Identification of pigments and binders of a 17th century mural painting (Bolivia). New report on pigments associated with Andean minerals

 Identification of pigments and binders of a 17th century mural painting (Bolivia). New report on pigments associated with Andean minerals. Eugenia Tomasinia*, Ilaria Costantinib, Valeria Careagac, Carlos Rua Landad, Kepa Castrob, Juan Manuel Madariagab, Marta Maiera,c, Gabriela Siracusanoa. a CONICET, Centro de Investigación en Arte, Materia y Cultura, IIAC, Universidad Nacional de Tres de Febrero, Avda. Antártida Argentina 1355 (C1104ACA), Ciudad Autónoma de Buenos Aires, Argentina. euge[email protected] , [email protected] b Departamento de Química Analítica, Facultad de Ciencia y Tecnología, Universidad de País Vasco UPV/EHU, P.O. Box 644, 48080 Bilbao, España. [email protected], [email protected], ju[email protected]. c UMYMFOR-CONICET, Departamento de Química Orgánica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellón 2, Ciudad Universitaria, (C1428EGA), Ciudad Autónoma de Buenos Aires, Argentina. [email protected], [email protected]. d Ministerio de Culturas y Turismo, Taller de Conservación y Restauración del Patrimonio Mueble, Ayacucho S/N, La Paz, Bolivia. [email protected] * Corresponding author. Tel.: +54 1148930625. [email protected] Abstract Research on the materiality of Andean Colonial artistic heritage contributes to the knowledge on the cultural and social history of the region. In this work we present the results of an interdisciplinary and multi-analytical investigation on a mural painting of the church of Curahuara de Carangas in Bolivia. This painting has been hidden for years behind the main 18th century wooden altarpiece. Its stylistic and pictorial characteristics place it in the early 17th century (1608), while other mural paintings of the temple correspond to a second stage, by 1777. Pigments were identified by scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDS) and Raman micro-spectroscopy. The application of imaging methods not only allowed the identification and characterization of the materials but also gave information on the manufacturing technique of this mural painting. An outstanding finding in our work was the identification of a mixture of two copper minerals, atacamite [Cu2Cl(OH)3] and antlerite[Cu3SO4(OH)4], as the green pigment. Atacamite and antlerite have been separately identified in a polychrome sculpture and a mural painting, respectively, but this is the first report on their use as a mixture. Other pigments identified in the mural painting were hematite, azurite, red lead, cinnabar, gypsum, cerussite, orpiment, and carmine lake. These results, together with the identification of traces of minerals of the region, indicate the use of local materials by Andean painters from the early Colonial period. This suggests continuity in the cultural practices from preHispanic times and an appropriation and reformulation of traditional European painting practices adapted to new resources. Analysis of lipid and protein materials by gas-chromatography coupled to mass spectrometry This is the accepted manuscript of the article that appeared in final form in Journal of Cultural Heritage 62 : 206-2016 (2023), which has been published in final form at https://doi.org/10.1016/j.culher.2023.05.030. © 2023 Consiglio Nazionale delle Ricerche (CNR). Published by Elsevier Masson under CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     (GC-MS) indicated the use of a secco painting technique involving the use of egg as pigment binder. Identification of animal glue suggested its use as the plaster primer. The knowledge of the materials and the painting technique of this artwork will allow the identification of similarities and differences with other mural paintings in churches of the Andean region and give useful information for conservation and preservation of this artwork. Keywords Pigments analysis; Mapping SEM-EDS; Raman imaging; Mural Painting, Spanish American Colonial art. Highlights   Analyses with complementary techniques were applied to the old painted altar corresponding to the first decoration of the church of Curahuara de Carangas (Oruro, Bolivia), apparently made around 1608 and which remained hidden for years behind the wooden altarpiece of the 18th century.  The application of SEM-EDS and Raman imaging methods provided useful information to identify and study the distribution of pigment within each paint layer as opposed to spot analysis.  