2015 International Nuclear Atlantic Conference - INAC 2015

São Paulo,SP, Brazil, October 4-9, 2015

ASSOCIAÇÃO BRASILEIRA DE ENERGIA NUCLEAR - ABEN
ISBN: 978-85-99141-06-9

EDXRF ANALYSIS OF SCULPTURES ON POLYCHROME WOOD

Valter de Souza Felix1,2, Cristiane Calza1, Renato P. Freitas2, Ricardo Tadeu Lopes1

1
Laboratório de Instrumentação Nuclear
PEN/COPPE/UFRJ
Caixa Postal 68509
21941-972 Rio de Janeiro, RJ
ccalza@lin.ufrj.br
ricardo@lin.ufrj.br
2
Instituto Federal do Rio de Janeiro (IFRJ)
Paracambi, RJ
renato.freitas@ifrj.edu.br
valter.felix@ifrj.edu.br

ABSTRACT

In the last years, the analysis of sacred sculptures, belonging to museums, churches and private collectors has
gained an increasing interest. This kind of analysis can be extremely valuable to conservation and restoration
treatments; it can also supply important information that makes possible to identify an artist, to date a sculpture
and to identify forgeries. By means of techniques such as XRF, PIXE or XRD, is possible to identify the
composition of materials employed in the manufacturing of the pieces, pigments used in the polychromy and
also the presence of ancient or modern retouchings. However, the fact that every artwork is a unique piece
emphasizes the necessity of working with non-destructive techniques. This work reports the analysis of two
sacred images on polychrome wood - portraying Our Lady (XIX century) and Saint Anthony (XVIII century) -
using EDXRF. The analyses were performed with a portable system, consisting of Amptek 123-SDD detector
and Mini-X X-Ray tube, with W anode, operating at 30 kV and 40 μA. The analysis of the sculpture of Our
Lady revealed the use of the following pigments: lead white, lithopone, vermilion, red ochre, umber and
Prussian blue. The analysis of Saint Anthony revealed: lead white, vermilion, brown ochre, bone black and
Prussian blue. In both sculptures the use of gold foils for gilding was proved by the presence of Au in the
spectra. These results were used to assist the restoration procedures of the sculptures.

1. INTRODUCTION

There is an increasing demand for non-destructive analysis of sacred sculptures, belonging to

museums, churches and private collectors. In this kind of analysis, is possible to identify the
composition of materials (terracotta, gypsum, gold foils, etc.), pigments employed in the
polychromy and also the presence of ancient or modern retouchings – by means of techniques
such as XRF, PIXE or XRD, for example. On the other hand, the conservation state of the
objects can be evaluated using radiography and tomography techniques, which provide the
visualization of structural damages and internal regions, revealing the use of metal structures,
nails and spikes, etc. Therefore, based on the results of the analyses, is possible to determine
the historical period of the artworks and, in some cases, to identify forgeries [1].

X-ray fluorescence (XRF) is the most used analytical technique in archaeometry, in order to
investigate the composition of pigments (in manuscripts, paintings, ceramics, and other
artifacts), metal alloys, coins, and statuary. It is a nondestructive technique that enables
qualitative and quantitative multielemental analysis with good precision and accuracy. The
latest developments in X-ray tubes and detectors have enabled the introduction of small and
portable equipments, which can be used for in situ analysis of cultural objects in museums
and galleries [1-6].

The analytical identification of a pigment by means of XRF involves its color and
composition. Since the chronology of pigments use is well-known, is possible to determine
the provenance, historical period and, consequently, the authenticity of an artifact [2, 7].
Organic pigments, or pigments composed only by light elements, cannot be directly detected
by XRF. However, their presence may, in some cases, be implied by the absence of certain
characteristic elements associated with the color under investigation. For blue colors, an
absence of copper would exclude azurite, verditer blue, etc.; an absence of cobalt would
exclude cobalt blue, cerulean blue, smalt, etc.; and an absence of iron would exclude Prussian
blue, for example [8].

This work reports the analysis of two sacred sculptures using Energy Dispersive X-Ray
Fluorescence (EDXRF). These sculptures are devotional images, made on polychrome wood,
portraying Our Lady and Saint Anthony and belong to private collectors. The first one was
sculpted in the XIX century and the second in the XVIII century. The results of the analysis
were used by restorers in the recovery procedures of both sculptures. Figure 1 shows the
analyzed sculptures.

Figure 1. The sculptures analyzed in this work: Saint Anthony and Our Lady.

