See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/233198931

Analysis of Minerals and Rocks by Laser-Induced

Breakdown Spectroscopy

Article in Spectroscopy Letters · September 2011

DOI: 10.1080/00387010.2011.573048

CITATIONS READS

13 463

6 authors, including:

Diego Díaz Pace Mayra Garcimuño

National University of Central Buenos Aires National University of Central Buenos Aires
17 PUBLICATIONS 188 CITATIONS 6 PUBLICATIONS 28 CITATIONS

SEE PROFILE SEE PROFILE

Daniela Bertuccelli
National University of Central Buenos Aires
20 PUBLICATIONS 198 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Diego Díaz Pace
letting you access and read them immediately. Retrieved on: 08 September 2016
Spectroscopy Letters, 44:399–411, 2011
Copyright # Taylor & Francis Group, LLC
ISSN: 0038-7010 print=1532-2289 online
DOI: 10.1080/00387010.2011.573048

Analysis of Minerals and Rocks by

Laser-Induced Breakdown Spectroscopy
Diego M. Dı́az Pace1,2,
Norberto A. Gabriele3, ABSTRACT Laser-induced breakdown spectroscopy (LIBS) technique was
Mayra Garcimuño1,4, applied for rapid analysis of major and minor elements composing geologi-
Cristian A. D’Angelo1,4, cal samples including minerals, rocks, and a soil sample. The plasma was
Graciela Bertuccelli1,2,
produced in air at atmospheric pressure by focusing on the targets a pulsed
and Daniela Bertuccelli1,2
1 infrared Nd:YAG laser in open-path configuration. The emitted light in the
Instituto de Fı́sica ‘Arroyo Seco’,
UV-Vis was analyzed by a compact LIBS system to measure spectral emis-
Facultad de Ciencias Exactas,
U.N.C.P.B.A., Campus sion lines of Si, Al, Fe, Ca, Na, K, Mg, C, Cu, Mn, and Ti. The experimental
Universitario, Paraje Arroyo Seco, issues relevant for the analysis of the different samples were investigated by
Tandil, Buenos Aires, Argentina taking into account their peculiar features: drilling through a weathered
2
CONICET, Rivadavia, Buenos layer, roughness and grain-size considerations, statistical averaging, and
Aires, Argentina accuracy of the measurements. In this approach, the characterization of
3
Facultad de Ciencias Exactas,
the samples was achieved by studying the relative variations of the emission
U.N.C.P.B.A., Campus
Universitario, Paraje Arroyo Seco, intensities of each element normalized with respect to an internal standard.
Tandil, Buenos Aires, Argentina The present study shows the usefulness of LIBS as a tool for reliable identi-
4
CICPBA, La Plata, Buenos Aires, fication of field samples.
Argentina
KEYWORDS laser-induced breakdown spectroscopy, LIBS, minerals, rocks

INTRODUCTION
Minerals and rocks that are found at a particular location contain impor-
tant clues about the geologic history of that site and how it originated and
has evolved with time. Field geology relies greatly on identification based
on a geologist’s assessment of certain physical properties and later classi-
fication at a laboratory facility, which is not always accurate. Due to its
advantages of lack of any sample preparation and in situ measurements,
Laser-Induced Breakdown Spectroscopy (LIBS) offers a tool for more-
reliable identification of field samples.
The aim of this work was to carry out a rapid analysis of major and minor
elements present in geological samples by LIBS technique. In this approach,
we employed a compact system equipped with a nonintensified detector
and without temporal resolution, and the results allowed the compositional
Received 8 February 2011; analysis of several minerals, rocks, and a soil sample. Furthermore, the LIBS
accepted 14 March 2011. capability to drill through weathered layers on the sample surface to deter-
Address correspondence to Diego M. mine its composition as well as that of the underlying bulk composition
Dı́az Pace, Pinto 399, B7000GHG
Tandil, Buenos Aires, Argentina. was addressed, and the experimental issues relevant for the analysis of
E-mail: [email protected]

