See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/227210787

Laser induced breakdown spectroscopy (LIBS)

Chapter · January 2006

DOI: 10.1007/1-4020-4789-4_13

CITATIONS READS
19 412

6 authors, including:

L. Caneve Francesco Colao

ENEA ENEA
80 PUBLICATIONS   687 CITATIONS    173 PUBLICATIONS   3,009 CITATIONS   

SEE PROFILE SEE PROFILE

L. Fornarini Violeta Lazic

ENEA ENEA
84 PUBLICATIONS   699 CITATIONS    74 PUBLICATIONS   2,031 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Many projects View project

CO.B.RA. Regional Project View project

All content following this page was uploaded by L. Fornarini on 03 December 2015.

The user has requested enhancement of the downloaded file.

LASER INDUCED BREAKDOWN SPECTROSCOPY (LIBS)

Process and applications to artwork, environment and security

R.FANTONI, L.CANEVE, F.COLAO, L.FORNARINI, V.LAZIC,

V.SPIZZICHINO*
ENEA – Applied Physical Technologies Dep.- Laser Applications Section
V. E. Fermi 45, I-00044 Frascati (Rome) Italy
* ENEA Fellow

Abstract
In this chapter, the laser ablation process involved in laser induced breakdown occurrence,
with plasma formation and relaxation is shortly described. The fundamentals of laser induced
breakdown technique are revised presenting current laboratory instrumentation and methods
for qualitative, semi-quantitative and quantitative analyses. Performances in air, low-pressure
gasses and underwater together with the analysis of different materials are discussed.
Some applications ranging from artwork characterization to environmental analyses are
reviewed with special attention to fragile substrates, ceramics, marbles, rocks relevant to
planetary exploration and underwater measurements of environmental interest. Additional data
relevant to monitor industrial process and organic matter characterization are presented.

1 Introduction

Laser Induced Breakdown Spectroscopy (LIBS) is an elemental analysis based on the atomic
emission from a plasma generated by focusing a powerful laser beam on a sample, which can
be solid, liquid or gaseous. First experiments, aimed to detect atomic line emission from laser-
induced plasma at metal surfaces, date back to the early eighties [1]. However before the
technique became appealing for quantitative analytical applications, instrumental and data
elaboration methods had to be considerably developed to overcome detecting apparatus
limitations arising from the intense background plasma emission dominating the process in its
first stages. Fundamental steps in this direction were achieved in the nineties, with the
development of intensified optical multichannel analyzers as detectors. They are able to offer
both fine time gating and high spectral resolution, the latter being obtained with the coupling to
a high-resolution monochromator. Since then, the process of laser ablation at high laser power
could be studied in detail in order to establish the fundamental parameters of the technique.

9-1
Particular attention has been devoted to the need of preserving the sample stoichiometry during
surface evaporation and the occurrence of local thermal equilibrium conditions in plasma, for
selected time windows, during the analysis.

1.1 THE PROCESS

The laser interaction with matter always involves an energy exchange. When a laser beam is
focused onto a dense matter the laser energy can be scattered at the same wavelength, re-
emitted at different wavelengths, absorbed or transmitted through the sample. The interaction
may lead to a partial evaporation originating a process known as laser induced ablation. Its
main characteristics are:
• Localized heating at the surface, both in space and in time;
• Fast process evolution, with a complete cycle shorter than 10 ms (for pulse widths < 100
ns) [2].
At high laser irradiances (e.g. > 1MW/cm2, depending on the target properties) the
evaporation is followed by plasma formation through several subsequent processes: surface
melting, vaporization, vapor ionization, surrounding and evaporating gas breakdown, free
electron acceleration within the evaporated cloud, plasma heating. The plasma growth and
decay evolves from initial expansion, shock waves formation and propagation in the ambient
gas, deceleration of free electrons with inverse Bremsstrahlung emission, particles collisions in
the gas cloud with excitation and relaxation of atoms/ions, chemical and radiative
recombination.
At first, the plasma temperature (typically T > 30000 K), electron and ion densities (1017 -
10 cm-3) are very high and plasma is characterized by high auto-ionization. The continuum
20

emission is the dominant feature of the spectrum [3]. As the plasma expands and cools,
emission lines, from ions and atoms, start appearing superimposed on the strong continuum
background. Owing to the high electron density, the emission lines are broadened by the Stark
effect. Afterwards, the continuum background rapidly decays; the atomic lines become
narrower and weaker while the atomic emission decays slowly. Figure 1 shows LIBS emission
spectra recorded on a steel sample after UV laser excitation (λ=308 nm) at different delay
times. As shown in the Figure, a properly selected detection window must be chosen to
improve the signal-to-background ratio of data. This is very important to obtain quantitative
measurement.
The analytical measurements are generally performed after the initial stages of plasma,
when a state of local thermal equilibrium (LTE) is achieved. LTE is a dynamic status reached
during the plasma relaxation when plasma parameters slowly vary in a selected temporal
range. At LTE condition, electrons have a Maxwell energy distribution, all the species have the
same temperature, and analyte intensity is proportional to the relative population of the level
and follows the Boltzmann statistics:

9-2
 n jp,s g jp,s
= exp[ −( E jp,s − E 1j ,s ) / k BTe ] (1)
n1j ,s g 1j ,s
where the thermodynamic temperature is replaced by the electronic one Te, kB is the
Boltzmann’s constant, gpj,s and g1j,s are the considered states degeneracy, and Epj,s-E1j,s is the
energy difference between excited p and ground state 1.
The Saha’s law, on the other hand, rules the distribution of species at different ionization
levels of the same element:
n e ni , s U i ,s ( 2π me k B Te ) 3 / 2
1
=2 3
exp[ − E s∞ / k B Te ] (2)
na ,s U a ,s h
where h is the Planck’s constant, me the electron mass, ne the electron density, na,s and ni,s are
the atoms and ions densities, Es∞ is the effective ionisation energy and Ua,s and Ui,s are partition
functions of atoms and ions.: U j ,s = ∑ p g jp,s exp( − E jp,s / k BTe )
Upon LTE conditions, characteristic line emission spectra mostly derive from the produced
neutral atoms and species in the first ionized state. Note that a high electron density is needed
to reach LTE: electron collisional rates must exceed radiative rates by at least one order of
magnitude [4]. Further plasma cooling and recombination cause its departure from LTE. The

Intensity (a.u.)

50
s)
100 (n
y
la
De
200

500

402 404 406 408 410 412

W avelength (nm)

Figure 1 - Example of LIBS spectra from stainless steel as a function of acquisition delay
from onset of the laser pulse. The signal acquisition gate is 50 ns.

9-3
 process, in its final stages, is characterized by cluster and nanoparticles ejection in the dark
slow expanding tail of the plume [5]. In laser ablation at high energy densities the bright
plume is observed together with the formation of a crater on the sample surface. During the
evaporation atoms are ejected perpendicularly to the target. The higher particle density along
this direction is due to the crater shape, which conveys material towards the laser beam
direction, thereby increasing excitation. Consequently, LIBS emission intensity is highest at
normal incidence. The central plume region has the most intense emission due to the higher
density of the emitting species.

1.2 DOUBLE PULSE EXCITATION

Thermal, mechanical and chemical properties of the medium can strongly affect the time
evolution of the process and may limit its analytical applications. In fact the time window,
useful for data acquisition, is dramatically reduced when increasing the density of the
surrounding medium. This problem originally prevented quantitative analyses in liquid phase.
The use of trains of laser pulses, suitably delayed, in order to initiate and sustain the process,
has been successfully exploited in order to overcome this limitation [6]. In particular the
double pulse (DP) technique, consisting of focusing two pulses, delayed by 1 - 100 µs, onto the
target, has been successfully applied to the analysis of water solutions and submersed samples
[7]. The first pulse creates a bubble in the dense medium, while the second, coming during its
expansion, creates a secondary plasma in the developed rarefied medium. The main
characteristics of the double pulse LIBS for application in dense media are summarized in the
following:
• LIBS signal is detected after the second pulse, during the expansion of the secondary
plasma in the low-density medium generated by the first pulse. The first one produces
only weak plasma.
• The expansion of the secondary plasma contained in the low-pressure bubble occurs on a
time scale longer than what expected in the high-density medium.
• Narrow emission lines are detected due to the relatively low pressure inside the bubble.

1.3 GENERATION OF LIBS SIGNALS

Once a high irradiance pulsed laser beam is focused on a dense medium or onto a surface, a
laser-induced plasma is formed and LIBS signal can be detected during plasma expansion and
decay. In case of double (multiple) pulse excitation, LIBS signal is always related to the
second laser pulse, which sets the zero delay time for data acquisition. In order to allow for
reliable analytical applications the proper LTE conditions must be selected for data acquisition,
which consists of recording spectrally resolved signal averaged in the pre-selected time
window.

9-4
2 LIBS Instrumentation

Before describing in further detail how a standard LIBS set-up for analytical applications is
built, it is worthwhile to remind that one of the main advantages of the technique relies on its
characteristic micro sampling. In fact, the technique can be considered as only micro-
destructive since just less than 1 µg of material is ablated at the sample surface.
The main constituents of a LIBS set-up are a pulsed laser source, a spectrometer (dispersive
element plus multichannel detector) with suitably designed optical systems (laser focusing and
collecting plasma emission) and electronics components (trigger and delay generators) (see
Figure 2). The experiment is usually fully computer controlled.

Time gateable detector

PC
Electronics
Monochromator
Laser
Optical fiber
Sample
Receiver
optics

Figure 2 – LIBS analysis set up.

