1. Introduction

2The mural paintings were in poor conservation condition with extensive loss of paint and overall flaking. Defective conservation treatments and various interventions between 1970-1994 contributed immensely to the loss of a major part of the original painting. Immediate rescue conservation interventions had to take place based on scientific methods of analysis, that would help guide decisions about care and conservation treatment.

3Our present work refers to the so far obtained results from archaeometrical techniques applied to minute samples and on a non-destructive manner and testing applied directly on the mural paintings in order to identify pigments, binders and substrate matter, a pre-requisite step to any further conservation work (Cotte et al., 2007; Schwan, 1998; Baumann et al., 1998; Mora et al., 1984).

2. Modern science and contemporary mural paintings

4Amongst the frequently used methods in wall painting, pigments, lacquers and media, are, Polarizing Light Microscopy (PLM), Scanning Electron Microscopy (SEM) and SEM-EDS, X-ray fluorescence (XRF), X-ray diffraction (XRD), Binocular Microscopy, Fourier Transform Infrared Spectroscopy (FTIR), Gas-Chromatography- Mass Spectrometry (GC-MS), High Performance Liquid Chromatography (HPLC), Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES), ultraviolet/visible absorption spectroscopy (UVS), visible absorption spectroscopy (VS), Raman Spectroscopy (Welcomme et al., 2007; Kakouli, 2002; Laurie, 1910 a, b; Mills and White, 1994, 1977; Lucas and Harris, 1962; Lorusso and Schipa, 2001; Ferretti, 1993).

5In particular XRF, XRD and mineralogical microscopic examinations of wall paintings have been reported (e.g. Leona and Winter, 2001; Fillipakis et al., 1976, 1979; Cameron et al., 1977, Calamiotou et al., 1983).

6Projects concerning characterization of wall paintings range from prehistoric right through Classical, Hellenistic to Roman, Medieval and contemporary and modern times (e.g. Lucas and Harris, 1962; Mantler & Schriner, 2001; Wallert, 1995; Leona and Winter, 2001).

7Early works and historical reports, such as Lucas on Egyptian materials, Church and Laurie and Cennini on artists’ techniques, the 19^th century explications of early manuscripts and printed works by Merrifield and by Eastlake, as well as, original sources themselves, such as, Theophilus and Pliny; all these are valuable guides as to what organic materials might be expected to be found in museum objects and mural paintings (Lucas and Harris, 1962; Church, 1915; Merrifield, 1849; Eastlake, 1967; Hawthorne and Smith, 1963; Cennini in 14^th c translated in 1933; Pliny).

8During the whole of the 20th century, it has been a usual practice that the artists continually redefined the conception of what constitutes art, a process that included a proliferation in the employment of materials not previously known or used. Since never before in history, artists have had at their disposal a tremendous assortment of natural and synthetic material, and moreover the license to use them. This profusion of products has been wholeheartedly embraced by modern artists. Art now incorporates new materials that include pigments with unprecedented variety, plus a plethora of synthetic paint media, exquisitely transparent plastics, fabrics, exotic metal alloys, quick-setting adhesives, and electronic devices, to name merely a few.

9Moreover, the long-established field of painting has seen a change. For example, paintings created in earlier eras reflected a relatively limited supply of artists’ materials, therefore the only available paint media were waxes, plant gums, egg, milk, animal hides, vegetable oils, and plant resins. Pigments came from mineral deposits properly processed (see Theophilus’ and Pliny’s treatise on this) or were extracted from plants, insects, and animals. Today, however, artists are not limited to these traditional materials and processing techniques but may also choose from a variety of commercial paint media, as well as a profusion of new synthetic pigments.

10Taken for granted that research into artists’ materials and their use plays an important role in conservation projects, the tremendous increase in the number of available materials has created new challenges for conservation professionals (Grattan, 1993).

11The yielded information, from the examination and analysis of artists’ paints, about artistic techniques and materials helps the taking of decisions about care and conservation treatment. Therefore, conservators of modern art also take a keen interest in art as a process, carefully researching the ideas behind the specific techniques that the artist used to create the work. The challenge is finding conservation solutions that preserve a painting without disregarding the artist’s intent.

12Nevertheless, even with these resources, much work remains to be done in identifying the vast number of materials used by contemporary and modern artists and in developing a better understanding of the properties of these materials. Archaeometry and Conservation science can play a significant role in this effort. There are now a number of scientific analytical techniques, such as those abovementioned, to aid and identify artists’ materials and techniques – and although most were refined and developed to study works of art made with more traditional materials, they can also be applied to 20th century artworks and their materials.

