Analysis of rock samples collected from rock hewn

Journal of Archaeological Science 40 (2013) 2570e2578
Contents lists available at SciVerse ScienceDirect
Journal of Archaeological Science
journal homepage: http://www.elsevier.com/locate/jas
Analysis of rock samples collected from rock hewn churches of Lalibela, Ethiopia
using laser-induced breakdown spectroscopy
Alemayehu Kiros a, *, Violeta Lazic b, Giovanni E. Gigante c, A.V. Gholap a
a
Addis Ababa University, Science Faculty, Physics Department, P.O. Box 1176, Addis Ababa, Ethiopia
ENEA, UTAPRAD-DIM, Via E. Fermi 45, Frascati (RM), Italy
c
Università di Roma “La Sapienza”, Rome, Italy
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 October 2012
Received in revised form
18 January 2013
Accepted 24 January 2013
With the aim to study alteration processes of the rock hewn churches from Lalibela (Ethiopia), we
applied Laser Induced Breakdown Spectroscopy (LIBS) technique to measure the elemental composition
both of the bulk rock materials and their external layers, exposed to the environmental factors. The
analytical plasma was generated by nanosecond pulses of an Nd: YAG laser emitting at 1064 nm. Different major and minor sample constituents were detected, including Ca, Mg, Na, Fe, Ti, Al and K. The
detected O emission originates both from air surrounding and the sample, while the intensity of N lines,
coming exclusively from air, was used for the LIBS signal normalization. By depth profiling of the
weathered basalt rock, we observed a lower presence of K in the external layers, corresponding to the
first 5 laser shots. The emission from this element is anti-correlated with the line intensities from O, and
this was attributed to the variations in relative abundances of clay minerals and K-feldspar. The analogue
measurements were performed on the tuff rock, and compared to the spectra from powder samples
containing only the external soft material, scratched from the rocks. These analyses show an abundance
of H in the weathered, wetted layers and suggest that cations are lost from the constituent primary
minerals and replaced by Hþ; this process disrupts the lattice structure and causes a marked loss of
strength. The studies presented here demonstrate that LIBS is a useful technique for studying the
alteration processes in the rocks, caused by environmental factors.
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Laser-induced breakdown spectroscopy
(LIBS)
Encrustation
Depth profiling
Rock hewn churches
1. Introduction
Ethiopia is one of the countries in the world known for its
ancient rock-hewn churches. The largest groups of the rock hewn
churches are situated in central and eastern Tigray (Asrat, 2002)
and Lalibela. The latter location is famous for its extraordinary
complex of monolithic churches, carved about 800 years ago during
the Empire of King Lalibela. The main cluster of 11 churches is
located in the middle of the village and they are: Bete Medhane
Alem, Bete Michael Golgotha, Bete Maryam, Bete Meskel, Bete
Danaghel, Bete Amanuel, Bete Merkorios, Bete Aba Libanos, Bete
Gebriel Rufael, and Bete Giyorgis. Since 1978, these churches and
their surrounding area have been included in UNESCO’s World
Heritage List (Asrat and Ayallew, 2011) (Fig. 1). The churches are still
used daily for religious practices and ceremonies, while during
major religious occasions large crowds of believers and pilgrims
travel to the site. The churches have been exposed to different
* Corresponding author. Tel.: þ251 911863347; fax: þ251 111223931.
E-mail address: ftkiros@gmail.com (A. Kiros).
0305-4403/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jas.2013.01.028
environmental factors, both physical (temperature and humidity)
and biological (biodeteriogens), and also to human impact (pollutant releases) (Delmonaco et al., 2009). Water absorption represents the main factor contributing to a deterioration of the exposed
rock materials and structures, in terms of direct rainfall, soil infiltration, capillarity and diffuse humidity, for the alteration of basalt
(Delmonaco et al., 2009). Another study from Asrat and Ayallew
(2011) reported material loss due to deep weathering activated
by rain water penetration, while the leakage of groundwater affects
most of the Lalibela churches too. This exposure has resulted in
severe degradation of the churches, most of which are now considered to be in critical conditions. According to the state party
report, presently the conservation and restoration of rock hewn
churches of Lalibela is one of the main world concerns. The first
restoration attempt of the Lalibela rock-hewn churches was carried
out in 1920 (Delmonaco et al., 2009). Currently, most of the
churches are protected by shelters in order to prevent the erosive
effects of rainfall, especially on the roof rock cover, and by temporary scaffolding to prevent collapse of the most exposed
structures.
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
2571
Fig. 1. a) Location of the Lalibela site (UNESCO); b) distribution of the 11 rock-hewn churches; powder samples collected from c) Bete Gebriel Rufael-internal; d) Bete Gebriel Rufael
(BGR-E); e) Bete Giyorgis-Mahilet (BG-I); f) Bete Giyorgis-external; g) Bete Michael Golgotha.
