This is the author-manuscript version of this work - accessed from

http://eprints.qut.edu.au
Goodall, Rosemary A and Hall, Jay and Sharer, Robert J and Traxler, Loa and
Rintoul, Llew and Fredericks, Peter M (2008) Micro-ATR spectral imaging in
combination with Raman spectroscopy and ESEM-EDX in Archaeology: Application
to Maya Paint and Plaster Wall Decorations . Applied Spectroscopy 62(1):pp. 10-16.

Copyright 2008 Society for Applied Spectroscopy

Micro-ATR spectral imaging in combination with Raman

spectroscopy and ESEM-EDX in Archaeology: Application to Maya
Paint and Plaster Wall Decorations

ROSEMARY A. GOODALL*, JAY HALL, ROBERT J. SHARER, LOA TRAXLER,

LLEW RINTOUL and PETER M. FREDERICKS
School of Physical and Chemical Sciences, Queensland University of Technology, GPO Box
2434, Brisbane, Qld 4001, Australia (R.A.G., L.R., P.M.F.); School of Social Sciences, University
of Queensland, 4067, Brisbane, Australia (J.H.); and University of Pennsylvania Museum,
Philadelphia, PA, 19104-6324, USA (R.J.S., L.T.)

Fourier transform infrared - attenuated total reflection imaging has been successfully used
to identify individual mineral components of ancient Maya paint. The high spatial
resolution of a micro FT-IR-ATR system in combination with a focal plane array detector
has meant that individual particles in the paint have been resolved and identified from
their spectra. This system has been used in combination with micro-Raman spectroscopy
to characterize the paint which was found to be a mixture of hematite and silicate particles
with minor amounts of calcite, carbon and magnetite particles in a sub-micron hematite
and calcite matrix. The underlying stucco was also investigated and found to be a
combination of calcite with fine carbon particles, making a dark sub-ground for the paint.

Index Headings: FT-IR-ATR Imaging; Raman microscopy; paint; pigments; stucco; Copan;
Maya
INTRODUCTION

The paints used by ancient people for the decoration of ceramics, buildings and cave
walls, have been studied intensely over the last 20 years.1-4 Knowledge of the mixtures used to
create paint blends can provide insight into, the type of materials utilized, the binding materials
required, where the materials were sourced, and is also of great assistance in the conservation of
the painted surfaces. To understand the mixtures used by ancient artisans in the formulation of
paints the individual components need to be identified. While many analytical techniques can
give valuable information on the elemental composition and structural makeup of the materials,
few have the spatial resolution required to identify the individual particles. Fourier transform
infrared imaging using an ATR attachment can provide both spatial and spectral information
about the sample area.5 When combined with techniques such as micro-Raman spectroscopy and
environmental scanning electron microscopy (ESEM) coupled with energy-dispersive X-ray
microanalysis (EDX), detailed information regarding paint components, their size, composition
and structure can be obtained.

*Author to whom correspondence should be sent.

E-mail: r.goodall@student.qut.edu.au

FT-IR imaging couples an FT-IR microscope system with a focal plane array (FPA) detector to
enable the simultaneous collection of multiple spectra across the sampled area. The number of
spectra collected per image is dependent on the size of the FPA; for example, a 32x32 array
allows 1024 independent spectra to be collected in a single measurement. The images generated
from individual spectra are spatially resolved maps providing the location of each component in
the sampled area.5 When operated in reflection or transmission mode FT-IR imaging has a
maximum lateral spatial resolution of around 5-10 m.6 This spatial resolution is limited by
diffraction and furthermore may not always be achievable because of signal-to-noise ratio
considerations. Micro-ATR imaging offers an increased lateral spatial resolution of around 23m,5 because of the change in wavelength of the light as it passes through the internal reflection
element, and can be applied to samples that are non-reflective or not suitable for transmission
measurements. FT-IR imaging has so far been applied mainly in polymer7,8 and
biological5,6,9,10,11 studies with a small number of geological studies12 and a single study of an oil
painting fragment.13 For archeological paint samples the particle size is typically in the range
from 1 to 20 m, making this technique ideal for the identification of individual particles. For
paint mixtures this permits an understanding of the spatial makeup including individual particle
size, shape and distribution. This in turn leads to a better understanding of individual pigments,
and can help to identify if they are natural mineral complexes or deliberate mixtures of minerals
to achieve a definite color or effect. Micro-Raman spectroscopy has also been used in similar
studies14,4 because of the high spatial resolution of the system, allowing spectra to be recorded on
individual particles of 1-2 m. Raman spectroscopy is particularly useful for the identification
of iron oxide minerals, which are often the main constituent of ancient red and yellow paints,
because of the restricted spectral range of FPA detectors. Generally, FPA detectors are not
3

