OPTICS LETTERS / Vol. 29, No. 19 / October 1, 2004
Temporally and spectrally resolved imaging of
J. Siegel, G. Epurescu, A. Perea, F. J. Gordillo-Vázquez, J. Gonzalo, and C. N. Afonso
Laser Processing Group, Instituto de Óptica, Consejo Superior de Investigaciones Cientificas, Serrano 121, 28006 Madrid, Spain
Received June 18, 2004
We report a hybrid imaging technique capable of performing measurements of the spatial, temporal, and
spectral emission characteristics of laser-induced plasmas by use of a single detection system.
this technique to study the plasma produced by laser ablation of LiNbO3 and observe phenomena not seen
in such detail with standard instruments.These include extreme line broadening up to a few nanometers
accompanied by self-absorption near the target surface, and expansion dynamics that differ strongly between
the different species.Overall, the wealth of quantitative information provided by this novel technique sheds
new light on processes occurring during plasma expansion.
110.0110, 160.3730, 260.6580, 300.3700, 300.6500.
© 2004 Optical Society of America
Plasma emission spectroscopy (PES) is a powerful
optical tool for remote material analysis and is re-
ceiving growing interest in a wide range of research
It consists of irradiating a material with a
high-energy laser pulse and measuring the emission
spectrum of the laser-induced plasma formed near the
material surface. Because each atom in every state
of excitation has its own characteristic spectral finger-
print, PES (also known as laser-induced breakdown
spectroscopy) potentially permits the identification
of the elemental composition of a material.
can also be applied to the study of plasmas produced
during pulsed laser deposition, because the physical
properties of the deposited films strongly depend on
the plasma expansion dynamics.2
form, PES is performed with a spectrometer, and
time resolution can be added by use of a suitable
photomultiplier. A drawback of this approach is the
restriction to single-point measurements, providing
the spectral information of only a single spatial point
of the plasma or a space-averaged emission spectrum
of the entire plasma.However, space resolution is
clearly desirable since the plasma expands from the
sample surface and the intensities of the emission
lines become a function of space and time.
using an imaging detector, for instance, a time-gated
intensified charge-coupled device (ICCD), snapshots of
the expanding plasma at different stages of the expan-
sion process can be obtained, but spectral information
Spectral information can be added by use of
another imaging approach based on Fourier-transform
visible spectroscopy, providing the emission spectrum
for each pixel of the image.4
is time consuming because it relies on a mechanically
driven interferometer and requires an image to be
collected for each interferometer step (100 images in
In this work we combine the ability of PES
to distinguish between different species with the
capability of time-gated imaging to localize the
species in space and time.
our knowledge, the first time that precise mea-
surement of all three parameters (spectral, spatial,
In its simplest
However, this technique
This is, to the best of
and temporal) of the emission characteristics of a
are incorporated into a single imaging detector.
This detector, which can acquire the complete spatial
evolution of the spectral plasma emission lines induced
by a single laser pulse, is based on an imaging spec-
trograph and a time-gated ICCD.
fiber-based approach of Pérez-Tijerina et al.,5which
is able to reconstruct a low-spatial-resolution two-
dimensional (2D) image (25 pixels) of the plasma with
wavelength resolution, we deliberately trade in lateral
resolution of the plasma for high resolution in the
expansion direction, limited only by the ICCD pixel
size (24 mm). Harilal et al.6reported a system similar
to ours, but theirs requires the image to be scanned
over the entrance aperture.
benefits of parallel acquisition of emission spectra
along the entire expansion direction.
The experimental apparatus is shown in Fig. 1.
An ArF excimer laser (l ? 193 nm, pulse duration
of 20 ns) is focused on a target located inside a
vacuum chamber (1026mbars) generating a plasma
that expands along the normal to the target surface
(z direction).The emitting plasma is imaged onto
a Czerny–Turner imaging spectrograph (f?4 aper-
ture ratio, 25-cm focal length, 0.15-nm resolution,
2.5-nm?mm dispersion) by means of two quartz lenses
(f1? 160 mm, f2? 100 mm) and a periscope.
purpose of the periscope is to rotate the plasma image
by 90±such that the expansion direction (z) lies along
the orientation of the entrance slit of the spectrograph.
For the results presented here, a 475-nm long-pass
filter was installed in front of the spectrograph to
prevent second-order diffraction of short-wavelength
light. The spectrograph incorporates three grat-
ings (G1, 600 grooves?mm, blaze of 750 nm; G2,
300 grooves?mm, blaze of 1700 nm; G3, 150 grooves?
mm, blaze of 500 nm). At its output the spectro-
graph produces a one-dimensional (1D) spatial and
spectral image of the expanding plasma, in which
the vertical axis corresponds to expansion direction
z and the horizontal axis to wavelength l of the
emission of the species. In addition, if the gratings
Unlike the elegant
It therefore lacks the
0146-9592/04/192228-03$15.00/0© 2004 Optical Society of America
October 1, 2004 / Vol. 29, No. 19 / OPTICS LETTERS
imaging with spectral resolution of laser-induced plasmas.