2.1 LASER CHOICE

The laser choice is a major issue in developing the apparatus, e.g. the choice of the main
parameters (wavelength and pulse duration) of the high power source (> 10 MW, at any
repetition rate compatible with the use of the spectrometer).

9-5
2.1.1 Wavelength selection
In the selection of the appropriate wavelength the following features should be considered:
• Q-switched Nd:YAG laser sources (with fundamental emission at 1064 nm in the near
IR) are usually used for plasma generation. Their operation at the second (532 nm), third
(355 nm) or fourth (266 nm) harmonics allows to analyze regions of shorter wavelength,
up to the UV, but it reduces the available power.
• The UV laser generated plasma is characterized by more intense emissions at equivalent
laser photon density, shorter plasma durations and less intense continuum emission
compared to the plasma generated by IR radiation.
• Less intense ion emission and narrower emission lines are observed in plasmas generated
by an UV laser with respect to those generated by an IR one [9] as the temperature and
the electron density reached are usually lower [8].
• The UV laser excitation is preferable for substrates having either a high evaporation
threshold (ceramics, stones) or high IR reflectivity (metals).
• Lower thermal effects are induced at the sample surface by the UV radiation, with a
reduced probability of secondary chemical reactions (oxidation in air), which could
modify surface composition.
• Generally, better plasma stoichiometry is achieved with UV excitation due to a more
efficient surface vaporization.

4 + 355 nm
9.0x10 Cu
1064 nm
4
6.0x10 + Cu
+ Pb Sn
Fe Sn
4
3.0x10

0.0
275.0 277.5 280.0 282.5 285.0 287.5
Intensity (a.u.)

5
6.0x10
5
Cu Cu
4.0x10
Cu
5
2.0x10
0.0
500 520 540 560 580 600
5
6.0x10
O N
5 Cu O
4.0x10
5
Cu
2.0x10

0.0
760 780 800 820 840 860
W avelength (nm )

Figure 3 - Comparison between spectra produced by a Nd:YAG laser operating at 355 nm and

9-6
 at 1064 nm for a bronze sample. From top: magnifications of the UV, visible and IR
spectral regions respectively.

As an example, LIBS spectra obtained on a bronze sample by laser emitting in the UV and
IR are compared in Figure 3. Three different spectral regions have been selected to show in
detail differences in signal intensity and background emission.
For in water analyses the following additional issues in wavelength selection must be taken
in to account:
• Plasma ignition is easier to achieve by higher incident photon energy, e.g. UV excitation,
which is required to break the water bond.
• Strong absorption of water in the UV and red wavelength regions suggests the use of
visible (400-500 nm) excitation. It attenuates the emission spectra with analytical
limitations.
• The energy per pulse available from commercial lasers suggests to work only in the near
IR with the Nd:YAG fundamental excitation line.
An accepted compromise for experiments in water is the use of near IR (1064nm) or
visible (532 nm) sources.

2.1.2 Pulse duration

Another choice of the laser characteristics is the pulse duration; it can range from few
femtoseconds to tens of nanoseconds. In this respect the following points are now fully
established:
• Stochiometric evaporation of the sample is better preserved when using ultrashort laser
pulses; less material is ablated, a cleaner crater is formed on the substrate (Figure 4).

Figure 4 – SEM pictures of craters formed a bronze surface after 200 shots of ns (left) and fs
(right) laser pulses. Laser parameters: 90 mJ, 8ns at 532 nm (l); 2.4 mJ, 250 fs at 527 nm
(r). Operation at 10Hz with the same focusing.

• Atomic lines should be favored at the same irradiance (note that higher irradiance are
easily reached with fs for the very effective pulse compression); however Stark effect

9-7
 distorts atomic line profiles at high irradiances (Figure 5).

LIBS intensity (Normalized)

Pb

Sn
fs
1.0 ns

Pb
Cu

Sn
0.5

Pb
Pb

Mg
Sn
M g+

0.0
278 280 282 284 286 288 290
Wavelength (nm)

Figure 5 - Comparison between ns and fs LIBS spectra in the UV region nm on the same
bronze sample. Laser parameters as in Figure 4. Main Cu, Sn and Pb line assignments are
indicated.

2.2 SPECTROMETER CHOICE

Two options are generally available for laboratory systems: a high-resolution monochromator
coupled to an Intensified CCD detector or an Echelle spectrometer coupled to a large ICCD.
The latter, which is the most expensive combination, has the advantage that different spectral
regions are dispersed at different wavelength orders on selected areas of the detector. This
allows collection of the entire UV-near IR spectrum at medium-high resolution, for example
200-1000 nm, in just a single shot. However, the signal output is lower than by using a
monochromator. This can be a problem for minor/trace element detection, particularly when
considering low intensity and short-duration plasmas as in the case of underwater LIBS
analyses.
On field systems often have the mandatory requirement to stand harsh environments and/or
to use very compact configurations. To this respect, whenever possible, series of very small
linear arrays detectors are preferred, although this choice implies a reduced sensitivity due to
the lack of the intensifier and a less effective plasma background filtering

9-8
2.3 EXPERIMENTS UNDER CONTROLLED ATMOSPHERE AND IN WATER

2.3.1 LIBS in controlled atmosphere

The ambient conditions, e.g. pressure and type of background gas, influence spectral emission
intensity of laser-produced plasmas. Decreasing the background pressure plasma is able to
freely expand into a larger volume above the target. It decreases in density, number of
collisions, excitation of the atomic species and analyte signals also decrease. Plasma decay is
faster due to faster expansion (Figure 6) while emission intensity is lower due to the less
efficient inverse Bremsstrahlung heating during the laser pulse and plasma temperature is
consequently lower (Figure 7).

Fe I
background
1.0 Fe II

Air

0.5
Normalised LIBS intensity

0.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0

1.0

CO 2 P=6 m bar

0.5

0.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Delay ( µs)

Figure 6 - Time evolution of the plasma integral line intensities for Fe I and Fe II lines in air
and CO2 atmosphere, as measured on a Mars analogue rock sample (Andesite). Laser
parameters: wavelength 355nm, pulse energy 21 mJ, pulse duration 8 ns; acquisition
parameters: gate width 100 ns.

9-9
 No enhancement in the intensity of weak lines could be observed by multiple pulse
techniques applied on the samples examined in rarefied atmosphere, since the bubble
confinement mechanism is effective only at high pressures [10].

Air
6
1.0x10 Al Al CO 2 (6 m bar)

Ti II Ti II
5
8.0x10
Intensity (a.u.)

5
6.0x10

5
4.0x10

5
2.0x10

0.0
307 308 309 310
W avelength (nm )

Figure 7 - Part of LIBS spectra detected in CO2 and air on the same Mars analogue sample
of Figure 6. Laser parameters as in Figure 6, acquisition parameter: delay 1 µs, gate 0.1 µs.

2.3.2 LIBS in water

The peculiar optical characteristics of surface analysis of submersed samples can be
summarized as follows:
• The double pulse emitted from the laser source horizontally enters the cuvette through the
lens sealing one of its sidewalls.
• Signal collection is perpendicular to the laser beam across the solution.
• Signal is driven into the entrance of the monochromator by a fiber glass and detected at
its exit by the ICCD.
LASER LASER LASER

a b c

9-10
 Figure 8 – Images of Double pulse LIBS in water: Plasma shape photographed by a digital
camera: a) t1=145 µs, ∆t=80 µs, 5 pre-pulses, E1Tot=72 mJ, E2=140 mJ; b) t1=155 µs, ∆t=80 µs,
2 pre-pulses; E1Tot=110 mJ, E2=123 mJ; c) t1=165 µs, ∆t=80 µs, 1 pre-pulse, E1=181 mJ, E2=57
mJ
The use of the double pulse LIBS technique in water is mandatory whenever the detection
of minor/trace elements is required, as it allows generating plasma in a reduced medium
density (see sect. 1.2). The occurrence of intense plasma after the second laser pulse can be
recorded in collecting LIBS images at selected wavelength. As an example LIBS images
collected in different operating conditions after the second pulse are shown in Figure 8.

3 LIBS data analysis procedure

A complete LIBS spectrum, recorded in all UV – visible range (typically from 200 nm to 850
nm) is made of thousands of spectral lines, even for rather simple alloys with a few (2 to 4)
main constituents, such as commercial aluminum or bronzes (see Figure 9).

5
2.5x10

5
2.0x10
LIBS Intensity (a.u.)

5
1.5x10

5
1.0x10
4
6.0x10
4
5.0x10
CuI
LIBS Intensity (a.u.)

4
0.0
4.0x10
400 800
Wavelenght (nm)
SnI
PbI
4
2.0x10 ZnI
CuI

0.0
315 330 345 360 375
Wavelenght (nm)

9-11
Figure 9 - LIBS spectrum of a Bronze sample (quaternary ,Cu/Sn/Zn/Pb, alloy). Full spectrum
recorded with the ICCD behind the Echelle spectrometer. Section used for analyses with the
assignment of main atomic lines.