The use of a mixture of atacamite with antlerite and the presence of minium and vermilion in a mural of the Colonial period has never been reported before.  PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     1. Introduction 1.1. Previous studies of Andean Colonial mural paintings The study of Andean Colonial artistic heritage contributes to the knowledge of the cultural and social history of the region. As part of an international and interdisciplinary project on the mural paintings of several Andean Colonial churches, built between the 17th and 18th centuries, several chemical and art historical studies were carried out [1–7]. The churches were located on the Ruta de la Plata (Silver Route), a road that was used to transport mercury and silver minerals from the city of Potosi (Bolivia) to Arica (Chile) on their way to Spain via Panama. This route also served to trade possessions and products, as it communicated important cities such as Cuzco, Arequipa, La Paz and Sucre. This movement of people and cultures promoted, in the Andean communities along the road, the architectural and artistic styles of the main churches of the Silver Route [6]. Churches with mid-18th century Colonial mural programmes are preserved in this area: Nuestra Señora de Copacabana de Andamarca, Santiago de Curahuara de Carangas, Santiago de Callapa, and San José de Soracachi in Bolivia, as well as San Andrés de Pachama in Chile, among others. These paintings are found on the walls of the naves and feature decorations with vases of flowers, fruits, and birds, along with religious iconographies developed to contribute to the evangelisation of the local population. Previous art historical studies have paid attention to their iconographic programs, their correspondence with written sources and oral practices like sermons, as well as their conservation status. [8–12] With the aim of investigating the materials and the pictorial technique of the murals of the churches on the Silver Route, questions were asked about history, chemistry and conservation, which gave rise to several research projects published in specialised journals[1–3,13–15]. The original remaining sections of these paintings (other parts have been restored) show little chromatic changes, alterations, or degradation phenomena thanks to the climatic and geographical conditions of the Andean highlands, which proved to be natural agents of preservation. In order to identify and characterise the colour palette and investigate the technique of these mural paintings, analyses were carried out using different techniques. Optical microscopy (OM) was used for visual and morphological analysis and scanning electron microscopy with a coupled probe for energy dispersive X-ray spectroscopy (SEM-EDS) for an elemental analysis of the area under study. Fourier transform infrared spectroscopy (FTIR-ATR) and micro-Raman spectroscopy were used for molecular analysis. On the other hand, analysis of organic dyes was accomplished by high performance liquid chromatography (HPLC) while lipidic and protein binders were analysed by gas chromatography coupled to mass spectrometry (GC-MS). The development and application of a methodology that emphasises the interdisciplinary and multi-analytical approach made it possible to establish the pigment palette and obtain precise and detailed information on the technique of mural paintings in 16th-18th century churches on the Silver Route. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     A series of iron oxides used as ochre pigments were revealed. Yellow or red earths (limonite, goethite, and hematite) were widely used as pigments for these colours in the Andean region. Hematite or almagre has been identified as a pigment highly used in Colonial paintings [16–19] as well as for the base preparation in a polychrome sculpture, being part of the "bol armenio" along with other compounds such as clays [20]. These have been identified in most of the murals where ochre shades were used, usually with quartz or mixed clays that are available in the region. Orpiment, an arsenic sulphide (As2S3), is shown in its yellow colour in some of the murals. This was one of the most appreciated pigments used in the Andean Colonial palette, despite its poisonous character [18]. It was not only used as a yellow colour, but also in mixtures with indigo or Prussian blue to obtain different shades of green [16–19,21]. Orpiment has also been detected in most of