Optimizing Energy Absorption for Ultrashort Pulse Laser Ablation of
Fused Silica
Nicolas Sanner, Maxime Lebugle, Nadezda Varkentina, Marc Sentis and Olivier Utéza
Aix-Marseille Univ., CNRS, LP3 UMR 7341, 163 avenue de Luminy C.917, 13288 Marseille, France
Keywords: Ultrashort Laser-matter Interaction, Dielectric Materials, Ablation.
Abstract: We investigate the ultrafast absorption of fused silica irradiated by a single 500 fs laser pulse in the context
of micromachining applications. As the absorption of the laser energy is rapid (~fs), the optical properties of
the material evolve during the laser pulse, thereby yielding a feedback on the dynamics of absorption and
consequently on the amount of energy that is absorbed. Through complete investigation of energy
absorption, by combining “pump depletion” and “pump-probe” experiments in a wide range of incident
fluences above the ablation threshold, we demonstrate the existence of an optimal fluence range, enabling to
turn transiently the material into a state such that each photon is optimally utilized for ablation.
1 INTRODUCTION
Ultrashort laser pulses are extremely interesting and
powerful tools for laser-matter interaction. The
spatial accuracy of energy deposition into matter
combined with the shortness of the energy
deposition step enable to reach relatively high
intensities (10
13
-10
14
W.cm
-2
) while using low
energies, capable to push matter into strongly non-
equilibrium conditions (Gamaly, 2011). The fast
creation of highly excited plasmas at solid densities
provides notably cutting-edge capabilities for
controlled modification of matter. Near-IR
wavelengths of most current femtosecond laser
sources are particularly well adapted to trigger
highly-nonlinear absorption in transparent dielectric
materials (silica, sapphire, diamond…), whose range
of applications for material science and photonics is
extremely extended.
First, the possibility to rapidly transform such
insulating materials into plasmas within 10
-15
second
timescale and on micrometric dimensions is of
strong interest for a vast panel of future ultrafast
applications (Sugioka and Cheng, 2013): change of
electrical properties for ultrafast laser-induced
electronic switches, change of optical properties for
ultrafast plasma mirrors and plasma optics, structural
change of atomic lattice (ultrafast melting) enabling
to change the refractive index, etc. Second, for
intensities above the material ablation threshold, the
final result is highly interesting for material
processing at micro- and nano-scales: voids for
memories, channels for microfluidic,
ophthalmic/neuronal surgery, material cutting,
drilling, surface structuration (e.g. for metamaterials
or plasmonics), etc. All these cutting-edge
applications for future technology and industry are
based on ultrafast laser-induced ionization of
dielectric solids, which provides time- and space-
confinement of energy for matter transformation.
Even for ablation, for which several temporal
orders of magnitude separate the very first stage of
interaction (ionization) and the removal of matter
produced by one single pulse, experimental
observations establish that the result of ablation is
closely linked to and driven by the mechanisms and
the dynamics of absorption (Balling and Schou,
2013). Yet, it turns out that for material ablation, the
total amount of energy that is deposited in the
material is of highest importance, since part of the
interacting energy is not optimally used for ablation
and is finally relaxed into unwanted effects in the
material (recast products, rims, thermal load,
residual mechanical constraints, etc.) (Ben-Yakar et
al., 2007). Although these effects are reduced in the
femtosecond pulse regime compared to longer pulse
durations, they are nevertheless no more negligible
when dealing with the micrometer scale. For the best
control of the outcome of ablation, it is therefore
important to provide no more than the amount of
laser energy necessary to machine the calibrated
modication of the material surface.
Sanner, N., Lebugle, M., Varkentina, N., Sentis, M. and Utéza, O.
Optimizing Energy Absorption for Ultrashort Pulse Laser Ablation of Fused Silica.
DOI: 10.5220/0005685300690073
In Proceedings of the 4th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2016), pages 71-75
ISBN: 978-989-758-174-8
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
71
In this context, it is important to precisely know
the response of the material, both temporally and
under various intensities of irradiation, to optimize
the result of ablation. Dielectric materials being
initially transparent to near infrared laser light, non-
linear absorption mechanisms are required to bridge
the band gap and promote free electrons in the
conduction band (Stuart et al., 1996). These seed
electrons are further accelerated through inverse
Bremsstrahlung (IB) and multiplied by impact
ionization (II) if their kinetic energy is sufcient,
leading to an avalanche multiplication phenomenon.
Strong laser absorption therefore takes place in the
rst (hundred) nanometers at the surface of the
material, where the material has been turned opaque.
It may yield to damage or ablation of the material, if
the quantity of laser energy absorbed by the material
is sufciently high and properly adjusted (Jia et al.,
2004, Chimier et al., 2011). At densities higher than
the critical plasma density, the material optical
properties evolve during the laser pulse, thereby
yielding a feedback on the dynamics of absorption
(Chowdhury et al., 2005, Hernandez-Rueda et al.,
2012) and consequently on the amount of energy
that is absorbed. This actually depends on the
characteristics of the plasma layer which is created
at the surface of the sample by the pulse itself.
To explore this, we present here a complete
investigation of energy absorption, by combining
“pump-probe” (Lebugle et al., 2014) and “pump
depletion” (Varkentina et al., 2013) experiments in a
wide range of incident fluences above the ablation
threshold. The first offers the ability to follow in
time the free-electrons plasma buildup and its
transient optical properties, while the latter provides
information concerning the absorbed energy. This
extensive information finally provides a
comprehensive picture of material response and
strategies for optimizing the amount of energy that is
eventually deposited in the material.
2 DYNAMICS OF ABSORPTION
2.1 Setup and Determination of Time
Zero
In order to probe the absorption during the
irradiating pulse, and therefore provide insights into
the mechanisms and dynamics of absorption, two
main conditions must be fulfilled (Lebugle et al.,
2015):
(i) Use a probe pulse much shorter than the pump,
to obtain high temporal sampling
(ii) Accurately determine the ‘time zero’, i.e. the
zero-delay instant between the peaks of the pump
and the probe pulses.
To address the first condition, we use a 500-fs pump
pulse, delivered by a commercial source (emitting at
1025 nm) and a 60-fs probe pulse at the same
wavelength, issued from self-phase modulation in a
microstructured fiber and further compressed by
chirped mirrors. The probe pulse duration is
therefore a factor of eight shorter than the pump
pulse, thus allowing accurate temporal sampling of
the optical transient states. The experimental setup is
sketched in Figure 1. The pump beam is focused
with normal incidence at the surface of the SiO
2
sample with horizontal polarization, and probed by
the s-polarized pulse at an angle of 20°. The two
beams are carefully mode-matched, with radii of ~13
µm at 1/e² in intensity, measured by imaging of the
beams onto a CCD camera with adapted
magnification. The reected and transmitted parts of
the probe pulse are recorded by photodiodes after
spectral, spatial (lens and pinhole) and polarization
ltering.
The second condition is of highest importance to
retrieve the exact temporal dynamics of mechanisms
involved in the interaction at the timescale of the
pulse. Instead of using nonlinear crystal in place of
the SiO
2
sample to perform SHG or THG for
determining the ‘time zero’, we directly rely on the
nonlinear second order coefficient of fused silica
itself (
310

