Journal of Archaeological Science 35 (2008) 2474–2485

Contents lists available at ScienceDirect

Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas

From Byzantine to post-Byzantine art: the painting technique of St Stephen’s wall paintings at Meteora, Greece Sister Daniilia a, *, Elpida Minopoulou a, Konstantinos S. Andrikopoulos a, b, Andreas Tsakalof a, c, Kyriaki Bairachtari a, d a

‘Ormylia’ Art Diagnosis Centre, Sacred Convent of the Annunciation, 63071 Ormylia, Chalkidiki, Greece Physics Division, School of Technology, Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece University of Thessaly, Department of Medicine, Papakiriazi 22, 41222 Larisa, Greece d National Centre for Scientific Research ‘Demokritos’, Aghia Paraskevi, GR-15310 Attiki, Greece b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 June 2007 Received in revised form 26 February 2008 Accepted 21 March 2008

The old katholikon of St Stephen’s monastery at the Meteora (site of the most important complex of monasteries in Greece after Mount Athos) is decorated with wall paintings that date from the beginning of 17th century. In terms of style, the artistic ensemble is altogether characteristic of the period. The painting technique has been examined by means of mRaman and mFTIR spectroscopies, gas chromatography–mass spectroscopy (GC/MS), optical microscopy (OM) and scanning electron microscopy (SEM). Prior to the commencement of restoration treatment, and in order to optimise its effect, it was considered prudent to identify the materials and ascertain the techniques that had been used to apply the plaster and the paint layers. It was noted that whereas the ariccio consists of yellow clay and straw, the intonaco contained calcite. The painter’s palette is made up of eight pigments: calcite, carbon black, yellow ochre, haematite, green earth, cinnabar, smalt and malachite. The stratigraphy and the scale of the shades differ significantly from those in works of the Palaeologan period (1261–1453) – indicative both of evolution in Byzantine iconography as a result of gradually changing religious and social circumstances, and of the skill and vision of the painter. In addition, some decay products, such as gypsum, were detected. In that they conceal important artistic details, this necessitates proper consolidation, cleaning and conservation treatment in order to restore to some degree the original splendour of the wall paintings. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Post-Byzantine art Meteora Egg Fresco–secco Smalt Efflorescence Spectroscopies GC/MS

1. Introduction Meteora (Greek: M3seura, ‘suspended in the air’) is one of the largest and the most important complexes of monasteries in Greece after Mount Athos. Its edifices are constructed on spectacular natural sandstone rock megaliths at the northwestern edge of the plain of Thessaly. The history of these monasteries was first mentioned from the 11th century when the first hermits settled on these ‘columns of the sky’. Although more than 20 monasteries were built at the time of the great 14th century revival of the monastic life, only six remain today: the Great Meteoron (or Holy Transfiguration), Varlaam, St Stephen, Holy Trinity, St Nicholas Anapafsas

Abbreviations: PCA, principle component analysis; Ala, alanine; Gly, glycine; Val, valine; Leu, leucine; Ile, isoleucine; Ser, serine; Met, methionine; Thr, threonine; Phe, phenylalanine; Asp, aspartic acid; Glu, glutamic acid; Tyr, tyrosine; Pro, proline; Hyp, hydroxyproline. * Corresponding author. Tel.: þ30 23710 98400; fax: þ30 23710 98402. E-mail address: [email protected] (S. Daniilia). 0305-4403/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2008.03.017

and Rousanou. Their 16th–17th century murals mark a key stage in the development of post-Byzantine art. The rock monasteries have been regarded by UNESCO as a unique phenomenon of the world’s cultural heritage and they form one of the most important stations on the cultural map of Greece. Built in the 14th century, the monastery of St Stephen is located near the outskirts of modern Kalambaka (Fig. 1). Its old katholikon, dedicated to St Stephen the Protomartyr, was probably built at the time of the monastery’s foundation or shortly thereafter. In 1545, it was rebuilt by St Philotheos, second founder of the monastery. This second building phase is a small, low, timber-roofed, single-naved basilica with a narthex. During the World War II (1939–1945) and the Civil War (1946–1949) the church suffered more damage than in the five previous centuries together, for example, the defacing of all the saints, especially their eyes. The wall paintings of the interior, initially retouched and restored three decades ago, form an interesting ensemble of postByzantine painting. Scenes of the Virgin ‘Platytera’ and of the Communion of the Apostles (above the sanctuary), the full-length

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

2475

Given the absence both of bibliographic resources and of recent studies on the painting techniques in the wider region of Epirus of the post-Byzantine period (17th century), the results of this research achieve their anticipated significance. Finally, comparison of the results with those from the Protaton (1290), Mount Athos (a representative monument of the Byzantine era) reveals significant stylistic and technical differences. 2. Experimental 2.1. Samples

Fig. 1. The Holy monastery of St Stephen at the Meteora, Greece.

