-ray intensity decreased sharply, from 20% to 60%, during 350 fs; after that the intensity remained
constant during the following 4 ps of observation. The authors attributed the sharp decrease in the intensity to the atomic
disordering. The average magnitude of displacement, approximately one angstrom, was deduced from the intensity fall-off
by applying the classical DebyeWaller factor. The magnitude of displacement significantly exceeds the phonons amplitude
and signifies the strong thermal expansion. It should be reminded that the classical DebyeWaller factor is invalid in such
non-equilibrium and inhomogeneous conditions.
The entropy increase due to the change in the inter-atomic separation on angstromis ten times less of catastrophic value
that necessary for disordering. On the other hand one can easily see that at the superheating seven times of the melting
point the entropy rise due to the thermal disordering (lattice heating) reaches the catastrophic value, signifying disordering
associated with melting, well before the temperature maximum and the full electron-to-lattice equilibration. Hence the
ultra-fast melting occurs in these experimental conditions due to the lattice overheating, and is of entirely thermal nature.
Thus, the experiments of Rousse et al. (2001) gave evidence that the ultra-fast melting in semiconductors occurs in
conditions of strong lattice superheating when electron and lattice temperatures are already equilibrated.
4.8. Conclusions
Summing up, the experimental results show that optical, spectroscopic, X-ray and electron diffraction measurements
of different ultra-fast laser excited solids clearly indicate that disordering (e.g. melting) commences and continues at the
thermal stage, when electron and lattice temperatures are equilibrated. Even in the case of strong overheating of Al over the
ablation threshold melting occurs before ablation in picoseconds after the end of the pulse (Kandyla et al., 2007).
The correct account for the lasersolid interaction processes allows better understanding of the ultra-fast melting of met-
als and dielectrics. The high intensity necessary to achieve the melting conditions with the short laser pulse usually results in
the breakdownof dielectrics making the interactionandthe following melting close tothat withmetals. Infact, careful analy-
sis shows many similarities even in equilibriummelting of metals and dielectrics as it documented in Table 4.1. The enthalpy
of melting and the ratio of melting point to the binding energy have similar values for both the metals and the dielectrics.
The degree of the material disordering caused by ultra-short laser excitation is characterised by time-dependent entropy
at non-thermal and equilibrium stages. The local equilibrium in electronic and lattice subsystems establishes early in the
pulse time making viable the thermodynamic description of the physical processes in each sub-system before electron and
lattice temperature equilibration. The entropy increase caused solely by the electronic excitation is less than one tenth of
the critical value for the major lattice disordering associated with melting. Structural stability of a swiftly excited solid also
enhanced due to the fact that the time for the thermal defect formation appears to be much longer than that for all other
relaxation processes. It immediately follows that fast thermal defect contribution into disordering is inhibited. Therefore,
only the superheating of the lattice over the equilibriummelting point can drive the disordering. The superheating also leads
to the fast electron-to-lattice temperature equilibration ensuring a purely thermal character for melting of a solid heated by
an ultra-short laser pulse of duration shorter than the major relaxation times. The catastrophic disordering leading to the
liquid state is a consequence of the superheating of atoms in the crystal, and/or the formation (or presence) of characteristic
point defects.
4.9. Summary
Enthalpy of fusion in equilibrium (specific heat of
melting) for metals:
H
f
= (1.071.86) kJ/cm
3
Entropy of fusion in equilibrium (metals): S
m
=
H
f
T
m
= (1.11.38)k
B
Lindemann criterion for meting:
q
2
m
d
2
0
k
B
T
m
b
3%
Heat for a point defect formation in metals: H
vac
= 9.28k
B
T
m
Entropy catastrophe value for metals: S
catastrophe
S
lattice
+S
defects
(67)k
B
The point defects (vacancies) formation time in
metals ( = T/T
m
):
t
def
t
tft
4.58 10
2 h
k
B
T
m
exp
_
18.56
_
5. Ablation of solids
5.1. Introduction
Atoms in a laser-irradiated surface layer gain energy in excess of binding energy and escape the surface when absorbed
laser energy increases above a certain threshold (Afanasiev and Krokhin, 1967; Afanasiev et al., 1999; Gamaly, 1994). Atoms
possessing a kinetic energy in excess of the binding energy are then removed from a sample. Removal of material by laser
action was termed laser ablation in order to distinguish it fromevaporation in equilibriumconditions (Chrisey and Hubler,
1994; Miller and Haglund, 1998; Bauerle, 2000; Eason, 2007). The ablation of materials by powerful femtosecond lasers
has attracted significant attention in recent decades due to many potential applications in industry, medicine, material
science and technology. This section is devoted to a thorough description of mechanisms of ablation by ultra-short laser
pulses, establishing thresholds, ablation rates, ablated plume composition, and other ablation parameters as a function of
laser irradiation and material properties. Let us briefly discuss the main characteristics of laser ablation, which distinguish
it from familiar evaporation in equilibrium.
In order to remove an atomfroma solid by means of a laser pulse one should deliver an amount of energy in excess of the
binding energy to that atom. Therefore the absorbed energy density per atom in a laser-heated layer, E
abs
= 2AF(t
p
)/n
a
l
s
,
should at least be comparable to the heat of vaporisation in equilibrium; here A is the absorption coefficient; F is the incident
laser fluence, the energy per unit surface area during the pulse of duration t
p
; n
a
is atomic number density and l
s
is the skin
depth, as in previous sections. Typical ablation threshold fluence, as shown later in this section, is of the order of 1 J/cm
2
.
For example, laser ablation with a 100 fs pulse requires the intensity in the range above 10
13
W/cm
2
(Perry et al., 1999; Du
et al., 1994; Stuart et al., 1995), while a 10 ns pulse ablates the same material at much lower intensities 10
8
10
9
W/cm
2
(Tien et al., 1999; Gamaly et al., 1999a; Rode et al., 1999). It has been shown that intensities of the order of 10
13
W/cm
2
correspond also to the ionisation threshold (Arnold and Cartier, 1992; Tien et al., 1999; Gamaly et al., 1999b; Hallo et al.,
2007; Mzel et al., 2010). The interaction mode changes and an ultrafast laser pulse interacts with plasma early in a pulse
time in non-equilibrium conditions in contrast to low intensity long pulse regime. The material removal with long low-
intensity pulse proceeds in equilibrium where heating, melting, and evaporation of the target occur in succession, without
plasma formation.
A distinctive feature of the ultra-short interaction mode is that the laser pulse is shorter than the major relaxation
processes in the irradiated material. Thus, the non-equilibrium phenomena play an important role, modifying significantly
the mode of material removal. One of these phenomena, which we discuss in detail later in this section, is the fast formation
of lattice distribution function with a truncated high-energy tail. The implication of this effect is that the solid should be
superheated beyond the equilibrium enthalpy of evaporation, in order to reach the non-equilibrium ablation threshold.
As a consequence of non-equilibrium conditions, there are three modes of ablation, depending on the laser and
target parameters and on the ambient conditions of the experiment. The extreme ablation mode, electrostatic ablation, is
completely non-equilibrium and non-thermal. This mode is realised when a short powerful pulse elevates average electron
energy during the pulse in excess of the sum of the binding energy of the ions plus the energy necessary for the electron to
escape from a solid. The lattice remains cold during the pulse. The energetic electrons escaping from a solid create a huge
electrostatic field of charge separation, which pulls the ions out of a solid. The second mode, non-equilibrium and semi-
thermal, is realised when electrons have enough time to transfer the energy to the lattice and the average energy of the ions
(temperature) exceeds the binding energy but the distribution function is far from the equilibrium Maxwell distribution.
In these conditions, the majority of ions escape the solid before the equilibrium distribution is established. The third mode,
thermal equilibrium ablation, coincides with evaporation in equilibrium conditions. This mode may occur after the pulse
and after the full Maxwell distribution is established, including the high-energy tail with energy in excess of binding energy.
In this mode the ablation is the same as evaporation in equilibrium: only particles with energy in excess of the binding
energy from the high-energy tail can escape the solid. An ambient gas next to the ablation surface also helps to establish
the equilibrium distribution. This effect leads to a decrease in the ablation threshold; we will present this in Section 5.5.3.
It is noteworthy that absorbed energy density at the ablation threshold in extreme non-equilibrium conditions is 23 times
higher than the equilibrium enthalpy of evaporation.
An ultra-short laser pulse heats a solid to a depth of 1050 nm. Heat conduction is usually negligible during the pulse
time. On the other hand, the heat conduction dominates the ablation process during the long (nanosecond) pulses at the
equilibriumconditions. A smooth transition between two regimes was established and the formulae for ablation thresholds
and ablation rates for metals and dielectrics, combining the laser and target parameters, were derived and compared to
experimental data (Gamaly et al., 2002). The calculated dependence of the ablation thresholds on the pulse duration is in
agreement with the experimental data in a femtosecond range, and it is linked to the dependence for nanosecond pulses.
Control over the state of the expanding laser-produced plume is important for many applications of laser ablation,
particularlyfor laser deposition of optical films (Chrisey and Hubler, 1994; Gamaly et al., 1999a; Eason, 2007). Such control
can be achieved through the proper choice of pulse duration, wavelength and energy, along with a proper spatial and
temporal distribution of the laser intensity during the pulse and across the focal spot, while keeping the absorbed laser
energy above a specific threshold (Gamaly et al., 2004). The optimal conditions for creating a fully atomised laser-ablation
plume are defined later in this section. An essential part of these conditions demands that spatial (over the laser focal
spot) and temporal intensity distributions during the pulse should be flat-top-like in the domains of space and time. It was
Table 5.1
Thermal characteristics of Al, Cu, and Pb at standard conditions; = (1/V
0
)(V/T)
p
is the coefficient of thermal expansion;
0
= (V
0
) is the Gruneisen coefficient
defined as a ratio of the thermal pressure to the thermal energy per unit volume
p
th
= (V)
th
/V; c
0
is the speedof sounddeterminedby the isothermal compressibility
(Zeldovich and Raizer, 2002, p. 698).
Al Cu Pb
(g/cm
3
) 2.71 8.93 11.4
10
5
(K
1
) 2.31 1.65 2.9
0
2.09 1.98 2.46
c
0
(m/s) 5200 3950 1910
particularly established that pre-pulses and after-pulses have a detrimental effect on the composition of the laser plume
(Pronko et al., 1998; Gamaly et al., 2004).
High repetition-rate lasers (up to 100 MHz) are used in practical applications of the laser ablation process (Luther-Davies
et al., 2005; Budd et al., 2001). A focused laser beamis scanned over the target surface with a velocity up to 10 m/s, but it can
still dwell at a particular spot for longer than the time gap between the pulses. This leads to coupling between successive
laser pulses and thereby to incubation of the heating and evaporation processes. At some repetition rate the multiple-pulse
action may result in a reduction of the laser-ablation threshold. Practical recipes are provided that lead to efficient ablation
and control the phase-state of the vapour as a self-consistent function of laser parameters, namely, the pulse duration,
the energy per pulse, and the wavelength, all adjusted to the target parameters. The implementation of these recipes for
deposition of high-quality micron-thick optical films for optical waveguide applications is demonstrated in Section 7.
This section is organised as follows. First, the evaporation in thermal equilibrium is revisited in order to remind readers
of the main features of the equilibrium process. Then, the mechanisms of non-thermal ablation, the thresholds for ablation
to commence in metals and dielectrics are determined, and the ablated rates, mass and depth of the ablated material as
a function of laser and material parameters are presented in succession. The problems of control over the ablated plume
content in a single-pulse ablation mode are followed by discussion of interaction between multiple successive pulses in the
high-repetition rate regime. The effect of ambient gas on the ablation threshold and rates is then discussed. In conclusion, we
discuss the unresolved problems in practical applications of laser ablation such as angular dependence and stoichiometry
of the ablated plume and the prospects for future studies.
5.2. Evaporation in equilibrium
It is instructive to revisit the evaporation in thermal equilibrium for better understanding the distinctive characteristics
of non-equilibrium ablation by short-pulse laser. The knowledge of relations between the pressure and the specific energy
in a solid as function of temperature and density (so called the equation of state (EOS)) is necessary for characterisation of
evaporation parameters. The process of evaporation is poorly investigated in the conditions when the solid expands into
the volume exceeding only a few times (510) the initial solid density volume. The equation of state (EOS) of the material
for this transitional stage from solid to vapour is very important for understanding the evaporation (ablation) process. The
EOS for this range of parameters is known only on a qualitative level for the most of the materials. However, one can draw
important conclusions even with a qualitative EOS.
Qualitatively correct equation of state can be obtained under assumption that in a solid at temperature in excess of the
Debye temperature (T > T
D
) its atoms are still undergoing the harmonic vibrations. The Helmholtz free energy then reads
(Landau and Lifshitz, 1980a):
F
s
=
c
(V) +c
s
TN ln

h
T
. (5.1)
Here

h = e
1/3
T
D
is the average energy of atomic vibrations (phonons) directly related to the Debye temperature, T
D
.
The total number of particles, N, is constant. Energy at T = 0, so called cold energy,
c
, is temperature independent
and it defines a cold pressure, P
c
. Using conventional thermodynamic relations for pressure P =
_
F
V
_
T
, and energy
= F T
F
T
, one arrives to the equation of state in the following form:
=
c
(V) +c
s
k
B
TN
c
(V) +
th
P =

th
V
(V) +P
c
(5.2)
where P
c
= d
c
/dV, and (V) = ln
h / ln V is the Gruneisen coefficient, > 0, which defines the ratio of the

thermal pressure to thermal energy. The Gruneisen coefficient changes from 2 at the solid state (see Table 5.1), to 2/3,
which corresponds to the ideal gas during the evaporation process. The thermal term in general includes the ionic and
electronic parts.
Fig. 5.1. Qualitative dependence of the cold pressure on the specific volume in solids.
Table 5.2
The enthalpy of equilibriumvaporisation per atomH
boil
/n
a
, the binding energy
b
, and the ratio k
B
T
boil
/
b
for metals (Kittel, 1996; Weast and Astle, 1981).
Al Si Fe Cu Mo Ag W Pb
H
boil
/n
a
; (eV) 3.05 4.09 3.53 3.11 6.158 2.953 8.284 1.783
b
; (eV) 3.065 4.118 3.695 3.173 6.200 2.972 8.34 1.795
k
B
T
boil
/
b
0.071 0.074 0.073 0.077 0.068 0.071 0.060 0.097
5.2.1. Cold pressure and energy
The solid at zero temperature is in a state of mechanical equilibrium: the inter-atomic forces of attraction and repulsion
are mutually balanced. The minimum of the inter-atomic potential in equilibrium equals to the binding energy, U
min
(d) =
b
, d is the inter-atomic distance in equilibrium. The dependence of elastic pressure, P
c
, and specific energy,
c
, on density
(or specific volume V = NM/) has qualitative character anda curve shape similar to the dependence of the potential energy
of the interaction between two atoms on the inter-atomic distance in a molecule, a Morse-like potential (see Fig. 5.1).
When expansion starts due to heating, the attractive forces first increase; correspondingly a cold pressure becomes
negative and reaches its minimum. With further expansion the inter-atomic interaction rapidly decreases and the elastic
pressure tends to zero. The negative sign for the elastic pressure under expansion reflects the physical fact that in order to
expand a solid fromthe initial state of mechanical equilibriuma tensile force must be applied to overcome the binding force,
which tends to return the atoms into the equilibrium positions. By definition, the work produced against the cold pressure
in order to expand a solid from the initial volume to infinity equals to the binding energy:

V
0
P
c
(V)dV =
b
. (5.3)
The energy of evaporation per atom (the enthalpy of vaporisation per atom) equals to the cohesion (binding) energy is
presented for some metals in Table 5.2 (Kittel, 1996).
The cold pressure and cold energy have direct connection to the critical parameters in equilibrium vaporisation. The
critical point is at its maximum on the phase equilibrium curve in PV and TV planes (Landau and Lifshitz, 1980b). The
importance of the critical point relates to the fact that difference between solid, melt and vapour ceases to exist in a critical
state, i.e. the density of all phases has the same value, V
cr
. The pressure at the solidvapour interface in equilibrium grows
up with the temperature increase until the critical point. The boiling curve in PT-plane ends in a critical point, at P
cr
and
T
cr
. The critical point is a single point (and a singularity point) of thermodynamically stable homogeneous states of matter
characterised by T
cr
and V
cr
. The critical values are poorly known for many materials. It is established, however (Zeldovich
and Raizer, 2002), that critical temperature in energy units constitutes 0.10.2 of the binding energy. One can see that for
good metals the ratio of boiling energy per atom to the binding energy is less than 10% (see Table 5.2). However, the work
necessary to achieve the critical state is proportional to the binding energy, P
cr
V
cr

b
. That is another indication that
evaporation occurs by the atoms from the high-energy tail at
b
> k
B
T
boil
.
5.2.2. Saturated pressure of vapours in equilibrium
Solid and vapour parameters near the solidvapour interface are explicitly linked for the evaporation at thermodynamic
equilibrium. Thermodynamic equilibrium implies that evaporation takes place in a period of time much longer than all the
relaxation times. It also means that temperature, pressure and chemical potential at the solidvapour interface are equal for
both phases. The pressure at the solidvapour interface in equilibrium follows from the condition of equality of chemical
potential for solid to that for liquid
s
=
g
(Landau and Lifshitz, 1980a,b):
P = C
_
T
T
D
_
c
p
c
s
exp
_
c

g
k
B
T
_
. (5.4)
Here c
p
, c
s
are respectively the specific heats of vapour and solid.
c
,
g
are cold energy in a solid and energy in a gas per
atom. The Debye temperature for a solid, T
D
, is used for normalisation, C is a constant. The heat of phase transition equals
to the difference between the enthalpy in the gas and in solid states: w
g
w
s
= (c
p
c
s
)k
B
T +
g

c
. One can neglect
the interaction of atoms in vapour in comparison to that in solids,
g

c
. Thus, the energy necessary to detach atom from
a solid at zero temperature equals to the binding energy in solid, because
c
(V 0) =
b
. Then Eq. (5.4) converts into
familiar form (Landau and Lifshitz, 1980a,b):
P = C
_
T
T
D
_
c
p
c
s
exp
_

b
k
B
T
_
. (5.5)
One can express the specific heat for the vapour, c
p
=

1
, through the adiabatic constant, = c
p
/c
V
. For = 5/3 as in
ideal gas it makes c
p
= 5/2, while c
s
changes from c
s
= 3 for a solid up to c
v
= 3/2 in a vapour state.
5.2.3. Evaporation rate
The equilibrium evaporation rate for a solid in vacuum at temperature T follows from conventional thermodynamics
(Landau and Lifshitz, 1980a,b):
(n
i
v
i
)
equilibrium
=
P
2m
i
k
B
T
(1 R). (5.6)
Here P is pressure of saturated vapour at the temperature T defined above, R is the average coefficient for vapour atoms
reflection from a solidvapour interface, and m
i
is atomic mass. Inserting Eq. (5.5) into Eq. (5.6) one obtains the explicit
expression for the evaporation rate in equilibrium:
(n
i
v
i
)
equilibrium
(1 R)T
c
p
c
s
1/2
exp
_

b
k
B
T
_
. (5.7)
Close to the critical point we take c
s
3/2 and c
p
= 5/2. Let us denote some characteristic density of the saturated vapours
near solidvapour interface as n
vap
(that density can be obtained only fromexperiments). Then the equilibriumevaporation
rate can be expressed in terms of mass, temperature, and binding energy as a scaling with clear physical meaning:
(n
i
v
i
)
equilibrium
n
vap
_
k
B
T
M
_
1/2
exp
_

b
k
B
T
_
. (5.8)
The pre-exponential factor presents a free flow of evaporated atoms through the solidvapour interface. Thus, only a
small amount of particles with energy
b
from the high-energy tail in the Maxwell distribution evaporates from a solid in
conditions when solid and vapour are in equilibrium (Landau and Lifshitz, 1980a,b). The rate of evaporation in equilibrium
is low due to condition, k
B
T
evap

b
, however velocity of evaporating particles equals to a local speed of sound.
Similar estimates could be obtained from kinetic considerations. Probability of evaporation of a single atom by the
kinetic theory (Frenkel, 1948; Anisimov et al., 1971; Afanasiev and Krokhin, 1971) is the following, w
D
exp{
b
/T}.
Time for establishing the MaxwellBoltzmann distribution is inverse of the Debye frequency, t
eq

1
D
, while the time
necessary to remove a single atom from a solid defined by, t
evap
w
1
. Therefore, in equilibrium the establishment of the
Maxwell distribution goes much faster than evaporation, t
eq
t
evap
. Note that the velocity of evaporated atoms expresses
as v
evap
d
D
v
sound
. One can see that in non-equilibrium conditions the particles removal and establishment of
distribution function goes together, t
eq
t
evap
. This means that the majority of atoms in non-equilibrium ablation should
have enough energy to break bonds, and almost mono-energetic distribution, therefore the exponential term is close to
unity. These conditions will be presented later in the studies of non-equilibrium ablation.
5.2.4. Number of particles in the high energy tail of the Maxwell distribution
Let us calculate the number of atoms in the high-energy tail, at
b
T range of the Maxwell energy distribution,
dN
v
= n(M/2T)
3/2
exp(Mv
2
/2T)4v
2
dv, here n is the number density of atoms. The fraction of high-energy particles
then reads:
n
b
n
=
1
n
_

b
/T
dN
v
_
= 1 erf(
b
/T)
1/2
+
2
(
b
/T)
1/2
exp(
b
/T) (5.9)
where erf is the error function (erf (x > 2) = 1). One can see from Eq. (5.11) that in conditions of thermal evaporation,
T
evap

b
, the number of energetic particles in the high-energy tail complies with those following from conditions of
evaporation in equilibrium equation (5.10):
n
b
n

2
(
b
/T)
1/2
exp(
b
/T). (5.10)
Thus, the study of the equilibrium evaporation teaches us two lessons: only atoms with the energy in excess of binding
energy are leaving a solid; the number of evaporating atoms corresponds to small fraction of those in the high-energy tail
of the Maxwell distribution with >
b
, thus providing the low evaporation rate.
5.3. State of solid excited by ultra-short pulse at the energy density around the equilibrium enthalpy of vaporisation
The major difference in the state of the ultra-fast laser excited solid from that in equilibrium relates to the fact that the
energy density comparable to the enthalpy of vaporisation delivered into the electronic component much faster than the
major relaxation times. The electrons are reaching the maximum temperature (average energy per electron) at the end of
the pulse while the lattice remains cold. The maximum electron temperature in these conditions is comparable to the first
ionisation potential and the Fermi energy in metals. Thus any material is swiftly ionised in these conditions. Then electrons
transfer the energy to the lattice, raising the lattice temperature (the average energy per atom) to the maximum value by
the time of electronlattice temperature equilibration. Note that maximum electron temperature is always larger than the
maximum lattice temperature. However the ion distribution function remains different from the Maxwell form due to the
absence of high-energy tail long after the end of the laser pulse and after the electronlattice temperature equilibration: the
high-energy tail needs longer time to be built. Therefore several mechanisms of ablation are contributing into the material
removal depending on the interplay of many competing processes, which we discuss below. Let us recall the major physical
processes that define the conditions created by a powerful short laser pulse depositing energy density in excess of enthalpy
of vaporisation.
5.3.1. Swift ionisation of dielectrics
It has been shown (Du et al., 1994; Stuart et al., 1995; Perry et al., 1999), that most materials can be ablated using ultra-
short pulses of 100 fs. The average intensity used for ablation is above 10
13
W/cm
2
, this intensity exceeds the ionisation
threshold. Therefore, the target material is ionised at the beginning of the laser pulse, which creates high-density plasma.
Free electrons absorb the laser energy in the plasma via inverse Bremsstrahlung and resonance absorption.
Dielectrics are transparent at lowintensity for the laser wavelength down to the ultraviolet range. Lowabsorption at low
laser intensity implies that large real and small imaginary parts characterise the complex dielectric function. The increase in
laser intensity, and as a resultthe increase of the energy available for electron excitation, leads to ionisation of the target
in a skin layer. The ionisation occurs through single-photon ionisation, multi-photon ionisation or tunnel ionisation, see
Keldysh (1965); Oppenheimer (1928), and ionisation by electron impact (avalanche ionisation) mechanisms. The imaginary
part of dielectric function, and hence the absorption, both increase due to ionisation while the absorption depth decreases.
The degree of ionisation increases with increasing laser intensity and with decreasing of the laser wavelength.
The relative roles of impact ionisation and multi-photon ionisation depend dramatically on the relation between the
electron quiver energy in the laser field and the ionisation potential (Perry et al., 1999; Gamaly et al., 2002). Electron impact
ionisation is the main ionisation mechanismin the long (nanosecond) pulse regime. Multi-photon ionisation dominates the
laser-interaction at intensities above 10
13
10
14
W/cm
2
that are characteristic of the short pulse regime, it also depends on
the laser wavelength and the ionisation potential of the material. For a 100 fs pulse at 800 nm this corresponds to a laser
fluence of 110 J/cm
2
. The ionisation time can be shorter than the pulse duration, in which case the ionisation threshold
depends on the laser intensity, and it decreases with the increase in the photon energy.
It is suggested conventionally that the ionisation threshold is achieved when the electron number density reaches the
critical density corresponding to the incident laser wavelength (Raizer, 1978; Kruer, 1988). The ionisation threshold for the
majority of materials lies at intensities between 10
13
and 10
14
W/cm
2
(at a laser wavelength 1 m) with a strong non-
linear dependence on intensity and wavelength. For example, for a silica target at an intensity of 210
13
W/cm
2
, avalanche
ionisation dominates, and the first ionisation energy is not reached by the end of a 100 fs pulse at 1064 nm. At 10
14
W/cm
2
multi-photon ionisation dominates and the full first ionisation is completed in the first 20 fs of the laser pulse. When the
ionisation is completed, plasma is formed in the skin layer. This plasma has a free-electron density comparable to the ion
(solid) density of about 10
23
cm
3
. Hence, for the derivation of scaling relations the electron number density (and thus the
electron plasma frequency) is considered to be constant and equal to the atomic number density.
5.3.2. Relaxation processes in laser-excited solids
In order to meet the ablation conditions the average electron energy should increase from the initial room temperature
up to the Fermi energy
F
in metals. Note that the Fermi energy, the binding (cohesion) energy and the first ionisation
potential are of the same order of several electronvolts. The electronelectron equilibration time is proportional to inverse
electron plasma frequency, i.e.
1
pe
10
2
fs. Therefore the electron energy distribution during the pulse is close to the
equilibrium one and follows the laser intensity evolution in time adiabatically adjusting to any changes. The same applies
to the phononphonon collision rate, which is responsible for the time of establishing the average energy density (lattice
temperature), t
eq
(
phph
)
1

h/T
L
. This time comprises a few femtoseconds in the conditions close to the ablation
threshold.
5.3.3. Skin-layer approximation
The laser pulse is shorter than the expansion time, which estimates as a skin depth divided by the sound velocity:
t
exp
l
s
/v
sound
; it is of the order of a few tens of picoseconds in metals. Thus, a femtosecond pulse interacts with a solid
target whose density remains almost constant during the laser pulse. The major process during the lasertarget interaction
is the heating of electrons by the laser electromagnetic field. However, the electrons number density n
e
, the electronion
collision frequency
ei
, the absorption coefficient A, and the skin-depth l
s
, all are time and laser intensity dependent. It
has been shown (Gamaly et al., 2002) that the above quantities, whilst changing rapidly during the very early part of the
pulse, become approximately constant for most of the remaining period up to the end of the pulse. Thus, the skin-effect
approximation can be used to describe the interaction of a sub-picosecond pulse with matter. As a result the laser electric
field E(x) as a function of the penetration depth x into a solid target obeys the relation E(x) = E(0) exp(x/l
s
).
The electronatom (or ion) momentum exchange rate changes from the low-temperature limit, where the lattice
temperature dominates the electronphonon interaction, to the high-energy limit, where the electronion interaction
depends on the electron temperature (see Section 2). The effective collision frequency has a broad maximumat the electron
energy of the order of the Fermi energy in metals, k
B
T
e

F
(see Section 2). Following the literature (Perry et al., 1999;
Eidmann et al., 2000; Gamaly et al., 2002) it seems reasonable to assume that the electronion momentum exchange rate is
proportional to the electron plasma frequency
mom
ei

pe
near the ablation threshold, and as the first order approximation
does not depend on the temperature. Thus, at k
B
T
e

F
the electronion momentum transfer rate is much larger the laser
frequency
mom
ei

pe
. Therefore the electron mean free path, l
mfp
= v
e
/
mom
ei
is of the order of a few angstroms,
much smaller than the skin depth. Hence the conditions for the normal skin effect to be valid are justified, and one can use
the exponential spatial decay with depth for the laser field.
5.3.4. Electron-to-ion energy-transfer time
The electronion energy transfer time in plasmas expresses through the momentum exchange rate by the familiar
formula, t
en
ei
(m
e
mom
ei
/M
i
)
1
(see Section 2). In accord with the previous paragraph the minimumenergy transfer time at
the maximum of momentum rate reads, (t
ei
)
min
M
i
/
pe
m
e
(Malvezzi et al., 1986; Luther-Davies et al., 1992; Perry et al.,
1999; Eidmann et al., 2000). Taking Cu as an example (M
Cu
= 63.54 a.u., n
e
= 8.4710
22
cm
3
,
pe
= 1.6410
16
s
1
), we
estimate the minimumionheating time as t
ei
= 710
12
s, whichis inagreement withexperimental values inthe literature
(Eidmann et al., 2000; Elsayed et al., 1987). Similar estimates for silver gives 15 ps, which qualitatively correlates with the
measured 71 ps (Groeneveld et al., 1990), and 27.4 ps for gold which should be compared with the measured 70120 ps
(Schoenlein et al., 1987). A similar estimate for fused silica yields 6.4 ps. These estimates demonstrate that the ions remain
cold during a sub-picosecond laser pulse and long after the end of the pulse for both metals and dielectrics. For this reason
one can apply a step-like plasma density profile for the laser absorption calculations as we will consider explicitly in the
next section.
5.3.5. Heat diffusion time
The electron thermal diffusion coefficient, D, is expressed through the electronion momentum transfer rate in accord
with general kinetic expression as follows (Lifshitz and Pitaevski, 1981):
D = l
e,mfp
v
e
3
=
v
2
e
3
mom
ei
T
5/2
. (5.11)
Here l
e,mfp
= v
e
/
mom
ei
and v
e
are the electron mean free path and velocity, respectively. The electron heat conduction time
t
heat
, which is the time for the electron diffusion across the skin-depth l
s
, now becomes:
t
heat

l
2
s
D
. (5.12)
For copper brought close to the ablation conditions by 780 nm laser (l
s
= 67.4 nm), the heat diffusion coefficient estimates
as the following, D v
2
e
/3
pe
1 cm
2
/s. Therefore the electron heat conduction time by Eq. (5.12) equals to t
heat
45 ps.
Thus, all the absorbed energy is confined in the electron component during the lasermatter interaction time with sub-
picosecond and picosecond pulses, and the energy losses are negligible. One expects that in these conditions only energetic
electrons might push the ultra-fast material removal.
5.3.6. Absorbed energy density in dielectrics below and at the ablation threshold
Absorbed energy density in dielectrics, especially in wide band gap dielectrics, is a very strong function of laser intensity.
It is instructive to consider the absorbed energy density in dielectrics during the ionisation process and at the state when
the first ionisation is almost completed and the electron number density is close to saturation.
(a) Approaching the breakdown threshold
The permittivity of a dielectric experiencing ionisation can be presented as a sum of the dielectric (polarisation)
contribution and the Drude-like part due to excitation of electrons into the conduction band, =
0
+
D
. The
electronphonon collision rate
eph
2k
B
T/
h < is less than the laser frequency. Therefore below the breakdown
threshold the number density of electrons in conduction band is less the critical density and therefore the laser frequency
is larger the plasma frequency (
2
pe
<
2
n
e
< n
cr
). In these conditions the contribution of excited electrons to the real
part of permittivity, (
D
)
re
1(
pe
/)
2
< 1, is positive and small. Hence, we take the real part for the permittivity and
for the refractive index equal to that of the unperturbed state. The imaginary part in these conditions takes the following
form:
(
D
)
im

2
pe
eph
=
n
e
n
cr
eph
.
One can see that imaginary part of the refractive index, which is responsible for the absorption length, grows with the
increase of electron density, and remains much lower than the real part. The relations between the permittivity and the
refractive index in these conditions are the following: n
2
0

re
; 2n
0
k (
D
)
im
. Thus the absorption coefficient can be
taken in the unperturbed form, while the imaginary part can be neglected: A
0
4n
0
/(n
0
+ 1)
2
. The main dependence
of the absorbed energy density on the number density of excited electrons comes from the absorption length through the
imaginary part of the refractive index:
l
abs
=

2
=
n
0
(
D
)
im
=
2cn
0
eph
n
cr
n
e
.
Now the absorbed energy density, E
abs
, in a dielectric approaching the optical breakdown reads:
E
abs
= A
0
eph
cn
0
n
e
n
cr
F. (5.13)
The number density of excited electrons is a strong non-linear function of laser intensity when multi-photon absorption
mechanism dominates, n
e
I
n
, n > 1. Temnov et al. (2006) measured the dependence of the number of excited electrons
as a function of the laser intensity in fused silica irradiated by 50 fs pulses at 800 nm ( = 2.356 10
15
s
1
). The measured
electron density dependence on fluence can be approximated as n
e
= 10
19
(F[J/cm
2
])
6
. At higher intensity the growth
saturates approaching to the critical density of n
cr
=
2
m
e
/4e
2
= 1.745 10
21
cm
3
at 800 nm. The dependences of the
absorbed energy density on the laser fluence for Aluminium (example of a good metal) and for fused silica (as a dielectric)
are presented at Fig. 4.2 for comparison. One can clearly see the dramatic difference in the interaction and therefore in the
ablation thresholds for metals and dielectrics: the dielectrics require much more energy to be deposited, and thus much
higher laser fluence for ionisation. As soon ionisation achieved, the absorbed energy density increase sharply at the ablation
threshold.
(b) Absorbed energy density above the breakdown threshold
At intensities above 10
14
W/cm
2
the ionisation time for a dielectric is just a few femtoseconds, shorter than a typical
pulse duration of 100 fs. The electrons produced by ionisation in dielectrics then dominate the absorption in the same
way as the free carriers do in metals, and the characteristics of the lasermatter interaction become independent of the
initial state of the target. As a result, the inverse Bremsstrahlung and resonance absorption (for p-polarised light at oblique
incidence) become the major absorption mechanisms for both metals and dielectrics. For this reason one can use the Drude-
like dielectric function for description of the optical properties of the ionised dielectric. The condition
mom
ei

pe
is
fulfilled if the first ionisation is completed at the early stage of the pulse duration. The link between the real and imaginary
parts of the Drude dielectric function and refractive index can be then simplified:
(
D
)
re

2
2
pe
;
(
D
)
im

pe
;
n k =
_
(
D
)
im
2
_
1/2
=
_
pe
2
_
1/2
.
The absorption coefficient, following fromthe Fresnel formulae, and the absorption depth (the skin depth) are the explicit
functions of the refractive index:
A = 1 R
4n
(n +1)
2
+n
2
;
l
s

c
n
.
(5.14)
Thus, the ratioof the absorptioncoefficient tothe skindepthappears tobe a weak functionof material properties (Rozmus
and Tikhonchuk, 1990; Luther-Davies et al., 1992):
E
abs
=
2AF
l
s
C
n
8
F. (5.15)
The function C
n
=
_
1 +
1
n
+
1
2n
2
_
1
depends weakly on the material and laser parameters. For example, for copper ablation
at 780 nm (w = 2.415 10
15
s
1
; w
pe
= 1.64 10
16
s
1
) it has the value 0.585, while for gold ablation at 1064 nm
(w = 1.77 10
15
s
1
; w
pe
= 1.876 10
16
s
1
) the value is 0.65. Thus, one can assume for the further estimates that the
absorbed energy density near the ablation threshold is directly proportional to the incident fluence and inverse proportional
to the laser wavelength. The numerical coefficient C
n
0.6 could be corrected in (5.15) for any particular material.
5.3.7. Maximum electron and ion temperatures
The specific heat of the atoms and electrons becomes comparable when electrons and lattice temperature in the energy
units approaches to the binding energy, C
atom
C
e
= 3k
B
/2 at T
e
= T
L
T
b
(see Section 2). The energy conservation
law for the conditions of normal skin effect and under the assumption that all the parameters but laser intensity are
time independent during the pulse takes the simple form of an equation for the change in the electron temperature, T
e
,
due to absorption in a skin layer (Gamaly and Tikhonchuk, 1988; Rozmus and Tikhonchuk, 1990, 1992; Luther-Davies et al.,
1992):
k
B
T
e
=
2E
abs
3n
e
exp
_
2x
l
s
_
; E
abs
=
2AF
0
l
s
; F
0
=
t
p
0
I
las
(t)dt; (5.16)
where E
abs
, F
0
are respectively the absorbed energy density and incident fluence in the skin layer, A = I/I
0
is the absorption
coefficient, I
0
= cE
2
/8 is the incident laser intensity, n
e
and C
e
3/2k
B
are the number density and the specific heat of
the conducting electrons.
The ion temperature reaches its maximum at the moment of electronion temperature equilibration. Because the
heat conduction time is much longer than that for equilibration, the heat losses are negligible and maximum ion surface
temperature can be found from the energy conservation as follows:
k
B
T
max
(x = 0) =
2E
abs
3(n
e
+n
a
)
. (5.17)
We assume here that the ion heat capacity also has the ideal gas value as we suggested for electrons. Thus the maximum
electron temperature is always higher than the ion temperature:
T
e,max
T
i,max
x=0
=
(n
e
+n
a
)
n
e
. (5.18)
The electron temperature is two times higher the ion temperature for one conductivity electron per atom. If the maximum
ion temperature is below the binding energy then the contribution of thermal ablation may only occur after the time
necessary for the building up the high-energy tail in the Maxwell distribution for ions. This case will be discussed later
in this section.
5.4. Mechanisms of ablation by ultra-short laser pulses
An atom (ion) can be removed from a solid if its total energy exceeds the binding energy, which is the energy of
vaporisation per particle,
tot
>
b
. The kinetic energy of a free particle should be
kin
=
tot
b
> 0 allowing the atom(ion)
to leave the solid. On the basis of the previous analysis one can show that three modes of ablation are possible, depending
on the relations between the absorbed energy density at the end of the pulse, the pulse duration, and the electron-to-ion
energy transfer time.
Let us outline three major ablation mechanisms, in a sequence from higher to lower laser intensity (Gamaly
et al., 2002); see also Zavestovskaya et al. (2000, 2008). The extreme non-equilibrium and non-thermal mechanism of
material removal, the electrostatic ablation, takes place when electron temperature exceeds some threshold defined
in the following section. The second ablation mode, the non-equilibrium ablation, occurs after the electron-to-
ion energy transfer. Non-equilibrium ablation takes place if electrons deliver the energy in excess of the binding
energy T
ion
>
b
to ions, and if it happens faster than thermodynamic equilibrium can be established, i.e. faster
than the time period required for building up, through collisions, the high-energy tail in the Maxwell distribution.
The ions energy distribution function in this period is still far from equilibrium. A purely thermal evaporation
may contribute to the total outcome after establishing the full Maxwell distribution. In either case the removal
of atoms requires the atom (ion) to acquire energy above the binding energy. These three cases are considered
below.
5.4.1. Electrostatic ablation: ions pulled out by energetic electrons
The electrons in the skin layer can gain the energy exceeding the threshold energy required to leave a solid target, which
is a work function,
esc
, for electrons in metals during the pulse time. The excited solid emits electrons while remaining
intact. The energetic electrons with T
e

b
escape the solid and create a strong electric field due to charge separation with
the parent ions. We are considering here the processes during the period longer than characteristic time for the electrons
oscillations, t
1
pe
0.1 fs, when electron inertia can be ignored (see Section 2). Then the electrostatic field of charge
separation between electrons and ions expresses through the gradient of the electronic pressure as follows:
E
elst
=
P
e
en
e
. (5.19)
The same electrostatic field enters into momentumequation for ions. The ions are cold during the pulse, thus one can ignore
the ion pressure force. During the pulse time the ponderomotive force of the laser electric field and polarisation force in
general should be included. Thus the equation of motion for ions reads:
M
u
i
t
=
P
e
n
i
+F
pond
+F
pol
. (5.20)
Here the charge conservation is taken into account: n
e
= Zn
i
. It is easy to show(Gamaly et al., 2002) that the ponderomotive
and polarisation forces are comparable to the electrostatic force at the very beginning of the laser pulse. However, when
electrons gain enough energy both forces can be neglected in comparison with the electrostatic force for the calculation of
the ablation threshold.
The electrostatic force, F
elst
= eE
elst
= P
e
/n
e
can be evaluated under the assumption that the density gradient of
escaping electrons is inverse proportional to the Debye length, l
D
= v
e
/
pe
(and taking for simplicity that Z = 1, so that
n
e
= n
i
). It is also necessary to account for the electrons energy losses for the work function, hence taking the electron
velocity as:
v
e
=
_
2(T
e
esc
)
m
e
_
1/2
.
Taking into account Eq. (5.16), the electrostatic force near the solidvacuum boundary at the end of the pulse reads:
F
elst
=
P
e
n
e
= k
B
T
e
ln n
e

2AF
0
n
e
l
s
l
D
. (5.21)
This force depends explicitly on time through the electron temperature, thus it reaches its maximumat the end of the pulse.
The electrostatic force of Eq. (5.21) pulls the ions out off the solid if the electron energy is larger than the sum of binding
energy plus work function. The maximum energy of ions dragged from the target reaches:
i
(t) = Z
e
= k
B
T
e

esc

b
(for Z = 1). One can see that the electron temperature of T
e
=
esc
+
b
sets the threshold for ablation of metals. Note
that the maximum electron temperature in energy units at the ablation threshold for metals is approximately twice the
binding energy. An atomis separated froma solid if it is dragged away to the distance larger than the Debye length, l
D
u
i
t.
The time necessary to move an ion to this distance could be estimated with the help of the equation of motion for ions
Eq. (5.20) with the force from Eq. (5.21) corrected for the energy losses for electron (work function) and ion detachment
(binding energy) from a solid. However, the simple and reasonable estimate can be made taking the maximum ion velocity
as u
i
[2(k
B
T
e,max
esc
)/M
i
]
1/2
and, correspondingly, the electron velocity as v
e
[2(k
B
T
e,max
esc
)/m
e
]
1/2
. Then
the time required for the removal of an ion from the target, the ion ablation time, reads:
t
abl

l
D
u
i
2
3
1/2
pe
_
M
i
m
e
_
1/2
_
T
e
esc
T
e
esc
b
_
1/2
. (5.22)
This qualitatively correct expression formally diverges at the ablation threshold, T
e
=
esc
+
b
, indicating the absence of the
removal of the ions as the ablation time goes to infinity. However, when the laser fluence even slightly exceeds the ablation
threshold, this time becomes comparable and even shorter then the pulse duration. For example, for Copper (M
i
= 63.5 a.u.;
pe
= 1.64 10
16
s
1
) at k
B
T
e
=
esc
+
b
+ (taking = 0.2
b
) this time comprises 57.6 fs, while at the extreme case
of k
B
T
e

b
it equals to 23.5 fs. It should be noted that for high intensities well over the ablation threshold the energy
conservation equations should include the energy losses for the ion heating, and for electron ionisation and emission. This
effect of electrostatic acceleration of ions is well known from the studies of the plasma expansion (Lifshitz and Pitaevski,
1981) and ultra-short intense lasermatter interaction (Luther-Davies et al., 1991, 1992; Gamaly, 1993).
Under the non-equilibrium conditions close to and above the ablation threshold, electrostatic ablation is the only
mechanism for atoms removal during the laser pulse. Two processes are responsible for terminating the electrostatic
ablation: a space charge that builds up in the plasma plume and two-dimensional effects associated with the plume
expansion. Following the above estimates, the characteristic time for the electrostatic ablation comprises several tens of
femtoseconds.
5.4.2. Non-equilibrium, semi-thermal ablation (k
B
T
ion

b
)
Removal of ions by escaping electrons becomes impossible when the maximum electron temperature is below the
electrostatic ablation threshold. The maximum electron temperature is more than two times larger than the maximum
ion temperature, T
e,max
/T
i,max
x=0
= (n
e
+ n
a
)/n
e
, depending on the number of conduction electrons per atom. Therefore
the removal of atoms is possible only after the transfer of energy from electrons to ions. That usually takes several
picoseconds and completed well after the end of the pulse. Thermal losses are negligible during much longer time than
the electronion energy equilibration. The surface atoms can leave a solid when the average ion temperature in the surface
layer becomes comparable to the binding energy. This process can contribute to the total ablation only if the maximum
electron temperature is 23 times higher than the binding energy. Note that this ablation mode can be efficient even if
only the main part of equilibrium distribution (no high energy tail) is established. Therefore, this mode can be termed as
semi-thermal and non-equilibrium one.
5.4.3. Thermal evaporation (k
B
T
ion
<
b
)
Finally, the ionion collisions establish full equilibrium distribution including the high-energy tail and this requires the
occurrence of many collisions. In plasma, where the Coulomb collisions dominate, the tail building time is (
b
/k
B
T
ion
)
3/2
times longer than that for establishing the temperature. Then conditions in ablating target and in the vapour next to a
solid surface are close to those in equilibrium; the ablation characteristics can be estimated by the results taken from
conventional thermodynamics. The ablation rate in these conditions can be estimated on the basis of thermodynamics as
for the equilibrium evaporation rate of a solid into vacuum in accord with formulae presented in the first paragraphs in this
section. However, caution should be exercised when applying thermodynamic relations to laser ablation with short pulses
because conditions in the expanding plume are usually far from equilibrium.
5.5. Single pulse ablation thresholds
The ablation threshold is defined as a minimum amount of energy that initiates the material removal process. The
definition of the ablation threshold relates to the obvious physical limit independently of any subjective views: at least one
atom should be removed from the target surface to mark a beginning of ablation process. In real experiments the focal spot
diameter is much larger than the absorption depth, which is usually of a fewtens of nanometres. Therefore the description of
the process as one-dimensional is a good approximation. We define the ablation threshold as amount of energy necessary for
removal of one mono-atomic layer. We showlater that such definition complies well with the classical theory of evaporation
and with the experimental results. Note that the thickness of the monolayer is comparable to the mean free path of atomic
collisions in a solid: l
mfp
= (n)
1
(10
23
10
15
)
1
10
8
cm. Therefore, only kinetic approach should be used for the
description of any phenomena in the outermost surface layer, even in the close to thermal equilibrium conditions.
It is well known that the surface atoms are loosely bound to the bulk making part of bonds dangling or saturated with
foreignatoms (Zangwill, 1988; Prutton, 1994). The removal of atoms fromthe surface layer at the ablationthresholddepends
on the energy distribution in the outermost surface layer. The energy distribution is responsible for the relative contribution
of non-equilibriumablation and thermal evaporation. The establishment of quasi-equilibriumdistribution in the outermost
surface layer also strongly depends on the environment: either the solid is placed in vacuum or it is in an ambient gas of
specific atomic content and pressure.
The question that nowarises is whether the time to create the Maxwell distribution in the bulk also applies to the surface
layer. The atoms in the outermost surface layer next to vacuum are in fact in a different condition compared to the atoms
in the bulk. Below we consider the processes responsible for the removal of swiftly heated atoms from the surface layer
into vacuum, and consider relative contribution from thermal and non-thermal processes near the ablation threshold. First,
the process of non-thermal ablation in vacuum is considered. Next, the processes of the energy transfer from the bulk to
the outermost surface layer (bulk-to-surface transfer time) are accounted for. These processes appear to be important for
understanding the experimentally measured ablation thresholds with the pulse duration transitional from short pulse to
long pulse regimes. Then, it is demonstrated that the presence of an ambient gas (air) next to the ablating solid leads to a
decrease of the ablation threshold due to significant contribution of thermal evaporation.
5.5.1. Ablation thresholds in vacuum
Let us consider first the thresholds for the electrostatic ablation in extreme non-equilibrium conditions for metals and
dielectrics (Gamaly et al., 2002).
(a) Ablation thresholds for metals
The minimum energy that electron needs to escape the solid equals to the work function. In order to drag cold ion out
of the target by the force of electrostatic field the electron must have an additional energy equal to or larger than the ion
binding energy. Hence, the ablation threshold for metals can be defined as the condition for the electron energy to reach,
in a mono-atomic surface layer d l
s
, the value equal to the sum of the atomic binding energy and the work function by
the end of the laser pulse. Using Eqs. (5.16) and (5.17) for the electron temperature one obtains the energy condition for the
ablation threshold:
e
=
b
+
esc
=
4
3
AF
0
l
s
n
e
.
The threshold laser fluence for ablation of metals is then defined as:
F
m
th
=
3
4
(
b
+
esc
)
l
s
n
e
A
. (5.23)
It is assumed that the number density of the conductivity electrons is unchanged during the lasermatter interaction process
and electrons heat capacity equals to that of ideal gas. In dense plasma the approximate relation holds (see Eq. (5.15)):
A/l
s
4C
n
/. Hence the threshold fluence is proportional to the laser wavelength: F
m
th
. We demonstrate below that
this relation agrees well with the experimental data.
(b) Ablation threshold for dielectrics
The ablation mechanism for the ionised dielectrics is similar to that for metals. However, there are several distinctive
differences (Gamaly et al., 2002). First, an additional energy is needed to create the free carriers, i.e. to transfer the electron
from the valence band to the conductivity band in order to increase the absorbed energy density to the level needed for
ablation. Therefore, the energy equal to the ionisation potential J
I
, should be delivered to the valence electrons. Second,
the number density of free electrons depends on the laser intensity and on time during the interaction process as has been
shown above. However, the intensity necessary for ablation significantly exceeds the breakdown threshold, therefore the
first ionisation is completed well before the end of the pulse. The number density of free electrons then saturates at the level
n
e
n
a
, where n
a
is the number density of atoms in the target. Under these assumptions the threshold fluence for ablation of
dielectrics F
d
th
is defined as follows:
F
d
th
=
3
4
(
b
+J
i
)
l
s
n
e
A
. (5.24)
Therefore, in general the ablation threshold for dielectric in the ultra short lasermatter interaction regime is higher than
that for the metals. The absorption in the ionised dielectric also occurs in a skin layer; thus one can use the relation
A/l
s
4C
n
/ for estimates of absorbed energy density in dielectrics. The ablation thresholds of Eqs. (5.23) and (5.24)
do not depend explicitly on the pulse duration and intensity. However it is just a first order approximation. Certain, though
weak, dependence is hidden in the absorption coefficient and in the number density of free electrons.
(c) Link between the short pulse and the long pulse regimes of ablation
Many experimental and theoretical studies of the ablation threshold and the ablation rate for metals irradiated with laser
pulses clearly demonstrate the presence of two different ablation regimes depending on the pulse duration (Afanasiev and
Krokhin, 1971; Anisimov et al., 1971; Malvezzi et al., 1986; Luther-Davies et al., 1992; Nolte et al., 1997; Eidmann et al.,
2000; Gamaly et al., 2002). The ablation threshold for ultra-short pulses is practically independent of the pulse duration.
The ablation with pulses longer than 100 ps proceeds in equilibrium conditions because the electronion temperature
equilibration establishes early in the pulse time. The heat conduction and hydrodynamic motion dominate the long pulse
ablation process, t
pulse
> t
ei
, t
therm
. Therefore long pulse ablation threshold becomes directly dependent on the pulse
duration. The heat conduction depth exceeds the field penetration depth l
therm
l
s
/2. The ablation threshold for this case
is defined by condition that the absorbed laser energy, AF, is fully converted into the energy of broken bonds in a layer with
the thickness of the heat diffusion depth l
therm
=
_
D
th
t
p
_
1/2
during the laser pulse (Afanasiev and Krokhin, 1971; Anisimov
et al., 1971):
AF

= (D
th
t
p
)
1/2
b
n
a
.
Here D
th
is the thermal diffusion coefficient. The ablation threshold fluence for the long pulses (>100 ps) immediately
follows from this equation:
F
th

(D
th
t
p
)
0.5
b
n
a
A
. (5.25)
This is a well-known square-root time dependence of ablation threshold on the pulse duration. The difference in the
ablation mechanisms for the thermal long pulse regime and the non-equilibrium short pulse mode is two-fold. Firstly, the
laser energy absorption mechanisms are different. The intensity for the long pulse interaction is in the range 10
8
10
9
W/cm
2
for the pulse duration range fromnanoseconds to several tens of picoseconds. The ionisation is negligible, and the dielectrics
are almost transparent up to the UV-range. The absorption is weak, and it occurs due to the inter-band transitions, defects
and excitations. At the opposite limit of the femtosecond lasermatter interaction the intensity is in excess of 10
13
W/cm
2
and any dielectric is almost fully ionised in the interaction zone. Therefore, the absorption due to the mechanisms of inverse
Bremsstrahlung andresonance absorptiononfree carriers dominates the interaction, andthe absorptioncoefficient amounts
to several tens percent. Secondly, the electron-to-lattice energy exchange time in a long-pulse ablation mode is shorter than
the pulse duration. By these reasons, the electrons and ions are in equilibrium, and ablation has a conventional character of
thermal evaporation.
An intermediate regime occurs for the conditions when the pulse duration is longer than electronion energy transfer
time and heat conduction time, t
p
> (t
heat
, t
ei
). This regime holds for the laser pulse durations t
p
> 100 ps and at the
Fig. 5.2. An example showing the difference in the number of close neighbours for an atom in the bulk and at the surface: twelve close neighbour atoms,
coloured in dark blue, for an atomin the bulk (left) and only four close neighbours for an atomon the surface (right) in a face-centred cubic (fcc-) structure.
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
intensities less than 10
11
W/cm
2
, where the heat conduction should be taken into account during the pulse. The thickness
of a target layer heated during a pulse longer than a picosecond becomes l
therm
+ l
s
/2. The ablation threshold for this case
can be obtained with the help of Eq. (5.23), by replacing l
s
/2 with l
therm
+l
s
/2:
F
m
th