INAC 2015, São Paulo, SP, Brazil.

 2. EXPERIMENTAL

The analyses were carried out with a portable system (figure 2) that comprises a detector 123-
SDD Amptek, with a 12.5 µm thick Be window and a resolution of 140 eV at 5.9 keV. The
X-ray tube is a Mini-X model, also from Amptek, with tungsten anode, which can be
operated in a range of current from 5 to 100 μA and voltage from 10 to 40 kV. The
collimation of the X-ray beam is obtained by means of a collimator supplied together with the
tube by Amptek, which is easily removable and allows the use of internal filters. Another
collimator, made in aluminium with a diameter of 2 mm, designed in the Nuclear
Instrumentation Laboratory, is placed over the detector window. The detector and the tube
are fixed over a basis with an angle of 60º between them. In each one of the sculptures were
analyzed several points, with a counting time of 120 seconds, using a voltage of 30 kV and
current of 40 µA.

Figure 2. The EDXRF portable system used in the analyses.

3. RESULTS AND DISCUSSION

The pigments employed in the polychromy of the sculptures were identified based on its key
elements (detected by EDXRF) and characteristic color of the analyzed regions. Table 1
shows the pigments, key elements, color, chemical composition and period of use. The
element tungsten (W) presented in the spectra is due to the X-ray tube anode.

3.1. Our Lady sculpture analysis

The presence of calcium (Ca) and sulphur (S) in the spectra indicates the use of gypsum
(CaSO4) in the preparation layer, under the polychromy. The presence of strontium (Sr) -
identified in most spectra - can be explained by the chemical similarity between the atoms of
Ca and Sr, which enables the replacement of Ca atoms by Sr ones in natural compounds as
gypsum, calcite and marble.

INAC 2015, São Paulo, SP, Brazil.

The analysis of the Armenian bole, applied under the gilding, revealed the presence of high
intensities of Fe, Ca and traces of K, Ti, Mn and Sr. The presence of Ti, usually associated
with modern paintings, has been reported in antiquity pigments by some authors [9-11], as a
common impurity in earth colors. Elements such as Ti, Mn, Zn, K, etc. can be found as
impurities in earth pigments like ochres, umber and Sienna, since the minerals used as raw
materials for these pigments are commonly associated to rutile (TiO2), cuprite (Cu2O),
microcline, (KAlSi3O8), etc. [11]. Ti-rich orange-red boles were probably formed by lateritic
weathering of neutral or mafic rocks, but such materials are probably not currently available
commercially to artists and restorers. Therefore, Ti-contents can be considered as a
fingerprint of the individual ochres and an indication of their mineralogical origin [11, 12].
The presence of Ti in the Armenian bole had been reported in the analysis of a Portuguese
baroque sculpture [13] and in another sacred sculpture, portraying Our Lady of Conception,
probably from XVIII century, manufactured in Portugal [1]. Figure 3 shows a characteristic
XRF spectrum of the Armenian bole.

20000 Fe

17500

15000

12500
Counts

10000

7500
Ca
5000 Ti Fe
S Ca W
2500 K Mn Sr

2 4 6 8 10 12 14 16 18 20
Energy (keV)

Figure 3. XRF spectrum of the Armenian bole.

In the white regions, as the sculpture basis, it was characterized the use of lead white pigment
(2PbCO3.Pb(OH)2), identified by the presence of Pb in the spectra. Lead white is one of the
oldest manufactured pigments, which use dates back to antiquity, remaining the most used
white pigment until the XIX century. Like cinnabar, it was used as cosmetic (face powder)
from the antiquity to the middle ages. The pigment was made by stacking lead strips in
porous jars with vinegar and burying in animal manure, which generated the heat necessary
to speed up the reaction [14].

In the mantle of the saint, presenting blue color, it was characterized the use of Prussian blue
pigment (Fe4[Fe(CN)6]3.14-16H2O), identified by the presence of Fe in the spectra. The
following elements were also identified in these regions: Ca (from the preparation layer) and
traces of S, K, Ti, Mn, Pb and Sr. In the light blue area of the mantle were identified the
following elements: Pb in high intensities, Fe and traces of Ca and Ti. This result indicates

INAC 2015, São Paulo, SP, Brazil.

the use of lead white and a small portion of Prussian blue. Prussian blue was the first modern
artificially manufactured color, discovered accidentally in 1704 by Diesbach, in Germany,
when he was trying to produce red lake pigments using potash and alkali as substrate. When
using a batch contaminated with animal oil, he accidentally made a purple and then a blue
pigment. The pigment was available to artists in 1724 and it has been extremely popular since
its discovery [14, 15]. Figure 4 shows a characteristic XRF spectrum of the blue regions.

