SR-XRD and SR-FTIR study of the alteration of silver foils in medieval paintings
Altarpieces and polychrome carved wood from the fifteenth century AD usually exhibit golden and silvery areas by the application of a very thin foil of metal. The metal foils were normally protected from the atmosphere by a varnish or resin which maybe either preserved or absent. Moreover, they were glued to the background surface by adhesive substances (egg yolk, drying oil or animal glue). The high proportion of the glueing substances often renders the development of reaction compounds. With time, silver alters blacken or simply disappear completely. In this paper, we study the alterations to metal foils from a selection of fifteenth century artworks showing different glueing agents, organic coatings and several degrees of conservation of the organic coatings and metal leafs. The submillimetric layered structure and the high variability and low amount of most of the compounds present in the different layers, as well as their differing nature (organic and inorganic) make the use of micron-sensitive high-resolution techniques essential for their study. In particular, the high resolution, high brilliance and small footprint renders synchrotron radiation most adequate for their study. SR-XRD was performed to identify the reaction compounds formed in the different layers; μFTIR was used at to identify the silver protecting organic coatings, the metal foil glueing layers and the corresponding reaction compounds. The results obtained suggest that atmospheric corrosion is the dominant mechanism, and therefore that the degree of corrosion of the metal foils is mainly related to the conservation state of the protecting coatings.
KeywordsSilver foil Silver chloride Silver sulphide Cultural heritage XRD FTIR
Silvery and golden finishes were frequently used in altarpieces in the fifteenth century to afford the aspect of metal in certain ornamentation, garments and mouldings [1, 2]. In most cases, both silver and gold were produced by the application of a very thin foil of metal. Metal leafs were applied directly over the ground layer (made of gypsum mixed with animal glue) with the help of an adhesive such as animal glue, egg yolk, drying oil, etc. or over an special preparation layer called bole, constituted by a clay enriched with iron oxides and mixed with proteinaceous substances. These glueing layers usually contain a high proportion of the adhesives. Silver foils, contrarily to gold foils, were always protected by a resinaceous coating to prevent their alteration. Moreover, silver foils were also used to give the appearance of gold in some areas and mouldings. This appearance of gold (simulating gold) was achieved by the application of a transparent yellow-coloured organic coating (typically a natural resin).
The phenomenon of atmospheric silver corrosion has been studied by several authors [3, 4, 5, 6, 7] and the stability of the different compounds formed (presence of sulphides and chlorides and absence of carbonates, sulphates, nitrates and organic salts) has been established mainly resulting from humid environments. Although sulphur pollutants are very reactive to silver, silver chlorides and oxides are more stable products with time . Silver corrosion in the field of cultural heritage has been devoted chiefly to coin corrosion [8, 9, 10]. The formation of silver chlorides has been associated to the oxidising properties of near-surface aerated sea water and land burials and together with the presence of copper oxide precipitates, responsible for the embrittlement of the coins. The coins contain copper and lead and thick crusts of copper corrosion products (Cu2O) are formed. Synchrotron studies of silver alloys in silverpoint drawings have also been performed; the presence of small amounts of mercury has been associated to atmosphere contamination [11, 12]. Our study proposes to elucidate the processes involved in the alteration of the silver leafs applied on altarpieces and polychrome carved wood. The extreme thinness of the silver leafs applied make their alteration highly visible. Over time, silver alters and results in the blackening of the silvered areas or, in the most extreme case the disappearance of the metal silver foil revealing the preparation layer, i.e. the reddish colour of the bole. Some of these artworks also have original varnishes (organic coatings) applied either to protect them from weathering, to accentuate the lustre or to imitate gold. In some cases, the removal or loss of these protective layers along the artwork life facilitated the alteration of the silver surfaces. Moreover, the mere presence of glueing substances and varnishes may also affect the characteristics and degradation mechanisms involved in silver corrosion. Finally, the presence of impurities in the original materials, the existence of former restorations, the incorporation of products resulting from ageing and degradation, often due to inadequate conservation conditions, are also some of the typical difficulties encountered in the study of these materials. Silver corrosion may be affected by all this and the object of this study is to elucidate it.