399
the different samples such as the accuracy of the analysis is based on obtaining the overall compo-
measurements, sample homogeneity, and grain size sition through inspection of the target by producing
were investigated. a small crater on its surface that is replicated at differ-
ent places, assuming that the plasma composition is
representative of that of the sample previous to the
LIBS Technique laser ablation (stoichiometric ablation). Moreover,
LIBS is a very useful technique to survey, identify, natural materials usually present different degrees
and quantify the chemical composition of a wide range of homogeneity, surface roughness, and grain size,
of samples in solid, liquid, or gaseous state by perform- which affect the plasma formation, causing strong
ing spectral analysis of the light emitted from a plasma shot-to-shot variations in the measured spectral
created on the target.[1] Thus, the resulting UV-Vis intensities.[12]
emission spectrum contains characteristic lines from In addition, the use of LIBS for field analysis
the atoms, ions, and simple molecules present in the imposes some constraints on size, weight, and
plasma. LIBS possess several advantages compared simplicity of the setup for placement on a rover or
to other conventional techniques: (1) simple experi- rugged portable equipment in geological prospect-
mental setup, (2) few or no sample preparation ing.[13] Therefore, laboratory tests are of major impor-
needed, (3) in situ and stand-off analysis capabilities, tance for evaluating and optimizing the performance
(4) microdestructiveness, (5) real-time measurement of the technique with the goal of obtaining useful
of many elements simultaneously, and (6) depth profil- analytical information under several environmental
ing and spatial resolution. These features allow LIBS to conditions, such as at close or stand-off distances
perform measurements under conditions not feasible and in atmospheres with different pressures and
with conventional analytical techniques (e.g., compositions. Furthermore, a spectral library under
ETA-AAS, ICP-AES, and ICP-MS) that, supported by adequate conditions can be constructed in order to
the actual technology development in the compo- analyze unknown samples by comparing the mea-
nents, make it a very promising technique to be inte- surements with reference spectra from a spectral
grated with the established atomic spectroscopic database.[15,16]
methods.[2] Hence, it has been recently employed as Several researchers have developed transportable
an analytical tool for field applications with the aim instruments to promote LIBS as an analytical tool in
of providing rock and soil analysis, as well as examin- the field. The first reported portable system was
ation and characterization of archaeological find- made by Yamamoto et al.[17] who presented the
ings.[3–7] In the stand-off approach, direct remote development of a portable instrument and its feasi-
sensing, in which the laser and signal are transmitted bility for environmental monitoring. It consisted of
through a transparent atmosphere, is performed.[8] a small laser to generate the laser sparks and a
The use of LIBS for material analysis is based on the fact fiber-optic cable to collect the light, which was spec-
that all the elements are characterized by their distinc- trally resolved with a compact spectrograph and
tive spectra. Thus, the elements present in the sample detected with an unintensified, nongated CCD. Pre-
are identified using their characteristic spectral lines. liminary tests were successfully performed measur-
Nevertheless, quantifying their concentration is a diffi- ing metals in soil, paint, and particles collected on
cult task. The most common approach is using cali- filters. In the work of Sallé et al.,[14] LIBS was evalu-
bration curves constructed with reference samples, ated for qualitative analysis of different rock types
but these are not possible when dealing with unique, under Mars atmospheric conditions using a compact
complex, or a priori unknown samples (e.g., museum experimental setup. Spectral lines from Ca, Mg, Fe,
pieces, archaeological findings). Moreover, the and Si were detected in basalt and dolomite in the
ablation process depends on the target properties.[9,10] 225–320 spectral range, without gating the detection
In this direction, several methods for overcoming system, and using a nonintensified detector. The
matrix effects have been successfully employed, but results were compared with spectra recorded from
these are time-consuming.[11] samples constructed with rock powder standards.
In the case of samples collected from the field, that In the paper of Palanco et al.,[13] the design and con-
is, geological and archaeological fragments, LIBS struction considerations of field-deployable LIBS

D. M. Dı´az Pace et al. 400

equipment provided with an intensified CCD detec- may too contain different groups of minerals with
tor with time resolution for remote elemental sensing very similar characteristics.
were reported. Moreover, the main parameters influ- Mineral compositions are more complex than their
encing its performance at stand-off distances were chemical formulas indicate and are heterogeneous
studied. Harmon et al.[12] reported the development within them. Hence, their compositions vary by sub-
of a prototype LIBS system as a real-time man- stitution of one or more of its elements by another
portable geochemical sensor. Time-resolved broad- atom that takes its place in the crystalline framework,
band spectra in the region 200–980 nm under normal for example the isomorphic series.[24]
atmospheric conditions were acquired from soil, A rock is a solid part of the Earth’s crust that is gen-
minerals, and metal samples. Results were encour- erally composed of more than one mineral. Then,
aging for the use of this prototype in geochemical, rocks cannot be classified in the same way as can
mineralogical, and environmental applications. An minerals because they may contain an inhomoge-
example of the application of LIBS to provide online, neous mixture of minerals in the form of grains. Geol-
real-time analyses was reported by Gaft et al.,[18] who ogists classify rocks according to how they were
developed an industrial LIBS machine based on a created: Igneous rocks, which are further divided into
small and robust CCD spectrometer with time resol- intrusive or plutonic rocks and extrusive or volcanic
ution, which was successfully tested for quantifying rocks; sedimentary rocks; and metamorphic rocks.
minerals in raw ores and main ash-forming elements
in coal. Due to technology development, modern
and compact LIBS instruments have become com- MATERIALS AND METHODS
mercially available in the market,[19,20] in which
the use of low-cost nongated detection systems
Experimental Setup
have proved very good performance for specific The setup used to analyze the samples in the lab-
applications.[21–23] oratory is shown in Fig. 1. A Q-switched Nd:YAG
pulsed laser (Laser Big Sky, Model ultra, Laser Tech-
nologies, Inc., USA) operating at the fundamental
Minerals and Rocks wavelength 1064 nm, at a repetition frequency of
Minerals are naturally occurring solids with an 2 Hz and with a pulse width of 5–7 ns and an energy
organized crystalline structure composed of either a of 50 mJ was used to produce the plasma. The sam-
single chemical element or a compound, and they ples were placed on a translator stage allowing the
are the fundamental constituents of rocks and soils. displacement in a plane normal to the incidence
They can be accurately identified by both X-ray direction in order to control the position of the sam-
and chemical analysis that reveal their structure and pling point and obtain spatially resolved measure-
chemical composition, which are the ‘‘fingerprints’’ ments. The laser pulses were focused with a quartz
of a mineral. Nevertheless, these tests require lens (L1) of 200-mm focal length onto the sample sur-
expensive equipments, are time-consuming, and face. The plasma light was collected at a right angle
often destroy the specimen. In practice, since struc- with respect to the laser beam and focused with a
ture and composition affect certain physical proper- second quartz lens (L2) of 100-mm focal length onto
ties of minerals, geologist use a number of criteria the entrance slit (100-mm width) of a small mono-
such as shape, hardness, and color to identify miner- chromator (Model 504, Acton Research Corporation,
als in the field. But these features usually do not USA) of 50-cm focal length, provided with a
necessarily lead to the correct recognition of any 1200-grooves=mm diffraction grating and spectral
unknown material. On Earth exist thousands of dif- coverage over the 200–1000-nm range, over which
ferent mineral species, which are classified according all elements have emission lines. In contrast with
to their chemistry into the following classes: native the gated, intensified CCD cameras usually used for
elements, sulfides, halides, oxides, carbonates, sul- LIBS experiments, light detection was performed by
fates, phosphates, silicates, and mineraloids. More- a nonintensified nongated linear photodiode array
over, each class is further divided into several detector (Model RY-1024, PDA, Princeton Instru-
subclasses by their structures, and each subclass ments, Inc., USA) covering a spectral window of

401 Analysis of Minerals

FIGURE 1 Experimental setup and observation geometry.

