Femtosecond laser processing of crystalline
silicon
D. V. Tran (a), Y. C. Lam(a),(c), H. Y. Zheng(b), V. M. Murukeshan(c), J. C. Chai(c), D. E. Hardt(d)
(a)
Singapore-MIT Alliance, Nanyang Technological University, Singapore 639798
Singapore Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075
(c)
School of Mechanical and Production Engineering, Nanyang Technological University, Singapore 639798
(d)
Singapore-MIT Alliance & Department of Mechanical Engineering, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
(b)
Abstract— This paper reports the surface morphologies
and ablation of crystalline silicon wafers irradiated by infrared 775 nm Ti:sapphire femtosecond laser. The effects of
energy fluences (below and above single-pulse modification)
with different number of pulses were studied. New
morphological features such as pits, cracks formation, LaserInduced Periodic Surface Structures (LIPSS) and ablation
were observed. The investigation indicated that there are two
distinct mechanisms under femtosecond laser irradiation: low
fluence regime with different morphological features and high
fluence regime with high material removal and without
complex morphological features.
Index Terms— crystalline silicon, femtosecond laser,
morphology, ablation, incubation effect.
S
I. INTRODUCTION
ILICON is an important material in semiconductor
industry and useful for MicroElectroMechanical
Systems (MEMS ) devices. Laser processing of silicon (Si)
has received significant attention in the last several
decades. In the past, one of the interests in laser processing
of Si is the “laser annealing” process in which lattice
damage by ion implantation of dopants such as Boron (B)
or Phosphorus (P) into the crystalline Si could be removed
and electrically activated [1].
Manuscript received November 19, 2004.
D. V. Tran is a Ph.D. student with the Singapore-MIT Alliance (SMA)
under the Innovations in Manufacturing Systems and Technology (IMST)
program hosted at Nanyang Technological University, Singapore. (email:
PBG0287913@ntu.edu.sg).
Y. C. Lam is with Singapore-MIT Alliance (SMA) and Professor of
School of Mechanical and Production Engineering, Nanyang
Technological University, Singapore. (email: myclam@ntu.edu.sg).
H. Y. Zheng is with Singapore Institute of Manufacturing and
Technology (SIMTech), 71 Nanyang Drive, Singapore. (email:
hyzheng@SIMTech.a-star.edu.sg)
V. M. Murukeshan and J. C. Chai are Assistant Professor and
Associate Professor with School of Mechanical and Production
Engineering, Nanyang Technological University, Singapore. (email:
Mmurukeshan@ntu.edu.sg and MCKChai@ntu.edu.sg ).
D. E. Hardt is with Singapore-MIT Alliance (SMA) and Professor of
Department of Mechanical Engineering, Massachusetts Institute of
Technology (MIT), Cambridge, Massachusetts. (email: hardt@mit.edu)
Laser annealing of Si has been demonstrated to offer
advantages over furnace annealing such as exceeding the
solid solubility limit of dopants in Si and having less
defects after annealing, leading to the improvement in the
electrical properties and hence the function of the junction
subjected to ion implantation. Different laser sources with
wavelengths from far infrared to ultraviolet (UV), for
example, continuous wave (CW) CO2 laser (λ = 10.6 µm)
to pulsed lasers such as ruby (λ = 0.694 µm), Nd: YAG (λ
= 0.1064 µm), and excimer (λ = 0.249 µm) lasers have
been used [1]. The fundamental frequency, for example,
Nd:YAG laser (λ = 0.1064 µm), could be doubled, tripled
and quadrupled to 0.532 µm, 0.353 µm and 0.265 µm
respectively to enhance absorption on crystalline Si
(indirect band gap ≈ 1.13 µm).
One of the important parameters in laser processing is the
laser pulse width (or pulse duration). The first laser
invented in 1960 by Maiman [2], based on a ruby crystal
and pumped by a xenon flash discharge, created a laser
pulse lasting between millisecond to microsecond. New
laser generation techniques such as Q-switching and modelocking has been developed since the first ruby laser
reduced the pulse width to nanosecond (1 ns = 10-9 s),
picosecond (1 ps = 10-12 s), femtosecond (1 fs = 10-15 s)
and recently attosecond (1 as = 10-18s) [3] duration. Among
these pulse durations, fs laser has been attracted many
interests in the last decade. By shortening the pulse width
to fs duration (sometime called ultrashort or ultrafast
pulses) it follows the studies of ultrashort phenomenon in
matters and laser-induced phase transformations. The
former related to the capability to use fs laser to observe
phenomenon with fs time resolution, for example, the
energy transfer from electrons to the lattice in solids [4].
New subfields of science have been created, namely
femtochemistry [5] and femtobiology [6] in which the
phenomena could be resolved on ultrashort time-scale
resolution. The focus of this paper is the capability of an
ultrashort laser pulse to induce a novel phase transitions on
matters, for example Si.