the illustrations in the 17th century manuscript of Getty Murúa's Historia General del Piru [22] and its discovery in funerary objects of a woman found in an ancient burial site at Chorrillos in Chile indicates the availability of orpiment as a pigment since pre-Hispanic times in the Andean region [23]. A red pigment widely used in the Andean area was also vermilion (HgS). It was found in the murals of the churches of San Andrés de Pachama, Curahuara de Carangas and San José de Soracachi, and in others along the Silver Route together with the presence of haematite [2,24]. Although in general the analytical studies carried out do not differentiate between the mineral (cinnabar) and the synthetic compound, the samples showed the presence of impurities, such as silicates, which may occasionally be associated with the mineral, indicating a natural origin. The source of cinnabar in America was the mines of Huancavelica and others near Guamanga in Peru [25], and its use prior to the arrival of the Spanish in the region is proven in mortuary contexts and especially as body paint in Peru in rituals and sacred ceremonies [18,26,27]. Later, it reappears in several Colonial polychrome paintings and sculptures as is well known [16,17]. On the other hand, in the mural paintings of the churches of Pachama in Chile and Orurillo and Marcapata in Peru, the use of a dark red colour was observed and the use of carmine lake was identified [28]. This dye is a red lake prepared from carminic acid, which is found in the insect Dactylopius coccus Costa known as cochineal. Its presence was confirmed by HPLC in a micro-destructive way and with previous preparation hydrolysis in acidic methanol [29]. Cochineal has been used as a textile dye since pre-Inca times and as a pigment in paintings, sculptures and Colonial textile objects from the 17th and 18th centuries [16,18,29]. This finding provided new information on the use of carminic acid lake as a pigment in Colonial mural paintings. Another natural organic dye identified was indigo, which is prepared from the fermentation in water of plant leaves, mainly of the species Indigofera tinctoria, and has been one of the most widely used blue pigments as evidenced by its identification in several 18th century paintings in South America [22,30–32]. Although indigo was traded from different parts of the Americas, mainly Central America, the existence of a local market in the study area cannot be ruled out due to the high demand, especially for dyeing and painting canvases [18]. The presence of indigo was a constant discovery in murals in Bolivia, Chile and Peru, and in all cases its great stability and deep tone is a phenomenon that is still being investigated [1,2,24]. Another blue pigment found, less frequently, was smalt, which was produced in Europe by roasting cobalt minerals to PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     obtain cobalt oxide. It was not produced in the Viceroyalty of Peru, so it probably had to be imported from Europe [33]. As part of our research on the chromatic palette of Colonial works of art, smalt has been identified in several 17th and 18th century Andean paintings and in the mantle of the Virgin of the gilded polychrome sculpture of Our Lady of Copacabana in Bolivia, among other research [30,34]. The identification of the use of two basic copper sulphates: antlerite (Cu3SO4(OH)4) and brochantite (Cu4SO4(OH)6) as green pigments is highlighted [1,2,6]. The local availability of both sulphates, their heterogeneous composition, the natural mixture with other oxides and the crystallinity in the wall painting samples point to a mineral origin of the pigment antlerite. This was the first time these minerals were identified in Colonial art. This finding expanded the knowledge about the palette of green pigments that our interdisciplinary group had already identified in Colonial easel paintings and polychrome sculptures [16,17,19,20,35]. These minerals were available in the Atacama Desert in northern Chile, which is rich in copper mines exploited since pre-Hispanic times [36,37]. On the other hand, copper mineral-based pigments have been identified in green paints on pre-Hispanic animal hide masks in the region [38]. The study of the black pigments of the murals gave rise to several investigations into carbon-based black pigments in order to better define their material characterisation and provide possible paths for identifying them through the use of different