13Identification of the materials – and their properties – in contemporary objects and new developments is being pursued at several major institutions. Thomas Learner, Tate Gallery, London, developed a technique to identify modern paint media using pyrolysis gas chromatography- mass spectrometry. Paul Whitmore, Carnegie Mellon Research Institute, developed a device for assessing, in a microscopic-sized spot, the lightfastness of contemporary colorants early in an object’s life. Scientists at the Canadian Conservation Institute constructed a transportable FTIR spectrometer that has been used to differentiate traditional materials such as Japanese lacquer from imitations made from cashew oil or alkyds.

14The application of scientific analytical techniques to contemporary works of art has increased understanding of artists’ materials and techniques (pigments, binding media, natural ageing materials). Polarized light microscopy, XRF, XRD, GC, LC, MS, FTIR were applied to the paintings of the 20th-century U.S. artists – Willem de Kooning and Jacob Lawrence – conducted in the laboratories of the Getty Conservation Institute and the National Gallery of Art (Schilling et al., 2002). This research illustrated how modern science can reveal new insights about contemporary works of art, which ultimately can aid in the conservation of these works (Doehne et al., 1989; Bower et al., 1994)

15An integrated collaborative project initiated in 2002 by Tate, London, the National Gallery of Art (NGA), Washington, D.C., and the Getty Conservation Institute (GCI) gave answers to some of the many questions regarding the character of modern paint materials, focusing on the cleaning of modern paintings, chemical analysis, and physical characterization (Learner et al., 2002. In 2000, 11 European museums, coordinated by the ICN and Tate Modern, established the International Network for the Conservation of Contemporary Art (INCAA) (see, www.incca.org).

3. Historical and architectural context

16In order to appreciate the impact of Papaloukas on Greek early 20^th century painting movement, some basic factual elements on major artistic movements, for the non-specialist, are necessary. He was born in 1892 at Desfina, a village close to the archaeological site of Delphi. The environmental landscape and the archaeological site of Delphi were the first impressions and later inspirations depicted in a large part of his works including the decorative and figurative painting on the interior walls and ceiling of the Amfissa Cathedral.

17In Greece, Byzantine art was adhered to rigid rules. It was a Christian art dedicated to the service of the Church controlled at a major part by the Church. In that concept any attempt for evolution and renovation was not acceptable. Papaloukas’ revolutionary approach, although appreciated at an early stage by the local community, it was ended up as a hostile attitude against him, egged by the clergy.

4. Original Techniques Used By Papaloukas

18This section contains visual and macroscopic observations made in situ by E. Polychroniadou, in conjunction with data available from the present analysis. The paintings were carried out on a single layer of lime mortar applied directly onto the wall, that is, a secco. The pigments were applied to the dry ground mixed with lime or casein, egg, animal glue, linseed oil or vegetable gum. The artist renders the contour of the figures by an arbitrary thick line which has a vital quality and applies rich amount of pigments on particular areas of his compositions to achieve distinctive visual effects.

20The anthivolo (Fig. 2) was applied for the illustration of the Archangel on the sofa of the narrow upper arch at the right aisle.

21He transferred the pre- drawings on the wall by piercing a series of small holes with a pointed instrument, actually the one used at the time by dress makers, so that the drawing would be reproduced by dabbing it on the back side with a gauze sachet filled with charcoal powder.

22The anthivola, property of Spyros Papaloukas’s daughter, stand on their own as works of art, evidence of his creative thought processes and his technical virtuosity as a painter. Mrs Mina Papalouka offered them to the Amfissa Municipality as the main subject of the Spyros Papaloukas Gallery in Amfissa.

5. Current condition and previous interventions

23The serious defects in the initial structure of the building, the low quality of materials provided by local shops, the experimental techniques the artist applied to achieve distinctive visual effects, the input of rain water from the roof, and defective conservation treatments and various interventions between 1970-1994 contributed immensely to the loss of a major part of the original painting noticeably worse in areas at the ceiling of the dome, the apses and the ceiling of the gynaeceum (the apartment reserved for women at the upper level of the church) at the right and left nave (Fig. 3). The paint layer was in an unstable and deteriorated condition, many areas were insecurely attached, flaking and extremely friable and they tented to shatter even when slightly touched.

