Applied Surface Science 252 (2006) 2131–2138

www.elsevier.com/locate/apsusc

Laser fluence, repetition rate and pulse duration

effects on paint ablation
François Brygo a,*, Ch. Dutouquet a, F. Le Guern a, R. Oltra b,
A. Semerok a, J.M. Weulersse a
a
CEA Saclay, DEN/DPC/SCP/LILM, Bat 467, 91191 Gif-sur-Yvette, France
b
LRRS-UMR 5613 CNRS, Université de Bourgogne, Dijon, France
Received 17 January 2005; received in revised form 8 February 2005; accepted 8 February 2005
Available online 17 May 2005

Abstract

The efficiency (mm3/(J pulse)) of laser ablation of paint was investigated with nanosecond pulsed Nd:YAG lasers
(l = 532 nm) as a function of the following laser beam parameters: pulse repetition rate (1–10,000 Hz), laser fluence (0.1–
5 J/cm2) and pulse duration (5 ns and 100 ns). In our study, the best ablation efficiency (h ffi 0.3 mm3/J) was obtained with the
highest repetition rate (10 kHz) at the fluence F = 1.5 J/cm2. This ablation efficiency can be associated with heat accumulation at
high repetition rate, which leads to the ablation threshold decrease. Despite the low thermal diffusivity and the low optical
absorption of the paint (thermal confinement regime), the ablation threshold fluence was found to depend on the pulse duration.
At high laser fluence, the ablation efficiency was lower for 5 ns pulse duration than for the one of 100 ns. This difference in
efficiency is probably due to a high absorption of the laser beam by the ejected matter or the plasma at high laser intensity.
Accumulation of particles at high repetition rate laser ablation and surface shielding was studied by high speed imaging.
# 2005 Elsevier B.V. All rights reserved.

PACS: 81.65 Cf; 42.62 Cf; 61.82 Ms

Keywords: Laser ablation; Repetition rate; Paint stripping; Pulse duration; Thermal confinement regime

1. Introduction surfaces [1–9], tissue ablation [10–12] or paint

For the last few years, material processing with
nanosecond pulsed laser for the etching of polymer
* Corresponding author. Tel.: +33 169087742;
fax: +33 169088691.
E-mail address: Brygo@carnac.cea.fr (F. Brygo).
removal [13–20] has been widely studied. The
development of new high repetition rate and high
power lasers allows to use lasers for large surface
cleaning and large area depainting. For industrial
depainting, laser application seems advantageous
compared with chemical or mechanical processes.
Laser application may offer a dry process with a lower
waste volume and may be automated. Experiments on