The object of this study is to identify the compounds that are produced in the process of alteration of the silver foils as a first step to describe these mechanisms. Understanding this phenomenon from a chemical point of view may lead to the development of a better conservation strategy for the art work and may help to choose the best method for an appropriate restoration. Taking into account that the layers are submillimetric and that both glueing agents and reaction compounds are often confined to the contact area between layers, the use of micron-sensitive techniques is fundamental for their study. Moreover, the coexistence of a large number of substances makes the use of high-resolution techniques highly advisable. For these reasons, the high resolution, high brilliance and small footprint render synchrotron X-ray diffraction and infrared spectroscopy radiation most adequate [13, 14, 15, 16].
A selection of silver decorations from fifteenth century AD artworks which were kept in different atmospheric conditions and showing various degrees of conservation (some showed a completely darkened surface other a nice metal finish, cracked areas, loosing areas), glueing procedures (bole, egg yolk and oil), protecting layers (resins and painted areas) was sampled. The examples shown here are a selection of samples from the following artworks: the altarpiece of “Santa Llúcia” by Bernat Martorell (1400?–1452), the altarpiece “Sant Vicenç de Sarrià” by Jaume Huguet (141?–1492), both currently on display at the Museu Nacional d’art de Catalunya in Barcelona , “La Mare de Déu” altarpiece dated 1459 and attributed to Pasqual Ortoneda  from the Vinseum Museum of Vilafranca del Penedès (Catalonia) and the sculpture of “Sant Joan” by an unknown author currently at the church of Arties in the Val d’Aran (Occitania). All these works are from the period of the old Crown of Aragon and, in most of the cases, we took advantage of the fact that they were undergoing restoration. With the exception of the sculpture of “Sant Joan” that to our knowledge has been kept always in the same church (a small church without any environmental control in the Pyrenees), the other pieces have been moved during their life from the original churches to the most various places (the altarpiece from Vilafranca has been kept for nearly 100 years on a family tomb in a cemetery) until they have been moved to their final places in the Museum. Therefore, they have been kept in several conditions (for instance the Museum of MNAC in Barcelona is near the sea while the church of “Sant Joan” is placed in the mountains, 1,144 m high, far from the sea influence). Therefore, differences in the corrosion products found should also help in assessing the effect of the environmental conditions on the corrosion of silver foils.
The study requires sampling of the artworks which are quite often exhibited in a Museum with the limitation in the number and size of the samples that this implies. Sample taking is a key step in the study as the samples must be obtained from as few points as possible and at the same time must be representative. The necessary small size of the samples extracted, below 1 mm2, the layered structure (micrometric layers) formed by a mixture of heterogeneous compounds and the poor crystalline nature of some of them were the main reasons for using complementary analytical techniques; each analytical technique having specific geometric requirements and demanding different sample preparation procedures.
The samples taken were small in size (surface area of a few hundred micrometres side), from which two basic preparations were made: fresh-fractured fragments and polished cross-sections. For the cross-sections, the samples were embedded in a polyester resin polymerized by a peroxo-organic catalyser under low humidity conditions. After polymerization, cross-sections were obtained; a first cut with a diamond saw was afterwards polished with 1-μm size diamond paste and the block was cut again in a slice of about 200 μm in thickness.
Optical microscopy and scanning electron microscopy (SEM), JEOL-5600, with elemental analysis using the PCXA LINK EDS microanalyzer is used in order to obtain information of the composition, size distribution and homogeneity of the particles in the samples. Small fragments or cross-sections were carbon coated to ensure the electrical conductivity necessary to perform the SEM-EDS analysis. The EDS spectra were obtained at 20–25 keV and 1 nA current.
Micro-SR-XRD is the most adequate technique to resolve complex mixtures of crystalline compounds from micrometric samples where the several substances are present in very low amounts such as happens in paintings. The sample peculiarities make the brilliance and collimation of the synchrotron beam fundamental.