24Furthermore, the seismic activity of the region was the main cause for the move of structural materials resulted as exhibit of numerous cracks and deterioration of mortar joints

25The conservation treatments and several interventions occurred during 1970-1994 were aimed at the stabilization of the structural elements of the building and the mural paintings. There was extensive use of cement for the reinforcement of structural elements, joints of vertical walls and replacement of wooden structures such as the wooden staircase at the left side of the nave leading to the upper floor. Cement was also used as an ingredient of the mortar to fill the losses of the paint. Besides, structural work on the roof of the Cathedral required partial removal of the tiled roof and replacement with new tiles with cement rendering (Fig. 3). Loss or weakening of cohesion of the paint layer and flaking were mainly treated with Primal AC-33, Polyvinyl Acetate Emulsions and Casein in cases, since the latter has not yet fully located. But such a treatment concerns the next stage of the project which is beyond our prime scope of the present work (Baumann et al., 1998).

6. The conservation project

26Lengthy discussions and the obvious need for restoration of the Amfissa Cathedral interior mural paintings led to the appointment of a conservation team and conservation scientists to undertake the preliminary study before the actual conservation procedure. The conservation project started on May 2001 under the auspices of the mayor of Amfissa, the daughter of the artist and the general director, General Directorate of Cultural Development, Hellenic Ministry of Culture.

7. Sampling

27Sampling locations ranged from the ceiling to ground level and on various decorative motifs and compositions.Only minute samples (millimeter scale) were permitted to obtain. Any perceptible invasive sampling was avoided. Non-destructive EDXRF measurements were performed on hagiographic representations in the pulpit, the women’s gallery, and the left aisle. In all 26 samples (minute sampling) were analysed and 114 measurements (minute sampling & 10 non-destructively) with all methods were made; 64 Raman, 14 FTIR, 27 EDXRF (10 on the painting), 7 GC-MS and 2 IRPAS.

8. Methodology

XRF (X-Ray Flourescence)

29In situ XRF measurements, as well as some laboratory measurements, were performed for a semi-quantitative recognition of pigments, from major and minor element presence. The EDXRF field portable analyzer Spectrace 9000 TN was used with a mercuric iodide (HgI₂) detector, which has a spectral resolution of about 260 eV FWHM at 5.9 keV, and three excitation sources of radioisotopes within the probe unit – Americium Am-241 (26.4 KeV K-line and 59.6 KeV L-lineV) measuring Ag, Cd, Sn, Ba, Sb; Cadmium Cd-109 (22.1 K-line, 87.9 K- & L-line KeV) measuring Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Zr, Mo, Hg, Pb, Rb, Th, U; and Iron Fe-55 (5.9 KeV K-line) measuring K, Ca, Ti, Cr. The wide range of excitation x-ray energies theoretically can measure all chemical elements from sulfur to uranium. However, limitation factors produce a lower threshold of detection, and in practice 26 simultaneous elements in soil are measured, in surfaces up to a depth of a few millimeters (thickness depends on the atomic number of the element measured and sample density). The minimum detection limits of measured elements were bound by three criteria, as defined by EPA, a) three times the standard deviation taken from a blank sample, b) a statistically significant number of measurements are taken to determine the actual standard deviation which is then multiplied by three, and c) the minimum detection limit will decrease as measurement time increases (4 x counting time approx.= ½ detection limit). Prior to any course of measurements reliability tests were applied to reassure the instrumental stability of the analyzer (gain and energy drifts, calibration of energy axis), though the instrument had an automatic gain compensation). The instrument is powdered by rechargeable batteries which have a field endurance of 4-5 h. Small artifacts are analyzed by placing them over the 25 mm diameter analyzer window. When an analysis sequence is initiated, each source is exposed automatically in turn behind the analyzer window to excite the 25 mm diameter area on the surface of the sample. Count times for individual radioisotope sources may be optimized according to the application, but total count times selected for this work were less than 10 min. At the end of each analysis sequence, spectra from each source are deconvoluted by the data processing software to determine fluorescence intensities for a pre-selected range of elements. These intensity data are then quantified automatically using a fundamental parameter correction procedure, incorporating sample matrix corrections, based on factors that are pre-programmed for the appropriate application (in this case, silicate rocks). Results were viewed on the integral display screen and are stored for subsequent down-loading to a PC and printer through an RS232 port. The performance of the portable XRF instrumentation in the laboratory has already been reported (Liritzis, 2007; Mantzourani & Liritzis, 2006), while the entire concept of the present investigation may be applied to any available portable XRF instrument that requires an irradiated area than a point focus. Here, a wide range of silicate rock reference materials were analyzed as powder pellets to evaluate accuracy, precision and detection limits. Reference samples included (rhyolites, metals, clays, soils, low radioactivity reference samples). The study showed the capability of the instrument to determine major and minor elements (K, Fe, Ca, Mn, Ti) and selected trace elements (Sr, Zr, Ba, Rb, Pb, Mo) in typical silicate (rhyolitic) rocks. Other trace elements were not measured because their lower counting sensitivities mean that the concentrations were near to or below detection limits. Samples measured in the laboratory were dried overnight to oven at 75 ^oC prior to any measurement.