0169-4332/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.apsusc.2005.02.143
2132 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138
laser ablation depainting have been made with
nanosecond Nd:YAG laser [13–15], CO2 laser [15–
18], excimer laser [19] and high power diode lasers
[20]. CO2 and excimer laser are commonly used when
it is necessary to prevent a metal substrate from
damages: with CO2 laser (10.6 mm), the ablation
threshold of a metallic substrate is higher than the one
of paint. Excimer laser (UV wavelength leading to
photochemical ablation) allows to decrease signifi-
cantly the thermal effects. High repetition rate
Nd:YAG lasers are very useful for the beam transport
by an optical fibre on a long distance. This is important
for laser decontamination as it allows to move the laser
away from the contaminated surface and to avoid its
pollution during ablation.
Laser–matter interaction is known to depend on
several laser parameters: wavelength, fluence, pulse
duration and repetition rate. Paint ablation with a
532 nm wavelength results mainly in thermal
ablation, because no apparent electronic transitions
inducing direct bond breaking are reported at this
wavelength. Studies on laser ablation efficiency are
frequently presented with the ‘‘etching curves’’.
These curves give the dependence of the removed
material depth (crater depth) on the laser fluence.
They make possible to define the threshold fluence
for ablation, i.e. the fluence from which the onset of
the ablation is noticeable or becomes efficient.
Indeed, with a thermally activated process, no real
threshold exists [2]. However, significant ablation is
observed only above a certain fluence, defined as the
threshold fluence that corresponds to a critical
temperature reached in the surface layer of the
material [21]. Etching curves usually show a
saturation behaviour or a negative inflexion at high
fluence due to plasma, vapor or released fragments
shielding of the surface.
For materials with a low thermal conductivity and
with a low optical absorption (similarly to the paints
under our study), the heating regime is presented
as a ‘‘thermal confinement regime’’ [10–12,21–24]
because the heat diffusion can be neglected during
the pulse absorption. The condition for thermal
confinement
p ffiffiffiffiffiffiffiffi regime can be expressed as LT ¼
Dtp  1=a, where LT is the thermal diffusion
length, D the thermal diffusivity of the irradiated
material, a the optical absorption and tp is the pulse
duration. Thus, the initial temperature distribution is
directly correlated with the deposit of the laser beam
energy in the bulk. Therefore, no influence of the
pulse duration on the ablation threshold fluence in
nanosecond regime is expected in this case.
The effect of the repetition rate on the ablation
efficiency is almost not presented in the literature.
Studies on laser ablation of polymer in UV spectral
range are reported by Burns and Cain [8] and Yung
et al. [9]. In these studies, a decrease of the threshold
fluence between 10 Hz and 300 Hz repetition rate is
observed and the repetition rate effect on the ablation
products is described. Roberts [16] presented results
on paint removal with a TEA CO2 laser, and
measured higher efficiencies with 1500 Hz than
with 1 Hz.
In our studies, laser removal of grey epoxy paint
has been investigated with two pulsed repetition rate
Nd:YAG laser systems at 532 nm. Epoxy coatings are
widely used in the civil industry (e.g. aeronautic or
automobile coatings) and in the nuclear one.
Specifically designed experimental set-ups were
provided to obtain accurate measurements of the
laser parameters and crater depths. Special efforts
were made to obtain a homogeneous (flat top) laser
beam distribution on the interaction zone. The
repetition rate and the pulse duration effects on the
ablation threshold and on the ablated depth are
presented in 0.1–5 J/cm2 laser fluence range. Surface
shielding by accumulation of particles in front of the
surface was investigated with a high speed imaging
system. The air jet effect on the high repetition rate
laser ablation is studied.
2. Experimental set-ups
Two specially developed laser benches (5 ns and
100 ns pulse duration) were provided for our studies
on paint ablation with homogenised laser beam (flat