effective below about 800 cm-1 and hence many important vibrational modes, for example of
oxides, are not seen. However, many minerals found in ancient paints are poor Raman
scatterers, or give rise to fluorescence, and are often more easily identified by infrared
spectroscopy, making the two techniques complementary.

In this work FT-IR-ATR imaging was used to characterize Maya paint samples from an
ancient building at Copan, Honduras. The Maya decorated the external and internal surfaces of
their structures by first coating them with stucco and then applying colored paints. The objective
of the work was to identify as many of the mineral phases in the paint and stucco as possible.
Elemental maps recorded using EDX analysis in an environmental scanning electron microscope
(ESEM-EDX) together with backscattered electron images from the ESEM, were used to
confirm the composition and morphology of each particle. The ESEM and FT-IR-ATR images
were matched for complete characterization of the paint particles.

EXPERIMENTAL
Samples. The sample is from the Sub Jaguar Tomb15 and consists of three separate pieces of
painted stucco which have exfoliated from the wall painting of the tomb. Each sample piece
consists of a thick layer of coarsely ground stucco up to 8 mm thick, covered by a thin fine layer
of stucco and then by a layer of red paint. All pieces were examined initially as received and
then a part of one of the pieces was set in Araldite M resin, and polished to a smooth surface
using diamond paste. This polished cross-section was used in subsequent mapping and imaging
analysis.

Attenuated Total Reflection FT-IR Spectral Imaging (FT-IR-ATR spectral imaging). Data
were collected using a Varian FT-IR imaging system. The system consists of a rapid scan
Varian 3100 FT-IR spectrometer, a Varian 600 UMA FT-IR microscope equipped with an ATR
objective and a 32x 32 liquid nitrogen cooled mercury-cadmium-telluride (MCT) focal plane
array detector. A germanium slide-on crystal was used in the ATR objective. The crystal was
pressed onto the sample in various locations within an area defined by the environmental
scanning electron microscopy- energy dispersive X-ray microanalysis (ESEM-EDX)
experiments. The area analyzed each time was approximately 40 x 40 m. Spectra were
acquired at 8 cm-1 resolution with 512 scans co-added, over a range of 4000-850 cm-1. Spectra
were processed and images created using Varian Resolution pro software. Band positions were
identified where necessary using band fitting of individual spectra with Grams/AI software
(Thermo Scientific, Madison WI). Band fitting was carried out using a Gaussian curve function
with the minimum number of bands used. The iterations were limited to 50 with a squared
correlation of r2 greater than 0.995. Unless otherwise stated, spectra are presented without
processing.

Raman spectroscopy. Spectra were recorded directly from the pigment and stucco samples
using a Renishaw inVia spectrometer (Renishaw plc, Gloucestershire, UK) consisting of a Leica
microscope with a 50x objective attached to a spectrograph equipped with an electrically-cooled
charged coupled device (CCD) detector. Excitation was from a Renishaw diode laser emitting at
785 nm. The laser power was reduced using an optical filter to approximately 0.7 mW at the
sample. The lower laser power was used to avoid inducing thermal changes in the mineralogy of
the iron oxide minerals. Integration time was 50 s with 5 accumulations.
5