Single line shape can be largely affected by plasma formation and expansion as well as by
instrumental factors. Main line broadening mechanisms can derive from collisions due to the
surrounding gas pressure, Stark effect due to the electrical field, Doppler shift due to the
temperature, plasma non-homogeneity due to the optical thickness, and finally from
instrumental factors related to the overall spectrometer resolution.
In general LIBS spectral lines show both homogenous and non-homogenous broadenings,
thus the most appropriate line fitting is obtained using a Voight profile, although Lorentzian
line fits are suitable for most practical purposes in high-density plasma.
With a zero line intensity at the sample surface and a negligible induced emission [11], the
line intensity at frequency ν which is emitted by a homogeneous plasma of thickness L, is:
2 hν 3 1
I αki (ν ) = − hν / k BT
(1 − e − kν L ) (3)
c 1− e
2

where c is the light velocity and kν is the plasma absorption frequency dependent coefficient
LIBS plasma is usually treated in first approximation by assuming a simplified model that
considers observing optical thin plasma (kν L <<1) during a time window where LTE
assumption holds and Saha equation rules the atom to ions distribution.
Under LTE assumption the integrated line intensity for a selected α species derived from
eq. (3) gives:
g k Aki e − Eki / kBT
I α = KCα (n)
(4)
U (T )
where C(n)α is the element concentration in the (n-th) ionization state, Aki the transition rate
coupling k and i levels, gk the degeneracy of the k-th initial level, U(T) the partition function
(eq. 2) at the equilibrium temperature T and K a system constant (optical efficiency, etc.).
ln(Iλ /gk Ak

Upper energy level (eV)

9-12
Figure 10 - Temperature determined from atomic Fe line intensities in an Fe sample
T is determined by means of a Boltzmann plot of the line intensities versus the initial state
energy for a given species (Figure 10).
In order account for atom/ion distribution for each α species, the sum over all the n-th ionic
states of each group must be performed. This in turn depends on the electron density relevant
to the plasma in the observed time window. The electron density ne can be retrieved with an
accuracy of about 20-30% from the measured Stark broadening:
5/ 4
∆w Stark = 2W
Ne
1016
⎛ n ⎞
+ 3.5 W ⎜ e16 ⎟ (1 − 0.75N ) A
1/ 3
D (5)
⎝ 10 ⎠
T 3/ 2
where N D = 1.72 10 9 is the Debye shielding parameter, W and A are Stark parameters
n1e / 2
that weakly depend on temperature [12].Atom (a) to ion (i) ratios for the same s species are
obtained from the electron density assuming the validity of Saha equation (see eq. 2) [13],
where the electron temperature matches the atomic temperature T in eq. (2).
Eqs. (1) - (5) allow to handle LIBS spectra and to support LIBS applications as analytical
technique. Should non-equilibrium phenomena occur, the concentration values retrieved from
ionic and atomic lines coupled in the plasma decay may be affected by large errors [14]. Data
for atomic iron lines in a stainless steel sample are reported in Figure 11 at different time
delays from plasma onset.

50 ns
100 ns
200 ns
350 ns
2000 ns
) FeI
p
/g FeI
p
log(n

4 .0 4 .5 5 .0 5 .5 6 .0 6 .5 7 .0 7 .5
p
E FeI
(e V )

9-13
 Figure 11 – Atom state distribution function (see eq.1) for Fe at different time delay from
plasma onset. Different data series are shifted on the ordinate axis avoiding data superposition.
Data collected on a stainless steel sample excited at 355 nm.
The first population density estimate, made at 50 ns from plasma onset, shows
overpopulation in the low energy levels indicating that plasma core is still in the ionization
stage. Population distributions for subsequent time delays, from 100 ns on, appear as straight
lines in the semi-log plot, this trend is approximately maintained for all delays until 1 µs. In
this time window, plasma moves from partial LTE, characterized by space non-homogeneities
to LTE. After 1.5 µs for Fe I, a second departure from LTE occurs, with a peculiar under-
population of low energy levels, caused by radiative losses and associated to the recombining
condition of plasma. Temperature estimates depend on the energy range chosen, thus showing
the failure of the unequivocal definition of temperature. Departure from LTE conditions can
also be expected from pressure and analyte concentration gradients along the edge of the
plasma plume. In these regions the temperature sharply decreases and the energy transfer to
surrounding atmosphere is more effective [15]. The transport of charged particles out of the
active region also contributes to modify the local Boltzmann and Saha equilibrium. These data
demonstrate that the LTE time window may be quite narrow and considerable deviations can
be observed both at short and long time delays after plasma ignition.

2 .5

2 .0
Cm(I)+Cm(II)

1 .5

1 .0

0 .5 F e I(3 .6 e V )+ F e II(5 .5 e V )
F e I(5 .5 e V )+ F e II(5 .5 e V )
F e I(3 .6 e V )+ F e II(8 .5 e V )
0 .0
0 500 1000 1500
D e la y (n s )

Figure 12 – Retrieved total Fe concentration for steel sample in the listed pairs of starting
levels

Total iron density (FeI + FeII) for different starting levels is reported in Figure 12 for a
steel sample. As shown, total iron concentration measured on FeI at Ep=3.6 eV and on FeII at
Ep=5.5 eV reaches a quasi steady state for times longer than 100 ns till 1500 ns, within an error
bar of 15%. When different energy levels are considered, a much longer delay is necessary to

9-14
reach a quasi steady state because of modifications in Boltzmann populations due to radiative
processes. In the same figure, the worst situation is represented by the intermediate energy
levels (Ep=5.5 eV for both atomic and ionic species), which are more affected from imbalances
between radiative and collisional processes. In this case a significant interaction occurs
between atomic and ionic levels of the same energy for short time delays, whereas after 500 ns
the increase of ions recombination becomes the dominant process repopulating atomic iron.

3.1 ELEMENT RECOGNITION: QUALITATIVE ANALYSIS

Qualitative analysis is based on the recognition of the most intense characteristic lines of each
element with the help of current line emission databases [16]. The procedure can be carried out
in a semi-automatic way, by means of a custom designed software (base on LabView programs
in our case), after the identification of a major element (usually a metal, e.g. Iron or heavy
element) with a sufficient number of well identified lines, which can be used in Boltzmann plot
to estimate plasma temperature. Different portions of the experimental spectrum are compared
to a simulation obtained from the data base line positions, where the intensities are calculated
at the estimated temperature and the instrumental line width is taken into account.
An example of qualitative analysis performed on materials suspended in seawater,
collected on a polycarbonate film after filtration, is given in Figure 13. Labels related to the
elements identified in different spectral regions, free from substrate C+, C and C2 emissions,
are given. Spectra are assigned on the basis of a simulation at 8400 K, consistent with Fe line
intensities.

5
6x10 Mg+ Si 5
1.5x10
Fe
LIBS Intensity /100 shots
LIBS Intensity /100 shots

5
5x10

5
4x10 10
5 Si

5 Mg
3x10
Si
5 Fe 5
2x10 5x10
Al
5
10

0 0
278 280 282 284 286 288 290 298 300 302 304 306 308 310
Wavelength Wavelength

9-15
 Figure 13 - Identification of different elements Fe, Mg, Si, Al found in suspended
organic matter collected from Antarctic seawater on a polycarbonate filter, other regions
showed the presence of Na, Cr and Cu, too (dots simulation, solid line experimental data).

3.2 ELEMENT DETERMINATION: SEMI-QUANTITATIVE AND QUANTITATIVE

ANALYSIS

The use of RU lines (Rayon Ultimes - well known intense emission lines used for each element
quantification in emission spectroscopy) for analytical determinations is strongly
recommended. Generally, significant lines for quantitative analysis must be selected on LIBS
spectra according to the following criteria:
• Well-resolved lines – e.g. isolated lines.
• Free from self-absorption and self-reversal lines [12] – e.g. lines with a medium-low
intensity.
• Related to transition not involving the ground state – mostly to avoid the onset of non-
linearity due to spatial non-homogeneities phenomena.
Lines selected for quantitative analysis can be integrated either numerically or by fitting the
experimental data to an appropriate line shape function. Background is then subtracted from
the integrated intensity. To this aim background can also be evaluated by numerical integration
performed on the same number of nearby channels, free of line peaks.

3.2.1 Calibration curves

In order to achieve quantitative results, the most common approach implies the construction of

1 .0
Ca I
0 .5

0 .0
Corrected intensity

0 200000 400000
C a (p p m )
4 Mg I
2

0
0 1000 2000
M g (p p m )

0 .2 Mn I
0 .1

0 .0
0 1000 M n (p p m ) 2000

9-16
 Figure 14 – Example of calibration plots from calcium carbonate references for some marble
characteristic elements

calibration curves performed on targets with known composition (standards). Unwanted matrix
effects can be limited by selecting standards with a similar matrix, both in major constituents
and in total morphology, the latter mainly affecting the sample physical properties. A few
points must be kept in mind in creating calibration curves:
• Calibration curves are built in order to normalize the line intensities measured with
respect to the same ones relevant to reference samples.
• Calibration curves may include saturation effects [17] (calibration plots from calcium
carbonate references for some marble characteristic elements are shown in Figure 14).
• In order to reduce the matrix effect in substrates for which standards are not available,
calibration curves may be corrected to account for significant differences in temperature
and electron density [18]. In combination with calibration curves, internal
standardization to one major element may help in case of fractionation, when the plume
has not the same stochiometric composition as the target [19].
An example of normalization to internal standards is shown in Figure 15 for a quaternary
bronze alloy containing different amounts of Cu, Pb, Sn and Zn. In this case the
standardization to Zinc is more effective in reducing fractionation problems than the
standardization on Copper.