30This non-destructive technique was used as sampling was restricted in view of the value and the condition of the wall paintings. Particularly in the scene of the Nativity on the right barrel-vaulted arm of the cross, and on the anthivola preparatory drawings papers (see, footnote 4).

31However, limitations due to x-ray attenuation by absorption make the technique ideal for smooth surface analysis. Care is taken to avoid any surface secondary products from weathering / erosion effects, while for very thin layers, such as the painted layers, the obtained data provide semi-quantitative results, while in cases of multi-layered compositions, the recorded spectra are composite from all layers without separation. It was also applied on some minute samples from small loose fragments or deteriorated areas on the wall, but mainly on the actual paintings by placing the probe in contact with the point of interest, in a non-destructive manner (The present XRF has been applied extensively in archaeological materials, e.g. Liritzis et al., 2002; Mantzourani and Liritzis, 2006; Papageorgiou and Liritzis, 2007; Liritzis, 2007, and in mural paintings, e.g. Agnoli et al., 2007; Capitan-Vallvey et al., 1994). Therefore, the present analysis represents an area on the mural painting covered by the diameter of the counting window of around 25 mm and in a variable depth of some millimeters which accounts for any more elements detected, beyond those representing the pigment minerals. Simulation experiments and numerical calculations have shown this result (Liritzis, 2007).

FTIR (Fast Fourier Infrared Transform)

32FTIR (Low and Baer, 1977) and a first use of Infrared Photoacoustic Spectroscopy (IR-PAS) were used to define any organics (lacquers such as binders, varnishes), and inorganic pigments. The tiny sample becomes a thin section less than 1 mm thick or for crystals they become compressed or in powder form 1% diluted in KBr for disc preparation, and the IR beam passes through. Here, samples were mounted on resin, cut with diamond saw and grinded by means of successively finer grades of abrasive. This way fine superimposed painted layers (micro-stratigraphy) were possible to be seen from CCD digital images in the microscope under visible or UV light (see, d below). FTIR sample sizes 50-100 µm in thickness were cut, compressed in a diamond compressed cell and spectra were obtained. The FTIR used was that of Biorad (FTS-45A coupled with microscope UMA500). In several cases IR is used to identify water in natural glasses, gypsum or ambers. Hydration is used in ageing experiments and may aid authenticity of some objects and pigments and even their dating (Stevenson et al., 2002; Goodall et al., 2004; Low and Baer, 1977).

IR-PAS (Photoacoustic Infrared Spectroscopy)

33The determination of the binders and pigment concentration level within a painted layer can also be obtained with an infrared photoacoustic accessory. In this technique, a sample less than 10 mm in diameter is placed in a helium-filled chamber equipped with a high sensitivity microphone. The specimen absorbs the infrared radiation and undergoes an oscillatory heating and becomes a source of thermal waves. The waves within the bulk sample propagate to the surface and into the helium atmosphere. The photoacoustic signal is generated at the surface and is caused by the thermal expansion of the gas (McClelland et al., 1993; Newman et al., 1986).

34The magnitude of the signal varies in proportion to the concentration of the searched species. The absorbance values are calculated from integrated peak areas. Each absorbance value represents an averaged measure for the entire surface of the sample exposed to the 5 mm diameter infrared beam.

35IR-PAS used to measure the concentration of various species in the painted surface layer is a MTEC 300 photoacoustic cell accessory for a Bomem MB-120 spectrometer. Each sample will be run at a resolution of 8 wave numbers for 256 scans. Carbon black, supplied by the manufacturer, is used to establish the background conditions for the instrument. Several absorption spectra were obtained on standard pigments which served as a data bank, and the identification of pigments and binding media of an unknown composite spectrum was achieved with appropriate spectrum standardization and multivariate analysis (Liritzis et al., 2008a) (Fig. 9). The results of analysed samples are referred to Tables 3, 4.

36Here, the IR and XRF techniques were used to recognize a) pigments, b) lacquers, c) oils or resins (Gettens and Stout, 1966; Doerner, 1984; White, 1981).