3
2
(
b
+
esc
)
n
e
A
_
l
s
2
+l
heat
_
. (5.26)
The long-pulse threshold fluence of Eq. (5.25) immediately follows from this equation in the limit l
therm
l
s
/2. Thus,
formulae for the ablationthresholds for the ultra-short andfor the long laser pulses are naturally linkedthrougha continuous
function of the pulse duration. However, the experimentally observed transition from the ablation threshold for the non-
equilibrium regime to that for the thermal regime occurs at much longer pulse durations, for example, at up to 100 ps in
gold (Stuart et al., 1996; Gamaly et al., 2002). This indicates that the thermal mechanismfor some reason does not contribute
to the ablation rate at fluences near threshold, as might be expected, even when the pulse-width is up to ten times longer
than the electronlattice equilibration time.
It was found (Gamaly et al., 2004) that the time to establish the high-energy tail of the Maxwell energy distribution of
atoms at the surface must be considered along with the time for equilibration of the electron and lattice temperatures.
Specifically in vacuum, the time needed to transfer energy from the high-energy Maxwell tail from atoms in the bulk to the
atomic layer at the surface (bulk-to-surface energy transfer time t
bs
) becomes the crucial parameter that determines the
relative contribution of equilibrium(thermal) evaporation and non-thermal ablation to the material removal rate, especially
near the ablation threshold. The analysis, therefore, suggests that thermal ablation will only dominate when the pulse
duration is comparable to or longer than the bulk-to-surface energy transfer time. This problemis addressed in the following
section below.
5.5.2. Long-lived non-equilibrium state in the outermost surface layer
The energy distribution in the outermost surface layer is responsible for the relative contributions of non-equilibrium
ablation and thermal evaporation. Atoms fromthis layer will immediately leave a solid if energy in excess of binding energy
is instantaneously deposited into this layer. This is the process of non-equilibriumablation considered above. In equilibrium
conditions, the evaporation can proceed at a much lower temperatures than that corresponding to the binding energy due
to the existence of high-energy atoms in the Maxwellian tail with
b
. However the presence of the free surface prevents
the equilibrium from being established in the surface layer itself whose thickness is comparable to the mean free path for
atomic collision. Indeed, if the energy of an atom in this layer reaches the binding energy due to collisions with the atoms
fromthe bulk, this atomimmediately escapes fromthe solid. Thermal evaporation fromthe surface heated to a temperature
below the binding energy can therefore, proceed only when energy is supplied to the surface from the bulk via atom-to-
atom collisions. Thus, the time for the energy to increase from = k
B
T <
b
to
b
in the surface layer (that is the
bulk-to-surface energy transfer time, t
bs
) determines the onset of the thermal evaporation at solidvacuum interface. This
time is analogous to the time needed to establish the Maxwell tail in isotropic conditions in the bulk. The probability of
energy transfer from the bulk to the surface can be found from a solution of the time-dependent two-dimensional kinetic
equation, which is a formidable problem. However, one can make a reasonable estimate as follows.
(a) Bulk-to-surface energy transfer
It is clear that the probability of energy transfer in excess of
b
from atoms in the bulk to those at the surface should
be lower than that between atoms in the bulk due to a decrease in the number of close neighbours around the surface
atom capable of such energy transfer. For example, the number of close neighbours (n
bulk
) equals six in the bulk of a closely
packed solid whereas the number of close neighbours from the bulk (n
surf
) for a surface atom is only one because the other
four closest neighbours are also surface atoms (see Fig. 5.2).
The number of collisions required to increase the energy of a surface atom, N
surf
, will need to be larger compared to that
in the bulk, N
surf
bN
bulk
where b n
bulk
/n
surf
. Correspondingly, the probability of energy transfer from the bulk to the
surface should be lower: W
bulk
W
N
bulk
1
; W
bs
W
N
surf
1
W
b
bulk
; here W
1
is a probability of a single independent collision
resulting in energy transfer. Then the cross section for a collision between atoms in the bulk with the surface atoms can be
presented in the form:
bs
=
0
W
bs

0
W
b
bulk
. (5.27)
Now one can arrive to the following estimate for the cross section for the bulk-to-surface energy transfer:
bs

0
W
bs
(k
B
T
b
)
0
exp
_
b

b
k
B
T
_
. (5.28)
The bulk-to-surface energy transfer time thus reads:
t
bs
= [n
a
v
bs
]
1
t
main
exp
_
b

b
k
B
T
_
. (5.29)
According to Eq. (5.29), the bulk-to-surface energy transfer time increases dramatically with decreasing temperature.
One can see that the bulk-to-surface energy transfer time exceeds markedly the electron-to-lattice energy transfer time and
the heat conduction time at fluences that are below the threshold for non-thermal ablation. In other words, as the surface
starts cooling by thermal conduction, the bulk-to-surface energy transfer time increases to such extent that the surface
atoms cannot gain energy above the binding energy. Hence, thermal evaporation does not occur.
(b) Ablation threshold for intermediate pulse duration: contribution of thermal evaporation at t > t
bs
It was experimentally observed that pulse-width-independent ablation threshold predicted for the non-equilibrium
regime holds at unexpectedly large pulse durations, for example, up to 100 ps in gold where the thermal regime is
expected to dominate. However, as it follows from the preceding analysis the time for establishing high-energy tail in the
outermost surface layer is crucial for correct calculation of thermal evaporation contribution. Thus the thermal evaporation
contributes in the total outcome if the pulse duration is longer than the bulk-to-surface energy transfer time responsible for
the establishing the full Maxwell distribution.
The total ablation rate is the sum of contributions from non-equilibrium mechanism at t < t
bs
and thermal ablation
at t > t
bs
provided that the threshold condition for the non-thermal ablation in vacuum is achieved. To quantify this
process, let us consider the relative contribution from thermal and non-thermal ablation mechanisms in vacuum when the
non-thermal threshold condition is satisfied. The outermost atomic layer, where T
max

b
, is removed, thus the ablation
depth equals the inter-atomic distance, d. Thermal ablation starts after a time t
bs
when the energy in excess of the binding
energy is delivered to the surface layer from the bulk through atomic collisions. The depth of material removed by thermal
evaporation can be expressed through the time- and space-dependent distribution function as follows:
l
th
=
1
n
a

t
bs

0
vf ( v, t)d
3
vdt. (5.30)
The transient distribution function differs from the equilibrium one only by the high-energy tail. Therefore, the
average atomic number density, n
a
=
_
0
f ( v, t)d
3
v, and the average velocity, v[T(t)] =
_
0
vf ( v, t)d
3
v, are close to their
equilibriumvalues. The number density of evaporating atoms (analogous to the saturated density of vapour in equilibrium)
can be approximated as n n
a
exp(b
b
/T). Then the evaporation depth in Eq. (5.30) is expressed as follows:
l
th

t
bs
_
2T
M
_
1/2
exp
_
b
b
T
_
dt. (5.31)
The temperature decreases in accordance with linear heat conduction T = T
bs
(t
bs
/t)
1/2
; T
bs
T(t
bs
). The latter
expresses as follows:
T
bs
= T
m
_
t
th
t
th
+t
bs
_
1/2
. (5.32)
Then, Eq. (5.39) can be immediately integrated to obtain (see Appendix D):
l
th
t
bs
_
2T
bs
M
_
1/2
k
B
T
bs
2
b
exp
_
b

b
k
B
T
_
. (5.33)
The maximum temperature at the end of the pulse in the absence of losses is proportional to the total absorbed fluence,
T
m
F
a
. Hence, thermal evaporation depth scales with the absorbed fluence as l
th
F
3/2
a
exp(const/F
a
).
A conservative estimate can be made by taking the maximumsurface temperature at the non-thermal ablation threshold
T
m

b
, t
main
= 0.2 ps, b = 3. Then the bulk-to-surface energy transfer time appears to be much larger the heat conduction
time: t
bs
80 ps t
th
30 ps, the surface temperature decreases two times while the high-energy tail builds up,
T
bs
= 0.52T
m
. Assuming
b
3 eV; M = 60 a.u. one gets v 2.2 10
5
cm/s. Calculation by Eq. (5.33) gives that thermal
evaporation depth much smaller of the monolayers thickness, l
th
4.410
11
cm d
a
. Thus, non-equilibriumablation in
vacuum completely dominates thermal evaporation for pulse durations shorter than bulk-to-surface energy transfer time.
In other words, in vacuum, thermal evaporation at the ablation threshold and below is completely negligible for t
pulse
< t
bs
.
This result explains the experimentally observed fact that the ablation threshold independence on the pulse duration holds
for up to 100 ps in non-equilibrium ablation.
5.5.3. Ablation thresholds in ambient gas
The experiments revealed that the ablation thresholds for several metals in air are less than half of those measured in
vacuum (Gamaly et al., 2005). The analysis shows that this difference is caused by the existence of a long-lived transient
non-equilibrium surface state at the solidvacuum interface. The energy distribution of atoms at the surface is Maxwell-
like but with its high-energy tail truncated at the binding energy. It was shown in the previous section that in vacuum
the time needed for energy transfer from the bulk to the surface in order to build the high-energy tail exceeds other
characteristic timescales such as the electronion temperature equilibration time and surface cooling time. This prohibits
thermal evaporation in vacuumfor which the high-energy tail is essential. In air, however, collisions between the gas atoms
and the surface markedly reduce the lifetime of this non-equilibriumsurface state allowing thermal evaporation to proceed
before the surface cools down. In order to understand these differences we shall consider how the presence of air can effect
thermal evaporation, the only process that can occur below the vacuum ablation threshold. The question is how is thermal
evaporation turned on by the presence of air when we concluded it is negligible in vacuum? It is worth noting that the
analysis below applies to any ambient gas with different from air molecular content and pressure.
After the laser pulse, the air next to the heated surface layer gains energy through collisions with the solid target. This
results in the establishment of a Maxwell distribution in the air near the airsolid interface. Hence it is possible that air
plays the same role as the saturated vapour in classical thermal evaporation. The presence of air introduces a new pathway
allowing the creation of the high-energy tail of the Maxwell distribution in the surface layer augmenting the bulk-to-surface
energy transfer process discussed earlier. Thus, there are now three processes acting at the same time which determine the
ablation conditions at the solidair interface:
(i) Evolution of the Maxwell distribution at the surface due to airsolid collisions;
(ii) Evolution of the Maxwell distribution at the surface due to bulk-to-surface energy transfer, and
(iii) Cooling of the surface layer by heat conduction.
We concluded that mechanism (ii) was too slow to result in thermal evaporation when T <
b
. Therefore the presence
of air significantly increases thermalisation rate reducing the time for building up the high-energy tail at the surface. The
ablation rate then can be calculated using the equilibrium relations. Let us consider all these processes in sequence.
The airsolid equilibrium energy distribution is established by collisions of air molecules with the solid. The gas-kinetic
mean free path in air in standard conditions is l
gk
= 610
6
cm(Zeldovich and Raizer, 2002). Therefore, the equilibration
time t
eq
needed to establish a Maxwell distribution in the gas can be estimated as t
eq
t
gk
l
gk
/v
th
, where v
th
is the
average thermal velocity in air. We estimate this time at room temperature (v
th
= 3.3 10
4
cm/s) as t
eq
1.8 10
10
s.
The bulk-to-surface energy transfer time calculated by the Eq. (5.29) at the maximum temperature (k
B
T
max

b
/2) for the
threshold fluence conditions in air constitutes t
bs
t
main
e
12
30 ns t
eq
for Cu, Al, and Fe after the pulse. Thus, only
the air-surface collisions could lead to the formation of the high-energy Maxwell tail, and therefore to thermal evaporation
from the surface.
The evaporation rate can be calculated in the following way. During the time when the solidair temperature
equilibration is completed the surface temperature has dropped due to thermal conduction to the level T
eq
T
m
(t
p
/t
eq
)
1/2
.
Here T
max
F
abs
(absorbed fluence) is the maximum temperature at the end of the laser pulse at the experimentally
determined threshold fluence for ablation in air. These values were presented in Table 5.4. Thermal evaporation starts after
the equilibration time t > t
eq
and the temperature at the solidair surface continue to decrease in accordance to the linear
heat conduction law. We suggest that thermal evaporation proceeds at a vapour density corresponding to the temperature
at the solidair interface. The number of atoms ablated per unit area after establishing the Maxwell equilibrium can be
estimated as follows:
nvt
therm
=

t
eq
(nv)
therm
dt. (5.34)
A reliable estimate of the evaporation rate can be obtained with the numerical coefficients extracted from the known
experimental data in equilibrium at the temperature close to our experimental conditions. We also assume that the
vapourair mixture with a predominance of air plays the role of the saturated vapour over the ablated solid. Thus, we take:
(n
a
v)
therm

n
air
k
B
T
eq
(2M
a
k
B
T
eq
)
1/2
.
Then, Eq. (5.34) takes the following form:
nvt
therm
=

t
eq
(nv)
therm
dt n
air
_
k
B
T
eq
2M
a
_
1/2
t
eq
_
atoms
cm
2
_
. (5.35)
Table 5.3
The predicted numbers of atoms thermally evaporated after the pulse once the Maxwell distribution has been established compared with
the number of atoms in a monolayer.
Al Cu Fe Pb
nvt
therm
, 10
15
cm
2
; Eq. (5.35) 2.4 5.28 1.67 0.45
n
a
d
mono
, 10
15
cm
2
(number of atoms in a monolayer) 1.72 2.16 2.0 1.15
Fig. 5.3. Threshold laser fluence for ablation of gold targets (mirror and grating) is plotted versus laser pulse duration.
Source: The experimental points are from Stuart et al. (1996). Theory (dotted line) is from Gamaly et al. (2002).
The ablation threshold is achieved when the above number of thermally evaporated atoms equals to the number of atoms
in the mono-atomic surface layer n
a
d
mono
:
nvt
therm
= n
a
d
mono
. (5.36)
The above expression combines air, laser and material parameters at conditions of ablation threshold. The values predicted
by Eq. (5.35) are presented in Table 5.3 for comparison with the areal density of a monolayer. It is clear that the predicted
number of the thermally ablated atoms is, in fact, close to the number of atoms in a mono-atomic layer.
Table 5.3 suggests that thermal evaporation well after the end of the laser pulse can be responsible for the removal of
a mono-atomic layer for Al, Cu, and Fe at the fluences corresponding to the threshold measured in air (see Appendix D).
This is in a good agreement with the experiments, as the threshold fluence was introduced as the fluence needed to remove
a single atomic layer. Therefore, we can conclude that the presence of air decreases the single pulse ablation threshold by
approximately a factor of tworelative tothe vacuumcase due tothe contributionof thermal ablationassistedby the presence
of the air well after the end of the pulse. Numerical studies of this effect were recently presented in Bulgakova et al. (2008).
5.5.4. Comparison with the experimental data
Now we compare the above theory to the different experimental data. We present the full span of pulse durations, from
the femtosecond to the nanosecond range, for ablation of metals and dielectrics, where the data are available. First, we
present data on ablation in vacuum and then we compare the theoretical predictions to the experiments of ablation in air.
(a) Ablation thresholds in vacuum
Metals: Let us apply Eq. (5.23) for calculation of the ablation threshold for Copper (density 8.96 g/cm
3
, binding energy
b
= 3.125 eV/atom, work function
esc
= 4.65 eV/atom, n
a
= 0.845 10
23
cm
3
) ablated by 780 nm laser. The
calculated threshold F
th
0.51 J/cm
2
(for A

= 1) is in agreement with the experimental number 0.50.6 J/cm
2
given
in Momma et al. (1997), though the absorption coefficient was not specified there. For the long pulse ablation Eq. (5.25)
predicts F
th
(J/cm
2
) = 0.045 (t
p
(ps))
1/2
. The thermal diffusion coefficient of Copper equals to D = 1.14 cm
2
/s.
For Gold target
b
= 3.37 eV/atom,
esc
= 5.1 eV, n
e
= 5.9 10
22
cm
3
. The ablation threshold for Gold evaporated by
1053 nm laser calculated with help of Eq. (5.23) equals to F
th
= 0.5 J/cm
2
. That number compares well to the experimental
value of 0.45 0.1 J/cm
2
(Stuart et al., 1996). For the long pulse ablation (assuming the absorption coefficient of A = 0.74)
Eq. (5.25) gives F
th
(J/cm
2
) = 0.049 (t
p
(ps))
1/2
. The thermal diffusion value of D = 1.3 cm
2
/s for Gold corresponds
to equilibrium conditions with ion-dominated heat capacity. The experimental points from Stuart et al. (1996) and the
calculated curve are presented in Fig. 5.3.
Silica: The ablation threshold for silica evaporated by laser with = 1.053 m ( = 1.79 10
15
s
1
; l
s
/A
1.6 10
5
cm) is calculated with the help of Eq. (5.24) taking n
e

= 7 10
22
cm
3
, and (
b
+ J
i
)

= (3.7 + 13.6) eV
(Sosman, 1965). This threshold equals to F
th
= 2.35 J/cm
2
, which is in a qualitative agreement with the experimental
Fig. 5.4. Threshold fluence for laser ablation of fused silica target as a function of the laser wavelength for sub-picosecond laser pulses.
Source: The experimental points are from Perry et al. (1999) and Stuart et al. (1996). Theory (solid line) is from Gamaly et al. (2002).
Fig. 5.5. Threshold laser fluence for ablation of fused silica target versus laser pulse duration.
Source: The experimental points are from Perry et al. (1999) and Stuart et al. (1996). Theory (solid line) is from Gamaly et al. (2002).
results 2 0.5 J/cm
2
(Perry et al., 1999). Eq. (5.24) with relation l
s
/A taken into account also predicts the correct
wavelength dependence of the threshold: F
th
= 1.84 J/cm
2
for = 825 nm and F
th
= 1.17 J/cm
2
for = 526 nm
(cf. Fig. 5.4). The experimental threshold fluences for the sub-picosecond laser pulses (Perry et al., 1999) are: 22.5 J/cm
2
( = 1053 nm), 2 J/cm
2
( = 825 nm), and 1.21.5 J/cm
2
( = 526 nm).
Using the following parameters for the fused silica at wavelength of 825 nm( = 0.0087 cm
2
/s,
b
= 3.7 eV/atom; n
a
=
0.710
23
cm
3
; and A 310
3
) one obtains a good agreement with the experimental data collected in Perry et al. (1999)
for the laser pulse duration from 10 ps to 1 ns. For the long pulse regime Eq. (5.25) holds: F
th
(J/cm
2
) = 1.29 (t
p
(ps))
1/2
(see Fig. 5.5).
The ablation threshold of 4.9 J/cm
2
for a fused silica with the laser t
p
= 5 fs, l = 780 nm, intensity 10
15
W/cm
2
has
been reported in Lenzner et al. (1999). This value is almost three times higher than that of Perry et al. (1999) and Stuart et al.
(1996) and from the prediction of Eq. (5.24). However, the absorption coefficients as well as the pre-pulse to main pulse
contrast ratio were not specified in the paper (Lenzner et al., 1999).
In the paper of Sokolowski-Tinten et al. (1998) the crater depth of 120 nm was drilled in a BK7 glass by a 100 fs 620 nm
laser at the intensity 1.510
14
W/cm
2
. Assuming that the number density, the binding and the ionisation energy in the BK7
glass target are the same as in fused silica, the Eq. (5.32) for the ablation threshold predicts the value of 1.0 J/cm
2
. This is in
a qualitative agreement with the number measured in Sokolowski-Tinten et al. (1998): F
th
= 1.4 J/cm
2
. Eq. (5.24) predicts
F
th
= 1.34 J/cm
2
for this experiment.
It should be noted that the definition of the ablation threshold implies that at the threshold condition at least a mono-
atomic layer x l
s
, of the target material should be removed. As one can see from above comparison, the experimental
data on the ablation threshold determined this way are in excellent agreement with the theoretical formulae. It should be
particularly emphasised that there were no any fitting coefficients in the calculations presented here.
Table 5.4
Threshold fluence for ablation of metals by 12 ps pulses measured in air and in vacuum.
Metal, M
a
(a.u.) Al, 26.98 Cu, 63.54 Fe, 55.85 Pb, 207.19
F
thr
in air, (J/cm
2
) 0.17 0.03 0.23 0.03 0.19 0.02 0.008 0.002
F
thr
in vacuum, (J/cm
2
) 0.32 0.04 0.41 0.05 0.36 0.04 0.08 0.02
(b) Comparing ablation of metals in air and in vacuum
The experimental observations that the ablation thresholds of metals in air and in vacuum in similar irradiation
conditions are significantly different were reported by Gamaly et al. (2004). These experiments were carried out with laser
pulses generated by a 50 W long-cavity mode-locked Nd:YVO
4
laser (Luther-Davies et al., 2004; Kolev et al., 2003) using a
number of Al, Cu, steel (Fe), and Pb targets. The samples were exposed to 12 ps 532 nm pulses at a pulse repetition rate of
4.1 MHz; the energy per pulse on the target surface was E
p
= 6.5 J. The energy per pulse and the pulse duration were
fixed, whilst the energy density (fluence) was varied from 5.4 10
2
to 1.3 J/cm
2
(or, of intensities from 4.2 10
9
to
1.0 10
11
W/cm
2
).
Since many pulses hit the same region of the target in succession at 4.1 MHz repetition rate, it is important to note
that provided the target cools completely between consecutive pulses, then should be no interaction between them. The
laser interaction with a target then proceeds in the same way as for a single pulse provided that any crater formed by the
preceding pulses is insignificant. This is the case near the ablation threshold. The characteristic cooling time (3 10
11
to
210
10
s) for laser heated skin layer in the metals studied in these experiments is much shorter than the time gap between
the pulses of 2.5 10
7
s. Therefore, target surface cools down completely between consecutive pulses and lasertarget
interaction near threshold proceeds in the single pulse mode. Another important condition was fulfilled in the ablation
threshold measurements: a high intensity contrast (the ratio between the peak pulse power to that of the background)
was maintained during ablation experiments to ensure that no surface modification occurs between the pulses due to, for
example, amplified spontaneous emission.
The ablation depth and mass per single pulse were obtained from the following procedure. Total amount of material
ablated over a 60 s period was measured by weighing the sample with the accuracy 10
4
g before and after the ablation.
The ablated mass per single pulse, m
av
, was determined by averaging the mass difference over the 2.46 10
8
pulses. Then
the ablation depth per single pulse was obtained as follows:
l
abl
=
m
av
S
f
. (5.37)
Here is the target mass density and S
f
is the focal spot area. The measured ablation depths for various ablated materials
as a function of the incident laser fluence are shown in Fig. 5.6.
The measured ablation depth as a function of fluence allows one to determine the ablation threshold. The focal spot
diameter is muchlarger thanoptical absorptiondepth, whichis a fewtens of nmcorresponding tothe skindepthof the metal.
Thus, ablation can be considered as a one-dimensional process. The fluence at the ablation threshold can be determined by
extrapolating the ablation depth dependence to the zero depth such as it was used in a number of reports (Stuart et al., 1996;
Nolte et al., 1997; Perry et al., 1999). However it appears that the threshold obtained this way may depend strongly on the
extrapolation procedure since the fluence dependence is not a simple linear function. Indeed, it is known from statistical
physics (Landau and Lifshitz, 1980a) that the relative average fluctuation in the number of particles in an open systemgrows
up as the average particles number goes to zero:
_
(N)
2
/N = 1/
N. Therefore the relative error in experiments trying to

measure the ablation threshold for a decreasing number of ablated atoms will increase. In practical terms repetition of an
ablation experiment at the same fluence close to the ablation threshold should produce randomly scattered results in terms
of particle removal. This is reflected by the fact that is impossible to find a physically justifiable extrapolation to zero depth.
As a result it seems reasonable to define the ablation threshold as the energy density needed to remove a single atomic
surface layer.
The ablation threshold fluence, F
thr
, derived in this manner from the experimental data for different metals is presented
in Table 5.4. We note that the threshold for Cu in vacuum, for example, of 0.41 0.05 J/cm
2
is in good agreement with the
results of Nolte et al. (1997) (F
thr
= 0.375 J/cm
2
for 9.6 ps and 0.423 J/cm
2
for 14.4 ps pulses).
These experiments were carried out using identical laser and focusing conditions so that the only significant variable was
the presence or absence of air. Thus, there is clear experimental demonstration of existence of non-equilibriumand thermal
ablation regimes.
5.6. Ablation rate, mass, and depth by a single pulse
The dependence of over-threshold ablation on laser and material parameters can be obtained on the basis of the
general features of the ultra-short pulse energy absorption discussed above. The electron temperature grows up linearly
in time (assuming the intensity is constant during the pulse) and its magnitude decreases exponentially inside a material:
T
e
(x, t) = (2I
o
t/C
e
n
e
) exp(2x/l
s
). Thus, the electron temperature at the outermost surface layer reaches its maximum to
the end of the pulse. At the threshold fluence, F
thr
, this outermost atomic layer is removed fromthe bulk by the electrostatic
Fig. 5.6. Ablation depth per pulse vs. fluence for Al; Cu; Fe; and Pb; in experiments in air (triangles) and in vacuum(circles) using 12 ps 4.1 MHz repetition
rate laser. The horizontal lines and the numbers above the lines correspond to inter-atomic distances, while the arrows indicate the ablation threshold.
The dashed lines are the upper limits for the ablated depth determined using the energy conservation law.
Source: Taken from Gamaly et al. (2004).
field of the electrons escaping froma solid after the end of the pulse. The total absorbed energy should be increased in order
to remove the next layer. Therefore, ablation at F
a
> F
thr
starts before the end of the pulse at t = (F
thr
/F
a
)t
p
. Note that
the temperature calculation presented above does not take into account the energy losses on the ablation during the pulse.
Optimistic but rather exaggerating estimate can be based on the assumption that all absorbed energy at the above threshold
fluence is spent on ablation (i.e. on breaking bonds plus kinetic energy). Then the ablation depth estimates on the basis of
energy conservation is as follows:
l
max
abl
l
mono
+
F F
thr
n(
b
+
esc
)
; l
mono
n
1/3
a
. (5.38)
The losses for kinetic energy of ablating atoms and for the heating of the rest of non-evaporating material should be
added into parenthesis in the denominator in Eq. (5.38). However, these losses are omitted because it is difficult to make a
reasonable estimate for them. The above estimate gives linear dependence on the absorbed fluence and an upper limit for
the ablated depth of material set by the energy conservation for any ablation mode.
Conservative (minimising) estimate of the ablation depth may be obtained from the spatial temperature dependence
under the assumption that ablation wave propagating inside the solid stops at the depth where the electron temperature
corresponds to that at the threshold fluence. The maximum electron temperature is proportional to the total absorbed
fluence, F
a
. Therefore the ablation depth can be estimated as:
l
min
abl

l
s
2
ln
_
F
a
F
thr
_
+l
mono
. (5.39)
The ablated plume in this estimate is overheated at F
a
> F
thr
, which is reflected in the weak ablation depth dependence on
the absorbed fluence. The ablated mass, which could be measured in experiments, is connected to the ablation length, the
laser focal spot area, S
foc
, the atomic number density, and the atomic mass, M, by an obvious relation:
m
abl
= S
foc
n
a
Ml
abl
, (5.40)
where l
abl
is in the range l
min
abl
< l
abl
< l
max
abl
given by Eqs. (5.38) and (5.39).
5.7. Control over the ablation rate and phase state of laser-produced plume: spatial pulse shaping
There are two large areas of applications of laser ablation (Bauerle, 2000; Eason, 2007; Miller and Haglund, 1998; Chrisey
and Hubler, 1994). One is in micromachining as a tool for making cuts, holes, in surgery etc. To make a perfect cut (hole) with
well defined size and sharp edges the ablation rate over the focal spot should be above the ablation threshold and of the same
value over the whole laser-affected surface. Another broad area of applications involves the use of the laser-ablated plume
for deposition of thin films, producing nanoclusters and nanocomposite materials, making a medium for different chemical
reactions etc. In these applications it is necessary to control the composition of the laser-produced plume. It is well known
from numerous experiments that laser plumes contain macroscopic particles and liquid droplets. This is a clear indication
that a variety of different processes take place during the ablation and the plume expansion. This is also a major obstacle for
many potentially attractive industrial applications of laser ablation. In this section we analyse the processes affecting the
phase composition of the plume. Then the recipes for controlling the composition of the plume and for achieving the fully
atomised plume by a single sub-picosecond laser pulse are presented along with their experimental verification.
There are two major reasons for a complex composition in the plume. The first relates to the spatial intensity variations
across the laser focal spot. There are varieties of processes that transform the state of a target exposed to intense
laser radiation, and the onset of all these processes depends on the incident laser intensity. The processes that appear
progressively with increase of laser intensity are structural phase transitions, destruction of the target integrity (appearance
of cracks, flaking of the surface), melting, ionisation and ablation. The spatial intensity distribution across the focal spot
usually has a Gauss-like form. Therefore the whole variety of above-mentioned processes takes place across the laser-
affected area even if the average spatial intensity is higher the ablation threshold. A plume produced by a laser with such
intensity variations can contain material in different phase states: gas, melt, etc.
Another source of non-homogeneity in the plume, particularly the formation of liquid droplets, is condensation of vapour
during the expansion. As a result, an expanding plume consists of a mixture of gas and liquid phases. Therefore, in order
to ensure a homogeneity of a plume one needs two conditions to be fulfilled: (1) the local fluence at every point across
the focal area should exceed the ablation threshold, and (2) the conditions which maintain the expanding plume in the gas
phase should be satisfied. Below we formulate both conditions and demonstrate their implementation in an experiment.
5.7.1. Local energy thresholds for phase transitions
In order to define the local thresholds one needs to take into account the spatial distribution of the absorbed energy in
two dimensions: across the focal spot and inside the target surface:
E
e
(r, x, t) =
2F(r, t)
n
e
l
s
exp
_
2x
l
s
_
; (5.41)
where F(r, t) is the spatial distribution of the absorbed fluence: F(r, t) = (r)
_
t
0
A(t
)I(t
)dt
. We assume that the focal

spot is a circle. The r-coordinate corresponds to the distance from the centre of a circular focal spot on the target surface,
while the x-coordinate is normal to the surface; (r) is a dimensionless function (for example, Gaussian-like) describing the
axially symmetric spatial distribution. The absorbed energy in a mono-atomic surface layer with thickness d
a
= n
1/3
a
l
s
at the end of the laser pulse becomes:
E
e
(r, x, t)
2F(r, t
p
)
n
e
l
s
. (5.42)
A local threshold by definition depends on position within the focal spot. The condition that the absorbed energy defined
by Eq. (5.42) equals the energy,
transf
, required for the particular phase transformation to occur defines the local threshold
for any laser-induced transformation of the material:
transf
=
2F(r, t
p
)
n
e
l
s
. (5.43)
It is obvious that the threshold energy density (fluence), F
thr

transf
scales with the characteristic energy required for
a particular type of phase transition. For example, the ratio of ablation threshold to the melting threshold equals the ratio
of the heat of vaporisation H
vap
to the heat of melting H
melt
. This ratio varies for most materials within a range 530. For
example, for a silicon target used in the experiments below H
vap
= 10.6 kJ/g; H
melt
= 1.66 kJ/g (Weast and Astle, 1981),
and H
vap
/H
melt
= 6.4.
One can see that in order to control the phase-state of the plume, energy in excess of the phase transition threshold
should be deposited. This condition emphasises that the threshold energy should be significantly higher than that required
Fig. 5.7. PV diagram of vapour states in the plume at various levels of absorbed laser energy and thus at different initial normalised pressures. The total
atomisation (gas phase) is achieved at the conditions when the PV curve goes above the critical point and therefore above the curve of phase equilibrium
(following from the ClapeyronClausius equation, black curve) separating the states with the mixture of phases and the gas phase.
to just break the inter-atomic bonds. Sufficient kinetic energy should be additionally delivered to an unbound atomin order
to remove it from the solid and maintain it in the desired vapour state for a certain time. The production of a droplet-free
laser plume imposes the additional condition that the gas state of the vapour should be effectively collision-less.
5.7.2. Criterion for total atomisation of ablated plume
The energy threshold for total atomisation of the ablated plume can be calculated on the basis of thermodynamic
arguments similar to those used to derive the criterion for complete vaporisation of material by strong shock waves
(Zeldovich and Raizer, 2002). The deposited laser energy is used to break the inter-atomic bonds and also provides the
kinetic energy to the expanding plume. The magnitude of the kinetic energy sufficient to keep the expanding plume in a
gaseous state containing non-interacting atoms determines the absorbed energy threshold for target atomisation. Let us
define the required level of laser fluence, F
atom
, for total atomisation of the target material with sub-picosecond laser pulses.
The equation of state of the ablated material in conditions close to that for the solidvapour phase transition can be
presented as a sum of the elastic pressure (related exclusively to the inter-atomic interaction) and the thermal pressure in
the formof Eq. (5.2). The important practical issue of target ablation and further expansion of vapour relates to the definition
of the phase state of the ablated matter in the different areas of phase space in pressuredensity, P, n, or temperature,
density, T, n, parameter planes. The material is swiftly elevated to high temperature under the action of a short powerful
laser pulse while the initial density remains unchanged. The phase states of the expanding ablated material lie along curves
for the adiabatic expansion (entropy is conserved during expansion), as shown in Fig. 5.7.
If the initial temperature of expanding plume is lower than a specific value defined below, the expansion curves cross the
phase equilibrium curve that separates the states of a single phase from the states with a mixture of phases. The maximum
of the phase equilibrium curve in T, n (or T, V = n
1
) plane is the critical point, T
crit
. The phase equilibrium curve in PT-
plane ends in a critical point, at P
crit
, T
crit
. The difference between the solid, the melt and the vapour ceases to exist in a
critical state, i.e. the density of all phases has the same value, V
crit
= n
1
crit
. The states above the first adiabatic curve touching
the critical point represent the atomic state of a homogeneous phase at V > V
crit
(n < n
crit
). Although the critical values
are poorly known for many materials, it has been established, however, that T
crit
constitutes a small fraction of the binding
energy, usually T
crit
(0.10.2)
b
(Zeldovich and Raizer, 2002).
The adiabatic curve, which separates the single-phase area from the mixture of phases, and the phase equilibrium curve
have only one common point, e.g. the critical point. One can find the initial temperature for a material at the initial density
that begins to expand along this adiabatic curve by applying the above condition with the help of equation of state Eq. (5.2):
P
crit
= c
s
n
crit
T
crit
(n
crit
) +P
c
(n
crit
). (5.44)
Note that the subscript c marks the cold pressure. The critical and initial parameters are linked through the equation
of adiabatic expansion:
T
crit
T
0
=
_
n
crit
n
0
_
(n
crit
)
. (5.45)
We remind here that the density-dependent Gruneisen coefficient, , plays the same role as the adiabatic constant,
1. Note, that at the critical point, vapour is described as an ideal gas, (n
crit
) 2/3. Now, the initial temperature
T
0
of the laser-excited solidthat should pass throughthe critical point during adiabatic expansionis easily expressedthrough
the critical parameters from Eqs. (5.44) and (5.45):
T
0
=
3
2
P
crit
+|P
c
(n
crit
)|
c
s
n
crit
_
n
0
n
crit
_
2/3
. (5.46)
The critical parameters, at least in principle, can be related to the binding energy and initial density (Zeldovich and
Raizer, 2002; Landau and Lifshitz, 1980a,b). For example, using known experimental data for aluminium (More et al.,
1988) and interpolation for the cold pressure one obtains: P
crit
= 1.84 kBar

= 5.55 10
3
b
n
0
, n
0
/n
crit
26, and
|P
c
(n
crit
)| 0.015
b
n
0
. Thus, the adiabatic expansion curve for aluminium touching the phase equilibrium curve at the
critical point must start at k
B
T
0
4.9
b
(T
0
is the temperature to the end of the laser pulse) in agreement with the qualitative
estimates from Gamaly et al. (2004).
The practical conclusion from this result is that the aluminium skin layer heated homogeneously to k
B
T
0
5
b
at its
initial density then expands adiabatically in a homogeneous state of atomic vapour. Accordingly, all states of expansion
starting at lower temperature shall cross the phase equilibriumcurve inevitably entering into states containing a mixture of
phases. A similar qualitative conclusion for atomisation of the ablated plume can be made for all solids because the binding
energy for most solids lies in the range of 25 eV. Thus, the energy per atom approximately 35 times larger than the
binding energy, E
abs
> (35)
b
, should be deposited into a solid in order to transform the target into a fully atomised gas.
We should note here that the exact value of the above numerical coefficient depends strongly on the equation of state in
a range n
0
> n > n
crit
for a particular material. Therefore, the laser energy density necessary to transform the ablated
material into the atomised vapour should comply with the following condition: F
atomise
4F
thr
. The qualitative dependence
of normalised pressure, P/(
b
n
0
), on normalised volume n
0
/n during the adiabatic expansion starting at the same initial
density n
0
and with the different initial normalised pressure is plotted in Fig. 5.7. The black curve corresponds to the states
of phase equilibrium (ClapeironClausius curve). This curve separates the states representing the mixture of phases (at
P < P
crit
) from the states of a homogeneous phase or atomised vapours (at P > P
crit
). We should stress that this function
is poorly known for most materials. One clearly sees from Fig. 5.7 that the expansion from the initial state with energy
E
abs
> (35)
b
keeps the expanding vapour in a gas state whilst at a lower initial energy a mixture of states is unavoidable.
5.7.3. Surface damage and evaporation
If the total deposited laser energy per atom is close to the binding energy, E
abs

b
, then the heated target experiences
only a small decrease from the normal solid density. The pressure in the material in this case is comparable to the bulk
modulus. Therefore, the final state of the target affected by the laser at this energy level might be considered as damaged,
resulting in the formation of cracks, flakes, de-lamination of the surface, etc., depending on the presence of defects and
impurities in the initial state of the target. If the deposited energy is in a range
b
< E
abs
< (35)
b
, then the final state
of the expanding vapour may fall within a region of the PV plane where the mixture of phases is energetically favourable
(Fig. 5.7). This leads to condensation of vapour into liquid droplets. We should keep in mind that the thresholds above
were introduced locally, i.e. different regions of the beam can create different phase states depending on the local beam
intensity. One should also note that the energy estimates are rather conservative (overestimated) because they are based on
the assumption of thermodynamic equilibrium. In reality, and especially for short pulses, the expansion time is shorter than
the equilibration time. Therefore, the expansion of vapour proceeds in the kinetic regime, and thus condensation processes
are decelerated.
5.7.4. Optimum pulse profile for atomisation of the plume
(a) Temporal shape of short pulses
The chirped pulse amplification (CPA) technique (Maine et al., 1988) commonly used for the generation of energetic sub-
picosecond pulses often produces a pre-pulse, which may contain a significant amount of energy. It is well known that to
achieve the short pulse interaction mode a high contrast ratio is required between the energy of the short pulse and the pre-
pulse. A detailed discussion of the methods for providing a high contrast ratio can be found in Luther-Davies et al. (1991),
Luther-Davies et al. (1992) and Wharton et al. (2001). We note that efficient methods to suppress the pre-pulse include gain
narrowing, the use of a saturable absorber, and frequency conversion.
There is an additional source of such a pre-pulse, namely amplified spontaneous emission (ASE) fromthe laser amplifiers.
The intensity contrast ratio of the main pulse/ASE is generally about 10
6
, whilst the duration of the ASE pulse can be of
0.22 ns (Luther-Davies et al., 1991; Whartonet al., 2001; Luther-Davies et al., 1992). For example, if the pre-pulse associated
with the ASE is of nanosecond length, t one ns, and A 1 then the intensity in the ASE pre-pulse of 10
8
W/cm
2
can
damage the target surface before the main pulse arrives. Recently, ASE effects on femtosecond lasermatter interactions
have been revisited and it was once more confirmed that the ASE could significantly affect the interactions of a main short
pulse withthe target (Whartonet al., 2001). To eliminate ASE the laser designmust include successive stages of amplification
interspersed with spatial filters (Luther-Davies et al., 1991, 1992).
Fig. 5.8. The Gaussian beam profile in the focal spot with the maximum fluence of 5F
thr
and 10F
thr
. The beam radius is measured in the units of FWHM
radius. The increase of maximum fluence in the beam from 5F
thr
to 10F
thr
improves the energy expense for total atomisation from 20% to 60% and reduces
the mixture of phases energy part from 60% to 30%. Accordingly, the target area producing total atomisation (gas) phase in the plume is increased four
times.
On a much shorter, ps-range time scale, temporal shaping of the ultra-short laser radiation tightly focused by a high-
NA microscope objective inside transparent material provides an opportunity to control the ionisation rate and to reduce
the absorbing energy volume below the diffraction limit (Englert et al., 2007). It was shown that it was possible to control
filaments propagation in non-linear aqueous solution (Heck et al., 2006), to optimise the efficiency of white continuum
generation in air (Ackermann et al., 2006), and to compensate spherical aberrations by changing temporal shape of ultra-
short laser pulses (Mermillod-Blondin et al., 2008; Mauclair et al., 2008). The adaptive control is achieved with a help
of spatial phase modulator placed into the fs-laser dispersive system before the amplification stage. Temporal shaping is
modified by phase-only filtering in liquid crystal array without loss of total energy. The control over the ionisation rate and
deposited energy density offer potential benefits for laser-based 3D optical memory procession and 3D microstructuring.
As far as atomisation of the ablated vapour is concerned, temporal shaping of femtosecond pulses provides the ability
to control the efficiency of nanoparticle generation through variation of the spatial temperature and density profiles in the
laser ablated plume (Guillermin et al., 2010). The optimal tailoring of the laser pulses allows one to enhance the atomisation
of the supercritical fluid in the plume inducing a more efficient transition to the gas phase.
(b) Spatial shaping: top-hat distribution
Let us apply the above reasoning to the problem of obtaining efficient atomisation of an expanding plume produced
by ablation of a material in order to demonstrate the importance of the spatial distribution of the laser intensity. First, we
assume that the temporal pulse shape is Gaussian and the pulse is free of ASE, pre-pulses and post-pulses. In most practical
cases the laser intensity across the focal spot is assumed to have the Gaussian distribution with axial symmetry (beam axis
is at r = 0): (r) = exp[(r
2
/r
2
f
) ln 2]. The focal spot radius, r
f
, is defined by the condition (r
f
) = 1/2. The part of the
total energy confined in a circle of radius r is given as follows:
_
r
0
(r
)dr
2
_
_
ln 2
r
2
f
_
= 1 exp
_
r
2
r
2
f
ln 2
_
. (5.47)
On the basis of the results of the previous sections one can estimate the phase state of the ablated plume for any peak
absorbed laser fluence. Choosing, for example, a laser fluence, AI
max
t
p
, 5 times larger than the ablation threshold (Fig. 5.8),
one can easily calculate that only 20% of the pulse energy goes into full atomisation of the target material and this occurs
in the beam area where F > F
thr
(0 < r < 0.57r
f
). The mixture of phases is produced by ablation from the area where
F
thr
< F < 4F
thr
and this region contains 60% of the beam energy (0.57r
f
< r < 1.52r
f
). The part of the beam where the
surface damage is most probable at F < F
thr
(r > 1.52r
f
) consumes 20% of the total absorbed laser energy. Therefore,
the target area producing a mixture of phases in the plume is about 5 times larger than that where total atomisation occurs.
Thus this plume is highly likely to contain particulates or droplets.
The ratio of the beam area producing a mixture of phases to that producing the fully atomised vapour can be obviously
reduced by increasing the peak fluence. For example, the part of absorbed energy used up for full atomisation of vapour
increases by three times, from 20% to 60%, if the peak absorbed fluence is increased from 5F
thr
to 10F
thr
(Fig. 5.8).
Nevertheless, incomplete atomisation still remains.
Clearly, it is very difficult to obtain complete atomisation using a Gaussian beam and it is obvious that it would be
preferable to use a top-flat-hat intensity distribution where the absorbed fluence everywhere exceeds about four times
the ablation threshold. A simple way to move towards the top-hat profile is to truncate the low energy wings in the spatial
distribution with an aperture and employ a relay-imaging focusing scheme to image the top-hat beam onto the target.
Belowwe describe the experimental procedure and experiments where this idea has been implemented in fs-laser ablation
of silicon (Gamaly et al., 2004).
5.7.5. Experiments: ablation and deposition of silicon films by spatially shaped pulses
It is a very difficult experimental task to measure in fine details the time and space resolved composition of laser-created
expanding plume. Such experiment was not performed yet to the best of our knowledge. However, one of the most broadly
used applications of laser ablation is deposition (condensation) of laser-ablated material on a substrate in order to make
highly homogeneous thin films of high quality. Therefore a quality of deposited filmhomogeneity, surface quality, absence
of droplets, large particles and other defects is indirect evidence of close-to-atomic composition of a plume. This idea
was implemented in work (Gamaly et al., 2004) were judgement on the composition of ablated Silicon plume was made
comparing the quality of the films deposited by the plume generated by laser pulse with different spatial distribution of
intensity over the laser focal spot. Three series of deposition experiments were conducted using a 150 fs laser: (i) with the
target located at the focal spot at the maximum intensity; (ii) with the target located at the image plane of the 2.2 mm
diameter aperture; (iii) at the same target position but with the aperture removed. Introduction of the aperture practically
removes the lowintensity wings in the Gauss distribution of intensity over the focal spot making it close to the flat-top-hat
one. As a result, the density of droplets was reduced from>1000 cm
2
in case (i) to 10 cm
2
in case (ii); but increased to
100 cm
2
in case (iii) (Gamaly et al., 2004).
The analysis of these optical images revealed that the density of 110 m droplets on the substrate surface was above
1000 cm
2
when the target was positioned in the focal plane where the absorbed laser fluence was AF = 65 J/cm
2
0.64 =
41.6 J/cm
2
(I
abs
= 2.8 10
14
W/cm
2
). By placing the target in the image plane of the aperture the maximumfluence of the
truncated beam was reduced to AF = 1.6 J/cm
2
0.64 = 1 J/cm
2
(I
abs
= 6.8 10
12
W/cm
2
). The droplets were almost
eliminated from the deposited Si-film: the density of droplets was reduced below 10 cm
2
. In order to test the influence
of the low-intensity wings in the spatial pulse profile on droplet formation, another deposition was performed with the
target in the same position but with the aperture removed. The density of droplets then increased to above 100 per cm
2
.
This experiment provided a clear demonstration that the low intensity wings in the spatial distribution are responsible for
the formation of droplets on the deposited film surface.
Positioning the target in the image of the aperture may also eliminate the problem of droplet formation from any pre-
pulse. The effect of the pre-pulse could be estimated in the following way. The measured 1 ns pre-pulse containing 12.7 J
results at most in 5.4 10
7
W/cm
2
absorbed laser intensity in the target before the arrival of the main pulse. The depth of
the heat wave propagation, l
heat
can be estimated as l
heat
= (Dt
pp
)
1/2
; here D is thermal diffusivity (D
Si
= 0.85 cm
2
s
1
)
(Weast and Astle, 1981) and t
pp
is the pre-pulse duration, thus l
heat
= 0.29 m. This leads to a maximum absorbed energy
in the heated volume below the target surface of 1.26 10
3
J/g. This value is below the heat of melting of 1.66 10
3
J/g
for Si (Weast and Astle, 1981), thus the pre-pulse could not melt the target in the focal spot before the arrival of the main
sub-picosecond pulse. However, the energy of the pre-pulse would still be high enough to cause some transformations of
the material such as the formation of cracks, which may be the reason for the presence of a small number of micron-size
particles on the deposited film seen in the experiments in case (iii).
5.8. Accumulation effects in ablation of solids by high-repetition-rate short-pulse lasers
We demonstrated that the use of ultra-short pulses with the top-flat-hat intensity distribution allows good control over
the phase state of ablatedplume. Eachsingle short low-energy high-intensity pulse evaporates relatively a few(10
11
10
12
)
atoms per pulse (Rode et al., 1999; Perry et al., 1999). However, for industrial applications of laser ablation the high average
ablation rate per second is necessary. To compensate for the reduced ablated mass per pulse, high pulse repetition rates are
then used to achieve a high average ablation rate. The high repetition-rate up to 100 MHz maintains the average atomic
flow in a plume at a high level of 10
19
10
20
atom/s appropriate for laser deposition and micromachining applications.
On the other hand, the interaction of pulse trains from MHz repetition rate lasers with matter appears to be significantly
different from single pulse interaction due to accumulation of the effects of successive pulses. The number of pulses on the
target surface at high repetition rate may reach thousands per spot because the scanning speed of the laser over the target
surface is normally too slowto physically separate successive pulses (the time between the successive pulses for MHz lasers
is of the order of hundred of nanoseconds).
The coupling between the successive pulses, with single pulse energy below ablation threshold and thus insufficient
to produce ablation, gradually transforms the lasermatter interaction into a regime where much stronger lasertarget
coupling occurs that results in the ablation threshold being exceeded. Therefore, although the first m pulses from the total
N pulses hitting the same target spot do not lead to substantial ablation, the remaining N mpulses do ablate the material.
Thus, the ablation length averaged by the total N pulses yields the energy per pulse, which is lower than that for the single
pulse ablation threshold.
The mechanisms for the coupling of successive pulses in ablation of metals and transparent dielectrics with low heat
conduction (such as silica and glasses) are different. We explain cumulative ablation of metals by the gradual plume density
and temperature increase in the vicinity of the target surface due to slow thermal ablation between the successive pulses.
The build-up of temperature and density leads to a change in the lasermatter interaction mode from absorption in the
skin-layer to absorption on a plasma density gradient.
The time between pulses is too short for completing cooling of an irradiated spot on a transparent dielectric between
successive pulses. Thus, in this case, the laser irradiates a spot already heated by the previous pulses and the surface
temperature gradually rises until it exceeds the temperature required for evaporation. Thus, the energy accumulation within
the target is the major effect in cumulative ablation of poorly absorbing dielectrics. As a result, high repetition rate pulse
trains allow precise control over the temperature at the sample surface via the dwell time of many pulses at the same
spot, allowing ablation to occur under conditions when the single pulse fluence is insufficient to ablate the material.
The cumulative effect in high-repetition-rate-laser interaction with matter has been observed during ablation of carbon
by a laser with 76 MHz repetition rate (Rode et al., 1999). The effects of high-repetition-rate ablation that include smoothing
of the spatial intensity distribution and cumulative heating, have both been observed in ablation and deposition of
chalcogenide glasses (Rode et al., 2002a,b, 2001). Cumulative heating in the bulk of a transparent glass was also reported in
Schaffer et al. (2003), and the reduced ablation rate with increased number of pulses per spot due to absorption in vapour
was analysed in Gamaly et al. (2001). The cumulative effect of increasing the electron density above the plasma critical
density in a silica substrate heated by a train of 2030 fs pulses with a time gap of 500 fs was numerically demonstrated in
Itina and Shcheblanov (2010).
In this section we will consider briefly experiments on ablation of aluminium, copper, steel, lead, and silicon targets with
a 4.1 MHz pulse-rate 10 ps mode-locked laser in air and in vacuumat conditions above and belowthe single-pulse threshold.
5.8.1. Dwell time and number of pulses per focal spot
Many laser pulses arrive at the same spot on the target surface at repetition rate of 1100 MHz. From the known
oscillation mode of the scanning mirrors one can easily estimate the maximum, t
max
, and minimum, t
min
, times near the
turning points where the laser beamdwells over a focal spot of a diameter, d
f
, for a givenscanning frequency,
s
, repetition
rate R
rep
, and scanning area of size a. The laser beam spends the maximum time near the beams turning points because the
scanning velocity passes through zero while changing direction and because the beam crosses the same spot twice. For
simple harmonic oscillations the maximum dwell time is expressed as follows (
s
t
max
1):
t
max
=
4
s
_
d
f
a
_
1/2
. (5.48)
Similarly, the minimum dwell time near the centre of a scanning area becomes:
t
min
=
d
f
a
s
. (5.49)
For example, for the conditions of experiments (
s
60 Hz, a 15 mm, d
f
= 25125 m, R
rep
= 4.1 MHz) (Gamaly
et al., 2007) the number of laser pulses per spot varies from115 in the middle of the scanning area for the maximumfluence
with a small focal spot, to 2.5 10
4
for the lowest fluence with the largest spot. In experiments with a 76 MHz repetition
rate laser (d
f
= 25 m; a 50 mm;
s
100 Hz) the number of pulses per spot lies in a range from 38 to 7 10
4
(Rode et al., 1999). Therefore during ablation by high (MHz) repetition rate lasers using 100 Hz scanners, numerous pulses
interact with the same spot making the interaction regime drastically different from a single-pulse interaction mode.
5.8.2. Smoothing of the evaporation conditions on the surface
There is an appealing effect connected to both the high repetition pulse rate and the beam scanning speed. A single laser
pulse focused at a target surface produces the Gauss intensity distribution across the focal spot. If the maximum intensity
of a single pulse only slightly exceeds the ablation threshold, then the plume produced by such a pulse contains a mixture
of phases: liquid droplets, flakes, etc. The spatial intensity distribution must be modified in order to control the phase state
of the ablated vapours as demonstrated earlier.
The use of a high repetition rate is, however, a natural way to produce a plume of homogeneous phase. Indeed, the
scanning speed usually is much lower than the repetition rate in multi-MHz regime. The laser beammoves less than 0.1 m
between two successive pulses. As a result, the spatial intensity distribution over the target surface integrates over time,
and the evaporation conditions become almost homogeneous. All the imperfections of a single beamare smoothed with the
successive pulses heating the same surface spot several times. As the result, it may be unnecessary to control the intensity
distribution as required for single pulse irradiation.
5.8.3. Temperature accumulation in a multiple-pulse action on poor heat-conducting dielectrics
The primary mechanism of coupling between the multiple successive pulses in low absorbing poor-heat-conducting
dielectrics such as silica and glasses relates to the temperature accumulation. The cooling between the pulses is small if the
heat conduction is low allowing for the energy accumulation. The temperature accumulation can be estimated as follows.
Let us suggest that the temperature decreases after the end of the first pulse in accordance with the one-dimensional linear
heat conduction as follows:
T = T
1
_
t
th
t
th
+R
1
rep
_
1/2
T
1
. (5.50)
One-dimensional approximation is valid because the condition, l
th
l
abs
d
foc
, holds. The characteristic cooling time, t
th
,
of absorption layer with thickness l
abs
comprises t
th
= l
2
abs
/D, where D is thermal diffusivity in cm
2
/s. The time gap between
the successive pulses is inverse proportional to the repetition rate, R
1
rep
. One can present the temperature drop after a single
pulse action in a form:

_
1
R
1
rep
2t
th
_
. (5.51)
We took into account that condition R
1
rep
t
th
(and 1) holds. The temperature rise after the N-th pulse hitting the
same spot then can be presented in the form:
T
N
= T
1
(1 + +
3
+ +
N
) = T
1
1
N
1
. (5.52)
The effect of temperature accumulation is very strong in chalcogenide glass where the both heat conduction and absorption
are very low: D = 1.15 10
3
cm
2
/s; l
abs
= 2 10
4
cm; t
th
= 3.48 10
5
s. In these conditions the temperature in the
absorption layer grows up as T
N
NT
1
until the thermal ablation threshold can be reached.
5.8.4. Thermal ablation threshold in multiple-pulse action
The ablation threshold can be calculated, if the temperature dependence on time is known, by the general formula:
t
thr
0
(nv)
therm
dt n
0
d
a
. (5.53)
In experiments on ablation of chalcogenide glass by 76 MHz 60 ps laser (Luther-Davies et al., 2005) the accumulation
effect is strong and the temperature between the pulses is practically constant due to negligible thermal losses. One can
calculate the ablation length during the time gap between the pulses at the temperature sufficient for ablation T
N
=
293 K + NT
1
neglecting the contribution from the previous pulses due to exponential dependence of the ablation rate
on the temperature. Then the threshold condition (Eq. (5.53)) takes the following simplified form:
_
2k
B
T
N
M
_
1/2
exp
_

b
k
B
T
N
_
R
1
rep
d
a
. (5.54)
Let us relate the temperature after N-th pulse to the temperature after first pulse, = T
N
/T
1
. The above equation reduces
to the following:
b
k
B
T
1
= ln
_
d
a
R
rep
v
1
_
1
2
ln ; v
1
=
_
2k
B
T
1
M
_
1/2
. (5.55)
The numerical solution of Eq. (5.63) presented in Luther-Davies et al. (2005) produces = T
N
/T
1
8.3 for ablation of
chalcogenide glass (
b
/T
i
= 81.5, R
rep
= 76 MHz; T
1
= 26 K + 293 K = 0.027 eV; d
a
2.5 10
8
cm; v
1
10
5
cm/s).
Thus, the ablation threshold can be reached due to temperature accumulation from N
thr
= (T
N
thr
/T
1
)(1 + 293 K/T
1
)
293 K/T
1
90.5 pulses and subsequent thermal ablation during the time gap between 91th and 92th pulses.
5.8.5. The build-up of plume density near the target surface
We demonstrated above that the presence of air next to the solid surface increases the ablation rate due to thermal
evaporation after the pulse. A similar effect may take place when a high repetition rate laser is used for ablation because of
the accumulation of a dense vapour in front of the solid target surface fromsuccessive pulses. Increase in the vapour density
is equivalent to the increase in the saturated vapour density, and therefore it results in an increase in the thermal ablation
rate.
One can estimate the conditions for such accumulation effects as follows. Thermal ablation can be efficient once the
Maxwell distribution between the vapour and the solid has been established. Thus, the first condition for cumulative
evaporation is that the equilibration time should be shorter than the time gap between the pulses, t
eq
= (nv)
1
< R
1
rep
.
It follows from this condition that the vapour density should comply with the inequality n
vap
> R
rep
/v. Thus, taking the
data fromexperiments (Luther-Davies et al., 2005): repetition rate, R
rep
= 4.110
6
s
1
; cross-section for atomic collisions,
10
15
cm
2
; and atomic velocity, v
th
10
5
cm/s, the vapour number density should be n > 4 10
16
cm
3
. Thus, for
the ablation in vacuum in experiments (Luther-Davies et al., 2005) at P = 3 10
3
Torr (n = 1.8 10
14
cm
3
) the density
near the ablated surface should increase more than 200 times due to the action of many consecutive pulses.
Let us consider the conditions for such density build-up. Entropy and the energy are both conserved after the pulse.
Therefore the plume expands adiabatically. The specific features of the isentropic expansion are the follows: the density and
the temperature of a plume go to zero at the finite distance from the initial position (in contrast to isothermal expansion),
while the velocity is at its maximum(Zeldovich and Raizer, 2002). Therefore the density next to ablation surface has a steep
gradient. The size of the expanding plume grows linearly with time (and with the number of consecutive pulses hitting the
same spot):
R
N
N
v
th
R
rep
. (5.56)
In fact the expansion velocity slows down as the density of vapours next to the ablation surface grows up. The average
number density of vapours in the plume can be estimated as follows. The total number of atoms evaporated during N > N
thr
pulses after the ablation threshold be achieved can be calculated with the help of Eq. (5.54):
S
foc
t
t
th
(nv)
therm
dt
S
foc
n
0
v
1
R
rep
N
N
thr
N
1/2
exp
_

b
Nk
B
T
1
_
. (5.57)
The average atomic number density in a plume after action of N pulses then expresses as:
n =
S
foc
n
0
v
1
R
rep
N
N
thr
N
1/2
exp
_

b
Nk
B
T
1
_
4
3
_
v
1
N
R
rep
_
3
. (5.58)
Thus, in ablation of silicon by 4.1 MHz 12 ps laser (T
1
= 0.1 eV; v
1
10
5
cm/s; n
0
= 510
22
cm
3
; S
foc
= 510
6
cm
2
)
the average number density of ablated atoms in a plume after the action of less than hundred pulses exceeds the limit for
creation of a dense plume, n > 4 10
16
cm
3
, where collisions between atoms in a plume and with ablating solid are
essential (Luther-Davies et al., 2005).
Similar estimate for ablation of chalcogenide glass by 76 MHz 60 ps laser (T
1
= 0.049 eV; v
1
10
5
cm/s; n
0
= 3.9
10
22
cm
3
; S
foc
= 5 10
6
cm
2
) by Eq. (5.58) show that after 100 pulses per spot the density increase creates a vapour
dense enough to switch on thermal evaporation in the manner invoked in the presence of air (Luther-Davies et al., 2005).
5.8.6. Change in the interaction regime
After several tens pulses the temperature at the ablation surface increases up to several eV, that is close to the binding
energy, in both experimental situations with 4.1 and 76 MHz lasers, along with the build-up of vapour density next to a
solid. Now the interaction regime begins to change well before the cooling by heat conduction can occur. We remind that
the heat conduction time in considered experimental conditions comprises 4 10
5
s. At the temperature of several eV
the ionisation becomes significant and the plume converts into plasma. Let us approximate the electron density gradient
in plasma near the solidplume interface by a linear density profile, n
e
= n
c
x/L with characteristic space scale L (n
c
is the
critical density for the incident laser radiation). The absorption in plasma can be significant if the condition L
ei
(n
c
)/3c > 1
is fulfilled (Kruer, 1987); here characteristic electronion collision rate
ei
(n
c
) is taken at the critical density and c is speed of
light in vacuum. One can see that in order to have a significant absorption in plasma in the condition of experiments under
consideration the plasma density scale should be L 0.2 m(
ei
(n
c
) 4 10
15
s
1
; n
c
2 10
21
cm
3
). Therefore the
target absorption mode changes to the absorption on the critical density (laserplasma interaction), significantly increasing
the temperature in the absorption zone and near the ablation surface. The temperature can be estimated under assumption
that whole absorption occurs at the critical density:
M
a
n
c
v
3
AI; v =
_
2k
B
T
M
a
_
1/2
. (5.59)
Then the temperature reads:
k
B
T
e

M
a
2
_
I
abs
M
a
n
c
_
2/3
. (5.60)
Thus, in ablation of chalcogenide glass (M
av
= 49) by 76 MHz 60 ps pulses at the intensity level I = 2.65 10
8
W/
cm
2
(16 mJ/cm
2
; A = 0.8) the single pulse interacting in laserplasma mode increases the temperature up to 2.8 eV,
slightly in excess of the binding energy. Ablation nowcan proceed by non-equilibriummechanismwith a maximumablation
depth per pulse l
max
abl
= F
p
M
a
/
b
, close to that observed in the experiments (Luther-Davies et al., 2005). Efficient ablation
thencontinues during the subsequent several hundredpulses after the laserplasma interactionregime has beenestablished
because the scan rate allowed up to several thousand pulses to hit the same spot in these experiments.
5.8.7. Concluding remarks on cumulative ablation
The experimental results and analysis presented above demonstrate that ablation of transparent dielectrics (silicon and
chalcogenide glass) witha lowheat conduction(10
3
cm
2
/s) by highrepetitionrate 476MHz lasers withenergy per pulse
less that sufficient for ablation, proceeds in a mode when interaction between successive pulses dramatically increases their
combine action. At MHz repetition rate up to several thousands of pulses hit the same spot at the target surface. The coupling
between consecutive pulses occurs in three stages.
First, the temperature accumulation takes place when absorbed energy of many pulses simply sums up, while the heat
losses are negligible, until the thermal evaporationthresholdis achieved. Second, after ablationthe density of vapours begins
to build-up near the ablating surface due to short time interval between the consecutive pulses, this is a direct consequence
of high repetition rate of the laser pulses.
Finally, due to increase of both temperature and density, the vapour became ionised and lasermatter interaction
changes to the interaction with plasma. The energy absorption occurs near the critical density defined by the incident laser
wavelength. The temperature rises slightly above the material binding energy allowing for efficient material removal by
non-thermal and thermal mechanisms. This scenario and calculations fit well to experimentally observed efficient ablation
of chalcogenide glass.
Ablation of crystalline silicon is an example of combined action only two consecutive pulses. The first pulse transforms
crystalline silicon into amorphous phase with higher absorption while the thermal diffusion is diminished. The second pulse
raises the temperature tothe level whenthermal ablationinthe periodbetweensecondandthirdpulses results insurpassing
of the ablation threshold.
Thus, cumulative ablation by high repetition rate lasers allows efficient ablation of any transparent solids in the
conditions when single pulse energy is insufficient for evaporation. The high quality thin films of different materials
including low absorbing dielectrics (absorption length m) such as silicon, silica, chalcogenide glasses, and sapphire
can be deposited by using very high repetition rate lasers (1100 MHz) by an appropriate self consistent choice of laser
wavelength, repetition rate, and energy per pulse. The appropriate combination of laser and material parameters determines
the optimumconditions for ablation and deposition. Control over the phase state of a plume, the elimination of droplets, and
high ablation rate can all be simultaneously achieved as it has already been demonstrated using silicon and chalcogenide
glasses (Luther-Davies et al., 2005).
5.9. Summary
Evaporation rate in conditions of thermal equilibrium (n
i
v
i
)
equilibrium
n
vap
_
k
B
T
M
_
1/2
exp
_

b
k
B
T
_
The ratio of the absorption coefficient to the skin depth in
metals near the ablation threshold
A
l
s

2
c
=
4
Threshold laser fluence for ablation of metals in extreme

non-equilibrium conditions (electrostatic ablation)
F
m
th
=
3
4
(
b
+
esc
)
l
s
n
e
A
(J/cm
2
)
Threshold fluence for electrostatic ablation of dielectrics F
d
th
=
3
4
(
b
+J
i
)
l
s
n
e
A
(J/cm
2
)
Threshold of material ablation by long pulses
t
pulse
> (t
ei
, t
therm
); l
therm
l
s
/2
F
th

(D
th
t
p
)
1
2
b
n
a
A
(J/cm
2
)
Ablation depth by single pulse l
mono
+
FF
thr
n(
b
+
esc
)
l
abl
l
mono
+
l
s
2
ln
_
F
a
F
thr
_
Ablated mass per pulse m
abl
= S
foc
n
a
Ml
abl
6. Confined interaction: laser beams tightly focused inside transparent solids
6.1. Introduction
The interaction of intense laser radiation with matter when the beam is tightly focused inside a transparent material
is radically different from the case of focusing the beam onto a surface. In the lasersurface interaction the temperature is
highest at the outermost atomic surface layer. If the absorbed energy density in the surface layer is in excess of the ablation
threshold, the atomic bonding breaks and the ablated atoms leave the surface. On the other hand, in the tightly focused
interaction mode the focal zone with high energy density is confined inside the bulk of a cold and dense solid. The laser-
affected material remains at the focal area inside the pristine material. The laser-induced changes in material properties
depend on the laser intensity. Non-destructive and reversible phase transitions, such as photo-refractive effect in LiNbO
3
(Buse, 1996; Buse et al., 1998; Gamaly et al., 2010) and photodarkening inchalcogenide glasses (Elliott, 1986), canbe induced
at intensity below the damage threshold. These structural changes produced in a small, micron-size, volume can be used as
a memory bit for high-density 3D optical storage (Glezer et al., 1996; Watanabe et al., 1998; Qiu et al., 1998; Juodkazis et al.,
2003), for formation of optical waveguides (Miura et al., 1997; Zakery et al., 2003) and in other photonics, micromachining
and material-processing applications (Gattass and Mazur, 2008).
Irreversible structural changes occur at an intensity level above the optical breakdown threshold. It was demonstrated
(Glezer and Mazur, 1997; Juodkazis et al., 2006a,b,c; Gamaly et al., 2006) that unique conditions of extremely high pressure
Fig. 6.1. A principal scheme for the tightly focused lasermatter interaction confined in a transparent solid.
and temperature with record high heating and cooling rates are created in the energy deposition region. Astrong shock wave
generated in the interaction region expands into the surrounding cold material and compresses it, which may result in the
formation of new states of matter. Shock wave propagation is accompanied by a decompressing rarefaction wave behind
the shock front, leading to the formation of a void inside the material. The shock wave compressed material surrounds the
void in the form of a densified shell.
The physics of laser interaction with transparent dielectrics at high laser intensity above the ionisation threshold is
markedly different from that below the threshold. Transparent dielectrics have several distinctive features. First, they have
a wide optical band-gap (it ranges from 2.2 to 2.4 eV for chalcogenide glasses, and up to 8.8 eV for sapphire) that ensures
they are transparent in the visible or near infrared spectral range at low intensity. In order to induce material modification
with moderate energy pulses, the laser intensity should be raised in order to induce a strongly non-linear response fromthe
material, such as ionisation and plasma formation. At intensities in excess of 10
14
W/cm
2
, most dielectrics can be ionised
early in the laser pulse, so the interaction proceeds in the laserplasma mode, which is similar for all materials.
A second feature of dielectrics is their relatively low thermal conductivity with a low thermal diffusion coefficient, D,
which is typically 10
3
cm
2
/s, compared with 1 cm
2
/s for metals. Therefore, a micron-sized regions cool in 10 s
(t l
2
/D 10
5
s). Hence, the laser effect of multiple low-energy laser pulses focused at the same point in a dielectric
will accumulate if the period between the pulses is shorter than the cooling time. Thus, if the single-pulse energy is too
low to produce any modification of the material, a change can be induced using a high pulse repetition rate because of
this accumulation phenomenon. The local temperature rise resulting from energy accumulation eventually saturates as the
energy inflow from the laser is balanced by heat conduction, this typically taking a few thousand pulses at a repetition
rate in the 10100 MHz range. This effect has been experimentally demonstrated from measurements of the size of a void
produced inside a dielectric by a high-repetition-rate laser (Schaffer et al., 2003). The size of the damage zone increased
with the number of pulses hitting the same point in the material. However, successive interaction of multiple pulses in the
same focal zone creates a structure with properties and shape that is very difficult to predict and control.
Recent studies have demonstrated (Juodkazis et al., 2003; Schaffer et al., 2003; Glezer et al., 1996; Watanabe et al., 1998;
Qiu et al., 1998; Juodkazis et al., 2002) that a single short laser pulse tightly focused inside the bulk of transparent solids
such as silica glasses, crystalline quartz, sapphire, and polymers, can produce a cavity, or void, inside a pristine crystal.
Directed translation of the laser beam within a crystal forms three-dimensional (3D) sets of laser-produced structures with
a controlled size less than half of micron. We stress here that the size and the properties of the structure produced under the
action of a single pulse can now be predicted. It has also been demonstrated that these structures can be formed in various
spatial arrangements (Glezer et al., 1996; Watanabe et al., 1998; Qiu et al., 1998; Juodkazis et al., 2002; Sun et al., 2001b) and
be used to form photonic crystals, waveguides and gratings for application in photonics. A single structure can also serve as
a memory bit because it can be detected (read) by the action of a probe laser beam(Glezer et al., 1996; Juodkazis et al., 2003).
In this section describe lasersolid interaction when a laser beam is tightly focused inside a transparent dielectric in
two cases: (i) low intensity, well below the ablation threshold (non-destructive interaction); and (ii) tight focusing beam
to the high intensity resulting in high energy density, when all bonds are broken and the material is ionised (destructive
interaction). Then we compare the results with experiments and draw the conclusions.
6.2. Focusing of the beam inside a transparent solid
The general scheme of the experimental set-up for the tight focusing inside transparent media is presented at Fig. 6.1.
In order to create high energy density in the bulk of a transparent material one should minimise the energy losses
along the transport path and reduce the focal volume to the smallest possible. The major obstacle for a powerful laser
beam propagating over a long distance in a transparent medium is self-focusing caused by the intrinsic non-linearity of
the medium, which may result in optical breakdown above certain threshold power of the beam. In addition, spherical
aberrations are the barrier to achieve a diffraction-limited spot size.
6.2.1. Limitations imposed by the spherical aberrations
Application of fs-laser pulses for 3D optical data storage, micro-fabrication, and optical trapping, require the laser beam
to be tightly focused to achieve high laser intensity for inducing a material modification. High numerical aperture (NA) optics
Fig. 6.2. (a) Ray propagation due to spherical aberration when the beam is focused into a media with n
1
> n
2
. (b) Example of axial intensity distribution
in the focal area distorted by spherical aberration of a lens with NA = 1.35(
1
= 63) from immersion oil with n
1
= 1.515 into sapphire, n
2
= 1.7.
(c), (d) The normalised 3D plots of the point spread function (Eq. (6.1)) at 0 and 20 m depth, respectively; the corresponding lateral and axial spans are
4 m and 30 m.
is essential to attain high-energy concentration per unit volume with ultra-short pulses, as the radius of the beamwaist is in
inverse proportion to the NA of the focusing optics (see belowin Section 6.2.3). At the same time, the increase of NA leads to
the increase of spherical aberration in the focal area. This is due to the fact that after the focusing objective different parts of
the spherical wave front enter the media at different angles to the interface, and refracted rays intersect the beamaxis at the
different distance from the surfacesee Fig. 6.1. As a result, the focal area in the transparent media becomes elongated in
the direction of the beampropagation, and thus the energy density is reduced. The function (, n, L
f
) describing spherical
aberrations depends on the converging angle of a ray, refractive indices n, which in turn depends on the laser wavelength,
and on the geometrical depth L
f
of focusing into the refractive media (Trk et al., 1996; Min Gu, 2000):
(, n, L
f
) = kL
f
(n
1
cos
1
n
2
cos
2
); (6.1)
here k = 2/ is the wave vector in vacuum, and
1
and
2
are related through Snells law. The point spread function I
f
, the
intensity distribution in the focal area, can be presented in spherical coordinates (r, z), in the frames of scalar Debye theory
as (Min Gu, 2000):
I
f
(r, z, , L
f
) =
_

0
P(
1
) sin
1
[t
s
+t
p
cos
2
]J
0
(k
1
, r
1
, sin
1
) exp[i((, n, L
f
) +k
2
z
2
n
2
cos
2
)]d
1
_
2
(6.2)
where t
s
= 2 sin
2
cos
1
/ sin(
1
+
2
) and t
p
= 2 sin
2
cos
1
/[sin(
1
+
2
) cos(
1

2
)] are the Fresnel coefficients
for s- and p-polarisations, correspondingly; P(
1
) =

cos
1
is the apodisation function obeying the sine condition (the
commercial objective lenses are designed to satisfy the sine condition); J
0
(k
1
, r
1
, n
1
, sin
1
) is the zero-order Bessel function
of the first kind; and
1
is varied from zero to = arctan(a/f )see Fig. 6.2(a).
The effect of spherical aberrations on the depth at which the laser is focused and the ways to compensate aberrations
were thoroughly studied in Min Gu (2000), Mauclair et al. (2008) and Mermillod-Blondin et al. (2008).
These studies show that spherical aberrations in high-NA focusing impose significant depth-dependent constraints onto
the concentration of laser energy inside transparent dielectrics when the refractive index of the immersion oil does not
perfectly match the refractive index of dielectrics. In practice, to achieve high energy concentration of the order 500 kJ/cm
3
focusing 100 nJ 100 fs 800 nm pulses with NA = 1.35 to produce voids in sapphire, for example, the focusing depth should
typically be not deeper than 20 m under the surface (see Section 6.4 and Juodkazis et al., 2006a, 2009).
6.2.2. Limitations imposed by the laser beam self-focusing
Self-focusing of a powerful laser beamimpose a limitation to the energy concentration which can be achieved in the bulk
of a transparent solid. The laser power should be kept below the threshold for a particular medium, which depends on the
nonlinear part of the refractive index, n
2
, (n = n
0
+n
2
I), which in turn is determined by the laser wavelength and intensity.
The critical value for the laser beam power, W
cr
, reads (Akhmanov et al., 1988):
W
cr
=

2
2n
0
n
2
. (6.3)
Fig. 6.3. The light intensity for the aberration-free (in air) and inside LiNbO
3
focusing was simulated by a 3D finite difference time domain (FDTD) code
for a focusing angle = 35, = 800 nm, and for a linearly-polarised E-field E (1; 0; 0) of a Gaussian beam. The reflection from crystalair boundary (the
dashed line) reduced the intensity inside the crystal; the interference in front of sample is caused by reflection from the surface.
Source: After Gamaly et al. (2010).
The self-focusing of the beam begins when the beam power, W
0
, exceeds the critical value, W
0
> W
cr
. The Gaussian
beam then self-focuses after propagating along the distance L
sf
(Akhmanov et al., 1988):
L
sf
=
2n
0
r
2
0
_
W
0
W
cr
1
_
1/2
; (6.4)
here r
0
is the minimum waist radius of the Gaussian beam. For example, in fused silica (n
0
= 1.45; n
2
= 3.54
10
16
cm
2
/W) for = 800 nm, the critical power comprises 2 MW. The self-focusing distance for a laser pulse with
W
0
= 2W
cr
and for a minimum waist r
0
equals to 9. Therefore, one can obtain the intensities up to 10
14
W/cm
2
with 150 fs pulses and stay below the self-focusing threshold. Thus self-focusing seems to be not a limiting factor in these
conditions. The critical beam power of (6.3) can be considered as an upper limit. However, the caution should be exercised:
theory of Marburger (1975) indicated that self-focusing without beam collapse may occur even at W
cr
/4.
6.2.3. Laser intensity distribution in a focal domain
Intensity distribution in a focal spot could be described by a complicated space function (see Figs. 6.2 and 6.3). It is
instructive to have a simple formula for practical estimation deduced from the known analytical solutions. A spherical light
wave focused by a lens of radius a at the focal distance f creates the axially symmetric intensity distribution near the focus
(Born and Wolf, 2003). The intensity distribution in a focal plane coincides with a well-known Airy expression for refraction
of a plane wave on a round pinhole. We take the radius of a circle of the first Airy minimum, r
min
, as a minimum beam waist
radius at the focus that expresses as follows:
r
min
= 0.61
f
a
=
0.61
0
NA
; (6.5)
here =
0
/n is the wavelength of light in the medium with the refractive index n,
0
is the wavelength in vacuum,
NA = n sin[arctg(a/f )]
= na/f see Fig. 6.2.

The energy within the first Airy disc (83%) is close to that within 1/e
2
envelope of the Gaussian beam (86%). Therefore
it is usually assumed that Eq. (6.5) defines the waist of the Gaussian beam as well. The intensity decreases to a half of
maximum at the distance z
R
, and r
1/2
(z
R
) = (ln 2/2)
1/2
r
min
. The focal volume then is defined as a cylinder with the length
equal to the doubled Rayleigh length, z
R
= r
2
min
n/
0
, and with the radius equal to the waist value at r
1/2
(z
R
) fromEq. (6.5),
V
1/2
= 2z
R
r
2
1/2
. Then the focal volume reads, V
1/2
= 2
2
n(ln 2/2)r
4
min
/
0
. And finally one gets (Gamaly et al., 2006):
V
1/2
= 0.95
n
3
0
(NA)
4
. (6.6)
In the conditions of the experiments of Juodkazis et al. (2006a,b,c) one gets V
1/2
= 0.2 m
3
(n = 1.453; = 800 nm;
NA = 1.35). It is convenient to present it as equivalent sphere with a radius of r
sph
(r
1/2
+ z
1/2
)/2. If one can deliver in
this focal volume the energy of 100 nJ in 100 fs (that is easily achieved by the commercially available lasers) then the energy
density reaches 510
5
J/cm
3
which corresponds to pressure 0.5 TPa, which is higher than the bulk modulus for the majority
of materials. However, it is shown by Juodkazis et al. (2006a,b,c) that the optical properties in the interaction volume are
changing dramatically during the pulse. This results in further decrease of energy deposition volume and increase in the
energy density.
We should stress that Eq. (6.6) is an approximation of the focal volume, because the wavefront of the real laser beam is
not exactly a plane wave, and the intensity distribution in a focal plane is approximated by Gaussian distribution.
6.3. Non-destructive interaction: formation of diffractive structures in photo-refractive materials
We present a logical sequence of material modification following the increase of intensity in the focal region: firstly,
changes in optical properties occur due to moderate modification of electronic and structural properties; this is followed
by phase transitions (bonding and crystal structure rearranging, crystal-amorphous, semiconductor-to-metal, solidliquid,
etc.); and withfurther intensity increase the material decompositionoccurs accompanied by breaking of inter-atomic bonds,
and ionisation. Any further increase of intensity above the ionisation threshold results in the heating of the plasma created.
The lasermatter interaction and laser-induced phase transformations are considered in this section in the conditions when
the energy density in the interaction volume is below the structural damage threshold and ionisation. There are numerous
effects induced by fs lasers in different materials at the intensity belowthe damage threshold: the photorefractive effect, the
formation of colour centres, photodarkening in silica and chalcogenide glasses, etc. The analysis below is restricted to the
phenomena in fs-laser excited photo-refractive crystals. This particular material is chosen because the fs-laser interaction
with a photorefractive crystal is attractive for applications in 3-D optical memories with additional ability to design of
readwriteerase devices. The interactions of long laser pulses and CW lasers with the photo-refractive materials were
studied thoroughly during the past decades and the mechanisms for refractive index modification at low intensities are
well understood. Below we consider the interaction of powerful fs-pulses with these crystals and compare the results with
the low intensity long pulse interactions.
The major difference between the fs-laser interaction with photo-refractive crystals and that with the long pulse lasers
relates to the fact that the laser field of high intensity is applied during the period shorter than major relaxation times. The
fact that the field of spontaneous polarisation has no effect during the pulse is particularly important. Indeed, as it follows
from the experiments (Juodkazis et al., 2002), the lowest intensity at which changes of the refraction index produced by a
single 800 nm, 150 fs laser pulse could still be detected is of the order of 4 10
11
W/cm
2
(6 nJ per pulse). This intensity
is just 23 times lower than the ionisation threshold for the dielectrics. Therefore one may expect that electron excitation
occurs differently from the low intensity case. It is instructive to recollect the main processes contributing to formation of a
diffractive structure in the photo-refractive materials under light illumination at low intensity (Valley, 1983).
The photo-excitation of free carriers into the conduction band occurs during the interaction time. The excited free carriers
are subject for the following processes: recombination, drift in a local field of charge separation, in photovoltaic field and
in diffusion field related to the carriers density and temperature gradients, while ions remain fixed. The carriers recombine
in a different location from where they were created because the recombination time is longer than the pulse. Thus, the
gradients in space distribution of charge carriers and a corresponding electric field are created. This field then induces a
refractive index modulation via the electro-optical effect. It is also important to define the time for the transition to quasi-
steady charge distribution and total life-time of the charge distribution that defines the reliability of this process for 3D
optical memory applications.
6.3.1. Properties of lithium niobate
We consider the interaction with Lithium Niobate (LiNbO
3
), a photo-refractive crystal widely used in different
applications, as a typical example of reversible transformations induced by fs-lasers inside a transparent material. LiNbO
3
structure consists of Li
+
, Nb
5+
, and O
2
see Fig. 6.4. The band gap is 3.8 eV. In the iron-doped crystal Fe iron introduces
a sub-level for Fe
2+/3+
at 1.5 eV from the conduction band. Iron concentration at 400 ppm corresponds to the number
density of iron atoms of n
Fe
= 1.8 10
19
cm
3
.
Conventionally the photorefractive materials are doped with transitional metals in order to make easier the excitation
of the free carriers at low intensity. At high intensity all other constituent atoms in the photorefractive medium are excited
to the level exceeding the concentration of dopants by a couple of orders of magnitude. The high electrons excitation and
ionisation leads to modification of optical properties and increased absorption. However, the spontaneous polarisation,
which is the intrinsic properties of the ferroelectrics, and doping appear to be important for making sustainable reversible
changes in a crystal. We are going to show that a short, high-intensity, pulse action is drastically different from the low
intensity case. First, we revisit the electronic excitation as it is treated at lowintensity and with long pulses and demonstrate
that it is incompatible with short pulse experiments with photo-refractive crystals. Then we present the concept of short
pulse interaction with photo-refractive media that complies with experiments.
6.3.2. Electronic excitation by a low-intensity laser field
The excitation rate of electrons from valence to conduction band by the laser beam at low intensity, I (W/cm
2
), in
accordance with (Valley, 1983) reads:
Fig. 6.4. LiNbO
3
cell structure (calculated by Crystal Maker; ICSD code: 155360); space group R3m(161).
Table 6.1
Oscillation energy, impact and multiphoton ionisation rates and number density of excited electrons at the
end of 150 fs 800 nm laser pulse in Lithium niobate at various laser intensities.
I (W/cm
2
) 4 10
11
10
12
2 10
12
6 10
12
osc
(eV) 0.024 0.06 0.12 0.36
w
imp
(s
1
) 0.18 10
12
0.45 10
12
0.9 10
12
2.7 10
12
w
mpi
(s
1
) 2.96 10
9
3.15 10
10
1.88 10
11
3.2 10
12
n
e
(cm
3
) 4.24 10
19
0.46 10
21
2.5 10
21
0.56 10
23
dn
e
dt
=
I
h
; (6.7)
it is proportional to the number density of photons with the energy

h, arriving to the unit volume per unit time, I/
h;
here is the quantum efficiency of a single photon, = 5 10
4
, and is inverse of the absorption length (for Lithium
Niobate 0.0033 cm
1
). It is shown (Gamaly et al., 2010) that Eq. (6.7) underestimates the number of excited electrons
in lithium niobate affected by ultra-short laser (150 fs, 800 nm;

h = 1.55 eV; intensity of 10
12
W/cm
2
) by nine orders of
magnitude. As we show later the excitation at high intensity with short pulses should be treated in a quite different way.
Another issue relates to the relative role of laser field and inherent photovoltaic field on the modification of the refractive
index by the laser action. Buse presents the excitation of electrons at low intensity as a stationary process when excitation
of electrons from Fe
2+
centres (number density n
Fe
2+) is balanced by recombination on Fe
3+
(n
Fe
3+) centres (Buse, 1996):
dn
e
dt
= qSIn
Fe
2+ n
Fe
3+n
e
. (6.8)
The source terms in Eqs. (6.8) coincides with that of Eq. (6.7) assuming that absorption occurs on Fe
2+
centres, = l
1
=
n
Fe
2+S (S is the absorption cross section), and q = /
h. In a stationary case one obtains the number density as follows:

n
e
=
qI
n
Fe
3+
. (6.9)
The photovoltaic current has a conventional form:
j
phv
= en
e
v
e
. (6.10)
The electron acceleration by all the electric fields and stationary velocity follows from the Newton equation:
m
e
dv
e
dt
= eE
las
+eE
int
ef
m
e
v
e
. (6.11)
Here
ef
is the effective collision frequency responsible for resistance, m
e
is the electron mass, and E
int
represents all
fields (including photovoltaic E
phv
) in a medium except for the incident laser field in the first term. In conventional photo-
refractive effect the electron motion is stationary and laser field is neglected because laser intensity is low in comparison to
photovoltaic field. Thus the velocity of electron accelerated by the photovoltaic field reads:
v
e
=
eE
phv
ef
m
e
. (6.12)
Fig. 6.5. (a) Avalanche (solid line) and multi-photon (dashed line) ionisation rates for 150 fs, 800 nm pulses in LiNbO
3
. Note that multi-photon ionisation
takes over the avalanche ionisation at intensities above 5.5 TW/cm
2
. (b) Normalised electron density vs. laser intensity at 150 fs, 800 nm (Table 6.1); the
horizontal line marks the critical density. The optically detectable changes of refractive index in Fe:LiNbO
3
and undoped LiTaO
3
are marked by arrows at
1 TW/cm
2
and 1.37 TW/cm
2
, respectively.
On the other hand the photovoltaic current, j
ph
, excited in non-centrosymmetric crystal by the laser excitation with intensity
I, depends on the material properties through the form (Glass et al., 1974):
j
ph
= GI. (6.13)
Here G is the Glass constant, which is a characteristic of a particular material, and (cm
1
) is the inverse of the absorption
length as above. Nowinserting Eqs. (6.9) and (6.12) into Eq. (6.10) and comparing with Eq. (6.13) one obtains the expression
for the photovoltaic field in the form:
E
phv
=
G n
Fe
3+
q
ef
m
e
e
2
. (6.14)
One can see that the photovoltaic field does not depend on the laser intensity, and it actually represents the field of
spontaneous polarisation that is the inherent property of ferroelectric photo-refractive crystals.
6.3.3. Electrons excitation by the high-intensity ultra-short pulse
Three major mechanisms contribute to the light absorption in solids (Ilinskii and Keldysh, 1994): the inter-band
transitionsfor example, single and multi-photon absorption, intra-band transitions, which is absorption on electrons in
the conduction band, and absorption on the donor (acceptor) level that locates inside the band gap. It was found that the
diffractive structure formation by 150 fs pulses can be observed at intensity 4 10
11
W/cm
2
at 800 nm (6 nJ per pulse).
This intensity is just 23 times lower than the ionisation threshold for the dielectrics (Juodkazis et al., 2003).
It instructive to note that the critical electron density, which signifies breakdown threshold, for 800 nm ( = 2.35
10
15
s
1
) is n
c
= m
e
/4e
2
= 1.735 10
21
cm
3
. The electric field amplitude at the intensity of 10
12
W/cm
2
equals to
E = (8I/c)
1/2
= 9.15 10
4
CGS = 27.46 MV/cm, which is 2 orders of magnitude larger than the photovoltaic field of
10
5
V/cm in Lithium Niobate (Glass et al., 1974). The qualitative indication of the electric field strength is the oscillation
energy of free electron and oscillation amplitude in this field. The oscillation energy for linear polarised light reads (Luther-
Davies et al., 1991, 1992):
osc
= 9.3 10
14
I
2
; (6.15)
here I is intensity in W/cm
2
and
osc
is in eV, and is in m. At 10
12
W/cm
2
, 800 nm, this energy equals to 0.06 eV, while the
oscillation amplitude of free electron is around of an angstrom. One can see that the oscillation amplitude of bonded electron
is comparable to the displacements of Li of 0.9 and Nb of 0.5 in LiNbO
3
responsible for the spontaneous polarisation
(see Fig. 6.5). Therefore it might be expected that during the laser pulse of such intensity the oscillating electrons affect the
intrinsic crystal field and ferroelectric properties. In these conditions, consideration must be given to the major mechanisms
responsible for excitation of electrons to the conduction band from all constituent atoms in the crystal. It is known that
ionisation by the electron impact (avalanche ionisation) and the ionisation produced by simultaneous absorption of multiple
photons are the two most important mechanisms for electron excitation in close to the ionisation threshold conditions.
(a) Avalanche ionisation
A few seed electrons in the conduction band oscillate in the electromagnetic field of the laser in dielectrics (Eq. (6.15)).
Electron can gain net energy by multiple electronphonon (lattice) collisions and eventually be accelerated to the energy in
excess of the band-gap >
gap
. Energetic electrons create an avalanche of ionisation events. The probability of such event
per unit time (ionisation rate) estimates as follows (Raizer, 1978):
w
imp

osc
gap
eph
(
2
eph
+
2
)
; (6.16)
e, m
,
eph
, and are respectively the electron charge, effective mass, electronphonon momentum exchange rate and the
laser frequency. The electronphonon momentum exchange rate approximately expresses as
(mom)
eph
C
T
L
/
h (C
is the
constant, see Section 2) at the crystal temperature being larger than the Debye temperature T
L
> T
D
(Ilinskii and Keldysh,
1994; Gamaly and Rode, 2009). This rate at the room temperature of 293 K constitutes 3.83 10
13
s
1
. For LithiumNiobate
(
gap
= 3.8 eV) under the action of 800 nm laser ( = 2.35 10
15
s
1

eph
) the avalanche excitation rate as function
of laser intensity then expresses as
w
imp
4.46 10
13
I
10
14
s
1
.
Here the intensity I is inW/cm
2
. Note that ionisationrate of metal dopants is approximately twice higher because the doping
introduces an additional energy level in the band gap.
(b) Multi-photon ionisation
It is reasonable totake the multi-photonionisationrate (probability of ionisationper atomper second) inthe form(Raizer,
1978):
w
mpi
n
3/2
ph
_

osc
2
gap
_
n
ph
; (6.17)
where n
ph
=
gap
/
h is the number of photons, which an electron should absorb in order to be transferred fromvalence to
the conduction band. Again taking as an example Lithium Niobate under the action of 800 nm laser one gets the intensity-
dependent multi-photon ionisation rate in the form:
w
mpi
4.55 10
15
_
I
10
14
_
2.58
(s
1
).
(c) Number of excited electrons produced to the end of the pulse
The number density of electrons n
e
created to the end of the pulse jointly by the avalanche and multi-photon processes
can be obtained with the help of the simplified rate equation (Gamaly et al., 2002):
dn
e
dt
= n
e
w
imp
+n
a
w
mpi
. (6.18)
Let us assume that the laser intensity is constant during the laser pulse and the recombination is negligible. Then solution
to Eq. (6.18), with the initial condition n
e
(t = 0) = n
e0
and w
imp
and w
mpi
are both time independent, is straightforward:
n
e
(I, , t) =
_
n
e0
+
n
a
w
mpi
w
imp
[1 exp(w
imp
t)]
_
exp(w
imp
t). (6.19)
Let us find the intensity level for the optical breakdown to occur in Lithium Niobate under the action of 800 nm, 150 fs laser
pulse fromthe condition that the plasma frequency of excited electrons equals to the frequency of impinging laser. Intensity
dependent oscillation energy, ionisation rates and number density of excited electrons to the end of 150 fs laser pulse in
Lithium niobate are presented in Table 6.1.
The critical electron density at 800 nmis 1.73510
21
cm
3
. Thus one can conclude that the simple model presented here
predicts the breakdown threshold to be around 1.5 10
12
W/cm
2
in semi-quantitative agreement with 1.0 10
12
W/cm
2
from experiments of Juodkazis et al. (2003) (see Fig. 6.5).
(d) Recombination rate
Recombination in three body collisions may occur with electron, an ion, or a neutral atom, each of those can act as a
third body. One can take an atomic cross section for estimation of ei collision rate in the case when ionisation degree is
below the breakdown threshold. The probability of recombination per unit volume per unit time (recombination rate) then
is a product of probability of ei collision,
col
= n
e
n
+
v
e
, and the probability for a third body presence in the vicinity
of colliding particles, p
3b
. The last one can be approximated by the ratio of atomic volume to the average volume per
atom, p
3b
4r
3
at
n
a
/3, if an atom can be considered as a third body. In this case this probability is close to unity. Thus
recombination time estimates as t
rec
(n
e
at
v
e
)
1
(10
21
10
15
10
6
)
1
, and it is around 1 ps at under-threshold
condition (n
e
< 1.74 10
21
cm
3
).
At the electronand ion density close to the breakdownthreshold the recombinationin the triple Coulomb collisions when
the electron acts as a third body can be of importance. In this case the Coulomb collisions are characterised by the cross
section r
2
0
(e
2
/
el
)
2
; here
el
is the electron energy. Excited electrons in the conduction band can be approximately
treated as the degenerated Fermi gas with the energy
F
= (3
2
n
e
)
2/3
h
2
/2m = 5.83 10
27
n
2/3
e
erg, and heat capacity
C
e
=
2
k
2
B
T
e
/2
F
. At n
e
= 10
21
cm
3
el

F
= 5.83 10
13
erg = 0.36 eV, and v
F
= 2.5 10
7
cm/s. One can easily see
that the Coulomb cross section is of two orders of magnitude higher than the atomic cross section. The electron mean free
path l
mfp
= v
F
/
eph
10
6
cm is much shorter than the laser-modified absorption length as it is shown below. Therefore
the recombination time in the triple Coulomb collisions can be in the order of femtoseconds and hence the recombination
during the pulse time might be significant and fast. On the other hand the recombination time is inverse proportional to the
electron density. Taking recombination cross-section and electron velocity time independent one obtains, t
rec
t
0
n
0
/n
e
,
and very slow decrease in time for the electron number density, n
e
(t) = n
e
(t
p
)t
0
/t; here t
0
is recombination time at the
end of the pulse. For n
e
(t
p
) = 10
21
cm
3
t
0
= 1 ps. On the other hand if the electron number density comparable to that
for metal dopants, i.e. n
e
(t
p
) = 10
18
cm
3
, then the recombination time is t
0
1 ns as it suggested at low intensity studies
(Valley, 1983).
6.3.4. Modification of the optical properties of fs-laser-excited crystal
Now let us consider the change in the optical properties during the pulse under the assumption that intensity during
the pulse time is lower than the breakdown threshold and therefore the laser-induced modification could be reversible. The
total dielectric function for a dielectric modified by electrons excitation at high intensity is the sumof unperturbed function
and contributions from excited electrons:
=
0
() +i
4

0
()
4
im
m
e
+i
4
re
m
e

0
()
re
+i
im
;
here
re,im
are the real and imaginary parts of conductivity, respectively. It is reasonable to take the conductivity for excited
electrons in the Drude form:
=
e
2
n
e
m(
eff
i)
=
e
2
n
e
eph
m
e
(
2
eph
+
2
)
+i
e
2
n
e
m
e
(
2
eph
+
2
)
. (6.20)
The effective collision rate,
eff
, includes all collision processes. Then one can find the contributions to the real and
imaginary parts of the dielectric function above at
eph
as follows:
re

2
pe
2
;
im

2
pe
3
;
re
=

im
4
. (6.21)
Note that in the considered here conditions the inequality holds,
im

re
. The modified refractive index, N
=
n +ik, then reads:
n (
0
re
)
1/2
; k

im
2n
. (6.22)
The change in the refractive index due to electron excitation is negative:
n
n
0

re
2n
2
0
. (6.23)
The numerical values are estimated later in this section.
(a) Electron and lattice temperature
Now, the important characteristics of laser-excited crystal at the end of a 150 fs pulse can be calculated: the absorption
length, l
s
= c/k, the absorption coefficient from the Fresnel formula A 4n(n + 1)
2
(n k), and the absorbed energy
density, E
abs
= 2AF
p
l
1
s
. Let us assume that conduction electrons can be characterised by the Fermi energy
F
n
2/3
e
and the corresponding heat capacity. Under these assumptions the maximum electron temperature is as the following:
T
2
e,max
= 4
F
AF
p
/
2
n
e
l
s
.
To obtain numerical estimates let us assume that the number density of excited electrons is in a range 10
20
10
21
cm
3
<
n
crit
, lower but in the vicinity of the critical density in accord to experiments. Therefore squared plasma frequency is in a
range
2
pe
= (0.3183.18) 10
30
s
2
, of the same order of magnitude as the squared laser frequency, 5.52 10
30
s
2
, and
both frequencies are much higher than the electronphonon momentumexchange rate. The absorbed energy density to the
end of the pulse, F
p
, is taken as the average intensity during the pulse, F
p
= I t
p
. Dielectric function of unperturbed Lithium
Niobate at low intensity at 800 nm equals to
0
= 6 (n = 2.45; R
0
= 0.1766) (Nikogosyan, 1997). The characteristics of
laser-modified crystal estimated by the formulae given above are presented in Table 6.2 and illustrated in Fig. 6.5.
Note that intense laser pulse drastically modifies the optical properties. For example, the effective absorption that defines
the absorbed energy density equals to A/l
s
= 131 cm
1
at average intensity of one TW that should be compared to
0.0033 cm
1
(for 653 nm) for long pulse low intensity case. The absorbed energy density at I t
p
= 0.15 J/cm
2
comprises 39.4 J/cm
3
. This may result in only insignificant increase in the lattice temperature.
Table 6.2
Modified optical properties of Lithium niobate at laser intensity 0.55 and 1.5 TW/cm
2
with 150 fs, 800 nm pulses.
n
e
F (J/cm
2
)
re

im
n
1/2
re
k

im
2n
l
s
(cm) A E
abs
(J/cm
3
) T
e,max
(eV)
re
10
12
(s
1
)
10
20
0.083 0.057 0.9410
3
2.44 0.19310
3
6.6 10
2
0.82 1.25 4.5 10
2
0.176
10
21
0.225 0.576 9.4 10
3
2.33 2.0 10
3
6.4 10
3
0.84 39.4 0.28 1.76
(b) Relaxation processes in laser-excited photo-refractive crystal
The excited electrons gain significant energy to the end of the short pulse (see Table 6.2). However the absorbed energy
density is not high enough for inducing the structural changes. Therefore the electron-to-lattice energy transfer time can be
estimated as for the cold solid as:
en
eph

=
_
m
e
M
_
1/2
mom
eph
.
For Lithium niobate this rate is 1.63 10
11
s
1
, giving the lattice heating time of 6 ps. Thus the lattice reaches
the maximum temperature during this time taking into account that heat conduction in LiNbO
3
is poor (heat diffusion
coefficient is D = 1.5 10
2
cm
2
/s). One can also observe that cooling of micron-size laser-affected spot, d, takes time
t
cool
= d
2
/D 700 ns. Assuming that heat capacity obeys the DulongPetit law one can estimate the increase in the lattice
temperature as:
T
L,max
=
2AF
p
l
s
C
L
n
a
.
Thus, the lattice temperature of Lithiumniobate increases on 10 K when the absorbed energy density comprises 39.4 J/cm
3
.
One cannot expect any structural changes at such temperature; only the charge re-distributions are expected. It is also
follows from the above that the cooling can be neglected during the time required for establishing a quasi-stationary
distribution of charges.
6.3.5. Quasi-stationary changes in optical properties of fs-laser-affected crystals
The conditions in photo-refractive crystal created by the action of intense femtosecond pulse are in sharp contrast to
those produced by long pulse or cw lasers. Firstly all processes during the pulse are transient. The quasi-stationary state is
achieved long after the pulse. We shall consider nowall relevant physical processes contributing towards the changes in the
optical properties of the material during the pulse and after the pulse.
(a) Processes during the pulse
Let us estimate the electron current, j
ph
= en
e
e
created during the short (150 fs) and intense (10
12
10
13
W/cm
2
) laser
pulse. The pulse generates the number of excited electrons of the order of 10
21
cm
3
. The electrons oscillate during the pulse
in the high frequency electric field with amplitude (2.78.7) 10
7
V/cm, which is several orders of magnitude higher than
the inherent field of spontaneous polarisation in photo-refractive crystals. Therefore it is impossible to establish a stationary
distribution of electric field during the pulse. Indeed, the photoconductivity fromEq. (6.20) in the condition that the collision
rate can be neglected in comparison to the laser frequency, reads:
ph

e
2
n
e
eph
m
e
2
. (6.24)
It is a strong function of the carrier number density. For 800 nmlight ( = 2.35610
15
s
1
), collision rate of 3.810
13
s
1
,
and n
e
= 10
21
cm
3
it yields
ph
2
1
cm
1
(1.74 10
12
s
1
in the Gaussian units). If the recombination in
triple Coulomb collision is significant during the pulse time, the photoconductivity might be lower and photovoltaic
field higher. The above conductivity value defines the time for establishing quasi-stationary distribution of electric field,
t
stat

=
st
/4
ph
1 ps. Thus, stationary distribution could be established only after the end of the pulse.
(b) Processes after the pulse
After the pulse the electrons are subject for recombination, for a drift in a local field of spontaneous polarisation, which
is inherent to a photo-refractive crystal, and for diffusion under the electrons temperature and density gradients, while the
ions remain fixed in their positions. The diffusion field is much smaller than other fields, so it can be neglected. The essential
difference in comparison with conventional long pulse low intensity case is that the field of charge separation, which is
finally responsible for a quasi-permanent change inthe refractive index, establishes due tothe spontaneous polarisationfield
in the absence of the external field of laser irradiation. Because the recombination time is longer than the pulse duration, the
carriers recombine ina different locationfromwhere they were created, most probably inthe ironsites. Thus there are strong
grounds to assume that the charge separation field, E
cs
, equals to E
s
the internal field of spontaneous polarisation in the
ferroelectric crystal, is unaffected by the electron excitation. It is conventionally assumed that quasi-stationary distribution
establishes during the so-called Maxwell time (Petrov et al., 1991):
t
stat

st
/4
d
st
is the static dielectric function in the absence of external field. The conductivity in the absence of photo excitation, the
so-called dark conductivity, reads:
d

e
2
n
e
m
e
eph
. (6.25)
Note that
d

ph
because
eph
. Taking
st
= 29; n
e
10
18
cm
3
,
(mom)
eph
T
L
/
h = 3.83 10
13
s
1
, and the
electron mass as for a free electron, the time for establishing the stationary distribution is less than a picosecond.
Spontaneous polarisation P
s
in LiNbO
3
reaches its maximumof P
s
= 7.110
5
C/cm
2
at the Curie point of 1480 K(Kittel,
1996). Respective electric field equals to E
s
= P
s
/3
0
= 2.67 10
6
V/cm (
0
is the dielectric permeability of vacuum in
SI units). Spontaneous polarisation strongly depends on temperature. Glass et al. (1974) gave the value for the photovoltaic
field in iron-doped LiNbO
3
at room temperature E
s
(T
r
) 100 kV/cm, which we use for the following estimates.
6.3.6. Possible mechanisms for changes in the refractive index after the pulse
Generation of large amount of excited electrons immediately results in the decrease in the refractive indexsee
Eq. (6.14):
n
n
0

re
2n
2
0

2
pe
2n
2
0
2
. (6.26)
At the electron number density of n
e
= 10
21
cm
3
at the end of the pulse the change constitutes n/n
0
= 4.810
2
. This
index modification decreases in proportion to the decrease of the number density of free carriers and reaches undetectable
level of n/n
0
10
5
after several nanoseconds. However, we should note that the real recombination time is unknown.
This might be the case of a perfect undoped crystal free of defects, which serve as trapping centres.
Most probably, the field of spontaneous polarisation dominates the process of charge separation in a metal-doped crystal.
Then the change in the refractive index due to the electro-optic effect reads (Petrov et al., 1991):
n
n
3
r
2
E
cs
. (6.27)
We take the known value of n
3
r 3 10
8
cm/V for Lithium Niobate. Now, the refractive index changes can be estimated
as n 1.5 10
8
(cm/V) E
cs
. For E
cs
10
2
kV/cm these changes are expected to be around n (1.5) 10
3
.
6.3.7. Comparison with experiments
The interaction of single intense femtosecond laser pulses tightly focused in the bulk of LiNbO
3
and LiTaO
3
crystals (with
pure crystals and with crystals doped with Fe) was recently studied by Juodkazis et al. (2003, 2006a,b,c, 2008), Beyer et al.
(2006) and Gamaly et al. (2008, 2010). Single 150 fs laser pulses at 800 nmwith energy per pulse in a range of 350 nJ, were
focused inside a crystal to the depth of 50 m and to the focal spot with diameter of 1.8 m (S
foc
= 2.54 10
8
cm
2
), so
that intensity in the interaction region varied from 1.0 to 16.7 TW/cm
2
. The optically detectable change of refractive index
were observed in Fe: LiNbO
3
at the energy per pulse 3.8 0.5 nJ and at 5.2 0.5 nJ in pure LiNbO
3
crystal. These figures
correspond to intensity of 1.0 TW/cm
2
and 1.37 TW/cm
2
respectively. The permanent modification of LiTaO
3
was observed
at 32 5 nJ (10 TW/cm
2
) that might be considered as a result of breakdown. Therefore the breakdown threshold locates
at the average intensity 10 TW/cm
2
. It also has been found that the laser-induced transformation of the material is fully
reversible at the energy per pulse of 14.5 nJ (3.8 TW/cm
2
). One can see that the theoretical estimates for a breakdown are
in qualitative agreement with the measurements. Thus it is confirmed that generation of free carriers by intense fs-pulses
occurs due to intertwined avalanche and multi-photon processes.
The changes of refractive index were recovered from the light transmission measurements. In z-cut pure LiTaO
3
crystal
(cut is perpendicular to c-axis) the maximum measured relative change in refractive index was 2.5 10
4
. In y-cut pure
LiNbO
3
(cut is parallel to c-axis) the change was 5 10
4
. Maximum reversible change in Fe:LiNbO
3
(400 ppm doping)
constitutes 10
3
. The laser-affected area has characteristic index modulation pattern +n/n/+n that corresponds
to dark/bright/dark regions along the crystallographic c-axis. This occurs inbothFe-dopedandundopedcrystals. This feature
is a qualitative evidence of the fact that the photovoltaic process was followed by the electro-optic effect, and both are
responsible for the index modulation. There is however a substantial difference in the after-pulse behaviour of laser-affected
regions in Fe-doped and undoped LiNbO
3
. Index modulation in Fe-doped crystal is long-lived while the effect in undoped
crystal completely disappears in 0.250.3 s, most probably due to recombination of the free carriers. On the other hand the
long life of the charge separation and resulted modulation of refractive index in Fe-doped crystal most probably occurs due
to the presence of the metal trapping centres separated by distance of n
1/3
Fe
10
6
cm from each other, which is much
larger than free carrier mean free path of v
e
/
eph
. The last fact indicates that the non-local approach should be applied for
the current calculation instead of the Ohms law.
Thus, it has been demonstrated that the interaction of intense femtosecond pulse with photorefractive crystal in the
conditions close to the breakdown threshold has several distinctive features in comparison to that of the long pulse (or cw)
lasers. First, the high number density of excited electrons modifies the dielectric function and leads to the negative change
in refractive index, 4.8 10
2
, exceeding that due to the charge separation long after the pulse. The amount of change
depends on laser intensity and independent on the beampolarisation. However, this index change is transient, it disappears
when the recombination is completed at the nanosecond time scale. Second, the dominance of high frequency laser field,
which is two orders of magnitude higher than the field of spontaneous polarisation, makes the stationary charge distribution
impossible during the pulse.
Third, the diffusion and recombination of the charge carriers continue long after the end of the pulse (on the nanosecond
time scale). The maindriving force responsible for the current is the field of spontaneous polarisation: the current terminates
when the field of charge separation balances this field. Quasi-stationary distribution of charges that results in change of the
refractive index due to the Pockels effect occurs well after the pulse. Modification of refractive index derived from this
theory is in a semi-quantitative agreement with experiment. In the model presented here the modification of refractive
index should be independent on polarisation that is also in agreement with observation of index modifications in dielectrics
(Juodkazis et al., 2003). These findings suggest that the laser electric field with high amplitude modifies not only linear
properties of the material. Most probably strong AC field also induces transient ferroelectric and non-linear properties
of a crystal. Therefore a new avenue opens up for the studies of the intensity-dependent transient phase transformation
induced by femtosecond lasers at intensity close but belowthe damage threshold. Pumpprobe experiments might provide
the information of time-dependent dielectric function of the excited crystal with fs resolution, while harmonics generation
may provide information on transient non-linear properties.
6.4. Lasermatter interactions confined inside a solid at high intensity
At least two conditions should be fulfilled in order to confine lasermatter interaction inside transparent material. First,
the powerful laser pulse should be transported over a distance larger than self-focusing length without losses. Second, this
distance ought to be far enough from the outermost crystal surface in order the interaction region can be considered as
confined. In addition, the laser energy has to be focused to the smallest possible volume, often to the dimensions of the order
of the laser wavelength , to achieve optical breakdown threshold and significantly increase absorption due to drastically
changed optical properties under the laser action.
The major mechanism of absorption in the low intensity lasersolid interaction is the inter-band electron transition.
Since the photon energy is smaller than the band-gap, the electron transitions are forbidden in linear approximation, which
corresponds to a large real and small imaginary part of the dielectric function. The optical parameters in these conditions are
only slightly changed during the interaction in comparison to those of the cold material described in the previous section.
The absorption can be increased for shorter wavelengths where the photon energy becomes larger then the band-gap
value or if the incident light intensity increases to the level well over the ionisation and ablation thresholds and strong
non-linear processes become dominant. We explore the second possibility in this section. The interaction swiftly changes
to laserplasma mode above the ionisation threshold, the absorption increases and absorption depth shrinks. A localised
depositionof the laser light creates a regionof highenergy density. If the pressure inabsorptionvolume significantly exceeds
the Young modulus of the solid, a void in the bulk of material is created. Single pulse action thereby allows a formation of
various three-dimensional structures inside a transparent solid in a controllable and predictable way.
Single short pulse tightly focused inside the bulk of a transparent solid can easily generate the energy density in excess of
the Young modulus of any of existing solid within a focal volume less than a cubic micron (Glezer and Mazur, 1997; Juodkazis
et al., 2006a,b,c; Gamaly et al., 2006). The pressure of the order of several TPa inside a focal volume leads to formation of
a cavity (void) surrounded by a shell of compressed material. These two features of the phenomenon delineate two areas
of studies and applications. The first area relates to formation of different 3D structures, photonic crystals, waveguides,
gratings etc. making use of multiple voids (separated or interconnected) created in the different space points of a crystal.
For these studies the most important part is the void formation. As it is shown later in this section, in order to produce a void
one has to generate a pressure in excess of the strength (the modulus) of a material. The second area of research relates to
the studies of transformations of the materials under high pressuretemperature conditions, which are possible to create
in tabletop laboratory experiments. The interaction of a laser with matter at intensity above the ionisation and ablation
threshold proceeds in a way similar for all the materials (Gamaly et al., 2002). The material converts into plasma in a few
femtoseconds early in pulse time, changing the interaction to the laserplasma mode, increasing the absorption coefficient
and reducing the absorption length, which ensures a fast energy release in a very small volume. A strong shock wave is
generated in the interaction region and this propagates into the surrounding cold material. The shock wave propagation is
accompanied by compression of the solid material at the wave front and decompression behind it, leading to the formation
of a void inside the material. The laser and shock wave affected material is confined in the shell that surrounds the void and
this shell is the major object for studies of new phases and new material formation. This section is devoted to intense short
pulse lasermatter interaction confined inside a transparent solid.
Let us first to underline the differences between the intense lasermatter interaction at the surface of a solid and the case
when lasermatter interaction is confined deep inside a solid. We compare the pressure created at the absorption region at
the same intensity and total absorbed energy. At the intensity well over the ionisation and ablation thresholds any material
converts into plasma in a fewfs time. Therefore the interaction proceeds most of the time in laserplasma interaction mode.
In these conditions the pressure at the ablated plumesolid interface (in lasersurface interaction) constitutes fromthe sum
of thermal pressure of plasma next to the ablation boundary plus the pressure from the recoil momentum of expanding
plasma. Significant part of absorbed energy is spent on the expansion and heating of the ablated part of a solid. Therefore
the ablation pressure in this case depends on the absorbed intensity by the power law P
abl
I
m
abs
; m < 1. There is no
expansion loss in the case of confined interaction. Hence the maximum pressure is proportional to the absorbed intensity
P
conf
I
abs
and it s almost twice bigger than in the surface interaction.
Full description of the lasermatter interaction process and laser-induced material modification fromthe first principles
embraces the self-consistent set of equations that includes the Maxwells equations for the laser field coupling with matter,
complemented with the equations describing the evolution of energy distribution functions for electrons and phonons (ions)
and the ionisation state. A resolution of such a system of equations is a formidable task even for modern supercomputers.
Therefore, a thorough analytical analysis is needed. We split below this complicated problem into a sequence of simpler
interconnected problems: the absorption of laser light, the ionisation and energy transfer from electrons to ions, the heat
conduction, and hydrodynamic expansion.
6.4.1. Absorbed energy density
Let us first to define the range of laser and focusing parameters necessary for obtaining high pressure inside the
interaction region. A 100 nJ laser pulse of duration t
p
100 fs with the average intensity I > 10
14
W/cm
2
focused into the
area S
foc