portraits of the saints and the representations in the nave of the 24 stanzas of the Akathistos Hymn are all equally arresting. Finally, in the narthex, the holy founders Anthony and Philotheos and of the Dormition of the Virgin Mary are depicted. While no attribution to the founder (the righteous Philotheos) or to the artist of the paintings’ first stage is noted in the monastery’s dedicatory inscription, commemorations do exist for the patrons of the earliest artwork – the abbot Mitrophanis and the hieromonk Grigorios (Fig. 2). The same inscription refers to the name of the second painter, the priest Nikolaos from Kalambaka, who completed the subsequent phase of the wall paintings. On the bases first, of the manner in which the iconographic programme is worked out and secondly, of the stylistic and morphological features of the figures portrayed in the wall paintings, it is likely that the first phase of the artwork dates from between the second and third decade of the 17th century. The second phase was completed soon after, that is to say, around the middle of the same century. The wall painting complex in St Stephen’s katholikon clearly belongs to a tradition known to us from a great many painted churches in Epirotic Greece and beyond (Vitaliotis, 1998). The present study concerns itself with two important aspects of the wall paintings in St Stephen’s monastery: (a) disclosure and classification of their painting techniques as they relate to traditional practices of the Byzantine era proper and (b) description of their state of preservation aiming at the consolidation, removal of soot and salts, and restoration of the aesthetic integrity of the murals. Analyses were undertaken by means of mRaman and mFTIR spectroscopies, gas chromatography–mass spectroscopy (GC/MS), optical microscopy (OM) and scanning electron microscopy (SEM/EDS).

The samples under investigation were acquired from a variety of scenes and zones in the church (Table 1) and were chosen for the purpose of identifying the material elements that make up the plaster, the pigments, the binding media, and the residual salts on the wall painting surface. Moreover, attention was given to locate and describe the painting techniques (stratigraphy, pigment mixture, colour combinations) and more generally the particular artistic features in post-Byzantine wall paintings of this period. 2.2. Methodology 2.2.1. Optical microscopy Samples were mounted in polyester transparent resin and the cross-sections were ground and polished using a Struers PlanopolV machine. Observation and photography of the samples’ surface before embedding and of their cross-sections were achieved using a Zeiss Axiotech 100 HD polarising microscope, equipped with white reflected and ultraviolet light as well as with a SPOT 2 1.4 digital cooled camera (res.: 1315  1033 pixels, 12 bits per colour). 2.2.2. mFTIR spectroscopy FTIR spectra were measured with a Biorad FTS 175 FTIR spectrophotometer equipped with a UMA 500 microscope. Powdered samples pressed into KBr pellets and micro-samples placed in a DC2 Graseby-Specac diamond compression cell were analysed in absorbance mode at 4 cm1 resolution. 2.2.3. mRaman spectroscopy A Renishaw System 1000 micro-Raman spectrometer comprising an Olympus BH-2 imaging microscope, a grating monochromator and a charged-coupled device (CCD) Peltier-cooled detector were employed for Raman spectra acquisition. An HeNe laser (632.8 nm) served as the excitation source and the beam Table 1 Description of sampling position Sample Point in the church

Scene

Sampling position

STA1

Northern wall, sanctuary

St Germanus

STA2

Northern wall, sanctuary

Flesh tone in right hand Brown background

STA6 STA8 STA10 STA11 STA12 STA13 STA14 STB1

St Peter of Alexandria Eastern facet, north pessary The Virgin of Supplication Western wall, nave (1st zone) Archangel Gabriel Northern wall, nave (2nd zone) Resurrection Southern wall, nave (3rd zone) Palm Sunday Eastern wall, sanctuary St Basil the Great Southern facet, north pessary St Orestes Western facet, south pessary Western facet, south pessary Northern wall, nave (1st zone) St Mercurius

STB2

Western wall, nave (2nd zone)

STB3

Eastern facet, north pessary

STB4

Northern wall, nave (1st zone)

STA5

Fig. 2. The dedicatory inscription above the transom of St Stephen’s west door.

The Dormition of the Virgin The Virgin of Supplication St Demetrius

Flesh tone in neck Red sticharion Christ’s purple tunic Brown background Grey background Plaster in right hand Yellowish plaster Grey background Light in olive-green tunic Green background Red light in mantle Blue tunic

2476

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

emitted was focused using a 100 microscope objective. Low laser power (up to 1 mW) was used for spectra accumulation. A set of two notch filters with a cut-off edge of wþ100 cm1 was employed for the rejection of Rayleigh scattered light. The resolution was kept at w5 cm1. 2.2.4. Gas chromatography–ion trap mass spectrometry (GC/MS) A Polaris gas chromatograph–ion trap mass spectrometer provided with an AS2000 autosampler (ThermoFinnigan, San Jose, USA) was used for amino acid quantification in reference and real samples. The gas chromatographer was equipped with split/splitless injector and ATÔ-5MS 30 m  0.25 mm column with 0.25 mm film thickness of 5% phenyl–95% methylpolysiloxane stationary phase (Alltech Associates, USA). The samples ware extracted twice with CHCl3 in order to separate possible natural resins, lipids and free fatty acids from the proteinaceous fraction. The solid residue, once dried, was extracted twice with 2.5 N ammonia solution for 3 h at 60  C in an ultrasonic bath. The extract was subsequently subjected to acid hydrolysis assisted by microwaves in order to free the amino acids, which were derivatised with N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide and then 1 ml (of each derivatised solution) was analysed by GC/MS (Daniilia et al., 2007).