10000 Fe

8750

7500

6250
Counts

5000

3750

2500 Ca Ca Fe
Pb Pb
S Ti W
1250 K Mn Sr

2 4 6 8 10 12 14 16 18 20
Energy (keV)

Figure 4. XRF spectrum of blue regions.

In the regions of red color, as the edge of the mantle, it was characterized the use of vermilion
(HgS), identified by the presence of S and Hg in the spectra. Besides this pigment, it seem to
be used lead white (2PbCO3.Pb(OH)2), red ochre (Fe2O3) and lithopone (BaSO4.ZnS). This
last one was characterized by the presence of S, Zn and Ba in the spectra. Vermilion was
developed by the Chinese around 2000 years before the Romans used it. This pigment was
made by crushing, washing and heating the mineral cinnabar to give a strong red powder.
Alternatively, it was made by mixing mercury and molten sulphur with heating. In the Roman
age, cinnabar was mined at Almaden, Spain. It was extensively used in wall decorations,
statues and also by gladiators (as body paint) and Roman ladies (as lipstick) [14]. Lithopone
was commercially available in 1874 as a substitute or supplement for lead white [16], to
overcome its drawbacks of toxicity, poor weathering, and darkening in atmospheres that
contain sulfur compounds. The pigment is an insoluble mixture of barium sulfate and zinc
sulfide that precipitates upon mixing solutions of barium sulfide and zinc sulfate. The
precipitate is recovered by filtration and then calcined at temperatures above 600°C.

In the gilding areas, as the mantle of the saint, besides the use of Armenian bole, it was
identified the employment of gold foils, indicated by the presence of Au in the spectra. The
other elements identified in the spectra were: Fe, Ca and traces of S, K, Ti, Mn and Sr. The
presence of traces of Ti and other elements - with the exception of Ca and Sr, from the
preparation layer - are related to the Armenian bole. Figure 5 shows a characteristic XRF
spectrum of the gilding.

INAC 2015, São Paulo, SP, Brazil.

 12000
Fe
10500

9000

7500

Counts
6000

4500 Au
Ca Au
3000 Ca Fe W Au
S Ti Mn
1500 K Sr

2 4 6 8 10 12 14 16 18 20
Energy (keV)

Figure 5. XRF spectrum of gilding.

In the carnation layer of the saint and Infant Jesus was characterized the use of lead white and
a small portion of red ochre to provide the skin tone. These pigments were identified,
respectively, by the presence of Pb and Fe in the spectra. Ochre pigments are obtained from
naturally tinted clay since prehistoric times. The word “ochre” comes from the Greek ochros
(yellow) and the chemical responsible for this color is ferric oxide monohydrate (Fe2O3.H2O).
The colored clay is ground and washed to obtain the yellow ochre pigment and, to produce
the red one, is necessary to heat until remove the water. Controlling the heating it is possible
to produce a range of warm yellows to bright red. Red ochre (Fe2O3) occurs naturally in
volcanic regions, where thermal activity has caused the dehydration. The excellent
permanence of the color and the abundance of raw material allowed these pigments to remain
among the cheapest artists colors available [14, 17].

In the regions presenting dark brown color – as the hair of the saint and Infant Jesus – it was
characterized the use of umber (Fe2O3.MnO2), identified by the presence of Fe and Mn in the
spectra. The other elements identified were: S, K, Ca, Ti, Pb and Sr. The presence of traces of
Ti, K, etc. on earth pigments was discussed above. Umber pigment comprises hydrated iron
and manganese oxides. In its natural state, the pigment is known as raw umber and when
heated it becomes a richer brown, known as burnt umber. Its original source was Umbria
(Italy), although it is also mined in Devon and Cornwall. The word “umbra” comes from
Latin and means “shadow” [14, 17].

3.2. Saint Anthony sculpture analysis

In the preparation layer was identified the use of gypsum (CaSO4) based on the presence of
Ca and S in the spectra. In the basis of the sculpture, presenting a marbled effect in blue
color, it was characterized the use of lead white (2PbCO3.Pb(OH)2) and Prussian blue
(Fe4[Fe(CN)6]3.14-16H2O). These pigments were identified by the presence of, respectively,
Pb and Fe in the spectra.

