Painted statuary, icons and architecture Few differences exist between the analysis of pigments on fine art and manuscripts and those on three dimensional art and architecture except for considerations of the sometimes immense size and irregular shape of the latter. These physical constraints have meant that samples were normally taken from objects either too large or irregularly shaped to fit easily under a traditional microscope. However, fibre optics for remote laser Raman spectroscopy, side-looking microscope objectives and customized microscope stages are alleviating this stricture.
Raman microscopic analysis of pigments used to decorate two Egyptian cartonnage masks found a diverse palette in use in Egypt in the third to fourth centuries BC that included realgar, pararealgar, cinnabar (HgS), haematite (α-Fe2O)), Egyptian blue (CaCuSi4O|0), calcite (trigonal CaCO3) and gypsum (CaSO4.2H2O) . The pararealgar appears to have been intentionally applied rather than present as a degradation product of realgar; no residual realgar was detected in admixture with pararealgar although it was present elsewhere on the mask. Whether the ancient Egyptians realized that they were using pararealgar as a yellow pigment or whether they mistook it for the more common orpiment is not known. Proof of an early appreciation by the Egyptians for this material as a pigment per se awaits further identifications, but already such evidence is mounting [22, 23, 40, 47, 67, 95, 96]. Guineau has shown that a similar, but much older yellow funerary mask from thirteenth-dynasty Egypt (1780 to 1680 BC) was in fact painted with orpiment and calcite .
The compound specificity of Raman microscopy has recently been coupled to the elemental depth profiling of LIBS for the analysis of a wooden Rococo altarpiece , a Byzantine icon , a Greek icon  and two Venetian miniatures executed on ivory . The LIBS technique uses laser radiation to generate a plasma at the surface of the sample and then collects the atomic emission of the atomized pigment. Multiple laser shots ablate successive layers of paint, thereby obtaining a micro-destructive, in situ cross-sectional elemental profile of the painting materials. Figure 4 shows the combined results of a LIBS-Raman analysis of a brown painted area on a Russian icon . The topmost paint layer failed to give a Raman spectrum using a 780 nm excitation line, but the LIBS data showed the presence of Fe, which, in combination with the negative result from the Raman analysis, suggested the use of an iron earth pigment. The brown earth pigments are known to be weakly scattering samples in the NIR region. LIBS spectra from successive laser pulses revealed that the brown paint was applied over a very thin silver foil. Elemental analysis can be especially important in iconography where silver and gold leaf, as well as pigments used to imitate these metals, are used extensively. Pure metallic pigments are Raman silent, and so the coupling of the two techniques is particularly powerful for providing a comprehensive analysis of an artist's palette. Finally, the lowest layers of the painting revealed large concentrations of Ca (Fig. 4, bottom left), suggesting a gypsum, calcite or mixed ground layer. In instances where the elemental profiles are inconclusive as to the identity of a particular pigment, the Raman technique is definitive; the Raman spectrum from this area (Fig. 4, bottom right) proves that the ground was in fact a mixture of gypsum and anhydrite (CaSO4). Although the studies mentioned above involved the sequential use of the two methods, a tandem LIBS-Raman spectrometer has been reported .
A thirteenth-century Spanish statue from Sasamon has been shown by FT-Raman spectroscopy to have been painted using several complex mixtures of pigments including cinnabar and red lead to form an orange-red colour as well as red lead possibly mixed with aurum musivum (SnS2, mosaic gold) and litharge or massicot to form a golden brown colour . The absence of bands due to a-quartz (SiO2) and the presence of those due to calcite in the spectrum of the HgS have been interpreted as indicating a local source of cinnabar, the Tarna mines. Pigment obtained from the other large Roman-period mine at Almaden, Spain, always bears tell-tale traces of quartz due to its volcanic origin, while calcite has been found as a
component of cinnabar ore from Tarna. Although such interpretations are worth noting, one must be cautious, since the purposeful addition of quartz sand as an aid to pulverization of the pigment, or the adulteration of cinnabar with chalk either to lighten the colour or as a profiteering trick of the medieval apothecary, cannot be ruled out. Furthermore, the appearance of either of these ancillary components in the paint as a byproduct of its application to lime plaster or surfaces otherwise prepared to accept pigments would not be out of the ordinary, and therefore provenance based on these indicators alone could be misleading.
The aureate pigment mixture is noteworthy because of the suggested presence of mosaic gold. However, this conclusion is surprising since the identification of SnS2 was based on a single, weak Raman band at the same wavenumber as a weak band of red lead, 313 cm-1, shown unequivocally to be present in the mixture. Definitive elemental data were not collected, and so the identification of mosaic gold in this instance must be taken with caution. Massicot or even pale litharge could equally well account for the yellowish colour of the pigment without invoking the presence of mosaic gold. Although SnS2 itself has a metallic lustre, it was discovered that the grinding of that material with a lead oxide produced a rich golden mirror-like surface. Upon examination, the lustrous material was thought to be a Pb-Sn alloy since it failed to give a Raman spectrum, suggesting that the original pigment mixture, if composed of red lead and mosaic gold as claimed, could indicate
Fig. 4 LIBS spectra (left) from successive laser pulses reveal the elemental makeup of paint
layers on a Russian icon. The first two layers did not yield Raman spectra (right), while the bottom layer is shown by its Raman spectrum (a) to be composed of (b) gypsum (CaSO4.2H2O) and (c) anhydrite (CaSO4) .
a lost technology for forming a faux gilding material. However, why it apparently existed on the statue in its mixed components rather than as the Raman-silent alloy discovered in the grinding experiments was not addressed.
Marble from the facade of the Certosa of Pavia was sampled for analysis by Raman microscopy to identify the nature of red stains . These had been thought to arise from bacterial growth, but were instead shown to be red lead, most likely from oxidation of lead salts originating from the degradation of lead architectural elements. Small green spots affecting the surface were also identified by the Raman technique as being Chlorophyta micro-organisms based on their association with carotenoids as indicated in the Raman spectrum. In another study, the identification of PbO and red lead in underlayers of gilded stucco provided data on the Baroque technique for applying gold to plaster . The lead compounds were used as siccatives to aid the drying of linseed oil mordants for the gold leaf.