lable at ScienceDirect

Vacuum 83 (2009) S4–S8

Contents lists avai

Vacuum

journal homepage: www.elsevier .com/locate/vacuum

Ion beam analysis of ancient Egyptian wall paintings

Shaaban Abd El Aal a,1, A. Korman a, A. Stonert a, F. Munnik c, A. Turos a,b,*

a Soltan Institute of Nuclear Studies, 05-400 Swierk/Otwock, Polandb Institute of Electronic Materials Technology, 01-919 Warszawa, Wolczynska 133, Polandc Forschungszentrum Dresden, IIM, P.O. Box 510119, D-01314 Dresden, Germany

a r t i c l e i n f o

Article history:Received 16 June 2008Received in revised form19 January 2009Accepted 30 January 2009

Keywords:Egyptian wall paintingPigmentsPIXEOptical microscopy

* Corresponding author. Institute of Electronic MWarszawa, Wolczynska 133, Poland.

E-mail address: turos@fuw.edu.pl (A. Turos).1 On leave from Faculty of Archaeology, Fayoum Un

0042-207X/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.vacuum.2009.01.012

a b s t r a c t

Polychromatic decorations of ancient Egyptians tombs and temples have a long tradition over threemillennia but are hard to identify because many pigments have been subjected to severe chemicalreactions, which have entirely changed their original colours. Optical microscopy, PIXE and microbeam-PIXE have been used for determination of the nature of pigments, their chronology, and identification ofdomestic and imported materials. Paint flakes from various archeological sites in Egypt were analyzed:we report the results of the analysis of samples which were collected at the Mortuary Temple of Ram-esses III (Habu Town), and the tombs of Tuthmosis III (Valley of the Kings) and Sennefer (Valley of theNoblemen). The paint is composed of grains of sizes typically ranging from 50 mm to 300 mm embeddedin binding material and has great non-uniformity of pigment depth and lateral distributions anddiscontinuity of the paint layers. Qualitative analysis using broad beam PIXE has been performed to allowdetermination of the average composition of both support and pigments. Microbeam-PIXE has been usedfor mapping of selected grains. Goethite FeO(OH) (yellow), orpiment As2S3 (green), and the two blues:Egyptian Blue CaCuSi4O10 and Green Frit CaCuSiO3 (mixed with the red haematite Fe2O3) were identified,and interesting details of the painting technique of ancient masters, like blending of pigments and theuse of multilayer structures, were revealed.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Ion beam analysis is one of the currently used techniques forelemental analysis of surface layers. Here we report on the appli-cation of PIXE, microbeam-PIXE and optical microscopy for ancientEgyptian pigment analysis. The aim of the study was to providetypical data for archeological evaluation of pigments like: theirchronology or identification of materials, whether they aredomestic and imported [1,2]. Moreover, such a study is importantfor conservation of paintings [3]. In the course of their history,sometimes more than 4000 years long, wall paintings were sub-jected to many deterioration factors both natural and human,which led to their degradation by discoloration and mechanicaldecomposition. The hues of ancient colour cannot be determinedonly by visual examination, because the colour of some pigmentscan be changed by chemical degradation caused by environmentalconditions (including light, humidity, temperature, atmospheric

aterials Technology, 01-919

iversity, Egypt.

All rights reserved.

pollutants). Many misidentifications of materials were made byarchaeologists before the introduction of accurate instrumentalanalytical techniques. Hence, before starting any conservation orrestoration activity the precise knowledge of the pigments usedtogether with their deposition sequence is indispensable. On thisbasis one can select the suitable methods for conservation andprimarily the judicious choice of chemical treatment for cleaning. Itis worth pointing out that selection of a pigment of the same huebut different from the original one can lead to the chemical reactionbetween them with disastrous end effects [4].

It should be pointed out that such a study makes sense only if isembedded in the immense treasure of archaeological knowledgecollected in the last two centuries. During that period ancientEgyptian pigments have been extensively studied and are quite wellknown [5,6]. The great majority of painting materials used inancient Egypt can be classified into two distinct categories: naturalpigments which make up a palette of natural polychromaticminerals (or their mixtures) exploited from domestic or foreign oredeposits, and synthetic pigments like Egyptian blue or green fritprepared from natural raw materials and metal scrap by firing themixtures at moderate or high temperatures. Discovery of anunknown pigment is nowadays very improbable and the principaltask in Egyptian pigment research is to assign the observed colour

S.A. El Aal et al. / Vacuum 83 (2009) S4–S8 S5

to one of the already known pigments or their blends.Consequently, qualitative analysis is usually sufficient, providedsome key-elements can be detected that can be recognized asa ‘‘fingerprint’’ characteristic of a given pigment [7–9].