TABLE 1 Mineral Samples Analyzed

Mineral samples Class Subclass Group Formula Characteristics
Purchased gems
Quartz (Rock crystal) Silicate Tectosilicate — SiO2 Quartz crystal, prismatic,
pyramidal, hyaline,
without visible impurities.
Magnesite Carbonate — — MgCO3 Polished, cryptocrystalline,
milky white, hardness 6.
No chemical reaction
occurred with HCl.
Manganite –Calcite Carbonate — — [Mn,Ca]CO3 Polished, milky white,
semitranslucent, hardness
5. Chemical reaction
occurred with HCl.
Tourmaline Silicate Cyclosilicate Tourmalines (Na,Ca)(Mg,Al)6- Polished, black, crystalline,
[B3Al3Si6(O.OH)30 hardness >7. Chemical
formula corresponding to
the most common species.

Gems collected from the field

Quartz Silicate Tectosilicate — SiO2 Cloudy, prismatic, pyramidal
crystals and with a
superficial coating of
reddish–gray clusters.
Biotite Silicate Phyllosilicate Micas K(Mg,Fe)3AlSi3O10(F,OH)2 Translucent, metallic gray,
forming a thick-grained
association with quartz.
Microcline Silicate Tectosilicate Feldspars KAlSi3O8 Crystalline, pale pink, good
cleavage, with albite
lamella pertite, small
clusters (<1 mm in size) of
white mica (muscovite)
and quartz.
Malachite Carbonate — — CuCO3  Cu(OH)2 Green, soft bright, long
radial fibers, hardness <5,
and pale green streak.

D. M. Dı´az Pace et al. 402

50 nm. The PDA was triggered at the start of the laser RESULTS AND DISCUSSION
pulse in such a way that it recorded the plasma emis-
The major and minor components of the samples
sion during its entire lifetime. A computer was used
were analyzed by measuring line emission intensities
for data acquisition from the detection system.
from Si, Al, Fe, Ca, Na, Mg, and K, which form
The experimental arrangement was in an open-
the Earth’s crust; and other elements of interest to
path framework in which the laser beam and the
geologist, such as Mn, Ti, Cu, and C. Since
plasma light are transmitted through the air atmos-
time-integrated spectra contain contributions from
phere. This configuration was chosen for its sim-
elemental emission lines and the plasma continuum
plicity and versatility, which make feasible many
emission, the signal-to-noise ratio was improved by
applications. In fact, the analysis of environmental
accumulation of many spectra on each sampling
and geological targets in the field and remote sensing
point on the target surface. The compositional analy-
at real-time (open-path, stand-off LIBS) under
sis was carried out by measuring the relative varia-
reduced air pressure or simulated extraterrestrial
tions of line emission intensities from the analytes
atmosphere are envisaged.
between the different samples, selecting several
emission lines to represent the corresponding
Samples elemental concentration for each element.
LIBS analyses were accomplished on different
mineral and rock samples, which were classified into
Analysis of Minerals
three different groups. The first was composed by
purchased mineral gems: quartz (rock crystal), mag- In gem samples, spectra were recorded by
nesite, manganite–calcite, and tourmaline (Table 1). accumulation of 20 consecutive laser sparks deliv-
The second group was composed by minerals col- ered in one arbitrary point of the surface. Even
lected from the field: biotite, microcline, quartz, though their compositions are expressed in a simple
and malachite (Table 1). The third group was formed manner in Table 1, minerals are inhomogeneous in
by collected rocks: caliche, tonalite, and basalt. detail because of the large amount of ionic substi-
Moreover, a soil sample (loess) was investigated tution, which originates a natural compositional vari-
(Table 2). All the samples were previously examined ation on a micron spatial scale. Thus, in the case of
through a megascopic analysis to determine their natural minerals, the spectra were obtained by aver-
characteristics (shown in Table 1 and Table 2) and aging measurements accumulated from firing the
address the subsequent LIBS analysis. The samples samples on five different surface locations. On each
collected from the field were analyzed in their natu- location, the signal was integrated over 20 laser shots
ral state and without any previous treatment. at the same point to obtain the bulk composition

TABLE 2 Rock Samples Analyzed

Rock samples Classification Minerals Characteristics
Tonalite Igneous. Intrusive. Plagioclase, hornblende, Rough surface, crystalline,
biotite, quartz, opaque medium-sized grains, gray.
minerals. Foliate, affected by
metamorphism.
Basalt Igneous. Volcanic. Microcrystalline minerals, glass. Hard, dense, dark gray,
aphanitic crystalline with
glassy appearance and
vesicles.
Caliche Sedimentary. Chemical. Calcite, quartz. Silts. Earthy white, fined grains,
moderate consolidation,
chemical reaction occurred
with HCl.
Loess Sedimentary. Eolian. Quartz, k-feldspar, plagioclase, Friable, pale yellow (dried),
mica, hornblende. soft.