2
delivers the energy density >10 J/cm
2
, well above the ionisation and ablation thresholds for any material
(Gamaly et al., 2002). The focal volume has a complicated three-dimensional structure. As a first approximation (that is also
useful for scaling purposes) the focal volume is the focal area multiplied by the absorption length. The absorbed laser energy
per unit time and per unit volume during the pulse reads:
dE
abs
dt
=
2A
l
abs
I(r, z, t). (6.28)
The absorption depth is l
abs
= c/k. We assume that the electric field exponentially decays inside a focal volume,
E = E
0
exp{x/l
abs
} as it does in the skin layer; A is the absorption coefficient defined by the Fresnel formula (Landau
et al., 1984):
A =
4n
(n +1)
2
+k
2
=
2
|1 +
1/2
|
2
k
. (6.29)
The duration of a typical short pulse of 100 fs is shorter than the electronion collision times. Therefore the electron
energy distribution during the pulse has a delta function like shape peaked near the energy that can be estimated from
the general formula of Eq. (6.28) under the assumption that the spatial intensity distribution inside a solid, and material
parameters are time independent. We denote the energy per single electron by
e
(it should not be confused with the
dielectric function that is always without a subscript here). Then the electron energy density change reads:
d(n
e
e
)
dt
=
2A
l
abs
I(t).
We show later that the ionisation degree at I > 10
14
W/cm
2
is high, Z > 1, the number density of electrons is large, and
electrons heat capacity can be taken as that for ideal gas. Thus fromthe above one can make a rough estimate of the electron
temperature to the end of the pulse:
T
e

2A
1.5k
B
n
e
l
abs
I(t)t.
The electron temperature rises to tens of electron volts level early in the pulse time. The fast ionisation of a solid occurs
that affects absorption coefficient and absorption length. Thus, the next step is to introduce the model where the optical
properties are dependent on the changing electron number density and electron energy.
6.4.2. Ionisation
Optical breakdownof dielectrics andoptical damage producedby the actionof anintense laser beamhas beenextensively
studied over the several decades (see Section 2). It is well established that avalanche ionisation and multi-photon ionisation
are the major mechanisms responsible for the solidplasma conversion. The relative contribution of both mechanisms
depends on the laser wavelength, pulse duration, intensity, and the atomic number. Analytical estimates of the breakdown
threshold, ionisation rates and transient number density of electrons created in the absorption region allows obtain the
general picture of the processes in qualitative and quantitative agreement with computer simulations.
(a) Ionisation thresholds
The optical breakdown occurs when the number density of electrons in the conduction band reaches the critical density
expressed through the frequency of the incident light by familiar relation, n
c
= m
e
2
/4e
2
.
The ionisation threshold for the majority of transparent solids lies at intensities in between (10
13
10
14
) W/cm
2
(
1 m) with a strong non-linear dependence on intensity. The conduction-band electrons gain energy in an intense short
pulse much faster than they transfer energy to the lattice. Therefore the actual structural damage (breaking inter-atomic
bonds) occurs after electron-to-lattice energy transfer, usually after the end of the pulse. It was determined that in fused
silica the ionisation threshold was reached to the end of 100 fs pulse at 1064 nm at the intensity 1.2 10
13
W/cm
2
(Arnold and Cartier, 1992). Similar breakdown thresholds in a range of (2.8 1) 10
13
W/cm
2
were measured in
interaction of 120 fs, 620 nm laser with glass, MgF
2
, sapphire, and the fused silica (von der Linde and Schuler, 1996).
This behaviour is to be expected, since all transparent dielectrics share the same general properties of slow thermal
diffusion, fast electronphonon scattering and similar ionisation rates. The breakdown threshold fluence (J/cm
2
) is an
appropriate parameter for characterisation conditions at different pulse duration. It is found that the threshold fluence
varies slowly if pulse duration is below 100 fs. For example, for the most studied case of fused silica the following
threshold fluences were determined: 2 J/cm
2
(1053 nm; 300 fs) and 1 J/cm
2
(526 nm; 200 fs) (Stuart et al., 1995);
1.2 J/cm
2
(620 nm; 120 fs) (von der Linde and Schuler, 1996); 2.25 J/cm
2
(780 nm; 220 fs) (Lenzner et al., 1998);
3 J/cm
2
(800 nm; 10100 fs) (Tien et al., 1999).
(b) Ionisation rates
Avalanche ionisation (see Sparks et al. (1981)). In interaction of lasers in a visible range with wide band gap dielectrics
the direct photon absorption by electrons in a valence band is rather small. However, a few seed electrons can always be
found in the conduction band. These electrons oscillate in the laser electromagnetic field and can be gradually accelerated
to the energy in excess of the band-gap. Electrons with
e
>
gap
collide with electrons in the valence band and can transfer
them a sufficient energy to excite into the conduction band. Thus the number of free electrons increases, which provokes
the effect of avalanche ionisation. The probability of such event per unit time (ionisation rate) can be estimated as follows:
w
imp

1
gap
d
e
dt
=
2
osc

2
eff
(
2
eff
+
2
)
I
2
/
gap
. (6.30)
Electron is accelerated continuously in this classical approach. The oscillation energy is proportional to the laser intensity
and to the square of the laser wavelength. At relatively low temperature corresponding to low intensities below the
ablation threshold the effective collision rate,
eff
, equals to the electronphonon momentum exchange rate
eff
=
eph
.
The electronphonon momentum exchange rate increases in proportion to the increase of temperature. For example, the
electronphonon momentum exchange rate in SiO
2
is of the order
eph
= 5 10
14
s
1
(Arnold and Cartier, 1992) and
it is lower of the laser frequency for a visible light, 10
15
s
1
. The ionisation rate in Eq. (6.30) grows in proportion to
the square of the laser wavelength, in correspondence with the Monte-Carlo solutions to the Boltzmann kinetic equation
for electrons (Arnold and Cartier, 1992). With further increase in temperature and due to the ionisation, the effective
collision rate becomes equal to the electronion momentum exchange rate, and reaches its maximum approximately at
the plasma frequency (10
16
s
1
) (Eidmann et al., 2000; Gamaly et al., 2002). At this stage the wavelength dependence
of the ionisation rate almost disappears due to <
ei

pe
, as it follows from Eq. (6.30) and in agreement with
rigorous calculations of (Arnold and Cartier, 1992). It is instructive to estimate the ionisation rate for an extreme case of
high intensity and for one of the largest band gap of sapphire, when the oscillation energy of a free electron in conduction
band,
osc
(see Eq. (6.17)), equals to the band gap,
osc

gap
= 9.9 eV. Then the intensity at 800 nm laser wavelength
( = 2.356 10
15
s
1
) equals to 1.7 10
14
W/cm
2
. In the beginning of the ionisation process, when >
eph
, the
ionisation rate is w
imp
2
eph
10
14
10
15
s
1
. When the collision rate reaches its maximum, <
ei

pe
, the
ionisation rate equals to w
imp
2
2
/
ei
2
2
/
pe
5 10
14
s
1
.
Multi-photon ionisation (or tunnel ionisation, see Keldysh (1965); Oppenheimer (1928)) has no threshold and hence its
contribution to the ionisation process is important even at relatively low intensity. Multi-photon ionisation creates the
initial (seed) electrons with the density n
0
; this density grows through the avalanche process. It proved to be a reasonable
to estimate the ionisation probability per atom, per second, in the multi-photon form Raizer (1978):
w
mpi
n
3/2
ph
_

osc
2
gap
_
n
ph
; (6.31)
where n
ph
=
gap
/
h is the number of photons required for an electron to be transferred from the valence to the
conduction band. The multi-photon ionisation rate dominates, w
mpi
> w
imp
, for any relationship between the frequency
of the incident light and the effective collision frequency in conditions when
osc
>
gap
. However, the contribution of
the avalanche process even at high intensity is crucially important: at w
mpi
w
imp
the seed electrons are generated
by multi-photon effect whilst final growth is due to the avalanche ionisation. Such an inter-play of two mechanisms has
been demonstrated with the direct numerical solution of kinetic FokkerPlanck equation (Stuart et al., 1995). Under the
condition
osc
=
gap
,
h = 1.55 eV, n
ph
=
gap
/
h 6.4, and = 2.356 10

15
s
1
, the multi-photon rate comprises
w
mpi
5.95 10
15
s
1
. The ionisation time estimates as t
ion
w
1
mpi
. Thus, the critical density of electrons (the ionisation
threshold) is reached in a few femtoseconds at the beginning of a 100 fs pulse. As soon as the critical density has been
reached, the interaction proceeds in a laserplasma interaction mode.
(c) Ionisation state during the laser pulse
In order to estimate the electron number density generated by the ionisation during the laser pulse the recombination
processes should be taken into account. Electron recombination proceeds in dense plasma mainly by three-body Coulomb
collisions with one of the electrons acting as a third body (Zeldovich and Raizer, 2002). In this case the cross section for
the Coulomb collision reads
ei
(e
2
/
el
)
2
Z
2
, while the probability for a third body (electron) presence in the vicinity of
colliding particles is proportional to the cube of the Coulomb impact distance, p
3b
r
3
Coul
= (e
2
/
el
)
3
. The growth rate of
the electron number density is the balance of ionisation and recombination terms (Gamaly et al., 2006):
dn
e
dt
n
e
w
ion
e
n
i
n
2
e
; (6.32)
here the ionisation rate is w
ion
= max{w
mpi
, w
imp
} 10
15
s
1
, and the recombination rate (the number of recombination
events per second in a cm
3
) is
e
n
i
n
e
, where the coefficient
e
is expressed as (Zeldovich and Raizer, 2002):
e
= 8.75 10
27
ln Z
2
/
9/2
el
. (6.33)
We assumed that n
e
= Zn
i
, the electron energy
el
is in eV; ln is the Coulomb logarithm. One can see that ionisation
time, t
ion
w
1
ion
, and recombination time t
rec
(
e
n
i
n
e
)
1
, are of the same order of magnitude, 10
15
s, and both
are much shorter then the pulse duration. This is a clear indication of the ionisation equilibrium, and that the multiple
ionisations Z > 1 take place. Therefore, the electron number density at the end of the pulse can be estimated in a stationary
approximation as follows: n
2
e
w
ion
/Z
e
. Taking w
ion
10
15
s
1
;
e
30 eV; Z = 5; ln 2, one obtains, that number
density of electrons at the end of the pulse becomes comparable to the atomic number density n
e
10
23
cm
3
.
6.4.3. Increase in the absorbed energy density due to modification of optical properties
We demonstrated above that the swift ionisation during the first femtoseconds in the beginning of the pulse produces
the electron number density comparable to the critical density for the incident laser light, n
e
= n
c
. The free-electron number
density grows up and becomes comparable to the ion density to the end of the pulse. Respectively, the electronion collision
rate reaches its maximum that equals approximately to the plasma frequency in the dense non-ideal plasma. With further
increase of electron temperature the electronion exchange rates decrease due to domination of the Coulomb collisions. The
optical properties of this transient plasma are described by the Drude-like dielectric function; they are changing in accord
with the change in electron density and temperature. Let us estimate the absorption coefficient and absorption length in
the beginning of the laser pulse and at the end of the pulse. The dielectric function and the refractive index in the conditions
ei

pe
are estimated as:
re

2
2
pe
;
im

pe
; n k =
_
im
2
_
1/2
. (6.34)
For example, after the optical breakdown of fused silica glass by 800 nm laser at high laser intensity ( = 4.7
10
15
s
1
;
pe
= 1.45 10
16
s
1
) the real and imaginary parts of refractive index are n k = 1.18, thus giving the
absorption length of l
s
= 54 nm and the absorption coefficient A = 0.77 (Gamaly et al., 2006). Therefore, the optical
breakdown and further ionisation and heating all together convert silica into a metal-like medium. This decreases the energy
deposition volume by up to two orders of magnitude when compared with the focal volume, and correspondingly massively
increases the absorbed energy density and consequently the maximumpressure in the absorption region. For the interaction
parameters presented above (I = 10
14
W/cm
2
; A = 0.77; l
s
= 54 nm; t
p
= 150 fs) the pressure corresponding to the
absorbed energy density equals to 4.4 TPa, ten times higher than the Young modulus of sapphire, one of the hardest of
dielectrics. The general approach presented above is applicable for the estimations of any wide band gap dielectric affected
by high intensity short pulse laser.
6.4.4. Energy transfer from electrons to ions: relaxation processes after the pulse
The hydrodynamic motion starts after the electrons transfer the absorbed energy to the ions. The following processes are
responsible for the energy transfer from electrons to ions: recombination; electron-to-ion energy transfer in the Coulomb
collisions; ionaccelerationinthe field of charge separation; electronic heat conduction. Belowwe compare the characteristic
times of different relaxation processes.
(a) Impact ionisation and recombination
The electron temperature in the energy units at the end of the pulse is much higher then the ionisation potential.
Therefore, the ionisation by the electron impact continues after the pulse end. The evolution of the electron number density
can be calculated in the frame of the familiar approach (Zeldovich and Raizer, 2002):
dn
e
dt

e
n
e
n
a
e
n
i
n
2
e
; (6.35)
where
e
=
e
v
e
n
e
(I/T
e
+2) exp(I/T
e
)[cm
3
/s] is the impact ionisation rate and
e
is the recombination rate connected to
e
by the principle of detailed balance. One can see that for parameters of experiments in question (
e
210
16
cm
2
;
e

30 eV) the time for establishing the ionisation equilibrium is very short
eq
(
e
n
e
)
1
10
16
s. Thus the average
charge of multiple ionised ions can be estimated fromthe equilibriumconditions using Saha equations. Losses for ionisation
lead to temporary decrease in the electron temperature and in the total pressure (Gamaly et al., 2006). However the fast
recombination results in the increase in the ionic pressure.
(b) Electron-to-ion energy transfer by the Coulomb collisions
The Coulomb forces dominate the interactions between the charged particles in the dense plasma created by the
end of the pulse. The parameter characterising the plasma state is the number of particles in the Debye sphere, N
D
=
1.7 10
9
(T
3
e
/n
e
)
1/2
(Kruer, 1988). Plasma is in ideal state when N
D
1. In the plasma with parameters estimated above
for the fused silica (Z = 5, ln = 1.7; n
e
= 3 10
23
cm
3
;
e
= 50 eV) N
D
is of the order of unity, that is a clear
signature of the non-ideal conditions. The maximum value for the electronion momentum exchange rate in non-ideal
plasma approximately equals to the plasma frequency,
ei

pe
310
16
s
1
(Eidmann et al., 2000; Gamaly et al., 2002).
Hence electrons in ionised fused silica transfer the energy to ions over a time t
en
ei
(
en
ei
)
1
(M
i
/m
e
ei
) (12) ps
(Gamaly et al., 2006).
(c) Ion acceleration by the gradient of the electron pressure
Let us estimate the time for the energy transfer from electrons to ions under the action of electronic pressure gradient
when ions are initially cold. The Newton equation for ions reads:
M
i
n
i
u
i
t

P
e
x
.
The kinetic velocity of ions then estimates as follows:
u
i

P
e
M
i
n
i
x
t. (6.36)
The time for energy transfer from electrons to the ions is defined by condition that the ions kinetic energy compares to that
of the electrons, M
i
u
2
i
/2
e
. With the help of Eq. (6.36) one obtains the energy transfer time (Zn
i
n
e
):
t
elst

x
Z
_

e
2M
i
_
1/2
; (6.37)
here x l
abs
= 54 nmis the characteristic space scale. Then taking the time for the electron-to-ion energy transfer by the
action of the electrostatic field of charge separation equals to t
elst
1 ps.
(d) Electronic heat conduction
Energy transfer by non-linear electronic heat conduction starts immediately after the energy absorption. Therefore heat
wave propagates outside of the heated area before the shock wave emerges. The thermal diffusion coefficient is defined
conventionally as D = l
e
v
e
/3 = v
2
e
/3
ei
, where l
e
, v
e
and n
ei
are the electron mean free path, the electron velocity and
the momentum transfer rate respectively. The characteristic cooling time is conventionally defined as t
cool
= l
2
s
/D. For the
conditions of experiments in Juodkazis et al. (2006a,b,c): n
ei

pe
3 10
16
s
1
;
e
= 50 eV, and the cooling time is
t
cool
= 3
pe
m
e
l
2
abs
/2
e
= 14.9 ps.
Summing up the results of this section we shall note that the major processes responsible for the electron-to-ion energy
transfer in the dense plasma created by the tight focusing inside the bulk of a solid are different from those in the plume
created by laser ablation. The energy transfer from electrons to ions occurs after the end of the pulse in 1 ps. The thermal
ionisation and recombination are in equilibrium early in the pulse time; this permits the description of the plasma state by
the Saha equations. The electronic non-linear heat conduction becomes important in the first 15 ps after the pulse, and
dominates the return to the ambient conditions.
6.4.5. Shock-wave propagation, stopping, and void formation
We approximate (see Section 6.2) nearly cylindrical focal (and absorption) volume by a sphere with a radius of r
sph

(r
1/2
+z
1/2
)/2. For the focusing = 800 nminto fused silica glass (n = 1.453) with a microscope objective with NA = 1.35
it yields r
sph
= 0.33 m and a volume of 0.15 m
3
The original focal volume shrinks due to ionisation to a region with a
radius approximately 5 times less than the averaged focal radius. Modified shape of the absorption region has a complicated
shape, which is practically unknown. Therefore, it is reasonable to assume for the further calculations that the absorption
volume is a sphere of a smaller radius than the focal volume. One can see that 100 nJ of laser energy focused in the volume
of 0.15 m
3
create the energy density of 6.7 10
5
J/cm
3
equivalent to the pressure of 0.67 TPa (6.7 Mbar). However this
energy is absorbed in a much smaller volume thereby generating pressure in excess of 10 TPa. All the absorbed energy is
confined in the electron component at the end of the 150 fs pulse.
(a) Shock wave generation and propagation
The hydrodynamic motion starts after the electrons have transferred their energy to ions. This process is completed in
a few picoseconds time, however one should note that the energy transfer time through the Coulomb collisions increases
in proportion to the increase in the electron temperature. The higher is the absorbed energy, the longer is the time for the
hydrodynamic motion to begin. The pressure in a range of several TPa builds up after the electronion energy equilibration;
this pressure considerably exceeds the Young modulus for the majority of materials. For example, the Young modulus for
Fig. 6.6. Formation of void, shock and rarefaction waves propagation in laser-induced microexplosion. The scheme is the density distribution along the
radius of the energy deposition sphere (centre of symmetry is at r = 0).
sapphire equals to 0.30.4 TPa, and that for silica it is 0.01 TPa. The high pressure generates the shock wave, which
propagates from the energy absorption region into the surrounding cold material. The bulk modulus of the cold material
equals to the cold pressure, P
c
. This counter pressure finally decelerates and stops the shock wave. Because the shock wave
driving pressure significantly exceeds the cold pressure, P P
c
, the strong emerged shock compresses the solid to the limit
determined by the equation of state for the solid, which does not depend on the magnitude of the driving pressure. The
maximum density behind the shock in a perfect gas with the adiabatic constant is:
=
+1
1
0
e
. (6.38)
The adiabatic constant for a cold solid is conventionally estimated as 3 (Zeldovich and Raizer, 2002). Therefore, the
maximum density after the shock front is expected to be
max
= 2
0
. The compression ratio gradually decreases along
with the shock wave propagation, deceleration and transformation into a sound wave. Note, that the temperature in the
compressed solid behind the shock front in the limit of P P
c
grows in proportion to the increase of the driving pressure:
T =
_
1
+1
_
P
P
c
T
0
. (6.39)
The material is compressed and heated behind the shock wave front. Hence, the conditions for transformation to another
phase can be created, and this phase could be preserved after the unloading of pressure to the normal pressure. The final
state of matter may possess the properties different from those in the initial state. A scheme of formation the shock and
rarefaction waves in a spherically symmetric strong explosion is presented in Fig. 6.6.
(b) Shock wave propagation
The propagating shock wave loses its energy due to dissipation, e.g. due to the work performed against the internal
pressure (Young modulus) of the cold material which resists the material compression. The propagation distance at which
the shock wave effectively stops defines the shock-affected area. The shock wave converts into a sound wave at the stopping
point; the acoustic wave propagates further into the material without inducing any permanent changes in the solid. The
distance where the shock wave stops can be estimated from the condition that the internal energy in the whole volume
embraced by the shock wave is comparable to the absorbed laser energy: 4P
0
r
3
stop
/3 E
abs
(Zeldovich and Raizer, 2002).
The stopping distance obtained from this condition reads:
r
stop

_
3E
abs
4P
0
_
1/3
. (6.40)
In other words, at this point the pressure behind the shock front equals to the internal pressure of the cold material. One
can reasonably suggest that the sharp boundary observed between the amorphous (laser-affected) and crystalline (pristine)
sapphire in the experiments corresponds to the distance where the shock wave was effectively stopped (Juodkazis et al.,
2006a,b,c; Gamaly et al., 2006). The sound wave continues to propagate at r > r
stop
apparently not affecting the properties
of material. For 100 nJ of the absorbed energy and taking P
c
= 0.4 TPa for sapphire, one gets r
stop
= 180 nm, with the help
of Eq. (6.40), which is in qualitative agreement with the experimental values.
(c) Rarefaction wave: formation of void
The experimentally observed formation of void, which is a low-density region surrounded by a shell of the laser-affected
material, can be qualitatively understood from the following simple reasonings. Let us consider, for simplicity, spherically
symmetric explosion. The strong spherical shock wave propagates outside the absorbed energy region, and compresses the
surrounding material. At the same time, a rarefaction wave propagates towards the centre of symmetry, and decreases the
density in the area of the energy deposition. This problem qualitatively resembles the familiar hydrodynamic phenomenon
of strong point explosion (P P
0
) in homogeneous atmosphere with counter pressure taken into account. The material
density decreases rapidly in space and in time behind the shock front in the direction to the centre of symmetry; this is the
characteristic of a strong spherical explosion. The temperature increases in the energy deposition region inside a sphere of
radius r l
abs
, and the density decreases towards the centre of symmetry, while the pressure is almost constant along the
radius (Zeldovich and Raizer, 2002). After the explosion, almost the entire mass of material, which was initially distributed
uniformly inside the sphere of radius r, is concentrated in a thin shell surrounding the area of energy deposition. This picture
is qualitatively similar to that observed in the experiments as a result of fs-laser pulses tightly focused inside the sapphire,
silica glass, or polystyrene (Juodkazis et al., 2006a,b,c). A void surrounded by a shell of the laser-modified material was
formed at the focal spot. Following the strong point explosion model, one can suggest that the whole heated mass in the
energy deposition region was expelled out of the centre of symmetry and was frozen after the shock wave unloading in the
form of a shell surrounding the void.
One can apply the law of mass conservation to estimate the density of compressed material from the size of the void.
Indeed, the mass conservation relates the size of the void to the compression of the surrounding shell. We assume that no
mass losses could occur in the conditions of confinement. One can use the void size and size of the laser-affected material
measured in experiments to deduce the compression of the surrounding material. The void formation inside a solid only
possible if the mass initially contained in the volume of the void, was pushed out and compressed. Thus after the micro-
explosion, the whole mass initially confined in a volume with of radius r
stop
resides in a layer in between r
stop
and r
v
, which
has a density =
0
; > 1:
4
3
r
3
stop
0
=
4
3
(r
3
stop
r
3
void
). (6.41)
Now, the compression ratio can be expressed through the experimentally measured void radius, r
void
, and the radius of
laser-affected zone, r
stop
, as follows:
r
void
= r
stop
(1
1
)
1/3
. (6.42)
It was typically observed that r
void
0.5r
stop
in the experiments of Juodkazis et al. (2006a,b,c). By applying Eq. (6.42), one
obtains that compressed material in the shell has in average the density 1.14 times higher than that of a crystalline sapphire.
Note that the void size was measured at the room temperature long after the microexplosion. The schematic picture of the
initial stage for the shock and rarefaction wave formation and propagation during the spherical explosion is presented in
Fig. 6.6.
6.4.6. Computer modelling of a confined micro-explosion
Plasma created in a focal volume by the rapid laser energy deposition attains the local thermodynamic equilibrium in
time scale of the order of a picosecond. Therefore, the process of plasma expansion into a cold solid on the larger time scales
can be described in the frames of high temperature hydrodynamics. Two-temperature hydrodynamic calculations for micro-
explosion in silica glass were performed in the spherical geometry with the code Chivas (Gamaly et al., 2006). The radiation
hydrodynamics code Chivas was designed for numerical simulations of the laserplasma interaction and the target
compression for the inertial confinement fusion; it has been used for the numerical simulations of the micro-explosion. The
code Chivas (Jacquemot andDecoster, 1991) is a one dimensional, two-temperature (electrons andions) hydrodynamic code,
which accounts for the electron and ion thermal transport, electronion coupling, and transient ionisation (see Appendix E).
The ionisation states and the opacity data were calculated assuming a local thermodynamic equilibrium. The equation of
state (EOS) implemented inthe code, QEOS, (quotidianequationof state) is described inMore et al. (1988). Three parameters,
namely, the mass density 2.2 g/cm
3
, the bulk modulus Y = 75 GPa, and the binding energy 3.16 J/mol (3.29 eV/atom),
define the EOS for glass. The main aims of the calculations were to reproduce the experimental observations: the absence
of a void at low energy, the energy threshold for the void formation, and the dependence of the void size on the deposited
energy. The equivalent sphere approximates the cylindrical region, where the energy is deposited, with sufficient accuracy.
The absorbed laser energy in a range from one nanoJoule (10
9
J) to 100 nJ (10
7
J) was deposited homogeneously and
instantaneously in the spherical volume of a radius r
dep
= 0.13 m.
The calculations at lowdeposited energy of one nJ have shown that the pressure in absorption region was comparable to
the Young modulus and the void could not be formed. The instantaneous isochoric heating at the absorbed energy of 50 nJ
produced the average ionisation Z = 4.83 and the pressure of 2.6 TPa, which exceeded the bulk modulus more than 500
times. It is in good agreement with the simple estimate made above. The strong shock wave emerged at the outer surface
of the energy deposition sphere and compressed the material to the density twice of the initial density. Then, the pressure
behind the shock front rapidly decreased with distance, finally the shock transformed into the acoustic wave at 50 ps. The
Fig. 6.7. (a) Spatial profiles of pressure, temperature, and mass density at 2 ps and 3 ps after the laser-induced micro-explosion in sapphire with 50 nJ,
150 fs, 800 nm pulses deposited into a spherical area with radius 0.13 m (Juodkazis et al., 2006a,b,c). The initial density
0
= 3.89 g/cm
3
and Young
modulus Y = 0.4 TPa of sapphire are shownby the dashedlines. (b) 3D-presentationof the initial stage of the hydrodynamic motioninsilica glass irradiated
with 100 nJ, 150 fs, 800 nm pulses absorbed in 0.1 m
3
volume at the time moment of 1 ps;
0
= 2.1 g/cm
3
. Calculations were performed by code Chivas
(Jacquemot and Decoster, 1991).
spatial density profiles for the time moments up to 0.9 ns are shown in Fig. 6.7. The compression ratio at 1 ns after the micro-
explosion reached its asymptotic value of = (r/r
0
) = 1.1, which qualitatively complies with the density of amorphous
layer retrieved fromthe experiments,
exp
= 1.14 (Juodkazis et al., 2006a,b,c; Gamaly et al., 2006). However, the gas density
in the central void region at 1 ns time was rather high, above 0.1 g/cm
3
.
Summing up, one can conclude that the hydrodynamic calculations can predict the evolution of the laser-affected
solid during the period up to 1 ns, the calculation results are in qualitative agreement with the experimental data. The
hydrodynamic simulations in the present model are valid during several hundreds of picoseconds after the pulse, up until
the material motion ends. However, the process of transformation of the material to the denser phases and the processes
responsible for transition to the steady state at room temperature are still unknown.
6.4.7. Density and temperature in the shock- and heat-affected solid
There are two distinctive regions in the area affected by the laser action, which is confined inside the cold solid. First is the
area where the laser energy is absorbed. Second area is the zone where the shock wave propagates outside the absorption
zone, compresses and heats the initially cold solid, and then decelerates into a sound wave. In the first area, the solid is
heated to the temperature of tens of eV (10
5
K). It is swiftly ionised at the density close to that of a cold solid. The energy
is transferred to the ions within a picosecond time, and the shock wave emerges. Conservatively, the heating rate can be
estimated as 10 eV/ps = 10
17
K/s. The material undergoes fast compression under the action of the micro-explosion and
its density may increase to 2 times the initial solid density. The energy dissipation by the shock wave and by the heat
conduction takes about a nanosecond. The maximum temperature reaches several thousands Kelvin in the second zone,
which is heated and compressed exclusively by the shock wave; this temperature is approximately ten times less than in
the first zone. The heating and cooling rates are of the order of 10
14
K/s in this zone.
Phase transformations in quartz, silica and glasses induced by strong shock waves have been studied for decades
(see Zeldovich and Raizer, 2002; Luo et al., 2003), and references therein. However, all these studies were performed in
one-dimensional (plane) open geometry in the conditions when the shock unloading into the air was always presented.
The pressure ranges for different phase transitions to occur under the shock wave loading and unloading have been
establishedexperimentally andwell understoodtheoretically (Luo et al., 2003). Quartz andsilica converts to a dense phase of
stishovite (mass density 4.29 g/cm
3
) in the pressure range between 1546 GPa. The stishovite phase exists up to a pressure
77110 GPa. Silica and stishovite melts at a pressure above 110 GPa, which is in excess of the shear modulus for liquid silica
10 GPa. The dense phases usually transformed into the low-density phases (2.292.14 g/cm
3
) when the pressure releases
back to the ambient level. There are numerous observations of amorphisation upon compression and decompression. An
amorphous phase of silica, which is denser than the initial state sometimes forms when the unloading occurs in the range
1546 GPa. Analysis of experiments shows that the pressure release and the reverse phase transition follows an isentropic
path.
In the studies of shock compression and decompression under the action of shock waves induced by explosives, the
loading and release time scales are in the order of 110 ns. The heating rate in the explosive experiments is 10
3
K/ns =
10
12
K/s, i.e. five orders of magnitude slower than in the confined micro-explosion.
On the contrast, the peak pressure at the front of the shock wave driven by laser in the confined geometry, reaches the
level of TPa, that is, 100 times in excess of the pressure value required to induce structural phase changes and melting.
Therefore, the region where the melting occurs is located very close to the area where the energy is deposited. The zones
where the structural changes and amorphisation might occur are located further away. Super-cooling of the transient dense
phases may occur if the quenching time is sufficiently short. The very fast heating and cooling along with the small size of
the area where the phase transition takes place can affect the rate of the direct and reverse phase transitions. In fact, the
studies of phase transitions in these space and time scale domains are very limited.
The refractive index changes in a range of 0.050.45, along with the protrusions surrounding the central void, as a result
of laser-inducedmicro-explosionwere observedinthe bulk of silica (Glezer et al., 1996). This is the evidence of the formation
of a denser phase during the fast laser compression and quenching; however, little is known of the exact nature of the phase.
Thus, we can conclude that a probable state of a laser-affected glass between the void and the shock stopping distance may
contain the amorphous or micro-crystallite material (Glezer et al., 1996), which is denser, and with larger refractive index
than the initial glass.
6.4.8. Upper limit for the pressure achievable in confined interaction
The laser-induced micro-explosion can be considered as a confined one if the shock wave affected zone is separated
from the surface by the layer of pristine material m-times thicker than the size of this zone. On the other side, the thickness
of this layer should be equal to the distance at which the laser beam propagates without self-focusing L
sf
(W/W
c
)see
Eq. (6.4); W is the laser power, W
c
is the critical power for self-focusing from Eq. (6.3). This condition expresses as follows:
L
sf
= mr
stop
. (6.43)
The absorbed energy, E
abs
= AE
las
, expressed through the laser power W = E
las
/t
p
(E
las
, t
p
are respectively energy per pulse
and pulse duration), and the radius of shock wave affected zone are connected by the equation:
r
stop

_
3AWt
p
4P
cold
_
1/3
. (6.44)
The condition of Eq. (6.43) with Eq. (6.4) for the self-focusing length inserted then turns to the equation for the maximum
laser power at which micro-explosion remains confined and self-focusing does not affect the crystal between the laser
affected zone and outer boundary:
2n
0
r
2
0
m
_
4P
cold
3AW
c
t
p
_
1/3
=
_
W
W
c
_
1/3
_
W
W
c
1
_
1/2
. (6.45)
The left-hand side in Eq. (6.45) can be calculated if the laser pulse duration and the focal area are both known. Taking, for
example, sapphire (n
0
= 1.75); W
c
= 1.94 MW; = 800 nm; 4P
cold
/3 = 1.67 MJ; t
p
= 100 fs; r
2
0
= 0.2 m
2
; m = 3)
one gets 0.6 inthe LHS of Eq. (6.45). Thus the maximumlaser power allowedinthese conditions equals to 1.3W
c
= 2.5 MW
or 250nJ of the energy in100 fs laser pulse. The maximumpressure whichcanbe achievedinthe absorptionvolume confined
inside the transparent crystal in the conditions considered above can be up to 27 TPa, approximately three times higher than
it was achieved in the experiments (Juodkazis et al., 2006a,b,c).
6.4.9. The propagation of an optical breakdown wave from the laser-absorption zone inside a solid
There is an additional effect, which can affect the size of the absorbing volume, namely, the expansion of the surface
where the electron density equals to the critical density. The optical breakdown threshold, F
thr
, is achieved at the beginning
of the pulse in a time period defined by the following condition:
F
a
t
thr
f (t)dt = F
thr
; (6.46)
here F
a
= E
a
/r
2
f
is the pulse energy per the focal spot area of r
2
f
(the total absorbed laser fluence). After the threshold
the time t
ion
= w
1
ion
is needed to reach the electron density equal to the critical density; here w
ion
is the total ionisation rate
by all ionisation mechanisms. The laser pulse is characterised by the energy per pulse, E
p
, pulse duration, t
p
, focal spot area
and radius, S
f
, r
f
, and normalised
_
+
f (t)dt = 1 temporal shape of the pulse f (t). The threshold condition is fulfilled at
the beam cross section larger than the focal spot when the laser fluence is well in excess of the breakdown threshold. Thus
the wave of breakdown starts to move in the direction towards the laser beam at the time period t t
thr
+t
ion
. The optical
breakdown wave front is a surface where the electron density attains the critical density value. The velocity of the front is
obtained from the condition that the breakdown condition at the broader part of the beam is fulfilled when the fluence in
this cross section increases in accord with the laser pulse temporal shape:
E
a
_
t
f (t
)dt
r
2
(x, t)
_
f (t
)dt
= F
thr
; (6.47)
here r
2
(x, t) is the cross section of the conical laser beam at the distance x from the focus (r(x, t) > r
f
) in the direction
towards to the beam propagation at the moment t. The radius of this cross section in a conical beam is as follows:
r(x, t) = r
f
+xtg, (6.48)
here is the half of the focal cone angle. Then equation for obtaining the velocity of the breakdown wave can be obtained
by combining Eqs. (6.47) and (6.48) (Raizer, 1978):
_
t
f (t
)dt
f (t
)dt
=
F
thr
F
a
_
1 +
x
r
f
tg
_
2
. (6.49)
Temporal pulse shape, in general, is close to the Gaussian one:
f (t) =
_
ln 2
_
1/2
_
2
t
p
_
exp
_
4t
2
ln 2
t
2
p
_
.
The Eq. (6.49) then can be solved numerically. For the practical estimates, however, it is useful to derive a simple scaling
by combining all the major parameters. We approximate the Gaussian profile by a triangle of the same area:
f (t) = 1 +t/t
p
at t < 0
f (t) = 1 t/t
p
at t > 0.
(6.50)
Then at the first half of the pulse t
p
< t < 0 Eq. (6.49) expresses as:
(t +t
p
)
2
2t
2
p
=
F
thr
F
a
_
1 +
x
r
f
tg
_
2
. (6.51)
Taking the square root and differentiating by time one gets the velocity of the breakdown wave front as follows:
dx
dt
=
r
f
t
p
tg
_
F
a
2F
thr
_
1/2
. (6.52)
The time for absorbing energy equal to the breakdown threshold is:
t
thr
=
F
thr
F
a
t
p
. (6.53)
Thus, the maximum distance travelled by the breakdown wave to the end of the pulse reads:
x
max
=
dx
dt
(t
p
t
thr
t
ion
) =
r
f
t
p
tg
_
F
a
2F
thr
_
1/2
_
t
p
t
tp
F
thr
F
a
t
ion
_
. (6.54)
One can get an approximate estimate for the length of the wave front propagation by the end of the pulse as x
max

(F
a
/2F
thr
)
1/2
r
f
assuming tg > 1 (focusing with high numerical aperture lens and 45). Even for (F
a
/F
thr
)5 for a
100 fs, 100 nJ pulse, and S
f
0.7 m
2
, (F
a
15 J/cm
2
) the elongation of the energy absorption region along the beam axis
approximately equals to the focal spot radius. This effect should be taken into account for the estimates of the maximum
absorbed energy density.
6.4.10. Similarity laws of hydrodynamics: can we model macroscopic explosions in tabletop conditions?
The micro-explosion can be described solely within the frames of the ideal hydrodynamics if the heat conduction and
other dissipative processes, all characterised by the specific length scales, could be ignored. The hydrodynamic equations
contain five variables: the pressure P, the velocity v, the density , the distance r, and the time t. The last three of them
are independent parameters, and the first two can be expressed through the other three. The micro-explosion can be fully
characterised by the following independent parameters: the radius of the energy deposition zone R
0
, the total absorbed
energy E
0
, and the initial material density
0
. Then, the initial pressure P
0
= E
0
/R
3
0
, and the initial velocity v
0
= (P
0
/
0
)
1/2
of the material are combinations of the independent parameters. One can neglect the energy deposition time and the time
for the energy transfer fromthe electrons to the ions, which is in picosecond range, in comparison to the hydrodynamic time
of a fewnanoseconds. Then, the hydrodynamic equations can be reduced to a set of the ordinary equations with one variable
= r/v
0
t, which describes similar hydrodynamic phenomena (Zeldovich and Raizer, 2002). With the increase of the energy
of explosion, the space sale R
0
, and time scales are increased accordingly to R
0
= (E
0
/p
0
)
1/3
, and; t
0
= R
0
/v
0
. The similarity
between the laws of hydrodynamics in micro- and macroscopic explosion suggests that the micro-explosion in sapphire
(E
0
= 10
7
J;
0
4 g/cm
3
; R
0
= 1.5 10
5
cm; t
0
= 5.5 10
12
s) is a reduced version of a macroscopic explosion,
generating the same final pressure. For example, the explosion energy of 10
14
J, which is equivalent to 25 kilotons of high
explosive or a nuclear bomb released in a volume of 4 m
3
(R
0
= 1.59 m) during 20 s, exerts the same pressure of 12.5 TPa
as the laser-induced micro-explosion in sapphire in the conditions of E
0
= 10
7
J;
0
4 g/cm
3
; R
0
= 1.510
5
cm; t
0
=
5.5 10
12
s. Thus, exactly the same physical phenomena occur at the scale of 10
7
times smaller volume, 10
7
times faster,
and 10
21
times lower energy. Therefore, all major hydrodynamic aspects of a powerful macroscopic explosion in confined
conditions can be reproduced in the laboratory tabletop experiments using ultra-short laser pulses.
Fig. 6.8. (a) SEM image of a cleaved sapphire crystal showing a cross-section of the shock affected area with void in the middle; the laser irradiation
direction is shown by the arrow. (b) similar cross-section where the amorphous shock-wave affected area was etched out. The inset (c) shows AA
section
in transverse plain (in respect to laser beam propagation) across the void opened by fast ion beam (FIB) milling technique.
Source: Taken from Juodkazis et al. (2006a,b,c).
6.4.11. Experimental observation of void formation in different materials
(a) Laser focusing parameters
Formation of voids by ultrafast laser pulses has been studied in various dielectrics: in crystalline sapphire, amorphous
silica glass (viosil), and polystyrene (Juodkazis et al., 2006a,b; Gamaly et al., 2006). The choice of materials with different
structural, mechanical, and optical properties was made in order to test the basic principles of nano-void formation. Single
laser pulses (200 fs, 800 nm) were tightly focused inside a sample using an optical microscope (Olympus IX70) equipped
with an oil-immersion objective lens of numerical aperture NA = 1.35. The focal volume was approximately 0.15 m
3
(slightly varied due to refractive index differences), the beam waist radius was r
1/2
= 0.26 m; and the focal area was
r
2
1/2
= 0.21 m
2
. The average intensity in the focal area was up to 5 10
14
W/cm
2
for a 100 fs, 100 nJ pulse well over
the breakdown threshold. The peak laser power of 0.5 MW for a 100 nJ pulse was lower than the threshold for self-focusing
in sapphire (P
cr
= 1.94 MW) and silica (1.98 MW). Therefore, the laser energy could be delivered to the focal volume
located from 5 m up to 50 m below the crystal surface without inducing the self-focusing effect. The critical power of
self-focusing in polystyrene is several orders lower than that in sapphire and silica. For this reason the effect of self-focusing
in polystyrene revealed itself through the formation of strongly elongated voids. The refractive index mismatch between the
immersion oil (n = 1.52), and the dielectrics used in experimentssapphire (n = 1.75), silica (n = 1.45), and polystyrene
(n = 1.55) had caused strong dependences of the focal intensity distribution on the focused depth, especially in the case of
sapphire.
(b) Voids in sapphire
Experimental studies of voids in sapphire were performed in the following way. First, an array of laser-affected regions
was produced inside the sapphire crystal by a succession of single 100 nJ, 200 fs, 800 nm laser pulses focused with an
objective lens of NA = 1.35. The lateral and axial extent of laser-affected region then was carefully examined. Second, for
the axial cross section the sample was cleaved along the c-axis of sapphire and examined by a scanning electron microscope
(SEM)see Fig. 6.8. For observation in the transverse plane, a focused ion beam(FIB) milling was used to open the void at the
largest cross-section; Fig. 6.8(c) shows the presence of void at the focal region. Careful examination revealed that the laser-
affected regions consisted of voids surrounded by a shell extended to about twice the void diameter. This shell was identified
as amorphous sapphire by chemical etching, since amorphous sapphire has a much higher solubility in hydrofluoric acid
when compared with crystalline material (Juodkazis et al., 2005). Indeed, the shell was completely etched away by a 10%
aqueous solution of HF revealing a smooth boundary between the pristine sapphire crystal and the shock-affected zone
(Fig. 6.8(b)). This allowed the exact size of this zone to be measured.
The diameter of the laser-affected region was measured at the same focusing conditions and presented as a function
of increasing pulse energy. The dependence of the diameter on the laser energy is plotted in Fig. 6.9. The central void was
absent whenthe laser energy was belowthe thresholdof 35 nJ per pulse. The size of laser-affectedregionunder the threshold
equals to 250300 nm. The laser-affected material in this zone most probably is amorphous and it has lower density than
the initial crystal. The shape of the amorphous region becomes elliptical for higher pulse energies. The elliptical shape is an
indication of the influence of spherical aberrations, as well as a result of optical breakdown wave propagation in direction
towards the laser beam. The appearance of cracks surrounding the laser-affected region of the size exceeding 600 nm was
observed at the pulse energy above 170 nJ. One may suggest that ellipticity of the affected zone implies the anisotropy of
the pressure distribution in the surrounding area, and the pressure and stress gradients cause cracking in the crystal.
Fig. 6.9. Lateral cross sections of the void (curve 1, circles) and amorphous region (curve 2, squares) in sapphire versus the pulse energy at focus.
Experiments were carried out at 30 m depth below the surface. The grey region defines the domain affected by the shock wave.
(c) Voids in polystyrene and in glass
The nano-void formation in the silica glass (viosil) and polystyrene were also studied in the same irradiation conditions
(Juodkazis et al., 2006a). Viosil has similar band gap to sapphire, both materials are transparent down to 150 nm. The Young
modulus of viosil is Y
viosil
= 75 GPa, much lower than for sapphire Y
sapph
= 375 GPa. SEM examination of the irradiated
region confirmed the formation of nanovoid for the pulse energies above 30 nJ. The same wet etching procedure as for
sapphire could not reveal a clear boundary between laser affected and unaffected regions in silica.
In polystyrene the voids were observed at 20 5 nJ. Their shape was considerably elongated in direction of the beam
axis due to a strong self-focusing effect. Thus, it was confirmed experimentally that voids could be produced in crystalline,
glass, and polymer materials. Obviously, the final size, morphology, and shape of void and the surrounding material, all are
affected by the shock wave, and relaxation and de-excitation of the material. However, the main parameters of the voids
are qualitatively, and in many cases quantitatively, well described by the physics of laser-triggered micro-explosion.
6.4.12. Multiple-pulse interaction: energy accumulation
In order to produce an optical breakdown inside a transparent dielectric by a single tightly focused 100 fs short laser
pulse, the energy in excess of 2030 nJ per pulse has to be delivered to the sub-micron focal volume. However, the low
heat conduction of dielectrics implies an additional way to induce structural and optical modifications by hitting the same
place by many consecutive low energy laser pulses. The low heat diffusion 10
3
cm
2
/s results in a long cooling time over
10 s of a micron size region. Therefore, the laser energy accumulates in the absorption region when the period between
the pulses is shorter than the cooling time. The energy accumulates within the space size less than the heat diffusion length
where the material can be heated to a very high temperature. The accumulation effect was observed experimentally in the
bulk of glass (Schaffer et al., 2003). Voids were produced by 25 MHz, 30 fs, 5 nJ, 800 nmlaser pulses in the bulk of a Zinc doped
borosilicate glass (Cornig 0211). The radius of these voids was measured using interference contrast optical microscopy as
a function of number of pulses. One should note that multiple pulses focused in the same spot produce structures with
unpredictable and complicated shape. The number of pulses heating the same spot was varied from 10
2
to 10
5
. The focal
volume was estimated to be V
foc
0.3 m
3
(r
foc
= 0.4 m). The single pulse energy density was insufficient for formation
of void in this case. The dependence of the void size on the number of pulses heating the same spot was determined in these
experiments; see Fig. 6.10.
Using the approach developed in the previous sections, one can estimate the focal volume temperature of 0.3 eV at
the end of a single 5 nJ pulse. After several tens of consecutive pulses the absorbed and accumulated energy is enough for
ionisation of the material and, therefore, for the change in the laserplasma interaction mode. The heat conduction also
changes from a linear process with respect to temperature to a non-linear process. The molecular bonds in the material
are broken after several pulses, the atoms are ionised, and therefore we can apply the plasma-based approach to find the
affected volume parameters. However, the multiple pulse interaction with the same focal spot confined inside the bulk of
a solid is very sophisticated three-dimensional phenomena with complicated time history, much more complex that of a
single pulse interaction. Fortunately, this phenomenon is well restricted in space and thus the energy conservation sets clear
limits, allowing us to derive a general semi-quantitative dependence to fit the experimental results.
Let us first estimate the size of the heat affected region after N successive pulses heating the same spot. A single pulse
delivers the energy E
1
, which transferred to a distance r
1,th
t
1/2
1
by linear heat conduction during the period t
1
between
the pulses. This distance increases after N consecutive pulses with the increase of the number of pulses as r
N,th
N
1/2
.
However, this dependence is much stronger than that experimentally observedsee Fig. 6.10. Thus, it seems unlikely that
the heat conduction alone is responsible for the void size produced by many pulses.
The shock wave affected zone produced by single pulse grows up with the energy per pulse as r
1,sh
E
1/3
1
(see
Section 6.4.5(b)). Therefore, it is reasonable to suggest that the shock-affected zone produced by consecutive action of N
Fig. 6.10. Radius of the structure in the bulk of borosilicate glass produced with a train of 30 fs, 5 nJ pulses incident at 25 MHz repetition rate and
focused in the bulk of glass by a 1.4 NA microscope objective, as a function of number of laser pulses per spot; a solid linefitting the data by Eq. (6.55):
r
N,void

= 0.67N
1/4
m.
Source: Trianglesexperimental points from Schaffer et al. (2003).
pulses grows up with the number of pulses as r
N,sh
N
1/3
. This dependence is weaker than produced by heat conduction
but it is still stronger than that observed. Thus, this mechanism should also be dismissed.
Let us now apply the simplest model of the adiabatic expansion of the region heated by the N successive pulses to the
energy NE
1
(pressure P
N
= NE
1
/V) and then expanded adiabatically to the cold pressure P
cold
of the surrounding material:
P
N
V
= P
cold
V
N
. The final radius of a void (r
N,void
V
1/3
N
) after action of N pulses reads:
r
N,void
N
1/3
. (6.55)
Here is the adiabatic constant. The Eq. (6.55) transformed into dimensional form as r
N,void