During weathering of basaltic rocks induced by hydrothermal
processes, different major and trace elements in the material are
mobilized from the original rock (Ignacio et al., 2007; Aiuppa et al.,
2000; Greenough et al., 1990; Lagat, 2007; Franzson et al., 2010). In
another study (Gan-Lin et al., 2007), Fe and Al were relatively
enriched while Ca, Mg, K and Na and Si were strongly diminished
during the rock weathering and the corresponding soil formation.
In many cases the presence of biological colonization by lichens
and macroscopic plants is quite evident. For the study of biological
aspects of weathering of rocks and minerals, lichen-encrusted
rocks provide an ideal environment due to a direct relationship
between lichens and their substrates (Adamou and Violante, 2000).
Lichens alter the mineral substrates both through physical and
chemical processes. Lichens also have a significant impact in the
chemical weathering of rocks by the secretion of various organic
acids, particularly oxalic acid, which can effectively dissolve minerals and chelae metallic cations (Adamou and Violante, 2000;
Chen et al., 2000). As a result of the weathering induced by lichens,
many rock-forming minerals exhibit extensive surface corrosion.
LIBS is an analytical technique that enables the determination of
the elemental composition of materials on the basis of the characteristic atomic emission from a micro-plasma produced by
focusing a high-power laser on or inside a material. LIBS have been
used in a wide variety of analytical applications for the qualitative,
semi-quantitative or comparative analysis. Analysis can be carried
out without physical contact with the examined sample, since only
2572
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
optical access is necessary. The sensitivity is sufficient (in nanograms range), so once the current problem of experimental setup
optimization is solved, numerous applications become possible.
LIBS technique found many applications in the field of geology,
archeology and cultural heritage analysis due to its low invasiveness, the possibility to perform in situ measurements, high spatial
discrimination, rapidity and capability for direct analysis without
any sample preparation. Detection of elemental sample composition regarding both major and minor constituents and some trace
elements too was demonstrated on different minerals, rocks and
soil or sediment samples. Spectral lines from Si, Al, Fe, Ca, Na, K, Mg,
C, Cu, Mn, and Ti were detected in caliche, tonalit, and basalt.
Classification of different basaltic materials and starting from the
LIBS spectra was also investigated (Perkins et al., 2010; Clegg et al.,
2009) by using the principal component vectors (PCA) and partial
least squares regression methods. Classification of unknown materials could be obtained also by building initially the spectral libraries for different samples and then comparing the
measurements with reference spectra (McMillan et al., 2007;
Jurado-Lopez and Castro, 2003; Cousin et al., 2012), for example
through linear correlation between the line emission intensities.
LIBS can also be used for in-situ depth profiling where the elemental distribution of weathered rock or marbles can be retrieved
(Maravelaki-Kalaitzakia et al., 2001; Lazic et al., 2004). For example,
on weathered marble encrustation a progressive decrease in depth
of Fe, Si, Al and Ti content relative to Ca indicate surface contamination; the spectroscopic data also show when the bulk, unaltered sample layer is reached by the laser pulses (Diaz Pace et al.,
2011; Gaft et al., 2009; Gondal et al., 2009; Salle et al., 2005).
The aim of this work was to carry out a rapid LIBS analysis of
major and minor elements present in samples from three Lalibela
rock hewn churches. The measurements regard the samples collected from internal and external walls of the churches and from
two freshly cut surfaces. The method was validated by the comparison with XRD results.
1.1. Climatic and geological setting
The Lalibela area has two rainy seasons from June to September
and from March to April. The annual rainfall varies from 500 to
1000 mm, which is believed to be the main direct/indirect cause of
stone deterioration (Smith et al., 2008; McCabe et al., 2011). Lalibela
area is characterized by mean annual temperature of 24.5 C and
relative humidity is 52.9% (Asrat and Ayallew, 2011; Delmonaco
et al., 2009).
Northern Ethiopia’s geology is characterized by thick sequences
of tholeiitic to transitional continental flood basalts overlain by
minor rhyolitic-trachytic lavas and pyroclastic rocks of Oligocene to
Miocene belonging to the Northern Ethiopia Plateau (Beccaluva
et al., 2009; Natali et al., 2011; Asrat and Ayallew, 2011). Within
the plateau, three magma types have been distinguished: two highTi groups (HT1 andHT2), and one low-Ti group (LT). The centraleastern sector of the plateau in particular, i.e. the one comprising
the Lalibela area and facing the Afar triangle, is mainly characterized by a 1700 m-thick sequence of high TiO2 (HT2 magma type)
picrite/basalt lavas capped by about 300 m of rhyolites, linked to
the magmatic activity at the Afar plume axial zone (Beccaluva et al.,
2009). The HT2 volcanics (Pik et al., 1998) are represented by subalkaline olivine-clinopyroxene aphyric or glomerophyric basalts
and picrites with phenocrysts of euhedral olivine (FO86eFO81
range) and of MgeTieAl rich augite. The groundmass contains
olivine, clinopyroxene, FeeTi oxides and plagioclase (An85eAn48
range) microlites. The Oligocene volcanism of the northwestern
Ethiopian Plateau has been subdivided into three formations (Berhe
et al., 1987); the Ashangi and Amba Aiba basaltic units, separated by
an angular unconformity (Zanettin, 1992) from the upper ignimbritic Alaji unit. In the Lalibela monumental site, the Amba Aiba
basalts are typically overlain by a thick horizon of reddish highly
hydrothermally weathered basaltic scoria (commonly referred to as
“tuffaceous material”). It is within this “softer” unit that the rockhewn churches were carved.