INAC 2015, São Paulo, SP, Brazil.

In the regions of gilding was identified the use of Armenian bole, under the gold foils, which
presented Fe and traces of Ti and Mn in its composition. The use of gold foils was
characterized by the presence of Au in the spectra. The presence of traces of Ti and Mn in the
Armenian bole was discussed in the previous item (3.1). Figure 6 shows a characteristic XRF
spectrum of the gilding.

14000 Fe

12000
Ca Au
10000

8000
Counts

Au
6000
Ca Fe
4000 W
Mn Au
S Ti
2000

2 4 6 8 10 12 14 16 18 20
Energy (keV)

Figure 6. XRF spectrum of gilding.

The gilding in the hair of the Infant Jesus presented an external layer of commercial golden
paint, based on Cu and Zn, and traces of the original layer that was composed by gold foils
(identified by the presence of Au). This external layer was executed in the XX century. The
presence of Pb seems to be related to the use of lead white pigment. Figure 7 shows a
characteristic XRF spectrum of this region.

Cu
10000

8000

6000
Counts

Zn
Pb
4000 Cu Pb
Zn
Au
2000 Au
Fe Pb
Ca Ti

2 4 6 8 10 12 14 16 18 20
Energy (keV)

Figure 7. XRF spectrum of Infant Jesus hair.

INAC 2015, São Paulo, SP, Brazil.

In the carnation layer of the saint and Infant Jesus was characterized the use of lead white and
a small portion of vermilion (HgS), in order to obtain the slightly pink skin tone. These
pigments were identified by the presence of Pb, Hg and S in the spectra.

In the regions exhibiting dark brown color – as the hair of the saint and the clothes – it was
characterized the use of brown ochre (Fe2O3), identified by the presence of Fe in the spectra.
In the cord knotted at the waist of the saint was characterized the use of lead white due to the
presence of Pb in the spectra. In the red book was characterized the use of vermilion and lead
white, identified, respectively, by the presence of Hg, S and Pb in the spectra. In the golden
edges of the book, some areas exhibited the original gilding made by gold foils and others
presented modern commercial golden paint. Figure 8 shows a characteristic XRF spectrum of
the red book.

Hg
8000

7000

6000
Hg
5000
Counts

4000

3000 Pb
W Pb Hg
2000 S
Ca Fe
1000

2 4 6 8 10 12 14 16 18 20
Energy (keV)

Figure 8. XRF spectrum of the red book.

In the sandals of the saint, presenting Black color, it was characterized the use of bone Black
(Ca3(PO4)2 + C + MgSO4), identified by the presence of high intensities of Ca in the spectra.
Bone black is made of charred bones and contains mainly calcium phosphates and carbon
black (amorphous carbon). The origin from bone also explains the presence of a small
amount of magnesium, atomic mass ratio to calcium 1:18. This pigment has been identified in
prehistoric paintings and was also found in Egyptian, Greek and Roman art, throughout
European medieval and Renaissance and later in both oil and watercolor paintings until
modern times [18].

INAC 2015, São Paulo, SP, Brazil.

 Table 1. Possible pigments identified by XRF, key elements, color, composition and
period of use.

PIGMENTS ELEMENTS COLOR COMPOSITION PERIOD OF USE

Lead White Pb white 2PbCO3.Pb(OH)2 Antiquity/XX century

Lithopone S, Zn, Ba white BaSO4.ZnS 1874/Still in use

Vermilion S, Hg red HgS Antiquity/XIX century

Red ochre Fe red Fe2O3 Prehistory/Still in use

Brown ochre Fe brown Fe2O3.H2O Prehistory/Still in use

Umber Fe, Mn brown Fe2O3.MnO2 Prehistory/Still in use

Ca3(PO4)2 + C +
Bone black Ca black Prehistory/Still in use
MgSO4
Fe4[Fe(CN)6]3.14-
Prussian blue Fe blue 1704/Still in use
16H2O

4. CONCLUSIONS

A portable EDXRF system was used to characterize the pigments employed in two sacred
images executed on polychrome wood. The analysis of the sculpture of Our Lady revealed
the use of the following pigments: lead white, lithopone, vermilion, red ochre, umber and
Prussian blue. The analysis of Saint Anthony revealed: lead white, vermilion, brown ochre,
bone black and Prussian blue. In both sculptures the use of gold foils for gilding was proved
by the presence of Au in the spectra. These results were used to assist the restoration
procedures of the sculptures.