Experimentalsetup fortime-resolved plasma
are operated in reflection rather than in diffraction
and the entrance slit is fully open, a 2D spatial image
of the plasma can be formed.
spectrograph, in the image plane, a time-gated ICCD
camera is installed.This ICCD has a spectral range
from 180 to 870 nm, a resolution of 512 3 512 pixels,
with an effective pixel size of 24 mm, an analog-to-
digital resolution of 16 bits, and a minimum gate
width of 2 ns. A delay generator is employed to delay
the time gate of the ICCD with respect to the laser
pulse by a user-defined value.
To demonstrate the flexibility, sensitivity, and
large spectral window of our system, we acquired four
time-gated images of the expanding plasma of LiNbO3,
each obtained by irradiation with a single laser pulse
(fluence of 1.2 J?cm2).Figures 2(a) and 2(b) show
two images obtained in the 2D spatial imaging mode
and Figs. 2(c) and 2(d) show two images obtained
in the 1D spatial and spectral imaging mode.
images were obtained with a constant gate width
(Dt ? 30 ns) but different delay times [t1? 50 ns in
Figs. 2(a) and 2(c) and t2? 140 ns in Figs. 2(b) and
2(d)] with respect to the laser pulse.
stages of the plasma expansion process from the
target surface (z ? 0 mm) can be appreciated in
the 2D spatial images [Figs. 2(a) and 2(b)], showing
a directional expansion into the z direction with a
relatively low angular spread.
the target surface is caused by chromatic aberrations
occurring in the two lenses used, due to the broad
spectrum imaged (475–850 nm).
By switching to the 1D spatial and spectral imaging
mode, operating grating G3 in diffraction, and using
an entrance slit width of 50 mm, we obtained the cor-
responding images shown in Figs. 2(c) and 2(d).
vertical axis still corresponds to the z axis of plasma
expansion, but the horizontal scale now represents
the wavelength of the emitted light.
of each image pixel corresponds to the local emission
intensity at the given delay and constitutes the third
At the output of the
The apparent blur at
emission with respect to the horizontal plane is
caused by chromatic aberration.
expansion [t1, Fig. 2(c)] a continuous spectrum shows
up near the target surface, which can be attributed to
a bremsstrahlung generated by deceleration of plasma
In addition, three individual emission
lines can be clearly identified in Fig. 2(c).
at 610.3 nm (3d2D3/2,5/2! 2p2P1/2,3/2) and 670.7 nm
(2p2P3/2,5/2 ! 2s2S1/2,3/2) correspond
transitions of excited Li neutrals (Li?).
line (3p5P3,2,1! 3s5S2) corresponds to transitions of
excited oxygen neutrals (O?).
emission lines due to Nb neutrals (Nb?) are present,
but these are barely perceivable because they spatially
overlap with the bremsstrahlung.
At a later stage of the expansion process [t2,
Fig. 2(d)] the emitting species expand further and
individual Nb?lines can be resolved.
the intensity of Nb?lines is much lower than that
of Li?lines, together with a decreasing sensitivity of
typical photocathodes in the infrared region where O?
emit, makes the simultaneous detection of all species
involved a challenge. However, as can be seen in
Fig. 2, our system is capable of recording Li?, O?, and
Nb?lines simultaneously due to the large dynamic
range of our detector and allows two entirely different
expansion dynamics to be discerned.
that O?is ejected from the target at only the early
stages of the ablation process since the distribution of
the O?line 90 ns after the laser pulse [t2in Fig. 2(d)]
shows a distinct gap to the target surface.
trast, Li?and Nb?are still emitted from the target
surface at this delay.
Spectral resolution can be improved by use of a grat-
ing with more grooves per mm.
an image obtained with a single laser pulse with G1.
The time gate was extended to 5 ms to integrate over
all stages of expansion. The multitude of Nb?lines
particularly in the 570–600-nm spectral range can be
clearly seen.The strong Li?line at 610.3 nm extends
The slight tilt of the near-surface
In the early state of
Besides, a multitude of
The fact that
It can be seen
Figure 3(a) shows
the surface of a LiNbO3target (z ? 0 mm), each obtained by
irradiation with a single laser pulse by use of (a), (b) the 2D
spatial imaging mode and the (c), (d) 1D spatial and spec-
tral imaging mode.The gate width was constant (Dt ?
30 ns), and the gate delay with respect to the laser pulse
was changed from (a), (c) t1? 50 ns to (b), (d) t2? 140 ns.
The images are represented in a logarithmic intensity scale
to compensate for the large differences in emission inten-
sities between the different species (Li?, O?, and Nb?).
