- Email: [email protected]
A multi-analytical study of ancient Nubian detached mural paintings
Microchemical Journal 124 (2016) 719–725
Contents lists available at ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
A multi-analytical study of ancient Nubian detached mural paintings☆ Noemi Proietti a, Valeria Di Tullio a,⁎, Federica Presciutti b, Gennaro Gentile c, Brunetto Giovanni Brunetti d, Donatella Capitani a a Laboratorio di Risonanza Magnetica “Annalaura Segre”, Istituto di Metodologie Chimiche, Consiglio Nazionale delle Ricerche (IMC-CNR), Area della Ricerca di Roma 1, Via Salaria km 29,300, 00015 Monterotondo (Roma), Italy b Dipartimento di Scienze Farmaceutiche, Università di Perugia, via Fabbretti 48, 06123 Perugia, Italy c Istituto per i Polimeri, Compositi e Biomateriali, Consiglio Nazionale delle Ricerche (IPCB-CNR), Via Campi Flegrei, 34, 80078 Pozzuoli (Napoli), Italy d Dipartimento di Chimica e Centro SMAArt, Università di Perugia, via Elce di Sotto 8, 06123 Perugia, Italy
a r t i c l e
i n f o
Article history: Received 3 August 2015 Received in revised form 19 October 2015 Accepted 19 October 2015 Available online 24 October 2015 Keywords: NMR stratigraphy NMR spectroscopy SEM-EDS Raman and IR spectroscopy Detached mural paintings Egyptian pictorial technique
a b s t r a c t A multi-analytical study was carried out on materials, techniques, and state of conservation of a set of detached Nubian mural paintings, originally belonging to the 10th century pictorial cycle of the church of Sonqi Tino, Sudan (70 km south of the Egypt–Sudan border). These paintings represent one of the very few examples of Christian Nubian pictorial cycles preserved from the wide defacement experienced by these artifacts with the building up of the Aswan Dam. Non-invasive NMR depth profiling was used to preliminary study the hydrogen rich layers of the detached mural paintings. By this technique, it was possible to perform a virtual coring with a reconstruction of the complex series of layers that characterize the investigated artifacts. Then, micro-sampling and crosssection examination by SEM-EDX, micro-Raman, micro-FTIR, and 27Al and 29Si MAS NMR, allowed us to shed light on materials and technology exploited by the Medieval Nubian painters, with the identification of pigments and composition of the original primer, obtaining also information on layering of materials used by conservators during the detachment of the paintings and their transfer to a new support. The combination of preliminary noninvasive investigation with a set of micro-invasive analytical techniques allowed us to set up a protocol for a satisfactory decoding of the multilayered complex systems that characterize the detached paintings. © 2015 Elsevier B.V. All rights reserved.
1. Introduction A multi-analytical study was carried out to investigate constitutive materials and state of conservation of a set of detached ancient Nubian mural paintings in permanent exhibition at Near East Museum of Sapienza University of Rome. The paintings were originally belonging to the 10th century pictorial cycle of the Nubian church of Sonqi Tino, Sudan (70 km south of the Egypt–Sudan border) and were detached during a UNESCO rescue campaign, just before the flooding of the area due to the rising of the Aswan Dam. The paintings were hastily removed and transferred to a new support using the so-called strappo (i.e. tear) procedure, a method of detachment that is practically never used in conservation, because considered dangerous, drastic, and irreversible [1,2]. Its application is only restricted to cases where the unique alternative is the complete loss of the artifact. The strappo consists of a number of successive steps: 1) areas with de-cohesions of the pictorial film are preliminary fixed; 2) the surface to be detached is prepared (facing), gluing a thin canvas to the portion ☆ Selected papers presented at TECHNART 2015 Conference, Catania (Italy), April 27–30, 2015. ⁎ Corresponding author. E-mail address: [email protected] (V. Di Tullio).
http://dx.doi.org/10.1016/j.microc.2015.10.022 0026-265X/© 2015 Elsevier B.V. All rights reserved.