In the late seventies, one of the studies created many
interests is on the phase transformation processes by laser
on Si in the laser annealing process as mentioned above.
Most of the studies have been carried out on ns laser
pulses. This ns laser pulses provide fluences (energy
densities) high enough to melt a surface layer of
amorphous Si created by ion implantation of dopants on
single crystalline silicon. Resolidification of the molten
silicon removed the lattice damage and electrically activate
the dopants with free defects. Nanosecond laser pulses
could also be used to recrystalline fine-grained
polycrystalline Si films deposited on insulating substrate,
for example, SOS (silicon-on-sapphire) with practical
applications in silicon technology [1]. For ps laser pulses,
when the laser irradiates on the materials, heating and
cooling rates up to 1014 °C/s could be obtained [7]. With
cooling rate at these magnitudes, the kinetics of phase
transitions may lead to new phenomena. Amorphization of
single crystalline silicon or recrystallization of this
amorphous layer could be carried out by controlling the
energy density of the Gaussian laser pulse [7]. In a recent
experiment, near-infrared fs laser-induced crystallization of
amorphous silicon on insulator have been performed [8]. It
is shown that large grains of polycrystalline silicon on
oxidized silicon wafer deposited by low-pressure chemical
vapor deposition (LPCVD) could be achieved when fs
laser is scanned over the sample. Unlike ns laser annealing
(linear annealing), fs laser annealing (FLA) was a
nonlinear photoenergy absorption process where nonequilibrium between the electrons and the lattice dominates
[9]. Amorphization of silicon by fs laser was also observed
when studying ablation of crystalline silicon [10]. A layer
of amorphous silicon around the ablated holes was found
by using
cross-sectional Transmission Electron
Microscopy (TEM). This study was in contrast with the
postulation that direct solid-vapor transition [11] and
negligible hydrodynamic motion [9] were assumed on
femtosecond laser ablation of materials.
One of the novel phase transformations in using fs laser is
the ablation process. In the last decades, many interests
focus on the fs ablation with potential applications in
industry (high precision microfabrication) and in medicine
[12]. With the advent of the Chirped Pulse Amplification
(CPA) technique [13] to increase the average power for
micromachining applications, a wide set of parameters,
especially the pulse duration, could be varied while other
parameters, for example, the focusing condition, are kept
unchanged. With this capability, ablation materials with fs
laser has been studied and compared to that of ns ablation
in different materials (metals [11], semiconductors [14],
dielectrics [15], and polymers [16]). Advantages of fs laser
ablation over ns ablation have been demonstrated for high
precision in micromachining applications. Some of the
salient features of ablation with fs are (i) reduction of HeatAffected Zone (HAZ) and defects (burr, micro cracks), (ii)
deterministic ablation threshold, (iii) sub-micrometer
processing and (iv) ability to machine transparent
materials.
In this paper, we report the morphologies and ablation on
crystalline silicon irradiated by fs laser. The experiment is
studied with both low (below single-pulse modification
threshold) and high (above single-pulse modification
threshold). The effect of fs laser fluences with different
number of laser pulses on Si will be presented. New
features on morphologies of crystalline silicon are
discussed.
II. EXPERIMENTAL
The experimental setup consisted of an amplified kHz
Ti:sapphire femtosecond laser system (CPA-2001, ClarkMXR, Inc.) illustrated in Fig. 1. The laser emits pulses of
150 fs at Full Width Half Maximum (FWHM) with
linearly-polarized light at a central wavelength of
approximately 775 nm in the near infrared. The laser has
Transverse Electromagnetic Mode TEM00 and nominal
repetition rate of 1 kHz. The total pulse energy Epulse was
attenuated by a rotating half waveplate followed by a linear
polarizer (attenuator in Fig. 1). The repetition rate of the
laser was lowered to and kept at 250 Hz and synchronized
with a mechanical shutter, which allowed control on the
number of laser pulses on the target. The laser beam was
made circularly-polarized by using a quarter waveplate in
the beam path before the focusing lens (focal length f = 50
mm). The beam was directed perpendicular to the sample
surface, forming an almost circular spot on it. The radius of
the spot sizes ω 0 (defined at 1/e2 of peak intensity) was
measured by the knife-edge technique [17]. Two silicon
wafers, thickness of 100 µm and 460 µm, were subjected
to different pulse energy fluences. The fluences were
varied by changing the total pulse energy and the focusing
spot size. At the given laser fluence, the same experimental
condition was kept constant except for the number of laser
pulses. Each data point was repeated 5 times with a fixed
number of pulses. A computer-controlled xyz-stage
allowed positioning of the wafer at the desired location.
