Optica Applicata, Vol. XXXVII, No.

12, 2007

Calibration free laser-induced breakdown

spectroscopy (LIBS) identification
of seawater salinity
WALID TAWFIK Y. MOHAMED1, 2
1

National Institute of Laser Enhanced Science (NILES), Dept. of Environmental applications,

Cairo University, Cairo, Egypt; e-mail: Walid_Tawfik@hotmail.com

Faculty of Education for girls, Department of Physics, Gurayyat, North of Al-gouf,

Kingdom of Saudi Arabia
Laser-induced breakdown spectroscopy (LIBS) has been used as a remote sensing system to
analyze seawater samples and to identify their salinities without ordinary calibration curves.
The plasma is generated by focusing a pulsed Nd : YAG laser on the seawater surface in air at
atmospheric pressure. Such plasma emission spectrum was collected using wide band fused-silica
optical fiber of one-meter length connected to a portable Echelle spectrometer (Mechelle 7500
Multichannel Instruments, Stockholm, Sweden) with intensified CCD camera. Spectroscopic
analysis of plasma evolution of laser produced plasmas has been characterized in terms of their
spectra, electron density and electron temperature assuming the local thermodynamic equilibrium
(LTE) and optically thin plasma conditions. Three elements Na, Ca and Mg were determined in
the obtained spectra to identify the salinity of seawater samples. The electron temperature Te
and density Ne were determined using the emission intensity and Stark broadening. The obtained
values of Te and Ne for natural seawater sample (salinity 3.753%) are 11580 K 0.35% and
3.331018 cm3 14.3%. These values exhibit a significant change only if the matrix changes (i.e.,
the salinity changes). On the other hand, no significant difference was obtained if Te and Ne were
determined using any of the three elements (Na, Ca and Mg) in the same matrix. It is concluded
that Te and Ne represent a fingerprint plasma characterization for a given seawater sample and
its salinity could be identified using only one element without need to analyze the rest of elements
in the seawater matrix. The obtained results indicate that it is possible to improve the exploitation
of LIBS in the remote on-line environmental monitoring, by following up only a single element
as a marker to identify the seawater matrix composition and salinity without need to analyze that
matrix which saves a lot of time and efforts.

Keywords: laser-induced breakdown spectroscopy (LIBS), matrix effect, remote sensing, plasma
parameters, seawater.

1. Introduction
Although a vast majority of seawater is found in oceans with salinity around 3.5%,
seawater is not uniformly saline throughout the world. Most of the saline seawater is

W. TAWFIK Y. MOHAMED

due to sodium chloride in addition to calcium and magnesium and some other trace
elements like sulfur, potassium, bromine, iodine and carbon. Salinity measurement is
very important for many environmental fields, such as marine environment monitoring
and quality monitoring of water resources. The measurement of salinity is a complex
subject that has a long history. Very briefly, it can be concluded that, in the past, the
salinity in seawater was determined by hydrometric methods. In recent years, several
methods and technologies have been proposed for salinity measurement [15]. For
example, an ultrasonic technique [1] based on measurement of the travel-time of light
was presented to measure the salinity in solar pond in 1995. A chemical method [2]
based on a polyaniline matrix coated wire electrodes was developed for salinity
measurement within a range of 0.010% to 75%. The salinity in water was obtained by
measuring the potential difference between the coated wire electrode and an Ag /AgCl
saturated KCl reference electrode. MINATO et al. [3] proposed a fiber-optic system for
remote salinity measurement of seawater, which indicated a salinity measurement
resolution of 0.02%. ZHAO et al. [4, 5] proposed a novel method, which can
simultaneously measure the salinity and temperature in water based on optical
refraction method. They used a He-Ne laser at 632.8 nm and a position-sensitive
detector (PSD) to monitor the beam deviation caused by refraction due to the
salinity of measured water at the receiving end face of a measurement cell.
The obtained experimental results indicate the feasibility of the developed system with
a measurement resolution of 0.012%. In 2006, the author [6] proposed a novel
technique, which could simultaneously measure the concentration of Ca, Na, and Mg
in seawater. The used technique was laser-induced breakdown spectroscopy (LIBS)
that offers unique capabilities for on-line composition determination without sample
preparation. In that method, the elemental calibration curves were constructed using
the elemental concentration versus its spectral intensity. Moreover, in LIBS, a small
volume of the target is intensely heated by the focused beam of a pulsed laser, and thus
brought to a transient plasma state where the samples components are essentially
reduced to individual atoms. In this high-temperature plasma, atoms are ionized, or
brought to excited states. Such states decay by emission of radiation, which is observed
in the ultraviolet (UV), visible and near-infrared (NIR) regions of the spectrum. An
atomic spectrum is examined by means of a spectrograph, thereby allowing elemental
components of the target to be identified and, quantified using a calibration curve.
LIBS measurements are generally carried out in ambient air at atmospheric pressure.
For this reason, and also due to its rapidity, non-contact optical nature, and absence of
sample preparation, since the only requirement is the optical access to the samples,
LIBS is a useful technique for the on-line process analysis. The basic features of this
spectroscopic technique and its applications for the on-line measurements have been
reviewed in several papers [616].
In this paper, LIBS technique will be used as a remote sensor system to measure
the salinity of seawater by determining three main elements in saline seawater Na, Ca
and Mg without ordinary calibration curves. In the presented sensor system, the plasma