3 R=0.715 2 R=0.987
2
1
1
0
0
0.00 0.05 Pb/Cu 0.10 0.15 0.0 0.5 Zn/Pb 1.0
LIBS intensity ratio

1.5
R=0.897 6 R=0.959
1.0
3
0.5
0
0.08 Sn/Cu 0.12 0.0 0.5 Zn/Sn 1.0

2 R=0.966 2 R=0.968
1 1

0 0
0.00 Cu/Zn 0.05 0.00 Zn/Cu 0.05

9-17
 Figure 15 – Standardization to Copper (left) and to Zinc (right) of LIBS intensities from
quaternary bronze ref. samples. Laser parameters: wavelength 1064 nm, fluence 74 J/cm2,
pulse width 8 ns, rep. rate 10 Hz.

3.2.2 The calibration free approach

This approach requires the simultaneous determination of all species present in a sample, and
usually works well in the case of well-known matrices (e.g. alloys). As proposed by the group
at IPCF-CNR [20], the unknown experimental constant KCF in eq. (2) can be estimated from
the closure relation:
⎛C K ⎞
qα = ln ⎜⎜ α CF ⎟⎟ (6)
⎝ U α (T ) ⎠
1
∑
α
Cα =
K CF α
∑ U α (T )e qα = 1 (7)

where Cα is the concentration of each atom and qα is determined by a Boltzmann plot for each
element, after the temperature is determined by one of the main constituents, as intercept on
the y axis once the slope is fixed.

3.3 ELEMENT DISTRIBUTION: SPACE RESOLVED PROFILES AND

STRATIGRAPHY

Another great advantage of LIBS technique is its capability to perform space resolved
measurements on the target surface. Its resolution is only limited by the laser focusing
capability, typically in the micrometer range both in the x-y plane and along the z-axis.
Furthermore depth profiles can be obtained in a very limited destructive way (micro-
destructive indeed), when the laser ablating pulses are used to reach inner layers of the target.
This application is of special interest in the field of cultural heritage preservation and examples
will be given in section 4.
Stratigraphy is often used to assist surface cleaning processes, e.g. the mechanical or laser
removal of a deteriorated crust. When stratigraphy is performed by recording LIBS spectra,
vanishing of surface impurities is detected. Depth profile analysis of sample crusts can be
aimed to ascertain the laser cleaning effectiveness and to determine the right point where the
process must be interrupted. Quantitative profiles can be obtained for most elements detected
in encrustations, either as major components or traces [18].
In order to estimate the ablation rate during stratigraphy measurements, a thin layer of bulk
sample can be cut and the number of laser shots necessary to drill a hole in it can be measured.
Average ablation rates are useful when comparing data collected during different experiments
on the same kind of substrates (stones, metals, etc.).

9-18
3.4 UNDERWATER ELEMENT RECOGNITION AND ANALYSIS

The use of LIBS for analysis of submersed samples or liquids, mostly water solutions, still
suffers from severe limitations both in sensitivity and reproducibility, due to the nature of the
process itself. In fact plasma emission produced in bulk water is generally lower than the one
at the water-air interface, with a consequent drop of the analytical detection limit. This is due
to several factors, e.g., water absorption of the laser and plasma emission and their scattering
from suspended particles and micro-bubbles [21], radiation shielding by the high density
plasma [22] and fast quenching in the dense medium, even for the secondary plasma from dual
pulse excitation. The complex processes involving laser-induced breakdown in water are
extensively described in [23]. Experiments aimed to collect quantitative results in water must
be designed with a proper data handling suitable to overcome these limitations.

3.4.1 The filtering algorithm

The process of generation of LIBS signal in water has been carefully examined in order to
improve the achievable signal to noise ratio [24]. Strong plasma intensity shot-to-shot
oscillations occur both for water solutions and for immersed samples. For some laser pulses
LIBS emission is sometimes missing, even at high laser energy. Plasma generation in water is
in fact a statistic process, with a predefined probability of breakdown generation [25]. SNR can
be improved after identification and elimination of pulses not leading to the breakdown.
In order to take into account for process statistics a filtering algorithm has been developed,
which retains for averaging only sufficiently intense spectra, rejecting both those not showing

SNR=13
5 Mg
+ (a) Summed spectra
2.0x10

+
Mg
Intensity (counts)

5
1.0x10

4
4.0x10
SNR=58
(b) Filtering at 50% level

4
2.0x10
Mg

278 280 282 284 286 288

Wavelength (nm)

9-19
 Figure 16 - Comparison of DP-LIBS spectra obtained in a water solution after: (a) summing
over 1000 shots and (b) filtering at 50% of maximum peak; Mg concentration (from MgSO4) is
5 mg/l, laser pulse energies are E1=37 and E2= 156 mJ.

breakdown at all and weakest ones. In spectra acquired at a fixed monochromator position, the
discrimination factor for data filtering was the peak intensity P of a selected element line. The
reference intensity over a single measurement is the spectrum with the highest detected peak
intensity (Pmax). Filtering at 40% level corresponds to discarding all spectra with the peak
intensity P<0.4 Pmax. An example of the effectiveness of the filtering algorithm used in
analyzing a water reference solution is reported in Figure 16.

3.5 THE DOUBLE PULSE TECHNIQUE

Several efforts have been devoted to understand phenomena involved in the process of double
pulse plasma generation with LIBS. However a satisfactory theoretical description is not yet
available in air, and only semi-empirical models appear to work fine in water, where the
bubble confinement effect dominates. Experiments have been performed both in air and in
water with the main purpose to increase LIBS sensitivity, especially when trace detection is
attempted.

3.5.1 DP LIBS in air

Peculiar features are observed in DP excitation in air with respect to single pulse excitation at
the same total laser fluence [26]:
• Most intense spectra.
• Most intense ionic lines.
• Most intense and more persistent weak lines.
• Longer time decay of the line emission spectra.
• Lager LTE window for data acquisition.
• Narrower lines.
• Lowering of the ablation threshold.
• Higher ablation rates.
Some of the observed features, e.g. narrower lines, are consistent with the bubble
mechanism proposed for explaining the process in water. Experiments have been performed so
far mostly on metallic targets, and IR excitation has been used whenever the double pulse was
emitted by the same laser source.

3.5.2 DP LIBS in water

The DP technique has been developed for the detection of lines both from immersed targets
and from cations in water solutions, quantitative analysis have been achieved in both cases.

9-20
The bubble confinement improves the accuracy, and ensures a large enough LTE time window
with respect to single pulse excitation. Although non-equilibrium effects appear less
detrimental in DP LIBS in water, as a consequence of the effective plasma confinement in the
bubble, space non-homogeneities may give rise to peculiar perturbations related to the initial
plasma formation at the target surface and to its following expansion in water.
As an example, a section of the DP LIBS spectrum of a Titanium target immersed in
reproduced seawater is shown in Figure 17. Kinetic analysis shows the rich Ti emission
structure emerging from plasma breakdown at the surface combined with the Na absorption
doublet from colder atoms at the edge of the bubble.

5 1200 ns
4x10 1000 ns
800 ns
600 ns
200 ns
LIBS intensity (a.u.)

5
3x10

5
2x10

5
1x10

582 584 586 588 590 592

Wavelength (nm)

Figure 17 – A section of spectrum of Ti-laser induced plasma in seawater taken in the spectral
region of Na by DP LIBS at different delays from the second pulse (gate width 100 ns).

4 Applications to artwork

LIBS has found in the Cultural Heritage preservation a fertile field of application due to its
minimally invasive and only micro-destructive feature, the availability of portable instruments
and the possibility of stratigraphy with on-line monitoring cleaning processes. These
characteristics together with the option of combining LIBS analysis to other laser surface
diagnostics, such as Laser Induced Fluorescence (LIF) and Raman Spectroscopy makes the
technique suitable for qualitative and quantitative multi-elemental analyses on a large variety
of samples surfaces in most different stages of preservation. A number of examples are

9-21
discussed in the following.

4.1 QUALITATIVE ANALYSIS: FRAGILE SUBSTRATES

Qualitative analysis ancient substrates of completely unknown composition was attempted at

Histria (RO) museum during the CULTURE 2000 “Advanced On-Site Laboratory for
European Antique Heritage Restoration” (April 24-26, 2004).
In particular few results on classical ceramics from Greece and a prehistoric metallic found
are here presented. One of the ceramic samples studied is shown in Figure 18.

Figure 18 - fragment of a painted ceramic from classical age, probably of Greek origin.

As it is possible to notice, its surface is decorated with a black paint. Measurements have
been performed in order to analyze both painted surface and ceramic bulk and to highlight
their different composition. Results presented in Figure 18 have been obtained using the
experimental parameters: λlaser=355 nm, prr=20 Hz, energy=15mJ, number of spectra
accumulated=50, pulse width=6. In this figure significant parts of the recorded LIBS spectra
are shown. A contribution of the bulk to the pigment spectrum can not be neglected, because of
the small thickness of the paint layer. However, differences in relative intensities can give
information on the different composition of decoration and bulk. The analysis of the spectra, in
fact, shows how the ceramic is rich in Ca, Al, Ti, Mn, Si, Fe, V e Na and the black paint is
constituted by Fe and Al, probably as oxides.

+
paint
Ca ceramic Fe

Al
Intensity (a.u.)

4
1.8x106 8.0x10
Intensity (a.u.)

Mn
5 4
9.0x10 4.0x10
Ti
Ti

0.0 0.0
392 396 400 265 270

Wavelength (nm) Wavelength (nm)

9-22
 Figure18b - Comparison between signals from pigment and from ceramic bulk in two
significant spectral regions.

The metallic sample investigated is shown in Figure 19. It is completely covered by a thick
crusts. The LIBS analysis (see Figure 19b), performed with the same experimental parameters
listed above, has showed how the metallic bulk is constituted essentially of iron, while external
crusts are rich in Mg, Si, Ca, Al and Fe, i.e. constituted by corrosion products, silicates and
carbonates.