The experimental results were reproducible. Using an
optical microscope and a Scanning Electron Microscope
(SEM), the laser-induced surface damage morphologies
and ablation of the crystalline silicon wafers were
analyzed. We defined surface damage as a sudden change
in surface damage morphology, i.e., modification, not
necessarily ablation, when observed under optical
Fig. 1. Schematic of the experimental setup
microscope or SEM.
III. RESULTS AND DISCUSSION
The surface damage morphologies and ablation of
crystalline silicon depend on the energy fluences and the
number of pulses. As shown later, the surface
morphologies of silicon were different for energy fluences
below and above single-pulse modification threshold. The
effect of number of pulses will be discussed.
The relationship between the total pulse energy, E pulse ,
and fluence is derived as follows. For a Gaussian beam, the
spatial and temporal profile of the laser intensity is [18]
(
) (
I (r , t ) = I 0 exp − 2r 2 / ω 02 exp − 2t 2 / τ 2
)
(1)
where I 0 is the peak intensity at the center, ω 0 and τ are
the spatial radius and temporal radius, respectively, at the
(1/e2) intensity contour, r is the radial coordinate of
distance from the propagation axis and t is the time
variable. The spatial distribution of the energy fluence
E (r ) =
∫
+∞
−∞
(
I (r , t )dt = E 0 exp − 2r 2 / ω 02
)
(2)
where E 0 is the peak fluence at the center of the beam.
The peak fluence relates to the total pulse energy as
∞
E pulse = E (r )2πrdr =
∫
0
=
πω 02 E 0
∫
∞
0
(
)
E 0 exp − 2r 2 / ω 02 2πrdr
(3)
2
or
E0 =
2 E pulse
πω 02
(4)
From equation (4), the peak fluence E 0 of the laser could
be varied by changing the total pulse energy E pulse or the
spot size ω 0 . We will investigate surface morphologies
and ablation of Si from low energy fluence to high fluence
with different number of pulses.
A. Fluence below single-pulse modification threshold
We have studied damage morphologies of crystalline
silicon at energy fluence as low as 0.03 J/cm2 (pulse energy
≈ 0.24 mJ, spot size (1/e2) diameter ≈ 1.4 mm) [19]. SEM
images of silicon at this fluence regime are shown in Fig.
2.
For a single pulse, we did not observe any surface damage
at the processed region on the silicon surface under optical
microscope or SEM. When the sample was subjected to 10
pulses, no trace of surface damage was observed as well,
see Fig. 2(a). The SEM image was identical to that of the
original phase. For 20 laser pulses, surface damage
appeared as shown in Fig. 2(b). Surface damage was not
easily recognizable since the phase change of the treated
region was nearly identical with the original phase, which
was crystalline. However, as can be observed later, the
surface damage at this low number of pulses was important
for surface damage evolution at subsequent laser pulses.
The surface damage in Fig. 2(b) at sub-threshold pulses
shows the incubation effect, i.e. accumulation of energy, of
materials when irradiated with multiple pulses. This effect
was observed in many materials for ns, ps and fs lasers
[20]−[24]. Jhee et al. [20] observed this incubation effect
on silicon irradiated with ps laser pulses. They suggested
that damage precursors by multiple pulses were due to the
long-lived excitations or accumulation of permanent states.
The incubation period of these states could last as long as 3
seconds and independent of the repetition rate. They
showed that there existed a multiple pulse damage
threshold or fatigue limit below which no damage was
supposed to occur even at an infinite number of laser
pulses. They calculated this fatigue limit for silicon to be
about 0.32 J/cm2.
One of the characteristics differentiating fs to longer laser
pulses (including ps laser pulses) is that fs laser has lower
fluence damage threshold. It was demonstrated that fluence
breakdown threshold Fth did not follow the scaling of
Fth ∝ τ (where τ is the pulse duration) when the pulse
duration was shorter than 10 ps [25]. For long laser pulses,
heat diffusion could involve a much larger volume than the
laser focus. As the pulse width decreases, the heat-affected
volume due to conduction was smaller. As a result, a lower
fluence was sufficient to create surface damage or ablation
on materials with fs lasers.
The accumulation of sub-threshold laser fluence when
employing fs laser resulting in damage, i.e. incubation
effect, were also observed on dielectrics [22] and tissue
[23]. The damage on dielectrics could be explained by Fcenter formation created during excitation by first or
subsequent laser pulses. The optical break down in
dielectrics occurred when local stress created by these Fcenter formations was built up [26] and created damage at
a certain number of laser pulses. The surface damage on
tissue, on the other hand, could be explained by cumulative
thermal effect [23] when the material was irradiated with
high repetition rate laser pulses.