top spatial profile).
2.1. Five nanoseconds pulse duration laser bench
Paint ablation was carried out with a Q-switched
Nd:YAG laser (Brilliant, Quantel) with 5 ns pulse
duration (FWHM) at 1 Hz, 10 Hz or 20 Hz. With a
KDP non-linear crystal at 532 nm, the maximum
energy per pulse was 180 mJ. The beam was expanded
 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138
Fig. 1. Laser beam profile. (a) Square beam from the low repetition rate, 5 ns pulse duration (700 mm). (b) Circular beam homogenised by the
optical fibre for the high repetition rate laser bench, 100 ns pulse duration (1.3 mm).
with a f = 30 mm lens and sent on a square aperture
of 7 mm side. The aperture transmission of laser beam
was 30%. The aperture was imaged with two lenses
giving 0.1 or 0.18 magnification, and providing a
homogenised beam of 700 mm or 1.3 mm dimension
on the surface. The beam intensity profile (Fig. 1a)
was controlled with a camera (TM-6AS Pulnix) and
the beam profile acquisition system Spiricon. The
energy can be recorded on line with a joulemeter
(SCIENTECH) that was calibrated with a calorimeter
(COHERENT LM45). Laser pulse duration was
measured with a fast photodiode (EG&G Electro-
optics FND-100, rise time < 1 ns). The energy was
adjusted with a half-wave plate (in the 0–50 mJ energy
range) and a cube polarizing beamsplitters to maintain
the same pulse duration during the experiments. The
pulse per pulse energy fluctuations were less than
10%.
2.2. One hundred nanoseconds pulse duration
laser bench
A Q-switched diode pumped Nd:YAG laser
(THALES-CEA) was applied for high repetition rate
experiments. The second harmonic (532 nm) was
obtained with an intra-cavity frequency doubling in a
KTP non-linear crystal. The laser provided 100 W
2133
mean power at 10 kHz and 532 nm with 100  10 ns
pulse duration (FWHM), depending on the laser beam
power. The repetition rate was adjustable in 20 Hz–
10 kHz range. A 20 m long optical fibre of 1 mm core
diameter and the numerical aperture N.A. = 0.2 was
used to transport the beam in a three-lenses focussing
system to ensure 0.4 to 3 magnification. Laser
beam homogenisation is seen as one of the advantages
of optical fibre application. Multiple reflections of the
beam in the multimode optical fibre homogenises it at
the fibre exit (Fig. 1b). In this set-up, the methods and
devices for the beam characterisation were the same as
in the low repetition rate laser bench. The beam energy
was measured with a calorimeter (MOLECTRON
PM-150). A specific electronic device controlled the
acousto-optics switcher of the laser and allowed to
select an adjustable number of laser pulses. Typical
laser pulse per pulse energy stability was better than
10%. An air jet can be applied near the interaction
zone.
Time and spatial evolution of plasma or ejected
matter were recorded with a high speed imaging
system. The intensified camera is a Hamamatsu ICCD,
sensitive in a visible spectral range. The camera is on
side view of the laser beam, with a microscope
objective of 3.5 magnification (Fig. 2). The
2134 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138
Fig. 2. High speed imaging system with an ICCD camera (Hama-
matsu) and a microscope objective.
appropriate integrating time, delay and pulse number
were adjusted by a electronic device.
2.3. Paint sample and crater measurements
The studies were made on an epoxy grey paint
(thickness > 500 mm) on concrete. The crater depths
and spatial profiles were measured with a profilometer
(MAHR) with a mechanical sensor (MFW-250). The
spatial resolution of the measurements was better than
1 mm. The crater spatial profiles (Fig. 3) were directly
correlated with the beam profile (flat top profile with
the same diameter that the laser beam). Thus, the
Fig. 3. Example of a crater spatial profile made with 5 ns pulse duration at F = 5 J/cm2. Laser beam size was 700 mm.
crater depth in our studies is an average value of the
depths measured in different zones of the craters. The
ablation efficiency h (ablated volume per one laser
pulse energy and per number of pulses) is equal to the
crater depth divided by laser fluence and by number of
pulses.
3. Results and discussion
The ablation experiments were carried out in air at
normal atmospheric conditions. From 10 pulses to 300