ESEM and X-ray microanalysis. Backscattered electron imaging, X-ray microanalysis and Xray mapping of the cross-sections set in resin were carried out on an FEI Quanta 200
environmental scanning electron microscope (ESEM). The instrument was operated in
environmental mode using water vapor as the specimen chamber gas, chamber pressures of 1.52.0 Torr (200267 Pa), an accelerating voltage of 20 kV and a working distance of 10mm. Under
these conditions surface charge cancellation occurs and therefore the sample could be examined
directly without a conductive coating. Since the backscattered electron signal is dependent on
the average atomic number at each sample point, the backscattered images allowed
discrimination of the different phases in the sample cross-section. Elemental analysis of
individual constituents in the cross-section was carried out using an EDAX energy-dispersive Xray microanalysis system. X-ray maps (elemental images) were acquired over 256x200 pixels at
a magnification setting of 900x and a chamber pressure of 2.0 Torr. Under high vacuum
conditions, the analytical spatial resolution is limited solely by the electron beam spread within a
solid sample. For Ca X-rays produced in a calcite matrix the X-ray emission volume per analysis
point in the sample surface can be calculated to be 3.4 m, and for Si in quartz it is 3.6 m16.
Under ESEM conditions, some scatter of the incident electron beam takes place in the chamber
gas17, and therefore the spectrum at each analysis point will contain a small contribution from the
surrounding material18,19. Modelling of the beam scatter17 shows that 70% of the beam is
unscattered at 2 Torr, hence the extraneous contribution is limited to no more than 30%. The
effect of the beam scatter on the elemental X-ray maps is a slight softening or broadening of
detail such as grain edges.

Micro-attenuated total reflection infrared spectroscopy. ATR Spectra were recorded directly
from the polished surfaces of the mounted cross-sections by contacting the ATR crystal onto
selected areas of the samples. The FT-IR spectrometer was a Nicolet Nexus 870 with a
Continum infrared microscope equipped with an ATR objective incorporating a Si internal
reflection element (Thermo Electron Corp., Madison, WI). The system was equipped with a
liquid nitrogen cooled MCT detector. Spectra were recorded with a resolution of 4 cm-1 and a
range of 4000 - 650 cm-1 for 128 co-added scans. All ATR spectra are corrected for wavelength
dependence of depth of penetration using OMNIC software (Thermo Scientific, Madison, WI).

RESULTS
The stucco: The sample pieces were first examined using a light microscope and an estimate
made of the particle sizes in each layer. The stucco is a pink/buff color, with a large range of
particles 1mm down to 5-10 m sizes dispersed throughout. Particles vary in color from brown,
red, yellow to white. This large variety of colored material differs from the composition of the
stuccos found on the external surfaces of buildings20, which are white and are predominately
composed of calcite. Raman spectroscopy of both the raw sample and the polished cross-section
identified a range of minerals. These are predominantly calcite and quartz, with small particles
of rutile, hematite, goethite, magnetite, carbon, and some volcanic rocks, possibly feldspars or
plagioclase. The distribution of these small mineral particles and their low number suggests that
they are contamination from the preparation process and not deliberate inclusions. Many of the
minerals are used in other paints and some could be part of the grinding stones. The larger sand
and soil particles could have been included with the calcite as fill for the stucco. FT-IR-ATR
images of this layer (not shown here) confirm the distribution of a range of non-calcite particles
7

in the layer including clay, mica and quartz particles. Above this coarse layer is a finely ground
calcite layer with a smaller number of impurities. This layer is quite dark in comparison to the
lower layer and Raman spectra of the particles revealed that this layer consists of very fine
carbon particles dispersed in the calcite stucco. The fine nature of this amorphous carbon and its
even distribution suggests that the inclusion of the carbon was deliberate to create a dark underlayer for the paint. Again this contrasts with the fine white, almost pure calcite stucco layer used
under paint on the external surfaces of buildings. Unfortunately, the high levels of carbon from
the setting resin and carbonate in the stucco prevented the observation of the variation in the
carbon levels in the ESEM-EDX maps. Carbon is not easily detected by infrared spectroscopy
and so the distribution was not mapped by this technique either. This type of carbon rich stucco
layer has not previously been reported for Maya stucco material, or in other stucco contexts.