Micro-XRD data were obtained at Station BM16 (Spanish CRG) of the European Synchrotron Radiation Facility (ESRF, Grenoble). Small fragments cut from the samples were previously selected with the help of the optical microscope. These fragments were placed on an adhesive support and diffraction measurements were collected in transmission geometry using an X-ray footprint of 30 × 30 μm. Some other samples were thin cross-sections, in which case a footprint of about 20 × 50 μm was found to be the most adequate for the discrimination of the compounds present in the several layers. A smaller beam footprint may lead to a spotty single crystal-like X-ray diffraction pattern, dominated by only some of the big crystallites, therefore, precluding the identification of the other compounds. If the beam is bigger, it is not possible to separate the compounds present in the different layers. A map of compounds was obtained from thin cross-sections of the layered paints. The precise selection of the sample region, from which the 2D-diffraction pattern was obtained, was achieved by using a XYZ motorised goniometer head to hold the samples in combination with a motorised videomicroscope Ultrazoom 12×. The detector was a CCD ADSC Q210r and the X-rays were at 10 keV (λ = 1.24 Å) or 12.7 keV (λ = 0.98 Å). This setup provides high sensitivity to compounds present in very low amounts and a good low angular limit adequate for the organic compounds present but has some limitations in the angular resolution for highly overlapping diffraction patterns.
For adequate separation and unambiguous identification of the compounds high-resolution powder diffraction (HRPD) measurements were also performed at Station BM01 (Swiss/Norwegian CRG) European Synchrotron Radiation Facility (Grenoble). For these measurements, small fragments of the sample were put in a glass capillary to allow rotation avoiding preferred orientations in the diffraction patterns. The system consists of a two-circle diffractometer with six lined-up detectors covering 5.5° that ensures a very high angular resolution over the whole angular range (below 0.01°). Measurements were taken in transmission geometry using a 24.9 keV (0.50 Å) X-rays and collected during a period of 5–15 h. In this setup, the whole fragment is analysed but the high angular resolution minimises the overlapping of the diffraction patterns corresponding to the several compounds, which is particularly important when some the compounds dominate allowing unambiguous identification and quantification of the compounds. However, this setup has some limitations in detecting substances present in very small amounts. A combination of both techniques micro-XRD and HRPD has proved to be very useful for the identification and quantification of the different compounds present.
SR-FTIR microspectroscopy analysis was also performed to extract chemical and structural information on the different compounds present; both the silver protecting organic coatings, as well as the metal foil glueing the layers and corresponding reaction compounds. This technique is particularly useful due to its capability of mapping, on a micrometric scale, its sensitivity to organic materials and its one-to-one relationship between molecule type and spectral fingerprint [14, 15, 16]. μSR-FTIR measurements were taken at beamline B22 of the Diamond Light Source (the Bruker 80 V Fourier transform IR interferometer is equipped with a Hyperion 3,000 microscope, MCT detector, measuring range 4,000–650 cm−1). The spectra were obtained measuring one of the sides of pressed fresh-fractured fragments of the samples in a diamond cell. Spectra were obtained in transmission mode and from different areas using a small spot of 10 × 10 μm/15 × 15 μm. This setup is essential in order to separate the several compounds present in the layers (organic, inorganic and reaction compounds) allowing their unambiguous identification which otherwise is not possible. It increases the capability of detecting compounds present in very small amounts and allows working in transmission mode that produces higher quality spectra.
Results and discussion
The samples were extracted taking into account the fact that silver foils had been modified in different ways over the years consequence of the pictorial technique used and the state of conservation of the protective layers. As a result, the various degrees of alteration lead to different states of conservation: from unaltered protected to fully absent silver foils. When the silver foil has disappeared completely, it is still possible to find the silver alteration products as a residue; they appear mainly present in the cracks showing filiform crystal growths. The presence of hair cracks on the external layers, inherent to the ageing process, the occurrence of porosity or the removal of the protective transparent or yellow tinged varnish layers as a consequence of former restorations favour the alteration of the metal silver foil. Furthermore, the environment in which the artwork is preserved may trigger the degradation process.
SEM-EDS analysis from the remnants of the original silver foil and from the corrosion products formed is shown in Fig. 1a–c, showing the presence of large amounts of chloride. The corresponding micro-XRD patterns (Fig. 2) show the presence of silver alteration compounds, i.e. AgCl chlorargyrite in a large amount and the cubic silver sulphide Ag2S argentite and Ag8S but in smaller quantity, among other substances. Ag8S is an intermediate compound that precedes the formation of the cubic Ag2S [19, 20].