pulses were applied for a crater formation. The crater
depth is a linear function of the number of pulses in
1.2–5 J/cm2 laser fluence range for 100 ns pulses
duration and in the whole of the applied fluence range
for 5 ns pulses duration. That is why the presented
crater depths are normalized to one laser pulse. This
relationship is not valid for a low fluence regime
(F < 1.2 J/cm2) at high repetition rate ablation. The
pulse number effect in low fluence regime is still under
investigation.
3.1. The repetition rate effect
The repetition rate effect on the crater depth is
presented for 20 Hz–10 kHz repetition rate range
(Fig. 4). The fluence of each laser pulse is constant
(F ffi 0.7 J/cm2). One hundred pulses are applied for
each crater formation. An air jet is applied on the paint
surface to avoid the screening of the laser beam by
accumulation of ejected particles above the surface.
For 20–100 Hz repetition rates, no significant change
in the crater depths was noticeable. For 100–1000 Hz
 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138
Fig. 4. Depth per pulse as a function of the repetition rate for 100
pulses, at F = 0.7 J/cm2 and 100 ns pulse duration. Depth per pulse
increases as a function of the repetition rate due to heat accumulation
in the material pulse per pulse in a ‘‘thermal confinement regime’’. A
5 bar air jet is applied on the surface to prevent from the screening of
the surface by the ejected matter.
repetition rates, however, an increase of the crater
depth with the repetition rate was clearly observed
(Fig. 4). For the pulse repetition rate higher than
1 kHz, the crater depths remained constant.
The repetition rate effect might be due to the low
thermal diffusivity of the paint (D  0.0025 cm2 s1)
[16]. Our study suggests that at high repetition rate, the
time between two pulses is not long enough to cool the
material down to its initial temperature. Therefore,
heating by the next pulse starts at a temperature higher
than the ambient temperature.
3.2. The fluence effect
The hypothesis of heat accumulation with the
number of pulse at high repetition rate suggests that
the ablation threshold fluence should be lower at
high repetition rate than at low repetition rate. This
conclusion is confirmed by the etching curve (Fig. 5) for
different laser ablation regimes: (1) 5 ns and 20 Hz; (2)
100 ns and 20 Hz; (3) 100 ns and 10 kHz.
The etching curves obtained with 100 ns pulses
duration for the low repetition rate and for the high
repetition rate ablation are similar but strongly shifted
to lower fluence for 10 kHz experiments. This fact is
clearly associated to heat accumulation. At low
fluence (<1.2 J/cm2), successive pulse heating of
2135
Fig. 5. Depth per pulse as a function of the fluence for three sets of
experimental condition: (1) 5 ns pulse duration, 20 Hz; (2) 100 ns
pulse duration, 20 Hz; (3) 100 ns pulse duration, 10 kHz with 50
pulses. High repetition rate experiments were made with an air jet
(5 bar).
the material allows to reach a temperature high enough
to vaporise the material, whereas the heating by one
pulse at the same fluence is not sufficient to ablate the
material. With the previous discussion on the thresh-
old fluence, defined as the onset of a significant
ablation, the ablation threshold fluence is measured
as the linear extrapolation of the crater depth d(F)
to d = 0. The threshold fluence at 20 Hz (100 ns) is
F ffi 1.2 J/cm2. At high repetition rate with the
following experimental parameters (50 pulses,
10 kHz), the threshold fluence is found to be: F th 
0.2 J/cm2, that means six times lower than at low
repetition rate.
Thus, two considerations can be emphasized for
paint ablation with a high repetition rate laser: (1) high
repetition rate increases the ablation efficiency and (2)
removal of material like paint can be done at very low
fluence, much lower than the ablation threshold with
low repetition rate laser.
3.3. The pulse duration effect
The third set of experimental conditions (5 ns
pulses duration at 20 Hz) presented a different
behaviour than for 100 ns pulse duration ablation.
Particular attention should be given to the obtained
threshold fluence: F ffi 0.6 J/cm2 that is two times
2136 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138

