A Scientific Approach to Cultural Heritage Preservation: A Case Study

---

Chemistry Everyday for Everyone

A Scientific Approach to Cultural Heritage Preservation: A Case Study of Vandalistic Acts on Important Roman Mosaics Enrico Ciliberto and Giuseppe Spoto Dipartimento di Scienze Chimiche, Università di Catania, Viale A. Doria 6, 95125 Catania, Italy Mauro Matteini Opificio delle Pietre Dure, Viale F. Strozzi, Firenze, Italy Concetto Puglisi Istituto per la Chimica e Tecnologia dei Materiali Polimerici, Consiglio Nazionale delle Ricerche, Viale A. Doria 6, 95125 Catania, Italy

The use of chemistry in the study and preservation of works Case Study of art is today an important aspect of the relationship between The Villa del Casale, near Piazza Armerina (Italy), is the chemistry and society.1 The development of new analytical techmost important testimonial of the Roman occupation of Sicniques has enabled researchers to optimize new investigative apily. It was built in the late 3rd century and early 4th century proaches that minimize the damage to the artistic object under C.E. as a luxurious country villa and was inhabited until the examination and expand the set of chemical and physical infor12th century, when it was buried in mud by a flood. The mation. Complete characterization of the constituting materivilla consisted of an extensive network of rooms, galleries, als, study of the degradation processes, evaluation of the intercourtyards, and baths, but its main feature is the mosaic action between the objects and the deteriorating agents, together floors, which contain some of the largest and most beautiful with all of the investigations related to the planning of the resmosaics surviving from Roman times. toration projects, represent some of the goals to be achieved usOn the night of September 29, 1995, some unknown ing a microdestructive approach. vandals poured dark brown paint over several of the most As a general rule, we can affirm that the wide spectrum beautiful and important mosaics in the villa and dug up many of information calls for a multitechnique approach to the tesserae (for a definition of tessera, see below), fortunately in problems related to the study of artistic objects. In fact, the marginal zones. In particular, the Triclinium area, the Big usual heterogeneity of the materials used in the production Hunt Corridor, and the Room of Girls in Bikinis (see Fig. of the artistic objects or involved in their degradation often 6A) were severely damaged in appearance over a total surleads to the final systems being so complex in composition face area of approximately 100 square meters. that they require the use of different kinds of techniques for Chemical investigations were performed immediately in their characterization. Moreover, the need of nondestructive order to draw up a rapid restoration plan aimed at identifying analysis restricts the techniques to be considered to those that the substances used and proposing a correct restoration procecan be operated “in situ”—collecting the information from dure. The problem to be solved required a multitechnique apthe interaction of a variety of radiation with the artifact surface—or techniques that require tiny samples (milligrams). The former are called undestructive while the latter are defined as microdestructive techniques (3, 4). As an example of the way in which chemistry can merge with the world of art, we report a case study where complete characterization of the materials used in an act of vandalism on one of the most important Roman monuments in Italy allowed us to project a successful restoration plan. Our aim is to demonstrate how chemical investigation can greatly help historical culture maintain its heritage. In fact, a scientific approach is fundamental in restoration projects because incorrect procedures can often cause greater damage to the object than the alteration itself. Figure 1. Schematic representation of the analytical multitechnique approach. 1302

Journal of Chemical Education • Vol. 75 No. 10 October 1998 • JChemEd.chem.wisc.edu

Chemistry Everyday for Everyone

proach because of the highly complex heterogeneity of the materials studied. In fact, paint is a mixture of different inorganic and organic compounds that act basically as vehicles, dryers, and pigments (5). A mosaic is generally made up of little fragments of solid materials of different colors, called tessera, placed on a suitable substrate to form geometrical and artistic drawings. In the case of the Villa del Casale, following the typical Roman procedure for floor mosaics, the tesserae are marble (mainly constituted by calcium carbonate in the crystallographic form of calcite) and diaspore (constituted by aluminum hydrated oxides) and are placed on a substrate called “nucleos” which is a mixture of lime and terracotta fragments called “cocciopesto”. As the inorganic chemical nature of the mosaic interacted with the components of the paint used, it was necessary to identify: • • • •

