This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Applied Surface Science 257 (2011) 5168–5171
Contents lists available at ScienceDirect
Applied Surface Science
journal homepage: www.elsevier.com/locate/apsusc
Simulation of ultrashort double-pulse laser ablation
Mikhail E. Povarnitsyn a,∗ , Tatiana E. Itina a,b , Pavel R. Levashov a , Konstatntin V. Khishchenko a
a
b
Joint Institute for High Temperatures RAS, Izhorskaya 13 Bldg 2, Moscow 125412, Russia
Laboratoire Hubert Curien, UMR CNRS 5516, 18 rue Benoît Lauras, Bât. F, 42000 St-Etienne, France
a r t i c l e
i n f o
Article history:
Available online 2 December 2010
Keywords:
Double-pulse ablation
Crater formation
Suppression of ablation
a b s t r a c t
In this paper, we study the mechanisms of femtosecond double-pulse laser ablation of metals. It was
previously shown experimentally that the crater depth monotonically drops when the delay between
two successive pulses increases. For delays longer than the time of electron–ion relaxation the crater
depth can be even smaller than that produced by a single pulse. The results of the performed hydrodynamic simulation show that the ablation can be suppressed due to the formation of the second shock
wave. The modeling results of the double-pulse ablation obtained for different delays correlate with the
experimental findings.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
2. Model
An application of the double-pulse (DP) technique in laserinduced breakdown spectroscopy [1] can give higher luminosity
of plasma in comparison with a single pulse (SP) of the same entire
energy and thus improve the accuracy of measurements. Performing femtosecond DP irradiation of different metals several authors
[2–4] surprisingly noticed a monotonic decrease in the ablation
crater depth with increasing delay delay between pulses. This effect
was observed both in the presence of ambient air [2] and in vacuum
[3,4]. In the experiment [4] with copper targets the laser fluence
of each pulse was set to be Fsingle = 2 J/cm2 , the pulse width 100 fs
and the laser wavelength 800 nm. The ablation crater depth as a
function of the delay is shown in Fig. 1.
For delays much shorter than the electron–ion relaxation time
in material ( ei = 10 ps for copper [5]) the crater depth was the same
as in the case of a SP with the laser fluence 2Fsingle = 4 J/cm2 . For the
delays close to the electron–ion relaxation time ( delay ∼ ei ), the
crater depth monotonically decreased. Similar dynamics of crater
formation was also obtained for other metals (aluminum [2], gold
[4] and nickel [3]). The authors of Refs. [2,3] also claimed that at
delay > ei the DP crater depth could be smaller than that obtained
using a SP with the same fluence as for each pulse in the DP. The
first results on the simulation of DP ablation have been published
recently [6]. In this article we continue to investigate the interplay
of physical mechanisms in DP ablation and highlight the analysis of
compression, tensile waves and phase states in copper at different
delays between pulses.
A hydrodynamic model [7,8] was developed and previously
used for simulation of a SP laser ablation of metals. The model
contains the multiphase wide-range equation of state (EOS) in tabular form allowing an accurate treatment of metastable phases
and phase transitions [9]. The specific Helmholtz free energy has
a form F(, Ti , Te ) = Fi (, Ti ) + Fe (, Te ), composed of two parts
which describe the contribution of heavy particles and electrons,
respectively. Here, is the material density and Ti and Te are the
temperatures of heavy particles and electrons. The first item, Fi (,
Ti ) = Fc () + Fa (, Ti ), in turn, consists of the electron–ion interaction
term Fc (calculated at Ti = Te = 0 K) and the contribution of thermal motion of heavy particles Fa . The solid, liquid, and gas phase
equilibrium boundaries are determined from the equality conditions for the temperature Ti , pressure Pi = 2 (∂Fi /∂)T , and Gibbs
i
potential Gi = Fi + Pi / of each phase pair. The free energy of electrons in metal Fe has a finite-temperature ideal Fermi-gas form
[10].
