How to identify a lime-based painting technique in catacomb environment: well-established criteria applied on Roman catacomb wall painting samples D. Tapete 1,* , R. Piovesan 2 , E. Cantisani 1 , F. Fratini 1 , C. Mazzoli 2 , L. Maritan 2 , B. Mazzei 3 (1) National Research Council (CNR) of Italy - Institute for the Conservation and Valorization of Cultural Heritage (ICVBC), Florence, Italy (e.cantisani@icvbc.cnr.it); (2) Department of Geosciences, University of Padua, Padua, Italy (rebecca.piovesan@unipd.it); (3) Pontificia Commissione di Archeologia Sacra (PCAS), Rome, Italy (bmazzei@arcsacra.va); *Corresponding author: deodato.tapete@gmail.com; currently at Department of Geography, Durham University, UK Painting techniques are usually identified through microsamples investigated with already established analytical methods. Nevertheless, the interpretation of the painting microstratigraphy frequently lies more on operator’s expertise than objective criteria for discrimination. That can sometimes lead to different hypotheses, even on same painting samples. With specific regard to mural paintings, objective methodological approaches have been recently proposed, such as those by Mugnaini et al. (2006) and Piovesan et al. (2011, 2012) who distinguish fresco from a secco techniques by applying thin/cross section-based methods, respectively. The second experimental study in particular has set up distinction criteria from a systematic implementation on fresco and lime-painted replicas produced under aerial conditions. However, peculiar and extreme microclimate such as that characterising sub-aerial environments like catacombs and hypogeum systems (e.g., RH close to saturation, high rate of air CO 2 content) can have exerted an influence on the carbonation process, thereby adding onto the effects due to the technical choices of the painter. The microstratigraphic features can consequently be different from those expected, as also demonstrated by laboratory simulations (Tapete, 2007) and real-world diagnostic studies (Tapete et al., 2013). Introduction and research aim With the kind permission of the Pontifical Commission for Sacred Archaeology (PCAS), a set of historical samples were collected from mural paintings in three different Roman catacombs, Rome, Italy: • DOM-M8 from the Bakers’ Cubicle in the catacombs of Domitilla (second half - end 4 th century AD; Fig. 1a) • ST2, ST3 and ST6 from the double cubicle P in the catacombs of St. Tecla (late 4 th - early 5 th centuries AD; Fig. 1b) • P5 and P6 from the cubicle of Crescenzione (XXX centuries AD – per Barbara) in the catacombs of Priscilla The samples were selected taking into account the dating of the catacombs. From a technical point of view, archaeologists recognize differences among the mural paintings dating back to end 4 th - early 5 th century AD and those to early 3 rd -4 th century AD. Case studies and microsamples Fig. 1 a) View of the Bakers’ Cubicle in the catacombs of Domitilla; b) detail of the vault in the double cubicle P in the catacombs of St Tecla. Carbonate incrustations from which the sample ST6 was taken, completely covered the surface before the recent laser cleaning carried out by PCAS. a b Methodology and microstratigraphic features Following the methodology by Piovesan et al. (2012), all the microsamples were analysed in both thin and corresponding thick transversal polished section, by Optical (OM; in transmitted and reflected light) and Scanning Electron (SEM) Microscopy. The following objective criteria were applied (Fig. 2): 1. Thickness of the pigment-bearing layers (thinner in fresco) 2. Distribution and outline of the paint surface (rough for fresco, smooth for lime-paint; also depending on the pigment particle size) Petrographic study under Polarised Light Microscope (PLM) using the method by Tapete et al. (under review) strengthened the interpretation of the outer Ca-rich layers, whenever surface incrustations due to crystallisation processes covered the paint layers (Fig. 3). Scientific question: Can the criteria defined by Piovesan et al. (2012) be applicable to painting microsamples taken from hypogean environments (i.e. Roman catacombs)? Mugnaini S., Bagnoli A., Bensi P., Droghini F. , Scala A., Guasparri G. (2006) Journal of Cultural Heritage 7, 171–185. Piovesan R., Siddall R., Mazzoli C., Nodari L. (2011) Journal of Archaeological Science 38, 2633–2643. Piovesan R., Mazzoli C., Maritan L., Cornale P. (2012) Archaeometry 54 (4), 723–736. Tapete D. (2007) Analytical study of calcite recrystallization phenomena on hypogean mural paintings in relation to their conservation and restoration. MSc Thesis, University of Bologna, Italy. Tapete D., Mazzei