A varnish layer is present on top of the silver layer, which is not crystalline and therefore cannot be determined by XRD. The most external layer shows also the presence of materials deposited from the atmosphere (calcite and carbon dust) together with efflorescent silver corrosion products, AgCl and secondary metal silver. The presence of metal silver is, due to the well-known high photo-sensitivity of silver, most probably consequence of the reduction of silver chloride when exposed to sunlight. This silver appears in the form of fine particles giving the black colour to the surface.
The second case studied belongs to the altarpiece from “Santa. Llúcia” by Bernat Martorell, and corresponds to a background area of the overall. This area was originally silver coated but later recoated with a gold foil over which several paint layers were applied.
Chemical analysis by means of EDS attached to the SEM from the altered area in the silver foil shows the presence of sulphide and in a much larger proportion chloride associated with silver. Fragmentation and separation of the sample layers was performed to study them separately. Two fragments were analysed by micro-XRD, one containing the sequence comprising the gold foil, red layer between the gold foil and the silver foil and the silver layer and a second fragment comprising the silver foil, the bole preparation over which the silver foil was stuck to the ground. The first fragment shows the presence of gold, metallic silver, the monoclinic form of Ag2S, acanthite, and AgCl, chlorargyrite in a larger proportion together with some small amounts of calcite –CaCO3–, calcium oxalates, weddellite –CaC2O4·2H2O– and quartz –SiO2– related to dust deposition on the surface (Fig. 3b I). The lattice parameter corresponding to the gold foil is smaller than expected and relates to a copper containing gold alloy of composition Au9Cu . The second fragment shows metallic silver and the compounds forming the bole,, clay minerals (kaolinite), quartz, calcite and hematite –Fe2O3– (Fig. 3b II). The silver foil as observed with optical microscopy appears in good state of conservation and while silver chloride is found, sulphide is not present. In this case, the chloride is basically found in the contact area between the silver foil and the bole. The presence of chloride in layers protected from the atmosphere (inner layers) maybe explained taking into consideration that elemental analysis of reference animal glues showed the presence of a small proportion of chloride in them. However, the extent to which this can be related to the formation of silver chloride in the internal layers has yet to be demonstrated.
In this case, the silver foil was glued directly to the ground layer. Both SEM-EDS and XRD analyses of the sample show the presence of an small amount of lead white –PbCO3/2PbCO3·Pb(OH)2– in Fig. 4c IV. IR analysis shows the presence of a high amount of lead carboxylates (2,955; 2,918; 2,849; 1,540; 1,514; 1,472; 1,462; 1,419 cm−1) and the lack of proteinaceous substances which suggests that the glueing material used was a drying oil. Treatises from the period  describe the addition of small amounts of lead white to the drying oil to help its polymerization and drying process. XRD analysis (Fig. 4d) shows the presence of lead white and of a mixture of crystalline carboxylates of different nature.
All the examples shown correspond to artworks preserved under very different atmospheric conditions. The silver alteration compounds formed are the same indicating a similar corrosion process in all the cases studied, which seems to depend more on the state of conservation of the protecting layers and the glueing agents used than on the environmental conditions. However, taking into account that the number of cases studied is small, a more thorough study should be performed to ascertain this.
The mechanism of formation of the silver alteration compounds involves the previous release of silver ions and the subsequent formation of silver chloride as chlorargyrite and silver sulphide as acanthite. Beforehand, the formation of silver sulphide intermediate species is obtained, such as Ag8S [19, 20] which has been determined in some cases by XRD. Although, silver chloride is the most extended and abundant corrosion product found, the presence of sulphides is also observed although present in lower quantities or missing if the silver foil is properly protected from the atmosphere. The presence of cracks and pores helps the penetration of the atmosphere in the inner layers and the growth and efflorescence of silver corrosion products causing tensions and flaking of the layers.
The formation of silver chloride and silver sulphide implies the action of different substances present in the atmosphere such as hydrogen sulphide, carbonyl sulphide and chlorides [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. These substances can be found in highly variable concentrations and in various relative proportions and originate from the decomposition of organisms and sea spray. Chloride and sulphide are also found in the binding and glueing media present in the altarpieces, such as egg yolk and animal glue.