In Fig. 2 the phase diagram of copper with stable and metastable
phases is shown. The metastable phases are indicated by the corresponding first letters in square brackets, the stable ones—without
brackets. In simulation the metastable phases of overheated liquid [l] and overcooled gas [g] are taken into account. Currently we
do not consider overheated solid and overcooled liquid; while the
last is not significant for fast heating processes, the first can play a
noticeable role for femtosecond laser pulses with smaller fluences.
In the future we plan to investigate the influence of the solid phase
overheating on the ablation in metals.
A possibility of the existence of negative pressures in condensed
matter should be considered in fast and ultra-fast processes. In fact,
the target matter is always metastable under such conditions. The
used multiphase equation of state can work with three metastable
∗ Corresponding author.
E-mail address:
[email protected] (M.E. Povarnitsyn).
0169-4332/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apsusc.2010.11.158
Author's personal copy
M.E. Povarnitsyn et al. / Applied Surface Science 257 (2011) 5168–5171
5169
pulse with normal incidence [11] is used:
∂2 E
+ k02 ε(ωL , , Te , Ti )E = 0
∂z 2
Fig. 1. Crater depth as a function of the delay between the pulses. () Experiment
[4], (△) simulation (this work), solid line—fit of experiment [4], and dash line—crater
depth in SP simulation (this work).
phases at negative pressures: metastable solid [s], metastable melting [s + l] and metastable liquid [l]. The boundary between the
metastable state of liquid at positive and negative pressures can
be estimated by the position of the isobar P = 0 GPa in Fig. 2. The
regions of metastable liquid and metastable melting are artificially
restricted from below at some temperature (and the region of
metastable solid from the left at some density) to simplify the calculation of thermodynamic functions at negative pressure; these
restrictions have no influence on the results.
The multiphase equation of state returns the information about
the phase state and phase and metastable states boundaries; this
fact allows us to switch to the kinetics and fragmentation mechanisms depending on the phase state of matter.
In the simulations of SP experiments, the approximation of
laser energy absorption by a simple Lambert–Beer’s law is satisfactory; using this approach we obtained good agreement with SP
experimental findings for several metals [9]. In DP ablation experiments, however, this simplified technique fails for the second pulse
because of surface smearing and inhomogeneity of target properties produced by the first pulse. To overcome this problem the
Helmholtz wave equation for the electric component of the laser
(1)
in the region z1 ≤ z ≤ z2 for the first and second pulses. Here E(z, t)
is the x-component of the slowly varying in time laser field amplitude, Ex (z, t) = E(z, t) exp( − iωL t), k0 = ωL /c is the laser wave vector
with the laser frequency ωL and the light speed in vacuum c. The
boundary conditions for ∂E/∂ z at z = z1 and z = z2 are lengthy and
can be found elsewhere [11].
A wide-range model of complex permittivity of matter was used
in the form ε = (1/2)(εmet + εpl ) − (1/2)(εmet − εpl ) tanh(2Te /TF − 2),
where TF is the Fermi temperature. This approximation describes
the Drude-like limit εmet (ωL , , Te , Ti ) in solid state (for copper
εmet = − 27.6 + 2.7i at normal conditions and 800 nm wavelength
[12]) and the plasma limit εpl (ωL , , Te ) in hot ionized gas state [13].
In metals for chosen parameters of laser wavelength and intensity
the inverse Bremsstrahlung mechanism of absorption is dominant
and thus the heat source can be expressed as QIB = I0 k0 Im{ε}|E/E0 |2 ,
I0 is the peak laserintensity, E is the solution to the Helmholtz
8I0 /c is the maximum of laser electric field
equation and E0 =
in vacuum.
For realistic treatment of the fragmentation process, the model
based upon the homogeneous nucleation mechanism [9] is used.
Three basic stages are assumed to take place in the fragmentation model: (i) spontaneous appearance of critical size gas bubbles,
(ii) growth of these bubbles, and (iii) confluence of bubbles and
final relaxation of pressure and temperature. All these stages are
present in the code. Simulations of the DP ablation show that pressure in copper can be as low as Pmin = − 3.6 GPa [6]. Such pressure
values were shown to result in the mechanical spallation in the
liquid phase. In the solid phase, however, the spallation strength
for high strain rates (−1 ∂ /∂ t ∼ 1010 s−1 ) is close to its theoretical
limit (the pressure on the cold curve at the same density) which
is substantially lower than Pmin . For example, at = 8 g/cm3 the
multiphase EOS gives ≈−10 GPa at Ti = 0.