lower than the 100 ns pulse duration ablation thresh-
old at the same repetition rate.
Typical optical penetration depths 1/a in epoxy
paint are ffi1–10 mm (a is the paint absorption
coefficient for laser wavelength), that is much higher
pffiffiffiffiffiffiffiffi
than the thermal diffusion length LT ¼ Dtp for
pulse durations tp = 5 ns or 100 ns. The heat does not
diffuse in the material during the laser beam
absorption, and the temperature profile should follow
the laser beam intensity in-depth distribution. Thus, in
this hypothesis of ‘‘thermal confinement regime’’, the
pulse duration should not affect the ablation threshold,
as it has been suggested by several authors [25,26].
Experimental studies of Dreisewerd et al. [25] on the
effect of pulse duration (0.55 ns and 3 ns pulse
duration at l = 337 nm) in the desorption/ionization
process showed that the rate of energy deposition does
not play a major role. Independence of the threshold
fluence with the pulse duration is also reported by
Zhigilei et al. [26].
Two possible explanations can be proposed for this
difference in the obtained ablation thresholds. It could
be attributed to some modifications in the paint optical
properties with the laser intensity (for example, in the
paint absorption). Higher absorption leads to lower
threshold fluence. A stress confinement regime [26],
due to the heterogeneity of the material, could also be
considered. But these explanations have not been
proved and should be carefully studied.
Another difference between a long pulse and short
pulse ablation was observed at high laser fluences.
Despite higher threshold fluence for 100 ns pulse
duration, higher ablation efficiencies were obtained
for this pulse duration than for 5 ns pulse duration
experiments. This difference could be explained by
important plasma or ejected matter absorption during
laser–matter interaction which can vary with the pulse
duration [10].
The etching curve (Fig. 5) allows to determine
the best ablation efficiencies for paint removal for