Figure 19 - fragment of a metallic prehistoric found. Probably a part of a weapon.

internal
external
Fe
Mg
4
1.2x10
Intensity (a.u.)

Si
+
Fe
3
6.0x10

0.0
284 288 292

Wavelength (nm)

4
9.0x10
+
Ca

Fe
Intensity (a.u.)

4
6.0x10 Al
Mn

Fe
4
3.0x10

396 400

Wavelength (nm)

9-23
 Figure 19b - Comparison between signals from bulk and from external surface of the ancient
metallic found in two significant spectral regions.

Underwater recognition of metals and marbles could be demonstrated as well [27]. An

example on precious alloys is shown in Figure 20

+
Ag
(a) Silver alloy
120000
Intensity (a.u.)

100000 Zn +
Ni
80000
Ni
60000
Ag
40000 Ag Mg
+
Ni
20000
0
270 275 280 285

10000 Au
+
(b) Gold alloy
Au
Intensity (a.u.)

8000 +
+ Ag +
6000 Cu +
Au
Au Cu Zn +
4000 Zn Au Au
Au
2000

0
270 275 280 285
Wavelength (nm)

Figure 20 - Recognition of precious alloys: (a) Silver alloy: atomic and ionic emission
from Nickel and Silver; (b) Gold alloy: detection of Gold (Au and Au+), Copper (Cu+), Zinc
atomic and Silver ionic emission; the Mg+ line at 279.50 nm comes from the seawater.

4.2 SEMI-QUANTITATIVE ANALYSIS: BRONZE ARCHEOLOGICAL FRAGMENTS

LIBS analysis has been carried out on samples identified as bronze alloys coming from
«Antiquarium of Canne della Battaglia», Barletta, Italy. Measurements showed the possibility
to obtain qualitative and semi-quantitative information on metallic samples of unknown
composition in a micro-destructive way.
Results obtained on a basin and on a belt indicate they are made of a binary bronze alloy,

9-24
e.g. Cu and Sn, but its composition was different from those usually manufactured by Romans
in that period. The surface exhibited corrosion products of the same alloy and ground deposits.
The third sample, a sword fragment, was found not to be made of bronze, as initially suggested
by archeologist, but iron. Elements detected indicate presence of soil and iron oxide in
encrustations. By means of calibration curves built for the Sn/Cu ratio it has been possible to

1 cm

Figure 21a – Belt fragments. Top: side cleaned for LIBS bulk measurements. The
enlargement shows a pre-cleaned spot where LIBS measurements have been carried out.
Bottom: side where the incrustation measurements have been performed.

9-25
 Figure 21b – Basin fragment completely covered by encrustations (top-left) Sword fragment
(bottom-right).

estimate the Sn and Cu percentage in the studied alloys with an error of 20%. The analysis on
bronze samples revealed a bulk composition of Cu=80% and Sn=15%, Fe appears as only
minor constituent (< 2%). On encrustations Ca, Al, Mg, Si, Na, C from residual soil were
found. The third sample resulted to be made of Iron, with impurities of Ca and Mg, and traces
of Al and Ti. (Figures 21). The large uncertainties in the composition allowed only a semi-
quantitative analysis.

4.3 QUANTITATIVE ANALYSIS AND STRATIGRAPHY: MARBLE

Quantitative analysis on white marble crust and bulk samples, coming from ancient quarries,
was based on calibration curves similar to those already shown in Figure 14. Due to difference
in temperature and electron density between the marble samples and the standards (CaCO3
pellets pressed from ultra pure powders doped with the addition of known quantities of
reference soils), calibration curves were corrected according to the procedure described in sect.
3.2.1 [18].
LIBS diagnostics was also successfully applied during the laser cleaning of marble [28]
and bronzes in air [29], automatic algorithms suitable to switch off the cleaning action as soon
as the original surface is reached are under development in several laboratories.
By the analytical method here developed, which takes into account differences in plasma
parameter between CaCO3 standards and marble based on calibration curves, the quantitative
marble and crust compositions were retrieved, and quantitative stratigraphy along the first

9-26
Figure 22 – Comparison of LIBS, SEM-EDX and ICP quantitative analysis on a White marble
sample: bulk (left) and crust (right).
layers involved in the cleaning process was possible [30]. The method was validated from the
comparison with SEM-EDX and ICP results obtained on the same samples. The accuracy
obtained by LIBS was comparable with the SEM-EDX one for major constituents (within
10%) and with ICP (within a factor 2), as shown in Figure 22. Note that with respect to EDX,
which is an instrument with portable versions, LIBS technique had a much better sensitivity
for trace detection. The sample non-homogeneity can be considered responsible for minor
discrepancies with ICP data, especially in trace detection.
After quantitative analysis of crust and bulk, LIBS stratigraphy has been performed on the
same sample in order to monitor the disappearance of surface impurities at increased depth
during the laser ablation process. Some significant results collected on surface contaminants
for two different Greek marble samples are reported in Figure 23. Here elemental
concentrations (% or ppm in weight) are plotted as a function of the number of laser shots,
which are in turn proportional to the reached depth.
5 3500
Naxos
3000
4 Paros
2500
Ti (ppm)

3 2000
Si (%)

1500
2
1000
1
500

0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
5 5000

4 4000
Fe (ppm)

3 3000
Al (%)

2 2000

1 1000

0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140

number of laser shots number of laser shots

Figure 23 – LIBS stratigraphy on marble samples from ancient Greek quarries. Laser ablation
and LIBS detection have been performed upon UV excitation at 355nm (ablation rate about 5
µm/shot).

The removal of surface impurities can be indirectly related to the sulfated layer ablation,
since it is well known that transition metals catalyze the transformation of marble into gypsum
upon the action of atmospheric pollutants (e.g. SO2). However to an effective monitoring of

9-27
sulfated crust, direct stratigraphy on Sulfur is advisable.
Difficulties in Sulfur detection, related to the existence of a few useful weak lines from S
and S+ in the usual LIBS range, have been recently overcome by means of DP LIBS in air. The
results obtained for a heavily encrusted (up to 10% S on the crust) white marble sample
coming from Veneto are shown in Figure 24, where data for Sulfur and Iron are reported as a
ratio to Calcium, the latter used as internal standard. Although Iron emission reaches stable
values after about the same number of applied shots as Sulfur, the correlation between Sulfur
and Iron emission is very weak, namely 0.58 for the reported example. This means that iron
might play a role in the surface sulfation, but its distribution in the crust is not an appropriate
indicator for Sulfur distribution.

1.0
+
S /Ca
Fe/Ca
0.8
Normalized Peak ratio

0.6

0.4

0.2

0.0

0 10 20 30 40 50
Shot number

Figure 24 – LIBS stratigraphy on an encrusted marble sample. Laser excitation at 1064 nm,
the number of shot refers to each pair of double pulses (ablation rate about 50 µm/shot).

4.4 QUANTITATIVE ANALYSIS: ANCIENT PAPER AND INKS

Conservation of historical paper documents is very important in preserving the evidence of

evolution of human history and thought. The support material knowledge as well as
identification of possible degradation processes is important in safeguarding this information
and in keeping it accessible to the future generations. Characterisation of inks and pigments as
well as of papers is often necessary to address the questions raised by archaeologists and
conservators. Historical inks differ considerably, and their identification is essential for proper
conservation. Inks produced from Fe II and oak galls were among the most widely used and
also the most corrosive inks. The chemical reactions based on iron-mediated oxidation of

9-28
cellulose, greatly reduce paper stability and may lead to disintegration of the inked areas.
Therefore, identification of these inks using fast and possibly non- or micro-destructive
portable instruments is essential.
LIBS has been applied to examine both paper and inks, Figure 25 shows an example of
LIBS high-resolution emission spectra for different paper samples. The elements: Ca, Fe, Ti,
Cu, Zn, Na can be easily identified. Quantitative analysis was performed on Fe and Cu after
building calibration curves from paper matrices containing a known amount of metals.
Accuracy of ≈ 15% was determined on Fe concentration. Minor elements found are: Mg and

In k _ A
4 In k _ B
Ca II

6 .0 x 1 0
Ca II

In k _ C
In k _ D
Fe I

Cu I
In k _ a v e

Zn I, Na I

Fe I
Fe I
LIBS Intensity (AU)

Fe I

Zn I
4
Fe I

4 .0 x 1 0

Ti II
Ti II, Ca I
Cu I
Ti II
Ti I

Fe I
Fe I
Ti II
Fe I
4
2 .0 x 1 0

315 320 325 330 335 340

W a v e le n g th (n m )

4 N o In k _ A
6 .0 x 1 0
Ca II

N o In k _ B
Ca II

N o In k _ C
N o In k _ D
4 N o In k _ a v e
Ti II

5 .0 x 1 0
LIBS Intensity (AU)

Ti II

Ti II
Ti II

4
4 .0 x 1 0
Fe I

Ti I
Cu I
Ti II

4
3 .0 x 1 0
Ti I

Cu I

Fe I
Zn I, Na I
Fe I
Fe I

4
Fe I

2 .0 x 1 0

4
1 .0 x 1 0
315 320 325 330 335 340
W a v e le n g th (n m )

Ni. Quantitative analysis of minor elements could not be carried out.