2. Materials and methods
2.1. Experimental setup
The LIBS plasma was generated by one of the twin lasers
(Quantel, Twins Ultra 200), emitting 6.5 ns long pulses at 1064 nm.
The beam diameter is 6 mm and with the flat intensity profile,
which is important to obtain better resolution during the depth
profiling. The laser beam is passed through a pierced aluminum
mirror and focused by a quartz lens f ¼ 75 mm onto a sample. The
laser energy measured after the mirror was 210 mJ. A sample was
mounted on an XYZ positioning system and placed above the focal
plane in order to obtain a larger ablation area and to reduce the
effect of sample in homogeneities. During the measurements, the
laser flash lamp was running at 1 Hz while the Q-Switch trigger was
externally controlled in order to generate a single pulse after repositioning of a sample. The distance between two sampling
spots was 2 mm.
The plasma emission was collected by two optical systems: one
for the radiation reflected by the pierced mirror and focused onto
one fiber bundle by quartz, and another one, containing two quartz
lenses, was mounted at angle of about 45 with respect to the laser
axis. Both collecting systems terminate with a fiber optic bundle
formed by three quartz fibers with diameter of 0.6 mm. At the other
end, the bundle was separated into single fibers, and each of them
was connected to one of the five spectrometer channels (Stellar
Net). Channels 1e4 have 0.1 nm spectral resolutions and cover
spectral range 200e600 nm. Channel 5 has lower resolution
(0.2 nm) and covers the spectral interval 600e800 nm. The co-axial
collection system was used for the spectral interval 200e500 nm
while the angular collection was covering range 500e800 nm.
Each channel is equipped with a 2048 Photo-Diode Array, and
was triggered externally by an optical trigger. The detectors do not
allow for time resolved measurements and the minimum integration time, here used, was 30 ms. The best LIBS signal from basaltic
samples was obtained for the acquisition from the laser pulse
corresponding to 4 ms. This delay, produced by signal/delay generator (Quantum Composer 9600þ) was used in all the measurements. After one laser shot the whole spectrum was saved for
further analysis by custom written programs under Lab view. This
approach allows monitoring of shot-to-shot variations in the
spectral intensities and also the depth profiling.
2.2. Samples preparation
Samples were collected from internal and external walls of three
Lalibela churches: Bete Michael Golgotha (Fig. 1g), Bete Giyorgis
(BG-E and BG-I) (Fig. 1f and e) and Bete Gebriel Rufael (BGR-E and
BGR-I (Fig. 1d and c)). For comparison we also performed the
measurements on two rock samples from the surrounding: one
from the tunnel in front of Bete Amanuel (BAT) and the other one
collected on the north side of Bete Giyorgis (BGN). The rock samples, cut to 1 1 cm2 large blocks, were directly interrogated by the
laser both on the weathered surface and on the bulk material
(lateral cut). The weathered deposits scratched from the surfaces in
the nearby locations were first pressed pellets into pellets and then
analyzed by LIBS. The samples analyzed in this study are listed in
Table 1.
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
Table 1
Sample analyzed and their locations.
Location
Bete Gebriel Rufael
Bete Giyorgis
Bete Michael Golgotha
Tunnel in front of
Bete Amanuel
North side of
Bete Giyorgis
Sample
name
Wall
(surface)
Sample type analyzed
Rock
Encrustation
(pellets)
BGR-E
BGR-I
BG-E
BG-I
BMG-E
BAT-E1
External
Internal
External
Internal
External
External
Basalt
Tuff
Basalt
Basalt
Basalt
Tuff
X
X
X
X
X
BGN-E1
BGN-E2
External
External
Basalt
Basalt
3. Results and discussion
3.1. Analysis of samples
The LIBS signal from bulk (clean) stones was intense and almost
constant from one laser shot to another. For their analysis we
considered the average spectra obtained by sampling by five consecutive laser shots and in two different sites (total ten spectra). The
powder pellets, although obtained by mixing the powder, show
visible in-homogeneities with yellowish spots due to presence of
lichens. In this case, the spectra used in analysis were averaged over
ten consecutive shots applied at four different sites (total forty
spectra). The spectral lines considered are listed in Table 2 and
typical examples of the spectra measured from encrustation
(powder) and bulk (clean) rock samples, with the most intense
lines identified, are shown in Figs. 2 and 3.
Our previous XRD data on whole rock samples confirm the
presence of peaks attributable to zeolites (in order of peak strength:
analcime, heulandite, thomsonite, wairakite, pollucite, clinoptilite,
scolecite), clinopyroxene, hematite, calcite, magnetite, pyroxene,
sanidine, plagioclase and minor smectites.
