Journal of Cultural Heritage 8 (2007) 289e293

http://france.elsevier.com/direct/CULHER/

Case study

Painting technique and state of conservation of wall paintings

Abstract

The state of conservation of the frescoes at Qusayr Amra was investigated by integrated physico-chemical measurements, particularly optical
and scanning electron microscopy, m-infrared spectroscopy, mass spectrometry and X-ray powder diffraction.
The frescoes appear darkened and severely damaged owing to the deterioration of surface treatments and to the widespread presence of dif-
ferent salts. These caused significant detachments of the painting layers.
The materials used in the painting and preparation layers and in the wall plaster were characterized in order to clarify the painting technique.
Different pigments have been identified on the pictorial layers: the original ones correspond to green earth, yellow and red ochre, realgar, bone
black and lapis lazuli and others, such as titanium white and cobalt blue, ascribed to recent restoration works. Proteinaceous materials (egg or
yolk) have been identified as binders for the pigments, indicating a tempera painting technique.
Three different products, belonging to restoration interventions, have been also identified: the natural product shellac, ascribable to next-
to-last restoration and scarcely removed during last restoration also because of its insolubility, the synthetic vinyl acetate and acrylic polymers.
Particularly the latter one indicates an on-going degradation of the applied fixatives.
The extensive deterioration of the surfaces is also due to the widespread presence of salts, such as chloride, nitrate and sulphate clearly iden-
tified by XRD measurements.
The obtained results give a comprehensive overview on the employed painting technique and its state of preservation and on the causes of the
painting deterioration. Therefore, they are a fundamental tool to develop durable and compatible materials and methodologies for a future con-
servation strategy of this site.
Ó 2007 Elsevier Masson SAS. All rights reserved.

Keywords: Wall paintings; State of conservation; Organic fixatives; Efflorescence

1. Aim of the research processes and to detect the materials used in previous interven-
tions. The final goal is to offer scientific tools for a better un-
Integrated physico-chemical investigations were carried out derstanding of this pictorial cycle and to set up a scientific aid
on fragments, collected from the pictorial cycle at Qusayr and guide for a conscious intervention on these paintings,
Amra, one of the most important buildings of the Umayyad severely damaged not exclusively by natural decay processes.
dynasty (750 A.D.) in Jordan, with the aim to obtain useful in-
formation about the painting technique, to offer a preliminary
scientific view of the state of conservation of these wall paint- 2. Introduction
ings, to ascertain the type and the extent of the degradation
As other desert castles, Qusayr Amra was built during the
dynasty of Umayyad Caliphs (661e750 A.D.) [1,2]. Qusayr
* Corresponding author. Tel.: þ39 049 829 5962; fax: þ39 049 870 2911. Amra was a residence of vacation or a place for rest and there-
E-mail address: [email protected] (P.A. Vigato). fore equipped with baths. In fact, the Arabic term ‘‘qusayr’’

1296-2074/$ - see front matter Ó 2007 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.culher.2007.05.002
290 S. Bianchin et al. / Journal of Cultural Heritage 8 (2007) 289e293

Fig. 1. (a) View of the ‘‘Desert Castle’’ at Qusayr Amra; and (b) a painted vault showing different deterioration features.