9-29
 Figure 25 - Inks on paper (top) and paper substrate spectra (bottom)

5 Applications to environment

LIBS was successfully applied in laboratory analyses of soils and sediments, in order to trace
the presence of heavy metals as pollutants [31] or to investigate sedimentation processes in
natural ecosystems (Ross Sea, Antarctica) [32]. To this respect main advantages of the
technique were the minimal sample preparation required (dried powders pressed in pellets
were usually examined), and the portability of the apparatus, which could be installed on board
of the oceanographic vessel.
In laboratory it was also recently demonstrated the possibility of asbestos recognition by
LIBS [33]. To this aim an algorithm based on a couple of LIBS intensity ratio (Fe/Si and
Mg/Si) as asbestos identifier has been developed.
Environmental applications however push towards developments of techniques for remote
in situ analyses. The laser beam directionality offers the possibility of in situ and semi-remote
LIBS applications for environment and planetary exploration. A few results of laboratory
experiments, preliminary to the design and realization of field systems, are reported in the
following.

5.1 ASBESTOS

In order to test the possibility of detecting and identifying asbestos by LIBS technique in
samples of different kind, as for example building materials, some real asbestos samples have

Mg Amosite
Chrysotile
LIBS intensity (a.u.)

Crocidolite

4
1.0x10 Si
Fe

Fe Fe

0.0
285 290 295 300 305 310
W avelength (nm)

9-30
 Figure 26 - LIBS emission spectra of three real asbestos: amosite (dot line), crocidolite (solid
line) and chrysotile (thin solid line).
been investigated. The LIBS spectra of amosite, crocidolite and chrysotile have been measured
in the wavelength range 280 nnmm-310 nm as shown in Figure 26.
The emission lines of Mg at λ = 285.17nm, Fe at λ = 293.70 nm and Si at λ = 288.15 nm
are easily recognized and, as expected, they are unaffected by overlapping and auto-
absorption. LIBS measurements performed on crocidolite, amosite and chrysotile confirmed
that also in the case of real asbestos, as just found for synthetic materials asbestos-like, the
Mg/Si and Fe/Si intensity ratios remain constant for each sample (respectively within ±0.19%
and ±0.17%) by repeating the measurement on different sample points. It can be considered as
a marker for asbestos recognition. The difference in elemental composition of different
asbestos can be determined by LIBS, with accuracy comparable to the one obtained in SEM-
EDX data. Concentration ratios of magnesium and iron on silicon can be deduced from LIBS
measurements on unknown samples, once calibration curves are built.

5.2 HEAVY ELEMENTS IN SOILS

Heavy elements research in soils has received more and more attention during the last decades
for the greatly increased circulation of hazardous elements in soil, water and air produced by
human activities. LIBS is a well-established analytical method for the determination of metals
in solids (Radziemsky et al., 1983), but it has been scarcely applied to soil analysis. For this
kind of samples the main advantage of LIBS is to simplify the conventional methodology by
avoiding time consuming chemical steps, e.g., the preparation and dissolution of the soil
sample.

5
6.0x10
+ compost
Sr standard
5
5.0x10 Fe
LIBS Intensity (a.u.)

5
4.0x10

3.0x10
5 Fe

2.0x10
5 Pb

5
1.0x10

0.0
404 406 408 410 412 414 416
wavelength (nm)

9-31
 Figure 27 - LIBS emission spectra of a compost sample (dots) and a standard sample (solid
line).
In Figure 27 LIBS emission spectra of a compost sample in the wavelength region 400-420
nm of a standard soil sample (NIST 2709) are compared.
Figure 28 shows the kinetics series for As determination in soils in DP experiments. The
section of the high resolution UV spectrum shows the most intense As line is resolved from the
nearby Iron ion emissions better and better as time delay between laser pulses increases (top).
Due to the non LTE conditions encountered at short delays, the time decay of the intensity of
Fe+ and As lines, could be fitted only by a 2nd order non linear fit (bottom).

Rock Sample 1511

Rock Sample 1511
Fe II

Integrated LIBS intensity - back ground

100000
width 25000 ns Fe+ @234.79 nm
d = 1 µs As @234.96 nm
d = 4 µs 100000
As I d = 5 µs
d = 9 µs
LIBS intensity

10000

10000

1000 0 2 4 6 8 10
234.6 234.8 235.0 235.2 235.4
Delay (microsec)
Wavelength (nm)

Figure 28 – Kinetic series for As determination in soils (Rock sample 1511) Experimental
conditions: excitation 1064 nm, total energy ≈ 100 mJ (shared 1:3 between 1st and 2nd pulse),
spot on the sample 60 µm2, coaxial optics, ½ m monochromator + ICCD detector (grating
3600 grooves/mm), Gate acquisition: delay from 2nd pulse variable, width 25 µs. Average on
10 shots.

5.3 PLANETARY EXPLORATION (ROCKS, DUST)

As a result of LIBS development in environmental analysis, a LIBS/Raman spectrometer is

planned on board of the Rover, which will land on Mars soil in the next unmanned mission.
ESA and NASA are running projects to realize the instrument. Preliminary laboratory work
has been performed to ascertain LIBS capabilities in analyzing Mars analogue rock samples in
the rarefied Mars atmosphere.
To this aim in our laboratory we optimized the LIBS signal generation/acquisition in
Martian conditions [34]. The equipment shown in Figure 29 was used to conduct experiments

9-32
in simulated Mars atmosphere, consisting of CO2 (95.3%), N2 (2.7%), Ar (1.6%) e O2 (0.13%),
at low- pressure
6.3 mbar, with
controlled substrate
temperature.

Figure 29 – LIBS laboratory set-up for experiments on Mars analogue rocks. From left to
right: delay generator, ICCD, laser source, monochromator, low-pressure chamber containing
the sample holder at controlled temperature (from 213 K to 283 K).

After optimization of the experimental conditions (laser parameters, data acquisition

window) LIBS spectra were analyzed automatically by using the calibration free (CF) method
(described in sect. 3.2.2). LIBS results obtained on three different andesite samples are
compared in Figure 30 with the atomic composition derived from SEM-EDX analysis
performed on the same samples. Results obtained by both techniques were mostly in
agreement, although the CF-LIBS accuracy resulted to be element and concentration
dependent, ranging from a few percent up to 20-30% for major constituents. Larger
discrepancies were obtained mostly for Ca, whose concentration was systematically
underestimated by LIBS due to saturation (this aspect is not yet included in the CF model
used). Matrix effects and sample space non-homogeneity could be responsible for other
discrepancies. In spite of limitations in accuracy, CF-LIBS seems to be a suitable method of
analysis for a system remotely operating at the pre-selected experimental parameters and
without any reference, and producing real time data, as it is the case of space missions.
Differently, an automatic Calibration Curve LIBS analysis can be carried out even in space,
providing standards on board and suitably built data set library.

9-33
 EDX
LIBS
10
8
6
4
2
0
Fe Si Mg Ca Ti Mn Al Na K Cu
Concentration (wt% )

10
8
6
4
2
0
Fe Si Mg Ca Ti Mn Al Na K Cu

10
8
6
4
2
0
Fe Si Mg Ca Ti Mn Al Na K Cu Ba

Element

Figure 30 - Major element concentration in andesite (silicate minerals). SEM-EDX and CF-
LIBS concentration comparison for three different samples. All data are % in weight

5.4 SUB-GLACIAL LAKE BOTTOM (ROCKS, SEDIMENTS, WATER)

Another challenging task is to perform analysis of environmental interest directly in water,

either on the water itself or solid materials at basin bottom, by using a compact set-up installed
on a submersible carrier (e.g. a ROV remotely operated vehicle). With the aim to design an
apparatus to be used for sub-glacial lakes exploration in Antarctica, preliminary laboratory
experiments were carried out, analyzing rocks and sediments immersed in water, the latter
after pressing in pellets to avoid the immediate dispersion of the sample.

9-34
 Underwater rock analysis was straightforward, as already performed for marble recognition
[27]. Due to the tendency of the sediments to dissolve in water upon laser irradiation, single
shot spectra were collected. Some examples generated in DP-LIBS experiments, aimed to the
qualitative (semi-quantitative) analysis of Antarctic marine sediments, are shown in Figure 31.
On the investigated sediments Ca, C, Mg, Mn, Si, Al, Fe, Ba, Ti, Na and K were detected in
the investigated spectral range (240-750 nm). Results are consistent with former analysis
conducted in air on the same dried sample [32].
Water analyses were attempted, as well, on standard solutions and calibration curves were
built for several elements (Mg, Mn, Cr). By applying the filter algorithm described in sect.
3.4.1, it was possible to reach sensitivity at the ppb level for the investigated metals dissolved
in water.

Si
5000 2000 +
+ Ti
Ca
4000 +
Fe
1500
Al
Intensity (a.u.)

Intensity (a.u.)

3000 +
Ti
+ 1000
Mn
2000
+ +
Fe Fe
C 500
1000

0 0
250 255 260 265 270 300 305 310 315 320 325

Wavelength (nm) Wavelength (nm)

Figure 31 – Selected spectral region of single shot DP-LIBS spectra of Antarctic sediment
collected during the XVI campaign and analysed in water.