3.2. Characterization of LIBS precision
Intensities of the overall LIBS spectra were fluctuating from one
laser shot to another. In the case of powder sample, these fluctuations might be attributed to a non-uniform sample composition, as
mentioned before, while the laser pulse energy is stable inside 5%.
When analyzing the rock samples, beside the sample inhomogeneities, the LIBS signal is also affected by the surface
roughness and grain size. In order to minimize effects of the signal
instability, in following we consider the line intensities normalized
on the intensity of Ca I line at 422.8 nm. This line does not exhibit
saturation and it is present in all the acquired spectra. The other
detected lines for basalt, tuff and powder samples below to Mg, Si,
Al, Na, Ti, Ba, Fe, H and K. The emissions from N and O were also
Table 2
Elements detected from the encrustation and the bulk rocks.
Wavelength (nm)
Element
Wavelength (nm)
Element
274.63
279.27
284.95
287.88
315.96
318
323.5
407.31
407.88
Fe(II)
Mg(II)
Mg(I)
Si(I)
Ca(II)
Ca(II)
Ti(II)
Fe(I)
Sr(II)
422.81
643.8
656.4
670.7
766.4
777.26
396.18
589.01
Ca(I)
Ca(I)
H
Li(I)
K(I)
O
Al(I)
Na(I)
2573
observed, where the latter element originates both from the surrounding air and the samples.
Tuff is a volcanic indurate hardened by temperature, pressure,
and/or chemical reaction volcanic deposit. It is predominantly
composed of volcanic ash. The measured RSD of the normalized
lines, basing on ten spectra, was between 7.5% for Si and 28.7% for K,
indicating that the latter element is less uniformly distributed in
the sample (Fig. 4). The examined basalt rocks were fine-grained
fragment with a high degree of surface roughness. On the basalt,
the peak-intensity ratios have much higher RSD than in the case of
tuff, indicating less uniform composition (Fig. 4). Except for Ba, the
RSD was about 15%, all the other elements exhibit fluctuations with
RSD above 22%.
LIBS spectra of powder samples collected from external and
internal walls of Bete Gebriel Rufael have the highest RSD for Na
and K, with RSD up to 28% and for Al and Si up to 41% (Fig. 4),
sample Bete Giyorgis-Mahilet (not shown in Fig. 4) is the most
inhomogeneous with RSD above 29.1%. In particular, Na and K
distributions were very inhomogeneous, producing RSD of 84 and
86% respectively. The sources of in-homogeneities of these samples
are the deposits with yellowish, green and black spots due to
presence of lichens, green algae and black crust as shown in (Fig. 1).
3.3. Composition differences of the powder samples and the bulk
rock
To compare the composition of bulk and powder samples, we
have normalized all line intensity on N line at 744.2 nm. Table 3
shows the composition of samples listed in Table 1 relative to the
bulk materials. In the powder sample Bete Giyorgis-external, the
average intensity ratios of almost all the elements are lower with
respect to the basalt. The peak intensities of Mg and Fe measured on
the samples collected from external walls coincide with the same
from the bulk basalt. Differently, the sample belonging to Bete
Gebriel Rufael showed a higher content of Fe in the external layers.
Samples collected from interior walls show a higher presence of Fe
and Mg than the bulk basalt due to oxidizing conditions at the time
of the rock alteration (Maravelaki-Kalaitzakia et al., 2001). The
normalized emission peaks from the other detected elements are
relatively well correlated, as for instance Si and Ti in samples collected from the inner and outer walls of Bete Giyorgis respectively.
The decreasing ratio of Na and K for all the samples of powder
implies that these elements are depleted with respect to basalt.
In Table 3b the LIBS major elements composition of the samples
representing tuff and powder samples is given. Coarse differences
in composition are clear between the sample Tuff, and the others.
The high (Fe, O, H peaks) of samples collected from internal and
external of Bete Gabriel Rufael (BGR-I and BGR-E) relative to tuff
may be indicates the presence of Fe-rich oxide-hydroxides (hematite, magnetite and lepidocrocite).
The normalized peak intensity of Ca, Na and K of the powder
samples collected from Bete Michael Golgotha-external coincided
with Bete Gebriel Rufael-external, although being lower than for
the bulk rocks. This indicates similar composition of these samples;
moreover, these elements are depleted relative to the bulk rocks.
The normalize intensity of Al of Bete Gebriel Rufael-external
matched with Bete Giyorgis-external, and Bete Gebriel Rufaelinternal matched with Bete Giyorgis-internal while the ratio of Si
has a constant value for all samples collected from the outer wall of
the churches. This indicates constant deposit of aluminosilicates.
The variation in the distribution of Ca, Mg, Na and K in some
samples may be the monument mainly controlled by clay minerals.
Ca-rich plagioclase dissolved in organic acids more readily done
Na-rich plagioclase, while Na-rich plagioclases were more soluble
in water (Lopez et al., 2006). The samples collected from the outer
Fig. 2. aed. Average spectra of bulk tuff (lateral cut) from the tunnel in front of Bete Amanuel, bulk basalt from North side of Bete Giyorgis and powder sample collected from
internal wall of Bete Gebriel Rufael, Lalibela.