indicates a Roman style villa with a piece of land for growing. elemental composition were obtained using an EDAX DX
Also a facing wall has been found in the proximity of the bot- Prime X-ray energy dispersive spectrometer equipped with
tomland (wadi), which suggests a settlement devoted to agri- a thin beryllium window, using an accelerating voltage of
culture (Fig. 1). Very important mural paintings decorate the 25 keV.
central hall and the connected rooms of the building of Qusayr By stereomicroscope observation of the collected samples,
Amra, destined to alcoves and baths with tepidarium and cal- the different layers of the paintings have been mechanically sep-
idarium. Although they have been recently restored, the paint- arated each others and studied separately. The binding media of
ings appear severely deteriorated. The low part of the paintings the painting layers were identified by mFTeIR analyses, per-
(originally painted as imitation of marbles) are damaged by formed either on the different crushed layers and on their solu-
the rising damp and worsened by writings, incisions and other ble fractions extracted by solvents (CH2Cl2, H2O, EtOH,
vandalic actions, that have caused the total loss of significant Acetone). IR spectra were collected in reflection mode by
parts of them. Besides all, chromatic alterations are widely a Nicolet Magna IR560 FTeIR Spectrometer and Spectra
distributed, essentially greyish in colour that could be ascribed Tech microscope on the separated layers and the extracted frac-
to the deposition of soot and other atmospheric fine particu- tions flattened on a gold surface. The IRUG database was used
lates. Moreover, many areas show an extensive network of mi- for the identification of the material constituents of the samples
cro-cracks associated with a spread darkening of the painted [4]. X-ray diffraction (XRD) analyses of the ground samples
surface and detachments; this peculiar decay pattern is espe- were performed on a Philips X’Pert PW3710 diffractometer, us-
cially localised on paintings subjected to earlier restorations ing Cu Ka radiation (40 kV, 30 mA), high-resolution graphite
by different organic fixatives [3]. monochromator, rotating sample holder and proportional detec-
The frescoes of Qusayr Amra show details of the clothes, tor. Measurements were carried out in the range 5 < 2q < 90
i.e. the colors and the fabrics, that testified the social and eco- with a step of 0.02 ; the ICDD data bank of standard X-ray pow-
nomic life in the Umayyad period. Today the visitor must use der spectra was used for phase identification [5].
his imagination to reconstruct the original aspect of the paint-
ings because of the very poor state of presentation of many
scenes, as a diffuse dark patina overshadows the restored 4. Results and discussion
painted surfaces. An urgent restoration is needed to adequately
clean and free from the dirt the wall paintings. The collected samples have been examined as received and/
or as cross sections by integrated physico-chemical measure-
ments, particularly optical and scanning electron microscopy,
3. Experimental mFTeIR and, where possible, by X-ray powder diffraction in
order to verify the pictorial technique, the state of conservation
Twelve samples areas have been collected from the western of the frescoes, the damages caused especially by chemically
and northern walls, side-walls with respect to the main en- agents, particularly salts, detachments, repaintings and fixa-
trance. The sampling procedure deserved difficulty owing to tives also in consequence of the materials employed in previ-
the scarce cohesion of the painting layers and the crumbling ous interventions and to propose an appropriate intervention
of the plaster. The samples were observed as received and as project. Some Pyrolysis GCeMS measurements have been
cross-sections. The optical observations on the cross-section also included to corroborate the infrared indications about
were carried out by a microscope Olympus BX51 equipped the organic materials.
with UV-light source to prove the organic layers. The cross Optical and SEM-EDS microscopies show that the finishing
sections were analysed also by SEM-EDS, after metallization plaster below the painting layers contains calcium carbonate
through the deposition of a graphite film. The instrument was and a small amount of Si, mainly in the form of quartz clus-
Philips Model XL 40 LaB6. Semi quantitative analyses of ters. Some samples show a superficial finishing layer,
 S. Bianchin et al. / Journal of Cultural Heritage 8 (2007) 289e293 291

Fig. 2. Cross section of sample G4a2, magnification 500. Optical observation shows the finishing lime layer followed by the blue layer composed of calcium
carbonate and particles of natural lapislazuli and two superimposed restoration layers (a). Observation by UV light evidences the fluorescence on original and
restoration layers due to organic materials (b).