of the wall painting to be removed; 3) the strappo is carried out, with a removal that involves the pictorial layer which remains adherent to the glued canvas, but also thin and irregular layers of the underlying preparation; 4) the back of the detached mural painting is leveled off and consolidated; 5) another canvas is glued to the back of the consolidated painting, which is thereafter fixed to a rigid support (backing); 6) finally, the canvas of the facing is removed. After the long procedure of the strappo, together with the pictorial film and preparation, all the inorganic and organic substances used during the transfer to a new support become parts of the artifact, which is finally composed by a complex multi-layered system of materials. This complexity renders any study of this type of detached paintings a real challenge. In a previous paper [3], we have exploited different NMR analytical techniques to evaluate the feasibility of a non-invasive or minimally invasive approach to understand the sequence of layers of the final detached system and their composition. First, through in situ non-invasive single-sided 1H NMR depth profiling, the whole stratigraphy has been investigated, leading to the identification of six distinguishable layers (L1–L6), then by 13C CPMAS NMR spectroscopy the organic materials mainly constituting these layers (up to the new support) have been investigated through a detailed analytical examination via selective sampling.
720
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
In this paper, we extended and integrated the previous study to better check the variability of the various layers in the whole examined paintings, characterized the execution technique of the paints, and finally obtained a full overview of the exhibited artifacts in their current status. For this purpose, further areas were non-invasively examined by single-sided 1H NMR, some new regions were sampled, and in order to extensively identify the original inorganic and organic materials constituting the paintings, an integrated multi-technique analytical approach was followed exploiting scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDS), micro-Raman, micro-FTIR, and 27Al and 29Si MAS NMR. In particular, to investigate the composition of the original pictorial film (L1) and the distribution of its elemental and molecular components, reflection-FTIR and EDS mappings on cross-sections were carried out. In spite of the complexity of the systems investigated, the obtained results finally allowed us to shed light on the technique and materials used by Christian Nubian painters in Sonqi Tino. The data, compared with those recently obtained on other detached paintings from the area of Faras, Old Dongola, and Baganarti [4], greatly contribute in defining material and techniques practiced by Medieval painters of Nubian art.
the composition of paintings detached from churches of the Nubian areas of Faras and Dongola. The paintings of the church of Sonqi Tino represent a chronological cornerstone in the history of Nubian painting. In fact, they are precisely dated due to the wishing inscription “Long life to King George” who reigned between 970 and 1003. They show relevant examples of the Christian iconography typical of this and other relevant Nubian archeological sites, such as those of Faras, Abdallah Nirqi, Abdel-Qadir, and Tamit. The representations include saints, kings, theological icons, and episodes from Old and New Testament. During the removal work, the mural paintings appeared in a very compromised state of conservation, since they were found to be fragmented in many areas, at risk of collapse, and partly detached from the wall or pushed out by plant roots. Because of the forthcoming flooding, they were detached from the masonry and brought elsewhere. Nowadays, most of the detached Sonqi Tino mural paintings are in exhibition at the Sudan National Museum in Khartoum, one piece is at the Vatican Museums, and some others are at the Near East Museum of Sapienza University of Rome, where the preliminary in situ measurements of this work were carried out.
2. Materials and methods
1 H NMR stratigraphies were collected at 13.62 MHz with a unilateral NMR instrument from Bruker Biospin interfaced with a single-sided sensor by RWTH Aachen University, Aachen, Germany [5]. Experiments were carried out by repositioning the single-sided sensor in steps of 50 μm to cover the desired spatial range, from the outermost surface of the detached mural painting to a depth of 0.5 cm with a resolution of 57 μm, the magnetic field gradient was 14.4 T/m, 100 experimental points corresponding to 100 different depths were collected to obtain the profile; the number of scans was 256, the total scan time was 3.5 hours, the recycle delay was optimized at 0.5 s, the π/2 pulse width was 11 μs. The spatial resolution was obtained from the pulse width to the intensity of the magnetic field gradient ratio. The area of measurement was 2.5 × 2.5 cm2.