As such, although the laser fluence employed here was of
0.03 J/cm2, one order of magnitude lower than the fatigue
limit reported by Jhee et al. [20] for long pulse laser, it was
not surprising to have observed surface damage with
increasing number of laser pulses as shown in Fig. 2(b).
Although our experimental conditions were not exactly the
same as Bonse et al. [24], we expected our results to be
similar to theirs as they also employed fs laser to irradiate
on crystalline silicon. Their single pulse threshold for
surface modification was 0.26 J/cm2. They employed the
same equation in Jee et al. [21] to describe the incubation
effect for multiple pulses. Using the same constants and
equation of Bonse et al., it could be calculated that for 20
(a)
(c)
(b)
cracks
(e)
(f)
redeposit
materials
(g)
(d)
cracks
(h)
Severe
surface
damage
by pits
pits
Larger pits
at cracks
larger pits
at cracks
no pits
formation
Fig. 2. Normal view SEM images of the crystalline silicon subjected to sub-threshold multiple femtosecond laser pulses at peak fluence of
0.03 J/cm2: (a) 10 pulses, no surface damage was observed, (b) 20 pulses (c) 50 pulses, (d) 100 pulses, (e) 500 pulses, (f) 1000 pulses, (g)
2000 pulses and (h) 5000 ( λ = 775 nm, τ = 150 fs, 250 Hz)
laser pulses, the surface modification threshold would be
0.16 J/cm2. Alternatively, it could be obtained that for a
laser fluence of 0.03 J/cm2, the number of pulses required
for surface modification would be approximately 0.7x106.
Thus, with our laser fluence of 0.03 J/cm2, it was expected
that no damage could be observed for a single pulse, or for
ten pulses. However, as compared with Bonse et al. [24],
our
observation of surface damage for 20 pulses was rather
cracks
low either for the value of the fluence, or the number of
pulses. The differences in observations could be because of
the differences in experimental conditions as will be
elaborated later, and we could not rule out the possibility
that the equation employed by Bonse et al. might not hold
for our experimental conditions.
After surface damage occurred (Fig. 2(b)), increasing
number of laser pulses at the same laser fluence created
different surface damage morphologies on silicon, as seen
in Fig. 2(c). After 20 pulses (Fig. 2 (b)), the surface
damage was hardly recognized under SEM. However, at 50
pulses, the surface damage became easily recognizable
under SEM, see Fig. 2(c). The effect of previous pulses
had changed the properties of the materials and hence the
absorption of laser energy. At this pulse number, the
overall image of the surface damage was similar to that of
Fig. 2(b). In addition, some cracks were observed on the
surface.
At higher number of pulses, the surface damage area
became larger, see Fig. 2(d) and Fig. 2(e) after irradiation
with 100 and 500 pulses respectively. The surface damage
areas were larger at higher number of pulses because of the
effect of previous pulses. At 500 pulses, many pits began
to appear (Fig. 2(e)). The formation of pits at this number
of pulses was analogous to previous observation by Jhee et
al. [20] at the beginning of the surface damage of silicon
using ps laser pulses and could be explained by the
evaporation of silicon by the accumulation of laser energy.
As shown later, more pits were created at higher number of
pulses in contrast to that reported by Jhee et al. [20] that
these pits were grown and were the seeds for the formation
of ripples and crater-like structures with increasing number
of pulses. The diameters of these pits were larger at the
crack positions compared to the surrounding regions, as
can be seen in Fig. 2(f) and Fig. 2(g). In Fig. 2(f) and Fig.
2(g), more re-deposited materials appeared and were
located at the boundary between the damaged and
undamaged regions. At 5000 pulses, Fig. 2(h), the surface
damage developed at the pit positions. Interestingly, center
region had not developed any pits, see Fig. 2(h).
In contrast to reports on formation of ripples and columnar
structures when irradiated with multiple pulses [20]−[23]
we did not observe these self-organized structures at this
fluence with multiple pulses (from low number to high
number of pulses). Instead, many pits were created and
formed the seed for the development of surface damage, as
clearly seen in Fig. 2(h). In the fs regime, the different
observation of surface damage morphologies on silicon
from previous reports [20]−[27] might be explained by the
differences in the experimental set up as follows.
Firstly, comparing to the work of Bonse et al. [24], the
laser fluence employed here was lower and the spot size
was larger. As a result, the surface damage morphology
was not observed at single pulse or even at 10 pulses (Fig.
2(a)) while it was not the case in their experiments.