techniques [13,14,39]. This proposal was enriched by the cross-referencing of different historical sources that allowed the study of the practices associated with the manufacture of these pigments. At the same time, these provided indications of the symbolic significance of the colour black in the Spanish-American Colonial period [40]. This methodology was applied to different samples of black areas of several murals and it was found that all of them coincided with the use of charcoal. Regarding the manufacturing technique, we have determined in several Andean murals the use of gypsum as a preparation ground on the adobe walls of the churches and identified an animal glue, presumably used as a primer of the gypsum layer [1,2,28,42], in accordance with recommendations for a secco painting technique in art treatises [43]. Recently, we have identified muscle proteins together with collagen in microsamples from mural paintings in two 18th century Andean churches in Chile [44]. This finding points to a local preparation of the glue, presumably using waste from cow and vicugna, a domesticated camelid used as food and for transportation in the Andean region since pre-Hispanic times. In addition, fatty acids and cholesterol were also identified, revealing the use of a mixture of vegetable oil and egg (tempera grassa) as the pigments´ binder. These results and our previous findings suggest a preference for the use of minerals as pigments in Andean church wall paintings, suggesting a natural origin of the pigments and the development of technological practices for their production in the Andean region since pre-Hispanic times. The reason for this selection may be related to the local availability of these minerals compared to others. 1.2. Research aim This paper presents the results of the analysis with complementary techniques (micro Raman spectroscopy, scanning electron microscopy-energy dispersion spectroscopy (SEM-EDS), high performance liquidchromatography with diode-array detection (HPLC-DAD) and gas-chromatography coupled to mass PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     spectrometry (GC-MS) of the former painted altar found in the church of Curahuara de Carangas (Oruro, Bolivia) [45]. This mural became visible when restoration work was carried out in 2011, after having been hidden for years behind the main wooden altarpiece since the 18th century. It represents the presentation in the Temple and Purification of Mary, following a Flemish print by Lucas van Doetecum after Gerard Groenning, and it corresponds to the first decoration of the temple, apparently done by 1608 (Fig. 1). The characterization of the materials and pictorial technique in this mural painting will allow to identify similarities and differences with previous results on Andean Colonial mural paintings and to contrast them with the artistic literature of the period. Micro-sampling in damaged areas was carried out for this work. Although the sampling could be considered invasive, only a minimum number of materials was taken. Indeed, the microimaging and microspectroscopy techniques employed allowed extensive studies on a micrometric single sample. This analysis can be done directly on the surface or in each layer in crosssections of the sample. These cross-sections will be part of the sample patrimony archive of the Centro de Investigación en Arte, Materia y Cultura (Centro MATERIA) of the Universidad Nacional de Tres de Febrero (Buenos Aires, Argentina) and will be made available for future analysis. Many heritage objects are created with a wide variety of materials, and they often present heterogeneities at different scales. In this sense, spectroscopic techniques equipped with imaging capabilities are useful because the information obtained allows to understand the materials and production techniques, as well as the deterioration of the objects in order to plan better strategies for their conservation. The non-invasive microscale analysis used in this work involved imaging, using optical and electron microscopy, and for pigment identification, scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDS) and Raman micro-spectroscopy were used. Micro-destructive techniques such as HPLC-DAD and GC-MS contributed to the identification of organic compounds and the characterization of the painting technique. 