The stratigraphy of the plaster is clearly made out in the crosssection of the sample taken from St Orestes’s right hand (STA12). The high degree of diffusion in the interface of the two layers indicates that the intonaco was probably applied while the underlying ariccio was damp. Here, the presence of yellow clay in the ariccio is impressive; it replaces lime which, as is well known, constituted the chief ingredient of plaster in most Byzantine monumental painting. It is likely that yellow clay existed abundantly in the surrounding area and was therefore preferred for financial reasons. For the identification of plaster compounds mFTIR spectra were acquired from two homogenous, pulverized samples: (a) the yellowish ariccio and (b) the whitish intonaco. mRaman spectroscopy was applied additionally for the identification of certain compounds. In the FTIR spectrum of ariccio (STA13) (Fig. 4, Table 2) the characteristic peaks of kaolin and calcite were recorded denoting the presence of yellow ochre (Ganitis et al., 2004; Pavlidou et al., 2006; Zorba et al., 2006). mRaman spectroscopy, on the other hand, revealed the presence of goethite (main compound of yellow ochre) (Fig. 4, Table 2) (Koszowska et al., 2005). The FTIR spectrum of intonaco (STA14) consisted of calcite (Gilbert et al., 2000) and quartz (a natural admixture) (Fig. 5, Table 2) (Goodall et al., 2006). 3.2. Pigments and layer structure

2.2.5. Scanning electron microscopy–energy dispersive system (SEM/EDS) Scanning electron microscopy (SEM) was carried out using a JEOL 6300 scanning microscope equipped with an energy dispersive X-ray spectroscopy (EDS) ISIS 2000 microanalytical system. The elemental composition was determined using the prepared carbon coated cross-sections, which were bombarded by a strongly accelerated and focalised electron beam in a vacuum (105 torr). 3. Results and discussion 3.1. The plaster Macroscopic investigation of the wall paintings’ surface at points of extensive damage and detachment permitted the detection of two layers of plaster of differing composition. The first, ariccio, of a yellowish shade, contains a considerable amount of straw; the second, superficial intonaco has a whitish hue (Fig. 3).

3.2.1. The background 3.2.1.1. Blue. The background of Byzantine wall paintings is habitually divided into two zones of unequal height, the upper one blue (larger) and the other of an olive-green shade. In post-Byzantine monuments, however, one sees a penchant for separating the field into three zones: the upper blue, the middle green (larger) and the lower brown. This artistic particularity – aside from its symbolic character – is most probably motivated by the practical need to restrict the space allotted to the especially costly pigments in the blue field. The grey-blue background in St Stephen’s wall paintings is executed with a familiar, traditional technique that had been used for centuries in murals of the Comnenian, Palaeologan and Cretan styles (Daniilia et al., 2000; Zorba et al., 2006; Pavlidou et al., 2006). Applied over the uppermost white plaster is a primary, dark covering – a grey layer from carbon black with a very small amount of

Fig. 3. Detail of St Orestes’s right hand (detection of two layers of plaster and strands of straw in the yellowish one) and cross-section of the underlying plaster (magnification at 100 and 200).

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

2477

Fig. 4. mFTIR and mRaman spectra of ariccio indicating the characteristic peaks of (a) kaolin and calcite and (b) goethite.

lime – in which scattered grains of haematite, cinnabar and yellow ochre could be traced. Above this the blue pigment exhibits angular crystalline grains in a variety of sizes, in hues from deep blue to greyish azure (Fig. 6). This technique has also been reported in medieval wall paintings (Mugnainia et al., 2006).

Smalt and calcite were initially identified with mFTIR spectroscopy (Gilbert et al., 2000; Ganitis et al., 2004). Smalt is an artificial, glass-like potash silicate pigment, which is strongly coloured with cobalt oxide and reduced to a powder. Evidence of its use as a painter’s pigment dates from the 15th to early

Table 2 Materials identified with FTIR and Raman spectroscopies Technique

Sample

Material

Peaks (cm1)

Assignments

FTIR

STA13 ariccio

Calcite

1424w 874w 3698br, 3619br 1032br, 1011br, 913m 693w, 527m, 467s, 430s

CO2 3 , stretching CO2 3 , bending O–H, stretching Aluminosilicates, stretching Aluminosilicates, bend

1416br 874s 1034br, 800w

CO2 3 , stretching CO2 3 , bending Si–O–Si, stretching and bending

2516w, 1796w 1430br 874s, 712m 1084w 781w, 464w

CO2 3 , overtone and combination bands CO2 3 , stretching CO2 3 , bending Si–O, stretching Si–O, bending

2514w, 1800w 1437br 874s, 714m 1036br, 976m, 453m

CO2 3 , overtone and combination CO2 3 , stretching CO2 3 , bending Silicates, stretching and bending

3544m, 3403m, 3363s, 3197m 1661s, 1634s 1138m, 1117m 3008w, 2923s, 2853m 1742w 1470m, 1424m

O–H stretching O–H bending S–O stretching C–H stretching C]O stretching C–H bending

Kaolin

STA14 intonaco

Calcite Quartz

STA14 paint layer

Calcite

Smalt STB2

Calcite

STA11

Gypsum

Green earth

Wax

Raman

STA13 ariccio

Goethite

304w, 401br, 563m

Fe–O bands

STA10

Goethite Red ochre

303w, 407br, 560m 225m, 245w, 292s, 411m, 499w, 611m, 661w 1321br, 1600br

Fe–O bands Fe–O bands

Carbon black

C–C stretching

STB2

Celadonite

215m, 271m, 393m, 457w, 551br, 700m, 908w

R–O–H bands (R: Al, Feþ2, Feþ3, Mg)

STB1

Malachite

3350br, 1052m 1486w, 1085m 533w, 431s 354w 265m, 217w, 178m

O–H stretching CO2 3 stretching Cu–O stretching Cu–O bending O–Cu–OH bending

STB3

Calcite Cinnabar

1086w 253vs, 344m

CO2 3 stretching Hg–S stretching

vs, very strong; s, strong; m, medium; w, weak; br, broad.