6 Other analytical applications

6.1 TRACE ELEMENTS IN INDUSTRIAL PROCESSING

Applications very close to environmental analysis concern material contamination in industrial

processing. The capabilities of LIBS in monitoring main material constituents and impurities
have already been demonstrated, mostly for alloys [34], during industrial processing. The
possibility of a remote analysis of metals (up to 100 m distance), during or immediately after

9-35
solidification was also ascertained for steel [35]. The chance to reach even longer distances by
remote LIBS has already been demonstrated by use of femtosecond laser pulses [36], due to
the extreme self-focusing effects that may occur in air.
LIBS determination of material contamination from different elements in industrial
processing is of interest in view of in-situ, or even semi-remote applications. To this respect
we investigated the possibility to detect contamination in bricks used in industrial ovens.
Indeed the production of high quality glasses greatly relies on purity of raw materials as well
as lack of possible contamination. Phosphorous and Boron occasionally remaining in
refractory bricks may act as contaminant. To keep under control the production quality is
therefore necessary to have qualitative and quantitative determination by means of non-
invasive and fast analysis. As in case of Sulphur, LIBS detection of non metallic species is
rather difficult, therefore a feasibility study for phosphorous and boron determination was
carried out and a careful analysis was started to identify the more appropriate conditions for
detecting these elements at trace level. High resolution measurement in the spectral region
centered around 250 nm were performed on two different bricks A and O samples, additional
spectra (not shown) from reference sample (99.99% pure phosphorous salt) were also acquired
for spectral comparison and calibration.
The standard qualitative analysis on bricks revealed the presence of: Na, Al, Si, Ca, Fe, Zr,
Ti, Hf, Sc, Sr, Ba, Mg and B traces (the latter found only on sample A), while the quantitative
identification of K, B and P required special instrumental setting (wavelength range extension,
high resolution, proper time window choice). Mass content for P was 7000 ppm and 300 ppm
for samples A and O, respectively, while boron concentration was less than 100 ppm for
sample A and even lower for O. The quantitative analysis aimed at trace detection was
performed as already described in sect.3, while the large number of Zr I and Zr II emission
lines were used to derive the plasma temperature and electron density. Plasma parameters had
to be optimized to observe at least a suitable emission line from each non-metallic species.

5
1.0x10

A sample
Zr II

4
B sample
7.5x10
LIBS intensity (a.u.)

Zr II
BI

4
5.0x10
PI

4
2.5x10

0.0
249.4 249.6 249.8 253.2 253.4 253.6 253.8
Wavelenght (nm)

9-36
 Figure 32 – Sections of high-resolution LIBS spectra were boron and phosphorous traces
could be detected. Laser parameters: wavelength 355nm, energy 20mJ, pulse duration: 8ns,
repetition rate 10 Hz

5
Norm alized LIBS intensity (a.u.) 1.0x10
Zr II @ 253.51 nm
P I @ 253.56 nm

4
1.0x10

3
1.0x10

0 500 1000 1500 2000

Delay (ns)

Figure 33 – Kinetic analysis on P I and nearby Zr II lines.

Results reported in Figure 32 show the successful detection of boron and phosphorous
traces, achieved with specific acquisition time windows: delay 1.0 µs, width 1.0 µs for Boron
and delay 2.5 µs, width 10.0 µs for phosphor, respectively. Note that the LTE condition could
not be fulfilled in the latter case. As is shown in the selected spectral region there is strong
emission from Zr II, while P I emission appears just as a weak shoulder on a stronger ionic
line. Although LIBS line intensity increases with laser pulse energy, a simple rise of it cannot
be the correct way to enhance sensitivity. In fact, signal intensity's increase will also affects the
optical emissions of all the elements. What we get in this case is just a magnified version of the
spectrum, without any significant improvement. However since the interfering line is an ionic
one, it was checked if the kinetics of lines emission for ions different from those of atomic
elements. It is known, indeed, that sulphur optical emissions come later with respect to that of
other elements, and similar situation could also occurs in case of P. Figure 33 shows the
kinetics of P I and Zr II lines. It may be noticed that the atomic line has a time constant
definitively longer than the ionic one, presenting a possible way to reduce the ionic influence
by time discrimination. The observed difference in atomic and ionic time constants makes
possible the use of time discrimination to decrease the detection threshold for phosphorous,
thus increasing the sensitivity limit of the experimental apparatus. Using delays longer than 5
µs will cause stronger decrease of interfering ions emissions as well as reduction of overlap
due to further narrowing of emission lines. However still longer delay produce a reduction of
total LIBS signal, which might cause a significant worsening of the signal to noise ratio.

9-37
6.2 EXPLOSIVES

Explosive substances can be characterized by LIBS through their C-H-N-O ratios [38-39], and
in particular, they have a large N content. One of the most important parameters for their
recognition through the atomic content is N/O ratio. In LIBS spectra of organic compounds
and explosives, emission from atomic carbon and its molecule C2 can be also detected. In
presence of carbon in the sample, and of nitrogen in the same sample or in the environment
(example from air), the emission from molecular fragment CN also appears. Fragment CN is
formed from the reaction:

N2 + C2 → 2 CN

Consequently, CN emission is proportional to the presence of C2 molecules in the plasma.

These molecules can be formed both by recombination of the atoms and by complex
fragmentation processes in the case of organic materials. In [40] it was observed that there is a
quasi-proportional relationship between the number of double-bonded carbon in different
compounds and LIBS intensity of C2 emission, therefore of CN emission. The latter is also
influenced by N2 presence in the plasma. As some explosives contain double-bonded carbon,
the ratios of its atomic and molecular emissions (C2 and CN) are another parameters that might
help identification of the explosive material. In Figure 34, the characteristic emission lines for
recognition of the explosives are shown, although the spectra were recorded on one coal
sample (NIST 1632 b), which contains N (1.56%), C (78%), H (5.1%) and O (not certified
value). Here, the laser used was Nd:YAG, 1064 nm, pulse duration 8 ns, laser energy 40 mJ,
fluence on the sample was 28 J/cm2. The detector integration time was 10 µs, gate delay with
respect to the laser pulse was 500 ns.
As it can be observed, the spectral range for the detection of C, N, H and O in the sample
should cover 240 - 780 nm. The spectral resolution should be sufficiently high (here 0.12-
0.13), because in the case of complex matrices, such as the coal sample, here analyzed, the
spectral interferences might occur, particularly close to the emission lines of C (247.8 nm), CN
(close to 385 nm) and C2 (Swan band around 515 nm).
The line intensity ratios in the LIBS plasma are much influenced by the chosen acquisition
delay time and gate window. For example, we measured temporal behavior of different lines of
coal sample, by applying short gate widths (100-1000 ns). While the atomic lines appear
shortly after the plasma ignition and reach their maximum after about 500 ns, C2 emission
becomes detectable later, after about 1000 ns, CN fragment only appears in the spectrum after
about 500 ns. The measured emission decay is much faster for atoms than for molecules (or
fragments), and from the exponential decay the time constants are in the range from 1.0-1.5 µs
(an example is reported in Figure 35). The analogue decay constant for C2 molecule is about
12 µs (Figure 35). In Figures 35-36 we considered SNR rather than the line intensity, and we
applied normalization on the gate width, as it was extended for longer delays in order to obtain

9-38
detectable signal from the decaying plasma.

30000
4000 C
C2 Aluminium
3500 25000 Coal
3000
20000
2500
+
Fe 15000
2000

1500 10000

1000 Si
5000
500

0 0
240 245 250 255 260 265 270 505 510 515 520 525

CN
LIBS Intensity (a.u.)

O
40000 10000
+
Ca Ca
+

8000
30000

Al 6000

20000
4000

10000
2000

0 0
380 385 390 395 400 765 770 775 780 785

N
14000 H 7000

12000 6000

10000 5000 N

8000 4000
N
6000 3000

4000 2000

2000 1000

0 0
650 655 660 665 670 730 735 740 745 750

W avelength (nm)

Figure 34 - The most intense emission lines relative to C, H, N and O detection; samples are

9-39
 coal (solid) and Aluminium (dot) – used for comparison

350

300
Exp. Decay
250
SNR/gate

2
R = 0.999
200
t = 1480 ns
150

100

50

0
0 1000 2000 3000 4000 5000 6000 7000 8000
Central gate delay (ns)

Figure 35 - SNR normalized on gate width (microseconds) as a function of the central gate
delay, for H line at 656 nm, sample coal

80

70 Exp. Decay
60 2
R = 0.967
50 t = 12700 ns
SNR/gate

40

30

20

10

0
1000 10000
Central gate delay (ns)

Figure 36- SNR normalized on gate width (in microseconds) as a function of the central gate
delay, for C2 line at 515 nm, sample coal

9-40
7. Conclusions

LIBS technique is nowadays well developed as far as the understanding of the laser matter
interaction processes involved in the plasma generation and evolution are concerned, at least
when a single laser pulse is applied. This allows supporting its analytical applications in a wide
range of problems where the advantages of a minimal and easy sampling are important and a
fast multi-elemental analysis is required, even if its sensitivity and accuracy cannot compete
with other assessed laboratory analysis tools.
The possibilities of in situ and remote application of the LIBS technique are further
supporting the development of automatic field systems suitable to operate without human
control also in harsh environments.