An ESEM-EDX map of the area of stucco immediately under the paint layer was
recorded. The mapped area is predominately Ca, C and O with 5 small grains showing high Fe
levels and a range of grains containing high Si and Al. All other grains are Ca based in a high Ca
background making it difficult to differentiate individual grains. Micro-Raman spectroscopy
was used to confirm that most of the grains were calcite. The small number of silicate grains
suggests that these are accidental rather than deliberate additions. ATR spectral images of 40x40
m areas of this mapped section show strong calcite spectra for most of the image, with bands at
1400 and 870 cm-1. ATR spectral images generated using the band at 1030 cm-1 show a small
number of grains coinciding with the silicate grains identified in the ESEM maps.

Red Paint layer: Surface microscopic observations of this paint identified a coarsely ground
material of uneven texture. Initial Raman spectroscopy of the surface identified that the paint
was predominately made up of hematite particles with some magnetite particles. Contaminating
particles consisted of calcite, carbon, quartz and goethite in order of amount, along with a
number of particles which fluoresced or were Raman inactive. The presence of a large number
of different mineral phases suggested that this was a heterogeneous paint which differs from the
homogeneous paints used in the external building decoration21. Analysis of the paint layer in the
polished cross-section identified a significant number of distinct mineral particles with a wide
range of particle sizes. Mineral types identified were similar to those identified in the surface
analysis, but also included rutile and anatase. Micro-FT-IR-ATR analysis also identified the
presence of calcite and clay minerals in the paint. Preliminary ESEM analysis of the crosssection confirmed that the paint was made up of a range of particles with sizes from less than 1
m up to 50 m (Fig. 1). ESEM-EDX confirmed that the paint was composed predominantly of
Fe and Ca with some Si, Al and low amounts of Mg. The presence of high levels of Ca along
with crystalline calcite in the paint matrix confirmed that calcite stucco was used as a binding
agent in the paint layer.

In order to identify the particle types and morphology, an ESEM micrograph of a section
of the paint layer was recorded (Fig. 1). An ESEM-EDX map of this area was then carried out.
Thirteen elements were mapped, Ca, C, Fe, Al, Si, Mn, O, Na, Mg, K, P, Cl and Ti. The most
relevant elements for this determination are Fe, Ca, Si, Al and to a lesser degree Na and Mg, as
the other elements were not found in the sample or only one particle was identified, as in the case
of Ti. The matrix of the pigment layer is a mixture of Ca and Fe (Fig. 2 Ca and 2 Fe), with
9

individual particles containing high Fe or Ca clearly delineated. The stucco layer lies across the
lower right hand corner of the map and is very high in Ca, low in Si and Al and contains no Fe.
The Si map contains a range of particles with varying intensity, the brighter particles containing
a higher concentration of Si. Some particles are very high in Si and this is confirmed by the
Raman analysis which identified some of the particles as quartz. Al is present in varying
concentrations in some of the silica particles, along with Na and Mg which are present in very
low levels in many of the particles. A small number of silicate particles also contain K and could
be a form of mica. A number of layered particles conforming to the morphology of mica were
identified in the ESEM analysis.