2478

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

amounts of carbon black, yellow ochre, haematite and cinnabar, while the second contains green earth and lime, with the result that the final shade is more luminous. Green earth could be detected with both mFTIR and mRaman spectroscopic techniques (Fig. 7c, d, Table 2) (Daniilia et al., 2000; Zorba et al., 2006; Pavlidou et al., 2006; Daniilia and Andrikopoulos, 2007). 3.2.1.3. Brown. The background’s lower brown zone is rendered in carbon black with red and yellow ochres (Table 2). 3.2.2. The garments

Fig. 5. mFTIR spectrum from the intonaco with the characteristic peaks of calcite and quartz.

19th century, notwithstanding the employment of cobalt ores in ancient Egypt and in classical times for colouring glass. In the early 17th century smalt is mentioned as being of widespread use in oil painting, substituting lapis lazuli and azurite as they became more and more scarce (Gettens and Stout, 1966; Muhlethaler and Thissen, 1993). Today, unlike the grey background underpaint in the church’s wall paintings which is visible, the upper layer of smalt is preserved in but a few areas. Larger smalt grains have become detached from the painting and this may be attributed either to accidental removal during the several rescue interventions undertaken for the purposes of conservation or to a weakening of the binding power of the medium. Analogous detachment occurs in many medieval art monuments where azurite is used as the blue pigment (Daniilia et al., 2006; Sotiropoulou et al., 2007). 3.2.1.2. Green. In the background’s olive zone the paint layers are applied directly onto the white plaster without the intermediary grey layer commonly found in earlier wall paintings (Fig. 7b) (Daniilia et al., 2000). In the cross-section two paint layers can be discerned (Fig. 7a). In the first, green earth is mixed with small

3.2.2.1. Blue. The grey-blue hues of the garments are executed in a technique similar with that seen in the blue background. Carbon black and gains of haematite were found in the underpaint (the tunic of St Demetrius, STB4), whereas the lights were rendered with mixture of smalt and calcite (Fig. 8). Furthermore, many microscopic, almost colourless, grains of smalt, were recorded most clearly under ultraviolet light, allowing the charting of their distribution throughout the paint layer. The identification of smalt used at St Stephen’s was carried out by SEM/EDS analysis on a cross-section, verifying the results obtained by mFTIR spectroscopy (Fig. 6). In the EDS spectrum the peaks of Co, Si and K, main elements of smalt, as well as Al, Ca and Fe (in some cases As), were recorded (Fig. 8c). Elements such as Fe and As indicate the origin of CoO from the mineral cobaltite (Co,Fe)AsS, given the fact that there is no evidence of Ni (an element contained in the mineral smaltite [Co,Ni]As3–2) in the EDS spectrum (Muhlethaler and Thissen, 1993). According to scholarly opinion the composition of smalt varies considerably in SiO2 (65–71, 66–72), K2O (16–21, 10–21), CoO (6–7, 2–18) and in impurities of other oxides (Al, Ba, Ca, Cu, Fe, Mg, Mn, Ni, Na) (Muhlethaler and Thissen, 1993). EDS microanalyses of two grains of smalt show the following mean composition: (a) silicon (68.93%), potassium (22.59%), cobalt (4.09%), calcium (2.16%) and iron (2.492%); (b) silicon (85.42%), potassium (3.89%), cobalt (3.25%), calcium (1.11%), iron (2.05%), aluminium (1.63%) and arsenic (2.66%). The remarkable variations between the composition of the two types of smalt used suggest the existence of two different supplies of the pigment. The painter used a low quality, finelyground smalt in the underlayers, whereas the better quality, large blue grains were applied on the uppermost layers. Notably, recent studies have shown that smalt, when used in oil paint, gradually discolours owing to the loss of potassium (this is

Fig. 6. (a) Photomicrograph of the background sample where smalt is applied over a grey underpaint and (b) mFTIR spectrum depicting characteristic peaks of smalt and calcite.

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

2479

Fig. 7. Cross-section from the olive-green zone in the background in (a) St Stephen’s and (b) the Protaton. (c, d) mRaman spectrum of green earth and mFTIR spectrum of green earth and calcite.

Fig. 8. Cross-section from St Demetrius’s grey-blue tunic in (a) reflected and (b) UV light. (c) SEM spectrum from the blue pigment.

2480

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

designated as a leaching process) (Boon et al., 2001). Although analogous smalt deterioration has been identified in 16th century Roman wall paintings (whether fresco or secco is not reported) (Santopadre and Verita`, 2006), and whereas smalt discolouration also occurs in 17th century wall paintings executed both in fresco or fresco–secco technique (Ajo` et al., 2004), no indication of this phenomenon has been noted in St Stephen’s church. 3.2.2.2. Green. Of particular interest is the painting technique displayed in the olive-green garments. The dark olive tone of the underpaint is combined with intense green highlights that tail off into fine white brushstrokes (Fig. 9a). In the cross-section of the sample from St Mercurius’ short tunic (STB1) and above the layer of the preliminary drawing – a mixture of lime and carbon black – one detects the layer of the underpaint, a mixture of yellow ochre and carbon black (Fig. 9b). In the first light gradation use is made of coarse-grained malachite, which contains natural admixtures of azurite grains, haematite and yellow ochre. The mRaman spectrum of the green pigment was indicative of malachite (Fig. 10, Table 2) (Frost, 2006; Frost et al., 2007; Bordignon et al., 2007). The particular combination of cool green highlights and warm olive-green underpaints is a technique known to painters from all eras. In Byzantine times, however, monumental painting mainly employs green earth in unmixed gradations (Fig. 9c, d) (Daniilia et al., 2000). Again in the post-Byzantine period copper pigments (malachite, verdigris), which offer more intense and more saturated green shades (Lelekova, 2006), are introduced to painters’ palettes. 3.2.2.3. Red. For the underpaint of the chestnut-red garments (Virgin’s mantle, STB3) pure haematite, mixed with cinnabar in the first light, is used whereas the final highlight is executed in thick brushstrokes of pure cinnabar (Figs. 11a, b and 12, Table 2) (Vandenabeele et al., 2005; Daniilia and Andrikopoulos, 2007). The presence of grains of lime indicates the presence of lime water which is used as a binder, following fresco painting technique. Applying this particular stratigraphy in the red garments at St Stephen’s the painter underscores one of the most significant differences between this and the artistic style of the Byzantine period. In the latter, red ochre is found in the underpaints, while for the