²/). This offers the
advantage to work directly in-situ, and suppresses
the difculty to further position the surface of the
target material exactly in the same plane as the non-
linear crystal used once it has been removed. This
configuration is in fact a polarization-gated cross-
correlation FROG experiment (PG-XFROG). It
enables to characterize in situ the probe pulse
(intensity and phase), but more importantly to
accurately calibrate the temporal overlap between
the pump and probe pulses. For that, the polarization
of the pump pulse is set to be 45° with respect to that
of the probe one, and the photodiode monitoring the
p-component of the polarization of the probe pulse is
also replaced by a spectrometer. Upon irradiation by
a low-energy pump pulse, transient birefringence
appears inside the fused silica sample during the
pump pulse, so that the polarization of the probe
beam is slightly turned to allow spectral analysis by
the spectrometer of the signal coming through the p-
transmitting polarizing beam-splitter. By time
delaying the probe pulse with respect to the pump
pulse, the spectrally resolved PG-XFROG signal is
recorded. We finally retrieve the probe pulse
PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology
72
duration of 60 fs, in excellent agreement with
independent measurement performed with a second-
order autocorrelator. More importantly, the zero-
time delay is dened by the instant of superposition
of the peaks of the pump and probe pulses, with a
precision of ± 30 fs.
Figure 1: Experimental set-up. A 500-fs pulse is split into
two arms, and the probe pulse is further compressed down
to 60 fs using spectral broadening in a micro-structured
bre, with the second-order spectral phase adjusted with a
pair of chirped mirrors. The two pulses are then focused
on the surface of the fused silica target, controlling the
delay between them. This pump–probe set-up, which is
designed to measure the optical dynamics during laser
excitation, makes also possible the direct calibration of the
time zero in a PG-XFROG conguration as well as
measuring the probe pulse (intensity and phase), provided
that the polarizations are properly set (see insets).
2.2 Dynamics of Transient Absorption
Now turning back to pump-probe experimental
configuration, transient optical properties of the
fused silica sample that transforms into plasma
during the pump pulse are measured as a function of
time delay for a set of increasing uences. Note that
these experiments are performed in single pulse
regime. Moreover, although the peak powers
attained with the highest fluences used here (~60
MW) largely exceed the critical power for self-
focusing in the fused silica sample (~3 MW at this
wavelength), in our experiments we do not detect
any signs of sub-surface or bulk laser-induced
damage.
For pump uences below the ablation threshold