The previous formation of silver ions necessarily implies the action of atmospheric oxygen (O2) in adequate conditions (humidity and temperature). The porosity of the materials and the formation of ageing cracks on the several protecting and glueing layers favour the penetration of oxygen and of humidity into the inner layers. However, silver oxides have not been determined by XRD, and although their absence is a possibility that cannot be withdrawn from the present data, their detection may be prevented by the amorphous nature of such compounds.
The alteration of silver foils applied to fifteenth century altarpieces and wooden carved sculptures was studied by different and complementary techniques. The most frequent corrosion product found and present in larger amounts is silver chloride in its form chloragyrite. The presence of sulphides is also observed although they are formed in lower quantities or missing if the silver foil is properly protected from the atmosphere. They are also absent in the contact areas between the silver foil and the bole or ground layer. Silver sulphides appear as the monoclinic form of Ag2S, acanthite in most of the cases studied. However, the development of sulphides seems to be related to the formation of some precursor, Ag8S, evolving into the cubic form of Ag2S, before the most stable monoclinic acanthite is formed. The alteration products present on the surface are silver chlorides only, which are driven to the surface by efflorescence through cracks or pores. The white silver chlorides are easily reduced back to metal silver due to the action of light; the metal silver fine particles formed are black and give a dark colour to the layer.
The formation of carboxylates from the reaction between the free fatty acids when present in the adhesive layers and the metal ions has been determined. They appear uniformly distributed in the layer without forming protrusions. The formation of these substances does not seem to have any influence in the stability of the silver leafs. On the contrary, silver chlorides show crystalline growths through the cracks and pores causing tensions and flaking of the layers.
The examples shown correspond to artworks preserved under very different atmospheric conditions, in various geographic areas with different climates. However, the alteration compounds formed are those formed due to atmospheric corrosion and the degree of alteration shown mainly depend on the state of conservation of the organic protecting coatings.
This work was funded by SM2410 project (beamline B22) at Diamond Light Source, EC-69 (beamline BM01, Swiss/Norwegian CRG) and 16-01-733 (beamline BM16, Spanish CRG) projects at ESRF. Part of this work was carried out within the framework of agreements of collaboration between the Technical University of Catalonia (UPC) and the “Museu Nacional d’Art de Catalunya” (MNAC) and the “Centre de Restauració de Béns Mobles de Catalunya” (CRBMC). We wish to thank the TdArt restorers and Vinseum Museum for their collaboration in the study and for the access given to sampling of the paintings. Received financial support under “Ministerio de Ciencia e Innovación” (Spain) Grant HAR2009-10790
- 1.Thompson DV (1956) The materials of medieval painting. Dover, New YorkGoogle Scholar
- 2.Brunello F (1997) Cennino Cennini, Il Libro dell’Arte, Neri Pozza (ed) VicenzaGoogle Scholar
- 6.Leygraf C, Graedel TE (2000) Atmospheric corrosion. Electrochemical society series. Pennington, New JerseyGoogle Scholar
- 9.Bozzini B, Giovannelli G, Mele C, Brunella F, Goidanich S, Pedeferri P (2006) An investigation into the corrosion of Ag coins from the Greek colonies of Southern Italy. Part I: an in situ FT-IR and ERS investigation of the behaviour of Ag in contact with aqueous solutions containing 4-cyanopyridine. Corros Sci 48:193–208CrossRefGoogle Scholar
- 13.Greagh D (2007) Synchrotron Radiation and its Use in Art, Archaeometry, and Cultural Heritage Studies in: D. Bradley, D. Creagh (ed) Physical techniques in the study of art archaeology and cultural heritage, Elsevier 2:1–95Google Scholar
- 17.http:/www.mnac.cat/, 2010
- 18.Ruiz I, Quesada F (2005) Pasqual Ortoneda i els retaules de Cabassers, la Secuita i Santa Margarida in: Enciclopèdia catalana (ed) L’art gòtic a Catalunya. Pintura II: El corrent internacional. Barcelona 160–161Google Scholar
- 19.Yu YD, Guan R, Hashimoto H, Marita Y (1995) Electron-microscope study of the structure of Ag8S formed in the initial stage of silver sulfidation. Acta Cryst B51:149–155Google Scholar