FT-IR-ATR spectral imaging was carried out on three sub-sections of the ESEM mapped
section of the paint layer. These are designated 1, 2 and 3, (Fig. 1) and also marked on the Si
map (Fig. 2). The ESEM image of Area 1 has two large, approximately 22x15 m grains at the
top. An FT-IR-ATR spectral image generated using the intensity of the 1065 cm-1 band (Fig. 3b)
shows the left hand grain shape clearly differentiated from the matrix material, while the right
hand grain is much less distinct. ESEM-EDX analysis of the left hand grain gives a very high Si
level with no Al, hence the very bright appearance in the Si map (Fig. 2). To better understand
the variations in the silicate minerals present in this area, band fitting was used to clarify the peak
positions. This enabled the discrimination of silicate minerals from alumino-silicate minerals.
The band fitting analysis of a typical infrared spectrum for this particle is shown in Fig. 4a. The
strong band at 1030 cm-1 along with the bands at 1192 and 1154 cm-1 are assigned to Si-O
stretching and are associated with silicate minerals.22 The band at 1068 cm-1 assigned to the SiO bond stretching is lower than the 1080 cm-1 normally found for quartz suggesting that this is
10

another form of silicate structure, possibly an amorphous silicate glass.23 This material could be a
mixture of silicate phases, unfortunately the lack of other diagnostic bands in the spectrum
makes it difficult to correctly match the mineralogy of the spectra. The second (right-hand)
particle can be seen more clearly distinguished from the matrix material in Fig. 3c and 3d. These
images were generated using the intensity of the band at 1014 cm-1. The band fitting analysis of
a typical infrared spectrum (Fig. 4b) has an intense Si-O stretching band22 at 1014 cm-1 with a
very weak Si-O-Al band at 942 cm-1 and a weak Al-O-H band22 at 916 cm-1. The Al map (Fig.
2) shows a higher concentration of Al in this particle. The low levels of Al (approx 4%)
compared to Si (approx 24%) and the lack of clay bands present in the spectra suggests that this
is most likely an Al substituted silicate.22 Micro-ATR infrared spectra confirm the presence of
quartz in the paint layer with bands at 1075, 798 and 780 cm-1, silicates with bands at 1055,
1035, 1010 and 798 cm-1 and also the presence of clay silicates with -OH stretching bands at
3695, 3666 and 3618 along with bands at 1030, 1008, 936 and 913 cm-1. The rest of the image is
predominantly matrix material with a broad carbonate band at around 1400 cm-1 and in some
areas a weak broad band at 1033 cm-1. Individual calcite particles are not observed in this image.

The ESEM micrograph, Fig. 3e, of Area 2, shows a mixture of grains imbedded in a
finely ground matrix of sub-micron particles. These grains range in size from 6 m to 19 m.
The ATR spectral image (Fig. 3f and 3g) generated using the band at 1010 cm-1 shows distinct
grains with one area, lower left, which has not been fully resolved from the adjoining grain. This
unresolved area is a mixture of finely ground matrix particles. A typical spectrum of this area
contains both calcite, 1407, 870 cm-1 and silicate 1068 and 1032 cm-1 bands (Fig. 5a). The grain
in the bottom left of the image is clearly resolved. The typical spectrum for this grain has an
11

intense band at 1077 cm-1 typical of quartz (Fig. 5b), while some of the peripheral spectra also
have bands at 1010 and 915 cm-1 due to Al substituted silicate on the edge of the grain. ESEMEDX analysis of this grain shows high Si, some Al and relatively high Fe. The grain is
surrounded by small iron particles which are detected by the slightly scattered beam in the low
vacuum conditions of the ESEM. The small grain in the bottom right hand corner of the ATR
spectral image has bands at 1115, 1030, 1012 and 905 cm-1 which are indicative of an aluminosilicate. The high iron levels found in the ESEM-EDX analysis of this grain are due to the
surrounding sub-micron particles of iron (Fig. 2 Fe). The grain in the top centre of the FT-IRATR spectral image (Fig. 3f) is very high in Si with no Al and a typical spectrum shows an
intense band around 1060 cm-1. The FT-IR-ATR spectral image generated using the band at 1400
cm-1 (Fig. 3h) which is the carbonate band in calcite, shows two grains on the right hand side.
These grains both have typical spectra with intense bands around 1400 and at 870 cm-1 indicative
of calcite (Fig. 5c). The Raman spectrum of the top grain confirms that it is calcite. The lower
area appears to be a tightly packed agglomeration of the matrix material, which is high in calcite
stucco. This area is also visible in Fig. 3f which is based on the intensity of the 1010 cm-1 band,
confirming the presence of silicate material in the matrix.