Calibration free LIBS identification of seawater salinity

characterization parameters are used to identify the seawater salinity. In doing so,
spectroscopic analysis of plasma evolution of laser produced plasmas has been
characterized in terms of their spectra, electron density and electron temperature
assuming the local thermodynamic equilibrium (LTE) and optically thin plasma
conditions.

2. Methodology
A typical LIBS experimental setup, described in details elsewhere [6], is used
throughout the present investigations. The plasma formation was attained with the aid
of a Q-switched Nd:YAG laser (Continuum NY81.30, USA) delivering laser pulses
of 300 mJ/pulse, with 7 ns pulse duration at its fundamental wavelength 1064 nm
with adjustable repetition rate up to 30 Hz. The laser pulse energy was adjusted by
a suitable combination of neutral filters at constant operating high voltage (1.3 kV)
and Q-switch delay (1.5 s) to ensure spatial and temporal beam profile stability.
An energy meter (Nova 978, Ophir Optronics Ltd., USA) was employed to monitor
the shot-to-shot pulse energy. Only 180 mJ of the laser pulse energy was focused
onto the seawater surface via a quartz plano-convex lens of 100 mm focal length.
The emitted light from the plasma plume is collected via a one-meter length wide
band fused-silica optical fiber connected to a 0.17 m focal length Echelle spectrometer
(Mechelle 7500, Multichannel Instruments, Sweden). The Mechelle 7500 provides
a constant spectral resolution of 7500 corresponding to 4 pixels FWHM, over
a wavelength range 2001000 nm displayable in a single spectrum. A gateable,
intensified CCD camera, (DiCAM-Pro, PCO Computer Optics, Germany) coupled to
the spectrometer was used for detection of the dispersed light. The overall linear
dispersion of the spectrometer-camera system ranges from 0.006 nm /pixel (at 200 nm)
to 0.033 nm /pixel (at 1000 nm). To avoid the electronic interference and jitters, the
CCD intensifier high voltage was triggered optically. The ICCD camera control was
performed via Mechelle software (Multichannel instruments, Stockholm, Sweden).
The emission spectra display, processing and analysis were done using 2D- and
3D- GRAMS/32 version 5.1 spectroscopic data analysis software (Galactic Industries,
Salem, NH, USA).
To improve LIBS precision, spectra from several laser shots have to be averaged
in order to reduce statistical error due to laser shot-to-shot fluctuation. As found before
by the author [6], all the subsequent measurements must carried out by accumulating
fifty spectra for each seawater LIBS spectrum. Moreover, it is found that the optimum
delay time for the ICCD Camera is 2500 ns and the gate delay is 10 s with respect to
the laser pulse starting time. Fifty laser shots were fired on the sample surface and
the average was computed and saved to serve as the spectrum library. For each recorded
spectrum, the peak intensity, the Gaussian curve fitting, the full width at half maximum
(FWHM), and the center wavelength of each line, as well as the background emission
continuum were determined. Data treatment preprocessing of the averaged spectra was

W. TAWFIK Y. MOHAMED

T a b l e 1. Reference sodium, magnesium and calcium values in ppm for the seawater samples.
Element
Na
Mg
Ca