Secondly, the laser was made circularly-polarized instead
of linearly-polarized as in their work. It was expected that
polarization of the laser beam affected the surface damage
similar to what has been observed recently in the laser
drilling process with ultrashort laser pulses [28]. Finally, it
(a)
(b)
20 µm
ripples
pattern
(c)
(d)
20 µm
Fig. 3. Normal view SEM images of crystalline silicon with sub-threshold multiple femtosecond laser pulses at peak laser
fluence of 0.25 J/cm2 after (a) 5 pulses: no self-organized structures was observed and (b) 10 pulses: ripples pattern was
observed, (c) 80 pulses, (d) 150 pulses ( λ = 775 nm, τ = 150 fs, 250 Hz)
was also expected that the surface orientation of the silicon
influenced the surface damage morphology. The surface
orientation employed here was (100) rather than (111) as in
Bonse et al. [24]. Experiments with ns pulses had
demonstrated that [21] the accumulative damage by
multiple pulses depended on the surface orientation.
Differences in laser fluence, polarization, and surface
orientation might explain the differences in observation of
surface damage morphologies. These factors would
contribute to the differences between our observations.
In Fig. 3, we show SEM images at fluence of 0.25 J/cm2
(pulse energy ≈ 0.36 mJ, and spot size (1/e2) ≈ 300 µm in
diameter) [29]. Similar to morphologies in Fig. 2., for a
single laser pulse, no trace of surface damage morphology
could be detected when the processed region was observed
under optical microscope and SEM. However, after
irradiating with 5 pulses, surface damage morphology
appeared; see Fig. 3(a). The phenomenon is similar to the
incubation effect discussed above for fluence of 0.03
J/cm2. However, in this case since the fluence is higher the
surface damage of silicon starts at lower number of pulses
i.e., surface damage starts at about 5 pulses at 0.25 J/cm2
whereas it starts at 20 pulses at 0.03 J/cm2.
When the number of laser pulses was increased at constant
fluence (80 consecutive pulses) different surface damage
morphologies were observed, see Fig. 3(c). At this number
of pulses, the surface damage area became larger compared
to that of low shot numbers as in Fig. 3(a) (note the scale
bar on the SEM images). Obviously, the effect of previous
pulses had changed the surface microstructure and
morphology and hence the absorption. To qualitatively
illustrate the different surface damage morphologies on
Fig. 3(c), Fig. 4 was used to schematically characterize
different damage regions with respect to the energy fluence
distribution and thresholds of the Gaussian beam profile
[29]. As shown later, the patterns of damage morphologies
schematically illustrated in Fig. 4 were repeatable with
increasing number of pulses. The only difference is that
these damages were more distinguishable with higher
pulses.
Different surface damage morphologies in Fig. 4 could be
qualitatively explained (from center to outermost regions
of damages). At the center region where the laser fluence
was highest (denoted as ablation in Fig. 4), ablation was
expected to occur along with Laser-Induced Periodic
Surface Structure (LIPSS) patterns, which were clearly
observed. This was not surprising since the material in this
region was irradiated with the highest energy fluences of
the Gaussian beam. A similar pattern was observed
recently with femtosecond pulses by other researchers [24].
The major difference was that our peak fluence was below
the single-pulse surface damage threshold while the
patterns observed previously were at peak fluence above
the single-pulse damage threshold [24]. The next annulus
region was the re-deposited materials because of the
ablation at the center. This could be similar to the “rim” in
the recent study on indium phosphide (InP) irradiated by
single and two fs pulses. The formation of the rim may
presumably due to either the momentum transfer from the
ablated material (at the center region) to the molten surface
layer or the surface tension gradients [30].
The next two annulus regions (one inside the border of the
“rim” and other outside the border of the rim, denoted as
central modification region in Fig. 4) was attributed the
modification region. These two annulus could be expected
to be in the amorphous states or polycrystalline silicon
since there was a difference in reflectivity when observed
under SEM. As discussed in the introduction, amorphous
Fig. 4. Schematic illustration of different annulus regions of
surface damage morphologies and ablation observed on silicon
by multiple sub-threshold ultrashort laser pulses.
or polycrystalline silicon was observed on crystalline
silicon when irradiated with proper energy fluence with
both long and ultrashort pulses. The formation of
amorphous or polycrystalline silicon on single crystalline
silicon was a kinetic process of resolidification of materials
from melting. The amorphous phase could be formed on
crystalline silicon because of the cooling rate during the
solidification process is fast (about 1014 °C/s [7]) and there
is no time for the material to return to the crystalline phase.
This phase change on crystalline silicon was confined only
within a certain laser fluence range. With the laser fluence
higher than this range (for example at the center of the
Gaussian beam), the cooling rate on solidification was low
enough to allow epitaxially re-crystalline from the
substrate.
In the next annulus region, we expect that there was no
phase change. The phase of this region was expected to be
consistent with the original phase as observed under SEM.