ACKNOWLEDGMENTS

To the foundations of research support CAPES and FAPERJ for the postdoctoral fellowship
granted to Cristiane Calza.

REFERENCES

1. C. Calza, D.F. Oliveira, R.P. Freitas et al., “Analysis of sculptures using XRF and X-ray
radiography”, Radiation Physics and Chemistry, doi: 10.1016/j.radphyschem.2015.04.012.
2. C. Calza, A. Pedreira, R.T. Lopes, “Analysis of paintings from the nineteenth century
Brazilian painter Rodolfo Amoedo using EDXRF portable system”, X-Ray Spectrometry, 38,
pp.327-332 (2009).
3. C. Calza, D.F. Oliveira, H.S. Rocha et al., “Analysis of the painting ‘‘Gioventú’’ (Eliseu
Visconti) using EDXRF and Computed Radiography”, Applied Radiation and Isotopes, 68,
pp.861-865 (2010).

INAC 2015, São Paulo, SP, Brazil.

4. C. Calza, M.O. Pereira, A. Pedreira, R.T. Lopes, “Characterization of Brazilian artists
palette from the XIX century using EDXRF portable system”, Applied Radiation and
Isotopes, 68, pp.866-870 (2010).
5. R. Cesareo, C. Calza, M.J. Anjos et al., “Pre-Columbian alloys from the royal tombs of
Sipan: Energy Dispersive X-ray Fluorescence analysis with a portable equipment”, Applied
Radiation and Isotopes, 68, pp.525-528 (2010).
6. R.P. Freitas, C. Calza, T.A. Lima et al., “EDXRF and multivariate statistical analysis of
fragments from Marajoara ceramics”, X-Ray Spectrometry, 39, pp.307-310 (2010).
7. R. Klockenkämper, A. Von Bohlen, L. Moens, “Analysis of pigments and inks on oil
paintings and historical manuscripts using Total Reflection X-Ray Fluorescence
spectrometry”, X-Ray Spectrometry, 29, pp.119-129 (2000).
8. P. Moioli, C. Seccaroni, “Analysis of art objects using a portable X-ray Fluorescence
spectrometer”, X-Ray Spectrometry, 29, pp.48-52 (2000).
9. M. Uda, M. Nakamura, S. Yoshimura et al., “Amarna blue painted on ancient Egyptian
pottery”, Nuclear Instruments and Methods in Physics Research B, 189, pp.382-386 (2002).
10. G. Paternoster, R. Rinzivillo, F. Nunziata et al., “Study on the technique of the Roman
age mural paintings by micro-XRF with polycapillary conic collimator and micro-Raman
analyses”, Journal of Cultural Heritage, 6, pp.21-28 (2005).
11. C. Calza, M.J. Anjos, S.M.F. Mendonça de Souza et al., “X-ray microfluorescence with
synchrotron radiation applied in the analysis of pigments from ancient Egypt”, Applied
Physics A, 90, pp.75-79 (2008).
12. T. Grygar, J. Hradilova, D. Hradil et al., “Analyses of earthy pigments in grounds of
Baroque paintings”, Analytical and Bioanalytical Chemistry, 375, pp.1154-1160 (2003).
13. A. Pereira, “O oratório indo-português do Museu de Évora. Análise dos materiais e
técnicas”, Bulletim of the Évora Museum, 2 (2007).
14. J.R. Barnett, S. Miller, E. Pearce, “Colour and art: A brief history of pigments”, Optics &
Laser Technology, 38, pp.445-453 (2006).
15. M. Ortega-Avilés, P. Vandenabeele, D. Tenorio et al. “Spectroscopic investigation of a
‘Virgin of Sorrows’ canvas painting: A multi-method approach”, Analytica Chimica Acta,
550, pp.164-172 (2005).
16. R.J.H. Clark, “Pigment identification by spectroscopic means: an arts/science interface”,
Comptes Rendus Chimie, 5, pp.7-20 (2002).
17. D. Hradil, T. Grygar, J. Hradilova et al. “Clay and iron oxide pigments in the history of
painting”, Applied Clay Science, 22, pp.223-236 (2003).
18. A. Van Loon, J.J. Boon, “Characterization of the deterioration of bone black in the 17th
century Oranjezaal paintings using electron-microscopic and micro-spectroscopic imaging
techniques”, Spectrochimica Acta Part B, 59, pp.1601-1609 (2004).

INAC 2015, São Paulo, SP, Brazil.