Sample-1
(salinity 3.753%)
11580
5238
772.8

Sample-2
(salinity 1.876%)
5790
2619
386.4

Sample-3
(salinity 0.938%)
2895
1309.5
193.2

performed in the Microsoft Windows XP environment on a Pentium 4 PC using

GRAMS/32, Excel (Microsoft office Excel 2003) and Origin software version 7.0220
(Origin Lab corporation, USA). The averages of peak tables (lists of wavelengths and
intensities) of the averaged spectra were roll generated in GRAMS/32 and exported
for data evaluation.
Using a classical LIBS configuration, however, as used for LIBS on solids with
the laser beam perpendicular to the surface leads to splashing in the case of liquids.
The splashing results in covering the focusing optics with droplets and, therefore,
prevents further use of this technique. This can easily be explained by the fact that
the plasma expansion at atmospheric pressure is directed perpendicularly to the surface
[17]. A tilted configuration can minimize this phenomenon. Thus, the Nd:YAG laser
beam was focused onto the water surface at 45 angle. This was done using a 25 mm
diameter dichroic mirror that reflects 99% of high energy 1064 nm wavelength at angle
45. Moreover, a low laser pulse repetition rate of 0.2 Hz was used to get rid of any
shockwaves produced ripples on the water surface and minimize water splashing too.
On the other hand, the non-contact remote measurements are simulated by using
the 100 mm focusing lens and the one meter fused-silica optical fiber (for long range
measurements a 3 m focusing lens and a 3 m fused-silica optical fiber could be used
instead).
The seawater sample was taken from the Mediterranean sea in the north of Egypt.
Then the natural seawater sample was diluted many times with doubly distilled water
to obtain a range of salinity. Three samples were selected for measurements and
namely: sample-1, sample-2 and sample-3 for natural seawater sample and two diluted
seawater samples with dilution ratios (1:1) and (1:3), and salinity values of 3.753%,
1.876% and 0.938%, respectively. The salinities and concentrations of Mg, Ca and Na
in these samples were performed on a Varian Spectra AA-220 FS atomic absorption
spectrometer (Varian, Australia) and the corresponding values are listed in Tab. 1.

3. Results and discussion

3.1. LIBS spectrum
Figure 1 shows typical LIBS spectra for natural seawater. The three spectra in the
figure are plasma emission spectra recorded at 2.5 s delay time and 10 s gate width.
The panoramic Echelle spectra in the spectral range 250700 nm made it possible to
observe the UV-emission of Mg-lines at 279.55, 280.27 and 285.21 nm, the Ca-lines

Calibration free LIBS identification of seawater salinity

300000

Mg
279.55
279.55
nmnm

Mg

Intensity [arb. u.]

250000
Mg
280.27
280.27
nmnm

200000
150000
100000
50000

Mg 285.21
285.21 nm
nm
0
278

279

280

281

282

283

284

285

286

Wavelength [nm]

140000

Ca

Ca
393.37
393.37
nmnm

Intensity [arb. u.]

120000
100000
Ca
396.85
396.85
nmnm

80000
60000
40000
20000
0
390

Ca 422.67
422.67nm
nm
395

400

405

410

415

420

425

430

Wavelength [nm]

140000

Na 588.99
588.99nm
nm

Na

Intensity [arb. u.]

120000
100000

Na
589.59
589.59
nmnm

80000
60000
40000
20000
0
585

587

589

591

593

595

Wavelength [nm]

Fig. 1. Typical LIBS spectra for natural seawater sample in the spectral range 278600 nm for Mg, Ca,
and Na, respectively. The laser energy was 180 mJ at wavelength 1064 nm, plasma emissions are
accumulated with delay 2.5 s, and gate width 1 s.

at 393.37, 396.85 and 422.67 nm, and the well resolved visible Na-lines at 588.99 and
589.59 nm at the same time. The spectra reflect the wide spectral range and the high
resolution of the used spectroscopic system.
3.2. Plasma characterization measurements and matrix effect
One of the main problems in the use of LIBS is the necessity of making a calibration
curve with samples possessing the same matrix composition as that of the samples to
be analyzed. In 1998, BULATOV et al. [18] proposed a method in which the composition

10

W. TAWFIK Y. MOHAMED

of a sample could be determined without the need of calibration curves. They suggested
that the matrix effects are due to the higher emission from easily ionized elements
existing in the matrix, which contribute to the first stages of plasma formation.
However, this method is based on the measurement of the emission from all the species
present in the sample, a requirement difficult to satisfy when dealing with in situ
measurements of natural seawater samples. Understanding of the matrix effect is
important to maximize LIBS analytical performance and to determine the technique
limitations. The matrix effect can results in the sample being ablated differently from
the target sample. The interaction between the laser and the target in LIBS is influenced
significantly by the overall composition of the target, so that the intensity of the
emission lines observed is a function of both the concentration of the elements of
interest and the properties of the matrix that contains them. Plasma composition is
dependent not only on the composition of the sample, but also on laser parameters,
sample surface conditions as well as on thermal and optical properties of the sample.
Previously published works studied the matrix effect under different experimental
conditions to specify causes and find out the methods of correction [11, 1824].
The different approaches have been undertaken to discriminate the problems
resulting from the fractionation of the ablation and matrix effects. The most convenient
approach is to determine elemental abundance comparing the analyzed line intensities
with signals obtained from the proper reference standards having the similar matrix
composition. But, it is not always possible to obtain such calibration curves because
there are no available standard samples, or it is impossible to have an internal standard
of known concentration. In addition, plasma formation dynamics, sample ablation and
associated processes are highly non-linear and not fully understood and may also play
an important role as reasons of the matrix effect. Thus, in spite of many advantages of
LIBS the realization of a quantitative analytical method, which is able to measure main
constituents in samples from different matrices, still remains a difficult task because
of the complex laser-sample and laser-plasma interaction mechanisms. As a rule, laser
ablation plasma is characterized by complex spatial and temporal structures, and one
meets a wide range of varying of parameters during the plasma existence time [24].
Below, we will study the dependence of the plasma parameters (electron density,
electron temperature) on the matrix composition of the seawater samples. Because if
a relation could be found, one can use it to distinguish between different seawater
matrices.
3.2.1. Electron temperature and matrix effect