The typical structure of a wall painting is shown in Fig. 1. Itconsists of four layers of different composition and structure. Thefirst layer is the support, which simply is the wall of a tomb ortemple, the second one is a coarse ground layer usually being themixture of gypsum (CaSO4 2H2O), calcite (CaCO3) and quartz (SiO2).The third grounding layer consisted of a fine grain plaster made ofgypsum or lime and the last one was the paint. Paint is a mixture offine grain natural or synthetic pigments with organic binder. Thelast one enables the smooth distribution of pigments on the surfaceand assures adhesion of the paint to the substrate. Typical binderswere Arabic gum, animal glue, or egg yolk [1,2,5].

Continuous degradation of wall paintings leads also to itsmechanical disintegration: many samples exhibit a discontinuouspaint layer because at different places paint patches flaked offexposing the underlying grounding layer. This effect, if not takeninto account, can lead to erroneous interpretation of the experi-mental results.

2. Experimental

2.1. Objects studied

We have analyzed pigments used for wall paintings from variousarcheological sites in Egypt. We report here the results of analysis ofblue pigment samples collected at the Mortuary Temple of

SupportGroundLayers Paint

Fig. 1. Structure of ancient Egyptian wall painting.

Ramesses III (Habu Town), yellow pigment samples from the tombof Tuthmosis III (Valley of the Kings) and blue pigment samplesfrom the tomb of Sennefer (Valley of the Noblemen). All analyzedspecimen were loose pieces of wall paintings collected on the spotin tombs. They consisted of thick painted plaster with the face areausually less than 1 cm2. They were originally 1–10 mm thick,however to facilitate the analysis they were thinned by polishingdown to 1–2 mm thickness. Because of their great brittleness carehas to be taken during the preparation and handling.

2.2. Experimental methods

2.2.1. Optical microscopyAxiotron Zeiss Optical Microscopes at the Institute of Electronic

Materials Technology (ITME) in Warsaw and ForschungszentrumDresden (FZD) were used to visualize the surface of the paintedsamples. In order to visualize the paint cross-section some sampleswere prepared in a special manner. Samples were thinned to about1 mm and subsequently immersed in the fluid Meliodent resin.Meliodent can be obtained by mixing two components: the firstone is powder polymethyl methacrylate and other one is liquidmixture of dimethacrylate and methyl methacrylate. After beingmixed in the correct proportions it solidifies at RT. After solidifi-cation of the resin, the samples were fine grain polished to obtaina flat surface exposing the cross-section of paint. Cross-sectionobservations provided information both about the number ofpainted layers, their sequence and thickness, and also on the size ofpigment grains.

2.2.2. PIXE analysisPIXE is one of the most sensitive micro analytical methods

[10,11]. It is non-destructive and is currently used for multi-elemental analysis both in vacuum and in air [12]. The ‘‘in vacuum’’mode can be advantageous for detection of low atomic numberelements.

Broad beam PIXE analysis with 1.0 MeV protons was performedat SINS, Warsaw using the Lech Van de Graff accelerator. Protonsprior to entering the target chamber were collimated by twographite collimators with a diameter of 1 mm. The beam currentwas between 5 nA and 20 nA, the total charge during the irradia-tion ranged from 0.1 mC to 0.5 mC. The X-rays generated in thesamples were measured by a 7 mm2 Amptek Inc. (Bedford, USA) Si(Li) detector with a 25 mm Be window, placed at 90� or 180� to thebeam direction outside the experimental chamber and an energyresolution (FWHM) of 180–200 eV at 6.4 keV (checked duringexperimental runs by an 241Am standard radioactive source).Samples with dimensions from 2� 3 mm up to 10�10 mm weremounted on 8 position target holder perpendicular to the beamdirection for analysis. Rotation of the target holder by 180� enablessample analysis from both sides i.e. front and back. This wasespecially important for samples with a discontinuous paint layer:the final evaluation of data was performed for the spectrumresulting from subtraction of that taken for front side (face) fromthe other for the support (back). This procedure made it possible toperform qualitative analysis even of partially flaked off paint layer.