2.1. The case study In 1966, UNESCO promoted the rescue of several Nubian archeological sites which would have soon been flooded by the river Nile due to the Aswan Dam raising. Within this framework, an archeological mission was organized to recover artifacts from Sonqi Tino, a village on the west bank of Nile river, at the south of the Egypt–Sudan present border. Four excavation campaigns took place between 1967 and 1970 with the aim of recovering the ancient Christian Nubian vestiges from the whole area, which included the church, the adjacent burying ground, and the Diff, a fortress built in an area at the south of Sonqi. Fig. 1a shows a detail of the cob made church, photographed during the excavation campaign. According to the report of the archeological mission, the church was found to be partially sanded up, the masonry covered with a local mud for leveling off the surface and painted over a white thin preparation layer of lime. Apart the archeological reports based exclusively on visual observation, no written sources are available on the composition and technology of the Nubian wall paintings. To this lack of knowledge strongly contributed the wide, if not total, defacement experienced by most of these pictorial cycles during the time. Only very recently, a study parallel to this one has been carried out on
2.2. Unilateral NMR
2.3. Solid-state NMR Samples were packed in 4-mm zirconia rotors with an available volume reduced to 12 μl, and sealed with Kel-F caps. The spin rate was 12 kHz. Solid-state 13C CPMAS NMR spectra were measured at 100.63 MHz on a Bruker Avance III spectrometer. The contact time for the crosspolarization was 1.5 ms, the recycle delay was 3 s, and the 1H π/2 pulse width was 3.5 μs. Spectra were collected with a time domain of 1024 data points, and Fourier transformed with a size of 2048 data
b) a)
Fig. 1. a) A detail of Sonqi Tino church photographed during the excavation campaign. b) A mural painting found in Sonqi Tino Church. We acknowledge Prof. L.Sist and M.Necci for the pictures.
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
721
points. The ppm scale was externally referenced to tetramethylsilane (TMS). 27 Al MAS NMR spectra were recorded at 104.26 MHz. The π/2 pulse width was 1.5 μs, and the recycle delay was 3 s. The ppm scale was externally referenced to a 1 M aluminum nitrate aqueous solution. 29 Si MAS NMR spectra were recorded at 79.49 MHz. The π/2 pulse width was 4 μs, 512 data points were collected with a recycle delay of 60 s. The ppm scale was externally referenced to TMS. 27Al and 29Si MAS spectra were collected with a time domain of 512 data points, and Fourier transformed with a size of 1024 data points. 2.4. Scanning electron microscopy Morphological and stratigraphic investigations were carried out by using optical microscopy (OM) and scanning electron microscopy (SEM) coupled with energy dispersive X-ray spectroscopy (EDS). OM images were taken with a Leica DMR microscope equipped with a Leica DC300 digital camera. SEM analysis by secondary electrons (SE) and EDS analysis were carried out with a FEI Quanta 200 FEG SEM (Eindhoven, The Netherlands) equipped with an Oxford Inca Energy System 250 and an Inca-X-act LN2free analytical silicon drift detector. Sections were then mounted onto SEM stubs by means of carbon adhesive disks and analyzed at 20 kV acceleration voltage in low vacuum mode (PH2O = 0.6 ÷ 0.9 torr). 