The pigments present in the paints, in order to know their chemical properties The vehicles, in order to project suitable methods to remove the paint The inorganic, inert powders, in order to find possible correlations with commercial paints The composition of the paint–tessera interface, in order to better understand any possible chemical interactions

Figure 1 is a schematic representation of the analytical procedure adopted. Two different groups of techniques were used for the characterization of the organic and inorganic components of the paint, respectively. UV–vis spectroscopy coupled with electron-impact mass spectrometry (EIMS), direct-pyrolysis mass spectrometry (DPMS), nuclear magnetic resonance ( 1H NMR), and X-ray photoelectron spectroscopy (XPS) techniques provided characterization of the organic components of the paint and were used to determine the elements present at the interface of the paint films with the tessera, while the inorganic components were characterized using scanning electron microscopy (SEM) coupled with energy-dispersive X-ray microanalysis (EDX) for morphological and elemental information, and the X-ray diffraction technique (XRD) for structural information. Moreover, XPS provided further information about surface-compositional properties at the paint–tessera interface.

Some samples of paint were treated with a variety of solvents, chosen on the basis of their solvent power (6), in order to evaluate their behavior. Particularly evident softening of the paint was found when dichloromethane was used as a solvent. Complete solubility of the film was not found in any solvent. In all cases, the organic solvents extracted a red pigment. Aggressive, caustic reagents (NaOH solutions at different concentrations) obtained poor results regarding the softening of the paint film, but it extracted the organic pigment and colored the solution yellow. The change in color may have been caused by the neutralization of acid groups. Characterization of the Organic Constituents Since the red pigment, which was easily soluble in the most common organic solvents, jeopardized the mosaics, attempts were made to characterize it. As discussed in a previous article in this Journal (7), organic pigments can be identified by performing some solubility experiments coupled with spectrophotometric measurements. In our case, various conventional, analytical techniques were used to obtain complete characterization of the pigment. It was necessary to extract the red pigment from the paint using chloroform, purify it by chromatography on silica gel, and then use a set of analytical techniques capable of giving information about optical properties, molecular weight, molecular fragmentation, and proton distribution in the molecule. It was seen that the pigment was a single, molecular species; no multiplicity of the eluted spot was observed after chromatography. In this case, UV–vis, EIMS, and 1H NMR data (8) clearly showed that the pigment was toluidine red (MW 307.31; mp 270–272 °C; λmax, nm: 510), whose formula is reported in Figure 2. In particular, EIMS (Fig. 3) showed highly intense peaks at m/z =307, 290, 275, 260, 246, 171, and 143, attributing to positively charged fragments of the molecule,

Figure 2. Formula of toluidine red.

Preliminary Actions The first step for correct and informative analysis of the alteration to be studied was sampling. For artistic objects, destruction caused by sampling must be reduced to a minimum. In every kind of sampling, any contamination that can alter the significance of the samples themselves must be avoided. In our case, samplings were performed in all of the rooms where damage had occurred. The first evidence showed the paint adhering so strongly to the mosaic that it was difficult to remove it mechanically. Moreover, no macroscopic evidence of infiltration of the paint components in the spaces between the tesserae was found. The samples exhibiting the typical smell of oil paint and a brown-black color with reddish shades were taken by microlancets and placed in hermetically sealed glass containers in order to avoid further evaporation of traces of solvents.

Figure 3. Electron-impact mass spectrum of the red pigment, purified by chromatography.