403 Analysis of Minerals

after 30 previous cleaning shots to remove weath- Mn, the presence of emission lines of other elements
ered layers or any possible artifact due to handling was detected. We speculated that this sample is
the samples at the laboratory. probably a solid solution of several elements (Fe
Typical examples of the spectra measured from and Mg) and has some impurities of quartz and K
minerals, with the most intense lines identified, are feldspar (Si, Al, Na, and K). The detection of Fe in
shown in Fig. 2, Fig. 3, and Fig. 4. In Fig. 2a, it can tourmaline confirms, in agreement with its dark
be seen that a strong Si transition, a spectral signature tonality, that it corresponds to the Fe-rich variety
of silicates, is present in quartz, manganite–calcite, called ‘‘schorl.’’ In Fig. 3, it is shown that spectra from
and tourmaline. Moreover, Al and Si lines character- biotite contain Si, Al, K, Mg, and Fe. Moreover,
istic of aluminosilicates were detected in tourmaline, microcline exhibited lines of Si, Al, Ca, Na, and K.
and several Fe and Mg lines were observed in both The measurements on malachite allowed us to
tourmaline and manganite–calcite. In Fig. 2b, lines determine the presence of Mg, Ca, Si, and Fe, as well
of Si, Ca, and Fe were identified in tourmaline and as K. Additionally, Mn was detected in manga-
manganite–calcite. In this region, Si emissions are nocalcite, Ti and Cu were found in malachite,
weaker, while those from Mg and Fe remain unre- and C was in both of them and also in magnesite
solved. In addition, strong Ca lines were recognized (Fig. 4).
in manganite–calcite, and strong Al lines were recog- In plasma produced by the interaction of the laser
nized in tourmaline. In Fig. 2c, the well-known with a solid target in air, the surrounding atmosphere
doublet of Na was detected in tourmaline and near the plume is excited, allowing the detection of
manganite–calcite. In Fig. 2d, lines corresponding its major constituents. Moreover, hydrogen emission
to K were observed in tourmaline and manganite– is expected because of the water vapor due to the
calcite, as well as other lines of Si and Mg. In the natural humidity of the atmosphere.[25–27] In our
manganite–calcite sample, in addition to Ca and experiment, emission atomic lines of N, O, and H

FIGURE 2 Detected lines of the major elements of Earth’s crust in mineral samples of Table 1. Some spectra were shifted vertically for clarity.

D. M. Dı´az Pace et al. 404

FIGURE 3 Elements detected in minerals microcline, biotite, and malachite collected from the field.

were detected. O and H lines are the result of contri- layer of MgOx and FeOx was completely removed
butions of both the sample and the ambient air, and after 30 laser shots, and then the presence of Si
they should be analyzed at a reduced pressure to was detected clearly.
minimize the contribution of the atmosphere to line
intensities in order to estimate their concentration
in the sample.[8] On the other hand, N lines origi- Issues Affecting the Accuracy
nated only from the air, since the samples do not of LIBS Analysis
contain any nitrogen at all.
The characteristics of plasmas generated on the
surface of solid samples depend in a complicated
Drilling Through Weathered Layers way on the experimental conditions, including the
The natural quartz fragment presented a super- features of the material being analyzed, the laser
ficial coating that needed to be passed over and that parameters, and the detection system. Thus, the
needed its character to be determined. The cleaning reliability of the analytical results was studied
process was monitored by collecting the plasma through the different issues affecting the accuracy
emission layer by layer. Figure 5a shows the spectral of LIBS measurements.
measurements as the coating is drilled by repetitive In our experiment, variations in the measured line
sampling at the same spot, where emission lines of intensities were observed as a result of (1) the sample
Fe, Mg, and Si were identified. The changes in Mg features, (2) fluctuations intrinsic to LIBS method, and
and Fe line intensities with respect to that of Si (3) the instrumental noise. The sample characteristics
are shown in Fig. 5b. When the emissions of the first affect the accuracy of the results through systematic
elements stopped decreasing and the emission of the errors caused by matrix effects, surface roughness,
latter element stopped increasing, the cleaning pro- grain size, and compositional inhomogeneities. More-
cess was complete. It can be seen that a superficial over, laser energy fluctuations, focusing length, and

405 Analysis of Minerals

FIGURE 4 Detected lines of C, Mn, Ti, and Cu in manganite– FIGURE 5 (a) Depth profiling of the quartz sample collected
calcite, magnesite, and malachite. from the field through a weathered coating of MgOx and FeOx oxi-
des; (b) Peak-intensity ratios of Mg and Fe lines with respect to Si
as a function of the laser shot number.
the region of the plasma plume imaged onto the
entrance slit of the monochromator, which is
spatially inhomogeneous, cause both random and on the entrance slit of the monochromator, keeping
systematic variations related to the reproducibility of fixed the focusing length, in such a way that the
the method under different conditions of mea- detected signal was optimized. As an example,
surement. Additionally, the dark current and the pho- shot-to-shot variations in line intensities of Fe, Mg,
ton noise originate random intensity fluctuations Al, and Si were observed in one-shot spectra from
independent of the source that affect the instrumental biotite, acquired under the aforementioned pro-
repeatability under the same conditions of measure- cedure. The accumulated spectra had averaged peak
ment. intensities with an RSD  10% for all the lines mea-
In order to minimize the systematic errors, thereby sured. These variations arose mainly from the instru-
increasing the accuracy, and to evaluate the contri- mental noise and laser energy fluctuations, being
bution of each source of random errors (precision independent of the analyte. The experimental peak
given by the relative standard deviation [RSD] of suc- intensities of the lines measured (previously cor-
cessive measurements) in the measurements, the fol- rected by the quantum efficiency of the detector
lowing analysis was carried out: Samples providing a and by background subtraction) were normalized
plane side (i.e., polished gems, biotite) were exam- with respect to the N peak emission line at
ined to minimize changes in the signal originated 744.23 nm to compensate for energy variations. N
by laser defocusing due to either surface roughness was selected as an internal standard because it is
or repositioning of them on the sample stage. In the most abundant element present in the air ambi-
addition, the central zone of the plasma was imaged ent, and it is generally not present in minerals and