exclusively constituted by Ca (white lime). Mineral and syn- several samples show a carbonyl ester band at 1738 cm1 (n
thetic inorganic pigments have been identified on different C]O stretching) and the protein bands (broad band centred
painting layers: green earth (potassium and iron silicate), at 3308 attributed to stretching n N-H, 1650 stretching n
bone black, natural lapislazuli, red and yellow ochres, realgar C]O amide I and 1543 cm1 bending d N-H amide II)
and white lead [6e8] (Fig. 2). With the exclusion of layers (Fig. 3a) [9]. The comparison with reference spectra and the
containing white lead, the widespread distribution of Ca on data obtained by pyrolysis GCeMS investigations indicate
the painting layers proves the use of lime as a binder; notwith- the presence of egg [10]. Absorptions of calcium oxalate
standing organic binding as egg tempera, has also been bands at 1618 and 1316 cm1, respectively the symmetric
employed as identified by mFTeIR and GCeMS investiga- and asymmetric stretching of COO group [11] and absorptions
tions. Industrial pigments as cobalt blue and titanium dioxide of nitrate salts at 1386 and 837 cm1 (Fig. 3b), respectively
[8] have been also identified; they clearly belong to a pictorial the stretching and out-of-plane bending of NO 3 group [12]
retouching performed during recent interventions. are clear indication of transformation of constituent materials
mFTeIR investigations, carried out on powders collected and salt induced decay processes.
from the different painting strata, gave quite convincing infor- The mFTeIR characterization of powders and CH2Cl2 ex-
mation on the original binding materials and the organic prod- tracts from the superficial layers of painting samples shows
ucts applied as restoration fixatives; spectra collected on different absorption patterns: some spectra, collected from

0,4 (a)
0,3
Abs

0,2
0,1
0,0
0,3 (b)
0,2
Abs

0,1

-0,0
0,15
(c)
Abs

0,10

0,05

4000 3000 2000 1000

Fig. 3. FTeIR spectrum of the CH2Cl2 soluble fraction of the sample G3 (a) and of the H2O soluble fraction of the sample G8 (b, c). Infrared absorptions reflect the
composition of proteinaceous binding media (i.e. animal glue, egg) in the painting layers (a) and the presence of nitrates (b) and oxalates (c) due respectively to
transformation of constituent materials and salt induced decay processes.
292 S. Bianchin et al. / Journal of Cultural Heritage 8 (2007) 289e293

the outmost layers, show a strong absorption at 1739 cm1 due CH2 groups [13] (Fig. 5). The mixture of esters and acids re-
to the n C]O carbonyl stretching of an ester group, intense sembles the composition of the natural product shellac, whose
bands at 1235 cm1, corresponding to the C-O stretching reference IR spectrum fully agrees with the analysed unknown
band of the acetate CH3-CO-OR, and at 1021 cm1, corre- product [18]. Shellac’s good adhesive power and penetration
sponding to the O-C stretching of the acetate group O-CH2 made it widely employed as fixative for wall painting, not-
[13] (Fig. 4a). This characteristic absorption pattern is ascrib- withstanding the yellowing and increasing insolubility with
able to a vinyl acetate polymer, in good agreement with the ageing that darken the treated surfaces and strongly reduce
reference spectrum of the commercial synthetic resin poly(eth- the removability [15,19].
ylene/vinyl-acetate) (BEVA 371), widely employed in the re- XRD analyses of samples collected from the widespread ef-
cent past as fixative for mural paintings and canvas [14,15]. florescences, prove the presence of sylvite and halite, respec-
Acrylic polymers have also been identified on superficial tively a potassium and sodium chloride, aphthialite, a mixed
layers of few samples (Fig. 4b), these may be ascribed to sodium and potassium sulfate, and, finally, nitratine, sodium
the commercial product Primal, applied during the last resto- nitrate. Macroevidences of pulverization of the wall painting,
ration work. More remarkably, the IR spectrum of the acrylic detachments and superficial incrustations, induced by the
polymer shows a weak band occurring at 1776 cm1 related to disruptive effects of these soluble salts, have been also high-
stretching absorption of g-lactones (Fig. 4b), a common fea- lighted by micro morphological SEM observations on samples
ture of polymer photo-oxidative degradation [16,17], indicat- collected from these deteriorate areas.
ing the applied product is undergoing severe deterioration
processes. 5. Conclusion
Several samples also show a brownish superficial layer,
which appears bright fluorescent by UV light observation of The physico-chemical investigations, carried out on a series
cross sections (Fig. 2b) and consequently not ascribed to of fragments, collected from the wall paintings of the desert
a polymer used as restoration material. So far, it has been me- castle at Qusayr Amra in Jordan, allowed a comprehensive
chanically separated from the painting fragments and charac- characterization of the pictorial technique and of the nature
terised by mFTeIR. All attempts to solubilise it with and the degradation degree of the wall paintings. Different pig-
different solvents (CH2Cl2, acetone, H2O, CH3CH2OH) failed. ments have been identified on the pictorial layers: the original
The IR spectra, collected on the insoluble brownish product, ones, corresponding to green earth, yellow and red ochre, re-
show a strong band occurring at 1712 cm1 with a shoulder algar, bone black and lapis lazuli and those ascribing to recent
at 1732 cm1, related respectively to antisymmetric C]O restoration works, such as titanium white and cobalt blue. In
stretching of a carboxylic acid and to n C]O stretching of disagreement with other studies [3], mFTeIR and GCeMS
an ester group. Other bands in the range 1300e1150 cm1 analyses identified a proteinaceous medium, i.e. egg or yolk,
can also be attributed to different stretching modes of the ester employed as binder for the pigments. mFTeIR spectroscopy
and acid groups, while absorptions at 2929 and 2856 cm1 are has also identified three different fixatives. The natural prod-
related to asymmetric and symmetric stretchings of CH3 and uct shellac has been found in several samples ascribable to