GC-MS (Gas Chromatography with Mass Spectrometry)

37A Finnigan Gas Chromatography with Mass Spectrometry (GC-MS) was used for identification of the organics present (aminoacids etc), following standard procedures (White, 1978; Kitson et al., 1996) 

Raman

38Raman spectroscopy for the recognition of particular colors attributable to certain pigments chemical compounds through comparison with available data bank spectra (special data base software for arts developed on an d base compiler) (e.g. Clark, 2007). Equipments used are: a) 1000 µRaman, Renishaw, with a He-Ne laser light source (632.8 nm), and use of x50 and x100 objective lenses of Olympus, the focusing of around 1 mW power laser beam was made on single spots of a cross-sectioned sample, and on the surface, with a spectral resolution of 4 cm^-1, b) ExamineR High Performance Microscope System DeltaNu, 785 nm laser beam, BX51 Olympus camera, resolution ~ 4 cm^-1, diameter spot size ~1 µm, and c) the RockHound Raman System of DeltaNu with Nuscope, 785 nm laser beam FSX technology, resolution ~ 8 cm^-1 and diameter spot size ~35 µm (Liritzis et al., 2007; Guineau, 1989; Coupry and Lautie, 2001). In some samples with contaminated surface the raman spectra were mainly from fluorescence. In these cases laser beam persisted to remove this secondary product and reduce significantly fluorescence, and measure the raman spectrum of the actual underlying painted layer. All Raman spectra are baseline corrected for a better visualization of the results. Three types of Raman instruments were used. In the first instance (a) was used for an initial phase of the project, and (b, c) were applied in a later phase of the project pinpointing on certain mineral spots of same and newer samples, and for investigating the efficiency of the portable RockHound for in situ collection of spectra in a non destructive manner.

Imaging

39Photographic recording is a pre-requisite step prior to any further analysis. Photos of sample surface and sides to identify micro-stratigraphy were made using optical microscope stimulated by white polarized light and UV (Hg source), with objective lenses of x10, x20 and x50. The photos were obtained via a 3CCD camera (Gowing, 1999). The UV aids in some cases location of organic compounds from their fluorescent emission, as well as other binding media compounds with strong fluorescence (e.g. varnishes, lacquers, etc). By this way any painted layers as well as other carriers were identified, for further detailed microanalysis.

X Ray Diffraction (XRD)

40The XRD Powder Diffractometer was a Siemens D500, Cu-Tube, Graphite Monochromator, 2-theta range 4-70 grad, step 0.02 grad, Time/Step 3 sec.

41The mineral recognition was made by software DiffracPlus, Firma Bruker and the PDF-data Base (Powder Diffraction File).

42More analytical work on X-ray diffraction for mineral identification of plaster or substrate and Scanning Electron Microprobe analysis for point chemical analysis and observation of multi-layered paints is under study (Liritzis et al., 2008b).

9. Results

43The results as presented in Tables 1-4 are outlined below.

441) The painter used a wide ‘paletta’. Amongst the various pigments are: ultramarine, (Na_8-10A_l6Si₆O₂₄)S_2-4), carbon black, hematite or red ochre (Fe₂O₃), yellow ochre (Fe₂O₃*nH₂O) (though the principal matter is goethite, alpha-FeOOH) but kaolinite too, Prussian blue (Fe₄(Fe(CN)₆)₃), lead white (PbCO₃*Pb(OH)₂), used from ancient times till today, lead chromate sulfate (member of the generic name chrome yellow)(2PbSO₄*PbCrO₄). He also used barium carbonate as white but also to dilute darker pigments. In one case of mortar the orange spot was identified by Raman as a type of fluorite (CaF₂) which is known to have an extreme variability in colors.

452) Stains of similar pigments were found in the anthivolon (Table 1, Fig. 4). In particular, various areas were analyzed by EDXRF and Raman – the paper support, red and black stains. The black stain contains mostly iron probably an unknown red paint poured accidentally on the tamping (tampon) layer of the prickled paper which contained sienna, chrome yellow, white of titanium, and others. The red stain contains much more iron identified as iron oxide (sienna). The rest of the paper contains faint yellow, traces of sienna, chrome yellow and lead oxide (orthorhombic massicot).

Table 1: Non-destructive analysis for pigment identification of colours on two pieces of anthivolo, by Raman and EDXRF methods. Note that EDXRF gives only chemical element concentration, while Raman spectroscopy gives information about the structure.
Tableau1: Analyse non destructive pour l’identification de couleur de pigments sur deux pièces d’anthivolo, par les méthodes Raman et EDXRF. Notons que l’EDXRF donne uniquement les concentrations chimiques élémentaires, tandis que la spectroscopie Raman donne des informations sur la structure.