B., Fratini F., Riminesi C., Manganelli Del Fà R., Cantisani E., Sacchi B., Cuzman O.A., Patrizi M.G., Scaletti L., Tiano P. (2013) Monitoring hypogeum systems affected by crystallisation processes among conservation needs, microclimate factors and accessibility to heritage. In: In: G. Biscontin, G. Driussi (eds), Proceedings of the International Congress “Scienza e Beni Culturali XXIX Conservazione e Valorizzazione dei siti archeologici. Approcci scientifici e problemi di metodo”, Bressanone, Italy, 9-12 July 2013, Edizioni Arcadia Ricerche, pp. 899-910. ISBN 978-88-95409-17-7 Tapete D., Fratini F., Mazzei B., Cantisani E., Pecchioni E. (under review) Petrographic study of lime-based mortars and carbonate incrustation processes of mural paintings in Roman catacombs. Periodico di Mineralogia (paper presented at AMMC 2013 - Conference on Ancient and Modern Mortars, Florence, 7 February 2013). Fig. 2 Microphotograph interpretation keys. Reflected light and BSE reference images of fresco technique for Bavarian green earth (a) and Italian burnt sienna (b, c); and lime- painting technique again for Bavarian green earth (d) and Italian burnt sienna (e, f). 3. Number, textural properties and microscopic appearance of Ca-rich carbonation layers (single over the surface for fresco, one exterior and one interior for lime-paint). Fig. 3 Microphotograph of the sample ST3 (PLM, XPL) with a detail of the painted surface showing a clear example of the interference that surface carbonate crystallisations can create to the clear detection of outer Ca-rich layers (i.e. one of the objective criteria defined by Piovesan et al., 2012), especially in fresco samples. 50 μm RESULTS AND DISCUSSION Catacombs of Domitilla, DOM-M8 (red ochre - fresco) Catacombs of St Tecla, ST6 (green earth over carbon black layer - lime-paint) Objective criteria indicate the use of a fresco technique (Fig. 4a). Anyway, the long time taken by the mortar to harden due to the high RH in the hypogeum seemed to have allowed carbonation go deeper within the microstratigraphy. BSE image clearly shows the extent of the carbonation (Fig. 4b). The absence of other features attributable to an intentional lime-paint technique (e.g. the inner Ca- rich layer; outline and higher thickness of the paint layer) leads us to exclude this hypothesis. BSE image acquisitions on not perfectly smooth cut surfaces confirmed their less suitability for feature detection than finely-polished cross-sections (Fig. 5). References Fig. 4 a) PLM microphotograph (XPL) and b) the corresponding BSE image of the sample DOM-M8, showing: 1) micritic lime enrichment within the mortar layer due to carbonation penetration (max. 200 μm thick; marked here with yellow dotted line); 2) red ochre paint thinner than it appears under PLM; 3) surface microsparitic crystallisations including random particles of tuff dust. Fig. 5 a) BSE image of another surface portion of sample DOM-M8 and b) the corresponding zoom at higher magnification (1200x) acquired on the rough cut surface corresponding to the thin section. The surface roughness and higher relief of the coarser aggregates layer emerging from the matrix of the underlying mortar do not allow a clear detection of the inter-layer contacts, thereby affecting the quality of the painting technique assessment. 1 2 3 a 1 2 3 b a b The microstratigraphy shows the unusual application of bimodal coarse glauconite/ celadonite grains (30-50 μm, 200-400 μm) mixed with iron oxides and carbon black, over a lime-paint carbon black layer (Fig. 6a). The mortar carbonation and consequent surface Ca-enrichment proceeded in depth, but not homogeneously as found in DOM-M8 (Fig. 6b-c). Traces of intentional smoothing during the mortar application are visible, but a proper dealbatio can be excluded (Fig. 6d). The inner Ca-rich layer is easily detectable along the interface between the carbon black layer and the mortar (Fig. 6e). Conversely, the exterior one is masked due to the penetration of surface carbonate incrustations into the paint layer. Only a textural remnant (like a ‘ghost’) due to alteration of the exterior Ca-rich layer is appreciable within the incrustation itself. Lime-paint application can be consequently proved for this sample. Fig. 6 a) PLM microphotograph (XPL) and b) the corresponding Ca elemental map of the microsample ST6, with detailed BSE images of the same areas of the sample, showing the penetration depth of the Ca-enrichment (green arrows in c); traces of smoothing and remnant of a Ca-rich layer (yellow and red arrows in d, respectively); lime-paint inner Ca-rich layer and a ‘ghost’ of the exterior one (orange and blue arrows in e, respectively). c d e 0.5mm b 1mm a