0,5 (a)
0,4

0,3
Abs

0,2

0,1

0,35
(b)
0,30
0,25
0,20
Abs

0,15
0,10
0,05 1776
0,00
3000 2000 1000
Wavenumbers (cm-1)

Fig. 4. FTeIR spectrum of the CH2Cl2 soluble fractions of the sample G4. Both spectra indicate the presence of restoration polymers applied in the recent past:
a vinyl acetate polymer, in good agreement with the reference spectrum of the commercial synthetic resin poly(ethylene/vinyl-acetate) (BEVA 371) (a) and an
acrylic polymer tentatively ascribed to the commercial product Primal (b). The latter is undergoing severe deterioration processes as indicated by the absorption
at 1776 cm1, ascribable to g-lactones, a common feature of polymer photo-oxidative degradation.
 S. Bianchin et al. / Journal of Cultural Heritage 8 (2007) 289e293 293

0,25 (a)
0,20

0,15

Abs
0,10

0,05

(b)
8

6
Abs

10
4000 3000 2000
Wavenumbers (cm-1)

Fig. 5. FTeIR spectrum of powdered superficial brownish layer of painting sample G8 (a). The infrared absorption can be inferred to a mixture of esters and acids,
resembling the composition of the natural product shellac. Its spectrum fully agrees with IRUG reference IR spectrum (b).