ANTHIVOLO SAMPLE REF.	COLOUR	PIGMENT	METHOD
AMFA-1 (26x56,5-59 cm)	Yellow	Lead chromate sulfate	Raman, EDXRF
	Black Black stain	Carbon Contains iron (sienna)	Raman EDXRF
	Red stain	Iron Oxide (sienna) Iron Oxide + Titanium dioxide white (TiO2)	Raman EDXRF
	Blue	Ultramarine	Raman
	Red	Iron Oxide (sienna)	Raman
AMFA-2 (18x85 cm)	Red	Iron Oxide (sienna)	Raman
	Blue	Ultramarine	Raman
	Dark blue	Prussian blue	Raman
	Black	Carbon	Raman
	White	TiO2	Raman
	Yellow	Lead chromate sulfate	Raman

46In particular, two pieces were analyzed (AMF-A1 of 26x56.5-59 cm, and AMF-A2 of 18x85 cm). Color was obviously dropped from the artist’s brush while painting on the murals (Fig. 4). Anthivolo reference numbers AMF-A1 and AMF-A2 form part of linear design of a decorative motive. In AMFA2 several oily stains were observed. Portable EDXRF probe was focused on particular stains on their surface. For Raman spectroscopy also non-destructive analyses were performed scanning the surface and focusing laser beam HeNe (632.8 nm) on visually identified pigment mineral grains. Paper sample sizes 50-100 µm were cut, compressed in a diamond compressed cell and FTIR spectra were obtained. Paper and an oily stain were also confirmed with µFTIR.

^rd
Tableau2: Mesures Raman pour l’identification de pigments sur des sections des échantillons AMF29, AMF13 et AMF2 de la galerie des femmes, aile droite 3e abside, et préparés dans une matrice de résine (voir fig. 6).

Sample ref.	Layer	Analysed Point	Colour	Pigment
AMF29 ∆3, lower part .	A	1	Red	Haematite
		2	Yellow	Red & Yellow Ochre
		3	Red	Red&Yellow ochre & SiO2
		4	Black	Uknown*
		6	Black	Carbon
		7	Green	Unknown** Organic
		8	Blue	Ultramarine
	B	1	Blue	Ultramarine
		2	Blue	Ultramarine & Prussian blue
	C	7	Blue	Ultramarine
	D	3	Red	Red ochre & Prussian Blue
		4	Blue	Ultramarine
		5	Blue	Ultramarine & Prussian blue
		6	Green	Unknown Organic***
	E	1	Blue	Ultramarine & Gypsum
		2	Blue	Ultramarine & Gypsum
		3	Green	Unknown Organic as D6
		4	Red	Red ochre & Ultramarine
		5	Blue	Ultramarine & Gypsum
		6	Blue	Gypsum & Ultramarine
AMF13, ∆3 plaster (Mortar) withcolour from the window in the conch of the 3rd apse	A	1	Yellow Orange	Chrome yellow Fluorite
		2	Yellow	Yellow ochre
		3	White	Calcium carbonate
	B, 90-100 um	4	Red	Red ochre
		5	Blue	Ultramarine
		6	White
		7	black	Carbon
	C, 60-70 um	8	yellow	Chrome yellow
		9	red
		10	blue	Ultramarine
		11	white
		12	lacquer
		13	yellow
		14	black	Carbon
	D, 200 um	15	white	Calcium carbonate
		16	yellow	Chrome Yellow
		17	brown	Red ochre
		18	green	Prussian blue
	E, about 30 um	19	yellow	Yellow ochre
AMF2 Yellow pigment flake. Vault	A, >600 um	1	yellow	Chrome yellow +Yellow ochre
	B, about 50 um	2	red	Red ochre
		3	yellow	Chrome Yellow (FTIR: kaolinite)
		4	blue	Ultramarine