Although this annulus region was repeatedly illuminated
by multiple pulses the accumulation of energy was not high
enough to modify the surface. As schematically illustrated
in Fig. 4, the energy fluence was presumably lower than
the multiple-pulse modification threshold and hence do not
create any phase change. Interestingly, there was a phase
change to the annulus region next to the above no phase
change region (denoted as outer modification region). This
could be explained by the diffraction of the beam (denoted
as diffraction part in Fig. 4). The phase change occurred
because the energy fluence had reached the multiple-pulse
modification threshold and hence with repeated
illumination, a phase change similar to the central
modification was expected.
With increasing number of pulses, the pattern of surface
damage morphologies were similar to those of Fig. 3(c).
However, these surface damage morphologies were more
distinguishable. At the center region, LIPSS pattern could
be clearly seen similar to what had been observed in the
center region of Fig. 3(c). At these pulse numbers,
however, we did not observe columnar structures. Column
formation in crystalline silicon has been observed in the
past with different pulse durations, laser fluence, and the
ambient environments. In a recent work [31] on fs laser
with crystalline silicon, sharp spikes were observed when
the material was irradiated in SF6 or Cl2 environment but
not in vacuum, N2 or He. The chemical reactions were
suggested to be essential for formation of sharp spikes
[31]. In our study, we did not observe columnar structures
probably because the materials was irradiated with much
lower fluence. Comparing to Fig. 3(c), more re-deposited
materials appeared in the next annulus regions. At 150
pulses, Fig. 3(d), even more materials was re-deposited
since there was more ablation at the center. The next two
annulus regions (the no phase change and outer
modification regions) could be similarly explained.
B. Fluence above single-pulse modification threshold
At a higher fluence level, we have observed a different
surface damage and ablation mechanism on crystalline
silicon [32]. Fig. 5 shows SEM images of silicon when
subjected to single-laser pulse with increasing laser.
Comparing the SEM images of Fig. 5 when the silicon was
subjected to single-pulse with increasing fluence level with
Figs. 2 and Fig. 3, we could observe clearly the difference
in surface damage and ablation mechanism. The surface
damage and ablation was driven by more material ejected
out of the center of the illuminated regions, i.e. at the peak
fluence of the Gaussian beam. No features such as ripples
and column formation were observed in this regime. This
ablation mechanism in silicon could be analogous to the
strong ablation phase in dielectrics [33] and the absorption
mechanisms by heat penetration depth in metals [34] as
reported recently.
From Fig. 5, one could observe that the radii or diameters
of the damaged areas (at the center of the laser focus)
increase with increasing laser fluence. If we assume that
the laser pulse had a Gaussian distribution intensity profile,
the development of the damaged features (e.g. ablated hole
areas in Fig. 5) should be consistent with the distribution
given by equation (2).
From equation (2), the radius r or diameter d of the
damaged features relates to the peak fluence E 0 by
d 2 = 4r 2 = 2ω 02 (ln E 0 − ln E (r ))
(5)
Since there is a linear relationship between the peak
fluence and the laser pulse energy (equation (4)), and with
equation (5), one can obtain a linear relationship between
the diameter square of the surface damaged areas versus
the logarithm of the laser pulse energy. The laser focus
radius ω 0 could be calculated from the slope of this plot.
The peak fluence could then be calculated again using
equation (4). From equation (5), one could obtain the
fluence threshold (expressed in peak fluence of the beam)
by extrapolating the linear line to the horizontal axis. This
technique was first introduced by Liu et al. [18] when
studying laser-induced damage solid-liquid transformation
in silicon with ps laser pulses. Recently, the technique was
used successfully even with pulses as short as 5 fs on
dielectrics [35].
Using this technique we calculated the ablation threshold
for single pulses (about 7.8 J/cm2) and the laser radius
focus ω 0 (about 33 µm at 1/e2 intensity) within the
experimental error of the energy fluctuation of about 5%.
Fig. 6 shows the relationship between the peak fluence and
the diameter square of the ablated holes in Fig. 5. The
value of the laser focus ω 0 (obtained from the slope of the
fitted line) was comparable to the value obtained when
measuring by the knife-edge scanning method (about
30 µm ). The ablation threshold in this regime was
significantly different from the low fluence regime. As
mentioned above, the ablation mechanism in this regime
might be characterized by the higher removal of materials
than in the gentle phase ablation regime. This could be
verified by comparing the SEM images from Figs. 2 and 3
a)
b)
c)
d)
e)
f)
Fig. 5. SEM images of development of the radii of 460-µm crystalline silicon with increasing laser fluence for single laser pulse,
τ = 150 fs, (a) 13 J/cm2, (b) 17 J/cm2, (c) 22 J/cm2, (d) 26 J/cm2, (e) 28 J/cm2, and (f) 33 J/cm2.
in the gentle ablation phase and Fig.
ablation phase.
d2 (µm2)
nm,
5 in the strong
ACKNOWLEDGMENT
D. V. Tran would like to acknowledge the financial
sponsorship of the Singapore-MIT Alliance (SMA) in the
form of a research scholarship. The technical support
provided by SIMTech is also gratefully acknowledged.