= 0.67N
1/4
m with = 4/3,
which fits the experimental data of Schaffer et al. (2003) well up to 10
4
pulses per spot (Fig. 6.10). Thus, the energy
conservation and isentropic expansion allows one to apply a semi-quantitative description to the results of experiments.
6.5. Laser-induced forward transfer
The major part of the absorbed laser energy in conventional laser-surface-ablation mode is in the kinetic energy of
expanding overheated plume. The temperature has a maximum at the surface decreasing exponentially with depth into
the solid. The ablated target material expands in the direction towards the incident beam. In the confined lasermatter
interaction, in contrast, the absorbed energy is mainly the internal energy of the material while the kinetic energy part
is minimal. One can present another experimental geometry for the confined interaction: the laser beam focused to the
opaque layer to be removed through a transparent layer of unaffected materialsee Fig. 6.1. The transparent layer acts as
a shield preventing the expansion of the material in the direction opposite to beam. In this case the optimum thickness of
the transparent layer has to fulfil several conditions discussed later in this section. A thin layer of material (a black layer in
Fig. 6.11) to be removed by the laser action is deposited on top of the transparent sample. The pressure, which builds up in
the laser absorbing film, induces material expansion in the direction of the incident laser beam propagation (Fig. 6.11). The
interaction mode and the following material expansion in the direction of beam propagation was coined as Laser-Induced
Forward Transfer (LIFT) (Bohandy et al., 1988; Toth et al., 1993; Pimenov et al., 1995; Zegrioti et al., 1998). The LIFT technique
has many potential application is laser deposition of 2-D multilayers and 3-D structures consisting of inorganic, polymeric,
nanoparticle-based and biological materials onto virtually any substrate. The techniques hold promise to make significant
contributions across many scientific disciplines ranging from bio-tissue engineering to electronics (Chrisey et al., 2003;
Arnold et al., 2007; Kattamis et al., 2007).
The laser energy is absorbed in the skin layer and then transported to the cold parts of a target by means of heat
conduction and shock wave propagation. The hydrodynamic motion of a material starts when the absorbed laser energy
is transferred from the electrons to the lattice (ions). Hydrodynamic expansion of the laser-heated layer into a space is only
possible in the direction of the beam propagation, see Fig. 6.11, if the pressure in the heated layer is in excess of the internal
strength of a material (i.e. the bulk modulus). Therefore, the amount, velocity and spatial distribution of forward-transferred
material depends on the combination of the laser pulse duration, pulse energy, focusing conditions, thickness and properties
of the target layer. In what follows, we concentrate on the material removal by sub-picosecond pulses with duration pulse
t
p
shorter than the major relaxation times.
6.5.1. Threshold conditions
Let us consider expansion of a metal film in the direction of the beam propagation in the condition when a fs-laser
pulse heats a film through the transparent shield. Certain conditions combining the laser and material parameters should
Fig. 6.11. A general scheme for the Laser-Induced Forward Transfer (LIFT) process. Formation of a plane shock wave and its transformation into a spherical
wave with time is shown on the left scheme; the spatial distribution of density, temperature and pressure in the ablated target vapour and in the air are
schematically presented in the right graphs.
be fulfilled in order to remove the material from the substrate. The problem can be treated as one-dimensional at the initial
stage close to the threshold. A layer next to the solidair interface in Fig. 6.11 starts to expand if the pressure in this layer
exceeds the internal strength of this material (the Young modulus or the cold pressure). The energy density created by the
laser near this interface should be equal or larger than the cold pressure. This is the condition for a threshold of material
ablation and expansion in the direction of the laser beam in LIFT configuration of Fig. 6.11. To quantify this condition, we
need to find the threshold laser fluence as a function of the laser and material parameters. The lasermetal interaction
proceeds in the regime of normal skin effect. Thus, the absorbed energy density, electron and lattice temperature, all decay
exponentially with depth inside the laser-absorbing layer.
We present the major processes in a sequence. Immediately after the end of the pulse t t
p
all the absorbed energy
(energy density per unit area in 1D problem) is confined in the electron component,
_
0
C
e
n
e
T
e
(x)dx = AF, as the pulse is
shorter thanthe electron-to-ionenergy transfer time. The focal spot radius, typically 10
3
cm, is muchlarger the skindepth
in metals, which is (23) 10
6
cm. Therefore, it is legitimate to treat the expansion problem as a one-dimensional in
space. By the same reason, the lateral heat losses are negligible. The electrontemperature inthe skinlayer after the end of the
laser pulse has a conventional form: C
e
n
e
T
e
(x) = (2AF/l
s
) exp(2x/l
s
). It is reasonable to assume that the electron number
density in the skin layer does not change during the interaction. The electron to lattice energy transfer time is shorter than
the heat conduction and hydrodynamic times. Therefore, the spatial dependence of temperature after the electronlattice
energy equilibration, T
e
= T
L
= T, also remains exponential in the skin layer:
(C
e
n
e
+C
L
n
a
)T(x) = (2AF/l
s
) exp(2x/l
s
). (6.56)
The temperature has a minimum at the outer boundary of the absorbing layer (at the sampleair interface) at the end of
the pulse. Spatial smoothing of temperature distribution along the layer by the heat wave or shock wave leads to the increase
in temperature and in pressure at the solidvacuum interface. The temperature at the interface reaches its maximum when
the temperature equilibrates along the layer. The average temperature over the layer with thickness d expresses in the
following form:
T =
1
d
d
0
T(x)dx.
By averaging the exponential distribution of Eq. (6.56) one gets:
T = T
L,m
l
s
2d
_
1 exp
_
2d
l
s
__
; T
L,m
=
2AF
(C
e
n
e
+C
L
n
a
)l
s
. (6.57)
Now, the condition for the threshold pressure at the outer boundary of the laser-heated layer, which causes the material to
move, can be expressed as:
(C
L
n
a
+C
e
n
e
)
T(d) P
cold
. (6.58)
The threshold laser fluence to be delivered into a layer of thickness d to initiate the expansion of the outer boundary of the
layer then reads:
F
thr
=
dP
cold
A[1 exp(2d/l
s
)]
. (6.59)
The function d[1exp(2d/l
s
)]
1
gradually increases with the growth of d. Therefore, the minimumvalue of the threshold
fluence is approximately at d l
s
/2. It is instructive to compare the threshold for the LIFT ablation with the conventional
threshold for the surface ablation with ultrashort laser pulses (Gamaly et al., 2002). We present here the threshold for the
surface ablation in a general form:
F
(surf)
th
=
l
s
C
e
n
e
(
b
+
esc
)
2A
. (6.60a)
Correspondingly, the LIFT threshold from Eq. (6.59) for d = l
s
/2 takes a form:
F
(LIFT)
thr
=
l
s
P
cold
2A(1 e
1
)
. (6.60b)
Then the ratio between the thresholds is:
F
(surf)
th
F
(LIFT)
thr
= 0.63
C
e
n
e
(
b
+
esc
)
P
cold
. (6.60c)
Let us present some experimental examples. The Young modulus of gold, for example, equals to 78 GPa (
b
=
3.37 eV/atom,
esc
= 5.1 eV; P
cold
c
2
0
= 79.5 GPa; = 19.3 g/cm
3
; c
0
= 2.03 km/s; n
a
= 5.910
22
cm
3
). Taking the
same parameters as for calculation of the electrostatic ablation of gold one obtains that the threshold for the LIFT-ablation
of half-skin-depth thickness is approximately the same, 0.5 J/cm
2
, as that for the surface ablation which. Similar estimates
for copper (Young modulus 128 GPa;
b
= 3.125 eV/atom,
esc
= 4.65 eV/atom, n
a
= 0.845 10
23
cm
3
; P
cold
c
2
0
=
129 GPa; = 8.96 g/cm
3
; c
0
= 3.8 km/s) give the ratio F
(surf)
th
/F
(LIFT)
thr
= 0.776. The surface threshold for Copper ablated
by 780 nm laser comprises 0.50.6 J/cm
2
. Therefore, the LIFT-threshold for a half-skin-depth layer can be estimates as
0.4 J/cm
2
.
This conclusion should not be surprising. In surface ablation, the outermost atomic layer where the electron temperature
is a maximum can be removed at the electron energy almost two times larger than the binding energy. In the LIFT-ablation,
the total pressure (electrons plus ions) in the whole layer should be elevated above the cold pressure in order to move
it. It is worth noting that the cold pressure corresponds to the average energy per atom, which should be equal to the
binding energy. However, the strength of a real material (and therefore its cold pressure) strongly depends on the presence
of imperfections such as defects, cracks, impurities etc. It could be several times lower than the above presented values
of the Young modulus for pure materials. Therefore, the above-calculated thresholds present rather overestimated values,
which could be considered as the upper limit for the thresholds.
6.5.2. Electron and ion temperature, relaxation time in the skin layer
In sub-picosecond laser ablation at the fluences around J/cm
2
the maximum electron temperature in the skin layer is of
the order of the Fermi energy k
B
T
e

F
, thus the electron heat capacity can be taken as that for the ideal gas: C
e
= 1.5k
B
.
Indeed, the maximumtemperature inthe Copper skinlayer (A = 0.7; l
s
= 5.710
6
cm; n
e
= 0.84510
23
cm
3
) heatedby
532 nm laser at 0.5 J/cm
2
equals to T
e,m
= 2AF/l
s
C
e
n
e
= 6 eV. The copper atoms at this temperature are singly ionised, and
the electronion momentumexchange rate is close to its maximumlimit of the plasma frequency,
mom
ei
1.64 10
16
s
1
.
The electron-to-ion energy transfer time expresses as t
en
ei
= M
i
/m
e
mom
ei
. For Copper at the above conditions this time equals
to t
en
ei
= 7 ps, which is much longer than a conventional 100 fs pulse. We assumed here that n
e
n
a
and C
L
C
e
= 1.5k
B
.
Thus, the maximum ion temperature after electronion equilibration reaches T
L,m
= T
e,m
/2.
The electron heat conduction starts spreading energy fromthe beginning of the pulse. The heat diffusion coefficient in the
considered conditions is D
th
= l
e
v
e
/3 v
2
F
/3
ei
= 1 cm
2
/s, which results in temperature smoothing time over the skin
layer t
th
= l
2
s
/D
th
32 ps, which is longer than the electronion temperature equilibration. The temperature at the target
surface after smoothing over the skin depth d = l
s
, then equals to

T = T
L,m
l
s
[1 exp(2d/l
s
)]/2d = 0.43T
L,m
0.2T
e,m
.
6.5.3. Propagation of a heat wave and a shock wave formation
There is two processes whichdetermine the energy transport to the outer boundary of the laser-heatedlayer: the electron
heat conduction and the shock wave. The heat conduction starts at the beginning of the pulse. However, the electron-to-
lattice equilibration is faster than the heat transfer. We ignore here, for simplicity, the electronion temperature difference
while calculating heat conduction during the temperature equilibration time. Let us describe the heat conduction by a
conventional non-linear heat conduction equation in one dimension assuming T
e
= T
L
= T, in the following way:
T
t
= D
0
x
_
T
T
0
_
m
T
x
; D = D
0
_
T
T
0
_
m
. (6.61)
The solution for Eq. (6.61) describes the heat transfer as the propagation of the heat wave (Zeldovich and Raizer, 2002):
T(x, t) = T
m
_
t
0
t
_
1/2+m
f (); (6.62a)
=
x
x
front
; (6.62b)
x
front
= x
0
_
t
t
0
_
1/2+m
. (6.62c)
The function f () describes the spatial form of the heat wave; it depends on the value of m. The case m = 0 corresponds to
the conventional linear heat conduction. One can see from the general formulae that contribution of non-linearity leads to
the slowing down of the heat wave front, when compared to the linear case. For a conservative estimate, we assume linear
heat conduction. The heat wave front velocity then reads:
dx
front
dt
=
1
2
_
D
0
t
_
1/2
; x
0
= (D
0
t
0
)
1/2
; t
0
= t
p
. (6.63)
The heat wave effectively stops when the velocity of the wave front compares to the local sound speed of the heated
material, but the shock wave overcomes the heat front and continues to propagate further. Let us approximate the sound
velocity v
s
= (P/)
S
by the expression v
s
( k
B
T/M)
1/2
where 3 is the adiabatic constant for solids. Then the
condition for the heat wave transformation into a shock wave reads:
dx
front
dt
= v
s
;
1
2
_
D
0
t
sw
_
1/2
=
_
k
B
T
M
_
1/2
. (6.64)
The shock wave formation time, t
sw
, immediately follows from Eqs. (6.64):
t
sw
=
D
0
M
4 k
B
T
.
One can easily see that this time is of the order of several picoseconds, as expected. It is not surprising because the
hydrodynamic motion (the ions motion) begins when the energy is transferred from electrons to the ions. Therefore, the
shock wave formation time coincides with the electron-to-ion energy transfer time. Note, that the electronion energy
exchange time in plasma is connected to the electronion momentumexchange rate as: t
en
ei
= M
i
/m
e
mom
ei
. The momentum
exchange rate is close to the plasma frequency at the temperature conditions above the ablation threshold. The shock wave
emergence time in Gold, for example, is t
sw
= t
en
ei
M
i
/m
e
pe
18 ps (M
i
= 197 a.u.;
pe
2 10
16
s
1
). The distance
travelled by the heat wave in a solid with the thermal diffusivity D
diff
during the electronion energy transfer time expresses
as x
ei
= (D
diff
t
ei
)
1/2
. This distance in metals is comparable to the skin depth. Thus the energy is delivered by the heat and
shock waves deep into the skin layer in several picoseconds, and the material expansion begins.
6.5.4. Expansion of a laser-affected solid into vacuum
The initial discontinuity of density and pressure at the vacuum (air) boundary decays at the moment of the arrival of the
shock wave. The material starts unloading and the strong shock wave starts propagating into the air. Simultaneously, the
rarefaction wave begins to move with the velocity equal to the local sound speed in the opposite direction, into the target
material. One can apply the theory of a solid unloading for the description of material removal when the shock wave arrives
at the solidvacuum (air) interface (Zeldovich and Raizer, 2002).
Let us consider for the sake of simplicity the material expansion into vacuum when the pressure and density at the
front of expanding material is equal to zero. The expansion proceeds with the constant entropy (adiabatically). Therefore
the Riemann invariant expresses through the sum of the shock velocity u in a solid plus the additional velocity u
, due to
expansion:
u +
P
0
dP
c
u +u
= const. (6.65)
The driving pressure reads, P = P
0,t
P
c
. The shock is weak when the pressure behind the shock front is slightly in excess
of the cold pressure, P
c
, which is a characteristic of strength of the solid. The velocity-doubling rule suggests that u u
(Zeldovich and Raizer, 2002). For a weak shock wave the velocity of expanding material front can be expressed as follows:
u

P
0,t
P
c
0
c
0
. (6.66)
The pressure driving the shock wave at F > F
thr
; d l
s
/2 then reads:
P
ot
= (C
e
n
e
+C
L
n
a
)
T =
AF
d
[1 exp(2d/l
s
)].
Note that the threshold fluence is proportional to the thickness d of the layer to be ablated. One can see that at the
laser fluence slightly above the threshold the ablated material expands with the doubled local speed of sound, while
the rarefaction wave moves inside the material with the sound velocity. The rarefaction wave reaches the transparent
shieldtarget interface during the time t
rarefaction
d/c
0,t
. That takes 1015 ps from the beginning of the unloading. The
front of expanding material moves during the same period out of the target to the distance approximately twice the initial
target thickness R 2d. Thus, the whole thickness of the unloaded solid at 30 ps after the pulse end is 3d.
The expansion of unloaded vapour is isentropic (constant entropy) process. Therefore the phase state of the expanding
vapour lies on the isentrope, which corresponds to the initial state of the unloaded solid. The phase state of expanding
gas strongly depends on the initial (before the expansion) temperature and pressure in the laser-heated material. If the
temperature is low then gas expands along the low isentrope, which can pass through the phase states where the mixture
of phases is most probable. Therefore after some time the liquid droplets may appear in the expanding gas. Another
limiting case is the case when the temperature of a laser-excited solid exceeds 45 times the binding energy (the energy of
vaporisation per atom). In this case a phase state of the expanding vapour is described by the high isentrope, which is located
in the phase space containing only the gas states. The description of this case is identical to the problem of atomisation of a
plume created by surface ablation that described in Section 5.
6.5.5. Strong shock-wave created by a solid expanding in ambient gas
Shock wave created by the unloaded solid in the ambient gas next to the target may serve as a diagnostic tool for the
defining the parameters of expanding front of the ablated material. The pressure at the expanding front is always much
higher than that of the atmospheric pressure in air even if the shock wave in a solid is weak. Therefore the shock in air is
always strong and its propagation can be detected by interferometry with time resolution. This phenomenon was studied
thoroughly experimentally and theoretically (Zeldovich and Raizer, 2002) and procedure of recovering the ablation front
parameters from the shock wave parameters in air is simple and well developed. The problem of shock wave formation,
propagation and gradual transformation can be solved numerically in the frame of two-dimensional hydrodynamics. It
seems instructive to present simple analytical solutions for clear understanding of the underlying physics.
The plane wave in the course of propagation slows down, spreads laterally (along the target surface) and gradually
transforms into a spherical wave. This occurs when the mass of the air embraced by the shock exceeds the mass of expanding
target material, M
air
M
t
.
(a) Plane shock wave propagation
Initially the shock wave front in air is close to that of the plane wave due to the conditions of ablation. Indeed, the focal
spot diameter is always significantly larger than the skin depth and the target thickness d
foc
l
s
. Therefore the shock wave
can be considered as a plane wave at the distances less than the focal spot diameter, d
foc
. Correspondingly, the shape of the
removed material resembles a very thindisc withthe radius only slightly, by a disc thickness, larger thanthe focal spot radius
(we assume here that the laser focal spot is circular). The shock wave can be considered as a plane wave at the distances
less than the focal spot diameter. One can describe the propagation of a plane wave using an approach of infinitely-thin
sheet explosion similar to the familiar spherical point explosion of (Zeldovich and Raizer, 2002). Let us suppose that the
infinitely thin plane layer with the surface area S
foc
instantaneously starts moving in the direction perpendicular to the plane
with the velocity D
pl
covering distance H
pl
= D
pl
t. The total energy, E = S
foc
H
pl
0
D
2
pl
, and area of the expanding layer are
conserved during the explosion;
0
is the characteristic initial density. Now, the time dependences of the travelling distance
and the shock front velocity immediately follow from the energy conservation:
H
pl
t
2/3
; D
pl
t
1/3
.
These dependences will hold till the plane wave expands more than the size of the ablated area, where the wave gradually
converts into a spherical wave.
(b) Spherical shock wave propagation and conversion to the sound wave
At the distance H
pl
> d
foc
the shape of the shock wave front transforms from a plane wave to the expanding spherical
wave. At this stage a spherical point explosion approximation can be applied for the description of the spherical shock
wave propagation. Spherical shock moves with the velocity D
sph
along the radius R
sph
= D
sph
t. The energy conservation
now reads: E = (4R
3
sph
/3)
0
D
2
sph
. The time dependence of the radius of the expanding shock wave along with the shock
velocity now takes a familiar form (Zeldovich and Raizer, 2002):
R
sph
t
2/5
; D
pl
t
3/5
.
More and more of air mass is involved in the shock front propagation later, the shock wave loses its energy, slows down,
and finally transforms into a sound wave. The expansion of the rarefied target material ends when the pressure behind the
shock front becomes close to the atmospheric pressure in air. This is the moment when the shock wave transforms into a
conventional acoustic wave moving with speed of sound. Further material transfer is continuing only by slow diffusion of
the target vapour through air. The successive transformations of the plane wave into the spherical one and finally to the
sound wave are presented in Fig. 6.12.
In conclusion, it is established that the thresholds for the material ablation in the Laser-Induced Forward Transfer (LIFT)
ablation mode are similar to those for the ultra-short powerful pulse surface ablation. High initial velocity of the shock
wave and correspondingly a higher material transfer velocity is a principal advantage of the short pulse ablation mode in
comparison to the long pulses in the application of the LIFT technique for the deposition of thin films and reproducing
patterns.
Fig. 6.12. Shock wave transformation in air versus time. The initial plane wave R t
2/3
transforms into a spherical R t
2/5
, and finally to an acoustic
wave R t. Calculations were performed for ablation of 40 nm thick Chromium layer by a 150 fs 800 nm laser pulse with the fluence 0.5 J/cm
2
.
6.6. Summary
Let us summarise the main conclusions of this section on ultrafast laser-induced material modifications in confined
geometry:
(i) In the conditions below but close to the optical breakdown threshold, single femtosecond laser pulse creates optically
detectable changes of the refractive index inphotorefractive materials. The inducedmodifications inthe refractive index
are short-lived and quasi-permanent. The modifications occur due to the excitation of electrons of all the constituent
atoms. Permanent modification occurs in the doped sites after the end of the pulse when the field of spontaneous
polarisation balances the field of charge separation.
(ii) Femtosecond laser pulse tightly focused by means of high-NA optics to intensity well above the optical breakdown
threshold deposits the absorbed energy density in excess of the strength of any material. A void surrounded by a
compressed shell is formed as a result of the confined micro-explosion in the focal volume. Warm Dense Matter at
the pressure exceeding TPa and the temperature above 100,000 K is created, mimicking the conditions in the cores of
stars and planets.
(iii) Confined micro-explosion studies open several broad avenues for research, such as formation of three-dimensional
structures for applications in photonics, studies of new materials formation, and imitation the inter-planetary
conditions in the laboratory tabletop experiments.
(iv) The material removal in the Laser-Induced Forward Transfer starts in the confined geometry conditions. The ablation
threshold for the LIFT mode is close to the surface ablation threshold value. The expansion of material in the LIFT
conditions is successively described as driven by a plane shock wave, followed by a transition to a spherical shock,
and finally to a sound wave, making the basis for analysis of forward-transfer deposition studies.
7. Applications of ultra-fast laser interaction with matter
7.1. Introduction
Ultra-short lasermatter interaction has a number of distinctive features, which have attracted several specific
applications. First attribute to be mentioned is the ability to control the phase state of ablated vapours through the optimum
pulse and focusing conditions adjusted to the ablated material properties. This feature ensured the application of ultra-
short lasers for ablation of different materials with the subsequent condensation of the ablated plume on a substrate and
producing thin films of superb quality and preserved stoichiometry. The ablation of solids by powerful lasers with a wide
range in pulse duration has attracted a significant attention during the past decades due to many potential applications of
this effect in industry, medicine, material science and technology. However, pulsed laser deposition (PLD) in its conventional
form using low-repetition rate lasers with nanosecond-range pulses (Chrisey and Hubler, 1994; Miller and Haglund, 1998)
generally leads to poor quality films contaminated by particles. It has been shown that this disadvantage of conventional
PLD is a direct consequence of the ablation regime with far from the optimum laser parameters (Gamaly et al., 1999a,b;
Rode et al., 1999). The plume produced in the long-pulse regime expands as a super-saturated vapour. This results in vapour
condensation and formation of droplets from the vapour phase at the early stage of expansion being then deposited onto
the substrate. It has been shown in Section 5 that the ablated plume can be kept in the atomised state during the expansion
stage through the proper choice of pulse duration, wavelength, along with a proper spatial and temporal distribution of the
laser intensity, and by keeping the absorbed laser energy above some specific threshold. The experimental implementation
of these recipes for deposition of different materials is presented later in this section.
The second distinctive feature of ultra-short lasermatter interaction relates to the fact that the pulse duration is shorter
than the heat conduction time and the electron-to-lattice energy transfer (hydrodynamic) time. Consequently, there are no
heat and shock waves propagating in the material surrounding the focal spot where ablation occurs. Hence, in application
of ablation in micro machining for making holes, cuts etc., there is no collateral damage due to the absence of heating or
shocking in the surrounding laser-ablation area. The edges of laser-produced holes are sharp and well defined. This is also
very important in medical applications; the examples of ultra-fast lasers applications for removal of dental enamel are
presented. The ultra-short pulse conventionally has a lowenergy per pulse. Therefore, a small number of particles per pulse
(10
11
10
12
atoms per 100 nJ, 100 fs pulse) are produced in a vapour phase. The control over the spatial and temporal
temperaturedensity distributions in the expanded laser plume allow one to generate nanoparticles from the ablated
material with the particle size of several nanometres well defined by the combination of laser and material parameters.
The nanoclusters are formed due to sticky collisions of ablated atoms in the expanding plume during the period when the
density and temperature in a plume complies with the conditions for atoms to assemble into a cluster. These conditions
hold for a short period of time after the end of the pulse because the expansion is rather fast. Hence the time for aggregation
of atoms into a nanocluster is limited by a characteristic plume expansion time comparable to the pulse duration. Formation
of 48 nm carbon and silicon clusters and carbon-clusters-assembled foam is described in this section.
The experiments demonstrated a possibility to deliver a low energy ultra-short laser pulse deep inside the transparent
solid into a sub-micron focal volume by the means of tight microscope focusing with high NA optics. This results in either
reversible or permanent transformation of the material in the focal area as it has been described in Section 6. The reversible
changes of refractive index allow one to create a set of three-dimensional spots with a sub-micron size inside a transparent
solid as a writing process. Each spot serves as a memory bit. The probe beam then can detect each spotto read it. A low
intensity wide beam can be used to delete the refractive index modulations, offering the erasing process. Thus there is an
opportunity for creation of 3D writereaderase memory with unprecedented memory density of the order 10 TB/cm
3
.
The permanent changes in the form of a void at the focal area can be produced if the intensity of a laser beam tightly
focused inside a transparent solid is well over the breakdown threshold. There are several possibilities for scientific
and industrial application of this phenomenon (see, for example Misawa and Juodkazis, 2006). First, the permanent
structure of sub-micron size can serve as a bit for the non-erasable memory. Second, it is possible to form 3D structures
separated with different space periods to form a photonic crystal structure. By making interconnected voids it is possible
to form 3D waveguides for application in photonics. It was also demonstrated experimentally that extreme pressure and
temperature conditions with record high heating and cooling rates could be created using high intensities. Such conditions
are appropriate for formation of new super-dense super-hard materials. A super-dense body-centred-cubic modification of
aluminium(bcc-Al) was recently discoveredinfs-laser-inducedmicro-explosionconfinedinside a sapphire crystal (Vailionis
et al., 2011).
7.2. Experimental results: ablation and deposition of thin films
Laser ablation process should comply with two necessary conditions to be successful for producing high quality thin
films by laser plume deposition on a substrate. First, high average ablation and deposition rate should be achieved. Second,
the laser plume should be homogeneous and with well-defined atomic content in order to ensure the high homogeneity
and surface finish of the produced films. A single short low-energyhigh-intensity pulse evaporates a few hot atoms per
pulse, thereby inhibiting the condensation of droplets during the fast non-equilibrium expansion. To compensate for the
reduced ablated mass per pulse, high pulse repetition rates are then used to achieve the high ablation-deposition rates per
second. The goal can be achieved with the use of short, 0.110 ps, low energy (J) pulses delivered at high repetition rates
of 10 kHz100 MHz, at the average intensity per pulse 10
13
10
14
W/cm
2
(Rode et al., 1999; Perry et al., 1999; Rode et al.,
2002a,b). The high level of average intensity leads to removing 10
11
10
12
atoms per 100 fs pulse, which corresponds
to a number of atoms in one cubic micron of solid, while the high repetition-rate maintains the average atomic flow at a
sufficiently high level of 10
18
10
19
atoms/s for a reasonable deposition rate. The ultra-fast laser deposition method was
proposed to achieve this goal combining the short laser pulses (picosecond or femtosecond) with very high repetition rate
of up to tens of megahertz (Rode et al., 1999, 2001). The application of this method for deposition of thin films has resulted
in atomic surface quality of the diamond-like carbon films (Perry et al., 1999; Rode et al., 1999) with the total elimination
of the macroscopic particles from the film surface. In this set of experiments the ablation of carbon has been performed by
76 MHz laser in conditions when single pulse energy was above the ablation threshold. As
2
S
3
chalcogenide optical films
(Rode et al., 2002a,b), and gallium films (MacDonald et al., 2001; Uteza et al., 2004) have been successfully produced with
similar surface quality and high volume homogeneity. The ablation of chalcogenide glasses by MHz repetition rate laser was
performed in the conditions when the energy of a single pulse was insufficient to produce ablation. The ablation occurred
due to the energy accumulation from the successive pulses hitting the same spot on the sample surface. This new ablation
Fig. 7.1. AFM images of a 13 13 m
2
surface area of laser deposited carbon films. (a) 10 kHz, 15 ns pulses, large particles of up to 30 nm height are
clearly seen; (b) 76 MHz, 60 ps pulses, the surface roughness <1 nm over the area 225 m
2
demonstrates the deposition of particulate-free films with
down to atomic level surface smoothness.
mode allows one to ablate any material, including transparent dielectrics and semiconductors, with a very high ablation
rate. Below the results of these experiments are presented.
7.2.1. Ablation and deposition of carbon films with atomic surface quality
A Nd:YAG mode-locked laser with 76 MHz, 60 ps, 300 nJ pulses at = 1.064 m was used for the ablation of carbon
in the experiments of Rode et al. (1999). The average intensity on the target surface was 2.8 10
9
W/cm
2
, the fluence
of 1.44 J/cm
2
was well above the ablation threshold. The quality of the deposited films was examined with the Scanning
Electron Microscopy (SEM), Atomic Force Microscopy (AFM), and optical microscopy. The number of particles visible with
an optical microscope on a carbon film was less than one particle per mm
2
. SEM images demonstrated that the deposited
film had a very fine surface texture with nanoscale irregularities. AFM surface micro-roughness measurements revealed a
saturation-like behaviour of the RMS roughness almost at the atomic level (<0.5 nm). Raman spectroscopy of the deposited
films indicated that films consist of a mixture of sp
3
and sp
2
bonded amorphous carbon. The thickness of the amorphous
carbon film deposited simultaneously on two 4 inch silicon wafers varied by only 5% over an area of 250 cm
2
. The
deposition rate was 26 /s on a substrate located at a distance of 150 mm from the target, which is 10 to 25 times
higher than that achieved with conventional high energy low repetition rate nanosecond lasers. The AFM images of the
film surfaces deposited with long, ns-range 10 kHz, and 60 ps, 76 MHz repetition rate are presented at Fig. 7.1(a, b) for
comparison.
7.2.2. Ablation and deposition of homogeneous chalcogenide glass films
The films of chalcogenide glasses of the high optical quality and atomic surface quality were produced by ablating
solid glass targets in vacuum using a 7 W Nd:YAG frequency doubled mode-locked laser (532 nm, repetition rate 76 MHz;
50 nJ/pulse; 60 ps; focal spot 20 mm; intensity I = 2.65 10
8
W/cm
2
; Fig. 7.1) (Rode et al., 2001, 2002a,b). An accurate
transfer the stoichiometry of a multi-component target to a film deposited on a substrate has been clearly demonstrated.
This is particularly important in materials containing weakly bonded volatile components such as sulphur where thermal
evaporation often results in marked changes in stoichiometry. The surface quality of the deposited films was almost at the
atomic level with the rms surface roughness of the order of 0.42 nm. The resulting films were highly photosensitive and
could be used without thermal annealing for waveguide fabrication using the photo-darkening process. Quality of the film
for the waveguide application suggests that the light propagation losses are a minimum.
7.2.3. Ablation by powerful short-pulse multi-MHz-repetition-rate lasers
Thinfilmproductionby any methodfor applications inphotonics, optics, electronics andnano-technology shouldpossess
several important properties: the stoichiometry (atomic arrangement) of the initial material should be preserved; the films
should be homogeneous, defect and contamination free; the surface finish should approach that of an atomic monolayer;
and last but not least, the ablation-deposition rate should match the requirements for industrial applications. Laser ablation
by powerful short pulse multi-MHz repetition rate lasers meets all the above conditions and offers the best solution for ideal
ablation and deposition of any solid material (Gamaly et al., 2006). The main results and a practical guide for the deposition
of thin films are presented below.
Any material is transformed into a plasma state under the irradiation by an intense (>10
13
W/cm
2
) sub-picosecond
laser pulse at the very beginning of the pulse (Gamaly et al., 2002). The resulting laserplasma interaction ensures strong
absorption and heating of the surface layer to a temperature above the ablation threshold. Typically the mass of a material
ablated by a single short (100 fs) pulse is equivalent to about one cubic micron of a solid volume. Non-equilibrium
expansion after the pulse along with the small amount of the ablated material inhibits condensation of the plume into
droplets. Therefore such a plume used for thin film deposition can produce a droplet-free surface. The energy deposition
depends on the laser intensity distribution during the pulse and across the focal spot. To ensure the phase transformation
of the ablated material to the state of an atomised vapour, the intensity distribution over the focal spot should be a flat-top-
like (Gamaly et al., 2004). Thus, a practical recipe for choosing the single laser pulse parameters for ablation or machining of
any material, including semiconductors and transparent dielectrics, is as follows. The laser pulse duration should be short,
pre-pulse and post-pulse free, has a flat-top spatial intensity distribution across the focal spot. The wavelength should be
in the range where the material has strong absorption, and the laser fluence should be in the range 110 J/cm
2
, depending
on the nature of the sample to be used. The absorbed laser fluence has to be 35 times higher than the ablation threshold in
order to keep the expanding plume in an atomised gaseous state. With such a pulse one can deposit films of atomic surface
quality of very different materials such as carbon and chalcogenide glass (Rode et al., 1999, 2002a,b).
High repetition rate lasers should be used in order to compensate the small amount of material ablated per single short
pulse. Indeed, the use of short-pulse, few-MHz repetition rate lasers produces the highest ablation and deposition rate
when compared to any deposition methods. Studies of multi-MHz lasermatter interaction and ablation uncovered several
phenomena that bring applications of the laser ablation-deposition process closer. Firstly, fs-lasermatter interaction with
the pulse repetition rate in the MHz range involves accumulation processes due to a coupling between the successive pulses.
This coupling changes the interaction mode and thereby reduces the ablation threshold. As a result, high repetition rate
lasers can ablate a material at fluences below the single pulse threshold limit. The most striking example is ablation and
deposition of chalcogenide glass using a 76 MHz pulses with the energy per pulse being 10 times less than the single pulse
ablation threshold (Rode et al., 2002a,b). MHz-rate ablation results in smoothing of the evaporation conditions over the
focal spot because the heating proceeds by a succession of pulses shifted along the target surface by only a small fraction
of a micron during the time gap between the pulses. Therefore, the spatial pulse shaping requirement, which is critical for
a single pulse regime to control the phase state of the plume, is much less stringent in the MHz-rate ablation regime. This
was demonstrated during the deposition of chalcogenide glass films with high optical quality (Rode et al., 2002a,b).
Finally, the ablation by powerful short pulse multi-MHz repetition rate lasers appears to be the ultimate solution for the
ablation, deposition and micro-machining problems of any material.
7.3. Precise micro-machining: removal of dental enamel by ultra-fast laser ablation
Application of laser light for removal of hard dental tissue has been studied as a means for replacing conventional surgical
tools since the first use of lasers in the medical field. Until recently, lasers have not succeeded in replacing the dental drill in
many hard tissue applications due to unacceptable collateral damage and slowmaterial removal rates produced by relatively
long (from100 ps to microseconds) laser pulses. On the opposite, ultra-short lasermatter interaction occurs during a period
shorter than the time for the heat wave and shock wave to affect a material surrounding the laser focal spot. There is
no collateral damage by heating or mechanical effects. This feature makes ultra-fast laser ablation an indispensable tool
for application in micro machining of different materials. The absence of the collateral damage is especially important in
medical applications. The advantages of ultra-fast ablation for micro-machining were demonstrated with the experiments
for removal of the dental enamel (Neev et al., 1996; Feit et al., 1998; Kimet al., 2000; Serbin et al., 2002; Rode et al., 2002a,b;
Chan et al., 2003; Niemz et al., 2004; Chung and Mazur, 2009).
Understanding of the interaction physics (see Section 5) allows one to choose the optimum combination of laser
and material parameters in order to ablate a particular material. A number of experiments were performed in order to
demonstrate that ultra-short laser ablation complies with all demands for efficient and patient-friendly applications in
dental surgery. We describe one of those in more details.
The dental enamel has a very complex chemical composition and its properties, such as the enthalpy of evaporation,
the binding energy, and the ionisation threshold, vary from sample to sample and are virtually unknown. With the aim
to uncover at least some of these properties, laser ablation experiments were conducted with Ti:sapphire laser (Rode et al.,
2002a,b), with the wavelength 780805 nm, pulse duration 95 and 150 fs, power range 0.51.0 W, and repetition rate 1 kHz.
The ablation threshold and the ablation rate were measured as a function of laser fluence. An analytical model was proposed
to derive the effective binding energy of enamel fromthese data (Gamaly et al., 2002; Rode et al., 2002a,b). The analysis of the
ablation experiments allowed the authors to determine the average binding energy for the complex composition of dental
enamel andto predict the optimumablationconditions. Temperature measurements inthe root canal inside the toothduring
the ablation showed saturation at a level below the pulp thermal damage limit of 5.5 C with the use of air cooling at a rate
5 l/min. With a rise of intrapulpal temperature over 5.5 C, 15% of the tooth pulp may suffer irreversible thermal damage
(Zach and Cohen, 1965), so this level is regarded as critical. Considering the electrostatic ablation mechanism, a significant
temperature increase is not expected, however, the laser pulse contrast ratio, with a lowpower nanosecond pedestal on the
pulse, and the non-ideal Gaussian pulse profile may contribute to the heating observed. A straightforward way by which to
increase the laser contrast is to convert the laser radiation into the second harmonic. The improved contrast ratio together
with the reduced threshold fluence are clear indications in favour of using the second harmonic of fundamental frequency
for subpicosecond laser ablation of hard dental tissue.
The optical microscopy (Fig. 7.2) and scanning electron microscopy, see also Fig. 7.3 allows characterising the crater floor
and the collateral damage in the bulk of the ablated teeth. The images of the craters ablated with subpicosecond lasers show
the absence of any collateral damage at the crater edge as it can be seen at Fig. 7.2.
It can be seen from Fig. 7.3((b), (d), (f)) that the enamel ablated with 60 ps pulses showed clear evidence of thermal
damage to the tooth. Each ablated square had a darkened ring approximately 0.3 mmfromthe edge of the square, as well as
enamel droplets on the surface of the tooth next to the squares. The enamel surface near the edge of the crater was raised due
to heat deposition in the bulk of the tooth. The floor of the ablated craters appeared melted and uneven. In sharp contrast to
Fig. 7.2. Optical microscope (a) and SEM (b) images of 1 mm
2
craters on a tooth surface ablated with 150 fs laser pulses at 1 kHz repetition rate. There is
no collateral damage around the fs-laser ablated cavity; the edges are sharp and the crater floor is smooth. Average power 1 W, ablation time 2 min.
Fig. 7.3. SEMimages of 110.25 mm
3
crates in dental enamel produced by 150 fs pulses ((a), (b), (c)) and 60 ps pulses ((d), (e), (f)). There is no collateral
damage around the fs laser ablated cavity. With 60 ps laser there is a thermally modified area around the crater and some occasional cracks. The scale bars
are 10 m in ((b), (c), (e), and (f)).
the picosecond pulses, the enamel ablated by 95 and 150 fs pulses seen in Fig. 7.3((a), (c), (e)) showed no collateral thermal
damage at all. The walls and floor of these ablated squares were also smoother than those due to picosecond ablation and
the edges and corners were sharper than those ablated with 60 ps pulses.
The measured ablation rates over the intensity range of 10
13
to 4 10
14
W/cm
2
with both 95 and 150 fs pulses are
presented in Fig. 7.4. The ablation threshold fluence F
th
was found to be 2.2 0.1 J/cm
2
, practically the same for both
lasers. The ablation threshold with 400 nm was found to be 1.0 0.2 J/cm
2
, almost two times lower than for 800 nm in
accord with the theory, which predicts nearly linear threshold dependence on the wavelength (see Section 5). However,
prediction of the ablation threshold using an approximate interpolation for the ratio of the absorption coefficient to the skin
depth A/l
s

= 4/ underestimates the absorption coefficient and predicts a lower threshold in comparison to the measured
one. The ablation rate and ablation mass per pulse increases in linear proportion to the absorbed fluence up to 36.5 J/cm
2
,
i.e., up to 15F
th
.
Fig. 7.4. (a) Measured ablated rates of dental enamel vs. laser fluence with 95 fs and 150 fs laser pulses; the line is the theoretical prediction, taking
the experimentally determined ablation threshold of 2.2 J/cm
2
for both pulse durations. (b) Spatial distribution of the crater profile (crosses) and of the
inverted laser pulse profile (dots), illustrating the linear dependence of the ablated depth on laser intensity.
One of the aims was to test the linear dependence of the ablated mass on the laser fluence. The laser beam with a
fluence of 36.5 J/cm
2
was scanned along a single axis, which produced a channel on the enamel surface. The cross section
of the ablated channel was measured with a nanoprofilometer. This cross-section is presented in Fig. 7.4 along with spatial
intensity distribution in a laser beam across the focal spot superimposed over the channel shape. The laser beam spatial
profile is upturned and normalised to the maximum depth of the channel. As can be seen, the channel profile faithfully
reproduces the spatial intensity distribution in the laser beam. This result confirms the theory prediction that the ablated
mass is directly proportional to the laser fluence up to 15 F
th
in accord with the rate measurements. The ablation depth per
pulse can be expressed through laser and material parameters (see Section 5):
l
abl

=
AF
n
a
(
b
+J
i
)
.
On the base of these measurements it was suggested to take for characterisation of the dental enamel the following
parameters (Rode et al., 2002a,b): the atomic number density n
a
= 7.8 10
22
atoms/cm
3
, the average binding energy
b
5 eV and the average ionisation potential J
i
12 eV per atom, close to that for dielectrics. Then one gets for absorbed
fluence of 10 J/cm
2
the ablation depth of 0.5 m per 100 fs pulse with the average intensity on the surface of 10
14
W/cm
2
in qualitative agreement with experiments.
The recommendations for precise laser machining of dental enamel by subpicosecondlaser canbe formulatedonthe basis
of the experimental results and scaling from the references (Rode et al., 2002a,b, 2009). The dental tissue can be removed
with 150 fs pulses at 20 J/cm
2
at the rate of 1 m/pulse or 210
6
mm
3
/s (for the focal spot area of 2.210
5
cm
2
). The
removal rates for dental tissues by mechanical drill are of the order of 0.51 mm
3
/s, depending on the cutting tool chosen.
Thus tissue removal rates with subpicosecond laser ablation can be made comparable to those of mechanical with pulse
repetition rates of the order 500 kHz, on conditions that the contrast of the fs-laser pulse has to be carefully maintained to
avoid heat accumulation.
7.4. Formation of nanoclusters by ultra-short pulses
The major problem in nanotechnology resides in the development of efficient methods for controlled, in size and in
properties, formation of nanoparticles that could be then scaled to industrial production. The first step in a bottom-up
formation process is to decompose the initial material into a preferably atomic state via ablation, forma cloud of hot atoms,
andcreate temperature anddensity conditions necessary for the assembly the atoms into nanostructures by sticky collisions.
These conditions should be maintained during the period necessary for assembly of a nanocluster with desirable size. The
formation conventionally occurs during the diffusion of hot atomic cloud through the non-reacting ambient gas that serves
as a confinement.
Laser ablation has proven to be an efficient method for producing nanoclusters of different atomic content, shape and
internal structure. Such techniques have been used for production of nanoclusters since the early eighties. Initially long
pulse (1030 ns) low repetition rate (1030 Hz) commercially available lasers with the average intensities per pulse of
(24) 10
9
W/cm
2
have been employed. Metal and metal oxide clusters (Bondybey and English, 1982; Riley et al., 1982;
Ullmann et al., 2002; Itina et al., 2002), fullerenes (Kroto et al., 1985), carbon and boron nitride nanotubes (Iijima, 1991;
Ebbesen and Ajayan, 1992; Golberg et al., 1996; Ebbesen, 1997; Gamaly and Rode, 2004), silicon clusters (Marine et al.,
2000; Patrone et al., 2000; Amoruso et al., 2004, 2005b) and many other structures have been produced with laser ablation
in similar conditions. Special conditions are often required complementing laser ablation to promote a particular type of
structure to grow. For example, in case of carbon nanotube growth in a low repetition rate regime, the laser plume cools
down after the pulse below the minimal temperature required for the nanotube formation due to the long gap between the
successive pulses. Therefore, additional heating of the ambient gas was necessary in order to maintain the proper conditions
for cluster formation. Cluster formation by long pulse lowrepetition rate lasers conventionally occurs during the laser plume
interaction with an ambient gas in a chamber placed into a furnace with controlled temperature and under a continuous
flow of a noble gas.
Advent of the femtosecond lasers with repletion rate up to 100 MHz along with better understanding of physics of
lasermatter interaction and cluster formation, allowed the implementation of new approach to formation of nanoclusters
by short single pulses. It has been shown possible to control the cluster size by choosing the optimum combination of
lasertarget parameters, eliminating additional heating and even removing the necessity for ambient gas in the chamber
(Gamaly et al., 2009).
Laser ablation has proven to be an efficient method for producing nanoclusters of different atomic content, shape and
internal structure. The size of a cluster has a significant effect upon various material properties and, therefore, provides a
relatively simple experimental avenue to control those properties (Gamaly and Rode, 2004). There is a critical cluster size of
several nanometres that separates atomic ensembles composed from the same atoms but possessing the different material
properties. Melting and boiling points, conductivity, electronic structure etc. in a cluster with a size smaller than the critical
one are drastically different from the same material in the bulk. It appears that nanoclusters close to the critical size and
smaller canbe producedby ultra-short single laser pulse invacuumandinambient gas. Thus it is possible tocontrol the prop-
erties of nanoclusters, which are the building blocks for any nanomaterial, via control over their size (Gamaly et al., 2009).
A laser plume with an uncontrollable composition containing the mixture of liquid droplets, flakes, and wide range of
different macroscopic and microscopic particles is produced at the laser intensity close to the ablation threshold. Direct
removal of dimers or multi-atomic clusters is energetically unfavourable and kinetically improbable from the physical
viewpoint. Indeed, one needs to supply the energy in excess of their binding energy simultaneously to at least two bounded
atoms and at the same time keeping their mutual bonding with the two times lower energy intact. This only can happen
randomly when a specific arrangement of material defects in some place of the target occurs. High intensity well above
the ablation threshold should be used for producing the atomised plume. Total atomisation in the plume was proved in a
number of works on fs-laser ablation and deposition of high surface quality diamond-like films, where the laser intensity
was several times above the ablation thresholdsee, for example, (Amoruso et al., 2005a,b). Atoms re-ensemble into clusters
in an expanding plume in a process that controlled by the density and temperature in the plume directly steered by the laser
parameters.
The temperature and density of atoms in the adiabatically expanding plume remain appropriate for formation of clusters
through the atom-to-atom and atom-to-cluster attachments over a certain time period after the laser pulse. The longer this
period of high density and temperature, the more successive inelastic collisions occur in the plume, and hence the larger the
cluster may be formed. Thus the plume expansion time in vacuumis a major factor in determining the cluster size. This time
depends on the combination of laser and material parameters. Similarly, in the case of expansion in ambient gas the diffusion
time of the ablated material through the gas determines the cluster formation time and hence the cluster size. Clusters are
formed in the dense area near the laser focal spot close to the ablation surface. The cloud of clusters expands, and clusters
are deposited on substrates in the experimental chamber. Depending on the quantity of the ablated material, clusters are
deposited on a surface as individual particles or in a form of cluster-assembled foam-like substance. By applying a simple
kinetic model one can demonstrate that the formation of clusters is possible in a plume created by a single ultra-short laser
pulse. Then the predictions of the model are compared with the experiments where different clusters were produced in
vacuum and in ambient gases (Gamaly et al., 2009).
7.4.1. Kinetic model of cluster formation in a plume of laser-ablated atoms
We consider a simplified model of nanocluster formation aiming for prediction of the total number of atoms per cluster
as function of the average laser plume parameters. The price we pay for such a simplification is the missing information
about the internal structure of a cluster. We suggest that a cluster forms via the neutral-to-neutral inelastic sticky collisions
and ignore the ion-to-ion, ion-to-neutral and other complex dusty-plasma collision effects at temperature range above
ionisation. We assume that the atomic source and the time-dependent density of atoms, the building blocks for the clusters
formation n
1
(t), are known. Therefore, the number of atoms, their energy and momentum in a plume after the pulse
termination are conserved. We also ignore, as a first approximation, the spatial dependence of all quantities, which will
be accounted for later on.
The key assumption is that clusters are formed preferentially in a pair of inelastic (sticky) collisions and multiple
collisions are less probable. Second assumption is that building of a larger cluster occurs by a monomer addition. Therefore
the first important step on the way to creation of a large cluster is the formation of dimers. Generally a third body is required
ina two-body exothermic collisionproducing a single particle for fulfilment the momentumandenergy conservationlocally,
at the cluster formation spatial point. The third body should be another carbon atom or a cluster. The cross-section for
atomic collisions that responsible for cluster formation and its internal structure most probably is dangling-bond-direction
dependent. Hence, such a cross-section itself should be a very complicated function of atomic characteristics, which is
unknown to the best of our knowledge. A single cluster formation process, which is 3D problem with such a cross-section,
is a formidable task even for a modern-day supercomputer. However, the outcome would be unreliable providing that very
little is known about this cross-section function even for the triple collisions. Therefore a different way has been chosen
to solve the problem, where just a total number of atoms per cluster could be predicted as a function of average plume
parameters, based on the energy, momentum and number of atoms conservation in a laser created plume after the end of
the pulse. In this approach the cross-section is taken in the formapplied for collisions of hard spheres. Thus the cross section
for the cluster containing N atoms is simply proportional to N
2/3
, as we demonstrate below. In our approach the coupled
rate equations exactly comply with the conservation of the total number of ablated atoms in a plume while the energy and
momentumconservation are only implicitly (through the velocity of carbons) coupled to the rate equations. In exact atomic
formulation such coupling should be explicit through the energy and momentum exchange with a third bodies while in
considered case the energy and momentum is conserved separately. In our approach therefore the cross-section is simply a
parameter defining the time for formation of dimers, trimers etc. as a function of number of atoms per cluster. Comparison
to the experiments belowdemonstrates that the number of particles per cluster as a function of the laser and averaged laser
plume parameters can be predicted with sufficient accuracy.
(a) Cross section
The exact cross section for atom-to-atom attachment in a laser plume is unknown. It is reasonable to suggest that the
attachment cross-section is proportional to the elastic one. The cross section for neutral-to-neutral collision is taken as that
for collision of hard spheres. The diffusion coefficient in the mixture of two gases depends on the total cross section, which
expresses as (Lifshitz and Pitaevski, 1981):
transp
=
total
=

4
(d
2
1
+d
2
2
); (7.1)
where d
1
and d
2
are diameters of particles in the mixture. In the case of carboncarbon scattering
CC
= 1.86 10
16
cm
2
(the covalent radius 0.77 ; d
1
= 1.54 ). Diffusion of carbon atoms that is the case of cluster formation in ambient gas
(d
Ar
= d
2
= 3.76 ) depends on
CAr
= 1.310
15
cm
2
. In what follows we suggest that two-atomcollisions dominate the
formation process (e.g. multiple collisions are neglected). We take into account collisions of N-clusters (cluster containing
N atoms) with a single atom only (monomer addition), ignoring the collisions of heavy clusters with each other. The cross-
section of such an interaction then expresses:
N,1
=

4
(r
2
1
+r
2
N
)
=
1,1
N
2/3
(N 1)
r
N

= N
1/3
r
1
. (7.2)
We assume that the probability of sticky attachment <1 is proportional to the above atom-to-atom inelastic cross section.
(b) Particle conservation law
A set of coupled chemical rate equations for cluster number density, n
N
(number of N-clusters per unit volume) governs
the successive formation of clusters from dimers, n
2
, to N-clusters. The total number of atoms in all clusters at any time
moment in the entire interaction volume should be equal to the total number of single atoms produced by the ablation
source:
N
i=1
in
i
=
t
0
_
dn
1
dt
_
0
dt
. (7.3)
It is convenient to present this law also in the form:
N
i=1
i
dn
i
dt
=
_
dn
1
dt
_
0
. (7.4)
Eqs. (7.3) and (7.4) represent the law of mass conservation.
(c) Formation of N-clusters in pair collisions
The set of coupled rate equations describing N-clusters formation in a pair of inelastic collisions then reads:
i
dn
i
dt
= K
1,i1
n
1
n
i1
n
i
ik
k=1
K
k,i
n
k
; i = 2, 3, . . . , N. (7.5)
This set of equations represents just the atoms conservation law where the cross-section is a parameter proportional to the
number of carbons per cluster. The reaction rate conventionally expresses as:
K
i,k
=
ik
v
ik
; (7.6)
where v
i
k is the relative velocity of collisions of i-cluster and k-cluster. The differential equation (7.6) holds if any arbitrary
constant multiplier to reaction rate in Eq. (7.6) is included. It is assumed that velocity of single atoms v
1
v
i
(i > 2),
and the collisions between heavier clusters, 22; 23; 33 etc., are ignored because their velocities are lower, v
N
N
1/2
.
Therefore the reaction rate for the monomer addition, K
1,N
N
2/3
, is higher than that for the triple collisions or for pair
collisions of larger clusters, K
N,N
N
1/6
. The maximum cluster formed comprises N atoms. Then, the coupled set of rate
equations reads:
_
_
2
dn
2
dt
= K
11
n
2
1
K
12
n
1
n
2
3
dn
3
dt
= K
12
n
1
n
2
K
13
n
1
n
3

(N 1)
dn
N1
dt
= K
1,N2
n
1
n
N2
K
1,N1
n
1
n
N1
N
dn
N
dt
= K
1,N1
n
1
n
N1
.
(7.7)
Applying the particle conservation law one obtains the equation for the change in time of the number density of building
atoms:
dn
1
dt
=
_
dn
1
dt
_
0
K
11
n
2
1
. (7.8)
Eq. (7.8) can be immediately integrated if the source term is known. Afterwards, the set of Eqs. (7.7) could be successively
integrated.
(d) Simplified solution: time for the N-cluster formation
A simple solution for the set of Eqs. (7.7) is found under following assumptions. First, it is assumed that the relative
velocities are v
1,N
v
1
. Second, the attachment cross-section expresses as
N,1