In LIBS experiments, after the initial plasma decay and during the entire observation
interval, the local thermodynamic equilibrium (LTE) conditions are assumed to hold.
For optically thin plasma, the re-absorption effects of plasma emission are
negligible. So, the emitted spectral line intensity I is a measure of the population of
the corresponding energy level of this element in the plasma. For the LTE plasma,

Calibration free LIBS identification of seawater salinity

11

the population of an excited level can be related to the total density N(T ) of neutral
atom or ion of this element by Boltzmann equation as:
A ki g k
Ek
hc
I = ------------- N ( T ) ----------------- exp ----------
4
U(T )
KT

(1)

where is the wavelength, Aki is the transition probability, gk is the statistical weight
for the upper level, Ek is the excited level energy, T is the temperature (in LTE all
temperatures are assumed to be equal, i.e., Te Tion Tplasma ), K is the Boltzmann
constants, U(T ) is the partition function [25].
The emitted spectral line intensity from a given state of excitation can be used to
evaluate the plasma temperature. The lines must be well resolved in order to accurately

ln(I/gA)

-18

-18

-20

-20

-22

-22

-24

-24

-26

-26

-28

ln( I/gA)

40000

60000

80000

-10
-15
-20
-25
-30

Ca for sample-1

-16
-18

-20

-20

-22

-22

-24

-24

-26

-26

Mg for sample-2

-28

45000

65000

60000

80000

-16

-18

-18

-20

-20

-22

-22

-24

-24

Mg for sample-3

-28

60000

Energy [cm 1 ]

80000

400000

-5

-15
-20
-25
-30

Ca for sample-2

Na for sample-2

-35

45000

65000

85000

200000

300000

400000

0
-5
-10
-15
-20
-25

Ca for sample-3

-30

-30
40000

300000

-10

-26

-26

200000

25000

-16

-30
20000

85000

-30
40000

Na for sample-1

-35

25000

-18

-28

-5

-30

-16

-30
20000

ln ( I/gA)

-28

Mg for sample-1

-30
20000

-28

-16

-16

25000

45000

65000

Energy [cm 1 ]

Na for sample-3

-35
85000 200000

300000

400000

Energy [cm 1 ]

Fig. 2. Nine Boltzmann plots were determined form the emission line intensities of Mg, Ca and Na
observed in the laser-induced plasma of seawater samples; sample-1, sample-2, and sample-3 respectively.

12

W. TAWFIK Y. MOHAMED

T a b l e 2. A list of the spectroscopic data of the spectral lines used for the determination of plasma
temperature and density of seawater samples..
Element
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Si
Si
Si
Si
Si
Si
Si
Si
Si
Si
Si
Mg
Mg
Mg
Mg
Mg
Mg
Mg

Wavelength [nm]
240.19
244.62
247.63
257.73
261.37
261.42
262.83
265.71
266.32
280.19
282.32
283.31
287.33
357.27
363.96
367.15
368.35
373.99
401.96
405.78
406.21
221.89
243.87
250.69
251.43
251.61
252.41
252.85
288.15
300.67
302.00
390.55
277.66
277.82
277.98
278.14
278.29
279.07
279.55

Aki [s1]
7

2.7910
2.45107
3.78107
6.68107
1.87107
2.35108
5.59107
9.91105
1.01108
1.08108
3.04107
5.92107
4.15107
4.08108
3.20107
1.11108
1.70108
8.30107
3.55107
9.12107
1.07108
3.55107
9.12107
1.07108
1.50106
7.40105
4.66107
6.10107
1.21108
1.81108
7.70107
1.89108
1.32108
1.82108
4.09108
5.43108
2.14108
4.01108
2.60108

Ek [cm1]

gk

49439.62
48686.93
48188.63
49439.62
46068.44
46060.84
48686.93
45443.17
48188.63
46328.67
46060.84
35287.22
45443.17
49439.62
35287.22
48686.93
34959.91
48188.63
46328.67
35287.22
46068.44
46328.67
35287.22
46068.44
45276.18
40991.88
39955.05
39760.28
39955.05
39683.16
39760.28
40991.88
57873.94
57833.4
57873.94
57812.77
57833.4
71491.06
35760.88