5.8/², no signicant variations of the
probe signals are recorded, whereas for all uences
higher than the ablation threshold the plasma
properties strongly evolve during the pulse. Note
that we use a low-intensity probe pulse, whose
intensity (

210

/²) is well below
any material modication threshold for a 60 fs pulse.
Indeed, we veried that no depletion of the probe
pulse occurs when the pump beam is blocked. We
also experimentally checked that no change is
observable in the response of the plasma as a
function of the probe energy. Thus, the plasma
absorption read by the probe mainly corresponds to
absorption by IB, i.e. the heating of the free carriers
(produced by the pump pulse) through collisions of a
photon with an electron in the eld of the parent
atom. The magnitude of this one-photon absorption
of the probe is then directly representative of the free
electron density of the plasma.
Transient transmissivity and reectivity
recorded during the pulse are shown in Figure 2. A
large set of uences is explored, from 0.7

to
3.9

. All the curves show absolute values, so that
their initial values measured before the pump pulse
account for Fresnel reection. With these
measurements, absorptivity can be deduced using
the energy conservation law:  1 .
Temporally, the curves present swift changes in
the second half of the pump pulse. They saturate
after the pulse at a time delay around 0.5 ps for all
uences. Concerning the reflectivity , we observe
that a plasma mirror effect occurs during the pulse,
and therefore partially limits energy deposition. This
evidences the formation of a transient overdense
plasma during the excitation pulse.
Now concerning absorption and its transient
dynamics, instantaneous absorption is maximal at
the end of the pulse (delay of ~500 fs). Transient
plasma absorption is higher and starts earlier (before
pump maximum) for growing fluences. At this point,
it is important to recall that what is measured is the
absorption of the probe pulse by Inverse
Bremsstrahlung (IB) at 1025 nm, meaning that the
probe pulse “reads” the plasma absorption
properties. Yet, since the probe and pump pulses
have the same wavelength and same spot size, this
demonstrates that IB absorption of the pump pulse
itself has a growing contribution during the second
half of the pulse, beginning earlier and earlier upon
increase of the fluence. The magnitude of IB
absorption reaches 40% for the highest fluence
3.9