1,1
N
2/3
; (N 1). We also assume that
the number density of constituent atoms is time-independent and n
1
n
2
, n
3
, . . . .n
N
. Then set of Eqs. (7.7) reduces to the
following:
2
dn
2
dt
= n
2
1
11
v
1
n
1
/t
2
; t
2
= (n
1
11
v
1
)
1
3
dn
3
dt
= 2
2/3
n
2
/t
2

(N 1)
dn
N1
dt
= (N 2)
2/3
n
N2
/t
2
N
dn
N
dt
= (N 1)
2/3
n
N1
/t
2
.
(7.9)
The set of Eqs. (7.9) can be successively integrated by time to obtain the number density of clusters containing N atoms
each:
n
N
=
n
1
[(N 1)!]
1/3
N!
_
t
t
2
_
N1
. (7.10)
The case under consideration corresponds to N 1, thus from the particle conservation follows that n
N
n
1
/N. Now Eq.
(7.10) reduces to:
N
1
1
[N!]
4/3
_
t
N
t
2
_
N
; (7.11)
where t
N
is the time necessary for the N-cluster formation. In the further simplification the factorial function using the
Stirling limit case for N 1 reduces to N!
= (2)
1/2
N
N+1/2
exp(N). Then Eq. (7.11) transforms into:
N e
_
t
N
t
2
_
3/4
. (7.12)
Here e is the basis for the natural logarithm. Inserting the time for the dimer formation, t
2
, one obtains that the number of
atoms per cluster is directly proportional to the number density and velocity of the source atoms:
N e (t
N
n
1
11
v
1
)
3/4
. (7.13)
It was implicitly assumed in this derivation that the number density of atomic source n
1
and atomic velocity v
1
are constant
during the formationtime t
N
. Infact, the density spatial distributioninthe plume changes fromthe soliddensity to zero. Time
for N-cluster formation above is actually the time when the plume volume increases to the size L at which its temperature
drops down to the limit temperature for the cluster formation. For the one-dimensional adiabatic expansion that suffices to
the case considered, one gets the formation time t
N
L/v
1
. The expression in parenthesis in Eq. (7.13) then can be presented
as t
N
n
1
11
v
1
Ln
1
11
= L/l
mfp
N
aa
, that is the number of atomatom collisions in a plume; here l
mfp
= (n
1
11
)
1
is
the atomic mean free path in the plume. As the plume atomic density is space-dependent, N
aa
should be presented in the
form:
N
aa
=
L
0
dz
l
mfp
=
L
0
n
1
(z)
11
dz. (7.14)
One can see that for the linear density dependence n
1
(z) = n
0
xzL
1
, for the exponential dependence, n
1
(z) = n
0
exp(z/L),
and for the adiabatic expansion in vacuum n
1
(x) = n
0
(1 z/L)
2/( 1)
the number of collisions differs only by numerical
coefficient of the order of unity (b = 0.250.63) which we leave as a fitting parameter:
N
aa
= bn
0
11
L. (7.15)
Here n
0
is the initial density of the target. Formula (7.15) indicates clearly on the dominant contribution of the dense parts
of the plume into the cluster formation process. Now, the number of atoms per cluster expresses as:
N B(n
0
11
L)
3/4
. (7.16)
Here B = eb
3/4
is the fitting coefficient that should be extracted from comparison to the experiments (B changes within the
range 0.961.92).
(e) The number and temperature of ablated carbons
We apply now the above model to the cluster formation by a single pulse and estimate the cluster size as a function
of laser parameters. The cluster formation scenario is as follows. First, the flow of hot carbons is created during ablation.
The laser pulse duration is too short for the clusters to be formed during the pulse. Therefore, after the end of the pulse
the ablated vapour either diffuses, when the chamber is filled with a gas, or adiabatically expands into vacuum. The total
number of atoms ablated per pulse is N
abl
= n
0
V
0
n
0
S
foc
l
abl
; here S
foc
is the focal spot area, n
0
is a number density of a
target, l
abl
is the ablation depth, and V
0
is the ablated volume.
The laser plume has to be in a highly collisional state in order to support effective cluster formation process. This requires
high ablation rate and thus reasonably high laser ablation fluence well above the ablation threshold. In these conditions
the target material is at least single ionised and electron and ions specific heat can be taken as that for the perfect gas,
C
e
= C
i
= 3k
B
/2. We also suggest that at the beginning of the plume expansion the temperatures of electrons and ions are
equilibrated. Thus, the maximum initial energy per atom at the beginning of expansion reads:
T
0
=
2(F
a
F
thr
)
3n
a
l
abs
; (7.17)
F
a
, F
th
is respectively the absorbed and ablation threshold fluence, l
abs
is the absorption depth. The ablation threshold in
accord to Gamaly et al. (2002) reads:
F
th

3
4
n
a
l
abs
(
b
+J
i
)
A
; (7.18)
b
is the binding energy and J
i
is the first ionisation potential. For ablation of graphite target as a source of carbon atoms
(A 0.85; l
abs
30 nm; (
b
+J
i
) 15 eV; and n
a
= 10
23
cm
3
for graphite) one gets the threshold F
thr
= 0.8 J/cm
2
. The
ablated depth can be expressed as the average between two limit cases, l
neq
abl
< l
abl
< l
max
abl
where the limits are determined
by the non-thermal (non-equilibrium) mode of ablation (Gamaly et al., 2002). The maximum ablation depth is defined by
the condition all the absorbed laser energy spent on breaking bonds and t the kinetic energy of ablated atoms is zero:
l
max
abl
=
F
a
n
a
(
b
+J
i
)
. (7.19)
The lower limit for the ablation depth l
neq
abl
is determined by a condition that the kinetic energy of ablated atoms is at its
maximum, l
neq
abl
= 0.5l
abs
ln(F
a
/F
thr
).
Let us compare these results with the experimental data. The ablation depth can be estimated fromthe measured ablated
mass m
abl
per single pulse as:
l
exp
abl
=
m
abl
S
foc
.
Taking the incident laser fluence of 8 J/cm
2
, and the absorption A = 0.85 (F
a
= 6.8 J/cm
2
) one gets l
abl
= 160 nm, which
is in qualitative agreement with the experimentally measured 200 nm (Gamaly et al., 2009). The ion temperature in the
energy units at the beginning of expansion for the same parameters is T
0
= 21.3 eV.
(f) Expansion-limited aggregation in expanding plume of a single pulse
Let us assume that the ablated plume adiabatically expands after the end of the pulse in vacuumas a perfect gas with the
adiabatic constant , T = T
0
(V
0
/V)
1
. We suggest that clusters can be formed in an expanding plume during the period
when the carbon temperature exceeds the minimum temperature for cluster formation, T
min
= T. The plume volume at
that instance reads:
V
max
= V
0
(T
0
/T
min
)
1/( 1)
. (7.20)
There are two possible scenarios of expansion depending on the experimental conditions: plume propagation in one-
dimension (plane-wave expansion) or in three-dimensions (hemispherical expansion). At the initial expansion stage when
the dense plume thickness is much less than the focal spot size, which is of the order of tens of microns, the expansion is well
approximated as one-dimensional motion: (V
max
= S
foc
L
exp
; V
0
= S
foc
l
abl
). The expansion length from Eq. (7.20) expresses
as:
L
(1D)
exp
= l
abl
_
T
0
T
min
_ 1
( 1)
l
abl
F
1
( 1)
a
. (7.21)
In 3D-expansion case the parameters are as the following: V
(3D)
max
= (4/3)(L
(3D)
exp
)
3
; V
0
= S
foc
l
abl
; r
0
= [(3/4)S
foc
l
abl
]
1/3
.
We take the adiabatic constant as that for the ideal gas, = 5/3. Then the expansion length reads:
L
(3D)
exp
= r
0
_
T
0
T
min
_
1/2
l
1/3
abl
F
1/2
a
. (7.22)
Note that 3D-expansion length is in accord with hydrodynamic solution for the adiabatic expansion (Zeldovich and Raizer,
2002). Now, one can also calculate the cluster radius, r
cl
= (3N/4n
0
)
1/3
, with the help of Eq. (7.16) and assuming that the
cluster material density is known:
r
cl
= (3N/4)
1/3
n
1/12
0
(
11
L
exp
)
1/4
. (7.23)
Eq. (7.23) shows very weak dependence on the material density. We suggest that the minimum temperature for formation
of carbon clusters equals to the temperature of graphitisation of 1200 K = 0.1 eV.
Experiments at low pressure of 20 mTorr, which we can consider as experiments in vacuum (the atomic mean free
path is comparable to the targetsubstrate distance), with S
foc
= 10
5
cm
2
, F
a
= 6.8 J/cm
2
, ablation depth of 200 nm,
CC
= 1.8610
16
cm
2
, T
0
= 21.3 eV correspond to 3D-expansion because L
exp
r
foc
at these parameters in accord with
Eq. (7.22). By applying these parameters to the Eq. (7.22) along with Eq. (7.23) and taking numerical coefficient B = 1, the
cluster diameter is 2.68 nm, that agrees qualitatively well with the measured average diameter of 3.2 0.5 nm (Madsen
et al., 2006; Gamaly et al., 2009)
One should note that the attachment probability, as well as the minimum temperature for the cluster formation, is
unknown to the best of our knowledge. Those parameters were suggested above on the basis of general laws as well as
several approximations. Therefore it is advisable to deduce the fitting coefficient B from experiments, giving that the above
scaling reproduces the major trends qualitatively correct.
(g) Diffusion limited aggregation: cluster growth in ambient gas
Growth of clusters in the ambient gas was conceived as a basic procedure for any cluster formation process. Use of an
ambient gas has the advantages of a confinement that increases the lifetime of constituent atoms in formation region and
therefore increasing the probability of sticky attachment. On the other hand, the ambient gas acts as a heat sink decreasing
the temperature necessary for cluster formation. This was the reason for heating the argon gas to 1200 C1600 C during the
nanotube growthby the laser ablationprocess (Ebbesen, 1997). Another limiting factor is that the pressure of the ambient gas
cannot be increased over the threshold for optical breakdown, which would increase the absorption in plume and therefore
decreasing the ablation rate.
The shock wave forms and propagates into the gas immediately after the laser pulse. The shock front however is smeared
over a distance comparable to the carbon mean-free-path, l
mfp
= (n
Ar
CAr
)
1
. For the 501000 Torr argon pressure range
the number density range is n
Ar
= 3 10
18
to 6 10
19
cm
3
. The mean-free-path range l
mfp
= (0.153.33) 10
4
cm is
longer than the ablation depth, which is the thickness of the energy deposition region. Diffusion therefore dominates, and
thus we ignore the shock wave stage in our future estimates.
Let us first obtain the maximum diffusion length from condition that cluster formation stops at some temperature T
min
.
Diffusion of single carbons in argon of density n
Ar
proceeds with the diffusion velocity, D = lv
c
/3 v
c
/(3n
Ar
CAr
);
here
CAr
is the cross section for carbonargon elastic collisions, which is taken the same as that for the hard sphere
collisions. Note that CAr collision cross section is almost 10 times larger than that for CC collisions. We assume that
carbons cooling proceeds in two overlapping stages: a nearly adiabatic expansion accompanied by the additional cooling
due to carbonargon collisions. At the adiabatic expansion stage the plume cools down to the temperature T
C
= T
0
(r
0
/L
D
)
2
,
where r
0
= (3S
foc
l
abl
/4)
1/3
. Correspondingly, the carbon density at the same moment expresses as n
C
= n
0
(r
0
/L
D
)
3
.
Temperature in the carbonargon mixture after equilibration is T
mix
= T
C
[n
C
/(n
Ar
+n
C
)]. Now, the maximum diffusion
length for the cluster formation zone is defined from the condition that the temperature in the mixture equals to the
minimum temperature for the cluster formation, T
mix
= T
min
, which expresses by the algebraic equation:
T
min
=
T
0
n
C
n
Ar
+n
C
_
r
0
L
D
_
2
. (7.24)
It is convenient to introduce a new variable x = L
D
/r
0
, and then the above equation takes a form:
x
2
_
1 +x
3
n
Ar
n
0
_
=
T
0
T
min
. (7.25)
In the pressure range P
Ar
= 501000 Torr the argon density is n
Ar
= (360) 10
18
cm
3
. Thus one can see from
the above equation that the influence of ambient gas on diffusion length of ablated carbons (n
0
= 10
23
cm
3
) becomes
significant when the gas pressure approaches to 10% of the atmospheric pressure. One can present solution of Eq. (7.25)
for two limiting cases: a diffusion-dominated expansion, n
Ar
n
C
, and the opposite limit corresponding to expansion into
vacuum:
L
D
r
0
_
T
0
T
min
_
1/5
_
n
0
n
Ar
_
1/5
; n
Ar
n
C
L
vac
= r
0
_
T
0
T
min
_
2
.
(7.26)
Let us calculate nowthe cluster size inconditions whendiffusiondominates. The time for N-cluster formationnowequals
to diffusion time:
t
N
t
D
=
L
2
D
D
=
3n
Ar
CAr
L
2
D
v
C
. (7.27)
The carboncarbon collision time reads: t
11
= (n
C
CC
v
C
)
1
; here v
C
is carbon atom velocity and n
C
= n
0
(r
0
/L
D
)
3
as
above. Then the number of atoms in N-cluster in accord with Eq. (7.12) reads:
N

= e
_
t
N
t
2
_
3/4
= C
1
_
n
Ar
CAr
CC
n
0
r
3
0
L
D
_
3/4
. (7.28)
The cluster radius r
cl
= (3N/4n
0
)
1/3
can be expressed nowas a function of the target and gas parameters for the diffusion-
dominated growth process:
r
cl

= C
2
n
1/12
0
r
3/4
0
_
n
Ar
CAr
CC
L
D
_
1/4
; (7.29)
C
1
and C
2
are dimensionless numerical coefficients in Eqs. (7.28)(7.29), which should be extracted from the experiments.
Finally, we present an explicit scaling of the cluster radius in the expansion-limited aggregation conditions in the following
form:
r
cl
C
2
n
2/15
0
r
1/2
0
n
3/10
Ar
(
CAr
CC
)
1/4
_
T
min
T
0
_
1/20
. (7.30)
The experimental conditions at a pressure approaching the atmospheric pressure and with laser intensity in excess of
10
13
W/cm
2
are, however, close to those for the optical breakdown in argon (Mlejnek et al., 1998). Therefore there is a
dependence on the buffer gas pressure hidden in the ablation depth (that enters in r
0
) that we take from the experiments.
The experiments demonstrated that strong decrease of ablation depth occurs at pressure exceeding 100 Torr (see Fig. 7.8
later in this section). Thus, in conditions of diffusion dominated cluster growth the cluster radius scales with the buffer gas
density, the cross-section and the ablation depth as follows:
r
cl
C
2
n
2/15
0
l
1/6
abl
n
3/10
Ar
(
cAr
CC
)
1/4
_
T
min
T
0
_
1/20
. (7.31)
As one can see from Eq. (7.31), the cluster size is a very weak function of the plume temperature and the initial target
density. The main factor is the density of the ambient gas, which affects the size approximately as a cubic root of the gas
pressure.
7.4.2. Growth of nanoclusters in ambient gas and in vacuum
(a) Formation of carbon nanoclusters
Carbon-cluster-assembled nanofoams were produced using a high power frequency doubled Nd : YVO
4
laser system
consisted of an oscillator and power amplifier, generating an average power up to 41 W in 12 ps pulses at a wavelength of
Fig. 7.5. Dependence of the ablation depth of graphite in argon per pulse on the laser fluence at atmospheric pressure. The inter-atomic spacing of 1.46
in graphite is represented by the dashed line. The intersection of a fitting curve with this line determines the ablation threshold fluence of 0.23 J/cm
2
for graphite at atmospheric pressure. The dotted line above the experimental points is the maximum possible ablation rate followed from the energy
conservation law.
532 nmandMHz-range repetitionrate (Luther-Davies et al., 2004). The laser operatedat repetitionrates of 150 kHz, 1.5 MHz,
2.7 MHz, 4.1 MHz, and 28 MHz. The single pulse energy of the laser could be changed in the range from 0.1 J/pulse to
10 J/pulse. The laser beam was focused down to a 15 m spot onto a graphite target placed in a vacuum chamber, which
could be filled with various gases. With the maximum average laser power of 40 W this produced incident intensity
of 1.2 10
12
W/cm
2
with corresponding fluence of 15 J/cm
2
at the repetition rate of 1.5 MHz. The gas pressures in the
experimental chamber were varied from 20 mTorr to 1500 Torr (2 atm.). The mean-free path of hot carbon atoms in argon
at the pressure of 50 mTorr is of the order of 10 cm, larger than the target-to-substrate distance. For this reason carbon
expansion in the chamber at the pressure of <50 mTorr can be considered as expansion in vacuum. The clusters were
produced in various experimental conditions with different type of filling gas (He, Ar, Kr, Xe) and different gas pressure
(Madsen et al., 2006, 2007; Gamaly et al., 2009).
Ablation threshold was defined as a particular laser fluence at which a single atomic layer was removed from the target
surface by a single laser pulse. The amount of ablated material m
avg
was found by averaging the total ablated mass over
the total number of laser pulses. At 1.5 MHz and for ablation lasting 60 s, for example, the total number of laser pulses was
9 10
7
. The ablation depth d
abl
per single pulse is d
abl
= m
avg
/(s
f
), where s
f
is the laser spot area and is the target
material density. The ablation rate as function of laser fluence is presented in Fig. 7.5.
The ablation threshold fluence in the presence of argon at atmospheric pressure was found to be 0.23 J/cm
2
. This value
is significantly lower than the 0.8 J/cm
2
predicted by Eq. (7.18) above, which does not take into account the existence of
ambient gas. This discrepancy can be understood and corrected if the presence of a gas next to the ablation surface is taken
into account (see Section 5). The presence of a gas stimulates contribution of thermal evaporation to the non-equilibrium
ablation reducing the ablation threshold (Gamaly et al., 2005).
The ablation rate in terms of the number of atoms ablated per pulse N
abl
calculated as N
abl
= d
abl
s
foc
n
a
; where s
foc
is the
laser spot area. The ablation rate in Fig. 7.5 for both glassy carbon and graphite targets at 8 J/cm
2
translates to approximately
N
abl
10
13
atoms per pulse. The efficiency of nanocluster formation in a typical experimental laser ablation conditions, i.e. at
(2540) Wlaser power, 1.5 MHz repetition rate, 8 J/cm
2
, 50 Torr of Ar, was of the order (0.51.0) g of nanomaterial per hour
(Rode et al., 2001, 2002a,b).
As the gas pressure is increased, ablation becomes less efficient due to the laser scattering from the laser plume. At
300 Torr the number density of gas atoms is 1.8 10
19
cm
3
that is comparable to the density of carbons in the ablated
plume. However another effect becomes increasingly important in the ablation process with the rising gas pressure, namely,
the optical breakdown of the buffer gas. An estimate of the rise of the electron density in the plume through the avalanche
ionisation and multi-photon ionisation processes shows that it takes about two picoseconds to reach the critical plasma
density at 10
12
W/cm
2
for 532 nm laser (Mlejnek et al., 1998; Rode et al., 2005). The plasma formation in the gas results
in the significant decrease in laser absorption and in the consequent decrease in the ablation rate. The rate measurements
showed the ablation rate decreases almost five times as the pressure increases from 300 to 1500 Torr (Fig. 7.6).
A series of experiments were conducted to investigate the dependence of the cluster size upon various experimental
parameters (Madsen et al., 2007; Gamaly et al., 2009). These parameters included the laser scanning speed, three different
repetition rates of 150 kHz, 1.5 MHz, and 28 MHz; the background pressure variation in experiments with argon in the range
from 20 mTorr to 1500 Torr, and the influence of different filling gases at a fixed pressure of 50 Torr.
Carbon nanoclusters were deposited upon copper grids coated with holey carbon films (hole size 101000 nm), located
approximately 1 cm from the target, for further analysis under transmission electron microscope (TEM)see Fig. 7.7. Short
Fig. 7.6. Ablation rate vs. pressure of argon gas for graphite at the laser fluence 8 J/cm
2
.
Fig. 7.7. TEM images of carbon nanofoam material with the average cluster size of 5.5 nm (a) and a single nanocluster of 10 nm (b).
exposures, less than 10 s, were performed in order to coat the TEM grid to ensure individual clusters could be seen as-
deposited in the web-like arrangement, which arises due to the propagation of clusters through the background gas.
The experiments demonstrated that the average cluster size practically does not depend on the laser beam scanning
speed in the range from 0.05 m/s to 1.0 m/s, and on the laser repetition rate (from 150 kHz to 28 MHz). All measurements
were made at 50 Torr of Ar pressure (Gamaly et al., 2009). These results are in agreement with the theory presented
above. Indeed, the theory suggests that the cluster formation time, which coincides with the plume expansion time, is less
than 10 ns under the experimental conditions under consideration. Therefore, one should not expect the pulse-to-pulse
accumulation effect during the shortest 36 ns time gap between the consecutive pulses at the highest 28 MHz repetition rate
used in the experiments. The observed independence of the cluster size on the scanning speed is the indirect confirmation
of the fact that the cluster formation time is less the 36 ns between the pulses.
In addition, the cluster size distribution was measured at various distances from the target surface, namely, with a
collecting TEM grid located at 1 mm, 3 mm, 5 mm, 10 mm, and 20 mm. The average cluster size remained practically
unchanged within the experimental error of 0.5 nm. These results indicate that the cluster growth process takes place
in close vicinity of the ablation surface and that the average size of the nanoparticles does not change in further diffusion
through the buffer gas.
Analysis of images of cluster-assembled material produced at pressures from0.2 Torr up to 1500 Torr (2 atm) suggests
the following conclusions. Close inspectionof apparently thinsolidlayer depositedona substrate innearly vacuumcondition
(pressure below 20 mTorr) reveals that the film is composed of aggregated clusters similar in size to those observed in the
presence of a background gas (Fig. 7.8(a)). These clusters are visible around the edges of the film where the contrast is
great enough for them to stand out. In Fig. 7.8(b), the pressure has been increased to 200 mTorr and it is clear now that the
presence of a buffer gas is beginning to introduce a more foam-like appearance to the deposited material due to the presence
of collisions as the plume propagates to the substrate. Finally, the image on the right (Fig. 7.8(c)) shows material produced
at 2 Torr displaying the expected structure achieved through deposition of clusters in the presence of a background gas.
Fig. 7.8. TEMimages of carbon nanocluster formed material created at various pressures. Fromleft to right the images depict material created at 20 mTorr
(a), 200 mTorr (b), and 2 Torr (c) respectively. Scale bars are 25 nm in all images.
Fig. 7.9. Cluster size distributions at pressure of Ar from 0.2 Torr to 1500 Torr.
Cluster size distributions were obtained using a series of ten TEM images taken at magnifications of 300k to 400k, and
about 500 100 clusters were measured to generate a cluster size distribution, which was then fitted with a best-fit
Poisson distribution to obtain the average cluster size. The size distributions were measured at pressures from 0.2 Torr
up to 1500 Torr. The size distributions at various pressures presented in Fig. 7.9.
There are two clear effects of the increasing pressure on the cluster size. The average cluster size (maximum of
distribution) grows weakly with the pressure increase while the number of larger clusters in the distribution tail at pressures
in excess of 50 Torr grows faster than the average size. The distribution curve significantly broadens with the pressure
increase while inthe near-vacuumconditions it is narrowpeaked. These measurements allowone obtaining the dependence
of the average cluster size on the gas pressure and compare it with the theory predictions (Fig. 7.10)
Fig. 7.10. Average cluster size dependence on the gas pressure. The error bar is the accuracy of the cluster size measurements of 0.5 nm. A solid curve
is r
cl
n
0.2
Ar
. The laser fluence equals to 8 J/cm
2
.
Fig. 7.11. Average cluster size as function of the atomic radius of the buffer gas at 50 Torr pressure. The dashed line is a square root scaling based on
r
cl
r
1/2
at
.
The effect of the gas pressure on the cluster size became apparent at pressures larger than 50 Torr. The number density of
the buffer gas atoms at such pressures becomes comparable to the density of carbons in the ablated plume, which is of the
order 10
19
cm
3
. The diffusion through the buffer gas dominates the growth process above this pressure. As follows from
Fig. 7.10, the average cluster size increases with pressure r
av
n
0.2
Ar
while the maximum cluster size increases r
max
n
0.3
Ar
in accord with the prediction of Eq. (7.30). However the ablation rate significantly decreases with increase of pressure. As
a result the two competing factors of increased confinement and lower feedstock of ablated material from the target to the
cluster formation zone result in a reduced dependence of the cluster size upon pressure. The influence of the gas pressure on
the cluster growth process in the expanding carbon plume at the pressure range below 50 Torr practically disappears. Only
small change in the average size from 3.2 0.5 nm to 4.4 0.5 nm was observed for the pressure range 0.2 Torr50 Torr.
The theory presented in Section 7.4.1 predicts that the main influence of the buffer gas on the cluster size relates to
increase in the attachment cross section which is directly proportional to the atomic cross section of a buffer gas. This cross
section is proportional to the square of atomic radius, r
2
. The atomic radius of the inert gases increases from 1.40
for He, to 1.88 for Ar, to 2.02 for Kr, and to 2.17 for Xe (Kittel, 1996). The cluster size changes with the cross section
as r
cl

1/4
. Therefore one can estimate the expected cluster radius increase from the formation, for example, in helium
(r
He
= 1.40 ) to the formation in xenon (r
Xe
= 2.17 ) in similar ablation conditions as r
cl
Xe
/r
cl
He
(r
Xe
/r
He
)
1/2
= 1.24. The
experimentally measured ratio r
cl
Xe
/r
cl
He
(r
Xe
/r
He
)
1/2
= 1.29 is close to the theory prediction. The cluster sizes estimated
from other gas radii yields even closer predictions for the experimental values.
The average radius of clusters produced in He, Ar, Xe, and Ne at the pressure of 50 Torr was measured and the dependence
of the average cluster size on atomic radius of the buffer gas is presented in Fig. 7.11.
(b) Formation of silicon nanoclusters in vacuum
Simple kinetic theory developed above should be applicable to formation of nanoclusters fromany material in the mono-
atomic vapour. Therefore it is instructive to apply this theory to the formation of different nanoclusters by single pulses in
vacuum and compare the theory predictions to the experimental data.
Ablation of silicon was thoroughly studied in a number of experiments (see for example Pronko et al., 1998; Lukyanchuk
et al., 1998; Patrone et al., 2000; Marine et al., 2000; Bonse et al., 2002; Gamaly et al., 2002, 2004; Bulgakov et al., 2004;
Amoruso et al., 2004; Tull et al., 2006). The ablation threshold was experimentally measured to be of 0.20.3 J/cm
2
(Bonse
et al., 2002; Gamaly et al., 2004). To take an illustration, silicon nanoclusters with average radii 8 2 nm were formed by
laser ablation experiments in vacuum using 780 nm, 120 fs laser pulses at the repetition rate 1 kHz3 Hz; the laser was
focused to a spot of area S
foc
= 2 10
4
cm
2
yielding the laser fluence of F = 5 J/cm
2
and intensity of 4 10
13
W/cm
2
(Amoruso et al., 2004, 2005a,b). Taking the atomic number density of n
a
= 5 10
22
cm
3
, absorption coefficient A = 0.64;
binding energy of
b
= 3.7 eV, and the first ionisation potential J
i
= 13.6 eV for silicon, one can estimate the ablation depth
with Eq. (7.19) and maximum temperature in the skin layer by Eq. (7.17) in the experimental conditions of Amoruso et al.
(2004, 2005a,b). Then the laser pulse, which delivered 5 J/cm
2
of energy, ablates Silicon to the depth of l
abl
= 3.710
5
cm
and the maximum temperature at the end of the pulse rises to T
0
= 46 eV. Ablation in vacuum has a three-dimensional
character. We suggest, quite arbitrarily, that the minimum silicon cluster formation temperature is 0.1 eV, of the same
order of magnitude as for carbon. Then, with the help of Eq. (7.22) one obtains the 3D-expansion length of 2.4 10
2
cm.
Siliconsilicon cross-section assuming the hard sphere collisions equals to
SiSi
= 3.8 10
16
cm
2
. The average cluster
size can be estimated by Eq. (7.23) taking the unknown dimensionless numerical coefficient equal to unity. One obtains the
average cluster radius of 7 nm in a qualitative agreement with the experimental observations.
Summing up, it has been demonstrated experimentally, and semi-quantitatively described theoretically, that
nanoclusters of several nanometres in diameter can be formed in laser-ablated plume by single laser pulses both in vacuum
and in a buffer gas. The experiments were performed with laser fluences several times larger than the ablation threshold.
Therefore, the laser-produced plume contained mostly fully ionised atoms with small admixture of neutrals and of multiple
ions. The cluster formation occurs by a monomer addition (atom-to-atom, atom-to-cluster collisions) in a space close to the
ablationsurface ina fewtens of nanoseconds time after the pulse end. The cluster size is controlledby expansionof the plume
in vacuum, or by atomic diffusion in the ambient gas. Analysis shows that a single ultra-short pulse (100 fs duration) can
produce clusters withonly several nanometres indiameter, witha very weak dependence onthe laser fluence. Therefore, the
ultra-short pulses canbe usedfor productionof clusters of different materials withcluster size belowthe critical value, which
allows one to separate clusters withnano-properties fromthose of the material inthe bulk. The use of the highrepetitionrate
lasers also provides an opportunity to produce clusters with quite unusual properties. One of the examples is the production
of all-carbon nanofoamassembled from36 nmcarbon clusters, which possess strong paramagnetic properties (Rode et al.,
2004).
7.5. Three-dimensional optical memory
The ability of pinpoint delivery of laser energy into a diffraction-limited volume inside a transparent material opens
up considerable possibilities for three-dimensional (3D) material structuring without altering the surrounding medium
outside the focal volume. Interaction of an ultra-short laser pulse tightly focused into transparent dielectric changes the
refractive index in a micron-sized focal spot in the bulk of the material. The refractive index modification can be reversible
or permanent, depending on the laser intensity level and the dielectric properties.
The modified refractive index retains for a long time and thus can be used as a memory bit. The difference of refractive
index in laser-affected spot from that in the pristine material can be afterwards detected, or read, by refraction of a low-
intensity probe beam from such spots. Finally, the optical properties in laser-affected spots can be reversed back to those
of pristine crystal by the action of a specially adopted erasing laser, which smoothens the refractive index gradients over
the irradiated volume. The information can be written in many layers, allowing the density storage beyond 10
12
bits per
cm
3
(Tb/cm
3
).
The ultrafast laser-induced approach opens routes for development of 3D writereaderase data storage in
photorefractive or polymer media (Kawata et al., 2001; Day and Gu, 2002; Juodkazis et al., 2006a,b,c; Gamaly et al., 2010))
and for permanent optical memory through the formation of voids in the material (Glezer et al., 1996; Glezer and Mazur,
1997; Watanabe et al., 1998; Juodkazis et al., 2003; Schaffer et al., 2004). The ultrafast laser pulses tightly focused inside
the crystal volume also provide the means to form complicated 3D structures for applications in photonics: waveguides,
photonic crystals, diffractive gratings etc. Several experimental results are presented below to demonstrate the ability to
write, read, and erase information using single femtosecond pulses. These experiments set a base for possible application of
single fs pulses for development of optical patterning, encryption, and data storage.
7.5.1. Recording in photorefractive media
Juodkazis et al. (2006a,b,c), reported the formation of a three-dimensional set of laser-affected regions inside the
photorefractive crystal of LiNbO
3
doped with Fe (400 ppm) by 150 fs laser pulses at 800 nm wavelength. The single laser
pulse produces a spot of diameter of 1.1 m with the relative refractive index modulation of 10
3
in comparison to that
of a pristine crystal. The essential differences of fs laser-induced reversible modification of the optical properties fromthose
produced by long pulse and CW lasers are two fold. First, the writing speed with high-repetition rate lasers is rather fast,
potentially exceeding 10
9
bits per second. Second, the average intensity per pulse is very high, 10
12
W/cm
2
, which is
close to the dielectric breakdown threshold. Therefore the pulse energy span for producing the reversible changes has been
found in a range of 3.8 nJ12 nJ energy per 150 fs pulse tightly focused with high numerical aperture (NA = 1.4) optics. At
the energy 1112 nJ/pulse the permanent structural damage was observed. Image patterns with these memory bits were
produced in layers 3 m apart at the depth up to 50 m from the surface.
7.5.2. Reading, erasing, re-writing
Optical transmission was used to observe and digitise laser-modified regions. The diffraction pattern allows
distinguishing the bits separated by 0.5 m in plane. The recording of the sequence of the memory bits occurs along the c-
axis of a LiNbO
3
crystal. It is possible to produce a 3D stacking of planes each with the bits separated by 0.5 m. The authors
Juodkazis et al. (2006a,b,c) recorded the bits in planes separated by 3 m. Thus it is possible in principle to have a memory
with capacity of one bit per m
3
(Tb/cm
3
).
A complete erasing of volume memory element (voxel) was achieved by irradiation of low intensity laser beam shifted
along the c-axis. Such an erasing leaves a recognisable feature in the beginning and at the end of the recorded line. Thus
it was difficult to remove of a single voxel without damaging the neighbouring dots. Different well recognised patterns of
voxels were recorded after erasing the previous writings. A good quality of a record was obtained after more than 20 cycles
of re-writing (Juodkazis et al., 2006a,b,c). The memory bits have different lifetime in Fe-doped LiNbO
3
and undoped LiNbO
3
.
Full recovery of the laser-perturbed properties in undoped LiNbO
3
occurred in 0.2503 s after the pulse, while femtosecond
laser-induced changes in Fe-doped LiNbO
3
could retain as long as in the index modifications produced by long exposition
by cw low-intensity lasers (Kukhtarev et al., 1979, 1984; Sturman and Fridkin, 1992; Buse et al., 1998). The stability of the
recorded structures is limited by the dark current and photo-conductivity. There are also methods for so-called fixing the
refractive index changes, which allows one to storage information for months and probably years. Several refractive index
fixing techniques have been developed with the best results achieved by thermal fixing.
Recent development of optical memory recording in plasmonic gold nanorods based media offers an unparallel approach
of integrating polarisation and wavelength sensitivity into the 3D optical memory and expanding optical memory into a
five-dimensional domain (Zijlstra et al., 2009). The new approach makes use of photothermal reshaping of gold nanorods
immersed in a polymer layer mediated by surface plasmon resonance. Combined with the high cross-section of two-photon
luminescence, incorporation of two polarisation and three wavelength channels enabled non-destructive, crosstalk-free
readout of memory recording and reading with a bit density of 1.1 Tb/cm
3
. Combined with a recording speed of Gb/s
and supercontinuum light generation high repetition rate source, this five-dimensional optical memory has a potential to
increase storage capacity by several orders of magnitude.
7.6. Three-dimensional networks
Microfabrication of 3D periodic dielectric structures with sub-micron features is a promising technology for fabrication
of photonic crystals and metamaterials, which are expected to have an impact across the entire range of technologies where
electromagnetic radiation is used due to their unique capability of controlling light propagation via photonic band-gap
effect (Misawa and Juodkazis, 2006; Zheludev, 2010). Ultrafast laser based 3D microfabrication already demonstrated its
flexibility, effectiveness, and sub-micron precision in photonic crystals, micro-electro-mechanical systems (MEMS), lab-on-
a-chip, microfluidic devices, and biomedical applications (Toader and John, 2001; Kawata et al., 2001; Korte et al., 2003;
Serbin et al., 2003; Sugioka et al., 2005; Juodkazis et al., 2009; Farsari and Chichkov, 2009; Farsari et al., 2010; Schizas et al.,
2010).
Most prominent results achieved so far are based upon two-photon polymerisation in photosensitive materials such as
photoresists, under intense irradiation of femtosecond pulses. Two-photon absorption in laser irradiated photoresists acts
similar to a hardener in liquid resins, it causes the polymer chains to cross-link and form a polymerised rigid material, thus
offering anopportunity for direct laser writing of structures. The unexposedareas canbe washedout fromthe laser-modified
framework by a solvent. This type of process is being used with most commonly used negative photopolymers, such as the
UVlithographic photoresist SU-8 and the photosensitive solgel organic/inorganic hybrid material Ormoser (Houbertz et al.,
2003). In positive photoresists laser irradiation breaks polymer chains to become soluble, so inverse structures can be built
in the sample.
The optical setup for 3D direct laser writing is similar to the point-to-point sequence of laser exposures described in
Section 6 for microexplosion experiments (see Fig. 6.1), the main difference is the reduced average laser power to nJ per
pulse level and the increased pulse repetition rate to MHz range for building continuous features like woodpile and or spiral
structures (Sun et al., 2001a,b; Toader and John, 2001) and 3D micromodels (Kawata et al., 2001; Serbin et al., 2003).
The direct writing of periodic structures can be significantly simplified and accelerated by illuminating photoresist to
periodic field patterns generated by interference of multiple laser beams. An entire periodic structure can be recorded at
once (Kondo et al., 2001, 2003), and the symmetries of all fourteen Bravais point lattices can be generated by the interference
of four beams (Cai et al., 2002). Moreover, the interference regions in the pattern formed by diffractive beam splitters with
femtosecond laser pulses are limited by the temporal pulse width and not affected by the beam size (Nakata et al., 2002).
Careful care, however, is requiredto preserve the connectivity inthe self-supporting structure (Misawa andJuodkazis, 2006).
Solgel technology provides a very powerful tool for the development of photosensitive compounds. These materials
benefit from straightforward preparation, modification and processing and, in combination with their high optical quality,
post-processing chemical and electrochemical inertness, good mechanical and chemical stability. Nanoscale features with
line widths as small as 80 nm and 65 nm have been formed reproducibly by the photoinduced polymerisation using
520 nm 100 fs pulsed excitation (Haske et al., 2007). The fabrication of nanoscale lines via visible wavelength multi-photon
polymerisation has been utilised to fabricate initial polymeric photonic crystal structures with lateral periods of 500 nm,
which gave stop bands in the near infrared spectral region.
7.7. Creation of a super-dense material in the confined micro-explosion
7.7.1. Progress in formation of new super-dense materials
We discussed in Section 6 that femtosecond laser produced micro-explosion confined deep inside of the bulk solid
generates pressure in excess of 10 TPa and temperature of 10
5
K (Juodkazis et al., 2006a,b,c; Gamaly et al., 2006). These
conditions are beneficial for formation of new super-dense, super-hard, and super-strong materials. The uniqueness of
these conditions apart fromultra-high pressuretemperature relates to the record high heating and cooling rates and to the
confinement of the transformed material inside a pristine crystal. The studies of materials transformed under such unique
conditions are at the very beginning. In what follows we summarise briefly the theoretical and experimental search for the
optimum paths to super-dense, hard and strong materials and compare to those promising by confined micro-explosion.
First, one should note that such material properties as high density, hardness and strength (modulus) are interconnected
andalmost always going together. Hardness for ideal systems onmicroscopic level is determinedby the bulk modulus, which
in turn depends on the nature of chemical bonding. In the real materials the point and macroscopic defects, dislocations,
disclinations, and impurities can limit the hardness. The largest bulk moduli are found in covalently bonded materials
(diamond4.43 Mbar). For covalent solids, such as C, Si, and zinc blendes the bulk modulus depends on the bond length
and on dimensionless parameter
i
, the measure of the ionicity. Last parameter accounts for the depletion of the bond
charge with increasing ionicity. The following scaling for the bulk modulus has been proposed by Cohen (1985) and Liu and
Cohen (1989):
B =
19.71 2.20
i
d
3.5
; (7.32)
where the bulk modulus is in Mbars, the bond length d is in angstroms, the ionicity,
i
, changes in a range 02. For the
homopolar semi-conductors of groupIV(C, Si, Ge)
i
= 0. Taking tetrahedral covalent bondlengthfor diamondas d = 1.54
one gets from the above scaling B = 4.35 Mbar, which is within a few percents in agreement with the experimental figure.
Already discovered super-dense, super-hard, super-strong materials are tetrahedrally bonded covalent solids C, SiO
2
,
SiN, CN, BN with coordination number higher than that in the low-density phase. The known examples are diamond and
stishovite, a high-density phase of silica (SiO
2
) with coordination number 6; mass density 4.29 g/cm
3
; bulk modulus
281313 GPa; hardness H = 33 GPa (Leger et al., 1996). Many post-stishovite polymorphs were experimentally found. It
has been postulated (Teter et al., 1998) that virtually infinite number of stishovite polymorphs may exist and be metastable
as they are all derived through slight variations of a common parent structure. Thus silica is the natural candidate for the
studies of phase transformations under micro-explosion.
The application of high pressure and temperature may induce a phase transition to a structure with more tetrahedral
bonding. The transition would be similar to the transition of graphite to diamond or silica to stishovite.
Hardest oxides were discovered under pressure up to 20 GPa created by multi-anvil apparatus (Leger et al., 1996)
and heating up to 1100 C. By this method cBN and stishovite were produced. The research group of Zerr et al. (1999)
reported formation of the new phases of c-Si
3
N
4
(-Si
3
N
4
, 3.183 g/cm
3
; -Si
3
N
4
, 3.20 g/cm
3
) at pressure above 15 GPa
and temperature exceeding 2000 K. The samples were heated by YAG or CO
2
lasers at constant pressure for 110 min and
then quenched during the seconds after switching off laser power. The authors (Zerr et al., 1999) claim that more than 90%
of material was converted to the dense phase.
Conversion of the silica glass into the super-dense stishovite phase solely by the action of the powerful laser on the silica
surface was reported by Salleo et al. (2003). The authors demonstrated through combination of electron diffraction and
infrared reflectance measurements that silica transforms partially into defective form of the high-pressure stishovite phase
under high intensity (GW/cm
2
) laser radiation. At the pressure above 7 Gpa the stable polymorph is stishovite where Si is 6
fold-coordinated. Correspondingly, at 1020 Gpa the coordination in glasses increases from 4 to 6. The dense material was
found at the fringe of the crater produced by a laser pulse on a fused silica surface. The material surrounding the crater is a
uniform aggregate of small rounded particles of 10 nm or smaller.
It is instructive to compare the preliminary results obtained in studies of materials transformed by micro-explosions
(Juodkazis et al., 2006a,b,c; Gamaly et al., 2006) to the super-dense materials produced by the diamond anvil cell or
by laser ablation on the surface. As a result of micro-explosion all the compressed material remains in the form of a
shell surrounding a central void. Therefore the direct measurement of density of compressed material based on the mass
conservation =
0
, with the average compression ratio of = 1.11.14. The simplest interpretation is to suggest that
shell consists from the mixture of the material of initial phase with density
0
, and the compressed phase with density
D
,
giving the average density as
0
=
D
+ (1 )
0
. Then the concentration of high-density (
D
) phase expresses as:
= ( 1)
0
/(
D
0
). In the case of compressed shell in silica the density increase most probably can be associated with
Fig. 7.12. (a) X-ray diffractionrings acquiredinthe centre of laser-affectedarea; (b) the same inthe pristine area; (c) comparisonbetweenthe experimental
peaks (black dots), obtained from the experimental data shown in (a) by integration over the radius, and peaks theoretically predicted for bcc-Al (green
lines). The crystalline hkl plane indices are shown next to the peaks. The calculated peaks for sapphire (grey lines) overlapped with peaks of conventional
fcc-Al are also shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Source: Taken from Vailionis et al. (2011).
formation of stishovite with density
D
= 4.3 g/cm
3
(initial silica density
o
= 2.24 g/cm
3
) because the conditions for its
formationwere definitely achieved andovercame. The concentrationof dense phase constitutes 5%7%, most probably inthe
form of nano-crystallites as it was suggested by Salleo et al. (2003). The amount of laser-affected material in a compressed
shell is around 0.1 pg (10
13
g). The small amount of high pressuretemperature affected material is a major obstacle
in identification of new phase with micro-Raman, X-ray beams, SEM, AFM and optical microscope studies. However, the
significant progress has been made recently with the experimental evidence of formation of novel crystal by confined micro-
explosion, which is presented in the following paragraph.
7.7.2. Evidence of super-dense aluminium produced in confined micro-explosion
The formation of a new stable high-pressure phase of aluminium, bcc-Al, by ultrafast laser induced micro-explosion
confined inside sapphire, -Al
2
O
3
, has been reported recently (Vailionis et al., 2011). The new phase was found preserved
inside the amorphous shock-compressed sapphire shell produced by laser-induced micro-explosion, which is described in
details in Section 6. Synchrotron X-ray diffraction (XRD) microanalysis first revealed that there are pronounced diffraction
rings in the laser-modified regions, while unaffected area shows only pristine sapphire. Then, it was found that the observed
diffraction peaks in the sample perfectly match the bcc-Al structures predicted to exist at high pressures but which
were not yet experimentally observed (Boettger and Trickey, 1996; Friedli and Ashcroft, 1975; Moriarty and McMahan,
1982; Pickard and Needs, 2010). The bcc-Al (space group Im3m) is high-pressure Al phase with a lattice constant a =
2.864 (11.75
3
/atom), this phase is 41% more dense than the conventional fcc-phase of Al. One can see at Fig. 7.12 the
comparison of the diffraction peaks in pristine sapphire, conventional fcc-Al, predicted peaks for bcc-Al and experimental
points fromXRD microanalysis. The figure illustrates a perfect match between seven experimental XRD peaks with those of
predicted bcc structure. The widths of the diffraction peaks were used for estimation the size of new phase crystallites. The
best fit of all seven XRD peaks is achieved for a crystallite size of 18 2 nm. The authors suggested that the sapphire
constituents, aluminium and oxygen, become spatially separated in the process of shock wave expansion due to the
differences in the ion masses and consequently in atomic and diffusion velocity. A new bcc-Al phase is frozen within
the compressed sapphire shell after ultra-fast quenching.
In these experiments the laser pulse of 150 fs, 800 nm, was tightly focused inside sapphire crystal (at depths from
5 m to 20 m) with energy in the focal spot, 135 6 nJ. Laser radiation of the sapphire caused the formation of void
surrounded by compressed shell at focal spot within the crystal that is explained in details in Section 6. The size of the voids
and the compressed shell thickness were measured using transmission (TEM) and scanning (SEM) electron microscopes
after opening the enclosed laser-affected zone by focused ion-beam milling. The average radius of the combined void and
shell structures was R
am
360 nm, the void radius was, R
void
180 nm. The estimates of the absorbed energy density
and maximum pressure in the energy deposition region, responsible for conditions for formation a new phase, are made
based solely on the experimental data. Sapphire is one of the hardest materials in Nature. Therefore the formation of the
void is in itself the experimental evidence for creation of pressure in excess of the Young modulus of sapphire, which from
different sources is in a range of 239435 GPa. Suggesting that shock wave dissipated all the energy on the work against cold
pressure in sapphire when achieving the experimentally determined boundary between pristine and amorphous material,
R
am
, and taking conservatively Y
sapp
= 300 Gpa one obtains, E
abs
Y
sapp
4R
3
am
/3 = 58.6 nJ. This estimate reasonably
suggests that only 43% of laser pulse energy were absorbed. The conservative estimate for the maximum pressure suggests
that the absorption zone cannot be larger then the void, therefore the absorbed energy per void volume gives the pressure
P
max
E
abs
/4R
3
void
/3 2.4 TPa.
The observed formation of crystalline aluminiumout of aluminiumoxide is experimental evidence of a quite unusual and
unexpected phenomenon, namely, spatial separation of oxygen and aluminium ions in conditions of complete confinement
and preserved stoichiometry. The explanation of this effect presented by Vailionis et al. (2011) relates to the conditions in
high temperature/density non-equilibrium plasma created by micro-explosion. In such plasma the Coulomb interactions
dominate the ion transport. Therefore at the same temperature (energy per ion) the light particle moves faster than heavy
one andthus during the short time of hightemperature existence the ions move at the distance of several tens of nanometres.
Following fast cooling then allows forming a new crystal and preserving it in a new state after release back to the ambient
conditions.
7.8. Application of fs-lasers in art restoration
The use of laser technology for cleaning heritage artefacts is gaining increasing interest due to a number of advantages:
laser cleaning is a dry and contact-free process, which selectively removes contaminating dirt or coatings, including
hazardous contaminants such as radioactive and biological material, minimises mechanical and chemical disruption of
historic surfaces and generates minimal waste (Zafiropulos and Fotakis, 1998; Lukyanchuk, 2002; Delaporte et al., 2003).
Until now, lasers used to perform conservation tasks have had pulse durations of nanoseconds or longer, and parameters
such as wavelength, intensity and pulse length have had to be carefully chosen to tailor the laser to different materials and
tasks. However, the disadvantage of currently used lasers for conservation is that their relatively long pulse duration allows
heat and shock waves to travel into the substrate, potentially causing damage to historic surfaces. The ablation process tends
to remove chunks of material from the surface, because of shock-wave formation and surface heating. Moreover, the high
energy of ultra-violet (UV) excimer lasers can also induce photochemical effects and cause deterioration of the pigment or
binder in the paint, with both immediate and long-term consequences for the artwork (Lukyanchuk and Zafiropulos, 2002;
Zafiropulos et al., 2003).
The ultrafast lasers are too short to allow heat or shock waves to travel into the substrate. They ablate very thin surface
layers, this allows highly controlled cleaning, does not affect the underlayer material, and can be combined with real-time
monitoring with a system such as laser-induced breakdown spectroscopy to analyse the ablated products and prevent
damage to uncontaminated layers under the surface of an object (Anglos et al., 1997).
The first attempts to use femtosecond laser pulses for conservation were performed with paintings (Zafiropulos et al.,
2003; Pouli et al., 2005). Aged varnish was successfully ablated with 500 fs laser pulses, removing a layer of controlled
thickness. It was shown that it is possible to adjust the depth of each layer removal by choosing the right laser fluence and
number of laser pulses. Femtosecond pulses appeared to be better than longer pulses in terms of the spatial confinement of
the beam and etching resolutionprecision could be achieved down to 0.02 m per pulse (Rode et al., 2008). As a result of
the different physical processes induced by ultra-short laser pulses, the cleaning efficiency is less sensitive to the chemical
composition of the contaminants than has been observed with nanosecond lasers. Moreover, working with lowpulse energy
and at high repetition rate allows the removal of a very thin layer of contaminant with each pass. This provides better control
of the process and reduces the dot effect, producing a smoother final surface.
Femtosecond pulses have also been successfully used for cleaning varnish from painted artefacts, showing very precise
spatial resolution (only the desired material was removed) and minimal photochemical changes to the underlying surface
(Andreotti et al., 2005). Femtosecond lasers have also been successfully tested on bronze objects with respectively indoor
and outdoor patina; by adjusting the laser parameters, both the outer layers of soot and organic materials and the inner
layers of copper corrosion products, were able to be selectively removed with the same laser system(Burmester et al., 2005).
All these results suggest that the use of ultrafast lasers provides a viable approach to successful laser cleaning of heritage
artefacts (Pouli et al., 2009; Gaspard et al., 2008; Rode et al., 2008).
In spite of obvious advantages of using ultrafast lasers for conservation of artworks and heritage objects, there are still
some issues to be addressed. One of the major challenges is related to the laser pulse delivery to the treated surfaces, which
in many cases are not flat surfaces. Optical fibres and mobile handpieces allow great flexibility of use, facilitating both
treatment of large surface areas and precision work on extremely minute details. Nevertheless, the use of optical fibre for
ultra-short pulse delivery is still a non-trivial problem, because nonlinear processes in the silica fibres stretch the pulse
duration. Moreover, high laser intensities can damage the fibre core. However, there are a number of reports on successful
optical fibre delivery of intense laser pulses, for example in endoscopic medical applications using flexible hollow fibres
(Takeda et al., 2005; Matsuura et al., 2002) and gradient index multimode fibres (Kamino et al., 2003). It was demonstrated
that using photonic bandgap fibres, megawatt microjoule femtosecond pulses could be delivered over a 3 m long distance
(Ouounov et al., 2003), making a number of previously impossible applications quite achievable (Luan et al., 2004).
With the rapid development of powerful and compact femtosecond lasers, ultrafast laser ablation has the potential to
become a standard tool in the conservation armoury and a key technique for conserving previously untreatable artworks
and heritage objects.
8. Conclusion remarks and future directions
Summing up the major results of the studies covered by this reviewwe conclude that the major concepts of lasermatter
interaction at non-relativistic intensities with sub-picosecond pulse duration shorter than the main relaxation times are
well established, and in full agreement with the previous experimental and theoretical studies. Studies of ultra-fast laser
excitation of coherent phonons, laser superheating without melting, and laser ablation rather than equilibriumevaporation
conducted in the last two decades have all demonstrated significant progress in understanding and control of ultra-fast
laser-induced processes, from a delicate atomic motion to complete ionisation and ablation of the material.
The major concepts presented in this review describe the time sequence of the most important phenomena during and
after the interactionof ultra-short laser pulses withmatter. The prime process is the absorptionof the laser light by electrons,
while atoms in a lattice remain at their initial unperturbed state during the pulse. The distinctive characteristic of ultra-short
lasermatter interactionis that the local statistical distributions inthe electronand inthe lattice sub-systems are established
in two stages. In the first stage, both sub-systems have different temperatures, and then the exchange of momentum and
energy via electronphonon interaction with temperature-dependent rates gradually leads to equilibration of temperatures.
The longest of all the relaxation phenomena is the process of building up the high-energy tail in the MaxwellBoltzmann
distribution. These concepts are successively embedded into the qualitative and quantitative description of the sequence
of major phenomena in this review. They form a sound basis for future studies in search of a deeper understanding of
phenomena in shorter time and smaller space scales.
Ultra-fast lasers enable many newapplications inmedicine and biomedical applications (corrective eye surgery, two- and
three-photon imaging), precision micro-processing of materials, and all-optical signal processing in telecommunications,
which can only work with ultrafast pulses (Keller, 2003; Gattass and Mazur, 2008; Chung and Mazur, 2009). The general
aimof future studies might be formulated as to achieve the ability to generate, observe and control material transformation
phenomena at the atomic space and time scales, i.e. at angstrom and sub-femtosecond resolution. Clearly, further progress
towards this depends on progress in the development of the ultrafast lasers, sophisticated diagnostic tools, and in the further
improvement of theory. We briefly outline some emerging problems and areas of research, which are expected to shape the
field of ultra-short lasermatter interaction in the coming years.
8.1. Ultra-short lasers
8.1.1. Solid-state laser systems
Rapid progress in ultra-fast lasers is expected in further shortening of the pulse length, broadening of the spectral range
of available photon energy into deep infra-red, UV, and X-ray spectral ranges, and scaling the energy per pulse above the
100 J level with increased repetition rate into the GHz domain, subsequently increasing the average power to the kWlevel
and above (Keller, 2003, 2010a). The invention of Chirped Pulse Amplification technology (CPA and OPCPA) (Maine et al.,
1988), creation of Kerr-lens mode-locking (Spence et al., 1991) and development of the semiconductor saturable absorber
mirror (SESAM) (Keller et al., 1992) made it possible to produce pulses of a fewlight cycles with intense Ti:sapphire systems
(Sutter et al., 1999; Ell et al., 2001). The average output power has been pushed beyond the 100 Wlevel for the first time with
a SESAM modelocked Yb : Lu
2
O
3
thin-disc laser oscillator generating 141 W, with 738 fs pulses, at 60 MHz repetition rate
(Baer et al., 2010; Keller, 2010b), and the 100 GHz repetition rate barrier was overcome with 1.5 m 1.6 ps pulses (Oehler
et al., 2008). The adaptive optics and spatial light modulators make achievable the flat-top-hat intensity distribution over
the focal spot area, and assures that the focal spot size is close to the diffraction limit. Well-determined pulse duration,
with a controlled spatial and temporal intensity profile makes it an excellent tool for applications and for studies of the
transformations of the material at the time level of sub-femtoseconds time and space levels of nanometres.
8.1.2. Attosecond lasers
While atomic and molecular vibrations occur on a time-scale of tens to hundreds of femtoseconds, the motion of
individual electrons in molecular orbitals and the inner shells of atoms occurs in the sub-femtosecond time domain. The
characteristic length scale is comparable to interatomic distance, and the period of plasma oscillations is of the order of 0.1 fs.
In situ observation of a nanometre-scale excited area with spatial and temporal resolution is necessary for understanding
and control over the atomic excitation and ionisation processes. Generation of intense EUV laser pulses of attosecond-range
pulse duration (1 as = 10
18
s) approaching single optical cycle limit with high signal-to-noise ratio will undoubtedly be
available as a pumpprobe tool in the near future (Bandulet et al., 2004; Corkum and Krausz, 2007; Keller, 2010b; Chang
and Corkum, 2010). Real-time observation of valence electron motion was already achieved with 24 as resolution using a
streaking technique with XUV (80 eV), 80 as pulses (Gouleimakis et al., 2010). Attosecond pulse interferometry makes
it possible to directly observe chemical reactions in real time (Woerner et al., 2010). Recent developments in powerful
femtosecond fibre laser amplifier systems are expected to break the 1 kW power barrier (Eidam et al., 2010), and the first
MHz system generating high-order harmonics has already been demonstrated (Boullet et al., 2009). The development of
high-average power, high repetition-rate, compact, tabletop ultrafast VUV/XUV sources will enable new measurements in
photoelectron imaging spectroscopy, surface science, metrology, and biology.
One may expect that attosecond VUV/XUV pulses will become viable diagnostic probes to study the excitation and post-
excitation processes in solids. We refer our readers to an excellent review (Krausz and Ivanov, 2009) for a comprehensive
introduction to the exciting newarea of research on attosecond physics, which emerged at the beginning of the millennium.
8.1.3. Free electron lasers
Relativistic electron beams in free electron lasers (FEL) act as an active mediumwhere free-to-free energy transition leads
to self-amplified spontaneous emission. The beam is passed through a periodic array of magnets the undulator and the
electrons, while wiggling in the magnetic field, start emitting photons. A resonant frequency occurs due to the difference
between the geometric path length of wiggling electrons and a straight-line radiation path, and due to the difference in
speed of relativistic electrons and their radiation, the electrons are bunched at a radiation wavelength down to 1 , which
results in increased coherency of their radiation. For this reason FEL does not need a resonator cavity, unlike a conventional
laser; and it is particularly well suited to generating short-wavelength X-rays.
Recent progress in the development of X-ray FEL has demonstrated the capability to produce high-brightness X-ray
pulses of femtosecond duration, with the potential to achieve attosecond durations (Popmintchev et al., 2010; Nugent and
Barletta, 2010). With femtosecond pulses and peak brightness that exceeds the brightest synchrotron sources by ten orders
of magnitude, the FEL naturally lends itself to time-resolved studies on the nanoscale (Galayda et al., 2010; Sakdinawat and
Attwood, 2010; Chapman and Nugent, 2010). Many scientific studies will benefit from this new capability, such as phase
transitions in magnetic materials, electron dynamics in chemical systems, which occur on femto- and attosecond time-
scales, and studies of correlated solid-state phenomena, which require angular-resolved photoemission with both spatial
and temporal resolution.
8.1.4. Sculptured laser beams
Intensity distribution across the conventional laser beam has a Gauss-like form. In 1974, Nye and Berry considered basic
properties of dislocations in wavefronts and introduced a so-called vortex beam, where the light propagates with its phase
twisted around the beam axis (Nye and Berry, 1974; Allen et al., 2003). The phase on the axis of such a twisted beam is
indeterminable (it has a singularity), which implies that the intensity at the axis is equal to zero. The intensity distribution
across the beam has a doughnut-like shape with a dark hole in the centre. The spiral phase of vortex beams carries spin
and orbital angular momentum, which introduce torque on an electric dipole and can be observed in the orbiting motion of
trapped particles in optical tweezers (He et al., 1995; Friese et al., 1998; ONeil et al., 2002). Recent development of efficient
beamconverters generating powerful femtosecond optical vortex pulses (Shvedov et al., 2010) offers newopportunities for
studies of lasermatter interaction with such beams and formation of specific micro- and nano-structures.
8.2. Theory and computer modelling
The ultra-short lasermatter interaction with a solid surface generates a chain of interconnected non-equilibrium
processes, which in turn change the transient optical properties of laser-affected material, thus influencing the interaction
mode. Comprehensive computer modelling of these processes should include the set of the Maxwell equations coupled with
the material equations, as a compulsory element (suchas kinetic equations, those for the density matrix, molecular dynamics
simulation etc.). Detailed studies of time evolution in distribution functions from the non-equilibrium excited state to their
equilibrium form are required for understanding and control over the interconnected phenomena in the ultra-fast laser-
induced transformations of the material. It should be stressed that at high intensity the ultra-short pulse removes (ablates)
a material in completely non-equilibrium conditions, where traditional hydrodynamics is inapplicable. A good example of
self-consistent formulation of complex multi-disciplinary problems for computer simulation is the code LASNEX developed
for laser-fusion problems (Lindl et al., 2004).
There are problems related to light beams tightly focused deep within a transparent solid, which need further theoretical
and computational studies. One of the first relates to propagation and focusing of a beam with intensity close to a self-
focusing threshold. Another problem is the influence of the transient changes in optical properties and plasma expansion
on the laserplasma interaction process similar to inter-relation in lasersurface interaction mentioned above.
The problem of shock-wave generation by swiftly excited electrons in hot, short-lived, and multi-component solid-
density plasma created by a confined micro-explosion inside a solid appears to be quite different from conventional
hydrodynamic shocks. Experiments revealed that unexpected spatial separation of ions with different masses occurs during
the short time of shock-wave generation and decay (Vailionis et al., 2011; Juodkazis et al., 2010). This is evidence of the
complicated structure of the shock-wave front smeared by the ion diffusion, which evolves during the shock propagation.
Thus, it seems that kinetic calculations of electrons coupling with different ion species are necessary in order to understand
the complicated structure of the emerging shock wave.
8.3. Controlled excitation of phonons and coherent sound waves
The excitation of phonon modes in different solids along with observation of their evolution in time until decay has been
performed in many laboratories. However, selective and controlled excitation of the desirable phonon mode from 3n of all
available modes (n is the number of atoms in the primitive cell) remains difficult. If it were possible to excite and control a
specific phonon mode, then a range of possibilities would look viable. We mention some suggestions based on experiments
performed, and on theory.
With the help of molecular dynamics simulations, Dumitrica et al. (2004), found that selective excitation of specific
phonon modes by femtosecond laser could lead to the opening of the carbon nanotube cap. Ultra-fast bond-weakening and
simultaneous excitation of two coherent phonon modes of different frequencies localised in the spherical cap and cylindrical
body of a carbon nanotube might be responsible for the opening of the cap in non-equilibrium conditions.
Stimulated emission of terahertz phonons in a super-lattice under vertical electron transport has been observed by Kent
et al. (2006). The authors suggested that such a super-lattice might form the basis of a SASER device (Sound Amplification
by Stimulated Emission of Radiation)creating a directed flow of coherent phonons and generating coherent sound waves,
equivalent to a laser generating electromagnetic waves. One may conjecture that excitation of resonant phonons by short
pulse laser may also lead to a similar device.
Electronphonon interaction and electron and phonon spectra in general are responsible for the conductivity and optical
properties of solids. There were experiments using two pump pulses, which excited the same phonon mode. It is possible, in
principle, to excite different phonon modes simultaneously by multiple pumps thus creating a particular phonon spectrum.
Hence, there remains the challenge of pursuing changes in conductivity and optical properties by selective excitation of a
particular phonon spectrum.
8.4. Control over chemical reaction-rates and transient phase states
Transient phases produced by ultra-short pulses were observed in different laser-excited materials and some were
identified. For example it was found in Collet et al. (2003), that a 300 fs laser pulse transforms charge-transfer molecular
material from a paraelectric to a ferroelectric state, which was completed in 500 ps after the pulse. X-ray probing revealed
macroscopic ferroelectric reorganisation with long-range structural order to be fully established. The recovery of the
equilibrium state occurs in one millisecond.
It was found that ultra-short pulses produce transient phase transformations in Gallium (Uteza et al., 2004) and in
Bismuth (Boschetto et al., 2008a,b) at the absorbed energy-density in excess of the equilibrium enthalpy of melting into
a state different from the equilibrium liquid of a corresponding solid, or any other known phases of these materials. The
phase transitions are reversible: solids recovered to their initial state in several nanoseconds after the excitation. However,
the identification of these transient states is still elusive.
There are some results, which suggest that control over phonon modes and their coupling will allow one to control the
chemical reaction rates (Zewail, 2000). Hillebrand et al. (2002), found the strong enhancement of the near-field resonant
coupling in the infrared by phonons in polar dielectrics (SiC). This coupling proved to be very sensitive to the chemical and
structural composition of samples, permitting nanometre-scale analysis of minerals and semiconductors.
When a femtosecond pulse shorter than any phonon period hits a crystal, non-equilibrium phonons and a lattice
distortion are induced. Williams (1992) suggested that the controlled distortion would then initiate metal-insulation
transitionand anaccompanying charge-density wave. The challenge for future studies is to identify the newtransient phases
and find methods to preserve them in a meta-stable state.
8.5. Warm Dense Matter on the tabletop: mimicking conditions in stars and creating new materials
It was demonstrated that tight focusing of a conventional tabletop laser inside the bulk of a transparent solid creates
pressure, exceeding strength of any material, and the shock wave compresses a solid, which afterwards remains confined
inside a crystal. High pressure and temperature are necessary to produce super-dense, super-hard and super-strong phases
or materials, which may possess other unusual properties, such as ionic conductivity. In nature, such conditions are created
in the cores of planets and stars (see for example, Drake (2006)). Extreme pressures were created by strong explosions, by
diamond anvil presses and with powerful ns-lasers (Drake, 2010). All these methods are cumbersome and expensive. By
contrast, ultra-short lasers create extreme pressure and temperature along with record high heating and cooling rates by
focusing 100200 nJ of conventional femtosecond laser pulse into a sub-micron volume confined inside a solid (Section 5).
Recently, the novel crystalline phase of aluminium, bcc-Al has been discovered in ultra-fast laser-induced micro-explosions
(Vailionis et al., 2011). These results open the possibilities for formation of new high-pressure phases and prospects of
modelling inthe laboratory the conditions inthe cores of planets and macro-explosions. The first results might be considered
a proof-of-principles step. However, it is obvious that, with this simple and inexpensive method for creation of extreme
pressure/temperature, the focus in research is shifted to post-mortem diagnosis of shock-compressed material. Micro-
Raman, X-ray and electron diffraction, and AFM and STM studies with resolution on the sub-micron level are needed for
identification of the new phases. Another challenge is to develop a pumpprobe technique with time resolution capable if
in situ observation of shock-front propagation inside a crystal.
Summing up, the prospects for the fundamental study of ultra-fast lasermatter interaction and its technological
applications look extremely encouraging. As the technology becomes smaller, less expensive, more robust, less power-
hungry and more energy-efficient, it allows the increased exploitation of ultra-fast phenomena, ultimately entering our
everyday lives.
Acknowledgements
I amindebted to my colleagues with whomI had been working together on several projects included into this review, and
who helped me to shape my understanding of many phenomena involved. I would like to acknowledge Barry Luther-Davies,
Saulius Juodkazis, Vladimir Tikhonchuk, Wieslaw Krolikowski and Lewis Chadderton for many enlightening discussions.
I would like to express my special gratitude to Andrei Rode for his invaluable help in the preparation of this review. We
have had numerous discussions, which generated many useful comments and suggestions. He has advised for choosing most
relevant illustrative material, and indicated to papers with which the author was unfamiliar. Andrei was a key figure in the
majority of experiments cited in the review; therefore discussions of the experimental results and selection of those to be
referred was very important. The last but not least was his valiant proofreading of the manuscript and checking references.
I express my appreciationtomy colleagues andthe editors of journals whohave givenpermissiontoreproduce previously
published results. This research was supported under Australian Research Councils Discovery Projects funding scheme
(project numbers DP0666080 and DP0988054).
Appendix A. Properties of some elements
Here is thermal conductivity (W/(cmK)); is the heat of vaporisation J/g; C
p
is thermal capacity (J/(g K)); D is
thermal diffusivity in (cm
2
/s);
b
is the atomic binding (cohesive) energy (eV); E
F
is the Fermi energy in (eV); J
1
is the first
ionisation potential (eV); and
esc
is the work function in (eV) (Weast and Astle, 1981; Kittel, 1996; Gray, 1972; Ashcroft and
Mermin, 1976).
C
p
D
b
E
F
J
1