Fig. 10. mRaman spectrum of malachite.

two gradations of the lights pure cinnabar and minium are used, respectively (Fig. 11c, d). The bright red tones of the clothing are rendered in cinnabar which, for the highlights, is mixed with lime (Fig. 13). In the varied chestnut and pink shades of the different garments mixtures of haematite and yellow ochre are used in dissimilar proportions, with, in certain instances, the further addition of lime and carbon black. 3.2.3. The flesh tones The artist of the St Stephen wall paintings follows faithfully the traditional manner of fashioning the flesh areas, but with deviations in his choice of pigments and, accordingly, in the final shading. The tones both of the underpaint and of the flesh (Fig. 14a) vary significantly from those found in 14th century Byzantine wall paintings (Fig. 14c), and are characteristic of the style not only of the period but also of this particular painter (Daniilia et al., 2000, 2008).

Fig. 9. (a, b) Detail of St Mercurius’s olive-green tunic in St Stephen’s church and cross-section from the green light. (c, d) Detail of a green garment in the Protaton and cross-section from a highlight.

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

2481

Fig. 11. (a, b) Detail of the Virgin’s mantle in St Stephen’s church and cross-section of a highlight. (c, d) Detail of a red garment in the Protaton and cross-section of a highlight.

Unlike the mixture of green earth and yellow ochre used for the underpaint of the Protaton’s wall paintings (Fig. 14d), here, the dark brown underpaint consists of haematite, yellow ochre, carbon black and grains of cinnabar (Fig. 14b). Similarly, the characteristic pink shade for the flesh in St Stephen’s (a mix of lime and cinnabar) is significantly different from that used in Protaton (a mix of lime and yellow ochre). At the same time, it is noteworthy that differences in the thickness of the paint layers between the wall paintings of the Protaton and those in St Stephen’s katholikon – both for the underpaint and for the flesh – result not only from the kind of pigments used (their hiding power, granulometry) but also from the desired aesthetic effect and fingerprint of each artist.

of painting on damp plaster whereby a restricted number of pigments (chiefly earthy) is mixed simply with water or limewater; al secco, on the other hand, denotes painting executed on dry plaster using pigments that are blended with one or another organic binder (Winfield, 1968; Daniilia et al., 2007). Macroscopic examination of the wall paintings’ surface texture and of the modelling of the brushstrokes leads to the hypothetical conclusion that the St Stephen painter employed a mixed technique of fresco–secco. On the other hand, microscopic observation of the cross-sections together with mFTIR spectroscopy have verified (a)

3.3. Binding medium In order to introduce the section on binding media, it is useful to clarify certain specific terminology: al fresco indicates a technique

Fig. 12. mRaman spectrum of cinnabar (Fig. 11a). The characteristic peak of calcite at 1086 cm1 is also recorded.

Fig. 13. Detail of Archangel Gabriel’s garment and cross-section from the red underpaint.

2482

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

Fig. 14. (a, b) Detail of the Virgin’s face in St Stephen’s church and cross-section of the flesh tone. (c, d) Detail of a face in the Protaton’s wall paintings and cross-section of the flesh tone.

extensive use of lime both as a white pigment and as a binder (in the form of lime water) and (b) different degrees of diffusion between the paint layers and the underlying plaster (Daniilia et al., 2007). In a few circumstances the grains penetrate the plaster (Fig. 15, left), a fact that reinforces the aforementioned opinion about painting on damp plaster, while at other times the paint layers separate conspicuously, which is an indication of painting

employing al secco technique, that is, with the use of an organic binder (Fig. 15, right). According to received tradition and supported by scholarship and data stemming from analyses carried out on samples from Byzantine monuments, proteinaceous materials (egg, animal glue, casein) were used as binders (Winfield, 1968; Dionysios of Fourna, 1996; Pavlidou et al., 2006; Daniilia et al., 2007). Recent

Fig. 15. Cross-sections of several shades in the St Stephen wall paintings. Left: the first paint layer has been applied on wet plaster (intense diffusion). Right: the absence of diffusion implies the application of al secco technique.