Lettering (A, B, …) corresponds to analysed sandwiched layers, and the numbers refer to the particular point within the layer with obvious colour.(AMF: AMFISSA + numbering of sampling, ∆3:  for women gallery (gynaeceum), ∆ for right aisle, 3 for 3^rd apse). Spot colours analysed include red, green, black, yellow, blue, orange, white. [NOTE for the unknowns, underlined is the higher relative intensity. Raman Wavelengths (cm^-1). * A4:  233.8, 311.4, 675.2; ** A7 (rather some organic substance) 292.2, 332.4, 510.7, 546.9, 644.7, 686.3, 707.7, 741.2, 777.4, 819, 958.4, 979.8, 1083, 1213, 1281.4, 1377.7, 1392.7, 1444.3, 1481.1, 1537.4, 1596.4, 2822, 2873.8, 3074.9. *** D6: 414.2, 493.3, 546.9, 622, 671.6, 1008, 1099.1, 1306.9, 3404, 3490].
Les lettres (A, B…) correspondent aux couches analysées, et les numéros font référence aux points particuliers dans la couche ayant une couleur évidente (AMF : AMFISSA + numérotation de l’échantillonnage, ∆3:  pour galerie des femmes (gynécée), ∆ pour aile droite, 3 pour 3e abside). Les point colorés analysés incluent le rouge, le vert, le noir, le jaune, le bleu, l’orange, le blanc. [note pour les indéterminés : les longueurs d’onde Raman des intensités relatives les plus élevées sont soulignées : * A4:  233,8, 311,4, 675,2 ; ** A7 (plutôt une substance organique) 292,2, 332,4, 510,7, 546,9, 644,7, 686,3, 707,7, 741,2, 777,4, 819, 958,4, 979,8, 1083, 1213, 1281,4, 1377.7, 1392.7, 1444,3, 1481,1, 1537,4, 1596,4, 2822, 2873,8, 3074,9. *** D6: 414,2, 493.3, 546,9, 622, 671,6, 1008, 1099,1, 1306,9, 3404, 3490].

484) In AMF13 (mortar bearing painted layer, Fig. 6, 7, 9), the following pigments were recognized through IR-PAS, FTIR, Raman techniques: a) yellow ochre (possibly goethite) b) calcium carbonate for the white, c) red ochre d) black of carbon, e) ultramarine, f) Prussian blue and g) fluorite (~1880 cm^-1). Raman measurements were repeated by two different instruments, and an example is shown for Fig. 7, whereas the identification is based on the peaks found between 230 and 660 cm^-1. Nevertheless, for other spectra the spectral region to 2000 cm^-1 was also employed.

Figure7: Spectre Raman (intensité / longueur d’onde) de AMF13 couche b (la seconde parmi les 5 couches), minéral rouge pris par ExamineR.

495) For AMF29 the successive layers B, C, D contain mainly ultramarine and Prussian blue. They are recognized as distinct due to different contents of these two pigments. In layer E there exists strong mixing of red and blue pigment, but no Prussian blue has been used. The presence of organic substance in green is detected but could not be identified as one of the known pigments. Perhaps it derives from later interventions on the artwork. Similarly the dark pigment in layer A could not be identified.

517) Further X-Ray Diffraction was used in a few cases of tiny sample. The spectra have shown Gypsum substratum (CaSO₄*2H₂O) with quartz (AMF27), while in another case (AMF28) a variety of colored grains of blue, brown and white were located. The blue ones were due to lapis lazuli (sodalith, NaS(AlSiO₄)₆(ClO₄)₂), the brown ones were due to hematite diluted with barite (Liritzis et al., 2008b).

528) The cross section of AMF2 have shown two painted layers. Fig. 5 (AMF2) show typical Raman spectrum respectively, and Table 3 shows all analyzed samples. Pigments (AMF18-26) from a local shop were analyzed because it was thought to have been used by Papaloukas, as this was a traditional shop since the days of Papaloukas, forgotten in some old boxes thought to be as old as 70 years. None were used by Papaloukas.

Table 3: Sample location, color, method used and pigment identification by different methods of colors on minute samples from various points of the wall painting (high barium contents as BaCO3 white are present rather for weight increase and thus higher cost).
Tableau3 : Localisation de l’échantillon, couleur, méthode utilisée et identification du pigment par différentes méthodes de couleurs sur de minuscules échantillons provenant de différents endroits de la peinture murale (de fortes teneurs en barium blanc BaCO3 sont présentes plutôt pour un accroissement de poids, et ainsi pour un coût plus élevé

Sample ref.	Description	Color	Pigment	Method
AMF2	∆3, women gallery, right aisle 3rd apse, yellow flaked pigment	Yellow	In 4 successive painted layers: Chrome yellow+ Yellow ochre+ Ultramarine+ Red ochre (yellow ochre=kaolinite by FTIR) High content: Pb and Fe (mainly yellow from similar pigments as above)	Raman, FTIR EDXRF
AMF3	∆2, women gallery Right aisle, 2nd apse. Red Flaked pigment	Red	Iron Oxides + Ca + Sr. Red ochre	EDXRF
AMF4	Pulpit, red decoration	Red	High Fe, Ca, Sr, Ba. Red ochre	EDXRF
AMF5	Left aisle Internal side of column	Blue	High Fe, Ba, CaCO3, K, Prusian Blue	EDXRF, FTIR, IRPAS
AMF12	∆3 Women gallery, right aisle, 3rd apse )	Blue	High Fe and CaCO3. Red ochre	EDXRF
AMF13	∆3, Women gallery, right aisle, window in the coche of the 3rd apse	5 layers	Chrome yellow, Red ochre, Ultramarine, Carbon , etc (details in Table 2)	Raman, FTIR, IRPAS
AMF14	Pulpit Part of surface Flaky pigment on paper with lettering	Yellow	Yellow of Pb-Sn + PbO, K, Fe (a surprising result which can not at present be justified)	EDXRF
AMF15	Pulpit Paper with letters	Black		FTIR, IRPAS
AMF16	∆2, women gallery right aisle, 2nd apse.	Brownish	Red ochre	Raman
AMF17	∆3, women gallery right aisle,  3rd apse red from the vault	Red	Red lacquer + Red ochre. Iron oxides + MnO2+ Clay. Red ochre	Raman EDXRF
AMF18- AMF26	Local shop	Various colours (modern pigments)	Gypsum, minium, Oxides of manganese and iron, high barium	EDXRF