Fig. 3. aed. Comparative average spectra of powder samples collected from internal and external walls of churches, Lalibela.
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
2575
Fig. 4. Relative Standard Deviation RSD% calculated from intensities measured for elements detected in LIBS rock spectra. Bulk samples, basalt and tuff; powder samples from
external and internal walls of Bete Gabriel Rufael.
wall of the churches were colonized by lichens, green algae and
black crust. Bacteria, fungi, lichen and plants (Adamou and Violante,
2000; Chen et al., 2000) have a significant impact in the chemical
weathering of rocks by the excretion of various organic acids, particularly oxalic acid, which can effectively dissolve minerals and
chelae metallic cations. Transformation of primary minerals to
secondary clay minerals formed in the weathering crusts, is presumably related to the activity of organic acids (Adamou and
Violante, 2000; Chen et al., 2000). The decreasing ratio of Ca in
Table 3
Encrustation of samples, elements is normalized to N at 744.23 nm. (L) Indicates that
the element intensity ratio of the powder samples, lower than sample basalt (bulk)
Table 3a and tuff Table 3b, (H) Indicates the powder samples higher than sample
basalt and tuff and (C) constant value.
Sample
a)
BGR-E
BGR-I
BG-E
BG-I
BMG-E
b)
BGR-E
BGR-I
BG-E
BG-I
BMG-E
Element/N
Fe(II)
Mg(II)
Si
Ca(II)
Ti(II)
H
K
O
Al
Na
H
H
C
H
C
C
H
C
H
C
L
H
L
C
L
L
H
L
H
L
L
H
C
L
L
H
H
L
L
H
L
L
L
L
L
C
C
C
C
C
L
H
L
L
L
L
L
L
L
L
C
H
C
C
L
H
H
C
H
C
L
H
L
C
L
L
H
L
H
L
L
C
L
L
L
H
H
C
C
H
C
H
C
L
L
H
H
C
C
H
L
H
L
H
L
L
H
C
L
L
E e external, I e internal, BGR e Bete Gebriel Rufael, BG e Bete Giyorgis, BMG e Bete
Michael Golgotha.
powder samples collected from the outer wall of the churches relative to bulk samples (lateral cut) shown in anorthite dissolution
.The elements mobility aspect of the powder samples collected from
the outer wall of Bete Giyorgis, Bete Gabriel Rufael and Bete Michael
Golgotha are (K > Al > Na > Ca > Fe), (Na > Ca > K > Al > Fe) and
(Na > K > Ca > Al > Fe), respectively.
3.4. Depth profiling
Selected LIBS spectra obtained within the encrustation at different laser shots corresponding to the different depths from the
surface of weathered basalt are presented. Here, the considered
peak intensities in the individual spectra were been normalized in
the intensity of N line at 744.2 nm (Diaz Pace et al., 2011). Nitrogen
comes mainly from air and is suitable for the normalization of the
elements, and helps to see the differences clearly.
3.4.1. Weathered basalt
The surface of weathered basalt was in contact with percolating
rain water, and it is covered by lichens and black crust. The
encrustation was characterized by recording the LIBS spectra after
each laser shot during drilling a hole in the material. In this way it
was possible to observe surface impurities and ascertain their
disappearance in depth.
Fig. 5 shows the LIBS profiles of Fe, Ca, Ti, Al, Na and K for
weathered basalt. The normalized peaks of Fe, Ca, Ti, and Al show
a decreasing tendency in the range from 1 to 4 laser shots, followed
by constant trend except the maximum at sixth laser shot. However, the LIBS profiles of Na and K show their maximum intensities
2576
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
Fig. 5. LIBS measured relative element distribution inside the sample weathered basalt.
in the range from 5 to 7 laser shots, while out of this interval their
intensities are almost constant.
As mentioned above a significant decrease of Fe (II) can be
observed for the weathered basalt with increasing the ablation
depth. This element is mainly located in the surface crust and its
diminution of the normalized intensity with the ablation proceeds
toward the substrate. A higher presence of Fe in the top encrustation layers is a consequence of oxidizing conditions at the time of
alteration of the parent rock crystal (Maravelaki-Kalaitzakia et al.,
2001).
Depletion of the emission intensities of Na and K on the surface
of weathered basalt, compared to the bulk material (lateral cut)
indicates a loss of the mineral feldspars. The poor correlation between the line intensities from Na with respect to other metals (Al,
Mg, Ca, Ti, Fe, Ba, Mn and Sr) suggests that Na may be removed by
dissolution into water. Here, this corresponds to the profiling by the
first 4 laser shots and implies the greatest contribution of clay
minerals (for instance kaolinite). In the interval (5e7) and (8e10)
laser shots the emission from Na is correlated with Al; this may
be indicates the presence of Na-plagioclase in depth.
Lower content of K on the surface (the first 1e4 laser shot)
together with a progressively increasing oxygen emission in depth
was observed. In the range 5e7 laser shot there is an increase of K
emission with simultaneous lowering of O emission. Between 8 and
10 shots K emission decreases again while O shows the opposite
behavior. A strong inverse association between K and O observed
here may reflect the variations in clay minerals and K-feldspar
present in depth and on the surface of weathered basalt.