, which is an indication that the predominant
mechanism of pump absorption is impact ionization.
However, we must remind that this is a mean value,
averaged on the whole spatial extend of the 2D
Gaussian beam profile. Thus, this effect should be
much more pronounced at the center of the beam.
Finally, these combined observations of R and T
Optimizing Energy Absorption for Ultrashort Pulse Laser Ablation of Fused Silica
73
suggest the existence of a uence range that may
enable to maximize the transfer of energy from the
laser pulse to the material. Ideally, overall
reflectivity should keep moderate, together with high
level of absorption. In this perspective, it is
interesting to point out the saturation of absorptivity
when fluence increases. Indeed, only a small
variation is recorded between the curves for 2.5

and 3.9

. This is analyzed with respect to
integrated measurements in the next section, in order
to identify the optimal fluence range for ablation.
Figure 2: Transient optical properties of the fused silica
sample during the pulse (retrieved from pump-probe
experiments). Each data point is the average of 5
independent measurements. The pump pulse is also
represented with a Gaussian function of 500 fs duration
FWHM (arbitrary vertical units).
3 EFFICIENCY OF ABSORPTION
FOR ABLATION
With the aim to link absorption with ablation, we
also implemented a “pump-depletion” experiment.
In a complementary manner with the pump-probe
experiment, we thus explore the total absorption of
the pump beam, temporally integrated over the
whole pulse length.
This setup uses the same kind of experimental
arrangement as presented in Figure 1, but only the
pump pulse is present. The same incident laser beam
is focused on the sample surface, but here the beam
radius is smaller than in the previously presented
experiment (6.3 µm at 1/e²). However, the measured
fluence for ablation threshold is identical, making
therefore possible to compare directly these two
experiments. The transmitted and reflected fractions
of the incident pump pulse are measured with proper
photodiodes and filtering. Note that here these
signals are integrated both spatially (over the
radially Gaussian beam distribution) and temporally
(over the whole pulse length).
Post-mortem analyses of the ablated craters are
performed by atomic force microscopy, and we
define the absorption efficiency

as the ratio of
the ablated volume V to the absorbed pulse energy:

/

. Note that this quantity is different
from the usual ablation efficiency used in other
works (Utéza et al., 2011) based on incident rather
than absorbed energy. Figure 3 shows that the laser
energy is more and more efficiently absorbed for
growing fluences, until the curve reaches a plateau
(~0.65 µm
3
/µJ) followed by a slow decrease. This
saturation appears for fluences that are well in
accordance with the saturation of absorption
measured in the in the previous section, i.e. for
fluences around three times the ablation threshold.
A fluence range ΔF
eff,abs
corresponding to
efficient absorption for ablation can be interestingly
identified. Its limits are defined with respect to the
fluences yielding ~90% of the maximum absorption
efficiency. According to that criterion, the
absorption is efficient from ~2 to ~8

. It is
important to determine the range of ΔF
eff,abs
when
one is intended to develop micromachining
applications. Indeed, it helps to define an optimal
working fluence range, where maximal absorption is
attained and the energy deposition is not too strongly
disturbed by the plasma screening effect. From the
viewpoint of efficient material removal, reaching
this precise state of the material optimizes
absorption efficiency, i.e. enables the optimal
utilization of photons for material ablation.
Moreover, the best ablation quality is also obtained
for fluence values corresponding to the ΔF
eff,abs
range, as illustrated by the typical crater morphology
displayed on figure 3 (corresponding to 3.9

).
Below 2

, the energy is not sufficient to produce a
smooth ablation (high surface roughness), and above
8

we observe a higher level of debris and a lower
efficiency removal, which is often not suitable for
applications.
PHOTOPTICS 2016 - 4th International Conference on Photonics, Optics and Laser Technology
74
Figure 3: Absorption efciency

/

, as a
function of normalized laser uence. ΔF
eff,abs
determines a
uence zone of high efciency of absorption, by taking a
criterion of 90% of the maximum value. For illustration,
an AFM picture of a typical crater obtained in this range is
shown (corresponding here to 3.9