The ESEM image of Area 3 (Fig. 3i) shows two small grains of around 10 m and 8 m
respectively in a finely ground matrix of sub-micron particles. The FT-IR-ATR spectral image
generated using the 1012 cm-1 band (Fig. 3j), shows two grains clearly distinguished from the
matrix material. Both grains are visible in the ESEM-EDX Al and Si maps. The bands at 1064,
1046, and shoulder at 950 cm-1 of a typical spectrum (Fig. 6a) for the top left grain, are
suggestive of a disordered silicate material with the weak band at 915 cm-1 due to Al substitution
12

in the silicate22 matrix. The grain in the lower right of the image has a typical spectrum with an
intense band at 1012 cm-1 and a shoulder at 1022 cm-1 which is more suggestive of a layer
silicate (Fig. 6b). The ESEM-EDX analysis confirms higher levels of Al in the mixture and this
combined with a high level of K in the grain indicates a slightly different mineralogy to the
alumino-silicate material found in the other areas. The carbonate band at 1400 cm-1 in this
spectrum is due to contamination from the surrounding matrix as the ESEM-EDX result
confirms only a low level of Ca in this grain. The image generated using the 1400 cm-1 band
(Fig. 3l) shows carbonate containing particles as well as a less defined area which is part of the
finer matrix. The larger intense area at the top of the image appears to be a cluster of small
particles of 1-2 m each. Raman spectra of the particles confirm that they are calcite. Many
sections of the fine matrix material have spectra with a combination of carbonate and silicate
bands. The matrix in the ESEM-EDX maps contains a mixture of the elements Fe, Ca, Si and Al
and gives Raman spectra with weak hematite and calcite bands.

DISCUSSION

Conventional micro-ATR infrared spectroscopy can only provide a generalized spectrum

of a limited area and the overlapping of the strong Si-O bands make it difficult to identify
individual components. In contrast FT-IR-ATR spectral imaging of this heterogeneous paint
mixture reveals a complex mixture of individual particles of different mineral phases. Individual
spectra can be obtained for each particle allowing the differentiation of each of these mineral
13

types. The main coloring agent in the paint mixture, established using Raman spectroscopy, is
iron oxide, in this case predominantly hematite. This material, in combination with finely
ground calcite, makes up the majority of the very fine sub-micron matrix material. The calcite in
the matrix can also be clearly identified in the FT-IR-ATR spectral images along with calcite
particles distributed throughout. Clusters of this matrix have been identified in the images, and
spectra confirm that they are a mixture of calcite and silicate materials. These areas can be
differentiated from individual grains in the images because they lack the definitive edges and
shape seen in the grains. A small proportion of calcite stucco added to the paint mixture when
wet would aid in the adhesion of the paint layer to the stucco surface. The addition of stucco to
the paint also lightens the color resulting in the softer red observed in this sample. In this
instance there is a clear delineation between the paint layer and the stucco layer (Fig. 2 Ca)
indicating that the paint has been applied to a dry stucco surface. The FT-IR-ATR spectral
images of the stucco layer show a small number of silicate and alumino-silicate grains. Some as
small as 5 m have been differentiated in a predominantly calcite stucco matrix. Other grains
identified in this layer are calcite. The fine carbon particles throughout this layer are not
detected in the FT-IR-ATR spectral images but are identified using micro-Raman spectroscopy.
The presence of finely ground carbon in the fine stucco layer applied directly under the painted
surface indicates that a different surface appearance was sought. This was an attempt to darken
the walls of the tomb prior to painting and perhaps darken the appearance of the paint layer.