next-to-last restoration and scarcely removed during last res- [3] V.J. Medina Flores, A. Garcı́a Bueno, Metodologı́a en la intervención de
toration also because of its insolubility in the most common las pinturas de Qusayr Amra, Jordania, Boletı́n del Instituto Andaluz del
Patrimonio, Histórico 10 (39) (2002) 59e68.
solvents. It could be responsible of the widespread and se- [4] Infrared and Raman Users Group Spectral Database (IRUG). Philadel-
vere darkening of the wall paintings. Also vinyl acetate phia, 2000. Available from: <www.irug.org>.
and acrylic polymers have been identified on the painted sur- [5] International Centre for Diffraction Data (ICDD), The Joint Committee
faces and the decay products, especially of the acrylic poly- on Powder diffraction Standards (JCPDS), 1996. Available from:
mers, clearly indicate an on-going degradation of the applied <www.icdd.com>.
[6] R.L. Feller (Ed.), Artists’ Pigments. A Handbook of Their History and
fixatives. Characteristics, Vol. 1, Oxford University Press, 1986.
The pictorial cycle is indeed severely damaged also by solu- [7] R. Ashok (Ed.), Artists’ Pigments. A Handbook of Their History and
ble salts transported by rising damp and ground water; chloride, Characteristic, Vol. 2, Oxford University Press, 1997.
nitrate and sulphate salts have caused remarkable surface alter- [8] E. West Fitzhugh (Ed.), Artists’ Pigments. A Handbook of Their History
ations, efflorescences and detachments of painting fragments. and Characteristics, Vol. 3, Oxford University Press, 1997.
[9] R.J. Meilunas, J.G. Bentsen, A. Steinberg, Analyses of aged paint binders
The present work must be considered as a preliminary as- by FTeIR spectroscopy, Studies in Conservation 35 (1990) 33e51.
sessment to quantify and qualify the degree and type of damage [10] I. Bonaduce, M.P. Colombini, personal communication, 10-2005.
suffered by wall paintings, to evaluate the current effectiveness [11] R. Arbizzani, U. Casellato, E. Fiorin, L. Nodari, U. Russo, P.A. Vigato,
of the fixatives applied during the previous restoration and their Decay markers for the preventive conservation and maintenance of paint-
possible alterations. The obtained results clearly show the se- ings, Journal of Cultural Heritage 5 (2004) 167e182.
[12] V.C. Farmer (Ed.), The Infrared Spectra of Minerals, Mineralogical
vere deterioration that are suffering both the original materials Society, London, 1974, pp. 409e410.
and the applied products and must be use as a starting point to [13] D. Lin-Vien, N.B. Colthup, W.G. Fateley, J.G. Grasselli, The Handbook
develop durable and compatible materials and methodologies of Infrared and Raman Characteristic Frequencies of Organic Molecules,
for a future conservation strategy of this site. Academic Press Inc, 1991.
[14] G.A. Berger, Consolidation of delaminating paintings, in: ICOM Com-
mittee for Conservation 5th triennial meeting: Zagreb, 1e8 October
Acknowledgement 1978: preprints. International Council of Museums, 1978, p. 19.
[15] P. Mora, L. Mora, P. Philippot, Conservation of Wall Paintings, Butter-
We thank Progetto Finalizzato ‘‘Beni Culturali’’-CNR for fi- worths, London, 1984.
[16] O. Chiantore, M. Lazzari, Photo-oxidative stability of paraloid acrylic
nancial support. We are also indebted to Dr. Fawwaz Al-Khrey-
protective polymers, Polymer 42 (2001) 17e27.
sha, director of the Jordanian Authority of Antiquities who [17] M. Favaro, R. Mendichi, F. Ossola, U. Russo, S. Simon, P. Tomasin,
made possible this investigation. We also thank Dr. Bonaduce P.A. Vigato, Evaluation of polymers for conservation treatments of out-
and Prof. M.P. Colombini for the GCeMS measurements. door exposed stone monuments. Part I: photo-oxidative weathering, Poly-
mer Degradation and Stability, Polymer Degradation and Stability 91
(2006) 3083e3096.
References [18] M. Derrick, Fourier Transform infrared spectral analysis of natural resins
used in furniture finishes, Journal of the American Institute of Conserva-
[1] M. Al Khouri (Ed.), Limes Arabicus, CISU, Roma, 2003. tion 28 (1) (1989) 43e56.
[2] R.W. Dutton, J.I. Clarke, A.M. Battikhi, Arid Land Resources and Their [19] L. Masschelein-Kleiner, Ancient Binding Media, ICCROM eds, Rome,
Management, University of Durham, 1998. 1995, pp. 90e91.