2.2.3. m-PIXE analysisThe scanning microbeam-PIXE analysis allows identification of

elements by taking high resolution elemental maps of the sampleby scanning an area of 2 mm square with a narrow beam focused to1 mm [5]. A proton beam of 3.05 MeV from the 3 MV Tandetronaccelerator at FZD was used for m-PIXE analysis. An Ortec Si(Li)detector was located outside the chamber behind a 125 mm thick Bewindows allowing detection of characteristic X-rays for elementsheavier than Na.

S.A. El Aal et al. / Vacuum 83 (2009) S4–S8S6

Since pigment grain sizes are usually of 20–100 mm and therange of 3 MeV protons in a typical pigment compound is above50 mm even m-PIXE did not allow analysis of a single graincomposition without any interference of the substrate and/orneighboring pigments or binder. The analysis was performed inthree steps: first the region of interest was selected in situ with anoptical microscope equipped with a CCD camera, next the beam of5 mm� 5 mm size was scanned across the selected area, usually105�105 mm, to get an elemental map, and finally in the region ofmaximum abundance of elements typical for a given pigment thePIXE spectrum was measured with good statistics.

Although PIXE is in principle a quantitative technique, quanti-fying the spectra, especially for these thick layers that arenonuniform both laterally and in depth, is a tremendous challenge.However, full quantification is unnecessary in this case since thesimple recognition of key-elements is sufficient for pigmentidentification.

Although PIXE is in principle a quantitative technique its cali-bration especially for thick layers that are nonuniform laterally andin depth is a tremendous challenge. Moreover, the windows in frontof the detector and in the vacuum chamber in our experimentalsetup were so thick that transmission of X-rays characteristic for Sior S (about 2 eV) was approximately 80% and 90% respectively.Taking into account the absorption in the sample itself one canestimate that for elements lighter than Ca characteristic X-rays canonly be detected if produced in the surface layer of less than 1 mmthickness. Since for pigment identification the recognition of key-elements is typically sufficient no attempt to quantify PIXE analysishas been undertaken.

3. Results and discussion

From the very large number of analyzed samples here we haveselected only three with most characteristic colours to describe theanalytical procedure and to highlight the virtues of ion beamanalysis of ancient pigments.

3.1. Yellow pigment

The yellow sample was collected at the tomb of Thutmosis III(1504–1450 BC, 18th dynasty) located in the Valley of the Kings.This tomb is famous for its ceiling decoration presenting the blue

2 4 6

Yield

1

10

100

1000

10000

Face Back F-B

Energy (ke

Ti

Ca

Ca

Fe

Fe

Ti

S

Si

1

10

100

1000

Fig. 2. PIXE spectra obtained for yellow pigment sample. Two spectra were taken for each sone the paint composition was determined. m-PIXE spectrum obtained for one selected pig

sky with yellow stars and yellow walls with white stars. Thespecimen studied is a piece from the wall. Spectra of broad beamPIXE for this sample are shown in Fig. 2. The spectrum taken for thesample’s face reveals the presence of elements like Si, Ca, Ti, Mn, Fe,and As. After subtraction of the spectrum for the back side only thelast four were identified as elements present in the paint. m-PIXEspectrum taken for the region of highest As yield is shown in theinset. Somewhat surprisingly large amounts of Fe and Ti as well assome traces of Mn are also present in this location. Optical micro-graphs provided important clues for the solution of this puzzle. Thepaint layer is strongly degraded and discontinuous, however, it wasoriginally of the uniformly yellow colour. Under high magnificationa few darker inclusions in the upper paint layer of about 40 mmdiameter were visualized. In cross-sectional micrographs only onepaint layer can be distinguished. This yellow layer is rather thin i.e.about 30 mm thick with very small pigment grains, below theobservation limit of a few mm.

Basing on the archeological knowledge on the ancient Egyptianpigments [5,6] the following conclusions can be drawn from ouranalysis:

� since the paint layer is discontinuous, the signals from Si and Cacome from the plaster substrate� the presence of the intense Fe signal can be attributed to a deep

yellow paint layer whose pigment was most probably finepowdered mineral goethite, i.e. a-FeO� OH; this so called yellowochre was a frequently used pigment, commonly present inEgyptian soil,� the signal of As has to be attributed to the only known yellow

pigment containing this element, namely orpiment As2S3.� Ti, Mn, Zn and other elements can be found as common

impurities in earth colours like red and brown. In addition, theminerals haematite and goethite, used to obtain ochrepigments, can be found associated to rutile (TiO2), cuprite(Cu2O), and other minerals [14–16].