esc
Li-3 0.848 21200 3.582 0.443 1.686 4.74 5.3917
Be-4 2 32960 1.825 0.592 3.358 14.3 9.3227
B-5 0.274 51800 1.026 0.114 5.845 8.2980 4.45
C-6
(graphite)
0.063 59700 0.709 0.039 7.48 11.2603 5.0
Na-11 1.42 4238 1.228 1.191 3.24 5.1391
Mg-12 1.56 5288 1.023 0.877 7.08 7.6462 3.66
Al-13 2.37 10885 0.897 0.979 3.44 11.7 5.9858 4.28
Si-14 1.49 14046 0.75 0.853 4.12 8.1517 4.91
P-15 0.121 400 0.769 0.086 10.4867
S-16 0.00205 1403 0.71 0.001 10.3600
K-19 1.025 1967 0.757 1.571 0.935 2.12 4.3407
Ca-20 2.01 3860 0.647 2.017 1.86 4.69 6.1132
Sc-21 0.158 7401 0.568 0.093 6.5615
Ti-22 0.219 8977 0.523 0.093 6.8281 4.33
V-23 0.307 9010 0.489 0.103 6.7462
Cr-24 0.939 6712 0.449 0.291 4.154 6.7665 4.28
Mn-25 0.0781 4091 0.479 0.022 10.9 7.4340 4.1
Fe-26 0.804 6339 0.449 0.228 4.337 11.1 7.9024 4.5
Co-27 1.0 4802 0.421 0.266 4.471 7.8810 4.1
Ni-28 0.909 6483 0.444 0.230 7.6398 5.15
Cu-29 4.01 4784 0.385 1.165 3.524 7.00 7.7264 4.65
Zn-30 1.16 1755 0.388 0.419 9.47 9.3942
Ga-31 0.281 3643 0.371 0.128 2.856 10.4 5.9993 4.2
Ge-32 0.602 4601 0.32 0.354 3.909 7.8994 5.
As-33 0.502 464 0.329 0.264 3.157 9.7886 3.75
Se-34 0.0052 1209 0.321 0.003 9.7524
Rb-37 0.582 887 0.363 1.047 1.85 4.1771
Sr-38 0.354 1585 0.301 0.445 1.708 3.93 5.6949
Y-39 0.172 4420 0.298 0.129 6.2171 3.1
Zr-40 0.226 6281 0.278 0.125 6.6339
Nb-41 0.537 7426 0.265 0.236 5.32 6.7589 4.3
Mo-42 1.38 6191 0.251 0.538 6.869 7.0924 4.6
Ru-44 1.17 5853 0.238 0.397 7.3605
Rh-45 1.5 4801 0.243 0.497 7.4589
Pd-46 0.718 3402 0.244 0.245 8.3369
Ag-47 4.29 2640 0.235 1.739 2.974 5.49 7.5762 4.6
Cd-48 0.969 888 0.232 0.483 1.167 7.47 8.9938
C
p
D
b
E
F
J
1

esc
In-49 0.818 1968 0.233 0.480 2.536 8.63 5.7864 4.12
Sn-50 0.668 2494 0.228 0.402 10.2 7.3439
Sb-51 0.244 1589 0.207 0.176 10.9 8.6084
Te-52 0.0338 898 0.202 0.027 2.053 9.0096
I-53 0.0045 328 0.214 0.004 1.114 10.4513
Cs-55 0.359 481 0.242 0.792 0.816 1.59 3.8939
Ba-56 0.184 1099 0.204 0.257 1.856 3.64 5.2117 2.7
La-57 0.134 2895 0.195 0.112 5.5769
Ce-58 0.113 2841 0.192 0.087 4.415 5.5387
Pr-59 0.125 2349 0.193 0.096 5.473
Nd-60 0.165 2004 0.19 0.124 5.5250 3.2
Sm-62 0.133 1097 0.197 0.090 5.6436
Eu-63 0.139 1158 0.182 0.146 5.6704
Gd-64 0.106 1916 0.236 0.057 6.1501
Tb-65 0.111 1844 0.182 0.074 5.8638
Dy-66 0.107 1723 0.17 0.074 5.9389
Ho-67 0.162 1607 0.165 0.112 6.0215
Er-68 0.145 1674 0.168 0.095 3.31 6.1077
Tm-69 0.169 1462 0.16 0.113 6.1843
Yb-70 0.385 919 0.155 0.357 6.2542
Lu-71 0.164 2366 0.154 0.108 5.4259
Hf-72 0.23 3199 0.144 0.120 6.461 6.8251
Ta-73 0.575 4050 0.14 0.247 7.5496
W-74 1.73 4346 0.132 0.681 8.870 7.8460 4.9
Re-75 0.48 3781 0.137 0.167 7.8335
Os-76 0.876 3880 0.13 0.298 8.4382 4.83
Ir-77 1.47 2929 0.131 0.497 6.983 8.9670 5.27
Pt-78 0.716 2404 0.133 0.251 8.9587 5.65
Au-79 3.18 1647 0.129 1.277 3.843 5.53 9.2255 5.1
Hg-80 0.083 295 0.14 0.044 0.641 7.13 10.4375 4.49
Tl-81 0.461 807 0.129 0.302 8.15 6.1082
Pb-82 0.353 830 0.129 0.241 1.795 9.47 7.4167 4.25
Bi-83 0.0792 725 0.122 0.066 2.188 9.90 7.2856 4.22
Appendix B. Temperature-dependent bismuth properties from optical measurements in equilibrium
B.1. Electronphonon momentum exchange rate
The measurements of the reflectivity and the dielectric function by the ellipsometry technique give simultaneously the
real and imaginary parts of dielectric function and allow direct recovering the electron density and the electronphonon
momentum exchange rate. Indeed, the reflectivity, R, is directly related to the real
r
and imaginary
i
parts of the dielectric
function, =
r
+i
im
through the Fresnel formula:
R =
+1
2
=
|| +1
2(|| +
r
)
|| +1 +
2(|| +
r
)
(B.1)
where || =
_
2
r
+
2
i
. The real part of dielectric function of solid Bi
r
= 16.25 at 800 nm and at room temperature
indicating Bi has metallic property at this wavelength (Landolt-Brnstein, 1983; Garl, 2008), so we can safely assume it
can be described by the Drude form. It was experimentally proved that liquid Bi obeys the Drude-like form (Comins, 1972;
Hodgson, 1962; Smith et al., 1964; Lenham et al., 1965). The dielectric function in the Drude form is directly linked to the
plasma frequency
pe
and the electronphonon momentum exchange rate,
eph
:
r
= 1

2
pe
2
+(
mom
eph
)
2
;
im
=

2
pe
2
+(
mom
eph
)
2
mom
eph
(B.2)
here is the laser field frequency, = 2.43 10
15
s
1
for 775 nm, and squared plasma frequency is
2
pe
= 4e
2 n
e
m
e
, e is
the electron charge and m
e
is the effective electron mass. Thus the electronphonon momentum exchange rate and plasma
Fig. B.1. Temperature dependence of the momentum exchange frequency; solid line corresponds to linear dependence (B.4).
Source: The circles are the results from ellipsometry measurements from Garl (2008), trianglefrom Comins (1972), squarefrom Hodgson (1962).
frequency are measured in units of laser frequency, are directly connected to the real and imaginary parts of the dielectric
function:
mom
eph
=

im
1
r
;

2
pe
2
= (1
r
)[1 +(
mom
eph
/)
2
]. (B.3)
If the electron mass is known, than the number density of the conductivity electrons is directly retrieved from the plasma
frequency. Following this procedure Garl (2008) recovered the following Bi properties fromthe ellipsometry measurements
at the room temperature for laser light of 800 nm ( = 2.356 10
15
s
1
) :
2
pe
/
2
= 31; R = 0.74;
mom
eph
=
2.1 10
15
s
1
; n
e
= 5.34 10
22
cm
3
; Fermi energy 5.17 eV (v
F
= 1.35 10
8
cm/s) under assumption that electron
has a free electron mass. Comins (1972) found that Bi at 773 K has the following parameters: The reflectivity is R =
0.67;
2
pe
/
2
= 81.58 that corresponds to the number density of free carriers n
e
= 1.42 10
23
cm
3
. Therefore all 5
valence electrons are transferred into conduction band. Comins (1972) suggested that electron has a free mass value, then
Fermi energy 9.92 eV, and the Fermi velocity is v
F
= 1.87 10
8
cm/s. The electronphonon momentum exchange rate
reads
mom
eph
= 5.67 10
15
s
1
. Interpolation of the data extracted from all available experiments (Garl, 2008; Comins,
1972; Hodgson, 1962), and keeping in mind the theory from Appendix B shows that
mom
eph
grows up in direct proportion to
temperature (Fig. B.1):
mom
eph
= 2 10
15
(T/T
room
). (B.4)
The linear dependence of the momentum exchange rate on temperature fits the optical measurements for bismuth in
equilibrium with sufficient accuracy; the proportionality to the temperature holds for Bi well before and long after the
melting point.
B.2. Electronphonon energy exchange rate
Taking the theoretically established link between the temperature dependent energy exchange and the momentum
exchange frequency (see Section 2) and using the experimental temperature dependence equation (B.4) one obtains the
electronphonon energy exchange rate dependence on temperature:
en
eph
= 10
15
k
B
T
room
F
(T/T
room
)
2
= 1.17 10
13
(T/T
room
)
2
(s
1
). (B.5)
The results of theoretical calculations of the energy exchange rate are compared in Fig. B.2 to the experimental data.
Nowit is possible to calculate the dielectric function and reflectivity changes with temperature, the results are presented
in Fig. B.3. The dielectric function at room temperature was measured to be
r
= 16.25,
i
= 15.4 (Garl, 2008;
Garl et al., 2008) at 775 nm; the values are very close to the literature data (Lenham et al., 1965; Landolt-Brnstein,
1983). The electronphonon collision rate and plasma frequency at room temperature were found with the help of
Eqs. (B.3) as
mom
eph
/ = 0.893, and
2
pe
/
2
= 31.0. The theory predicts well the experimental results of dielectric function
measurements for liquid bismuth at 610 K (Hodgson, 1962) and 773 K (Comins, 1972).
Fig. B.2. Energy exchange rate: solid line corresponds to (B.5).
Source: Circles are the results from ellipsometry measurements (Garl, 2008), trianglefrom Comins (1972), squarefrom Hodgson (1962).
Fig. B.3. Dependence of real
r
(left low curve) and imaginary
i
parts (left upper curve) of the dielectric function and reflectivity (right) of bismuth at
775 nm on temperature in equilibrium. The solid lines were calculated using Eqs. (B.1) and (B.2).
Source: The circles are results of ellipsometry measurements from Garl (2008), trianglesfrom (Comins, 1972), squaresfrom Hodgson (1962), arrows
indicate the melting point T
m
= 544.7 K.
B.3. Number of electrons in the conduction band from the optical measurements
The electrons number density in the conduction band is retrieved fromthe plasma frequency under assumption that the
electron mass is equal to a free electron mass: m
= m
e
. The result at room temperature gives n
e
= 5.34 10
22
cm
3
(1.89
electrons out of total 5 are in the conduction band) and
F
= 5.17 eV. At 773 K in liquid bismuth all five electrons are in the
conduction band: n
e
= 1.42 10
23
cm
3
,
F
= 9.92 eV, and the dielectric function obeys the Drude-like form (Comins,
1972; Garl et al., 2008). Thus the experiments had shown that from 40% to 100% of the valence electrons are transferred to
the conduction band at the temperature increase from the room temperature to the melting point.
B.4. Electronic heat conduction and characteristic cooling time
Thermal diffusivity (coefficient of thermal diffusion) relates to both the Fermi energy and the electronphonon
momentum exchange rate: D = v
2
F
/3
mom
eph
. The diffusivity values recovered from the temperature dependency of the
momentum exchange rate in equilibrium from Fig. B.1, are D = (2.02.89) cm
2
/s for the temperature range 293 K793 K.
These results are in good agreement with the recent non-equilibrium measurements from the X-ray reflectivity data of fs-
laser excited bismuth giving D = 2.3 cm
2
/s (Johnson et al., 2008). Note the drastic difference of this value from0.067 cm
2
/s
given for the equilibrium conditions in the reference book (Landolt-Brnstein, 1983). Accordingly, the electron mean free
path, l
mfp
= v
F
/
mom
eph
= 0.67 nml
s
, is much less than the skin layer depth and the filmthickness, assuring the legitimacy
of the diffusion approximation for the electron heat transfer.
The time for the temperature smoothing across a 30 nmthick filmtaken in accord to the diffusivity value is, respectively,
t
smooth
= l
2
s
/D = 3.9 ps, which is a reason for preserving a high thermal gradient across the film thickness in the first ps
after the laser excitation (Table B.1).
Table B.1
Summary of optical properties of solid and liquid Bismuth at 800 nm. Laser frequency = 2.356 10
15
s
1
; photon energy 1.55 eV; electrons have free
electrons mass.
re

im
n
e
, 10
22
cm
3
mom
eph
/ R
2
pe
/
2
Solid Bi (Garl, 2008) 16.25 15.4 22.39 5.34 0.893 0.74 31.0
Liquid Bi (Comins, 1972) 11 28.9 30.92 14.1 2.408 0.67 81.58
Appendix C. Empirical pseudo-chemical inter-atomic potential
A simple Morse-like form of the empirical inter-atomic potential depends on four parameters (Abell, 1985):
V(r) = V
R
exp(r) V
A
exp(r); (C.1)
here V
R
, andV
A
, are respectively the repulsive andattractive parts of the potential andtheir gradients. These four defining
the shape of potential parameters are connectedto four experimentally or theoretically obtainedimportant characteristics of
a solid. The first parameter is the atomic separation in equilibrium, d
0
, that is the position where the potential has minimum.
The second parameter is the binding (cohesive) energy, which is equal to the minimumvalue of the potential in equilibrium,
b
= V
min
= V(d
0
). The third parameter is the characteristic phonons frequency, that is close to the Debye frequency, which
is defined as,
2
ph
M
1
(
2
V r
2
)
d
0
. Finally, the fourth is the ratio of the repulsive and attractive potential gradients,
s = /, it is a known number for metals and dielectrics. For example, the empirically determined ratio for transitional
metals falls in a range 3 s 5, while for some covalent systems the values are: (C) = 1.4; s(Si) = 1.9; s(Ge) = 2.0
(Abell, 1985). Note, the Morse potential corresponds to the case = 2.
The equilibrium inter-atomic spacing obtains from the condition of the potential minimum(V/r)
r=d
0
= 0:
d
0
= ln
_
V
R
V
A
_ 1
. (C.2)
Accordingly, the binding energy is as follows:
b
V(d
0
) = V
R
(s 1)
_
V
R
V
A
_
= V
R
(s 1) exp(d
0
). (C.3)
The squared phonon frequency expresses through the potential parameters in a form:
M
2
ph
=
_
2
V
r
2
_
d
0
=
2
b
_
s
2
s 1
(s 1)
1
s
V
A
b
_
b
V
R
_1
s
_
. (C.4)
Our main goal is to express the perturbed quantities through the equilibrium values. It is reasonable to suggest that the
change in the attractive potential only is responsible for small increase in the inter-atomic spacing and the decrease in
the binding energy due to the electrons or lattice heating. Then, the relation between changes in the binding energy and
inter-atomic distance follows from (C.3):
b
= |
b0
| |
b
| =
b0
d
. (C.5)
Weak electronic excitation, k
B
T
e

b
, results in the change in the binding energy
b
k
B
T
e
, in the increase of the atomic
space separation, and in decrease of the phonons frequency (the red shift or, mode softening). The increase in the spacing
reads:
d k
B
T
e
/
b,0
. (C.6)
The mode softening comes from (C.4) in the form:
2
ph
=
2
ph,0
b
M
. (C.7)
Note that the use of empirical potential for scaling allows one to take into account the gradient scales of the attractive and
repulsive potentials, , instead of the average inter-atomic distance d as it is done in Section 3. Because 1 < < , the
perturbations calculated with the help of empirical potential are larger then those calculated with the approximation used
in Section 3.
Table D.1
Physical properties of metals in the experiments for ablation in air and in vacuum.
Al Cu Fe Pb
Thickness of mono-layer, 10
8
cm 2.86 2.56 2.35 3.5
Electron density, 10
22
cm
3
18.6 8.45 16.8 13.2
Atomic density, 10
22
cm
3
6.02 8.45 8.5 3.3
Fermi energy, eV 11.63 7.0 11.1 9.47
Ionisation potential, eV 5.98 7.73 7.9 7.417
Binding energy, eV 3.44 3.524 4.337 1.795
Thermal diffusivity, cm
2
/s 0.979 1.165 0.228 0.241
Electron effective mass, m
/m
e
(Kittel, 1996) 1.48 1.38 8 1.97
pe
= (4e
2
n
e
/m
e
)
1/2
10
16
s
1
2 1.4 0.817 1.46
M (a.u.) 26.98 63.546 55.845 207.2
Table D.2
The optical parameters of Al, Cu, Fe, Pb at 532 nmat the roomtemperature;
here n andk are real andimaginary parts of the refractive index; A = 1R =
1
(n1)
2
+k
2
(n+1)
2
+k
2
is the absorption coefficient; l
s
= c/k is the skin depth;
en
eph

pe
(m
e
/M) is the energy exchange rate; t
en
eph
= (
en
eph
)
1
is the
energy exchange time.
Al Cu Fe Pb
a
n 0.85 2.60 1.05 2.01
k 6.48 2.58 3.33 3.48
A 0.075 0.487 0.432 0.38
l
s
, (nm) 13.11 32.85 25.43 24.3
t
en
eph
, (ps) 1.67 6 1.56 13
a
At 589.3 nm.
Table D.3
The optical parameters for hot metallic plasma
of Al, Cu, Fe, Pb at = 532 nm.
Al Cu Fe Pb
n k 1.67 1.45 1.81 1.46
A 0.673 0.716 0.648 0.714
l
s
, nm 50.7 58.4 46.8 58
Table D.4
The electron heat capacity near the ablation threshold at
k
B
T
e
=
b
.
Al Cu Fe Pb
Binding energy (eV) 3.065 3.173 3.695 1.795
C
e
(in units k
B
) 0.473 0.76 0.122 0.269
Appendix D. Ablation of metals
In the conditions close to the ablation threshold the modified optical properties of metallic plasma can be estimated as
follows. Near the ablation threshold the condition holds,
eff

pe
> . The dielectric function and the refractive index
then are as follows:
2
pe
;

pe
_
1 +

2
2
pe
_
1
; n k =
_
2
_
1/2
. (D.1)
The absorption coefficient follows from the Fresnel formulae (Table D.2):
A = 1 R
4n
(n +1)
2
+n
2
. (D.2)
The optical parameters for the hot metallic plasma at = 532 nm ( = 3.54 10
15
s
1
) are presented below in
Table D.3.
The electron heat capacity at k
B
T
e

b

F
is unknown. We interpolate its dependence on x = k
B
T
e
/
F
by the function
C
e

3
2
k
B
(2x x
2
) that attains the ideal gas value at x = 1. The electron effective masses for the threshold calculations were
taken equal to those from the thermal conductivity measurements in Table D.1 (Kittel, 1996) (Table D.4)
Table E.1
Material and optical properties of transparent solids at 700900 nm where micro-explosion experiments were produced (Gamaly et al., 2006).
Refractive
index
a
Young
modulus (GPa)
Formation threshold
b
(nJ)
Amorphous/Void
Depth(m) n
2
(cm
2
/W) Mass
density (g/cm
3
)
Sapphire 1.75 400 21/35 20 3 10
16
3.89
Viosil 1.47 75 13/30 30 3.5 10
16
2.2
Polystyrene 1.55 3.5 6/11 10 9.3 10
9
1.05
a
The approximate values for 700900 nm wavelengths. Refractive index of immersion oil was approximately 1.51.
b
Thresholds have been measured by the best fit procedure. The amorphous threshold value marks the optically recognisable breakdown threshold
in the viosil and polystyrene; this threshold corresponds to the amorphisation onset in the case of crystalline sapphire.
Table E.2
The ionisation potentials, J
Z
(in eV), for the silicon and oxygen fromVainshtein and Shevelko (1993) (number
of electrons per state is given in parentheses).
Electron state 1s 2s 2p 3s 3p
Si 1844 (2) 154 (2) 104 (6) 13.46 (2) 8.15 (2)
O 538 (2) 28.5 (2) 13.6 (4) . . . . . .
Table E.3
Ionisation energies Q
Z
(in eV) for the silicon and oxygen.
Z
O
, Z
Si
1 2 3 4 5 6 7 8
Q
z
(Si) 8.15 16.3 29.76 43.2 147.2 251.2 355 459
Q
z
(O) 13.6 27.2 40.8 54.4 82.9 111.4 649.4 . . .
Q
z
(Si) +2Q
z
(O) 35.35 70.7 111.3 152 313 474 1653.8 . . .
Analytical formula for the thermal ablation depth
By introducing new variable x =
_
t
t
bs
_
1/4
the formula for the ablation depth reduces to the following:
l
th
= 2t
bs
_
2k
B
T
bs
M
_
1/2

1
exp
_

b
k
B
T
bs
x
2
_
xdx
2
. (D.3)
The integration result is:

1
exp
_
4
b
k
B
T
bs
x
2
_
2xdx
2

1
exp{cx
2
}2xdx
2
=
2
c

1
xd exp{cx
2
}
=
2
c
_
x exp{cx
2
}
c
exp{u
2
}du
_

k
B
T
bs
2
b
exp
_
4
b
k
B
T
bs
_
. (D.4)
During thermal ablation the condition holds: c = 4
b
/k
B
T > 2. Therefore,
_
c
exp{u
2
}du 0. The thermal ablation depth
takes its final form:
l
th
t
bs
_
2k
B
T
bs
M
_
1/2
k
B
T
bs
2
b
exp
_
4
b
k
B
T
bs
_
. (D.5)
Appendix E. Micro-explosion in transparent solids: material parameters and ionisation losses
Ionisation losses in silica
One can estimate the degree of ionisation (Table E.1) and losses on the basis of data fromTable E.2, which are summarised
in Table E.3.
References
Abell, G.C., 1985. Empirical chemical pseudopotential theory of molecular and metallic bonding. Phys. Rev. B 31, 61846196.
Ackermann, R., Salmon, E., Lascoux, N., Kasparian, J., Rohwetter, P., Stelmaszczyk, K., Li, S., Lindinger, A., Wste, L., Bjot, P., Bonacina, L., Wolf, J.-P., 2006.
Optimal control of filamentation in air. Appl. Phys. Lett. 89, 171117.
Afanasiev, Yu.V., Demchenko, N.N., Zavestovskaya, I.N., Isakov, V.A., Kanavin, A.P., Uriupin, S.A., Chichkov, B.N., 1999. Simulation of metal ablation by the
ultra short laser pulses. Izv. Acad. Nauk Ser. Fiz. 63, 667675 (in Russian).
Afanasiev, Yu.V., Krokhin, O.N., 1967. Evaporation of matter by a laser radiation. Sov. Phys. JETP 52, 966975.
Afanasiev, Yu.V., Krokhin, O.N., 1971. High temperature and plasma phenomena arising in the power lasermatter interaction. In: Calderola, P., Knoepfel, H.
(Eds.), Physics of High Energy Density. Academic Press, New York, London.
Akhmanov, S.A., Vyspoukh, V.A., Chirkin, A.S., 1988. Optics of Femtosecond Laser Pulses. Nauka, Moscow.
Albrecht, W., Kruse, Th., Kurz, H., 1992. Time-resolved observation of coherent phonons in superconducting YBa
2
Cu
3
O
7
- thin films. Phys. Rev. Lett. 69,
14511453.
Allen, P.B., 1987. Theory of thermal relaxation of electrons in metals. Phys. Rev. Lett. 59, 14601462.
Allen, L., Barnett, S.M., Padgett, M.J., 2003. Optical Angular Momentum. Taylor and Francis, London.
Amoruso, S., Ausanio, G., Vitiello, M., Wang, X., 2005a. Infrared femtosecond laser ablation of graphite in high vacuum probed by optical emission
spectroscopy. Appl. Phys. A 81, 981986.
Amoruso, S., Ausanio, G., Bruzzese, R., Vitiello, M., Wang, X., 2005b. Femtosecond laser pulse irradiation of solid targets as a general route to nanoparticle
formation in a vacuum. Phys. Rev. B 71, 033406.
Amoruso, S., Bruzzese, R., Spinelli, N., Velotta, R., Vitiello, M., Wang, X., Ausanio, G., Iannotti, V., Lanotte, L., 2004. Generation of silicon nanoparticles via
femtosecond pulse in vacuum. Appl. Phys. Lett. 84, 45024504.
Andreotti, A., Colombini, M.P., Nevin, A., Melessanaki, K., Pouli, P., Fotakis, C., 2005. Laser pulse duration effects for the cleaning of wall paintings. In: 6th
International Congress on Lasers in Conservation of Artworks Book of Abstracts, Academy of Fine Arts, Vienna, 2126 September, p. 127.
Anglos, D., Couris, S., Mavromanolakis, A., Zergioti, I., Solomidow, M., Liu, W.-Q., Papazoglou, T.Z., Zafiropulos, V., Fotakis, C., Doulgeridis, M., Fostiridou,
A., 1997. Artwork diagnostics. Laser-induced breakdown spectroscopy (LIBS) and laser-induced fluorescence (LIF). In: Knig, E., Kautek, W. (Eds.), 1st
International Congress on Lasers in the Conservation of Artworks LACONA I. Verlag Mayer, Vienna, pp. 113118.
Anisimov, S.I., Imas, Ya.A., Romanov, G.S., Khodyko, Yu.V., 1971. Action of high intensity radiation on metals. National Technical Information Service,
Springfield, Virginia.
Anisimov, S.I., Kapeliovich, B.L., Perelman, T.L., 1974. Electron emission frommetal surfaces exposed to ultra-short laser pulses. Sov. Phys. JETP 39, 375377.
Anisimov, S.I., Lukyanchuk, B.S., 2002. Selected problems of laser ablation theory. Phys. Usp. 45, 293324.
Arnold, D., Cartier, E., 1992. Theory of laser-induced free-electron heating and impact ionization in wide-band-gap solids. Phys. Rev. B 46, 1510215115.
Arnold, C.B., Serra, P., Pique, A., 2007. Laser direct-write techniques for printing of complex materials. MRS Bull. 32, 2331.
Ashcroft, N.W., Mermin, N.D., 1976. Solid State Physics. Holt, Rinehart and Winston.
Baer, C.R.E., Kraenkel, C., Saraceno, C.J., Heckl, O.H., Golling, M., Peters, R., Petermann, K., Suedmeyer, T., Huber, G., Keller, U., 2010. Femtosecond thin-disk
laser with 141 W of average power. Opt. Lett. 35, 23022304.
Bandulet, H.C., Comtois, D., Bisson, E., Fleischer, A., Pepin, H., Kieffer, J.C., Corkum, P.B., Villeneuve, D.M., 2004. Gating attosecond pulse train generation
using multicolor laser fields. Phys. Rev. A 81, 013803.
Bauerle, D., 2000. Laser Processing and Chemistry, 3rd Ed. Springer, Berlin.
Beyer, O., Breunig, I., Kalkum, F., Buse, K., 2006. Photorefractive effect in iron-doped lithium niobate crystals induced by femtosecond pulses of 1.5m
wavelength. Appl. Phys. Lett. 88, 051120.
Blatt, F.J., 1968. Physics of Electronic Conduction in Solids. McGraw-Hill, New York.
Boettger, J.C., Trickey, S.B., 1996. High-precision calculation of the equation of state and crystallographic phase stability for aluminum. Phys. Rev. B 53,
30073012.
Bohandy, J., Kim, B.F., Adrian, F.J., Jette, A.N., 1988. Metal deposition at 532 nm using a laser transfer technique. J. Appl. Phys. 63, 11581162.
Bondybey, V.E., English, J.H., 1982. Laser excitation spectra and lifetimes of Pb
2
and Sn
2
produced by YAG laser vaporization. J. Chem. Phys. 76, 2165.
Bonse, J., Baudach, S., Kruger, J., Kautek, W., Lenzner, M., 2002. Femtosecond laser ablation of silicon-modifications thresholds and morphology. Appl. Phys.
A 74, 1925.
Born, M., 1939. Thermodynamics of crystals and melting. J. Chem. Phys. 7, 591603.
Born, M., Wolf, E., 2003. Principles of Optics, 7th ed. Cambridge University Press, Cambridge, UK.
Boschetto, D., Gamaly, E.G., Rode, A.V., Luther-Davies, B., Glijer, D., Garl, T., Albert, O., Rousse, A., Etchepare, J., 2008a. Small atomic displacements recorded
in bismuth by the optical reflectivity of femtosecond laser-pulse excitations. Phys. Rev. Lett. 100, 027404.
Boschetto, D., Garl, T., Rousse, A., Gamaly, E.G., Rode, A.V., 2008b. Lifetime of optical phonons in fslaser excited bismuth. Appl. Phys. A 92, 873876.
Boullet, J., Zaouter, Y., Limpert, J., Petit, S., Mairesse, Y., Fabre, B., Higuet, J., Mevel, E., Constant, E., Cormier, E., 2009. High-order harmonic generation at a
megahertz-level repetition rate directly driven by an ytterbium-doped-fiber chirped-pulse amplification system. Opt. Lett. 34, 14891491.
Bronson, S.D., Fujimoto, J.G., Ippen, E.P., 1987. Femtosecond electronic heat-transport dynamics in thin gold films. Phys. Rev. Lett. 59, 1962.
Budd, D.M., Horwitz, J.S., Callahan, J.H., McGill, R.A., Houser, E.J., Chrisey, D.B., Papantonakis, M.R., Haglund, R.F., Galicia, M.C., Vertes, A., 2001. Resonant
infrared pulsed-laser deposition of polymer films using a free-electron laser. J. Vac. Sci. Technol. 19, 26982702.
Bulgakov, A.V., Ozerov, I., Marine, W., 2004. Silicon clusters produced by femtosecond laser ablation: non-thermal emission and gas-phase condensation.
Appl. Phys. A 79, 15911594.
Bulgakova, N.M., Zhukov, V.P., Vorobyev, A.Y., Guo, C., 2008. Modeling of residual thermal effect in femtosecond laser ablation of metals: role of a gas
environment. Appl. Phys. A 92, 883889.
Burmester, T., Meier, M., Haferkamp, H., Barcikowski, S., Bunte, J., Ostendorf, A., 2005. Femtosecond laser cleaning of metallic cultural heritage and antique
artworks. In: Dickman, K., Fotakis, C., Asmus, J.F., Osnabrck, (Eds.), Lasers in the Conservation of Artworks LACONA V Proceedings. In: Springer
Proceedings in Physics, vol. 100. Springer-Verlag, Berlin, Heidelberger, pp. 6169.
Buse, K., 1996. Light-induced charge transport processes in photorefractive crystals I: models and experimental methods. Appl. Phys. B 64, 273291.
Buse, K., Adibi, A., Psaltis, D., 1998. Non-volatile holographic storage in doubly doped lithium niobate crystals. Nature 393, 665668.
Cahn, R.W., 1986. Melting and the surface. Nature 323, 668669.
Cai, L., Yang, X., Wang, Y., 2002. All fourteen Bravais lattices can be formed by interference of four noncoplanar beams. Opt. Lett. 27, 900902.
Car, R., Parinello, M, 1985. Unified approach for molecular dynamics and density-functional theory. Phys. Rev. Lett. 55, 2471.
Cauble, R., Rozmus, W., 1995. Two-temperature frequency dependent electrical resistivity in solid density plasma produced by ultra-short pulses. Phys.
Rev. E 52, 29742981.
Cavalleri, A., Tth, Cs., Siders, C.W., Squier, J.A., 2001. Femtosecond structural dynamics in VO
2
during an ultrafast solidsolid phase transition. Phys. Rev.
Lett. 87, 237401.
Chan, A., Rode, A., Gamaly, E., Luther-Davies, B., Taylor, B., Dawes, J., Lowe, M., Hannaford, P., 2003. Ablation of dental enamel using subpicosecond pulsed
lasers. In: Ishikawa, I., Frame, J.W., Aoki, A. (Eds.), Lasers in Dentistry: Revolution of Dental Treatment in the New Millennium, vol. 1248. Elsevier,
ISBN: 0-444-51163-6, pp. 117119.
Chang, Z., Corkum, P., 2010. Attosecond photon sources: the first decade and beyond. J. Opt. Soc. Amer. B 27, B9B17.
Chapman, H.N., Nugent, K.A., 2010. Coherent lensless X-ray imaging. Nat. Photonics 4, 833839.
Cheng, T.K., Vidal, J., Zeiger, H.J., Dresselhaus, G., Dresselhaus, M.S., Ippen, E.P., 1991. Mechanism for displacive excitation of coherent phonons in Sb, Bi, Te,
and Ti
2
O
3
. Appl. Phys. Lett. 59, 19231925.
Cho, G.C., Ktt, W., Kurz, H., 1990. Subpicosecond time-resolved coherent-phonon oscillations in GaAs. Phys. Rev. Lett. 65, 764766.
Chollet, M., Cuerin, L., Uchida, N., Fukaya, S., Shimoda, H., Ishikawa, T., Matsuda, K., Hasegawa, T., Ota, A., Yamochi, H., Saito, G., Tazaki, R., Adachi, S.-I.,
Koshihara, S.-Y., 2005. Gigantic photoresponse in 1/4-filled-band organic salt (EDO TTF)
2
PF
6
. Science 307, 86.
Chrisey, D.B., Hubler, G.K. (Eds.), 1994. Pulsed Laser Deposition of Thin Films. Wiley and Sons, New York.
Chrisey, D.B., Pique, A., McGill, R.A., Horwitz, J.S., Ringeisen, B.R., Bubb, D.M., Wu, P.K., 2003. Laser deposition of polymer and biomaterial films. Chem. Rev.
103, 553576.
Chung, S.H., Mazur, E., 2009. Surgical application of femtosecond lasers. J. Biophotonics 2, 557572.
Cohen, M.L., 1985. Calculation of bulk moduli of diamond and zinc-blende solids. Phys. Rev. B 32, 7988.
Collet, E., Lemee-Cailleau, M.-H., Buron-Le Cointe, M., Cailleau, H., Wuff, M., Luty, T., Koshihara, S.-Y., Meyer, M., Toupet, L., Rabiller, P., Techert, S., 2003.
Laser-induced ferroelectric structural order in an organic charge-transfer crystal. Science 300, 612615.
Comins, N.R., 1972. The optical properties of liquid metals. Phil. Mag. 25, 817831.
Corkum, P.B., Brunel, F., Sherman, N.K., Srinivasan-Rao, T., 1988. Thermal response of metals to ultra-short-pulse laser excitation. Phys. Rev. Lett. 61,
28862889.
Corkum, P.B., Krausz, F., 2007. Attosecond science. Nat. Phys. 6, 381387.
Day, D., Gu, M., 2002. Formation of voids in a doped polymethylmethacrylate polymer. Appl. Phys. Lett. 80, 24042406.
Debye, P., 1913. Interferenz von Rntgenstrahlen und Wrmebewegung. Ann. Phys. 348, 4992 (in German).
DeCamp, M.F., Reis, D.A., Bucksbaum, P.H., Merlin, R., 2001. Dynamics and coherent control of high-amplitude optical phonons in Bi. Phys. Rev. B 64, 092301.
Delaporte, P., Gastaud, M., Marine, W., Sentis, M., Uteza, O., Touvenot, P., Alcaraz, J.L., Le Samedy, J.M., Blin, D., 2003. Dry excimer laser cleaning applied to
nuclear decontamination. Appl. Surf. Sci. 208209, 298305.
Drake, R.P., 2006. High Energy Density Physics. Springer, Berlin.
Drake, R.P., 2010. High energy density physics. Phys. Today.
Du, D., Liu, X., Korn, G., Squier, J., Mourou, G., 1994. Laser-induced break down by impact ionisation in SiO
2
with pulse width from7 ns to 150 fs. Appl. Phys.
Lett. 64, 30713073.
Dumitrica, T., Garcia, M.E., Jeschke, H., Yakobson, B., 2004. Selective cap opening in carbon nanotubes driven by laser-induced coherent phonons. Phys. Rev.
Lett. 92, 117401.
Eason, R., 2007. Pulsed Laser Deposition of Thin Films. John Wiley and Sons, Hoboken, New Jersey.
Ebbesen, T.W. (Ed.), 1997. Carbon Nanotubes. Preparation and Properties. CRC Press, Boca Raton.
Ebbesen, T.W., Ajayan, P.M., 1992. Large-scale synthesis of carbon nanotubes. Nature 358, 220.
Eidam, T., Hanf, S., Seise, E., Andersen, T.V., Gabler, T., Wirth, C., Schreiber, T., Limpert, J., Tnnermann, A., 2010. Femtosecond fiber CPA system emitting
830 W average output power. Opt. Lett. 35, 9496.
Eidmann, K., Meyer-ter-Vehn, J., Schlegel, T., Huller, S., 2000. Hydrodynamic simulation of subpicosecond laser interaction with solid density matter. Phys.
Rev. E 62, 12021214.
Ell, R., Morgner, U., Krtner, F.X., Fujimoto, J.G., Ippen, E.P., Scheuer, V., Angelow, G., Tschudi, T., Lederer, M.J., Boiko, A., Luther-Davies, B., 2001. Generation
of 5 fs pulses and octave-spanning spectra directly from a Ti: sapphire laser. Opt. Lett. 26, 373.
Elliott, S.R., 1986. A unified model for reversible photostructural effects in chalcogenide glasses. J. Non-Cryst. Solids 81, 7198.
Elsayed, H.E., Norris, T.B., Pessot, M.A., Mourou, G.A., 1987. Time-resolved observation of electronphonon relaxation in copper. Phys. Rev. Lett. 58,
12121214.
Englert, L., Rethfeld, B., Haag, L., Wollenhaupt, M., Sarpe-Tudoran, C., Baumert, T., 2007. Control of ionization processes in high band gap materials via
tailored femtosecond pulses. Opt. Express 15, 17855.
Fann, W.S., Storz, R., Tom, H.W.K., Bokor, J., 1992. Direct measurements of non-equilibrium electron-energy distributions in sub-picosecond laser-heated
gold films. Phys. Rev. Lett. 68, 28342837.
Farsari, M., Chichkov, B., 2009. Two-photon fabrication. Nat. Photonics 3, 450452.
Farsari, M., Vamvakaki, M., Chichkov, B., 2010. Multiphoton polymerization of hybrid materials. J. Opt. 12, 124001.
Fecht, H.J., 1992. Defect-induced melting and solid-state amorphization. Nature 356, 133135.
Fecht, H.J., Johnson, W.L., 1988. Entropy and enthalpy catastrophe as a stability limit for crystalline material. Nature 334, 5055.
Feit, M.D., Rubenchik, A.M., Kim, B.M., Da Silva, L.B., Perry, M.D., 1998. Physical characterization of ultra-short laser pulse drilling of biological tissue. Appl.
Surf. Sci. 127129, 869874.
Fradin, D.W., Bloembergen, N., Letellier, J.P., 1973. Dependence of laser-induced breakdown field strength on pulse duration. Appl. Phys. Lett. 22, 635637.
Frenkel, Ya.I., 1948. Statistical Physics. Nauka, Moscow, (in Russian).
Friedli, C., Ashcroft, N.W., 1975. Aluminum under high pressure. I. Equation of state. Phys. Rev. B 12, 55525559.
Friese, M.E.J., Nieminen, T.A., Heckenberg, N.R., Rubinsztein-Dunlop, H., 1998. Optical alignment and spinning of laser-trapped microscopic particles. Nature
394, 348350.
Fritz, D.M., Reis, D.A., Adams, B., Akre, R.A., Arthur, J., Blome, C., Bucksbaum, P.H., Cavalieri, A.L., Engemann, S., Fahy, S., Falcone, R.W., Fuoss, P.H., Gaffney, K.J.,
George, M.J., Hajdu, J., Hertlein, M.P., Hillyard, P.B., Horn-von Hoegen, M., Kammler, M., Kaspar, J., Kienberger, R., Krejcik, P., Lee, S.H., Lindenberg, A.M.,
McFarland, B., Meyer, D., Montagne, T., Murray, .D., Nelson, A.J., Nicoul, M., Pahl, R., von der Linde, D., Hastings, J.B., 2007. Ultra-fast bond softening in
bismuth: mapping a solids interatomic potential with X-rays. Science 315, 633636.
Frohlich, H., 1949. Theory of Dielectrics. Clarendon Press, Oxford.
Galayda, J.N., Arthur, J., Ratner, D.F., White, W.E., 2010. X-ray free-electron laserspresent and future capabilities. J. Opt. Soc. Amer. B 27, B106B118.
Gamaly, E.G., 1993. The interaction of ultra-short, powerful laser pulses with a solid target: ion expansion and acceleration with time-dependent ambipolar
field. Phys. Fluids B 5, 944949.
Gamaly, E.G., 1994. Ultra-short powerful laser matter interaction: physical problems, models and computations. Laser Part. Beams 12, 185208.
Gamaly, E., 2010. Ultra-fast disordering by fslasers: lattice superheating prior to the entropy catastrophe. Appl. Phys. A 101, 205208.
Gamaly, E.G., Juodkazis, S., Misawa, H., Luther-Davies, B., Hallo, L., Nicolai, P., Tikhonchuk, V.T., 2006. Lasermatter interaction in a bulk of a transparent
solid: confined micro-explosion and void formation. Phys. Rev. B 73, 214101.
Gamaly, E.G., Juodkazis, S., Mizeikis, V., Misawa, H., Rode, A.V., Krolikowski, W.Z., 2010. Modification of refractive index by a single fs-pulse confined inside
a bulk of a photo-refractive crystal. Phys. Rev. B 81, 054113.
Gamaly, E.G., Juodkazis, S., Mizeikis, V., Misawa, H., Rode, A.V., Krolikowski, W.Z., Kitamura, K., 2008. Three-dimensional writereaderase memory bits by
femtosecond laser pulses in photorefractive LiNbO
3
crystals. Curr. Appl. Phys. 8, 416419.
Gamaly, E.G., Madsen, N.R., Duering, M., Rode, A.V., Kolev, V.Z., Luther-Davies, B., 2005. Ablation of metals with picosecond laser pulses: evidence of long-
lived non-equilibrium conditions at the surface. Phys. Rev. B 71, 174405.
Gamaly, E.G., Madsen, N.R., Golberg, D., Rode, A.V., 2009. Expansion-limited aggregation of nanoclusters in a single-pulse laser-produced plasma. Phys. Rev.
B 80, 184113.
Gamaly, E.G., Rode, A.V., 2009. Is the ultra-fast transformation of bismuth non-thermal? ArXiv e-prints: arXiv:0910.2150.
Gamaly, E.G., Rode, A.V., Luther-Davies, B., 1999a. Ultrafast ablation with high-pulse-rate lasers. Part I: theoretical considerations. J. Appl. Phys. 85,
42134221.
Gamaly, E.G., Rode, A.V., Luther-Davies, B., 1999b. Laser ablation of carbon at the threshold of plasma formation. Appl. Phys. A 69, S121S127.
Gamaly, E.G., Rode, A.V., Luther-Davies, B., 2007. Ultra-fast laser ablation and film deposition. In: Eason, Robert W. (Ed.), Pulsed Laser Deposition of Thin
Films: Applications in Electronics, Sensors, and Biomaterials. John Wiley and Sons Inc., Hoboken, New Jersey, pp. 99130.
Gamaly, E.G., Rode, A.V., Luther-Davies, B., Tikhonchuk, V.T., 2002. Ablation of solids by femtosecond lasers: ablation mechanism and ablation thresholds
for metals and dielectrics. Phys. Plasmas 9, 949957.
Gamaly, E.G., Rode, A.V., Perrone, A., Zocco, A., 2001. Mechanisms of ablation-rate decrease in multiple-pulse laser ablation. Appl. Phys. A 73, 143149.
Gamaly, E.G., Rode, A.V., Uteza, O., Kolev, V., Luther-Davies, B., Bauer, T., Koch, J., Korte, F., Chichkov, B.N., 2004. Control over a phase state of the laser plume
ablated by femtosecond laser: spatial pulse shaping. J. Appl. Phys. 95, 22502257.
Gamaly, E.G., Rode, A.V., 2004. Nanostructures created by lasers. In: Nalwa, H.S. (Ed.), Encyclopaedia of Nanoscience and Nanotechnology, vol. 7. American
Scientific Publishers, Stevenson Range, pp. 783809.
Gamaly, E.G., Tikhonchuk, V.T., 1988. Interaction of ultra-short laser pulses with matter. JETP Lett. 48, 453456.
Garl, T., 2008. Ultrafast dynamics of coherent optical phonons in bismuth. Ph.D. Thesis. Ecole Polytechnique, Palaiseau, France (unpublished).
Garl, T., Gamaly, E.G., Boschetto, D., Rode, A.V., Luther-Davies, B., Rousse, A., 2008. Birth and decay of coherent optical phonons in fslaser excited bismuth.
Phys. Rev. B 78, 134302.
Garrett, G.A., Albrecht, T.F., Whitaker, J.F., Merlin, R., 1996. Coherent THz phonons driven by light pulses and the Sb problem: what is the mechanism? Phys.
Rev. Lett. 77, 3661.
Gaspard, S., Oujja, M., Moreno, P., Mendez, C., Garsia, A., Domingo, C., Castilejo, M., 2008. Intrraction of femtosecond laser pulses with tempera paints. Appl.
Surf. Sci. 255, 26752681.
Gattass, R.R., Mazur, E., 2008. Femtosecond laser micromachining in transparent materials. Nat. Photonics 2, 219225.
Gibbon, P., 2005. Short Pulse Laser Interactions with Matter. Imperial College Press, London.
Glass, A.M., Von der Linde, D., Negran, T.J., 1974. High-voltage bulk photovoltaic effect and the photorefractive process in LiNbO
3
. Appl. Phys. Lett. 25,
233235.
Glezer, E.N., Mazur, E., 1997. Ultrafast-laser driven micro-explosions in transparent materials. Appl. Phys. Lett. 71, 882884.
Glezer, E.N., Milosavjevic, M., Huang, L., Finlay, R.J., Her, T.-H., Callan, J.P., Masur, E., 1996. Opt. Lett. 21, 20232026.
Golberg, D., Bando, Y., Eremets, M., Takemura, K., Kurashima, K., Yusa, H., 1996. Nanotubes in boron nitride laser heated at high pressure. Appl. Phys. Lett.
69, 2045.
Gong, X.G., Chiarotti, G.L., Parinello, M., Tosati, E., 1993. Coexistence of monoatomic and diatomic molecular fluid character in liquid gallium. Europhys.
Lett. 21 (4), 469475.
Gorecki, T., 1977. Comments on vacancies and melting. Scr. Metall. 11, 10511053.
Gouleimakis, E., Loh, Z.-H., Wirth, A., Santra, R., Rohringer, N., Yakovlev, V., Zherebtsov, S., Pfeifer, T., Azzeer, A., King, M., Leone, S.R., Krausz, F., 2010.
Real-time observation of valence electron motion. Nature 466, 739743.
Gray, D.E. (Ed.), 1972. American Institute of Physics Handbook, 3rd ed. McGraw-Hill Book Company, New York.
Grimvall, G., 1981. The electronphonon interaction in metals. In: Wohlfarth, E.P. (Ed.), Selected Topics in Solid State Physics, vol. XVI. North-Holland,
Amsterdam.
Groeneveld, R.H.M., Sprik, R., Lagendijk, A., 1990. Ultrafast relaxation of electrons probed by surface plasmons at a thin silver film. Phys. Rev. Lett. 64,
784787.
Gruzdev, V.E., 2004. Laser-induced ionization of solids: back to Keldysh. In: Exarhos, G.J., et al. (Eds.), Laser-Induced Damage in Optical Materials.
In: Proceedings of SPIE, vol. 5647. SPIE, Bellingham, WA, 2005.
Gruzdev, V.E., 2005. Laser-induced collective ionization in wide band-gap crystalline dielectrics. In: ICONO-2005, Ultra-Fast Phenomena and Physics of
Super-Intense Laser Fields. In: Proceedings of SPIE, SPIE, Bellingham, WA.
Guillermin, M., Colombier, J.P., Valette, S., Audouard, E., Garrelie, F., Stoian, R., 2010. Optical emission and nanoparticle generation in Al plasmas using
ultra-short laser pulses temporally optimized by real-time spectroscopic feedback. Phys. Rev. B 82, 035430.
Guo, C., Rodriguez, G., Lobad, A., Taylor, A.J., 2000. Structural phase transition of aluminium induced by electron excitation. Phys. Rev. Lett. 84, 44934496.
Hallo, L., Bourgeade, A., Tikhonchuk, V.T., Mzel, C., Breil, J., 2007. Model and numerical simulations of the propagation and absorption of a short laser pulse
in a transparent dielectric material: blast-wave launch and cavity formation. Phys. Rev. B 76, 024101.
Hase, M., Kitajima, M., Nakashima, S., Mizoguchi, K., 2002. Dynamics of coherent anharmonic phonons in bismuth using high density photoexcitation. Phys.
Rev. Lett. 88, 067401.
Hase, M., Misoguchi, K., Harima, H., Nakashima, S., 1998. Dynamics of coherent phonons in Bi generated by ultra-short laser pulses. Phys. Rev. B 58,
54485452.
Hase, M., Mizoguchi, K., Harima, H., Nakashima, S., Tani, M., Sakai, K., Hangyo, M., 1996. Optical control of coherent optical phonons in bismuth films. Appl.
Phys. Lett. 69, 24742476.
Haske, W., Chen, V.W., Hales, J.M., Dong, W., Barlow, S., Marder, S.R., Perry, J.W., 2007. 65 nm feature sizes using visible wavelength 3-D multiphoton
lithography. Opt. Express 15, 34263436.
He, H., Freise, M.E., Heckenberg, N.R., Rubinsztein-Dunlop, H., 1995. Direct observation of transfer of angular momentum to absorptive particles from a
laser beam with a phase singularity. Phys. Rev. Lett. 75, 826829.
Heck, G., Sloss, J., Levis, R.J., 2006. Adaptive control of spatial position of white light filaments in an aqueous solution. Opt. Commun. 259, 216.
Hillebrand, R., Taubner, T., Keilmann, F., 2002. Phonon-enhanced light-matter interaction at the nanometer scale. Nature 418, 159162.
Hodgson, J.N., 1962. Optical properties of liquid indium, cadmium, bismuth and antimony. Phil. Mag. 7, 229236.
Holway Jr., L.H., 1972. Temporal behaviour of electron distribution in electric field. Phys. Rev. Lett. 28, 280283.
Holway Jr., L.H., Fradin, D.W., 1975. Electron avalanche breakdown by laser radiation in insulating crystals. J. Appl. Phys. 46, 279291.
Houbertz, R., Frohlich, L., Popall, M., Streppel, U., Dannberg, P., Brauer, A., Serbin, J., Chichkov, B.N., 2003. Inorganicorganic hybrid polymers for information
technology: from planar technology to 3D nanostructures. Adv. Eng. Mater. 5, 551555.
Iglev, H., Schmeisser, M., Simeonidis, K., Thaller, A., Laubereau, A., 2006. Ultrafast superheating and melting of bulk ice. Nature 439, 183186.
Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature 354, 56.
Ilinskii, Yu.A., Keldysh, L.V., 1994. Electromagnetic Response of Material Media. Springer, New York.
Ishioka, K., Kitajima, M., Misochko, O.V., 2006. Temperature dependence of coherent A
1g
and E
g
phonons of bismuth. J. Appl. Phys. 100, 093501.
Itina, T.E., Hermann, J., Delaporte, Ph., Sentis, M., 2002. Laser-generated plasma plume expansion: combined continuous-microscopic modeling. Phys. Rev.
E 66, 066406.
Itina, T.E., Shcheblanov, N., 2010. Electronic excitation in femtosecond laser interactions with wide-band-gap materials. Appl. Phys. A 98, 769775.
Ivanov, D.S., Zhigilei, L., 2003. Combined atomistic-continuum modelling of short-pulse laser melting and disintegration of metal films. Phys. Rev. B 68,
064114.
Jacquemot, S., Decoster, A., 1991. Z scaling of collisional Ne-like X-ray lasers using exploding foils: refraction effects. Laser Part. Beams 9, 517.
James, D.B., 1968. The thermal diffusivity of ice and water between 40C and +60C. J. Mater. Sci. 3, 540543.
Johnson, S.L., Beaud, P., Milne, C.J., Krasniqi, F.S., Zijlstra, E.S., Garcia, M.E., Kaiser, M., Grolimund, D., Abela, R., Ingold, G., 2008. Nanoscale depth-resolved
coherent femtosecond motion in laser-excited bismuth. Phys. Rev. Lett. 100, 155501.
Juodkazis, S., Kohara, S., Ohishi, Y., Hirao, N., Vailionis, A., Mizeikis, V., Saito, A., Rode, A.V., 2010. Structural changes in femtosecond laser modified regions
inside fused silica. J. Opt. 4, 124007.
Juodkazis, S., Kondo, T., Mizeikis, V., Matsuo, S., Misawa, H., Vanagas, E., Kudryashov, I., 2002. In: Proc. Bi-Lateral Conf. Optoelectronic andMagnetic Materials,
Taipei, ROC, 2526 May 2002, pp. 2729. (Available as arXiv:physics/0205025v19 May 2002).
Juodkazis, S., Misawa, H., Gamaly, E., Luther-Davies, B., Hallo, L., Nicolai, P., Tikhonchuk, V.T., 2006a. Laser-induced micro-explosion inside of sapphire
crystal: evidence of multi-megabar pressure. Phys. Rev. Lett. 96, 166101.
Juodkazis, S., Misawa, H., Hashimoto, T., Gamaly, E., Luther-Davies, B., 2006b. Laser-induced micro-explosion confined in a bulk of silica: formation of
nano-void. Appl. Phys. Lett. 88, 1.
Juodkazis, S., Sudzius, M., Mizekis, V., Misawa, H., Gamaly, E., Liu, Y., Louchev, O., Kitamura, K., 2006c. Three-dimensional recording by tightly focused
femtosecond pulses in LiNbO
3
crystal. Appl. Phys. Lett. 89, 062903.
Juodkazis, S., Mizekis, V., Misawa, H., 2009. Three-dimensional microfabrication of materials by femtosecond laser for photonics applications. J. Appl. Phys.
106, 051101.
Juodkazis, S., Mizekis, V., Sudzius, M., Misawa, H., Kitamura, K., Takekawa, S., Gamaly, E.G., Krolikowski, W.Z., Rode, A.V., 2008. Laser-induced memory bits
in photorefractive LiNbO
3
and LiTaO
3
. Appl. Phys. A 93, 129133.
Juodkazis, S., Rode, A.V., Gamaly, E.G., Matsuo, S., Misawa, H., 2003. Recording and reading of three-dimensional optical memory in glasses. Appl. Phys. B
77, 361368.
Juodkazis, S., Tabuchi, Y., Ebisui, T., Matsuo, S., Misawa, H., 2005. Anisotropic etching of dielectrics exposed by high intensity femtosecond pulses.
In: Shcherbakov, I.A., Giardini, A., Konov, V.I., Pustovoyt, V.I. (Eds.), Advanced Laser Technologies. Sept. 1015, 2004, Rome and Frascati, Italy. In: SPIE
Proc., vol. 5850. pp. 5966.
Kaganov, M.I., Lifshitz, I.M., Tanatarov, L.V., 1957. Relaxation between electrons and the crystalline lattice. Sov. Phys. JETP 4, 173.
Kamino, L., Kanai, M., Obara, M., 2003. Optical fiber delivery of intense ultra-short laser for endoscopic medical applications. In: Lasers and Electro-Optics
Society, 2003: The 16th Annual Meeting of the IEEE, vol. 1, p. 376.
Kanavin, A.P., Smetanin, I.V., Isakov, V.A., Afanasiev, Yu.V., Chichkov, B.N., Wellgehausen, B., Nolte, S., Motta, C., Tnnermann, A., 1998. Heat transport in
metals irradiated by ultra-short laser pulses. Phys. Rev. B 57, 14698.
Kandyla, M., Shih, T., Mazur, E., 2007. Femtosecond dynamics of the laser-induced solid-to-liquid phase transition in aluminium. Phys. Rev. B 75, 214107.
Kanter, H., 1970. Slow-electron mean free paths in aluminum, silver, and gold. Phys. Rev. B 1, 522.
Kattamis, N.T., Purnick, P.E., Weiss, R., Arnold, C.B., 2007. Thick film laser induced forward transfer for deposition of thermally and mechanically sensitive
materials. Appl. Phys. Lett. 91, 171120.
Kautek , Krausz F., 1998. Femtosecond optical breakdown in dielectrics. Phys. Rev. Lett. 80, 40764079.
Kawata, S., Sun, H.-B., Tanaka, T., Takada, K., 2001. Finer features for functional microdevices. Nature 412, 697698.
Keldysh, L.V., 1965. Ionization in the field of a strong electro-magnetic wave. Sov. Phys. JETP 20, 1307.
Keller, U., 2003. Recent developments in compact ultrafast lasers. Nature 424, 831838.
Keller, U., 2010a. Ultrafast solid-state laser oscillators: a success story for the last 20 years with no end in sight. Appl. Phys. B 100, 1528.
Keller, U., 2010b. Femtosecond to attosecond optics. IEEE Photonics J. 2, 225228.
Keller, U., Miller, D.A.B., Boyd, G.D., Chiu, T.H., Ferguson, J.F., Asom, M.T., 1992. Solid-state low-loss intracavity saturable absorber for Nd:YLF lasers: an
antiresonant semiconductor FabryPerot saturable absorber. Opt. Lett. 17, 505507.
Kent, A.J., Kini, R.N., Stanton, N.M., Henini, M., Glavin, B.A., Kochelap, V.A., Linnik, T.L., 2006. Acoustic phonon emission from a weakly coupled superlattice
under vertical electron transport: observation of phonon resonance. Phys. Rev. Lett. 96, 215504.
Kim, B.M., Feit, M.D., Rubenchik, A.M., Joslin, E.J., Eichler, J., Stoller, P.C., Silva, L.B., 2000. Effects of high repetition rate and beamsize on hard tissue damage
due to subpicosecond laser pulses. Appl. Phys. Lett. 76, 4001.
Kittel, Ch., 1996. Introduction to Solid State Physics. Wiley and Sons, New York.
Kolev, V.Z., Lederer, M.J., Luther-Davies, B., Rode, A.V., 2003. Opt. Lett. 28, 12751277.
Kondo, T., Matsuo, S., Juodkazis, S., Misawa, H., 2001. Femtosecond laser interference technique with diffractive beam splitter for fabrication of three-
dimensional photonic crystals. Appl. Phys. Lett. 79, 725727.
Kondo, T., Matsuo, S., Juodkazis, S., Mizeikis, V., Misawa, H., 2003. Multiphoton fabrication of periodic structures by multibeaminterference of femtosecond
pulses. Appl. Phys. Lett. 82, 27582760.
Korte, F., Nolte, S., Chichkov, B.N., Fallnich, C., Tnnermann, A., Welling, H., 2003. Towards nanostructuring with femtosecond laser pulses. Appl. Phys. A
77, 229235.
Krausz, F., Ivanov, M., 2009. Attosecond physics. Rev. Modern Phys. 81, 163234.
Kroto, H.W., Heath, J.R., OBrien, S.C., Curl, R.F., Smalley, R.E., 1985. Buckminsterfullerene. Nature 318, 162.
Kruer, W.L., 1988. The Physics of Laser Plasma Interactions. Addison-Wesley, New York.
Kukhtarev, N.V., Dovgalenko, G.E., Starkov, V.N., 1984. Influence of the optical-activity on hologram formation in photorefractive crystals. Appl. Phys. A 33,
227230.
Kukhtarev, N.V., Markov, V.B., Odulov, S.G., Soskin, M.S., Vinetskii, V.L., 1979. Holographic storage in electrooptic crystals. 1. Steady-state. Ferroelectrics 22,
949960.
Landau, L.D., 1937. Kinetic equation in the case of a Coulomb interaction. Sov. Phys. JETP 7, 203209.
Landau, L.D., Lifshitz, E.M., 1980a. Theory of Elasticity. Pergamon Press, Oxford.
Landau, L.D., Lifshitz, E.M., 1980b. Statistical Physics. Pergamon Press, Oxford.
Landau, L.D., Lifshitz, E.M., Pitaevskii, L.P., 1984. Electrodynamics of Continuous Media. Pergamon Press, Oxford.
Landolt-Brnstein, , 1983. Numerical data and functional relationships in science and technology, group III. In: Madelung, O., Schulz, M., Weiss, H. (Eds.),
Semiconductors, vol. 17. Springer-Verlag, Berlin.
Leger, J.M., Haines, J., Schmidt, M., Petitet, J.P., Pereira, A.S., Jomada, J.A.H., 1996. Discovery of hardest known oxide. Nature 383, 401.
Lenham, A.P., Treherne, D.M., Metcalfe, R.J., 1965. Optical properties of antimony and bismuth crystals. J. Opt. Soc. Amer. 55, 1072.
Lennard-Jones, J.E., Devonshire, A.F., 1939. Critical and co-operative phenomena. III. A theory of melting and the structure of liquids. Proc. R. Soc. A 169,
317.
Lenzner, M., Kruger, J., Kautek, W., Krausz, F., 1999. Incubation of laser ablation in fused silica with 5 fs pulses. Appl. Phys. A 69, 465466.
Lenzner, M., Kruger, J., Sartania, S., Cheng, Z., Spielmann, Ch., Mourou, G., Kautek, W., Krausz, F., 1998. Femtosecond optical breakdown in dielectrics. Phys.
Rev. Lett. 80, 40764079.
Lide, L.D., 2008. CRC Handbook of Chemistry and Physics, 88th ed. CRC Press, Taylor and Francis Group.
Lifshitz, E.M., Pitaevski, L.P., 1981. Physical Kinetics. Pergamon Press, Oxford.
Lin, Z., Zhigilei, L.V., 2008. Electronphonon coupling and electron heat capacity of metals under conditions of strong electronphonon nonequilibrium.
Phys. Rev. B 77, 075133.
Lindemann, F.A., 1910. The calculation of molecular vibration frequencies. Phys. Z. 11, 609612 (in German).
Lindl, J.D., Amendt, P., Berger, R.L., Glendinning, S.G., Glenzer, S.H., Haan, S.W., Kauffman, R.L., Landen, O.L., 2004. The physics basis for ignition using
indirect-drive targets on the National Ignition Facility. Phys. Plasmas 11, 339.
Liu, A.Y., Cohen, M., 1989. Prediction of new low compressibility solids. Science 245, 841842.
Luan, F., Knight, J., Russell, P., Campbell, S., Xiao, D., Reid, D., Mangan, B., Williams, D., Roberts, P., 2004. Femtosecons soliton pulse delivery at 800 nm
wavelength in hollow-core photonic bandgap fibers. Opt. Express 12, 835840.
Lukyanchuk, B. (Ed.), 2002. Laser Cleaning. Worlds Scientific, Singapore.
Lukyanchuk, B.S., Marine, W., Anisimov, S.I., 1998. Condensationof vapor andnanoclusters formationwithinthe vapor plume, producedby ns-laser ablation
of Si. Laser Phys. 8, 291302.
Lukyanchuk, B.S., Zafiropulos, V., 2002. On the theory of discoloration effect in pigments at laser cleaning. In: Lukyanchuk, B. (Ed.), Laser Cleaning. Worlds
Scientific, New Jersey, pp. 393414 (Chapter 8).
Luther-Davies, B., Gamaly, E.G., Wang, Y., Rode, A.V., Tikhonchuk, V.T., 1991. Interaction of ultra-short powerful laser pulses with matter. Laser Phys. 1,
325365.
Luther-Davies, B., Gamaly, E.G., Wang, Y., Rode, A.V., Tikhonchuk, V.T., 1992. Matter in ultra-strong laser fields. Sov. J. Quantum Electron. 22, 289325.
Luther-Davies, B., Kolev, V.Z., Lederer, M.J., Madsen, N.R., Rode, A.V., Giesekus, J., Du, K.-M., Duering, M., 2004. Appl. Phys. A 79, 1051.
Luther-Davies, B., Rode, A.V., Madsen, N.R., Gamaly, E.G., 2005. Picosecond high-repetition rate pulsed laser ablation of dielectrics: the effect of energy
accumulation between pulses. Opt. Eng. 44, 051102.
MacDonald, K.F., Fedotov, V.A., Eason, R.W., Zheludev, N.I., Rode, A.V., Luther-Davies, B., Emelyanov, V.I., 2001. Light-induced metallization in laser-
deposited gallium films. J. Opt. Soc. Amer. B 18, 331334.
MacDonald, W.M., Rosenbluth, M.N., Chuck, Wong, 1957. Relaxation of a system of particles with Coulomb interactions. Phys. Rev. 107, 350353.
Madsen, N.R., Gamaly, E.G., Rode, A.V., Luther-Davies, B., 2007. Cluster formation through the action of a single picosecond laser pulse. J. Phys.: Conf. Ser.
59, 762768.
Madsen, N.R., Rode, A.V., Gamaly, E.G., Luther-Davies, B., 2006. Expansion-limited nanocluster formation through the action of a single laser pulse. In: High
Power Laser Ablation VI. In: Phipps, C.R. (Ed.), Proceedings SPIE, vol. 6261. Paper 6261-22.
Maine, P., Strickland, D., Bado, P., Pessot, M., Mourou, G., 1988. Generation of ultrahigh peak power pulses by chirped pulse amplification. IEEE J. Quantum
Electron. 24, 398403.
Maksimov, E.G., Savrasov, D.Yu., Savrasov, S.Yu., 1997. The electronphonon interaction and the physical properties of metals. Phys. Usp. 40, 337358.
Malvezzi, M., Bloembergen, N., Huang, C.Y., 1986. Time-resolved picosecond optical measurements of laser-excited graphite. Phys. Rev. Lett. 57, 146149.
Marburger, J.H., 1975. Self-focusing: theory. Prog. Quantum Electron. 4, 35110.
Marine, W., Patrone, L., Lukyanchuk, B., Sentis, M., 2000. Strategy of nanocluster and nanostructure synthesis by conventional pulsed laser ablation. Appl.
Surf. Sci. 154155, 345352.
Matsuura, Y., Miyagi, M., Shihoyama, K., Kawachi, M., 2002. Delivery of femtosecond pulses by flexible hollow fibers. J. Appl. Phys. 91, 887889.
Mauclair, C., Mermillod-Blondin, A., Huot, N.E., Audouard, E., Stoian, R., 2008. Ultrafast laser writing of homogeneous longitudinal waveguides in glasses
using dynamic wavefront correction. Opt. Express 16, 5481.
McCarty, K.F., Nobel, J.A., Bartell, N.C., 2001. Vacancies in solids and stability of surface morphology. Nature 412, 622625.
Merlin, R., 1997. Generating coherent phonons with light pulses. Solid State Commun. 102, 207220.
Mermillod-Blondin, A., Mauclair, C., Rosenfeld, A., Bonse, J., Hertel, I.V., Audouard, E., Stoian, R., 2008. Size correction in ultrafast laser processing of fused
silica by temporal pulse shaping. Appl. Phys. Lett. 93, 021921.
Mzel, C., Bourgeade, A., Hallo, L., 2010. Surface structuring by ultra-short laser pulses: a review of photoionization models. Phys. Plasmas 17, 113504.
Miller, J.C., Haglund Jr., R.F. (Eds.), 1998. Laser Ablation and Desorption. Academic Press, San Diego.
Min Gu,, 2000. Advanced Optical Imajing Theory. Springer-Verlag, New York, (Chapter 7).
Misawa, H., Juodkazis, S. (Eds.), 2006. 3D Laser Microfabrication: Principles and Applications. Weilley-VCH, Weinheim.
Misochko, O.V., Hase, M., Ishioka, K., Kitajima, M., 2004. Observation of an amplitude collapse and revival of chirped coherent phonons in bismuth. Phys.
Rev. Lett. 92, 197401.
Misochko, O.V., Lu, R., Hase, M., Kitajima, M., 2007. Coherent control of the lattice dynamics of bismuth near the Lindemann stability limit. J. Exp. Theor.
Phys. 207, 245253.
Miura, K., Qiu, J.R., Inouye, H., Mitsuyu, T., Hirao, K., 1997. Photowritten optical waveguides in various glasses with ultra-short pulse laser. Appl. Phys. Lett.
71, 33293331.
Mlejnek, M., Wright, E.M., Moloney, J.V., 1998. Femtosecond pulse propagation in argon: a pressure dependence study. Phys. Rev. E 58, 4903.
Momma, C., Nolte, S., Chichkov, B.N., Alvensleben, F.V., Tnnermann, A., 1997. Precise laser ablation with ultra-short pulses. Appl. Surf. Sci. 109110, 1519.
More, R., Warren, K., Young, D., Zimmerman, G., 1988. A new quotidian equation of state (QEOS) for hot dense matter. Phys. Fluids 31, 30593078.
Moriarty, J.A., McMahan, A.K., 1982. High-pressure structural phase transitions in Na, Mg, and Al. Phys. Rev. Lett. 48, 809812.
Muller, H.G., Agostini, P., Petite, G., 1992. Multiphoton ionisation. In: Gavrila, M. (Ed.), Atoms in Intense Laser Fields. Academic Press.
Murray, E.D., Fritz, D.M., Wahlstrand, J.K., Fahy, S., Reiss, D.A., 2005. Effect of lattice anharmonicity on high-amplitude phonon dynamics in photoexcited
bismuth. Phys. Rev. B 72, 060301(R).
Nakata, Y., Okada, T., Maeda, M., 2002. Fabrication of dot matrix, comb, and nanowire structures using laser ablation by interfered femtosecond laser beams.
Appl. Phys. Lett. 81, 42394241.
Neev, J., Silva, L.B., Feit, M.D., Perry, M.D., Rubenchik, A.M., Stuart, B.C., 1996. Ultra-short pulse lasers for hard tissue ablation. IEEE J. Sel. Top. Quantum
Electron. 2, 790800.
Niemz, M.H., Kasenbacher, A., Strassl, M., Bcker, A., Beyertt, A., Nickel, D., Giesen, A., 2004. Tooth ablation using a CPA-free thin disk femtosecond laser
system. Appl. Phys. B 79, 269271.
Nikogosyan, D.N., 1997. Properties of Optical and Laser-Related Materials, A Handbook. John Wiley and Sons, Chichester.
Nolte, S., Momma, C., Jacobs, H., Tnnermann, A., Chichkov, B.N., Wellegehausen, B., Welling, H., 1997. Ablation of metals by ultra-short laser pulses. J. Opt.
Soc. Amer. B 14, 27162722.
Nugent, K.A., Barletta, W.A., 2010. Short-wavelength free-electron lasers. IEEE Photonics J. 2, 221224.
Nye, J.F., Berry, M.V., 1974. Dislocations in wave trains. Proc. R. Soc. Lond. Ser. A 336, 165190.
Oehler, A.E., Sdmeyer, T., Weingarten, K.J., Keller, U., 2008. 100 GHz passively mode-locked Er:Yb: glass laser at 1.5 m with 1.6 ps pulses. Opt. Express
16, 21930.
ONeil, A.T., MacVicar, I., Allen, L., Padgett, M.J., 2002. Intrinsic and extrinsic nature of the orbital angular momentum of a light beam. Phys. Rev. Lett. 88,
0503601.
Oppenheimer, J.R., 1928. Three notes on the quantum theory of a periodic effect. Phys. Rev. 31, 66.
Ouounov, D.G., Ahmad, F.R., Mller, D., Venkataraman, N., Gallagher, M.T., Thomas, M.G., Silcox, J., Koch, K.W., Gaeta, A.L., 2003. Generation of megawatt
optical solitons in hollow-core photonic band-gap fibers. Science 301, 17021704.
Palik, E.D., 1998. Handbook of Optical Constants of Solids. Academic Press, Orlando.
Patrone, L., Nelson, D., Safarov, V.I., Sentis, M., Marine, W., Giorgio, S., 2000. Photoluminescence of silicon nanoclusters with reduced size dispersion
produced by laser ablation. J. Appl. Phys. 87, 38293837.
Penn, D.R., 1980. Mean-free path of very low energy electrons: the effect of exchange and correlations. Phys. Rev. B 22, 26772682.
Perry, M.D., Stuart, B.C., Banks, P.S., Feit, M.D., Yanovsky, V., Rubenchik, A.M., 1999. Ultra-short-pulse laser machining of dielectric materials. J. Appl. Phys.
85, 68036810.
Petrov, M.P., Stepanov, S.I., Khomenko, A.V., 1991. Photorefractive Crystals in Coherent Optical Systems. In: Springer Series in Optical Sciences, Springer-
Verlag, New York.
Pickard, C.J., Needs, R.J., 2010. Aluminium at terapascal pressures. Nat. Mater. 9, 624627.
Pimenov, M., Shafeev, G.A., Smolin, A.A., Konov, V.I., Vodolaga, B.K., 1995. Laser-induced forward transfer of ultra-fine diamond particles for selective
deposition of diamond films. Appl. Surf. Sci. 86, 208212.
Pines, D., 1964. Elementary Excitations in Solids: Lectures on Phonons, Electrons, and Plasmons. W.A. Benjamin.
Popmintchev, T., Chen, M.-C., Arpin, P., Murnane, M.M., Kapteyn, H.C., 2010. The attosecond nonlinear optics of bright coherent X-ray generation. Nat.
Photonics 4, 822832.
Pouli, P., Bounos, I., Georgiou, S., Fotakis C., Doulgeridis M., 2005. Femtosecond laser cleaning of painted artefacts: is this the way forward? In: 6th
International Congress on Lasers in Conservation of Artworks Book of Abstracts, Academy of Fine Arts, Vienna, 2125 September, p. 67.
Pouli, P., Nevin, A., Andreotti, A., Colombini, P., Georgiou, S., Fotakis, C., 2009. Laser assisted removal of synthetic painting-conservation materials using UV
radiation of ns and fs pulse duration: morphological studies on model samples. Appl. Surf. Sci. 255, 49554960.
Pronko, P.P., VanRompay, P.A., Horvath, C., Loesel, F., Juhasz, T., Liu, X., Mourou, G., 1998. Avalanche ionisation and dielectric breakdown in silicon with
ultrafast laser pulses. Phys. Rev. B 58, 23872390.
Prutton, M., 1994. Introduction to Surface Physics. Clarendon, Oxford.
Qiu, J.R., Miura, K., Inouye, H., Nishii, J., Hirao, K., 1998. Three-dimensional optical storage inside a silica glass by using a focused femtosecond pulsed laser.
Nucl. Instrum. Methods Phys. Res. B 141, 699703.
Quinn, J.J., Ferrell, R.A., 1958. Electron-self-energy approach to correlation in a degenerate electron gas. Phys. Rev. 112, 812827.
Raizer, Yu.P., 1978. Laser-Induced Discharge Phenomena. Consultant Bureau, New York.
Rethfeld, B., Sokolowski-Tinten, K., von der Linde, D., Anisimov, S.I., 2002. Ultra-fast thermal melting of laser-excited solid by homogeneous nucleation.
Phys. Rev. B 65, 092103.
Riley, S.J., Parks, E.K., Mao, C.R., Pobo, L.G., Wexler, S., 1982. Generation of continuous beams of refractory-metal clusters. J. Phys. Chem. 86, 39113913.
Rode, A.V., Baldwin, K.G.H., Wain, A., Madsen, N.R., Freeman, D., Delaporte, P., Luther-Davies, B., 2008. Ultrafast laser ablation for restoration of heritage
objects. Appl. Surf. Sci. 254, 31373146.
Rode, A.V., Boschetto, D., Garl, T., Rousse, A., 2009. Transient dielectric function of fslaser excited bismuth. In: Corkum, P., de Silvestri, S., Nelson, K.A.,
Riedle, E., Schoenlein, R.W. (Eds.), Ultrafast Phenomena XVI. Springer, New York.
Rode, A.V., Gamaly, E.G., Christy, A.G., Fitz Gerald, J.G., Hyde, S.T., Elliman, R.G., Luther-Davies, B., Veinger, A.I., Androulakis, J., Giapintzakis, J., 2004.
Unconventional magnetism in all-carbon nanofoam. Phys. Rev. B 70, 054407.
Rode, A.V., Gamaly, E.G., Luther-Davies, B., Taylor, B.T., Dawes, J., Chan, A., Lowe, R.M., Hannaford, P., 2002b. Subpicosecond laser ablation of dental enamel.
J. Appl. Phys. 92, 21532158.
Rode, A.V., Luther-Davies, B., Gamaly, E.G., 1999. Ultrafast ablation with high-pulse-rate lasers. Part II: experiments on laser deposition of amorphous
carbon films. J. Appl. Phys. 85, 42224230.
Rode, A.V., Luther-Davies, B., Gamaly, E.G., 2001. Method of deposition of thin films of amorphous and crystalline microstructures based on ultrafast pulsed
laser deposition. US Patent US6312760 (priority date 11 September 1997).
Rode, A.V., Madsen, N.R., Christy, A.G., Hermann, J., Gamaly, E.G., Luther-Davies, B., 2005. In: Kuzmany, H., Fink, J., Mehring, M., Roth, S. (Eds.), Electronic
Properties of Novel Nanostructures. In: AIP Conference Proceedings, vol. 786. pp. 9699.
Rode, A.V., Zakery, A., Samoc, M., Gamaly, E.G., Luther-Davies, B., 2002a. Laser-deposited As
2
S
3
chalcogenide films for waveguide applications. Appl. Surf.
Sci. 197198, 481485.
Rousse, A., Rischel, G., Fourneaux, S., Uschmann, I., Sebban, S., Grillon, G., Balcou, Ph., Frster, E., Geindre, J.P., Audebert, P., Gauthier, J.C., Hulin, D., 2001.
Non-thermal melting in semi-conductors measured at femtosecond resolution. Nature 410, 6568.
Rozmus, W., Tikhonchuk, V.T., 1990. Skin effect and interaction of short laser pulses with dense plasmas. Phys. Rev. A 42, 74017412.
Rozmus, W., Tikhonchuk, V.T., 1992. Heating of solid targets by subpicosecond laser pulses. Phys. Rev. A 46, 78107814.
Sakdinawat, A., Attwood, D., 2010. Nanoscale X-ray imaging. Nat. Photonics 4, 840848.
Salleo, A., Taylor, S.T., Martin, M.C., Panero, W.R., Jeanloz, R., Sands, T., Genin, F.Y., 2003. Laser-driven formation of a high-pressure phase in amorphous
silica. Nat. Mater. 2, 796800.
Schaffer, C.B., Garcia, J.F., Mazur, E., 2003. Bulk heating of transparent materials using a high-repetition rate femtosecond laser. Appl. Phys. A 76, 351354.
Schaffer, C.B., Jamison, A.O., Mazur, E., 2004. Morphology of femtosecond laser-induced structural changes in bulk transparent materials. Appl. Phys. Lett.
84, 14411443.
Schizas, C., Melissinaki, V., Gaidukeviciute, A., Reinhardt, C., Ohrt, C., Dedoussis, V., Chichkov, B.N., Fotakis, C., Farsari, M., Karalekas, D., 2010. On the design
and fabrication by two-photon polymerization of a readily assembled micro-valve. Int. J. Adv. Manuf. Technol. 48, 435441.
Schoenlein, R.W., Lin, W.Z., Fujimoto, J.G., Eesley, G.L., 1987. Femtosecond studies of nonequilibrium electronic processes in metals. Phys. Rev. Lett. 58,
16801683.
Sciaini, G., Harb, M., Kruglik, S.G., Payer, T., Hebeisen, C.T., Meyer zu Heringdorf, F.-J., Yamaguchi, M., Horn-von Hoegen, M., Ernstorfer, R., Dwayne Miller,
R.J., 2009. Electronic acceleration of atomic motions and disordering in bismuth. Nature 458, 5659.
Serbin, J., Bauer, T., Fallnich, C., Kasenbacher, A., Arnold, W.H., 2002. Femtosecond lasers as novel tool in dental surgery. Appl. Surf. Sci. 197198, 737740.
Serbin, J., Egbert, A., Ostendorf, A., Chichkov, B.N., Houbertz, R., Domann, G., Schulz, J., Cronauer, C., Frhlich, L., Popall, M., 2003. Femtosecond laser-induced
two-photon polymerization of inorganicorganic hybrid materials for applications in photonics. Opt. Lett. 28, 301303.
Shen, Y.R., Bloembergen, N., 1965. Theory of stimulated Brillouin and Raman scattering. Phys. Rev. 137, 17871805.
Luo, Sheng-Nian, Arens, T.J., Asimov, P.D., 2003. Polymorphism, superheating and amorphization of silica upon shock wave loading and release. J. Geophys.
Res. 108, 2421.
Shvedov, V.G., Hnatovsky, C., Krolikowski, W., Rode, A.V., 2010. Efficient beam converter for generation of high-power femtosecond vortice. Opt. Lett. 35,
26602662.
Siwick, B.J., Dwyer, J.R., Jordan, R.E., Miller, R.J.D., 2003. An atomic level view of melting using femtosecond electron diffraction. Science 302, 1382.
Smith, G.E., Baraff, G.A., Rowel, J.M., 1964. Effective g-factor of electrons and holes in bismuth. Phys. Rev. 135, A1118.
Sokolowski-Tinten, K., Bialkowski, J., Cavalieri, A., Boing, M., Schuler, H., von der Linde, D., 1998. Dynamics of femtosecond laser-induced ablation from
solid surfaces. In: Phipps, C. (Ed.), High-Power Laser Ablation. In: Proceedings SPIE, vol. 3343. pp. 4657 (Part 1).
Sokolowski-Tinten, K., Blome, C., Blums, J., Cavalleri, A., Dietrich, C., Tarasevitch, A., Uschmann, I., Foerster, E., Kammler, M., Horn-von-Hoegen, M., von der
Linde, D., 2003. Femtosecond X-ray measurement of coherent lattice vibrations near the Lindemann stability limit. Nature 422, 287.
Sosman, R.B., 1965. The Phases of Silica. Rutgers University Press, New Bruswick.
Sparks, M., Mills, D.L., Warren, R., Holstein, T., Maradudin, A.A., Sham, L.J., Loh Jr., E., King, D.F., 1981. Theory of electron avalanche in solids. Phys. Rev. B 24,
35193536.
Spence, D.E., Kean, P.N., Sibbett, W., 1991. 60 fsec pulse generation from a self-mode-locked Ti: sapphire laser. Opt. Lett. 16, 4244.
Stuart, B.C., Feit, M.D., Herman, S., Rubenchik, A.M., Shore, B.W., Perry, M.D., 1996. Optical ablation by high-power short-pulse lasers. J. Opt. Soc. Amer. B
13, 459468.
Stuart, B.C., Feit, M.D., Rubenchick, A.M., Shore, B.W., Perry, M.D., 1995. Laser-induced damage in dielectrics with nanosecond to picosecond pulses. Phys.
Rev. Lett. 74, 22482251.
Sturman, B.I., Fridkin, V.M., 1992. The Photovoltaic and Photorefractive Effects in Noncentrosymmetric Materials. In: Ferroelectricity and Related
Phenomena, vol. 8. Gordon and Breach Science Publishers.
Sugioka, K., Cheng, Y., Midorikawa, K., 2005. Three-dimensional micromachining of glass using femtosecond laser for lab-on-a-chip device manufacture.
Appl. Phys. A 81, 110.
Sun, H., Mizeikis, V., Xu, Y., Juodkazis, S., Ye, J.-Y., Matsuo, S., Misawa, H., 2001b. Microcavities in polymeric photonic crystals. Appl. Phys. Lett. 79, 13.
Sun, H., Xu, Y., Juodkazis, S., Sun, K., Watanabe, M., Matsuo, S., Misawa, H., Nishii, J., 2001a. Arbitrary-lattice photonic crystals created by multi-photon
micro-fabrication. Opt. Lett. 26, 325.
Sutter, D.H., Steinmeyer, G., Gallmann, L., Matuschek, N., Morier-Genoud, F., Keller, U., Scheuer, V., Angelow, G., Tschudi, T., 1999. Semiconductor saturable-
absorber mirror assisted Kerr-lens mode-locked Ti: sapphire laser producing pulses in the two-cycle regime. Opt. Lett. 24, 631633.
Takeda, S., Hagiwara, Y., Obara, M., 2005. Biomedical material ablation by flexible-hollow-fiber-delivered double pulses. In: Lasers and Electro-Optics,
CLEO/Europe Conference Proceedings, p. 632.
Tallon, J.L., 1984. Communal entropy in melting and the glass, fast-ion, and superfluid transitions. Phys. Rev. B 29, 41534155.
Tallon, J.L., 1989. A hierarchy of catastrophes as a succession of stability limits for the crystalline state. Nature 342, 658660.
Tas, G., Maris, H.J., 1994. Electron diffusion in metals studied by picosecond ultrasonics. Phys. Rev. B 49, 1504615054.
Temnov, V.V., Sokolowski-Tinten, K., Zhou, P., El-Khamhawy, A., von der Linde, D., 2006. Multiphoton ionization in dielectrics: comparison of circular and
linear polarization. Phys. Rev. Lett. 97, 237403.
Tersoff, J., 1986. New empirical model for the structural properties of silicon. Phys. Rev. Lett. 56, 632635.
Teter, D.M., Hemley, R.J., Kresse, G., Hafner, J., 1998. High pressure polymorphism in silica. Phys. Rev. Lett. 80, 21452148.
Tien, A.-C., Backus, S., Kapteyn, H., Murname, M., Mourou, G., 1999. Short-pulse laser damage in transparent materials as a function of pulse duration. Phys.
Rev. Lett. 82, 38833886.
Toader, O., John, S., 2001. Proposed square spiral microfabrication architecture for large three-dimensional photonic bandgap crystals. Science 292,
11331135.
Trk, P., Varga, P., Laczik, Z., Booker, G.R., 1996. Electromagnetic diffraction of light focused through a planar interface between materials of mismatched
refractive indices: an integral representation. J. Opt. Soc. Amer. 12, 325332.
Toth, Z., Szrenyi, T., Toth, A.L., 1993. Ar
+
laser-induced forward transfer (LIFT): a novel method for micrometer-size surface patterning. Appl. Surf. Sci. 69,
317.
Tull, B.R., Carey, J.E., Sheehy, M.A., Friend, C., Mazur, E., 2006. Formation of silicon nanoparticles and web-like aggregates by femtosecond laser ablation in
a background gas. Appl. Phys. A 83, 341346.
Uhlenbeck, G.E., Ornstein, L.S., 1930. On the theory of the Brownian motion. Phys. Rev. 36, 823841.
Ullmann, M., Friedlander, S.K., Schmidt-Ott, A., 2002. Nanoparticle formation by laser ablation. J. Nanopart. Res. 4, 499509.
Uteza, O.P., Gamaly, E.G., Rode, A.V., Samoc, M., Luther-Davies, B., 2004. Galliumtransformation under femtosecond laser excitation: phase coexistence and
incomplete melting. Phys. Rev. B 70, 054108.
Vailionis, A., Gamaly, E.G., Mizeikis, V., Yang, W., Rode, A.V., Juodkazis, S., 2011. Evidence of super-dense Al: synthesis by ultrafast microexplosion. Nat.
Commun. 2, article number 445, Published 23 August 2011; doi:10.1038/ncomms1449.
Vainshtein, L.A., Shevelko, V.P., 1993. Atomic Physics for Hot Plasmas. Book News, Portland.
Valley, G.C., 1983. Short-pulse grating formation in photo-refractive materials. IEEE J. Quantum Electron. QE-19 (11), 16371645.
von der Linde, D., Schuler, H., 1996. Breakdown threshold and plasma formation in femtosecond laser-solid interaction. J. Opt. Soc. Am. B 13 (1), 216.
Watanabe, M., Sun, H.B., Juodkazis, S., Takahashi, T., Matsuoto, S., Suzuki, Y., Nishii, J., Misawa, H., 1998. Three-dimensional optical data storage in vitreous
silica. Japan. J. Appl. Phys. 37, L1527L1530.
Weast, R.C., Astle, M.J. (Eds.), 1981. CRC Handbook of Chemistry and Physics, 60th ed. CRC Press, Boca Raton.
Wharton, K.B., Boley, C.D., Komashko, A.M., Rubenchik, A.M., Zweiback, J., Crane, J., Hays, G., Cowan, T.E., Ditmire, T., 2001. Effects of nonionizing prepulses
in high-intensity lasersolid interactions. Phys. Rev. E 64, 025401.
Williams, D.A., 1992. Coherent non-equilibrium phonons: an induced Peierls distortion. Phys. Rev. Lett. 69, 25512554.
Williamson, S., Mourou, G., Li, J.C.M., 1984. Time-resolved laser-induced phase transformation in aluminium. Phys. Rev. Lett. 52, 23642367.
Woerner, H.J., Bertrand, J.B., Kartashov, D.V., Corkum, P.B., Villeneuve, D.M., 2010. Following a chemical reaction using high-harmonic interferometry.
Nature 466, 604607.
Wu, A.Q., Xu, X., 2007. Coupling of ultrafast laser energy to coherent phonons in bismuth. Appl. Phys. Lett. 90, 251111.
Yablonovitch, E., Bloembergen, N., 1972. Avalanche ionization and the limiting diameter of filaments induced by light pulses in transparent media. Phys.
Rev. Lett. 29, 907910.
Zach, L., Cohen, G., 1965. Pulp response to externally applied heat. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endod. 19, 515530.
Zafiropulos, V., Balas, C., Manousaki, A., Marakis, Y., Maravelaki-Kalaitzaki, P., Melesanaki, K., Pouli, P., Stratoudaki, T., Klein, S., Hildenhagen, J., Dickmann,
K., Lukyanchuk, B.S., Mujat, C., Dogariu, A., 2003. Yellowing effect and discoloration of pigments: experimental and theoretical studies. J. Cult. Herit. 4,
249s256s.
Zafiropulos, V., Fotakis, C., 1998. Lasers in the conservation of painted artworks. In: Cooper, M. (Ed.), Laser Cleaning in Conservation: An Introduction.
Butterworth-Heinemann, Oxford, pp. 7990 (Chapter 6).
Zakery, A., Ruan, Y., Rode, A.V., Samoc, M., Luther-Davies, B., 2003. Low-loss waveguides in ultra-fast deposited As
2
S
3
chalcogenide films. J. Opt. Soc. Amer.
B 20, 19.
Zangwill, A., 1988. Physics at Surfaces. Cambridge University Press, Cambridge.
Zavestovskaya, I.N., Afanasiv, Yu.V., Isakov, V.A., Chichkov, B.N., 2000. Extended two-temperature model of laser ablation of metals. In: Phipps, Claude R.
(Ed.), High-Power Laser Ablation III. In: Proc. SPIE, vol. 4065. pp. 349354.
Zavestovskaya, I.N., Eliseev, P.G., Krokhin, O.N., Menkova, N.A., 2008. Analysis of the nonlinear absorption mechanisms in ablation of transparent materials
by high intensity and ultra-short laser pulses. Appl. Phys. A 92, 903.
Zegrioti, I., Mailis, S., Vainos, N.A., Papakostantinou, P., Kalpouzos, C., Grigoropoulos, C.P., Fotakis, C., 1998. Appl. Phys. A 66, 579.
Zeiger, H.J., Vidal, J., Cheng, T.K., Ippen, E.P., Dresselhaus, G., Dresselhaus, M.S., 1992. Theory for displacive excitation of coherent phonons. Phys. Rev. B 45,
768.
Zeldovich, Ya.B., Raizer, Yu.P., 2002. Physics of Shock Waves and High-Temperature Hydrodynamic Phenomena. Dover, New York.
Zerr, A., Miehe, G., Serghiou, G., Schwarz, M., Kroke, E., Riedel, R., Fues, H., Kroll, P., Boehler, R., 1999. Synthesis of cubic silicon nitride. Nature 400, 340342.
Zewail, A.H., 2000. Femtochemistry: atomic-scale dynamics of the chemical bond. J. Phys. Chem. A 104, 56605694.
Zheludev, N.I., 2010. The road ahead for metamaterials. Science 328, 582583.
Zhou, P., Rajkovi, I., Ligges, M., Payer, T., Meyer zu Heringdorf, Frank-J., Horn-von Hoegen, M., von der Linde, D., 2009. Ultrafast heating of bismuth observed
by time resolved electron diffraction. In: Corkum, P., de Silvestri, S., Nelson, K.A., Riedle, E., Schoenlein, R.W. (Eds.), Ultrafast Phenomena XVI. Springer,
New York.
Zijlstra, P., Chon, J.W.M., Gu, M., 2009. Five-dimensional optical recording mediated by surface plasmons in gold nanorods. Nature 459, 410413.
Zijlstra, E., Walkenhorst, J., Garcia, M., 2008. Anharmonic noninertial lattice dynamics during ultrafast nonthermal melting of InSb. Phys. Rev. Lett. 101,
135701.
Ziman, J.M., 1960. Electrons and Phonons. Clarendon Press, Oxford.
Ziman, J.M., 1964. Principles of the Theory of Solids. Cambridge University Press, New York.

Physics of Interaction of Ultrashort Laser Pulses With Solids RPP2011

Uploaded by

Document Information

Original Description:

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

Physics of Interaction of Ultrashort Laser Pulses With Solids RPP2011

Uploaded by

Copyright:

Available Formats

Physics Reports 508 (2011) 91243

Contents lists available at SciVerse ScienceDirect

Tel.: +61 2 6125 8659; fax: +61 2 6125 0029.

I during the pulse time, defined

/4. The value of the evanescent electric field at the

E/m. The energy of oscillations averaged

h. Then collision rate reads:

h, with the help of the adiabaticity principle, v

h) by electron with quasi-momentum p (energy ) and transition to the electron state p

, has familiar the

, is that of the BoseEinstein with the different temperatures, T

is the number of particles in the high-energy tail of the Maxwell distribution at k

being the numerical coefficient that must

atomic number density in liquid and solid.

h is the Debye frequency.

is the transient pressure in the skin layer, and P is the

h / ln V is the Gruneisen coefficient, > 0, which defines the ratio of the

N. Therefore the relative error in experiments trying to

. We assume that the focal

Threshold laser fluence for ablation of metals in extreme

= na/f see Fig. 6.2.

h. In a stationary case one obtains the number density as follows:

h 6.4, and = 2.356 10

You might also like