We compared the intensity ratio of hydrogen on the powder,
bulk rocks and during the depth profiling of the weathered basalt
samples. Hydrogen emission peak (656.4 nm) was first normalized
on the nearby line intensity of Ca (643.8 nm), belonging to the same
spectrometer channel. The measured H/Ca intensity ratio is higher
in the samples collected from external wall of the churches i.e.
exposed to the wetting. The emission from H in weathered basalt
was detected for all laser shots, with the maximum corresponding
to the seventh shot. The presence of hydrogen is the indicator of
humidity in dirty layer (Clegg et al., 2009). The measured variations
of H emission with the depth in the weathered basalt in-depth,
together with its high detected presence in the powder samples,
is an indicator of dissolving feldspars into secondary minerals, as for
example kaolinite (Madejova and Komadel, 2001) and montmorillonite (Delmonaco et al., 2009), caused by hydration of inorganic
cations. The presence of kaolinite and montmorillonite were confirmed in our previous FTIR and XRD work. Kaolinite detected only
at the surface and montmorillonite at the characteristic spectrum
n ¼ 3435 cm1 in depth up to 20 mm. Montmorillonite is a member
of the general mineral group of clays, typically the smectite family.
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
Table 4
Result of correlation coefficient between the selected line intensities of different
samples, collected from rock hewn churches of Lalibela.
Sample
Tuff
Basalt
BGR-E
BGR-I
BG-E
BG-I
BMG-E
1
0.89328
0.72649
0.87257
0.80562
0.86781
0.74997
Tuff
1
0.84112
0.85304
0.9335
0.90659
0.87319
Basalt
1
0.88465
0.96874
0.89451
0.99443
BGR-E
1
0.90548
0.96577
0.89688
BGR-I
1
0.9351
0.98573
BG-E
1
0.91065
BG-I
1
BMG-E
E e external, I e internal, BGR e Bete Gebriel Rufael, BG e Bete Giyorgis, BMG e Bete
Michael Golgotha.
2577
(Authority for Research and Conservation of Cultural Heritage) deposits like lichens and green algae were removed from the churches
using mechanical method. The use of rotating abrasive discs or
brushes to remove deposit is normally not advisable because they
are both inefficient and causing the additional damage.
In-situ investigation of the rock hewn churches of Lalibela using
LIBS (Salle et al., 2005) in depth plus removal of the biological and
other deposits with the non-destructive method (laser cleaning)
makes an important contribution to damage diagnosis, for the
cultural heritage sustainable conservation when coupled with understanding of the indicators of climatic and environmental change
(Smith et al., 2008; McCabe et al., 2011).
3.5. Sample classification
Chemically it is hydrated sodium calcium aluminum magnesium
silicate hydroxide (Na, Ca)x(Al, Mg)2(Si4O10)(OH)2$nH2O. The
important aspect of the montmorillonite is the ability for H2O
molecules to be absorbed between the TetrahedraleOctahedrale
Tetrahedral (TeOeT) sheets, causing the volume of the minerals
to increase when they come in contact with water (Madejova and
Komadel, 2001). The kaolinite found on samples of basalt stone
could have the origin: deposition of kaolinite bearing airborne soil
dust and/or surface wash on the monument surface and natural
weathering in the original basalt stone, such as organic acids
excretion by the mycobiont of lichens. Lichens enhance the rate of
dissolution of plagioclase (Steven, 2005).
The identification of montmorillonite as a weathering product
up to 10 laser shot from the rock/atmosphere interface, suggests
that montmorillonite may penetrate even deeper into the rock
substrate causing major rock decay. The presence of swelling
montmorillonite could result from surface weathering or hydrothermal alteration and the inexorable lateritization processes,
typical of regions with dry and rainy seasons, are all consistent with
the deterioration of the rocks of Lalibela. The recognition of this
process therefore has important applications in the field of conservation science.
Major problems that are currently affecting the conservation of
the churches and the stability of the whole church complex are
cracks, and human impact due to uncontrolled urbanization, religious sermonic and unwise restoration works carried out on an
uncontrolled way. Biological attack by green algae and lichens is
also currently responsible for severe stone surface physical and
chemical weathering leading to considerable weakening of the
churches walls. According to the Department of ARCCH, Ethiopia
Comparison between the line intensities of the elements
forming samples allows classification of the mineral type
(McMillan et al., 2007). This is demonstrated in Table 4, where the
names and correlation coefficients for the rock samples with the
most related LIBS spectra are listed for the 7 analyzed samples (rock
and powder).
The spectral lines of the samples (powder) collected from the
outer wall of the rock-hewn churches of Lalibela correlate well each
other. The correlation coefficient between the line intensities
measured on the powder samples collected from the outer wall of
Bete Gebriel Rufael and Bete Michael Golgotha, (r ¼ 0.99443) is
given in Table 4. The association between the samples collected
from the inner walls also shows a strong correlation (r ¼ 0.96577)
however, the correlation coefficient for samples collected from the
same church is relatively low. For example, powder samples from
internal and external walls of Bete Gebriel Rufael have only
r ¼ 0.88465. This suggests that although the samples are colonized
by similar deposits (lichens, green algae, and various fungi), the
rock-forming minerals of these samples are transformed into secondary minerals and there might be a disparity in substrate composition of the rocks exterior and interior of the churches.