The experiment [4] for copper target was reproduced numerically applying two Gaussian laser pulses with normal incidence,
100 fs width, 800 nm wavelength and fluence 2 J/cm2 each. The
peak intensity of the first and second pulses corresponds to
moments t = 0 ps and t = delay , respectively, and initial surface position is at 0 nm.
3. Results and discussion
Fig. 2. Phase diagram of copper. CP: critical point; bn: the binodal; sp: the spinodal;
g: stable gas; l: stable liquid; s: stable solid; s + l: stable melting; l + g: liquid–gas
mixture; s + g: solid–gas mixture; [g]: metastable gas; [l]: metastable liquid; [s + l]:
metastable melting; [s]: metastable solid. Solid (green) lines—isobars P = 0, −1, and
−3 GPa. Solid (red) line—phase trajectory of the target layer with initial coordinate
z = 60 nm for delay = 0, dash (magenta) line—for delay = 10 ps, and dash–dot line—for
delay = 100 ps. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of the article.)
To illustrate the process of the DP ablation we choose a z = 60 nm
depth layer of the copper target at initial time t = − 0.5 ps. Then the
evolution of thermodynamic and kinematic parameters of this layer
during the simulation with delay = 0, 10 and 100 ps is analyzed. In
Fig. 2 the phase trajectories of the referred above layer by solid
(0 ps), dash (10 ps) and dash–dot (100 ps) lines are shown.
All the trajectories originate in one and the same point at normal conditions for copper. The trajectory with delay = 0 (solid line
in Fig. 2) is equivalent to the SP ablation with the doubled fluence 4 J/cm2 . It crosses the melting region almost along the normal
isochore and then evolves into the liquid phase where it reaches
supercritical temperatures because of the fast heating of electrons
by the laser pulse and energy transfer from electrons to ions during
t ∼ ei . The heating of ions gives rise to the formation of the shock
wave (SW) spreading into the target; the compression of matter
is also visible along the delay = 0 trajectory. Then the tensile wave
(TW) arrives at layer under consideration, and the matter begins
to expand. The trajectory enters into the metastable liquid region
at negative pressures where the mechanical fragmentation process
begins; this leads to the fluctuations of thermodynamic parameters
along the trajectory.
Author's personal copy
5170
M.E. Povarnitsyn et al. / Applied Surface Science 257 (2011) 5168–5171
For the trajectory with delay = 10 ps (dash line in Fig. 2) the
fluence of the first pulse is 2 J/cm2 , therefore the heating and
compression of the considered layer is lower than in the case of
delay = 0 ps. At t = 10 ps (see the arrow in Fig. 2) the second laser
pulse gives rise to the sharp increase in temperature; for the dielectric permittivity model used we obtain even higher temperatures
than those in the case of delay = 0 ps. Then again the trajectory
comes into the metastable liquid region at negative pressures
because of the TW propagation.
Finally, the trajectory with delay = 100 ps (dash–dot line in Fig. 2)
shows the most peculiar behavior. During the first 10 ps it obviously coincides with the trajectory with delay = 10 ps, but then the
TW comes into play, and the density and temperature along the
trajectory diminish up to the metastable liquid region at negative
pressure. One can see that pressure on the trajectory drops below
−3 GPa, and then the fragmentation process gives rise to the relaxation of pressure up to 0 GPa. At this moment (100 ps, shown by the
arrow in Fig. 2) the second pulse arrives, thus establishing the rise
of temperature and density variation. The last can be explained by
the complex non-uniform distribution of density and temperature
in the fragmented material in which the second pulse is absorbed.
Nevertheless, the density higher than at t = 100 ps is reached before
the second TW provides for the drop of temperature and density and causes the trajectory to enter into the metastable liquid
phase again where fragmentation of the target material continues.