, with a crater
diameter of 9.3 µm and a maximum depth of 210 nm).
4 CONCLUSIONS
In conclusion, the dynamics of transient optical
properties of silica during a 500-fs laser pulse is
retrieved and compared to time-integrated ones,
obtained by independent energy balance
measurement. The high temporal resolution
achieved in this experiment enables to reveal the
fine dynamics of transient material states. In
particular, for ablation purposes, it appears that an
optimal fluence range (with respect to ablation
efficiency) is identified between two and four times
the ablation threshold. This corresponds to a very
particular situation: the material has already turned
strongly absorbing, whereas its reflective properties
are still low. This is taking place mostly in the
second half of the pulse. In this regime, the incident
laser energy is best absorbed.
By demonstrating that a careful choice of the
incident fluence promotes a large coupling of the
incoming laser energy to the excited material, this
work is of particular importance in the context of
micromachining process, for which a better control
of the material excitation is sought. Our results
highlight the crucial role of transient optical
properties during the laser-matter interaction in the
regime of ablation, and open a comprehensive way
toward designing dedicated user-defined temporal
excitation profiles.
ACKNOWLEDGEMENTS
Financial support of the French National Agency of
Research (ANR) - Nanomorphing-07-BLAN-0301-
03 and the Region Provence-Alpes-Côte d’Azur and
Department of Bouches-du-Rhône is gratefully
acknowledged.
REFERENCES
Gamaly, E., 2011. Femtosecond laser-matter interactions.
Pan Stanford Publishing.
Sugioka, K., Cheng, Y., 2013. Ultrafast Laser Processing:
From Micro- to Nanoscale, Pan Stanford Publishing.
Balling, P., Schou, J., 2013. Femtosecond-laser ablation
dynamics of dielectrics: basics and applications for
thin lms. Rep. Prog. Phys. 76 036502.
Ben-Yakar, A., Harkin, A., Ashmore, J., Byer, R.L.,
Stone, H.A., 2007. Thermal and uid processes of a
thin melt zone during femtosecond laser ablation of
glass: the formation of rims by single laser pulses. J.
Phys. D: Appl. Phys. 40 1447–1459.
Stuart, B.C., Feit, M.D., Herman, S., Rubenchik, A.M.,
Shore, B.W., and Perry, M.D., 1996. Nanosecond-to-
femtosecond laser-induced breakdown in dielectrics.
Phys. Rev. B 53, 1749.
Jia, T., Xu, Z., Li, R., Feng, D., Li, X., Cheng, C., Sun, H.,
Xu, N., and Wang, H., 2004. Mechanisms in fs-laser
ablation in fused silica J. Appl. Phys. 95(9), 5166.
Chimier, B., Utéza, O., Sanner, N., Sentis, M., Itina, T.,
Lassonde, P., Légaré, F., Vidal, F. and Kieffer, J.C..
2011. Damage and ablation thresholds of fused-silica
in femtosecond regime Phys. Rev. B 84, 094104.
Chowdhury, I.H., Wu, A.Q., Xu, X., and Weiner, A.M.
2005. Ultra-fast laser absorption and ablation
dynamics in wide-band-gap dielectrics Appl. Phys. A
81, 1627.
Hernandez-Rueda, J., Puerto, D., Siegel, J., Galvan-Sosa,
M., and Solis, J. 2012 Plasma dynamics and structural
modications induced by femtosecond laser pulses in
quartz Appl. Surf. Sci. 258(23), 9389.
Lebugle, M., Sanner, N., Varkentina, N., Sentis, M., and
Utéza, O. 2014 Dynamics of femtosecond laser
absorption of fused silica in the ablation regime J.
Appl. Phys. 116, 063105.
Varkentina, N., Sanner, N., Lebugle, M., Sentis, M., and
Utéza, O. 2013 Absorption of a single 500fs laser
pulse at the surface of fused silica: Energy balance and
ablation efficiency J. Appl. Phys. 114, 173105.
Lebugle, M., Utéza, O., Sentis, M., and Sanner, N. 2015
High temporal resolution and calibration in pump–
probe experiments characterizing femtosecond laser–
dielectrics interaction, Appl. Phys. A 120, 455–461.
Utéza, O., Sanner, N., Chimier, B., Brocas, A.,
Varkentina, N., Sentis, M., Lassonde, P., Légaré, F.
and Kieffer. 2011 Control of material removal of fused
silica with single pulses of few optical cycles to sub-
picosecond duration Appl. Phys. A 105, 131.
Optimizing Energy Absorption for Ultrashort Pulse Laser Ablation of Fused Silica
75