D. M. Dı´az Pace et al. 406

rocks. The stronger N line at 746 nm was discarded
because it is affected by interference of close Ca
lines. We observed that the normalized ratios had
an improved RSD <5% corresponding only to the
instrumental repeatability. Therefore, despite signifi-
cant shot-to-shot variations in the emission intensity
of the plasma, the line intensity ratios had good accu-
racy (5%) for compositional analyses just integrat-
ing several laser shots.

Analysis of Rocks
Rocks are intrinsically inhomogeneous since they
are a mixture of different particles of minerals.
Consequently, they usually present several degrees
of surface texture and=or grain size, which should
be considered in the analysis. Most rocks are formed
up by grains of minerals larger than 1 mm and up to
some centimeters in size. The mineral grains may be
large enough to be easily identified (e.g., tonalite),
barely distinctive grains (e.g., basalt), or microscopic
grains (e.g., slate). This gives rise to some concerns
since the crater originated by the focused laser spot
onto the sample surface is typically on the order of
100 mm to 1 mm wide and 10–20 mm deep, depend-
ing on the irradiance and the properties of the target. FIGURE 6 Spectra obtained from tonalite at different sampling
Therefore, measurements representative of only one positions.

mineral or perhaps of the interface between two

minerals could be obtained. Thus, the issues of inho- points of tonalite is shown in Fig. 6. It can be seen
mogeneity and grain size were considered in carry- that different emission lines with different intensities
ing out a grid of measurements consisting in a set are present at different positions on the surface. This
of sampling points regularly distributed over an area is clearly reflected in the statistics computed from the
with such a size and spacing depending on the grain grid with the experimental peak-intensity ratios of
size. The emission of the plasmas at each point of the the available lines (Table 3). The high RSD (70%)
grid was accumulated to produce each spectrum. is a joint consequence of mineral grains and surface
In this approach, two samples of igneous rocks, roughness (originating laser defocusing). For the
tonalite and basalt, and one sample of sedimentary purpose of this work, it was a good approximation
rock, caliche, were analyzed as examples of rocks considering the sampling area (16 mm2) as rep-
that could be commonly found in the field resentative of the overall features and the chemical
(Table 2). Additionally, a loess sample in the form composition of this kind of sample.
of a pressed pellet was inspected to assess LIBS for On the other hand, basalt was a fine-grained frag-
potential environmental applications. ment with a high degree of surface roughness. Spec-
Tonalite was the most heterogeneous sample in tra were averaged over five replicates at different
our work. It presented easily seen coarse mineral random positions of the sample surface. The statistics
grains >1 mm in size and a considerably rough sur- of the peak-intensity ratios are shown in Table 4. The
face. A grid of measurements was carried out with RSD was approximately 15%, as a result of the
a total of 25 sampling points, arranged in five rows combination of fluctuations due to laser energy
by five columns with 1-mm spacing between them. and roughness, while the grain-size effects were
An example of the spectra recorded from several grid negligible.

407 Analysis of Minerals

TABLE 3 Statistics of Line Peak-Intensity Ratios from Tonalite
in a Grid of Measurements

Peak-Intensity
ratio Min Max Range Mean RSD (%)
Fe=N I275.01=I744.23 0 2.9 2.9 1.6 70.7
Mg=N I279.55=I744.23 0 7.0 7.0 3.6 72.6
I280.27=I744.23 0 5.8 5.8 3.0 71.0
I285.21=I744.23 0 2.8 2.8 1.6 62.2
Si=N I285.21=I744.23 1.6 2.6 1.0 2.0 15.2
Al=N I307.47=I744.23 2.2 9.1 6.9 5.9 43.7
I308.85=I744.23 1.1 4.8 3.7 3.3 36.4
Ca=N I393.37=I744.23 0.6 4.2 3.6 1.6 53.5
I396.85=I744.23 0.7 3.1 2.4 1.2 50.2
Na=N I589.00=I744.23 0.8 5.9 5.1 3.5 49.4
I589.60=I744.23 0.8 4.6 3.8 2.8 45.2
K=N I766.49=I744.23 0.7 4.6 3.8 1.8 61.5
I768.90=I744.23 0.8 3.5 2.8 1.4 54.4