It should be noted however, that the second TW is weaker than the
first one, that is why the ablation process and fragmentation is less
pronounced in this case.
The previous calculations [8,9] have demonstrated that only the
melted region can be ablated due to the TW propagation. Further, if
the second pulse arrives when the TW goes through the liquid layer,
this second pulse reheats the nascent ablation plume. As a result, a
high-pressure region is generated in the vicinity of the initial target
surface. In this reheated region the SW forms and moves into the
bulk suppressing the action of the first TW. Therefore, the depth of
the crater drops. We are going to discuss this effect in detail using
the time-space diagrams (Figs. 3 and 4).
For delay = 0 ps the formation of melting front (s + l cyan region)
is seen. The thickness of the melted region reaches about 180 nm
by the time delay of 40 ps after the first pulse maximum. After the
heating of ions, the SW arises with the maximal pressure behind
the front ∼35 GPa (Fig. 3b). At the same time the high-pressure
heat-affected zone begins to expand in two rarefaction waves; one
appears on the free surface of the target and moves into the bulk, the
other moves from the end of the heat affected zone towards the free
surface. As a result of the rarefaction waves interaction a zone of
negative pressure arises and the minimal pressure trace in this TW
is schematically shown in Fig. 3a and b. Mechanical fragmentation
starts after the propagation of the TW and by the moment 60 ps
the voids in liquid layer have been formed. The speed of the fastest
liquid layer on the interface between liquid (yellow) and liquid–gas
mixture (orange) is about 500 m/s. The boundary between the highdensity (metastable liquid) and low-density (liquid–gas mixture)
phases is denoted by b1 in both Fig. 3a and b; it has a negative
slope and moves away from the target; the same can be observed
for the region of metastable melting [s + l] (magenta in Fig. 3a). The
liquid–gas mixture and gas (red for the metastable phase and dark
blue for the stable) fractions form the front of the nascent ablation
plume.
For 100 ps delay one can clearly see in Fig. 4b that two SWs
(SW1 and SW2) are formed; the pressure behind the first SW is
approximately 3 times higher than that behind the second one.
These two SWs are separated by the TW, the path of which (TW1)
is also shown in Fig. 4a and b.
The second pulse produces an extension of substance (dash line
TW2 in Fig. 4a and b), but this effect is weak and does not con-
Fig. 3. z–t diagrams of phase states (a) and pressure in GPa (b) in the target for
delay = 0. Presented phase states are: s: solid (olive); s + l: melting (light blue); l:
liquid (blue); l + g: liquid–gas mixture (orange); g: gas (dark blue); [s]: metastable
solid (green); [l + g]: metastable melting (magenta); [l]: metastable liquid (yellow);
[g]: metastable gas (red). For interpretation of the references to color in this figure
legend, the reader is referred to the web version of the article.)
tribute much in fragmentation of matter; the pressure behind the
second SW (SW2) drops more slowly than behind the first one
(SW1). The boundary between high-density and low density phases
is also shown in Fig. 4a and b by b1–b3. It is remarkable that after
the first TW the boundary b1 has a negative slope, then after the
second SW b2 has a positive slope and after the second TW b3
is almost vertical, or, in other words, its velocity is nearly zero. It
means that the second pulse produces an increase in the atomization of the ablation plume head and compression and recoil of the
Fig. 4. z–t diagrams of phase states (a) and pressure (b) in the target at delay = 100 ps.
Designations are the same as in Fig. 3.
Author's personal copy
M.E. Povarnitsyn et al. / Applied Surface Science 257 (2011) 5168–5171
plume tail. This, in turn, leads to the decrease of the crater depth
(see Fig. 1).
To compare the results of modeling with experiment of Ref. [4]
the crater depth for different delays has been estimated. To this
end we integrate the mass flux through the plane z = 0 (initial surt
face position) by using the formula (t) = 0−1 0 (u)|z=0 dt ′ , where
0 is the initial material density and u is the material velocity. The
dynamics of crater formation is presented elsewhere [6]; in this
work in Fig. 1 the ultimate crater depth for t → ∞ (triangles) is presented. Already for the 10 ps delay the crater depth is similar to the
one for a SP; for longer delays (50 and 100 ps) the crater depth was
obtained to be smaller than that in the case of SP.