3
3
5
3
3
5
3
5
5
7
5
3
5
3
3
3
1
5
7
3
3
7
3
3
9
7
7
9
9
7
11
3
5
3
5
1
3
4
4

Calibration free LIBS identification of seawater salinity

13

T a b l e 2. Continued.
Element

Wavelength [nm]

Aki [s1]

Ek [cm1]

gk

Mg
Mg
Mg
Mg
Mg
Mg
Mg
Mg
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn
Mn

279.79
280.27
281.11
281.17
285.21
291.54
292.86
293.65
258.97
401.81
403.08
403.31
403.45
404.14
404.88
405.55
405.89
406.17
406.35
407.92
408.29
408.36
423.51
441.49
445.16

4.79108
2.57108
1.96108
2.11108
4.91108
4.09108
1.15108
2.30108
2.6108
2.54107
1.70107
1.65107
1.58107
7.87107
7.50107
4.31107
7.25107
1.90107
1.69107
3.80107
2.95107
2.80107
9.17107
2.93107
7.98107

71490.19
35669.31
83520.47
83511.25
35051.26
80693.01
69804.95
69804.95
38543.08
41932.64
24802.25
24788.05
24779.32
41789.48
42143.57
41932.64
42198.56
49415.35
42053.73
42143.57
42053.73
41932.64
46901.13
45940.93
45754.27

6
2
5
3
3
5
2
2
7
8
8
6
4
10
4
8
2
6
6
4
6
8
6
6
8

evaluate their wavelengths , intensities I, and their transition probabilities Aki must
be well known [26].
Reformulating Eq. (1) gives:
I
CF
1
ln ----------------- = ----------- E k + ln -------------U(T )
KT
Ak i gk

(2)

where F is an experimental factor and C is the species concentration.

By plotting the left hand side of Eq. (2) vs. the excited level energy Ek, the plasma
temperature can be obtained from the slope of the obtained straight line.
The plasma temperatures were determined form the emission line intensities of Ca,
Na, and Mg observed in the laser-induced plasma of seawater samples. Here, we aimed
at studying the matrix effect on the plasma temperature. In doing so, the plasma

14

W. TAWFIK Y. MOHAMED

T a b l e 3. The electron temperature Te determined from the observed magnesium, calcium, and sodium
spectral lines in the three seawater samples.
Element
Mg
Ca
Na
Average temperature

Sample-1
11603 K
11547 K
11575 K
11580 K 0.35%

Sample -2
10697 K
10681 K
10665 K
10680 K 0.37%

Sample -3
9900 K
9983 K
9942 K
9940 K 0.43%

temperatures were determined for these elements in the three seawater matrices
sample 1, 2 and 3, respectively. Figure 2 shows nine Boltzmann plots of Eq. (2), for
these three elements in the three seawater samples where the data were fitted with the
least-square approximation. The spectral lines wavelengths, energies of the upper
levels, statistical weights, and transition probabilities used for each element were
obtained from GRIEM [26], NIST [27] and Kurucz Atomic Line Database [28], and
listed in Tab. 2. The slope of the plotted curves yields electron temperature values for
the elements Ca, Na, and Mg in each of the three seawater samples as listed in Tab. 3.
The average values of the plasma temperature are 11580 K 0.35%, 10680 K 0.37%,
and 9940 K 0.43% for seawater samples 1, 2, and 3, respectively. The obtained
plasma temperature values agree with the values obtained by the author before [6].
The slightly difference in the plasma temperature values of the three elements is very
small (< 0.5%) in the same matrix. Since we assumed the LTE conditions, the obtained
small difference in the plasma temperature could be attributed to some experimental
errors. On the other hand, there is significant change in the plasma temperature
(1025%) if the matrix changes. This could be understood as follows: for optically
thin plasma, any change in the sample matrix produces a change in the physical and
chemical properties of the target. This, in turn, affects the energy distribution during
the plasma generation and results in a change in the plasma parameters (Te, Ne) as
proved previously by the author group MARWA et al. [11] for the case of solid state
matrices under conditions similar to the present conditions.
3.2.2. Electron density measurements and matrix effect

The usual method for determination of electron density is the measuring of the
broadening of a suitable emission line of the laser-plasma spectrum. There are several
possible mechanisms of line broadening in plasma: self-absorption, pressure
broadening, Doppler broadening, Stark broadening, etc. Lida reported that the line
broadening and the spectral shift of the emission line are due mainly to self-absorption
phenomenon [29]. In the present study line splitting and the spectral shift, which
provide a good evidence of self-absorption, were monitored carefully. No evidence of
line splitting or spectral shift was observed.