Due to its brightness and gold shiny glaze ancient Egyptiansrecognized orpiment as a superior yellow pigment. Orpimentcannot be find in any ore deposit in Egypt and has to be imported.The use of this rare and expensive material was a royal privilege.Orpiment was seldom used as a principal pigment, but usuallyused to blend yellow ochre: however, not simply mixed with it

8 10 12V)

As

As

2 4 6 8 10 121

0

0

0

0

SiS

Ca

Fe

Fe

As

As

Ti

Ti

ample: for the face and back side. After subtraction of the face spectrum from the backment grain is shown in the inset.

S.A. El Aal et al. / Vacuum 83 (2009) S4–S8 S7

but applied in a distinct painting sequence, usually as a very thintop layer. Such a sequence has apparently been discovered in thisstudy.

3.2. Blue pigment

Habu Town is an archaeological locality situated on the WestBank of the River Nile opposite the modern city of Luxor. Althoughother structures are located within the area, the location is todayassociated almost exclusively with the Mortuary Temple of Ram-esses III. Ramesses III (1186–1155 BC, 20th dynasty) is consideredthe last great pharaoh. Although the great pharaohs of ancientEgypt were buried in the Valley of the Kings they also built greatmortuary temples to honor their memory and to host the cult thatconnected them with the gods. The investigated blue colour samplecomes from the great statue of Ramesses III.

Blue pigments were frequently used in ancient Egypt. The mostcommon were azurite 2CuCO3$Cu(OH)2 a natural pigment andsynthetic pigment cuprorivaite – CaCuSi4O10 called Egyptian Blue[13]. Egyptian Blue has been known since 2500 BC and is the oldestknown synthetic pigment. It was prepared from natural materialsand fine metal scraps by firing the mixtures at temperatures ofabout 800 �C.

As can be seen in Fig. 3 PIXE spectra revealed that Si, S, Ca, Fe, Cuare present in the surface layer, this was confirmed by the m-PIXEanalysis. Since Cu, Ca and Si are main components of Egyptian Bluethe answer to the question on the blue pigment nature seems to bestraightforward. But since one cannot expect Fe-containingcompounds in this configuration the large Fe signal in all spectra isquite surprising.

Optical microscopy helped to resolve this puzzle. Dark blueEgyptian Blue grains of the size of 10–40 mm are mixed with lightblue grains of more or less the same size. It is known that thetemperature regime of Egyptian Blue synthesis is critical. If the

2 4

Yield

1

10

100

1000

10000

Face

Back

F-B

Ener

1

10

100

1000

10000

Ti

Si

Ca

Ca

Fig. 3. PIXE spectra obtained for blue pigment sample. Two spectra were taken for each samthe paint composition was determined. m-PIXE spectrum obtained for one selected pigmen

temperature raises above 1000 �C instead of Egyptian Bluewollastonite (CaCu) SiO3 of light blue-greenish colour will besynthesized. As a consequence the chromatic hue of preparedpigment was not dark enough. Apparently, to save the preciouspigment the ancient master blended it with very fine grainhaematite Fe2O3 obtaining as expected deep dark blue hue. Veryfine haematite grains are below the optical microscope resolutionbut haematite was detected as the Fe signal by PIXE.

Another blending technique was observed in the preparation ofdark blue pigment used in the tomb of Thutmosis III (not shownhere). To obtain the desired effect the artist produced four-layerpaint structures: first the grounding layer was covered with verythin black layer (probably soot) next came a layer of Egyptian Blue,which in turn was over-painted by the thin haematite layer andfinally a layer of Egyptian Blue was deposited again.

3.3. Green pigment

Selected green pigment samples were also collected in thetomb of Thutmosis III. Corresponding PIXE spectra are shown inFig. 4. Detected elements are Si, S, Ca, Fe, and Cu with some tracesof Mn. From the beginning of the 18th dynasty (1550 BC) theprincipal green pigment was Green Frit (CaCu)SiO3 – wollastonite.It has been synthesized from the same components as EgyptianBlue but requires a very precise estimation of ratios of compo-nents and a much higher temperature. The resulting pigment ismulticomponent: containing Cu-bearing glass phase in variableproportions as well as Green Frit. The chromatic hue of thispigment is strongly dependent on the wollastonite to copper glassratio.

Here again the artist was not content with the pigmentprovided, and blended it with haematite to produce a deep greenhue. This time it was not very fine ground: haematite grains ofabout 10 mm size can be observed by the optical microscope.