2. Experimental 2.1. Painting samples Eight microsamples were extracted with a scalpel from coloured areas of the former painted altarpiece present behind the actual wooden altarpiece of the church of Curahuara de Carangas (samples CCG14 - CCG21) (Fig. 1). The location of the samples and their colours is shown in Table 1. Fragments of samples with an area of less than 1 mm2 were embedded into an acrylic transparent resin Subiton® (Buenos Aires, Argentina) and polished with sandpaper of decreasing size (until to 12.000 mesh) to prepare the crosssections (Fig.1). 2.2. Instrumental and methodology Optical Microscopy (OM) PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Observation and photography of the surface and cross-section of the samples was achieved using a Leica MZ6 stereomicroscope and a Leica DM 750 microscope equipped with visible and ultraviolet light sources in normal and polarized modes. Scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDS) SEM analysis by secondary electrons (SE) and backscattered electrons (BSE) were carried out using a Field Environmental Scanning Electron Microscope Zeiss Supra 40 coupled with an EDS microanalysis INCA X Sight, Oxford Instrument. Analyses were carried out using an accelerated potential of 20 kV. Cross-sections of the samples were coated by sputtering with a thin (less than 80 Ẳ) layer of platinum. Several measurements were performed on selected areas or certain grains of each layer of the samples. Micro-Raman spectroscopy Raman spectra of the cross-sections were acquired by a Renishaw InVia Raman confocal spectrometer, coupled to a Leica DMLM microscope using a 532 and a 785 nm diode pumped solid state Nd:YAG excitation laser source. The spectra were acquired with the Leica 50X N Plan (0.75 NA) lens with a 2 μm spatial resolution and the Olympus 100X (0.8 NA). The minimum theoretical spot diameter using the 532 nm laser was, for the Leica 50X and Olympus 100X, 0.9 μm and 0.8 μm, respectively, while using the 785 nm laser it was, for the Leica 50X and Olympus 100X, 1.7 μm and 1.1 μm, respectively. Additionally, for visualization and focusing, Leica 5X N Plan (0.12 aNA) and a 20X N Plan EPI (0.40 NA) lens were used. For focusing on and searching for points of interest, the microscope employs a motorized stage (XYZ). The power applied was set at the source at a maximum of 50 mW while on the sample it was always less than 20 mW. Normally, 10-300 scans, each lasting 1-20 s, were accumulated to achieve a suitable signal-to-noise enhancement at an operating spectral resolution of ≤1 cm-1. Spectra were acquired between 150 and 3200 cm-1. Stream Line technology was employed to generate high-definition molecular images of specific samples. Briefly, the motorized stage of the microscope coupled to the Raman spectrometer moves the sample beneath the lens so that the line of laser-light is rastered across the area under analysis. Afterwards, from the collected spectra, chemical images are generated by considering the position of main band of the compound of interest. In both cases, the Renishaw Wire 3.2 (Renishaw, UK) software was used for data collection and analysis. High performance liquid chromatography with diode array detector (HPLC-DAD) HPLC–DAD analysis was carried out with a Gilson 506C HPLC system using a Phenomenex Gemini 5μ column (25 cm length, 4.6 mm internal diameter). Gradient elution was performed using mixtures of MeOH and 1% (v/v) aqueous orthophosphoric acid, as described elsewhere [34]. The detection wavelength was 540 nm. A grain from sample CCG19 and a carmine lake standard (Carmine Naccarat 42100, Kremer Pigmente, Germany) were hydrolyzed in acidic methanol as reported previously [28]. Gas chromatography coupled with mass spectrometry (GC–MS) PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Samples were extracted with ammonia 2.5 M and chloroform to separate proteins (ammonia phase) from lipids (chloroform phase), as reported previously [2]. Fatty acids were derivatised as their methyl esters (FAME) [46] and analyzed by GC–MS, as described before [2]. Trimethylsilyl sterol derivatives were prepared and analyzed by GC–MS as reported in Lantos et al. [47]. Mass spectra for sterols were measured in the selected ion monitoring (SIM) mode with ions at m/z 129, 329, 368, and 458 for cholesterol. Amino acid derivatives were prepared following the procedure of Bersani et al. [48] and analyzed by GC-MS, as described in literature [2]. 