Calibration free LIBS identification of seawater salinity

15

NEMET and KOZMA [30] reported the broadening of transition lines as pressure,
Stark, and Doppler broadening. But pressure and Doppler broadening should not be
so much distinguished among transition lines as it is the case for plasma of solids.
KYUSEOK SONG et al. [31] stated that Stark broadening may be one of the reasons since
the broadening effect increases with the increase of the energy level. Stark broadening
results from Coulomb interactions between the radiator and the charged particles
present in the plasma. Both ions and electrons induce Stark broadening, but electrons
are responsible for the major part because of their higher relative velocities. Therefore,
in our conditions, the profile of a line is mainly contributed to line widths arises from
the Stark effect while the contribution of other mechanisms of broadening (Doppler
effect, Van der Waals broadening, and resonance broadening) can be neglected.
The electrons in the plasma can perturb the energy levels of the individual ions
which broaden the emission lines originating from these excited levels. Stark
broadening of well-isolated lines in the plasma is, thus, useful for estimating the
electron number densities provided that the Stark-broadening coefficients have been
measured or calculated.
Since the instrumental line-broadening exhibit Gaussian shape, then the Stark line
width FWHM can be extracted from the measured line width observed by subtracting
the instrumental line broadening instrument:
FWHM = observed instrument

(3)

In our case instrument was 0.05 nm (determined by measuring the FWHM of the
Hg lines emitted by a standard low pressure Hg lamp).
The width of Stark broadening spectral line depends on the electron density Ne.
Both the linear and the quadratic Stark effect are encountered in spectroscopy. Only
the hydrogen atom and H-like ion exhibit the linear Stark effect. For the linear Stark
effect the electron density should be deduced from H line width from the formula [26]
3 2

N e = C ( N e , T ) FWHM

(4)

The values of the parameter C(Ne, T ) are tabulated in [21], which determine
the relative contribution of the electron collision to the electrostatic fields, and depend
weakly on Ne and T.
For a non-H-like line, the electron density (in cm3) could be determined from the
FWHM of the line from the formula [26]:
FW HM
16
N e ------------------------- 10
2w
where w is the electron impact parameter (Stark broadening value).

(5)

16

W. TAWFIK Y. MOHAMED

T a b l e 4. The plasma electron density Ne determined from the observed calcium, magnesium,
and sodium spectral lines in the three seawater samples.
Element
Ca
Mg
Na
Average Ne

Wavelength
[nm]
422.67
285.21
297.50

Stark broadening
parameter [nm]
7.18104
4.13104
8.85105

Sample-1
[cm3]
3.751018
2.851018
3.381018
3.331018 14.3%

Sample-2
[cm3]
6.051017
5.921017
6.211017
6.031017 2.6%

Sample-3
[cm3]
2.421017
2.781017
2.821017
2.671017 9.4%

The same three elements were taken into consideration for electron density
measurements at wavelengths 422.67, 285.21 and 297.50 nm for Ca, Mg, and Na,
respectively. The three candidates lines data points were fitted with Gaussian fitting
function using the Origin software (version 7.0220, Origin Lab corporation, USA) to
determine 1/2 (the wavelength of full width at half-maximum). The Stark broadening
W values are taken from GRIEM [26] at electron temperature of 10000 K and listed in
Tab. 4. Substituting the values of 1/2 and the corresponding value of Stark
broadening W in Eq. (5), the electron densities for Ca, Mg, and Na were determined.
These steps were repeated for each of the seawater samples. The obtained electron
density values are listed in Tab. 4 and the average values of the plasma density are
3.331018 cm3 14.3%, 6.031017 cm3 2.6% and 2.671017 cm3 9.4% for
seawater samples 1, 2, and 3, respectively. Since we assumed the LTE conditions, the
slight difference in the plasma density values of the three elements (Na, Ca, and Mg)
is small (2.614.3%) in the same matrix which could be attributed to the experimental
errors. On the other hand, there is a significant change in the plasma density up to order
of magnitude if the matrix changes. This could be understood, as stated for the case
of plasma temperature, that any change in the sample matrix produces a change in the
physical and chemical properties of the target and, in turn, affects the energy
distribution during the plasma generation and results in a change in the plasma
parameters (Te, Ne) [6].
Finally, by knowing the electron density and the plasma temperature we can
determine whether LTE assumption is valid applying the criterion given by
MCWHIRTER [32]. The lower limit for electron density for which the plasma will be in
LTE is:
12

N e 16 10 E

13

12

(6)

where E is the largest energy transition for which the condition holds and T is the
plasma temperature [15]. In the present case E = 4.9 eV for Mg at 279.55 nm (see
Tab. 2) and its electron density lower limit value given by Eq. (6) is 1.881016 cm3
(see ref. [11]). The experimentally calculated densities are greater than this value,
which is consistent with the assumption that the LTE is prevailing in the plasma.