Fig. 6. Ejected matter recording. Integration time, 1 ms; laser fluence, F = 5 J/cm2; number of pulses, 50. (a) Diffusion of the laser beam
(l = 532 nm) by accumulated particles in front of the surface. By applying an air jet, 0.5 bar in (b), 1 bar in (c), 5 bars in (d), the signal decreases.
Accumulation of particles results in a shielding effect of the surface and in the ablation efficiency decrease.
 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138
the different experimental parameters and conditions.
The best efficiencies were found to be: (1)
h ffi 0.3  0.03 mm3/(J pulse) at laser fluence F ffi
1.5 J/cm2 for 10 kHz and 100 ns pulse duration; (2)
h ffi 0.22  0.03 mm3/(J pulse) at F ffi 2.2 J/cm2 for
20 Hz and 100 ns pulses; (3) h ffi 0.16  0.03 mm3/
(J pulse) at F ffi 1.3 J/cm2 for 20 Hz and 5 ns pulses.
3.4. The air jet effect
High repetition rate laser ablation can be affected
by the accumulation of ejected particles in front of the
surface. An ICCD camera was used to record the
matter ejection during paint ablation (Fig. 6). The
gated camera integrates the signal during 1 ms, from
the beginning of the 50th laser pulse. Fig. 6 gives a
side view of the laser beam interaction zone (surface
and near surface zones) for high repetition rate laser
ablation. The sample surface is on the left in Fig. 6.
The laser beam diffusion by the surface of the paint
is noticeable on the left of the pictures. The detected
light from the near surface zone in Fig. 6a and b is
mainly due to the laser beam diffusion by ejected
particles created during the previous pulses.
One can presume that ejected particles created
during high repetition rate ablation do not have
enough time to be extracted from the near surface
zone, thus reducing the ablation efficiency by
absorbing and diffusing the laser beam. This results
Fig. 7. Air jet effect on the depth of the craters for three repetition
rates: 100 Hz, 1000 Hz and 10,000 Hz at 1.5 J/cm2, 50 pulses.
Screening effect due to particles accumulation is only noticeable
with very high repetition rate (10 kHz).
2137
in the ablation efficiency decrease. An air jet was
applied at a few centimetres (3 cm) from the surface:
0.5 bar (Fig. 6b), 1 bar (Fig. 6c) and 5 bar (Fig. 6d).
The particles are swept off, thus decreasing the
diffusion signal.
The ablation efficiency (crater depth evolution) can
be measured as a function of the air jet pressure
(Fig. 7). This figure demonstrates that this screening
effect is significant only for a very high repetition rate
ablation. For 1 kHz and 100 Hz repetition rate
ablation, no noticeable air jet effect on the crater
depths was observed.
4. Conclusions
Investigations on the effects of repetition rate, pulse
duration and laser fluence on paint ablation with
nanosecond lasers are presented. High repetition rate
ablation allows to increase the ablation efficiency by
heat accumulation with each pulse in a ‘‘thermal
confinement regime’’. In our experiments, this
resulted in the significant (six times) decrease of the
ablation threshold fluence. At high repetition rate
ablation, accumulation of particles in front of the
surface can decrease the ablation efficiency due to
absorption/diffusion of the laser beam. This absorp-
tion/diffusion can be avoided by applying an appro-
priate air jet for micro-particles sweeping. The best
paint ablation efficiency h ffi 0.3 mm3/(J pulse) was
found for 10 kHz at 1.5 J/cm2 for 100 ns pulses
duration. Investigation on the pulse duration effect
showed that, despite the low thermal diffusivity of the
paint, a difference in the ablation threshold fluence and
on the efficiency of the ablation is noticeable between
100 ns and 5 ns pulse durations. It was demonstrated
that longer pulses result in deeper craters at high
fluence and higher threshold fluence. Additional
experimental and theoretical studies should be made
to obtain an adequate explanation of the obtained
results.
Acknowledgements
This research is supported by AREVA. The authors
would like to express their gratitude to Mr. G.
Decobert and Mr. H. Masson for financial support and
2138 F. Brygo et al. / Applied Surface Science 252 (2006) 2131–2138
fruitful discussions. We would like to thank Mr. C.
Hubert and Mr. C. Lascoutouna for technical
assistance in the laser benches development, Mr. D.
Farcage for providing the high repetition rate laser and
Mr. G. Brunel for the electronic devices development.
References
[1] T. Lippert, Polymers and Light, Springer, Berlin, 2004, Adv.
Polym. Sci.
[2] D. Bäuerle. Laser Processing and Chemistry, 3rd ed., Springer,
Berlin, 2004.
[3] N. Bityurin, B.S. Luk-yanchuk, M.H. Hong, T.C. Chong,
Chem. Rev. 103 (2003) 519–552.
[4] S.R. Cain, F.C. Burns, C.E. Otis, B. Braren, J. Appl. Phys. 72
(11) (1992) 5172.
[5] S.R. Cain, F.C. Burns, J. Appl. Phys. 71 (9) (1992) 4107.
[6] J.C. Miller, R.F. Haglund, Experimental Methods in the Phy-
sical Sciences, vol. 30, Academic Press, 1998.
[7] R. Srinivasan, B. Braren, Chem. Rev. 89 (1989) 1303.
[8] F.C. Burns, S.R. Cain, J. Phys. D 29 (1996) 1349.
[9] K.C. Yung, D.W. Zeng, T.M. Yue, Surf. Coat. Technol. 16
(2002) 1.
[10] B. Majaron, D. Sustercic, M. Lukac, N. Funduk, Appl. Phys. B
66 (1998) 479.
[11] E. Papagiakoumou, D.N. Papadopoulos, M.G. Khabbaz, M.I.
Makropoulou, A.A. Serafetinides, Appl. Surf. Sci. 233 (2004)
234.
[12] C. Appel, R. Franzen, J. Meister, H. Sarrafzadegan, S. Thelen,
N. Gutknecht, Laser Med. Sci. 17 (2002) 253.
[13] A. Costela, I. Garcı́a-Moreno, C. Gómez, O. Caballero, R.
Sastre, Appl. Surf. Sci. 207 (2003) 86.
[14] D.E. Roberts, Appl. Phys. A 79 (2004) 1067.
[15] G. Daurelio, G. Chita, M. Cinquepalmi, Appl. Phys. A 69
(1999) S543.
[16] D. Roberts, Conference on Lasers and Electro Optics Europe,
2000.
[17] A. Tsunemi, K. Hagiwara, N. Saito, K. Nagasaka, Y. Miya-
moto, O. Suto, H. Tashiro, Appl. Phys. A 63 (1996) 435.
[18] M.J.J. Schmidt, L. Li, J.T. Spencer, J. Mater. Process. Technol.
114 (2001) 139.
[19] L.M. Galantucci, A. Gravina, G. Chita, M. Cinquepalmi,
Polym. Polym. Composites 5 (2) (1997) 87.
[20] M.J. Schmidt, L. Li, J.T. Spencer, Opt. Laser Technol. 31 (2)
(1999) 171.
[21] L.V. Zhigilei, B.J. Garrison, J. Appl. Phys. 88 (3) (2000)
1281.
[22] Y.G. Yingling, P.F. Conforti, B.J. Garrison, Appl. Phys. A 79
(2004) 757.
[23] L.V. Zhigilei, B.J. Garrison, Appl. Phys. Lett. 74 (9) (1999)
1341.
[24] K. Dreisewerd, Chem. Rev. 103 (2003) 395.
[25] K. Dreisewerd, M. Schürenberg, M. Karas, F. Hillenkamp, Int.
J. Mass Spectrom. Ion Process. 154 (1996) 171.
[26] L.V. Zhigilei, E. Leveugle, B.J. Garrison, Y.G. Yingling, M.I.
Zeifman, Chem. Rev. 103 (2003) 321.