Tonalite and basalt are both igneous rocks with an

analogous elemental composition where the charac-
teristic spectral lines of the elements forming the
minerals quartz, mica, feldspar, and ferromagnesian
mineral were recognized. Like basalt, caliche and
loess were also fine-grained samples with moderate
consolidation. No significant fluctuations were
observed in this case, so measurements from both
FIGURE 7 Some spectra obtained from (a) caliche; and (b)
samples were not affected by roughness. Examples loess.
of spectra from caliche and soil are shown in Fig. 7a
and Fig. 7b, respectively. It is observed that Ca lines
are dominant in the spectrum of caliche, and strong Sample Characterization
Ti lines were detected in loess. Figure 8 depicts the results obtained in this work
from the analysis of several collected minerals, rocks,
and a soil sample. The analysis was based on com-
TABLE 4 Statistics of Line Peak-Intensity Ratios from Basalt in puting the ratios of the peak intensities of the most
Different Positions intense emission lines characteristic of each element
Peak-Intensity with respect to an internal standard (i.e., N). Differ-
ratio Min Max Range M RSD (%) ent lines of the same analyte having different intensi-
Fe=N I275.01=I744.23 2.6 3.1 0.5 2.9 7.8 ties but a similar qualitative behavior were averaged
Mg=N I279.55=I744.23 3.5 4.7 1.2 4.1 11.8 to give just one representative result for each sample.
I280.27=I744.23 3.0 4.0 0.9 3.4 10.8 In addition, it should be stressed that the results
I285.21=I744.23 1.7 2.2 0.4 1.9 9.3 belonging to different samples cannot be compared
Si=N I285.21=I744.23 1.9 2.4 0.5 2.2 9.9
in between because of matrix effects. For that reason,
Al=N I307.47=I744.23 2.8 3.5 0.7 3.3 8.9
I308.85=I744.23 3.3 4.4 1.0 4.1 10.7
the following analysis is based on the relative varia-
Ca=N I393.37=I744.23 3.7 5.5 1.8 5.0 15.0 tions of the intensity ratios of the elements of interest
I396.85=I744.23 3.1 4.4 1.3 4.0 13.0 observed in the different samples.
Na=N I589.00=I744.23 4.7 5.7 1.0 5.1 8.2 The major components of biotite are Si, Al, and
I589.60=I744.23 4.3 5.0 0.7 4.7 5.6 Mg, while Fe, K, and Na are also present but at minor
K=N I766.49=I744.23 1.7 2.4 0.6 2.0 12.8
quantities. The presence of Si and Al indicates that
I768.90=I744.23 1.5 2.0 0.4 1.7 11.3
this mineral is an aluminum silicate, while K, Mg,

D. M. Dı´az Pace et al. 408

FIGURE 8 Averaged peak-intensity ratios obtained for minerals and rocks collected from the field. In the case of tonalite, the maximum
and minimum values are shown. The error bars correspond to the standard deviations.

and Fe complete their elemental composition. In deposits by weathering processes affecting cal-
addition, a little quantity of K is replaced by Na. careous rocks containing copper sulfurs. The sample
Microcline contains Al, Si, and Na as major element analyzed is impure. We speculate that, in addition to
constituents, together with K and Ca, corresponding Cu (not shown in Fig. 8), it contains significant
to a variety of K feldspar (or alkaline feldspar) usually amounts of Si and Ca probably because they are often
found in pegmatite rocks, from which the sample found together with the mineral in the ganga.[28,29] Fe
seems to come. It is an aluminum silicate with a sig- could be due to limonite, which is formed in the
nificant amount of Na because the mineral was upper part of the oxidation zone (gossan), and which
sampled transversely to perthite lamellae, a structure forms a thin film that covers the malachite. Addition-
of layers of the mineral albite (Na-plagioclase) that ally, it is presumed that Mg, Na, and K were detected
is the Na-rich end of the Ca-Na feldspar series. At as impurities of the adjacent wall rock.
lower temperatures, the K feldspars separate the Na, Quartz sample collected from the field is a good
which crystallizes with AlSi3O8, forming, in this case, example of how LIBS can perform a depth-profiling
layers of albite inside the K-feldspar crystal, in a analysis (Fig. 5). The Si emission line, which was
process called ‘‘exsolution.’’ Moreover, it frequently hidden and hardly identified, could be detected cleanly
contains some amount of Ca since it is also a solid sol- after a coating of Fe and Mg oxides was identified and
ution of the plagioclase feldspar series. removed to obtain the underlying bulk composition.
Malachite is a secondary mineral formed typically The results obtained from basalt suggest a nearly
in the superficial oxidation zones of copper ore uniform distribution of its chemical constituents,