Calibration free LIBS identification of seawater salinity

17

4. Conclusions
In the present work, we used an accurate LIBS setup to identify different seawater
samples (with different salinities) using an optical fiber probe. In doing so, we study
the matrix effect on the plasma characterization of seawater samples. The obtained
results showed that both electron temperature and density are related to the matrix
composition and change if the matrix changes. Moreover, Te and Ne could be measured
using any of the three elements (Na, Ca, and Mg) in the seawater matrix. This means
that Te and Ne represent a fingerprint plasma characterization for a given seawater
sample and its salinity could be identified using only one element without need of
analysis of the rest of elements in the seawater matrix. This could be done by building
a database containing the determined values of Te and Ne for a wide range of seawater
salinities. Then the salinity of the unknown seawater sample could be identified just
by comparing its measured Te and Ne values with the previously stored values in our
database.
The obtained results indicate that it is possible to improve the exploitation of LIBS
in the remote on-line environmental monitoring, by following up only a single element
as a marker to identify the seawater matrix composition and salinity without need to
analyze that matrix which saves a lot of time and efforts.
Acknowledgment The author acknowledges Prof. M. Abdel Harith specially for offering the seawater
samples.

References
[1] MIN Z.J., LI J.P., JIANG S.H., Measurement of salt salinity in solar pond by supersonic method,
Acta Eneglae Solaris Sinica 16(2), 1995, pp. 2248.
[2] DINIZ F.B., DE FREITAS K.C.S., DE AZEVEDO W.M., Salinity measurements with polyaniline matrix
coated wire electrodes, Electrochemistry Communications 1(7), 1999, pp. 2713.
[3] MINATO H., KAKUI Y., NISHIMOTO A., NANJO M., Remote refractive index difference meter for salinity
sensor, IEEE Transactions on Instrumentation and Measurement 38(2), 1989, pp. 60812.
[4] ZHAO Y., LIAO Y.B., Novel optical fiber sensor for simultaneous measurement of temperature and
salinity, Sensors and Actuators B: Chemical 86(1), 2002, pp. 637.
[5] ZHAO Y., ZHANG BO, LIAO YANBIAO, Experimental research and analysis of salinity measurement
based on optical techniques, Sensors and Actuators B: Chemical 92(3), 2003, pp. 3316.
[6] TAWFIK Y. MOHAMED W., Quantitative elemental analysis of seawater by laser induced breakdown
spectroscopy, International Journal of Pure and Applied Physics 2(1), 2006, pp. 1121.
[7] RADZIEMSKI L.J., CREMERS D.A., Laser Induced Plasmas and Applications, Marcel Dekker, New York
1989.
[8] SOLIMAN M., TAWFIK W., HARITH M.A., Quantitative elemental analysis of agricultural drainage
water using laser induced breakdown spectroscopy, [In] First Cairo Conference on Plasma
Physics and Applications, Cairo, Egypt, Forschungszentrum Juelich GmbH, Bilateral Seminars of
the International Bureau, Vol. 34, 2003, pp. 2403.
[9] TAWFIK Y. MOHAMED W., ASKAR A., Study of the matrix effect on the plasma characterization
of heavy elements in soil sediments using LIBS with a portable Echelle spectrometer, Progress in
Physics 1, 2007, pp. 4652.