JChemEd.chem.wisc.edu • Vol. 75 No. 10 October 1998 • Journal of Chemical Education

1303

Chemistry Everyday for Everyone

while the 1H NMR revealed a typical spectrum for toluidine red as deduced from comparison with a spectrum acquired on a standard sample. The identification of the paint vehicle, which provides a fundamental clue for the discovery of the commercial product used, was the most complex procedure. In fact, it is well known that commercial paints are usually prepared using a variety of resins that form a matrix for the organic and inorganic pigments (5). DPMS (9, 10) measurements were performed on the paint as they have proved to be an extremely useful tool in analyzing large, nonvolatile molecules, such as natural and synthetic polymers, cross-linked drying oils, natural resins, etc. In our case, the unknown paint samples were placed in the ion source by a specific probe for solid samples that was heated up to 700 °C with a heating rate of 20 °C min᎑1. Positive mass spectra were acquired every 30 seconds during the heating process. The same kind of procedure was carried out on three samples of black commercial paint having different compositions (polyester, acrylic–water-based, epoxidic) and on a standard of linseed oil for comparison. The analyses of the unknown paint distinguished four different intervals of temperature of the heated probe, in which the different spectra

allowed identification: 50–70 °C, 150–170 °C, 300–400 °C, and 400–450 °C. Comparison among the spectra obtained at temperatures included in these four ranges and the spectra obtained under the same conditions from the four standards enabled us to deduce important considerations on the composition of the unknown paint. The first consideration was related to the presence of unsaturated fatty acids in the paint. This finding was obtained from the spectra in the 50– 70 °C range. In fact, mixtures of oils are generally used, in various percentages, in the production of paints for their cross-linking capacity due to the polymerization and oxidation at the carbon–carbon double bonds in the presence of atmospheric oxygen (7). A peak characteristic of phthalatecontaining compounds (m/z = 149) (11) also appeared in the same temperature range. The mass spectra recorded in the other temperature intervals showed peaks attributed to the toluidine red pigment (peaks at m/z = 275, 290, and 307, respectively, appeared in the 150–170 °C range) and peaks attributed to the presence of styrene (a peak at m/z = 104 appeared in the 300–400 °C range), the latter usually used as a solvent and for cross-linking. The 300–400 °C range appeared to be the most significant for discrimination among the three commercial, standard black paints. In fact, great similarities were observed in the spectra obtained from the standard polyester paint and the unknown paint, heated to temperatures in this interval. At higher temperatures (400–450 °C), the mass spectra for both the unknown paint and the standard paints showed the presence of ions resulting from the thermal decomposition of the siloxane units present in the samples (peaks at m/z = 207, 221, 281, 297, 341, and 355, respectively). This similarity was due to the fact that silicone oils are usually added to the recipe of the paint because they produce better mechanical and thermal properties, and gloss. The presence of silicone-based compounds in the unknown paint was confirmed by XPS analyses (see below). Figure 4 reports some of the most significant ionic fragments detected in the analysis of the unknown paint. In conclusion, critical evaluation of all of the data observed indicated that toluidine red was present as pigment in the paint and that the vehicle was made up of a mixture of alkyd resins, together with styrenated compounds and unsaturated long chain-containing oils. This complex formulation is common in commercial paints because it endows products with specific features, such as solubility in specific solvents, faster drying, gloss retention, and chemical resistance. Morphological Properties of the Film and Inorganic Components Present

Figure 4. Some of the most relevant ionic fragments identified by the DPMS analysis of the unknown paint. The temperature range in which the ionic fragments were revealed is reported.

1304

Fragments of paint film were investigated by SEM (12) in order to investigate the external surface and the internal region corresponding to the paint–tessera interface. These fragments were mounted on an aluminum holder with graphite paste and were successively covered with a conducting, gold film to prevent the charging effects—induced by the electron bombardment—that alter the electron image (12). Differences between the two regions were clearly evident. The outer surface was corrugated by streaks generated by the paint-drying process, while the internal surface had a microgranular structure resulting from the inorganic, inert powders present in the paint which formed sediment by gravity in the lower zones of the films. The section showed a strati-

Journal of Chemical Education • Vol. 75 No. 10 October 1998 • JChemEd.chem.wisc.edu

Chemistry Everyday for Everyone

A

Figure 5. XPS wide spectrum of the paint film surface. The most important ionizations are labeled.