2.5. Infrared spectroscopy Micro-FTIR investigation was carried out with a Jasco FTIR 4100 equipped with an IMV-400 optical microscope. In particular, the conventional spectra collected on the surface and on the back of samples were obtained using a single nitrogen-cooled mercury cadmium telluride (MCT) detector. Reflection measurements were recorded with a 16× Cassegrain objective collecting 800 scans with a spectral resolution of 4 cm− 1 on a spot of 625 × 625 μm2. The mapping measurements were carried out by a nitrogen-cooled linear array detector composed of 16 MCT detector elements. An x–y computer-controlled stage enabled scanning of the sample line by line according to the selected area. Reflection measurements were performed with a 16× Cassegrain objective collecting 10,000 scans with a spectral resolution of 8 cm−1 and a lateral resolution of 12.5 μm. 2.6. Micro-Raman spectroscopy Micro-Raman spectra were recorded by a Jasco NRS-3100 spectrometer coupled to an optical Olympus microscope (100× objective) and equipped with an argon laser source at 514 nm. The laser power at the sample was 1–2 mW. The instrument was equipped with a 1200 lines/mm grating, providing approximately a resolution of 3–5 cm−1, and a CCD detector Peltier cooled to − 50 °C. Spectra were acquired with 10 s and 5 scans. 3. Results 3.1. 1H NMR stratigraphy Following the 1H NMR method already described in our previous paper [3], stratigraphic determinations were carried out in situ in a fully non-invasive way. The method encodes the amplitude of the 1H NMR signal as a function of the depth scanned [6–9] allowing us to discriminate and visualize different layers made of different hydrogencontaining materials. The trend of the stratigraphy was found similar in all the scanned regions of the investigated paintings. As example of the numerous results obtained on different areas, two cases are shown in Fig. 2, chosen as representative of analogy and variability of the recorded sequences of layers. The stratigraphy showed generally a flat region (L1) with a
Fig. 2. 1H NMR stratigraphy of a wall painting belonging to Sonqi Tino pictorial cycle and exhibited at the Near East Museum of Sapienza University of Rome. The labeling of each layer (L1–L6) is reported. a)1H NMR stratigraphy where thin L1 and L4 layers are observed. b)1H NMR stratigraphy where thick L1 and L4 layers are observed.
very weak NMR signal, followed by a broad shoulder (L2) and a sharp peak (L3). After this maximum, a region (L4) characterized by a net decrease of the NMR signal was generally recorded, while at a deeper level, another sharp peak (L5) appeared, followed by a region (L6) with a very weak NMR signal. While this layers' sequence was substantially confirmed everywhere, the thickness of the various layers was found to be largely inhomogeneous. In particular, L1 and L4 were those showing the most variable thickness. Layer L1 was made of the inorganic materials constituting the pictorial layer together with the underlying preparation (primer). The large thickness' variability of L1, as will be demonstrated by the data of the following paragraph, is related to the different quantity of original primer removed by the strappo (Fig. 2). As example, in 2a, the thickness of layer L1 is about 100 μm whereas in 2b, it is about 500 μm, indicating that in the former case only the paint layer was detached from the original site while in the latter the detached original material is composed by the paint and the primer (or at least a fraction of it). On the opposite, layer L2 which is the first layer of material used by conservators for the adaptation of the removed painting to a new support [3], is thicker in 2a than 2b, indicating that a higher amount of consolidating material had been necessary to level off and consolidate the regions where the strappo removed a thinner layer of material. The thickness of layer L4 gives information on the level of adhesion between layers L3 and L5. Basically, the thick layer L4 observed in Fig. 2b is characterized by the absence of NMR signal (no organic materials) indicating a detachment at the interface between L3 and L5. On the opposite, the thin layer L4 in 2a, where an NMR signal is present, indicates that L3 and L5 are still joined together.