18

W. TAWFIK Y. MOHAMED

[10] MOHAMAD SABSABI, VINCENT DETALLE, MOHAMED A. HARITH, WALID TAWFIK, HISHAM IMAM,
Comparative study of two new commercial Echelle spectrometers equipped with intensified CCD
for analysis of laser-induced breakdown spectroscopy, Applied Optics 42(30), 2003, pp. 60948.
[11] ISMAIL M.A., IMAM H., ELHASSAN A., TAWFIK Y. MOHAMED W., HARITH M.A., LIBS limit of detection
and plasma parameters of some elements in two different metallic matrices, Journal of Analytical
Atomic Spectrometry 19(4), 2004, pp. 48994.
[12] CHARFI B., HARITH M.A., Panoramic laser-induced breakdown spectrometry of water,
Spectrochimica Acta Part B: Atomic Spectroscopy 57(7), 2002, pp. 114153.
[13] ARCA G., CIUCCI A., PALLESCHI V., RASTELLI S., TOGNONI E., Trace element analysis in water by
laser-induced breakdown spectroscopy technique, Applied Spectroscopy 51(8), 1997, pp. 11025.
[14] FICHET P., MAUCHIEN P., WAGNER J.-F., MOULIN C., Quantitative elemental determination in
water and oil by laser induced breakdown spectrodcopy, Analytica Chimica Acta 429(2), 2001,
pp. 26978.
[15] SAMEK O., BEDDOWS D.C., KAISER J., KUKHLEVSKY S.V., LISKA M., TELLE H.H., YOUNG J.,
Application of laser-induced breakdown spectroscopy to in situ analysis of liquid samples, Optical
Engineering 39(8), 2000, pp. 224862.
[16] WISBRUN R., SCHECHTER I., NIESSNER R., SCHROEDER H., Laser-Induced Breakdown Spectroscopy
for Detection of Heavy Metals in Environmental Samples, Vol. 1716, SPIE, Bellingham,
Washington, 1992, pp. 214.
[17] SALLE B., CHARLEARD C., DETALLE V., LACOUR J.L., MAUCHIEN P., NOUVELLON C., SEMEROK A., Laser
ablation efficiency of metal samples with UV laser nanosecond pulses, Applied Surface Science
138139, 1999, pp. 3025.
[18] BULATOV V., KRASNIKER R., SCHECHTER I., Study of matrix effects in laser plasma spectroscopy by
combined multifiber spatial and temporal resolutions, Analytical Chemistry 70(24), 1998,
pp. 530211.
[19] XU L., BULATOV V., GRIDIN V.V., SCHECHTER I., Absolute analysis of particulate materials by laser
-induced breakdown spectroscopy, Analytical Chemistry 69(11), 1997, pp. 21038.
[20] GOODE S.R., MORGAN S.L., HOSKINS R., OXSHER A., Identifying alloys by laser-induced breakdown
spectroscopy with a time-resolved high resolution echelle spectrometer, Journal of Analytical
Atomic Spectrometry 15(9), 2000, pp. 11338.
[21] EPPLER A.S., CREMERS D.A., HICKMOTT D.D., FERRIS M.J., KOSKELO A.C., Matrix effects in the
detection of Pb and Ba in soils using laser-induced breakdown spectroscopy, Applied Spectroscopy
50(9), 1996, pp. 117581.
[22] CIUCCI A., CORSI M., PALLESCHI V., RASTELLI S., SALVETTI A., TOGNONI E., New procedure for
quantitative elemental analyses by laser induced plasma spectroscopy, Applied Spectroscopy 53(8),
1999, pp. 9604.
[23] BULAJIC D., CORSI M., CRISTOFORETTI G., LEGNAIOLI S., PALLESCHI V., SOLVETTI A., TOGNONI E.,
A procedure for correcting self-absorption in calibration-free laser induced breakdown
spectroscopy, Spectrochimica Acta Part B: Atomic Spectroscopy 57(2), 2002, pp. 33953.
[24] CORSI M., PALLESCHI V., SALVETTI A., TOGNONI E., Making LIBS quantitative: a critical review of
the current approaches to the problem, Research Advances in Applied Spectroscopy 1, 2000,
pp. 417.
[25] SIMEONSSON J.B., MIZIOLEK A.W., Time-resolved emission. studies of ArF laser-produced
micro-plasmas, Applied Optics 32(6), 1993, pp. 93947.
[26] GRIEM H.R., Plasma Spectroscopy, McGraw-Hill, New York 1964.
[27] NIST National Institute of Standards and Technology, USA, electronic database, http://
physics.nist.gov/PhysRefData/ASD/lines_form.html.
[28] Kurucz Atomic Line Database, http://www.cfa.harvard.edu/amdata/ampdata/kurucz23/sekur.html.
[29] IIDA Y., Effects of atmosphere on laser vaporization and excitation processes of solid samples,
Spectrochimica Acta Part B: Atomic Spectroscopy 45(12), 1990, pp.135367.

Calibration free LIBS identification of seawater salinity

19

[30] NEMET B., KOZMA L., Time-resolved optical emission spectrometry of Q-switched Nd:YAG
laser-induced plasmas from copper targets in air at atmospheric pressure, Spectrochimica Acta
Part B: Atomic Spectroscopy 50(14), 1995, pp. 186988.
[31] KYUSEOK SONG, HYUNGKI CHA, JONGMIN LEE, YONG-ILL LEE, Investigation of the line-broadening
mechanism for laser-induced copper plasma by time-resolved laser-induced breakdown
spectroscopy, Microchemical Journal 63(1), 1999, pp. 5360.
[32] MCWHIRTER R.W.P., In Plasma Diagnostic Techniques, [Eds.] R.H. Huddlestone, S.L. Leonard,
Chap. 5, Academic Press, New York 1965, p. 206.

Received January 11, 2007

in revised form March 5, 2007