fied structure with layers of approximately 50 µm wide which may indicate a lack of homogeneity in the paint used. The EDX (12) spectra, acquired in the three regions to reveal the elemental composition,2 showed X-ray signals due to Al (Al K α,β radiation), Si (Si K α,β radiation), S (S K α,β radiation), Ca (Ca K α,β radiation), Ti (Ti K α,β radiation), Ba (Ba K α,β radiation) and, in certain spectra, Mg (Mg Kα radiation) coming from the inert powders added to the vehicle of the paint. XRD (13) analyses were carried out using a Ni-filtered, Cu Kα radiation and a Bragg–Brentano goniometer in order to identify the crystalline phases present. The XRD spectra for the samples indicated the presence of calcite (CaCO3), barium sulfate (BaSO4), and aluminum oxide (Al 2O3). No phases corresponding to titanium-containing compounds were identified in spite of the presence of titanium revealed by EDX. This evidence could therefore be indicative of the presence of an amorphous titanium-containing compound not detectable by XRD, since XRD is able to reveal and identify only crystalline compounds. The XPS technique (14) was used to determine the elements present at the interface of the paint films with the tessera. Figure 5 shows a representative XPS wide spectrum acquired on the surface of a paint film. Signals coming from O 1s, O 2s, Si 2p, Si 2s, C 1s, and Ag 3d ionizations were clearly evident in the spectra obtained from both the external and internal surfaces. Moreover, signals due to the emission of Auger electrons (14) from carbon (C KLL) and oxygen (O KVV) were present. In order to evaluate the chemical environment of each element, a specific study based on the evaluation of binding energies (15) enabled us to deduce conclusions about the organic and inorganic composition of the paint. In particular: • • •

Aliphatic, carbonylic and carbonatic carbons were present in the sample. Silicon was coming from silicone compounds. Silver was in the +1 ionization state.

Concluding Remarks The reported case shows the importance and the role of modern, analytical methods in solving relevant questions for restoration. Questions concerning restoration projects often appear complex and are not easily addressed by a chemical point of view. The potential of modern, analytical techniques

B

Figure 6. (A) Image of a particular of the mosaic present in the Room of Girls in Bikinis as it appeared before the restoration. A squared area in which cleaning tests were performed is visible. (B) Image of a particular of the mosaic present in the Room of Girls in Bikinis as it appears after the restoration was carried out, following the procedure described in the paper.

allows the chemist to obtain, as much as possible, complete knowledge of the object to be restored and to provide an effective and not “dangerous” restoration plan. In our case, the coupling of techniques that provided information about the organic components of the paint (DPMS, 1H NMR, UV– vis, EIMS, and XPS) with analytical techniques that clarified the composition of the inorganic constituents of the studied materials—both the paint and the tesserae—(SEM/EDX, XRD, and XPS) allowed us to obtain the necessary knowledge of both the object to be restored and the offending material. Once the chemical nature of the paint used was known, it was possible to decide on the removal methods to be adopted. Dissolving the paint with organic solvents did not appear straightforward because of its polymeric nature. Softening techniques using polar, organic molecules of low molecular weight, such as chloroform or methylene chloride, seemed more suitable, followed by mechanical removal of the paint layer. However, treatment with organic solvents could have led to the toluidine red pigment (together with the oligomeric parts of alkylsiloxane present in the paint) marking the stone substrate. For these reasons, adsorbent materials (cellulose) were used with chloroform in order to avoid damaging the stone substrate.

JChemEd.chem.wisc.edu • Vol. 75 No. 10 October 1998 • Journal of Chemical Education