3500
d 2 = 1900.22 ln E0 − 3907.63
3000
λ = 775
2500
2000
1500
REFERENCES
1000
[1]
500
0
1.5
2
2.5
3
3.5
4
2
ln(E0) (J/cm )
Fig. 6. Relationship between the developments of the diameter the
ablated holes (from Fig. 5.) and the peak fluence E0. Solid line was
obtained from linear regression analysis.
IV. CONCLUSION
In this paper, we report surface damage morphologies
and ablation of crystalline silicon interacted with
femtosecond laser pulses at different energy fluences and
number of pulses. At low energy fluence regime, we
observed new features such as pits and cracks formation
which exists even for a high number of pulses. We also
have observed similar morphologies, for example, LaserInudced Periodic Surface Structure (LIPSS) and columnar
formation. At high fluence regime, we did not observe
these morphologies but high removal of materials which
left a crater consistant with a Gaussian distribution of the
pulse.
R. F. Wood, C. W. White, R. T. Young, Semiconductors and
Semimetals, vol. 23 (Academic, Orlando 1984).
[2] T. H. Maiman, “Stimulated optical radiation in ruby lasers,” Nature
(London)., vol. 187, pp. 493–494, 1960.
[3] P. Agostini P. and F. D. Louis, “The physics of attosecond light
pulses,” Rep. Prog. Phys., vol. 67, pp. 813–855, 2004.
[4] J. Shah, Ultrafast Spectroscopy of Semiconductors and
Nanostructures. Berlin: Springer-Verlag, 1996.
[5] A. H. Zewail, “Femtochemistry,” J. Phys. Chem., vol. 97, pp.
12427–12446, 1993.
[6] R. A. Matthies, R. W. Schoenlein, L. A. Peteanu, Q. Wang, and C.
V. Shank, “Femtosecond reaction,” in Femtosecond Reaction
Dynamics, edited by D. A. Wiersman. North-Holland, Amsterdam,
p. 229.
[7] P. L. Liu, R. Yen, N. Bloembergen, “Picosecond laser-induced
melting and resolidification morphology on Si,” Appl. Phys. Lett.,
vol. 34, pp. 864−66, 1979.
[8] S. Jia-Min, C. Zun-Hao, T. Bau-Tong, W. Yi-Chao, Z. Alexei, and
Pan.
Ci-Ling,
“Near-infrared
femtosecond
laser-induced
crystallization of amorphous silicon ,” Appl. Phys. Lett., vol. 85, pp.
1232−1234, 2004.
[9] X. Liu, D. Du, and G. Mourou, “Laser ablation and micromachining
with ultrashort laser pulses,” IEEE J. Quantum. Electron., vol. 33,
pp. 1701−1716, 1997.
[10] J. Jimmy, L. Ming, and V. T. Carl, “Amorphization of silicon by
femtosecond laser pulses,” Appl. Phys. Lett., vol. 84, pp. 3205−3207,
2004.
[11] B. N. Chichkov, C. Momma, S. Nolte, F. von Alvensleben, and A.
Tünnermann, “Femtosecond, picosecond and nanosecond laser
ablation of solids,” Appl. Phys. A., vol. 63, 109−115, 1996.
[12] J. H. Niemz, Laser-Tissue Interactions. Berlin: Springer-Verlag,
1996.
[13] D. Strickland and G. Mourou, “Compression of amplified chirped
optical pulses,” Opt. Commun., vol. 56, pp. 219−221, 1985.
[14] A. Cavalleri, K. Sokolowski-Tinten, J. Bialkowski, M. Schreiner,
and D. von der Linde, “Femtosecond melting and ablation of
semiconductors studied with time of flight mass spectroscopy,” Jour.
Appl. Phys., vol. 6, 3301−3309, 1999.
[15] B. C. Stuart, M. D. Feit, S. Herman, A. M. Rubenchik, B. W. Shore,
and M. D. Peryy “Nanosecond-to-femtosecond laser-induced
breakdown in dielectrics,” Phys. Rev. B., vol. 53, 1749−1761, 1996.
[16] J. Krüger and W. Kautek, “Ultrashort pulse laser interaction with
dielectrics and polymers,” in Adv. Polym. Sci., vol. 168, 247−289,
2004.
[17] Y. Suzaki and A. Tachibana, “Measurement of the µm sized radius
of Gaussian laser beam using the scanning knife-edge,” Appl. Opt.,
vol. 14, 2809−2810, 1975.
[18] J. M. Liu, Simple technique for measurements of pulsed Gaussianbeam spot sizes, Opt. Lett., vol. 7, pp. 196−198, 1982.