Portable EDXRF analyses refer to a non destructive manner right on a larger area on the surface mural painting. Chemical elements are reported which reflect known compounds. In contrast to the tiny samples detached and analyzed by Raman and FTIR. The compounds here are recognized from a data bank spectral library. Manganese and iron oxides (ochres) and minion were used in ancient times. ). Les analyses EDXRF in situ se réfèrent à une approche non destructive sur une plus grande surface de la peinture murale. Les éléments chimiques reflètent des composés connus. Par contraste avec les échantillons minuscules détachés et analysés par Raman et FTIR. Les composés ici sont reconnus à partir d’une banque de données spectrale. Les oxydes de manganèse et de fer (ocres) et leurs dérivés étaient utilisés dans les temps anciens.

Table 4: Gas chromatography-mass spectrometry (percentage of aminoacids present and comparison with a data bank via principal component analysis score plots) and FTIR and IR-PAS screening (spectra compared with data bank) for detection of organics of samples from various parts of the wall painting.
Tableau4: Spectrométrie de masse en phase gazeuse (pourcentage d’acides aminés présents et comparaison avec la banque de données via un diagramme d’analyse en composantes principales), et filtrage FTIR et IR-PAS pour détection d’éléments organiques sur des échantillons de différentes parties du mur peint.

SAMPLE REF.	BINDING MEDIA
AMF-5, AMF-6, AMF-16	Egg (main stretching frequencies: 463, 530, 618, 794, 926, 1009, 1097, 1359, 1577, 3325, 3500, 3587 cm-1)
AMF-8, AMF10	Animal glue
AMF-15	Near Casein cluster
AMFA-1 (Anthivolo)	Paper / oil, Absorbance : 897, 1061, 1161, 1267, 1317, 1373, 1427, 1508, 1653, 2905, 3353 cm-1.

10. Conclusions

54The application of archaeometrical techniques increased understanding of the artist materials and working methods, thereby enhancing conservationist’ ability to preserve these paintings. All the above methods of analysis complement each other, reinforce reliability, and clarify the technique of execution of such an important work of art with minute or non-destructive sampling.

55The painter did not follow the usual a secco practice, instead he combined several painting techniques (pigments mixed with lime, tempera and oil) to open up new area of creative possibilities and mastered the natural characteristics of the pigments by the quality of his brushstrokes and the particular medium’s degree of gloss.

56The artist utilized unconventional materials to achieve desired visual and textural effects. However, his experimental techniques and low quality materials had a negative effect on the long term stability of his mural paintings, and pose major problems to conservators. It is indispensable the need to establish new conservation techniques to remove previous conservation materials and apply appropriate materials to stabilize the painting surface.

The authors wish to warmly thank late Mrs Mina Papalouka and Mrs Mary Michailidou for their encouragement of this project, the Amfissa Town Municipality and the Mayor of Amfissa for their aid and financial support. We are grateful to Mr Vassilios Theoharakis, Mr Ioannis Moscholios, the Cultural Foundation of the Alpha Bank, Greece, and the Silver and Baryte Ores Mining Co. S. A, for financial support. We extend our thanks to Mrs Ioanna Konstandinou painter, paper conservator for her advice and comments and Mr. Alexandros Polychroniadis for continuing support.

We appreciate Mrs Sibyle Nerger, Prof. Luigi Dei , David Park, Dr. Marcel Stefanaggi for their in situ comments and suggestions, and the Greek Ministry of Culture, Directorate of Conservation of Ancient and Contemporary Monuments, for granting permission to sampling.

Thanks to Mr Elias Magafas for technical assistant and the editor of FOKIS newspaper for permitting access to archives.