The FT-IR-ATR spectral images of the paint layer obtained here show a selection of
distinct particles distributed across the layer. In each of the areas examined grains ranging in size
from 6 to 22 m have been resolved from the surrounding matrix material. What makes this
14

paint mixture interesting is the variety of silicate minerals present making up a significant
percentage of the total paint. The components identified using FT-IR-ATR spectral imaging,
include quartz, disordered silicate and alumino-silicates combined with iron rich silicates. The
mixture of particulates and the coarseness of the grains, along with the presence of distinct grains
of high iron containing hematite, indicate that the silicates have been added to a higher grade
hematite as a grinding and filling agent. Clay material in particular can be added as a binding
agent to assist in the adhesion of the paint to the stucco under layer. The inclusion of small
contaminating particles of magnetite and goethite along with carbon particles in the paint,
suggests that little care has been taken to maintain the purity of this paint. This is in stark
contrast with other paints from the external surfaces of buildings examined in an earlier study21.
These paints consist of finely ground hematite with only traces of other contaminant materials.
This matches with what we know about the rest of the tomb. The roughness of the rock walls,
the coarse application of the stucco, and contaminants in all coating materials point to a hurried
preparation and application in this tomb.

CONCLUSIONS

The paper reports the successful application of FT-IR-ATR spectral imaging to layered
archaeological paint samples. The improved spatial resolution of this technique over
conventional micro-ATR infrared, allows the orientation and morphology of individual grains to
be distinguished. The heterogeneous mixture of mineral phases, including a high proportion of
other materials with the pigment, hematite either as a byproduct of the preparation process or to
deliberately extend the paint, was determined. By combining these data with the results from the
15

complementary technique of micro-Raman spectroscopy a comprehensive characterization of the

paint mineralogy has been established. This paint is a deliberate mixture of hematite, silicates,
calcite, carbon and magnetite particles in a fine submicron matrix of hematite and calcite stucco.
The interpretation of the FT-IR-ATR spectral imaging was confirmed with the use of ESEMEDX microanalysis, X-ray mapping and backscattered electron micrographs. The use of carbon
to darken the underlying stucco has not been previously reported.

ACKNOWLEDGMENTS

We wish to thank the Instituto Hondureno de Antropologia e Historia for its support in this and
other studies we have carried out. We also thank Dr Thor Bostrom (Analytical Electron
Microscopy Facility, QUT) for carrying out the ESEM-EDX analyses, for assistance in the
interpretation and presentation of the elemental maps, and for his advice and recommendations
on this paper.

1. D. Barilaro, G. Barone, V. Crupi, M.G. Donato, D. Majolino, G. Messina, and R.

Ponterio, J. Mol. Struct. 744-747, 827 (2005).
2. F. Ospitali, D.C. Smith, and M. Lorblanchet, J. Raman Spectrosc. 37, 1063 (2006).
3. P. Vandenabeele, S. Bode, A. Alonso, and L. Moens, Spectrochim. Acta, Part A 61, 2349
(2005).
4. A. Zoppi, G.F. Signorini, F. Lucarelli, and L. Bachechi, J. Cult. Heritage 3, 299 (2002).
16