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

publications have presented the results achieved by employing analytical methods both to identify organic materials in artworks and to measure the effect of ageing of those materials in their identification (Colombini et al., 1998, 1999a; Rampazzi et al., 2002). In order to study the GC/MS results of St Stephen’s wall paintings (samples STB3a, STB3b, STB4, STA5) it was first necessary to analyse the reference substances of egg, animal glue and casein (Daniilia et al., 2007). Recognition of the nature of the protein binding medium was achieved by a quantitative determination of amino acids (Table 3); their characteristic ratios were examined (Table 4) and principle component analysis (PCA) was applied in the percentage content of the amino acids (Fig. 16). Table 3 summarises the amino acids’ content, which fluctuated between 0.08–0.81% (w/w). The characteristic ratios (Pro/Asp and Glu/Pro) concur with those of egg. The small differentiations (Gly/ Glu, Gly/Asp) are most probably due to the presence of inorganic pigments in the St Stephen samples (Halpine, 1992; Colombini et al., 1998, 1999b; Daniilia et al., 2007). It is noticeable that in the analysed samples there is evidence neither of animal glue (absence of hydroxyproline) (Casoli et al., 1996; Halpine, 1992), nor of casein (ratio of Pro/Asp is <1) (Daniilia et al., 2007). In the PCA score plot of the relative percentage amino acids, the St Stephen samples are found nearest to the cluster of pure egg (Fig. 16). This observable shift is probably due to the influence of the inorganic pigment content in the samples (Daniilia et al., 2007). Apart from the inorganic pigments, the shift can also be attributed to natural ageing of the egg protein (Rampazzi et al., 2002), given that the wall paintings were completed at the beginning of the 17th century. Colombini et al. (2000) have demonstrated that artificial ageing does not significantly affect the amino acid profile of protein binders; consequently, protein binders in old paintings can be reliably identified by comparing the amino acid composition with that of reference paint materials which have not aged. In conclusion, the results of the entire analysis have shown that egg was used not only as a binding medium for the paint layers but also as a component of the final plaster (intonaco), a fact that concurs with the surviving oral tradition in the wider region of Thessaly.

3.4. Efflorescence A close examination of the nature of the decay in wall paintings reveals soluble salts as the main cause of deterioration. Nitrates, oxalates and sulphates are known to be among the most harmful soluble salts (Arnold and Zehnder, 1985; Wust and Schluchter, 2000). Under the influence of environmental conditions (temperature, humidity, light, atmospheric pollutants, micro-organisms, etc.) these salts are subjected to cycles of crystallization–dissolution, leading to mechanical stresses and chemical alterations that

2483

Table 4 Amino acid concentration ratios in samples from St Stephen’s Amino acid ratios of reference egg

Gly/Glu <1

Gly/Asp <1

Pro/Asp <1

Glu/Pro >2

STB4 STA5 STB3a STB3b x SD

0.9 1.3 1.2 0.9 1.1 0.2

2.7 2.7 2.4 2.3 2.5 0.2

0.6 0.4 0.4 0.4 0.5 0.1

4.9 4.9 4.9 6.8 5.4 1.0

can result in flaking and powdering of both paint layer and plaster (Mora et al., 1984). One of the most important questions that have to be settled, when planning a conservation methodology, is how to identify the salts present as contaminants and pollutants in the painting. Since their removal is crucial in cleaning painted surfaces, conservators need to know precisely the kind of salts in question. Calcium sulphate dehydrate (gypsum) is the most commonly encountered pollutant in carbonatic wall paintings. The formation of gypsum during sulphation involves the dry deposition reaction between calcite (CaCO3) and sulphur dioxide (SO2) gas, in the presence of high relative humidity, an oxidant and a catalyst (Fe2O3 or NO2). In this case gypsum is detected on the surface (Pavlidou et al., 2006). The widespread presence of salts, visible as a whitish irregular film in several places in St Stephen’s wall paintings, is, in certain representations, especially evident in that it screens the intensity of the original colour and masks details in the paint (Fig. 17a). The nature of the salts has been identified by mFTIR on a sample taken from the grey background beside St Basil (STA11). The mFTIR spectrum (Fig. 17b, Table 2) reveals the presence of gypsum (CaSO4$2H2O) (Grassi et al., 2007). The sample also contains wax (Mazzeo et al., 2006), which might have been used during a previous conservation attempt. 4. Conclusions The wall paintings in the katholikon of St Stephen’s monastery constitute a representative example of post-Byzantine monumental art at the beginning of the 17th century. Their anonymous artist echoes in general terms the tradition of contemporary monumental art in the wider region of central and northern Greece. He was evidently quite familiar with previous styles, which influenced his

Table 3 Mean values of the relative percentage content of amino acids in St Stephen’s samples Amino acids

STB4

STA5

STB3a

STB3b

Ala Gly Val Leu Ile Met Ser Thr Pro Phe Asp Glu Hyp Tyr

9.2 20.2 4.1 6.5 3.4 0.1 10.9 7.8 4.5 3.3 7.4 22.1 0.0 0.4

9.3 19.9 4.4 6.7 3.5 0.3 16.6 8.5 3.3 4.0 7.3 15.9 0.0 0.3

7.5 22.1 3.7 5.9 2.9 0.1 14.8 7.6 3.8 3.2 9.3 18.8 0.0 0.3

5.7 22.1 5.0 8.9 4.6 0.0 7.3 4.9 3.5 4.2 9.6 23.6 0.0 0.6

Fig. 16. PCA of amino acid percentage data of reference samples (egg, animal glue, casein) and wall painting samples from St Stephen’s church.