Fig. 6a shows coefficient of linear correlation (r2 ¼ 0.51566) of
line intensities from the selected elements measured on the outer
and inner walls of Bete Gebriel Rufael. The coefficient of correlation
is low, particularly due to difference in Na and Ca contents, higher
for the Bete Gebriel Rufael-internal (BGR-I). The peak intensities
clustered below the regression line show a relatively uniform distribution between the two samples. In Fig. 6b the line intensities
from different elements (Fe, Mg, H, O and Ti) detected in the bulk
Fig. 6. Correlation between selected peak line intensities, from the spectra of samples a) external and internal wall of Bete Gebriel Rufael; b) basalt (bulk) and Bete Giyorgis-external
(BG-E); the linear correlation coefficient r is given on the plot.
2578
A. Kiros et al. / Journal of Archaeological Science 40 (2013) 2570e2578
basalt and the same from the external wall of Bete Giyorgis, show
a relatively good correlation. In this comparative study LIBS line
intensity among the elements that constitute the samples help to
classify the type of mineral.
4. Conclusions
This study demonstrates that LIBS spectra can be used to match
rock samples collected from the inner and outer walls of rock-hewn
churches of Lalibela. Comparing mediated LIBS spectra against each
other and with tuff and basalt samples. The samples collected from
the outer wall of the churches are well correlated.
Selected LIBS spectra obtained within the encrustation at different laser shots corresponding to different depth from the surface
of weathered basalt is presented. A lower content of K on the surface together with increasing oxygen intensity in depth was
observed, except in the interval from 5 to 7 shot where K emission
has a maximum and O presence is low. Variations in depth of these
two elements, which are clearly anti-correlated, may reflect
changes in abundance of clay minerals and feldspar due to alteration of the basalt.
A causal relationship between the relative loss of cations and
higher presence of hydrogen in the samples collected from external
wall of the churches and in-depth profile of weathered basalt is
suggested. Cations are lost from the constituent primary minerals
and replaced by Hþ; this process disrupts the lattice structure and
causes a marked loss of strength. This study suggests that LIBS is
also the promise of identifying impurities characteristics of the
rocks and their deposits.
References
Adamou, P., Violante, P., 2000. Weathering of rocks and neogenesis of minerals
associated with lichen activity. Appl. Clay Sci. 16, 229e256.
Aiuppa, A., Allard, P., Alessandro, W., Michel, A., Parello, F., Treuil, M., Valenza, M.,
2000. Mobility and fluxes of major, minor and trace metals during basalt
weathering and groundwater transport at Mt. Etna volcano (Sicily). Cosmochimica Acta 64 (11), 1827e1841.
Asrat, A., 2002. The rock-hewn churches of Tigray, Northern Ethiopia: a geological
perspective. Geoarchaeol. Int. J., 649e663.
Asrat, A., Ayallew, Y., 2011. Geological and geotechnical properties of the medieval
rock hewn churches of Lalibela, northern Ethiopia. J. Afr. Earth Sci. 59, 61e73.
Beccaluva, L., Bianchini, G., Natali, C., Siena, F., 2009. Continental flood basalts and
mantle plumes: a case study of the northern Ethiopian Plateau. J. Petrol. 50 (7),
1377.
Berhe, S.M., Desta, B., Nicoletti, M., Teferra, M., 1987. Geology, geochronology and
geodynamic implications of the Cenozoic magmatic province in W and SE
Ethiopia. J. Geol. Soc. London 144, 213.
Chen, J., Blume, H.P., Beyer, L., 2000. Weathering of rocks induced by lichen colonization, a review. Catena 39, 121e146.
Clegg, M.S., Sklute, E., Dyar, M.D., Barefield, J.E., Wiens, R.C., 2009. Multivariate
analysis of remote laser-induced breakdown spectroscopy spectra using partial
least squares, principal component analysis and related techniques. Spectrochim. Acta B 64, 79e88.
Cousin, A., Forni1, O., Sautter, V., Fabre, C., Maurice, S., Gasnault, O., Wiens, R., 2012.
Classification of non-homogeneous basalts using independent component
analysis technique for MSL/ChemCam data. In: 43rd Lunar and Planetary Science Conference.
Delmonaco, G., Margottini, C., Spizzichino, D., 2009. Analysis of Rock Weathering
and Conservation Strategies for Rock-hewn Churches of Lalibela (Ethiopia).
Taylor & Francis Group, London,, ISBN 978-0-415-55803-7.
Diaz Pace, M.D., Gabriele, N.A., Garcimuno, M., Angelo, A.C., Bertuccelli, G.,
Bertuccelli, D., 2011. Analysis of minerals and rocks by laser-induced breakdown
spectroscopy. Spectrosc. Lett. 44, 399e411.