1305

Chemistry Everyday for Everyone

The presence of silver salts at the paint–stone interface of the tessera must be underlined. This element segregated on the surface of the paint, and if it were not detected in the bulk, could have jeopardized the mosaic. Therefore, treatment with ammonia or alkaline ammonium salts, routinely used by restorers as cleaning agents, could have extracted the silver in the form of the complex Ag(NH3) 2+ and transported it inside the stone to produce successive, irreversible darkening of the stone; this was strictly avoided. All of these considerations, together with the finding that simple, mechanical removal of the paint by lancet or similar tools appeared unsuccessful and jeopardized the tesserae, suggested that strictly controlled treatment using chloroform compresses to soften the paint film, followed by a washing with water, could achieve good results. In this way, the possible marking of the stone by the red pigment was stopped by the pigment’s insolubility in water. Although this method required accurate monitoring of the time necessary to soften the paint without marking the tesserae, it was very effective and achieved restoration of the original beauty of the mosaics (Fig. 6). Acknowledgment The authors thank the CNR–Progetto Finalizzato Beni Culturali (Rome) for financial support. The Soprintendenza ai Beni Culturali ed Ambientali of Enna (Italy) is kindly acknowledged for providing the images of the mosaics. Sebastiano Sciuto and Rosanna Chillemi, from the University of Catania, are acknowledged with thanks for valuable discussions and interest. Notes 1. The importance of the relationship among chemistry, the study and preservation of works of art, and society can be accounted by considering both the economical and cultural aspects related to the preservation of works of art. A more specific view of this aspect can be seen by reading excellent books (1) and various articles published in this Journal (2). 2. The coupling of the scanning electron microscope with an energy-dispersive X-ray fluorescence detector allows one to obtain an el-

1306

emental analysis of a selected area of the electron image. A rastered electron beam is used to stimulate the emission of both secondary electrons (used to reconstruct an electron image of the rastered area) and characteristic X-ray. The EDX detector allows the identification of the elements present in the rastered area by detecting and distinguishing the different energies of the emitted X-ray.

Literature Cited 1. Lambert, J. B. Traces of the Past. Unraveling the Secrets of Archaeology through Chemistry; Addison–Wesley Longman: Reading, Massachussetts, 1997. 2. Schmuckler, J. S. J. Chem. Educ. 1981, 58, 326–327. 3. Ciliberto, E.; Allen, G. C.; Di Stefano, C.; Fragalà, I.; Spoto, G. Anal. Chem. 1995, 67, 249A–253A. 4. Ciliberto, E.; Fragalà, I.; Mancini, N. A.; Spoto, G. Microsc. Microanal. Microstruct. 1995, 6, 533–543. 5. Weber, W. C. J. Chem. Educ. 1960, 37, 322–324. 6. Grulke, E. A. In Polymer Handbook, 3rd ed.; Brandrup, J., Immergut, E. H., Eds.; John Wiley & Sons, Inc.: New York, 1989; pp 519–559. 7. Billmeyer, F. W.; Kumar, R.; Saltzman, M. J. Chem. Educ. 1981, 58, 307–313. 8. Silverstain, R. M., Bassler, G. C., Morrill, T. C. Spectrometric Identification of Organic Compounds, 5th ed.; John Wiley & Sons, Inc.: New York, 1991. 9. Montaudo, G.; Puglisi, C. In Comprehensive Polymer Science, 1st Suppl.; Aggarwal, S., Russo, S., Eds.; Pergamon: Oxford, 1992; pp 227–235. 10. Shedrinsky, A. M.; Wampler, T. P.; Indictor, N.; Baer, N. S. J. Anal. Appl. Pyrol. 1989, 15, 393–412. 11. Sproch, N.; Begin, K. J.; Morris, R. J. J. Chem. Educ. 1996, 73, A33–A39. 12. Goldstein, J. I.; Newbury, D. E.; Echlin, P.; Joy, D. C.; Fiori, C.; Lifshin, E. Scanning Electron Microscopy and X-ray Microanalysis; Plenum Press: New York, 1989. 13. Jenkins, R.; Snyder, R. L. Introduction to X-Ray Powder Diffractometry, Wiley–Interscience: New York, 1996. 14. Seah, M. P., Briggs, D. In Practical Surface Analysis: Auger and Xray Photoelectron Spectroscopy; Briggs, D., Seah, M. P., Eds.; John Wiley & Sons Ltd.: New York, 1990; Vol. 1. 15. Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. In Handbook of X-ray Photoelectron Spectroscopy; Chastain, J., Ed.; Perkin Elmer Corporation: Eden Prairie, 1992.

Journal of Chemical Education • Vol. 75 No. 10 October 1998 • JChemEd.chem.wisc.edu