Acknowledgments: The authors wish to thank Prof M. Strlic and Dr J.Koller, Ljubljana,
Slovenia, for valuable works on ancient papers and inks

References

1. Radziemski, L.J., Loree, T.R. (1981) Laser-induced breakdown spectroscopy: time-

resolved applications, Plasma Chem. Plasma Proc., 1, 281-293.
2. Amoruso, S., Bruzzese, R., Spinelli, N., Velotta R., (1999) Characterization of laser-
ablation plasmas, J. Phys. B: At., Mol. Opt. Phys., 32, 131-172.
3. Russo, R.E., Mao, X., Liu, H., Gonzalez, J., Mao, S.S. (2002) Laser ablation in analytical
chemistry—a review, Talanta 57, 425–451.
4. Griem, H.R. (1964) Plasma Spectroscopy , McGraw-Hill.
5. Dolgaev, S.I., Simakin, A.V., Voronov, V.V., Shafeev, G.A., Bozon-Verduraz, F. (2002)
Nanoparticles produced by laser ablation of solids in liquid environment, Appl. Surf. Sci.
186 546-551.
6. Sattmann, R., Sturm, V., Noll, R. (1995), Laser induced breakdown spectroscopy of steel
samples using multiple Q-switch Nd:Yag laser pulses, J.Phys D: Appl.Phys 28, 2181-2187.
7. De Giacomo, A., Dell’Aglio, M., Colao, F., Fantoni, R. (2004) Double pulse laser
produced plasma on metallic target in seawater: basic aspects and analytical approach,
Spectrochim. Acta, part B 59, 1431-1438.
8. Dahmani, F. (1992), Experimental study of laser and wavelength dependences of laser–
plasma coupling, transport, and ablation processes in planar gold targets, Phys. Fluids B:
Plasma Phys. 4, 1943-1952.
9. Lee, Y-I., Song, K., Cha, H-K., Lee, J-M., Park, M-C., Lee, G-H., Sneddon, J. (1997)

9-41
 Influence of Atmosphere and Irradiation Wavelength on Copper Plasma Emission Induced
by Excimer and Q-switched Nd:YAG Laser Ablation, Appl. Spectrosc. 51, 959-964.
10. Cristoforetti, G., Legnaioli ,S., Palleschi ,V., Solvetti, A., Tognoni, E. (2004) Influence of
ambient gas pressure on laser induced breakdown spectroscopy technique in the parallel
double-pulse configuration, Spectrochim Acta, Part B 59, 1907-1917.
11. Conrads, H., (1973) Diagnostics of high temperature high density plasma by radiation
analysis, in E.G Brame (ed), Applied Spectroscopy Reviews, M. Dekker Inc, New York,
Vol. 6, pp. 135-188.
12. Griem, H. R. (1974) Spectral Line Broadening by Plasmas, London, Academic Press.
13. Pipemeier, E.H. (1986) Laser ablation for atomic spectroscopy, in E.H.Pipemeier (ed),
Analytical Applications of Lasers, John Wiley and Sons, New York, pp. 627-669.
14. Colao, F. , Fantoni, R., Lazic, V. (2002) LIPS as an analytical technique in non-equilibrium
plasma, in V. Ochkin (ed), SPIE Special Issue on Spectroscopy of non equilibrium plasma
at elevated pressures, SPIE, Bellinghan, Vol 4460, pp.339-348.
15. St-Onge, L., Sabsabi, M. (2000), Towards quantitative depth-profile analysis using LIPS:
investigation of galvannealed coatings on steel, Spectrochim. Acta, Part B 55, 299.
16. NIST electronic database, available at http:\\physlab.nist.gov\physrefdata\contents-
atomic.html; Kurucz database, available at http:\\cfa-
www.harvard.edu\amdata\ampdata\kurucz23\sekur.html
17. Lazic, V., Barbini, R., Colao, F., Fantoni, R., Palucci, A. (2001), Self-absorption model in
quantitative laser induced breakdown spectroscopy measurements on soils and sediments,
Spectrochim. Acta, Part B 56, 807-820.
18. Lazic, V., Fantoni, R., Colao, F., Santagata, A., Morone, A., Spizzichino, V. (2004),
Quantitative laser induced breakdown spectroscopy analysis of ancient marbles and
corrections for the variability of plasma parameters and of ablation rate, J. Anal. At.
Spectrom. 19, 429-436.Fornarini, L., Colao, F., Fantoni, R., Lazic, V., Spizzichino, V.
(2005), Calibration Analysis of bronze samples by nanosecond Laser Induced breakdown
spectroscopy: A theoretical and experimental approach, Spectrochim. Acta, Part B 60, 1186
– 1201.
20. Corsi, M., Cristoforetti, G., Palleschi, V., Salvetti, A., Tognoni, E. (2001), A fast and
accurate method for the determination of precious alloys caratage by LIPS, Eur. Phys., J. D
13, 373-377.
21. Kompitsas, M., Roubani-Kalantzopoulou, F., Bassiotis, I., Diamantopoulou, A.,
Giannoudakos, A. (2000), LIPS as an efficient method for elemental analysis of
environmental samples, EARSeL Proceedings of 4th Workshop on lidar remote sensing of
Land and Sea. Dresden June 16-17, 2000. EARSeL e-Proceedings http://las.physik.uni-
oldenburg.de/, vol. 1, pp 130-138.
22. Bunkin, N. F., Lobeyev, A. V. (1997) Influence of dissolved gas on optical breakdown and
small-angle scattering of light in liquids, Phys. Lett. A 229, 327-333.
23. Hammer, D. X., Jansen, E. D., Frenz, M., Noojin, G. D., Thomas, R. J, Noack, J., Vogel,
A., Rockwell, B. A., Welch, A. J. (1997) Shielding properties of laser induced breakdown
in water for pulse durations from 5 ns to 125 fs, Appl. Opt. 36, 5630-5640.
24. Lazic, V., Colao, F., Fantoni, R., Spizzichino, V. (2005), Laser Induced Breakdown

9-42
 Spectroscopy in water: improvement of the detection threshold by signal processing,
Spectrochim. Acta, Part B 60, 1002 – 1013.
25. Kennedy, P. K., Hammer, D. X., Rockwell, B. A. (1997) Laser-induced breakdown in
aqueous media, Prog. Quantum Electron. 21, 155-248.
26. Colao, F., Lazic V., Fantoni, R., Pershin S. (2002) A comparison of single and double pulse
laser-induced breakdown spectroscopy of aluminum samples, Spectrochim. Acta, Part B
57, 1167-1179.
27. Lazic, V., Colao, F., Fantoni, R., Spizzichino, V. (2005). Recognition of archaeological
materials under water by Laser Induced Breakdown Spectroscopy, Spectrochim. Acta Part
B 60, 1014 – 1024
28. Maravelaki-Kalaitzaki, P., Anglos, D., Kilikoglou, V., Zafiropulos, V. (2001) Composition
characterization of encrustation on marble with laser induced breakdown spectroscopy,
Spectrochim. Acta, Part B 56 887-903.
29. Colao, F., Fantoni, R., Lazic, V., Caneve, L., Giardini, A., Spizzichino V. (2004) LIBS as a
diagnostic tool during the laser cleaning of copper based alloys: experimental results, J.
Anal. At. Spectrom., 19, 502 – 504.
30. Colao, F., Fantoni, R., Lazic, V., Morone, A., Santagata, A., Giardini, A. (2004). LIBS
used as a diagnostic tool during the laser cleaning of ancient marble from Mediterranean
areas Appl. Phys. A 79, 213-219
31. Capitelli, F., Colao, F., Provenzano, M.R., Fantoni, R., Brunetti, G., Senesi N. (2002)
Determination of heavy metals in soils by laser induced breakdown spectroscopy,
Goederma, 106, 46 – 62.
32. Barbini, R., Colao, F., Lazic, V., Fantoni, R., Palucci, A., Angelone M. (2002) On board
LIBS analysis of marine sediments collected during the XVI Italian campaign in
Antarctica, Spectrochim. Acta, Part B, 57, 1203-1218.
33. Caneve, L., Colao, F., Fabbri, F., Fantoni, R., Spizzichino, V., Striber, J. (2005), LIBS
Analysis of asbestos, Spectrochimica Acta, Part B 60 1115-1120.
34. Colao, F., Fantoni, R., Lazic, V., Paolini, A. (2004) LIBS application for analyses of
Martian crust analogues: search for the optimal experimental parameters in air and CO2
atmosphere, Appl. Phys. A 79, 143-152.
35. Lopez-Moreno, C., Amponsah-Manager, K., Smith, B. W., Gornushkin, I. B., Omenetto,
N., Palanco, S., Laserna, J. J., Winefordner, J. D. (2005) Quantitative analysis of low-alloy
steel by microchip laser induced breakdown spectroscopy, J. Anal. At. Spectrom.20,552 –
556.
36. Palanco, S., Conesa, S., Laserna, J. J. (2004), Analytical control of liquid steel in an
induction melting furnace using a remote laser induced plasma spectrometer, J. Anal. At.
Spectrom. 19, 462-467.
37. Stelmaszczyk, K., Rohwetter, P., Méjean, G., Yu, J., Salmon, E., Kasparian, J.,
Ackermann, R., Wolf, J.-P., Wöste, L. (2004) Long-distance remote laser-induced
breakdown spectroscopy using filamentation in air, Appl. Phys. Lett. 85, 3977- 3979.
38. De Lucia Jr., F. C., Harmon, R. S., McNesby, K. L., Winkel Jr., R. J., Miziolek, A. W.
(2003) Laser-induced breakdown spectroscopy analysis of energetic materials, Appl. Opt.
42 (30) 6146-6152.

9-43
 39. Lo´pez-Moreno, C., Palanco, S., Laserna, J. J., De Lucia Jr, F., Miziolek, A. W., Rose, J.,
Walters, R. A., Whitehouse, A. I. (2006) Test of a stand-off laser-induced breakdown
spectroscopy sensor for the detection of explosive residues on solid surfaces, J. Anal. Atom.
Spectrom. 21 55-60.
40. St-Onge, L., Sing, R., Bechard, S., Sabsabi, M. (1999) Carbon emissions following 1.064
µm laser ablation of graphite and organic samples in ambient air, Appl. Phys. A 69 S913–
S916

9-44

View publication stats