Franzson, H., Gudfinnsson, G.H., Helgadottir, H.M., 2010. Porosity Density and
Chemical Composition Relationships in Altered Icelandic Hyaloclastites. Taylor
& Francis Group, London, ISBN 978-0-415-60426-0.
Gaft, M., Nagli, L., Fasaki, I., Kompitsas, M., Wilsch, G., 2009. Laser-induced breakdown spectroscopy for on-line sulfur analyses of minerals in ambient conditions. Spectrochim. Acta B 64, 1098e1104.
Gan-Lin, Z., Ji-Hua, P., Cheng-Min, H., Zi-Tong, G., 2007. Geochemical features of
a soil chronosequence developed on basalt in Hainan Island, China. Revisita
Mexicana de Ciencias Geol. 24 (2), 261e269.
Gondal, A.M., Nasr, M.M., Ahmed, Z., Yamani, Z.H., 2009. Determination of trace
elements in volcanic rock samples collected from Cenozoic lava eruption sites
using LIBS. J. Environ. Sci. Health A 44, 528e535.
Greenough, D.J., Fryer, J.B., Robinson, T.P., 1990. Geochemical effects of alteration on
mafic rocks from Indian Ocean site 7061. Proc. Ocean Drill. Program Sci. Results
115.
Ignacio, S., Torres-Alvarado, Pandarinath, K., Verma, S.P., Dulski, P., 2007. Mineralogical and geochemical effects due to hydrothermal alteration in the Los
Azufres geothermal field, Mexico. Revisita Mexicana de Ciencias Geol. 24 (1),
15e24.
Jurado-Lopez, A., Luque de Castro, M.D., 2003. Rank correlation of laser-induced
breakdown spectroscopic data for the identification of alloys used in jewelry
manufacture. Spectrochim. Acta B 58, 1291e1299.
Lagat, J., 2007. Hydrothermal alteration mineralogy in geothermal fields with case
examples from Olkaria domes geothermal field, Kenya. Presented at Short
Course II on Surface Exploration for Geothermal Resources. Organized by UNUGTP and KenGen, at Lake Naivasha, Kenya, 2e17.
Lazic, V., Fantoni, R., Colao, F., Santagata, A., Moroneb, A., Spizzichin, 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, 429e436.
Lopez, P., Navarro, E., Marce, R., Ordonez, J., Caputo, L., Armengol, J., 2006. Elemental
ratios in sediments as indicators of ecological processes in Spanish reservoirs.
Limnetica 25 (1e2), 499e512.
Madejova, J., Komadel, P., 2001. Baseline studies of the Clay Minerals Society source
clays: infrared method. Clays Clay Miner. 49 (5), 410e432.
Maravelaki-Kalaitzakia, P., Anglos, D., Kilikoglouc, V., Zafiropulos, V., 2001. Compositional characterization of encrustation on marble with laser induced
breakdown spectroscopy. Spectrochim. Acta Part B 56, 887e903.
McCabe, S., Smith, B.J., Adamson, C., Mullan, D., McAlister, J., 2011. The “Greening” of
natural stone buildings: quartz sandstone performance as a secondary indicator
of climate change in the British Isles. Atmos. Clim. Sci. 1, 165e171.
McMillan, J.N., Harmon, R.S., De Lucia, F.C., Miziolek, A.M., 2007. Laser induced
breakdown spectroscopy analysis of minerals: carbonates and silicates. Spectrochim. Acta Part B 62, 1528e1536.
Natali, C., Beccaluva, L., Bianchini, G., Siena, F., 2011. Rhyolites associated to Ethiopian CFB: clues for initial rifting at the Afar plume axis. Earth Planet. Sci. Lett.
312, 59.
Perkins, J.J., Sharma1, S.K., Lienert, B.R., Misra1, A.K., Clegg, S.M., Wiens, R.C., 2010.
Improvement in qualitative and quantitative LIBS analysis of elemental compositions of basalts. In: 41st Lunar and Planetary Science Conference.
Pik, R., Deniel, C., Coulon, C., Yirgu, G., Hofmann, C., Ayalew, Y., 1998. The northwestern Ethiopian Plateau flood basalts: classification and spatial distribution
of magma types. J. Volcan. Geotherm. Res. 81, 91.
Salle, B., Cremers, A.D., Maurice, S., Wiensd, R.C., Fichet, P., 2005. Evaluation of
a compact spectrograph for in-situ and stand-off laser-induced breakdown
spectroscopy analyses of geological samples on Mars missions. Spectrochim.
Acta Part B 60, 805e815.
Smith, B.J., Gomez-Heras, M., McCabe, S., 2008. Understanding the decay of stonebuilt cultural heritage. Prog. Phys. Geogr. 32, 439e461.
Steven, J., 2005. Effect of environmental factors on the chemical weathering of
plagioclase in Hawaiian basalt. Phys. Geogr. 26 (1), 69e84.
Zanettin, B., 1992. Evolution of the Ethiopian volcanic province. Mem. Lincee Sci. Fis.
Nat. 1, 155.