6 8 10gy (keV)

2 4 6 8 10

Fe

Fe

Cu

Cu

Cu

Cu

Fe

Fe

Ca

Ca

Si

ple: for the face and back side. After subtraction of the face spectrum from the back onet grain is shown in the inset.

FaceBackF-B 0 2 4 6 8 10

1

10

100

1000

10000

Si

Ca

Ca

Fe

Fe

Mn

Cu

Cu

Cu

Cu

Fe

Fe

Mn

Ca

Ca

Si

Yield

1

10

100

1000

10000

2 4 6 8 10Energy (keV)

Fig. 4. PIXE spectra obtained for green pigment sample. Two spectra were taken for each sample: for the face and back side. After subtraction of the face spectrum from the backone the paint composition was determined. m-PIXE spectrum obtained for one selected pigment grain is shown in the inset.

S.A. El Aal et al. / Vacuum 83 (2009) S4–S8S8

4. Conclusions

Qualitative PIXE, m-PIXE and optical microscopy weresuccessfully applied to identify the pigments used in ancientEgyptian wall paintings. Our observations can be summarized asfollows:

� the yellow pigment was identified as goethite FeO(OH) coveredwith thin surface layer of orpiment As2S3 to give it more shineand brightness,� the dark blue pigment was composed of a mixture of Egyptian

Blue (CaCuSi4O10) and Green Frit blended with haematiteFe2O3,� the green pigment is Green Frit (CaCuSiO3) blended with

haematite to obtain a specific hue of the green colour.

Besides pigment identification we are also able to draw someconclusions about the painting technique of ancient masters. Theyused various methods to change the chromatic hue of principalpigment, including:

� adjusting grain size by appropriate grinding,� blending with other pigments,� creating multilayer structures.

Analysis of these selected samples clearly demonstrates theusefulness of ion beam techniques for this kind of analysis.Although the results have not always led to unambiguous conclu-sions many important clues were provided by microscopic obser-vations. Moreover, complementary data from X-ray diffraction andRaman spectroscopy would greatly enhance this analysis. Sucha study is currently being carried out.

Acknowledgment

Sincere thanks are due to Ms. M. Mozdzonek (ITME) andDr. B. Schmidt (FZD) for their help with optical microscopy.

Microbeam-PIXE analysis has been carried out at the AIM of theInstitute of Ion Beam Physics and Materials Research of the For-schungszentrum Dresden-Rossendorf within the framework of thespecific research and technological development program of theEuropean Community ‘‘Structuring the European Research Area:Research Infrastructures Transnational Access’’ (RITA ContractNumber 025646).

References

[1] Nicholson PT, Shaw I. Ancient Egyptian materials and technology. CambridgeUniversity Press; 2000.

[2] Lucas A. Ancient Egyptian materials and industries. Kessinger Publishing; 2003.[3] Lindborg U. Conservation of mural paintings. National Heritage Board; 2001.[4] Clark JH, Chimie CR. 2002; 5:7–20.[5] El Goresy A. In: Nomikos PM, editor. Proceedings of the First Thera Interna-

tional Symposium; 1997. p. 49–70. Thera, Greece.[6] Eastaugh N, Walsh V, Chaplin T, Siddall R. Pigment compendium: a dictionary

of historical pigments. Butterworth-Heinemann; 2004.[7] Uda M. Nucl Instrum Methods 2004;B226:75–82.[8] Adriens A, Dowsett MG. Appl Surf Sci 2006;252:7096–101.[9] Vandenabeele P, Garcia-Moreno R, Mathis F, Leterme K, Van Elslande E,

Hocquet F-P, et al. Spectrochim Acta A: Mol Biomol Spectrosc 2008;.doi:10.1016/j.saa.2008.07.028.

[10] Johansson SAE. Analyst 1992;117:259–65.[11] Johansson SA, Campbell JL, Malmqvist KG. Particle-induced X-ray emission

spectrometry (PIXE). Wiley-Interscience; 1995.[12] Neelmeijer C, Mader M. Nucl Instrum Methods 2002;B189:293–302.[13] Bianchetti P, Talarico F, Vigliano MG, Ali MF. J Cult Herit 2000;1:179–88.[14] Calza C, Anjos MJ, Mendonça de Souza SMF. Appl Phys A 2008;90:75–9.[15] Paternoster G, Rinzivillo R, Nunziata F. J Cult Herit 2005;6:21.[16] Uda M, Sassa S, Yoshimura S, Kondo J, Nakamura M, Ban Y, et al. Nucl Instrum

Methods 2000;B161–B163:758–61.