5. K.L.A. Chan, S.G. Kazarian, A. Mavraki, and D.R. Williams, Appl. Spectrosc. 59, 14
(2005).
6. D. Faibish, A. Gomes, G. Boivin, I. Binderman, and A. Boskey, Bone 36, 6 (2005).
7. C.M. Snively and J.L. Koenig, J. Polym. Sci. Part B 37, 2353 (1999).
8. O.S. Fleming, K.L.A. Chan, and S.G. Kazarian, Polymer 47, 4649 (2006).
9. K. Verdelis, M.A. Crenshaw, E.P. Paschalis, S. Doty, E. Atti, and A.L. Boskey, J. Dental
Res. 82, 697 (2003)
10. J. van der Weerd and S.G. Kazarian, J. Parmaceutical. Sci. 94, 2096 (2005).
11. C. Ricci, K.L.A. Chan, and S.G. Kazarian, Appl. Spectrosc. 60, 1013 (2006).
12. R. Wysoczanski and K. Tani, J. Volcanic and Geotherm. Res. 156, 302 (2006).
13. J. van der Weerd, R.M.A. Heeren, and J.J. Boon, Studies in Conserv. 49, 193 (2003).
14. H.G.M. Edwards, P.S. Middleton, S.E. Jorge Villar, and D.L.A. de Faria, Anal. Chim.
Acta 484, 211 (2003).
15. E.E. Bell, R.J. Sharer, L.P. Traxler, D,W. Sedat, C.W. Carrelli, and L.A.Grant, Tombs
and burials in early classic acropolis Copan in Understanding early classic Copan, E.E.
Bell, M.A. Canuto and R.J. Sharer, eds. (University of Pennsylvania Museum of
Archaeology and Anthropology, Philadelphia, 2004) p.151.
16. J.I. Goldstein, D.E. Newbury, P. Echlin, D.C. Joy, A.D. Romig Jr, C.E. Lyman, C. Fiori,
E. Lifshin, Scanning electron microscopy and X-ray microanalysis: a text for biologists,
materials scientists, and geologists, (Plenum Press, New York, 1992) p.131.
17. S.A. Wight, Scanning 23, 320 (2001).
18. D.C. Sigee and C. Gilpin, Scanning Microscopy Supplement 8, 219 (1994)
19. C. Mathieu, Mikrochimica Acta (Suppl) 15, 295 (1998)
17

20. R.A. Goodall, J. Hall, H.G.M. Edwards, R.J. Sharer, R. Viel, and P.M. Fredericks, J.
Arch. Sci. 34, 666 (2007)
21. R.A. Goodall, J.Hall, R. Viel, R. Agurcia F., H.G.M. Edwards, and P.M. Fredericks, J.
Raman Spectrosc. 37, 1072 (2006)
22. V.C. Farmer, Infrared Spectra of Minerals, (Mineralogical Society, London, 1974) p.
367.
23. H.W. van der Marel and H. Beutelspacher, Atlas of infrared spectra of clay minerals and
their admixtures, (Elsevier Scientific, Amsterdam, 1976)

18

Figure Captions
Fig. 1 ESEM backscattered electron image of paint layer covering the area of elemental mapping
Fig. 2 Elemental maps for Al, Si, Mg, Ca, Fe, Na. The Si map includes the areas imaged using
the FT-IR-ATR spectral imaging system
Fig. 3 ESEM and FT-IR-ATR spectral images of the studied areas. a) ESEM image of area 1, b)
ATR spectral image generated using the intensity of the 1065 cm-1 band, c) ATR spectral image
generated using the intensity of the 1014 cm-1 band, d) ATR 3D spectral image generated using
the intensity of the 1014 cm-1 band, e) ESEM image of area 2, f) ATR spectral image generated
using the intensity of the 1010 cm-1 band, g) ATR 3D spectral image generated using the
intensity of the 1010 cm-1 band, h) ATR spectral image generated using the intensity of the 1400
cm-1 band, i) ESEM image of area 3, j) ATR spectral image generated using the intensity of the
1012 cm-1 band, k) ATR 3D spectral image generated using the intensity of the 1012 cm-1 band,
l) ATR spectral image generated using the intensity of the 1400 cm-1 band
Fig. 4 a) FT-IR-ATR spectrum of particle top left area 1 with band fitting of Si-O region
identifying bands at 1174, 1065 cm-1 of an amorphous silicate, b) FT-IR-ATR spectrum of
particle top right area 1 band fitting of Si-O region identifying bands at 1117, 1014, 942, 916
cm-1 of an alumino-silicate
Fig.5 FT-IR-ATR spectrum of a) the unresolved region, bottom right area 2, containing a
mixture of calcite and silicate bands b) particle of quartz lower centre of area 2, c) calcite
particles right hand side area 2
Fig. 6 FT-IR-ATR spectrum of a) disordered silicate particle, top left area 3, b) layer silicate
particle, lower right area 3

19

Figure 1

Figure 2
20

Figure 3

Figure 4

21

Figure 5

Figure 6

22