409 Analysis of Minerals

indicating a fine-grained volcanic rock where the collected from the field and a soil sample in the form
spectra recorded from each sampling spot involve of a pressed pellet. We measured spectral lines
several minerals of microscopic size. The strong pres- characteristic of the major components of the Earth’s
ence of Na, Ca, and Al, together with Si, indicates that crust including Si, Al, Fe, Ca, Na, K, and Mg and lines
the plagioclase of this igneous rock has a near inter- from other elements of interest, that is, C, Cu, Mn,
mediate composition (andesine–labradorite) that and Ti, and from the atmoshere, that is, N, O, and H.
defines an andesitic–basaltic rock; the other chemical The LIBS capability to drill through a weathered
elements (Mg, Fe) are usually present in ferromagne- layer allowed determining the character of the oxides
sian minerals (pyroxenes, amphiboles) formed at covering superficially a quartz fragment as well as its
higher temperatures, and K belongs to an amphibole. bulk composition. The influence of the samples’ fea-
The spectra recorded from tonalite shows a het- tures, such as roughness, and the accuracy of the mea-
erogeneous distribution of the different chemical ele- surements were also investigated. Thus, in natural
ments. Fe and Mg, constituents of biotite and minerals the spectra were averaged over different loca-
hornblende, are not detected in some positions of tions to determine their composition, and in rocks their
the grid, as shown in Fig. 6, indicating a grainy rock intrinsic characteristics of inhomogeneity and=or grain
where the felsic minerals (plagioclase, K feldspar, size were taken into account by carrying out a grid of
quartz) have a grain size greater that the laser spot. measurements to obtain the overall composition. After
Moreover, the larger amount of Na with respect to that, an internal standard normalization with respect to
Ca indicates that the sample is an acid plagioclase, a proper N emission line was carried out to compen-
K is due to an alkali feldspar, and the lack of Si seems sate laser energy variations. Since matrix effects pre-
to show that the rock is close to diorite. vent the comparison of absolute intensity ratios, a
The loess sample has the typical elements of compositional analysis was performed by studying
igneous rocks. In fact, the mineral components char- the relative variations of line emission ratios of the
acteristics of the Argentine pampean loess are andes- chemical elements between the different samples.
itic plagioclases, quartz, orthoclase, volcanic glass The results obtained allowed the characterization of
shards, altered feldspars, volcanic rock fragments, the samples and illustrated the importance of some
organic opal, iron ores (magnetite, hematite, ilmenite, relevant experimental issues for LIBS geochemical
etc.), illite, amphiboles, and pyroxenes.[30,33] The pam- applications. Nevertheless, the LIBS measurements
pean loess is formed by eolic silts from the Tertiary and presented provide a basis for discrimination of some
Quaternary Ages. These sediments came from dry but not all of the samples. The discrimination can be
areas at the Patagonia and the Andes Range mixed improved using different experimental conditions or
with volcanic ash from the Andean Arc, which had a spectral analysis techniques, such as statistical multi-
major role in the formation of loess. It suffered a sub- variate analyses. Moreover, stricter research of the
sequent fluvial transformation, giving rise to loessoid issues of line sensitivity, matrix effects, ionization
mud sediments. Additionally, Ti (not shown in Fig. 8) degree of the elements in the plasma, and self-
comes from the titaniferous-variety magnetite.[30–33] absorption should be accomplished in order to achieve
Large amounts of Ca and Mg were detected in a semiquantitative or quantitative analytical determi-
caliche, locally called ‘‘tosca,’’ because it is a carbonate nation. In this case, the compositions obtained by LIBS
concretion developed in the pampean loess by adding should be compared to those obtained by other stan-
grains from its supporting material. In addition, the dard techniques such as X-Ray Fluorescence (XRF),
elements of loess are found as impurities. inductively coupled plasma-atomic emission spec-
trometry (ICP-AES), and inductively coupled plasma-
mass spectrometry (ICP-MS).
CONCLUSIONS
The analysis of geological samples was carried out
by using a compact, laboratory-based LIBS system in
ACKNOWLEDGMENTS
open-path configuration equipped with a noninten- The authors thank Dr H. F. Ranea Sandoval and M.
sified, nongated detector. The samples analyzed Giaconi for their valuable suggestions about this
were purchased gems, minerals, and rock fragments work.