3. Results and discussion 3.1 Cross-section analysis by mapping Blue sample and ground layer The blue sample (CCG14) taken from the Virgin mantle (Table 1) showed mostly Cu in the SEM-EDS mapping of the stratigraphy. In addition, minor elements such as S, Si, Ca and Al suggested the presence of a copper pigment, likely azurite (Cu3(CO3)2(OH)2) (Fig.2). Unfortunately, carbon signal of the pigment would be overlapped by the carbon signal coming from the binder. Certain greenish crystalline zones at the OM showed vanadium content, and a punctual EDS analysis yielded higher amounts of Cu, Pb and V and lower amounts of As, Fe, Ca, Mg, Al, Si and Cl. Figure 3 shows the images obtained from the Raman spectroscopy mapping by using the 532 nm laser, showing the presence of azurite in most of the bluish crystals on the surface, but also mottramite, in the green crystals, corresponding to a basic vanadate of copper and lead (PbCu(VO4)) that is dispersed on the surface. Additionally, hematite (Fe2O3) and anatase (TiO2) were identified by Raman spectroscopy, in accordance with the minor presence of elements such as Fe, Ti, etc. Mottramite is a vanadate mineral of the descloizite group that may be associated with azurite in the Andean area, whose structure forms a series with duftite (PbCuAsO4(OH)), in which vanadium is replaced by arsenic [49], an element that also appears as a trace element in the pictorial layer. The Raman spectra at 532 nm of the minerals present in both the pictorial layer and the preparation layer are shown in Figure 4. Spectra recorded on the pictorial layer showed Raman bands at 180, 238, 333, 401, 541, 764, 840, 1096, 1430 and 1578 cm-1 characteristics of azurite [50,51]. In addition, the spectrum of hematite (Fig. 4a) shows a band profile compatible with a natural origin [52,53]. Figure 4b shows the Raman spectra of mottramite obtained with 532 and 785 nm laser sources with main bands at 113, 293, 333, 366, 617, 714, 804 and 828 cm-1 [54]. The heterogeneity of the pictorial layer of azurite suggests a natural origin, but also a local one considering the presence of the minerals described above. To our knowledge, this is the first time reporting the presence of mottramite as a minor mineral in a blue paint layer of azurite. On the other hand, Figure 4a shows the main Raman bands of anhydrite, mostly in its type I (cubic) polymorph [55] in the ground layer, which is consistent with the presence of the CaSO4.xH2O system already PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     observed in other mural paintings where the use of a local mineral mixture is proposed [1,3]. In addition, a minor presence of clay component elements such as Si, Al, K and Mg is observed at the ground layer of all samples, showing the heterogeneity of the mixture. Green sample The cross-section of the green sample (CCG15) showed a pictorial layer composed mainly of Cu according to the SEM-EDS analysis, but in certain crystals it is accompanied by S and Cl (Table 1). The SEM-EDS mapping image clearly shows that these elements are found in different crystals associated with copper (Fig. 5). Two green copper-based minerals, atacamite [Cu2Cl(OH)3] and antlerite [Cu3SO4(OH)4], were identified from the image obtained from the Raman mappings (Fig. 6) by selecting the main Raman band of each compound. This indicates the use of a mixture of green pigments that had been characterised separately in a wooden sculpture [20] and in a mural painting [2,3], but had never been found together until now. The Raman spectra of each mineral (Fig. 4c) show bands at 226, 357, 452, 516, 821, 912 and 976 cm-1 for atacamite [20,56] and bands at 269, 418, 989 and 1077 cm-1 for antlerite [2,3,57]. For the antlerite identification, the possible thermal decomposition of the mineral under the laser beam, previously observed, was considered, so the power and time of measurement were controlled to avoid such decomposition [3]. Flesh colour sample SEM-EDS mapping of the flesh of the Child´s leg sample (sample CCG16), shows the content of Pb and Hg, probably a mixture of red cinnabar or vermillion [HgS] with lead white (albayalde) [Pb3(CO3)2(OH)2], which is to be expected for flesh of that period (Fig. 7). However, it is observed in the images of the Raman mapping in figure 8, that in the area where only lead is found, there is a mixture of cerussite [PbCO3] with the characteristic band at 1054 cm-1 [58] and minimum [Pb3O4], with bands at 227, 314, 478 and 551 cm-1 [59] (Fig S1). In addition, the presence of cinnabar (or vermilion) is confirmed by the Raman bands at 108, 255, 280, 289, 343 and 352 cm-1 [60] (spectrum not shown). This mixture of vermilion and minium has not been found in other murals in the region and could have been missed without Raman mapping. Orange sample SEM-EDS mapping of the orange sample CCG20 only identified lead as the chromophore element (Table 1). Raman spectroscopy analysis (Fig.S2) shows the presence of two lead compounds, the red minium [Pb3O4], with bands at 124 and 551 cm-1 [59] and massicot, a yellow lead oxide [PbO], with bands at 142, 289 and 392 cm-1 [61]. Raman mapping yielded the images in figure 9, which show the mixture of pigments distributed in the pictorial layer. This combination of pigments has not been identified in Colonial Andean murals in the region until now. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Table Table 1. List of samples, colour, location, and results of SEM-EDS analysis on the pictorial layers of the cross-sections of the wall painting samples. Major elements are marked in bold style. The elements at trace levels are shown between parenthesis. Table (Supplementary Material) Table 1S. Relative amino acid composition (in percentages) of samples CCG15 and CCG21. Sample Location Colour Elements of the pictorial layer CCG14 Vir g in mantle Blue Cu, S, Si, Ca, Al ( V, Pb, Ti, As, Fe, K, M g) CCG15 Table Green Cu, S, Cl, Si, Al, Ca ( Fe, K, M g) CCG16 Child's le g Flesh Pb, Ca, S, Al, ( H g , Fe, K, M g , Na ) CCG17 Child cloth White Pb, Ca, S, Al, ( Fe, K, M g , Na ) CCG18 Man's cape Yellow As, S, Ca, Fe, Si ( K, Al, M g , Na ) CCG19 Curtain Red N, P, Na, Al, S, Ca, K ( M g , Si, Cl ) CCG20 Man behind vir g in Oran g e Pb, S, Ca, Si, Al ( Fe, K, M g) CCG21 Frame Black Si, Ca, Al, Fe, Na, M g , S, Cl Sample Ala Gl y Th r Se r V al Leu Ile Pro H y p Asp Glu Phe CCG15 6.8 12.5 5.5 9.2 12.5 5.9 0.9 7.8 2.5 16.6 14.8 5.0 CCG21 10.9 19.1 4.7 4.8 4.9 6.0 0.7 12.9 10.3 7.6 13.8 4.3 PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure Captions Figure 1. Image of the mural paintings found behind the altarpiece in the church of Curahuara de Carangas. (Ph. Gabriela Siracusano) and cross-section of the samples. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure 2. a) BSE micrograph, b) stereoscopic image, c) MO image of an area of the cross-section of sample CCG14 and d) mapping of copper (Cu), vanadium (V), silicon (Si), aluminium (Al), sulfur (S), calcium (Ca). Figure 3. Raman mapping image of the pictorial layer of the stratigraphy of the blue sample CCG14. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure 4. Raman spectra of the blue sample CCG14. a) azurite and hematite (pictorial layer) and gypsum (ground layer) at 532 nm, and b) mottramite at 532 nm and 785 nm (pictorial layer). d) Raman spectra of the green minerals identified in sample CCG15 at 532 nm. Figure 5. a) BSE micrograph, b) stereoscopic image of an area of the cross-section of the green sample CCG15 and mapping of copper (Cu), calcium (Ca), chlorine (Cl) and sulfur (S). PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     F igure 6. Raman mapping image of the pictorial layer in the cross-section of the green sample CCG15 (Magnification 100X). a) Cross-section, b) antlerite and c) atacamite Raman mapping images. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure 7. a) BSE micrograph of the cross-section of the flesh sample CCG16 and b) detail of the mapped zone. c) Mapping of lead (Pb) and d) mercury (Hg). PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure 8. Raman mapping image of the pictorial layer in the cross-section of the flesh sample CCG16 (Magnification 100X). a) Cross-section, b) cinnabar c) lead white and d) minimum. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure 9. Raman mapping image of the pictorial layer in the cross-section of the orange sample CCG20. (Magnification 100X). a) Cross-section, b) minium and c) massicot. Figure Captions (Supplementary Material) PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure S1. Raman spectra of the lead-containing minerals identified in the flesh sample CCG16 at 532 nm. Figure S2. Raman spectra of the lead-containing minerals identified in the orange sample CCG20 at 785 nm. PapersubmittedtoJournalofCulturalheritage https://doi.org/10.1016/j.culher.2023.05.030     Figure S3. HPLC chromatogram of hydrolysed sample CCG19 at 540 nm and UV spectrum of the peak at 13.8 min. Figure S4. Raman spectra of the minerals identified in the a) white sample CCG17, b) yellow sample CCG18 and c) black sample CCG 21.