2484

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485

Fig. 17. (a) St Basil the Great and (b) mFTIR spectrum of the whitish film formed over the paint surface.

productions. But in certain aspects he took an independent path. For example, he not only augmented his palette by importing new pigments of more recent circulation in Europe, but also developed new stylistic features that characterise his epoch and his own, personal e´lan. Noteworthy particularities vis-a`-vis medieval Byzantine paintings have been revealed. These relate to the materials as well as their technical application. For the plaster, by way of example, use is made on the one hand of the ariccio yellow clay with a modest amount of lime mixed with straw, and on the other of the intonaco lime. The artist’s palette includes (besides what is customary): lime (CaCO3), carbon black (C), yellow ochre (Fe2O3$H2O), haematite (Fe2O3), green earth (K[(Al(III), Fe(III))(Fe(II), Mg(II))], [(AlSi3, Si4)O10(OH)2]), and cinnabar (HgS), smalt (CoO$SiO2$K2O) and malachite [CuCO3$Cu(OH)2]. An interesting point is the presence of smalt in the blue field, instead of the more commonly used readily available azurite or the rarer lapis lazuli. Moreover, aside from green earth, employed very widely and the sole green pigment for wall paintings in the Byzantine period, St Stephen’s representations make use of malachite, a pigment especially selected for highlighting the garments in order to achieve further saturation of the shades. The artwork begins on damp plaster and is completed by means of secco technique which uses egg as the binding medium. It is important to note that egg is also seen to constitute a component of the intonaco. The wall paintings under surveillance, beyond the vandalism which they suffered during the World War II and the Civil War, exhibit problems in preservation as a result of the salt (gypsum) that has formed on the painted surface. This whitish layer of salt, covering many areas of the wall paintings, affects significantly the aesthetic impact of the post-Byzantine, old katholikon of the monastery of St Stephen at the Meteora. Acknowledgements The authors would like to thank the Holy Monastery of Agios Stephanos at Meteora for the approval to analyze samples from the wall paintings, as well as Mrs Dimitra Lazidou, conservator, for sample acquisition. References Arnold, A., Zehnder, K., 1985. Crystallization and habits of salt efflorescence on walls. Part II. Condition of crystallization. In: Proceedings of the 5th International Congress on Deterioration and Conservation of Stone, Lausanne, pp. 269–277.

Ajo`, D., Casellato, U., Fiorin, E., Vigato, P.A., 2004. Ciro Ferri’s frescoes: a study of painting materials and technique by SEM–EDS microscopy, X-ray diffraction, micro FT-IR and photoluminescence spectroscopy. J. Cult. Herit. 5, 333–348. Boon, J.J., Keune, K., Van Der Weerd, J., Geldof, M., Van Asperen De Boer, J.R.J., 2001. Imaging microspectroscopic, secondary ion mass spectrometric and electron microscopic studies on discoloured and partially discoloured smalt in crosssections of 16th century paintings. Chimia 55 (11), 952–960. Bordignon, F., Postorino, P., Dore, P., Trojsi, G., 2007. Raman identification of green and blue pigments in Etruscan polychromes on architectural terracotta panels. J. Raman Spectrosc. 38, 255–259. Casoli, A., Musini, P.C., Palla, G., 1996. GC/MS approach to the problem of characterizing binding media in painting. J. Chromatogr. A 731, 237–246. Colombini, M.P., Fuoco, R., Giacomelli, A., Muscatello, B., Fanelli, N., 1998. Characterization of proteinaceous binders in samples of Giudizio Universale wall paintings at Florence Cathedral. Sci. Technol. Cult. Herit. 7 (1), 49–58. Colombini, M.P., Modugno, F., Giacomelli, A., Francesconi, S., 1999a. Characterization of proteinaceous binder and drying oils in wall painting sample by gas chromatography–mass spectrometry. J. Chromatogr. A 846, 113–124. Colombini, M.P., Modugno, F., Giacomelli, A., 1999b. Two procedures for suppressing interference from inorganic pigments in the analysis by gas chromatography– mass spectrometry of proteinaceous binder in paintings. J. Chromatogr. A 846, 101–111. Colombini, M.P., Modugno, F., Menicagli, E., Fuoco, R., Giacomelli, A., 2000. GC–MS characterization of proteinaceous and lipid binders in UV aged polychrome artifacts. Microchem. J. 67, 291–300. Daniilia Sr., Sotiropoulou, S., Bikiaris, D., Chryssoulakis, Y., Salpistis, C., Karagiannis, G., Price, B.A., Carlson, J.H., 2000. Panselinos’ Byzantine wall-paintings of the Protaton Church, Mount Athos, Greece: a technical examination. J. Cult. Herit. 1 (2), 91–110. Daniilia Sr., Sotiropoulou, S., Karagiannis, G., Salpistis, C., Bikiaris, D., 2006. The pathology of the wall paintings in the Protaton Church, Mount Athos, Greece. In: Hanna, H. (Ed.), Proceedings of ICOM-CC-Wood, Furniture and Lacquer ‘Icon and Portrait International Conference’. 18–20 September 2006, King Maryut, Alexandria, Egypt, pp. 126–135. Daniilia Sr., Andrikopoulos, K.S., 2007. Issues relating to the common origin of two Byzantine miniatures: in situ examination with Raman spectroscopy and optical microscopy. J. Raman Spectrosc. 38, 332–343. Daniilia Sr., Tsakalof, A., Bairachtari, K., Chryssoulakis, Y., 2007. The Byzantine wall paintings from the Protaton Church on Mount Athos, Greece: tradition and science. J. Archaeol. Sci. 34, 1971–1984. Daniilia Sr., Minopoulou, E., Demosthenous, Fr.D., Karagiannis, G., 2008. A comparative study of wall paintings at the Cypriot monastery of Christ Antiphonitis: one artist or two? J. Archaeol. Sci. 35, 1695–1707. Dionysios of Fourna, 1996. The Painter’s Manual. Oakwood Publications, California. Frost, R.L., 2006. A Raman spectroscopic study of selected minerals of the rosasite group. J. Raman Spectrosc. 37, 910–921. Frost, R.L., Wain, D.L., Martens, W.N., Reddy, B.J., 2007. Vibrational spectroscopy of selected minerals of the rosasite group. Spectrochim. Acta A 66, 1068–1074. Ganitis, V., Pavlidou, E., Zorba, F., Paraskevopoulos, K.M., Bikiaris, D., 2004. A postByzantine icon of St Nicholas painted on a leather support. Microanalysis and characterisation of technique. J. Cult. Herit. 5, 349–360. Gettens, J., Stout, G.L., 1966. Painting Materials – a Short Encyclopaedia. New York, pp. 157–159. Gilbert, M., Sutherland, I., Guest, A., 2000. Characterization of coated particulate fillers. J. Mater. Sci. 35, 391–397. Goodall, R.A., Hall, J., Viel, R., Agurcia, F.R., Edwards, H.G.M., Fredericks, P.M., 2006. Raman microscopic investigation of paint samples from the Rosalila building, Copan, Honduras. J. Raman Spectrosc. 37, 1072–1077.