D. M. Dı´az Pace et al. 410

 REFERENCES 16. McMillan, N. J.; Harmon, R. S.; De Lucia, F. C.; Miziolek, A. M.
Laser-induced breakdown spectroscopy analysis of minerals: Carbo-
nates and silicates. Spectrochim. Acta Part B 2007, 62, 1528–1536.
1. Tognoni, E.; Palleschi, V.; Corsi, M.; Cristoforetti, G. Quantitative
17. Yamamoto, K. Y.; Cremers, D. A.; Ferris, M. J.; Foster, L. E. Detection
microanalysis by laser-induced breakdown spectroscopy: A review
of metals in the environment using a portable Laser-Induced Break-
of the experimental approaches. Spectrochim. Acta Part B 2002,
down Spectroscopy instrument. Applied Spectroscopy 1996, 50,
57, 1115–1130.
222–233.
2. Winefordner, J. D.; Gornushkin, I. B.; Correll, T.; Gibb, E.; Smith, B.
18. Gaft, M.; Sapir-Sofer, I.; Modiano, H.; Stana, R. Laser induced break-
W.; Omenetto, N. Comparing several atomic spectrometric methods
down spectroscopy for bulk minerals online analyses. Spectrochim.
to the super stars: Special emphasis on laser induced breakdown
Acta Part B 2007, 62, 1496–1503.
spectrometry, LIBS, a future super star. J. Anal. At. Spectrom.
19. Ocean Optics (Dunedin, United States) at http://www.oceanoptics.
2004, 19, 1061–1083.
com/Products/libs.asp
3. Lazic, V.; Colao, F.; Fantoni, R.; Spizzicchino, V. Recognition of arche-
20. Marwan Technology (Pisa, Italy) at http://www.marwan-technology.
ological materials underwater by laser induced breakdown spec-
com/english/prodotti-modi.htm
troscopy. Spectrochim. Acta Part B 2005, 60, 1014–1024.
21. Sirven, J. B.; Bousquet, B.; Canioni, L.; Sarger, L. Time-resolved and
4. López, A. J.; Nicolás, G.; Mateo, M. P.; Piñón, V.; Tobar, M. J.; Ramil,
time-integrated single-shot laser-induced plasma experiments using
A. Compositional analysis of Hispanic Terra Sigillata by laser-
nanosecond and femtosecond laser pulses. Spectrochim. Acta Part
induced breakdown spectroscopy. Spectrochim. Acta Part B 2005,
B 2004, 59, 1033–1039.
60, 1149–1154.
22. Cristoforetti, G.; Legnaioli, S.; Palleschi, V.; Salvetti, A.; Tognoni, E.;
5. Vieitez, M. O.; Hedberg, J.; Launila, O.; Berg, L. E. Elemental analysis
Benedetti, P. A.; Brioschi, F.; Ferraro, F. Quantitative analysis of alu-
of steel scrap metals and minerals by laser-induced breakdown spec-
minium alloys by low-energy, high-repetition rate laser-induced
troscopy. Spectrochim. Acta Part B 2005, 60, 920–925.
breakdown spectroscopy. J. Anal. At. Spectrom. 2006, 21, 697–702.
6. De Giacomo, A.; Dell’Aglio, M.; De Pascale, O.; Longo, S.; Capitelli,
23. Barbieri Gonzaga, F.; Pasquini, C. A Complementary Metal Oxide
M. Laser induced breakdown spectroscopy on meteorites. Spectro-
Semiconductor sensor array based detection system for Laser Induced
chim. Acta Part B 2007, 62, 1606–1611.
Breakdown Spectroscopy: Evaluation of calibration strategies and
7. Pérez-Serradilla, J. A.; Jurado-López, A.; Luque de Castro, M. D.
application for manganese determination in steel. Spectrochim. Acta
Complementarity of XRFS and LIBS for corrosion studies. Talanta
Part B 2008, 63, 56–63.
2007, 71, 97–102.
24. González Bonorino, F. Introducción a la Geoquı´mica; Programa
8. Sallé, B.; Mauchien, P.; Maurice, S. Laser-Induced Breakdown Spec-
Regional de Desarrollo Cientı́fico y Tecnológico, Dpto de Asuntos
troscopy in open-path configuration for the analysis of distant
Cientı́ficos, Sec. Gral. De la O.E.A.: Washington, DC, 1972.
objects. Spectrochim. Acta Part B 2007, 62, 739–768.
25. Harris, R. D.; Cremers, D. A.; Ebinger, M. H.; Bluhm, B. K. Determi-
9. Yoon, Y. Y.; Kim, T. S.; Chung, K. S.; Lee, K. Y.; Lee, G. H. Appli-
nation of nitrogen in sand using Laser-Induced Breakdown Spec-
cation of Laser Induced Plasma Spectroscopy to the analysis of rocks
troscopy. Applied Spectroscopy 2004, 58, 770–775.
samples. Analyst 1997, 122, 1223–1227.
26. Baudelet, M.; Boueri, M.; Yu, J.; Mao, S. S.; Piscitelli, V.; Mao, X.;
10. Anzano, J. M.; Villoria, M. A.; Ruı́z-Medina, A.; Lasheras, R. J.
Russo, R. E. Time-resolved ultraviolet laser-induced breakdown spec-
Laser-induced breakdown spectroscopy for quantitative spectroche-
troscopy for organic material analysis. Spectrochim. Acta Part B 2007,
mical analysis of geological materials: Effects of the matrix and
62, 1329–1334.
simultaneous determination. Anal. Chim. Acta 2006, 575, 230–235.
27. De Giacomo, A.; Dell’Aglio, M.; Gaudiuso, R.; Cristoforetti, G.;
11. Laville, S.; Sabsabi, M.; Doucet, F. R. Multi-elemental analysis of sol-
Legnaioli, S.; Palleschi, V.; Tognoni, E. Spatial distribution of hydrogen
idified mineral melt samples by Laser-Induced Breakdown Spec-
and other emitters in aluminum laser-induced plasma in air and conse-
troscopy coupled with a linear multivariate calibration.
quences on spatially integrated Laser-Induced Breakdown Spectroscopy
Spectrochim. Acta Part B 2007, 62, 1557–1566.
measurements. Spectrochim. Acta Part B 2008, 63, 980–987.
12. Harmon, R.; De Lucia, F.; Mc Manus, C.; Mc Millan, N.; Jenkins, T.;
28. Betejtin, A. Curso de Mineralogı´a; 3rd Ed.; Mir: Moscow, Russia,
Walsh, M.; Miziolek, A. Laser-induced breakdown spectroscopy: An
1977.
emerging chemical sensor technology for real-time field-portable,
29. Hulburt, C. S.Jr.; Klein, C. Manual de Mineralogı´a de Dana; 3rd Ed.;
geochemical, mineralogical, and environmental applications. Applied
Reverté: Alexandria, VA, USA, 1982.
Geochem. 2006, 21, 730–747.
30. Teruggi, M. E. The nature and origin of Argentine Loess. Journal of
13. Palanco, S.; López-Moreno, C.; Laserna, J. J. Design, construction and
Sedimentary Petrology 1957, 27, 322–332.
assessment of a field-deployable laser-induced breakdown spec-
31. Zárate, M. A. Loess of Southern South América. In Encyclopedia of
trometer for remote elemental sensing. Spectrochim. Acta Part B
Quaternary Sciences; Elias, S., Ed.; Elsevier Ltd.: Burlington, MA,
2006, 61, 88–95.
2007, pp. 1466–1479.
14. Sallé, B.; Cremers, D. A.; Maurice, S.; Wiens, R. C. Laser-induced
32. Etchichury, M. C.; Tofalo, O. R. Mineralogı́a de arenas y limos en sue-
breakdown spectroscopy for space exploration applications: Influ-
los, sedimentos fluviales eólicos actuales del sector austral de la
ence of the ambient pressure on the calibration curves prepared from
cuenca Chacoparanaense: Regionalización y áreas de aporte. Revista
soil and clay samples. Spectrochim. Acta Part B 2005, 60, 479–490.
de la Asociación Geológica Argentina 2004, 59, 317–329.
15. Ferrero, A.; Lucena, P.; Herrera, R. G.; Doña, A.; Fernández-Reyes,
33. Teruggi, M. E. Apéndice. In Loess y limos pampeanos; Frenguelli, J.,
R.;. Laserna, J. J. Libraries for spectrum identification: Method of nor-
Ed.; Universidad Nacional de la Plata. Fac. de Cs. Naturales y Museo,
malized coordinates versus linear correlation. Spectrochim. Acta Part
Serie Técnica y Didáctica N 7, 1955, 85–88.
B 2008, 63, 383–388.

411 Analysis of Minerals