Time-gated images of the plasma expanding from
OPTICS LETTERS / Vol. 29, No. 19 / October 1, 2004
high-resolution image of the expanding plasma obtained
by irradiation with a single laser pulse.
intensity profiles of the Li?line at 610.3 nm at a distance
of 0.62 mm (dashed curve) and 10 mm (solid curve).
arrow indicates the region of self-absorption (small dark
triangle). (b) Linewidth of the 610.3-nm line determined
from the data in (a) as a function of target distance z.
The shaded area very close to the surface (0.5–0 mm)
indicates the region of self-absorption, and for z . 2 mm
the values are limited by the spectral resolution of our
(a) Time-integrated 1Dspatial andspectral
The inset shows
to more than 10 mm, and two interesting processes can
be observed that occur close to the target surface:
broadening and self-absorption.
be caused by different mechanisms.8
broadening under these conditions is typically limited
to the subangstrom range, Stark broadening due to col-
lisions between Li?and electrons dominates for plasma
densities above 1014cm23.8
We extracted from Fig. 3(a) the evolution of the
linewidth with target distance z.
representative spectra obtained at two distances,
each one obtained by averaging over two horizontal
lines in Fig. 3(a). Strong line broadening can be
observed very close to the target compared with a
narrow line at a target distance of 10 mm.
spectra were fitted to a Lorentzian function, whose
width Dl was plotted in Fig. 3(b) as a function of
z. For large distances the measured linewidth is
limited by the resolution of the spectrograph to
Dl ? 0.5 nm. However, for z , 2 mm a dramatic
increase in linewidth can be easily resolved, yielding
a maximum value of Dl ? 3.3 nm for z ? 0.5 mm.
Because Stark broadening dominates under these
conditions by electron impact, local electron density
Necan be calculated directly from Dl with the relation
Dl ? 0.2W?Ne?1016? nm.9
parameter, which for the Li?(610.3 nm) line is weakly
dependent on temperature.8
we obtain Ne ? 7.6 3 1017cm23at z ? 0.5 mm.
This value can be compared to the one obtained in
a recent study9on LiNbO3 limited to a minimum
target distance of z ? 2 mm.
Line broadening can
The inset shows
W is an electron impact
Using this relation,
The authors employed
a single-point scanning technique and reported a
maximum value of Ne ? 2.9 3 1016cm23. Because
Neis expected to decrease with 1?z, our value, which
is more than 1 order of magnitude higher, is consistent
with the much shorter target distance.
For shorter distances [shaded area in Fig. 3(b)]
the linewidth cannot be measured precisely, because
self-absorption begins to strongly affect the line shape,
showing a dip in the central part of the spectrum.
Self-absorption may occur when the density of a
given species is high enough for light emitted by
this species to be reabsorbed by the same species.10
Self-absorption of the 610.3-nm Li?line can be seen in
Fig. 3(a) by the small dark triangle close to the target
In conclusion, we have developed a novel technique
that allows precise measurement of the spectral,
spatial, and temporal emission characteristics of
laser-induced plasmas by use of a single detection
system. By applying it to the study of LiNbO3, we
have demonstrated the system’s sensitivity, its large
spectral window, and its flexibility to switch between
the 2D spatial and 1D spatial and spectral imaging
mode. With this technique we have been able to
separate populations of different species by their dif-
ferent expansion dynamics.
to observe bremsstrahlung, strong self-absorption,
and Stark broadening effects near the sample surface,
allowing a quantitative determination of the electron
density as a function of target distance, obtaining
values of up to 7.6 3 1017cm23near the target surface.
It has also been possible
G. Epurescu acknowledges funding by the European
Paul Tadrous for providing the software11for logarith-
mic intensity representation of the images in Fig. 2.
J. Siegel’s e-mail address is firstname.lastname@example.org.
We are grateful to
1. For a review see, for instance, the feature issue of Appl.
Opt. 42, 5937–6225 (2003).
2. D. B. Geohegan, in Pulsed Laser Deposition of Thin
Films, D. B. Chrisey and G. K. Hubler, eds. (Wiley,
New York, 1994), Chap. 5.
3. D. B. Geohegan, Appl. Phys. Lett. 62, 1463 (1993).
4. V. Bulatov, L. Xu, and I. Schechter, Anal. Chem. 68,
5. E. Pérez-Tijerina, R. Machorro, and J. Bohigas, Rev.
Sci. Instrum. 75, 455 (2004).
6. S. S. Harilal, C. V. Bindhu, M. S. Tillack, F. Najmabadi,
and A. C. Gaeris, J. Appl. Phys. 93, 2380 (2003).
7. H. R. Griem, Principles of Plasma Spectroscopy (Cam-
bridge U. Press, New York, 1997).
8. H. R. Griem, Spectral Line Broadening by Plasmas
(Academic, New York, 1974).
9. F. J. Gordillo-Vazquez, A. Perea, J. A. Chaos, J.
Gonzalo, and C. N. Afonso, Appl. Phys. Lett. 78, 7
10. T. Sakka, T. Nakajima, and Y. H. Ogata, J. Appl. Phys.
92, 2296 (2002).
11. Bialith image processing software (www.bialith.com).