Sr. Daniilia et al. / Journal of Archaeological Science 35 (2008) 2474–2485 Grassi, S., Carretti, E., Pecorelli, P., Iacopini, F., Baglioni, P., Dei, L., 2007. The conservation of the Vecchietta’s wall paintings in the Old Sacristy of Santa Maria della Scala in Siena: the use of nanotechnological cleaning agents. J. Cult. Herit. 8, 119–125. Halpine, S., 1992. Amino acid analysis of proteinaceous media from Cosimo Tura’s, the annunciation with Saint Francis and Saint Louis of Toulouse. Stud. Conserv. 37, 22–38. Koszowska, E., Wese1ucha-Birczynska, A., Borzecka-Prokop, B., Porebskaa, E., 2005. Micro and FT-Raman characterisation of destinezite. J. Mol. Struct. 744–747, 845–854. Lelekova, O., 2006. Dionysiy’s paints 2006. Exhibition catalogue ‘‘Russia. Dionysiy’s Frescoes for the World’’. In: Holdin, Y. (Ed.), Frescoes of Old Russia, Dionysiy. Golden Age of the Icon-Painting in the 14–15th Centuries. 16 February–19 March 2006. State Tretyakov Gallery, Moscow, pp. 75–76. Mazzeo, R., Joseph, E., Minguzzi, V., Grillini, G., Baraldi, P., Prandstraller, D., 2006. Scientific investigations of the Tokhung-Ri tomb mural paintings (408 A.D.) of the Koguryo era, Democratic People’s Republic of Korea. J. Raman Spectrosc. 37, 1086–1097. Mora, P., Mora, L., Philippot, P., 1984. Causes of alteration in wall paintings. In: The Conservation of Wall Paintings. Butterworths, London, pp. 165–215. Mugnainia, S., Bagnolib, A., Bensic, P., Droghinia, F., Scalaa, A., Guasparria, G., 2006. Thirteenth century wall paintings under the Siena Cathedral (Italy). Mineralogical and petrographic study of materials, painting techniques and state of conservation. J. Cult. Herit. 7, 171–185. Muhlethaler, B., Thissen, J., 1993. Smalt in. In: Roy, A. (Ed.), Artists’ Pigments, vol. 2. Oxford University Press, New York, pp. 113–130.

2485

Pavlidou, E., Arapi, M., Zorba, T., Anastasiou, M., Civici, N., Stamati, F., Paraskevopoulos, K.M., 2006. Onoufrios, the famous XVI’s century iconographer, creator of the ‘‘Berati School’’: studying the technique and materials used in wall paintings of inscribed churches. Appl. Phys. A: Mater. 83 (4), 709–717. Rampazzi, L., Cariati, F., Tanda, G., Colombini, M.A., 2002. Characterisation of wall paintings in the Sos Furrighesos necropolis (Anela, Italy). J. Cult. Herit. 3, 237–240. Santopadre, P., Verita`, M., 2006. A study of smalt and its conservation problems in two sixteenth-century wall paintings in Rome. Stud. Conserv. 51 (1), 29–40. Sotiropoulou, S., Daniilia, Sr., Miliani, C., Rosi, F., Cartechini, L., Papanikola-Bakirtzis, D., 2007. Investigation of degradation issues in Byzantine wall paintings through non-invasive and microscopic multi techniques approach. In: E-MRS Spring Meeting, Science & Technology of Cultural Heritage Materials: Art Conservation and Restoration, 28 May–1 June 2007, Strasbourg, France. Vandenabeele, P., Lambert, K., Matthys, S., Schudel, W., Bergmans, A., Moens, L., 2005. In situ analysis of mediaeval wall paintings: a challenge for mobile Raman spectroscopy. Spectrochim. Acta A 61 (10), 2349–2356. Vitaliotis, I., 1998. Le vieux catholikon du monaste`re Saint-Etienne aux Me´te´ores, la premie`re phase des peintures murales. PhD thesis, University of Paris 1. Wust, R.A.J., Schluchter, C., 2000. The origin of soluble salts in rocks of the Thebes Mountains, Egypt: the damage potential to ancient Egyptian wall art. J. Archaeol. Sci. 27, 1161–1172. Winfield, D.C., 1968. Middle and Later Byzantine Wall Painting Methods – a Comparative Study. Dumbarton Oaks Papers 22, Washington. Zorba, T., Pavlidou, E., Stanojlovic, M., Bikiaris, D., Paraskevopoulos, K.M., Nikolic, V., Nikolic, P.M., 2006. Technique and palette of XIIIth century painting in the monastery of Mileseva. Appl. Phys. A: Mater. 83 (4), 719–725.