4. Summary
We summarize here the basic stages of the DP ablation process.
The first laser pulse creates the SW, which propagates into the target. Behind the SW a TW forms and propagates through the liquid
layer resulting in the mechanical fragmentation and ablation of this
layer. When delay is much shorter than ei , only one SW and one
TW appear. In this case, the ablation crater is formed by both pulses
simultaneously as in the case of a SP of the same energy. When
the delay is on the order of the relaxation time, the second pulse
creates the second SW thus reducing the intensity of the first TW
and the depth of the ablation crater decreases. Finally, for delays
longer than the electron–ion relaxation time the ablation crater is
formed by the first pulse only, whereas the second pulse reheats
and decelerates the ablated material; the crater depth in this case
can be smaller than that for the SP.
Acknowledgments
The authors thank Dr. V.V. Milyavskiy for helpful discussions.
This work was supported by the Centre National de la Recherche
5171
Scientifique and the Russian Academy of Sciences (French-Russian
collaboration project CNRS-ASRF 21276), the Russian Foundation
for Basic Research (grants 08-08-01055 and 09-08-01129), and the
Ministry of Education and Science of the Russian Federation (the
Federal targeted program ‘Scientific and scientific-pedagogical personnel of the innovative Russia’ 2009–2013).
References
[1] D.A. Cremers, L.J. Radziemski (Eds.), Handbook of Laser-Induced Breakdown
Spectroscopy, Wiley, New York, 2006.
[2] A. Semerok, C. Dutouquet, Ultrashort double pulse laser ablation of metals, Thin
Solid Films 453–454 (2004) 501–505.
[3] T. Donnelly, J.G. Lunney, S. Amoruso, R. Bruzzese, X. Wang, X. Ni, Double pulse
ultrafast laser ablation of nickel in vacuum, J. Appl. Phys. 106 (2009) 013304.
[4] S. Noël, J. Hermann, Reducing nanoparticles in metal ablation plumes produced
by two delayed short laser pulses, Appl. Phys. Lett. 94 (2009) 053120.
[5] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 87th ed., CRC Press,
2007, pp. 2006–2007.
[6] M.E. Povarnitsyn, T.E. Itina, K.V. Khishchenko, P.R. Levashov, Supression of
ablation in femtosecond double-pulse experiment, Phys. Rev. Lett. 103 (2009)
195002.
[7] M.E. Povarnitsyn, T.E. Itina, P.R. Levashov, K.V. Khishchenko, Multi-material
two-temperature model for simulation of ultrashort laser ablation, Appl. Surf.
Sci. 253 (2007) 6343–6346.
[8] M.E. Povarnitsyn, T.E. Itina, M. Sentis, P.R. Levashov, K.V. Khishchenko, Material decomposition mechanisms in femtosecond laser interactions with metals,
Phys. Rev. B 75 (2007) 235414.
[9] M.E. Povarnitsyn, K.V. Khishchenko, P.R. Levashov, Phase transitions in femtosecond laser ablation, Appl. Surf. Sci. 255 (2009) 5120–5124.
[10] L.D. Landau, E.M. Lifshits, Statistical Physics, Pergamon Press, Oxford, 1980.
[11] N. Andreev, V. Fortov, E. Kazakov, K. Khishchenko, O. Kostenko, T. Kuehl, P.
Levashov, M. Povarnitsyn, O. Rosmej, Y. Zhao, J. Duan, J. Liu, Interaction of
phelix laser pulses with nanostructured targets: experiment and modeling,
in: V.E. Fortov, al. et (Eds.), Physics of Extreme States of Matter—2009, IPCP,
Chernogolovka, 2009, pp. 18–20.
[12] E.D. Palik, Handbook of Optical Constants of Solids, Academic Press INC, 1985.
[13] M. Veysman, B. Cros, N.E. Andreev, G. Maynard, Theory and simulation of short
intense laser pulse propagation in capillary tubes with wall ablation, Phys.
Plasmas 13 (2006) 053114.