[19] Y. C. Lam, D. V. Tran, H. Y. Zheng, V. M. Murukeshan, J. C. Chai,
and D. E. Hardt, “Surface damage of crystalline silicon by low
fluence femtosecond second laser pulses,” Surf. Rev. Lett., vol. 11,
pp. 217−221, 2004.
[20] Y. K. Jhee, M. F. Becker, and R. M. Walser, “Charge emission and
precursor accumulation in the multiple-pulse damage regime of
silicon,” J. Opt. Soc. Am. B., vol. 2, pp. 1626−1633, 1985.
[21] Y. Jee, M. F. Becker, R. M. Walser, “Laser-induced damage on
single-crystal metal surfaces,” J. Opt. Soc. Am. B., vol. 5, pp.
648−689, 1988.
[22] D. Ashkenasi, M. Lorez, R. Stoian, and A. Rosenfeld, “ Surface
damage threshold and structuring of dielectrics using femtosecond
second laser pulses: the role of incubation,” Appl. Surf. Sci., vol.
150, pp. 101−106, 1999.
[23] B. M. Kim, M. D. Feit, A. M. Rubenchik, E. J. Joslin, J. Eichler, P.
C. Stoller, L. B. Da Silva, “Effects of high repetition rate and beam
size on hard tissue damage due to subpicosecond laser pulses,” Appl.
Phys. Lett., vol. 76, pp. 4001−4003, 2000.
[24] J. Bonse, S. Baudach, J. Kruger, W. Kautek, and M. Lenzner,
“Femtosecond laser ablation of silicon-modification thresholds and
morphology,” Appl. Phys. A., vol. 74, pp. 19−25, 2002.
[25] D. Du, X. Liu, G. Korn, J. Squier, and G. Mourou, “Laser-induced
breakdown by impact ionization in SiO2 with pulse widths from 7 ns
to 150 fs,” Appl. Phys. Lett., vol. 64. pp. 3071−3073, 1994.
[26] S. C. Jones, P. Braunlich, R. T. Casper, X. A. Shen, P. Kelly,
“Recent progress on laser-induced modifications and intrinsic bulk
damage of wide-gap optical materials,” Opt. Eng., vol. 28, pp.
1039−1068, 1989.
[27] G. Herbst, M. Steiner, G. Marowsky, and E. Matthias, “Ablation of
Si and Ge using UV femtosecond laser pulses,” in Mat. Res. Soc.
Sym. Proc. Vol. 397, pp. 69−74, 1996.
[28] S. Nolte, C. Momma, G. Kamlage, A. Ostendorf, C. Fallnich, F. von
Alvensleben, and H. Welling, “Polarization effect on ultrashortpulse laser drilling,” Appl. Phys. A., vol. 68, 563−567, 1999.
[29] D. V. Tran, H. Y. Zheng, Y. C. Lam, V. M. Murukeshan, J. C. Chai,
and D. E. Hardt, “Femtosecond laser-induced damage morphologies
of crystalline silicon by sub-threshold pulses,”, published on Optics
and Lasers in Engineering, 2005 (accepted)
[30] J. Bonse, J. M. Wrobel, K. W. Brzezinka, N. Esser, and W. Kautek,
“Femtosecond laser irradiation of indium phosphide in air: Raman
spectroscopic and atomic force microscopic investigations,” Appl.
Surf. Sci., vol. 202, pp. 272−282, 2002.
[31] T. H. Her, R. J. Finlay, C. Wu, and S. Deliwala, “Microstructuring of
silicon with femtosecond laser pulses,” Appl. Phys. Lett., vol. 73, pp.
1673−1675, 1998.
[32] D. V. Tran, Y. C. Lam, V. M. Murukeshan, J. C. Chai, H. Y. Zheng,
and D. E. Hardt, “Femtosecond Laser-Induced Damage
morphologies and ablation of crystalline silicon: a study on the
incubation effect,” Proceeding of International Connference on
Precision Engineering, ICoPE 2003-2004, pp. 265−272,
[33] D. Ashkenasi, R. Stoian, and A. Rosenfled, “Single and multiple
ultrashort pulse abltion threshold of Al203 (corundum) at different
etch phases,” Appl. Surf. Sci., vol. 154-155, pp. 40−46, 2000.
[34] S. Nolte, C. Momma, H. Jacobs, A. Tünnermann, B. N. Chichkov, B.
Wellegehausen, and H. Welling, “Ablation of metals by ultrashort
laser pulses,” J. Opt. Soc. Am. B., vol. 14, 2716−2722 (1997).
[35] M. Lenzner, “Femtosecond laser-induced damage of dielectrics,”
International Journal of Mordern Physics B., vol. 13, 1559−1578,
1999.