722
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
3.2. Pigments and inorganic materials characterized by SEM-EDS, micro-Raman and micro-IR spectroscopy In order to investigate the execution technique of the detached wall paintings, a few micro-samples were collected from red (1P), black (2P), green (3P), and yellow (4P1 and 4P2) painted areas and analyzed by SEM-EDS, micro-Raman, and micro-IR spectroscopy. Samples were examined directly and in cross-section, with the aim to identify pigments (and possibly binder), the original primer, and the backing. Fig. 3a shows the map of elements of sample 1P obtained by SEMEDS. The pigment layer was constituted mostly by aluminum, silicon, and iron. Iron-rich particles were found from the outermost surface to a depth of about 20 μm, although some aggregates containing iron were also found at higher depth. Specific attention was focused on the characterization of the primer whose thickness in sample 1P was found to be about 200 μm by SEM-EDS analysis. The elemental imaging shows that S, Si, Al, and Ca are distributed in the whole primer layer, see Fig. 3a. Below this layer, at a deeper level and beyond the primer, a white thick layer very rich in calcium (96 wt%) was observed. This latter layer is composed by the calcium carbonate used to level off and consolidate the detached original paint. The micro-Raman measurements performed on the painting layer showed the characteristic peaks of hematite, see Fig. 4a, typical of a red ochre. The bands at 298, 410, and 610 cm−1 are ascribed to the hematite Eg symmetry vibration modes, while the band at 660 cm−1 is attributed to the relaxation of the selection rules induced by disorder effects in the crystal structures of hematite for the possible substitution of Fe3+ by Al3+ cations [10]. The IR spectrum of the sample acquired in reflection mode from the surface of the painting layer, without any treatment, is shown in Fig. 4b. The spectrum revealed the presence of aluminum silicates, mostly kaolin and quartz [10], calcium carbonate [11], and calcium sulfate with different levels of hydration (gypsum CaSO4•2H2O, bassanite CaSO4•0.5H2O, anhydrite CaSO4) [12]. A proteinaceous substance was also clearly revealed on the surface of the painting layer. This latter material might be ascribed either to the original binder or, most probably, a residue of the glue used in the first steps of the strappo. The FTIR spectrum acquired in reflection mode from the back side of the sample is shown in Fig. 4c. In this case, calcium carbonate and an acrylic resin [13], used as consolidant after the strappo, were identified. The SEM-EDS analysis of sample 2P (sample with a black pigment), revealed the presence of a layer about 150 μm thick, constituted mainly by Al and Si, whereas Ca was the main element of the inner part of the
sample, see Fig. 3b. The surface of the sample was quantitatively analyzed by EDS without embedding it in resin. Apart from C and O with relative amounts about 26.5 and 44.4 wt% respectively, Si (15.7 wt%) and Al (7.6 wt%) were the main constituents of the surface. Ca (1.8 wt%) and Fe (1.4 wt%) were also present on the external surface, together with a low amount (b1 wt%) of potassium, sulfur, and sodium. These data suggested the presence of carbon black, as effectively found by micro-Raman spectroscopy. In fact, the Raman spectrum of Fig. 3d shows two bands at 1585 and 1350 cm−1 known as G (graphite) and D (disorder) bands due respectively to E2g and A1g modes [14]. The SEM-EDS analysis of sample 3P (sample with a green pigment) revealed the presence of an external layer about 20 μm thick, constituted by Cu, Cl, and Mg, see Fig. 3c. Again Ca was the main element of the inner part of the sample along with a very small amount of Si, S, Al, and Fe. The Raman spectrum showed bands at 976, 914, and 830 cm−1 ascribable to hydroxyl deformation bands, and bands at 511 and 150 cm−1 due to copper oxygen vibrations typical of atacamite, see Fig. 4e. Therefore, Raman spectra along with SEM analysis of the external layer clearly indicate the use of atacamite as green pigment. The cross-section of sample 4P1 (first sample of the two with a yellow pigment) was investigated by SEM-EDS, see Fig. 3d. This sample was obtained from the region where the stratigraphy reported in Fig. 2a was collected. A first layer about 40 μm thick was found to be mainly constituted of Al, Si, and Fe. A second layer about 250 μm thick mostly constituted of Ca was observed. A very thin layer (b5 μm) of Ti was present at the interface between the first and the second layer. The Raman spectrum did not allow for the identification of the pigment because of a very intense florescence which covered any other signal. However, from the elemental composition obtained by SEM-EDS, it was reasonable to hypothesize that the pigment employed was a yellow ochre. The presence of the thin Ti layer just below the paint might be ascribed to a polymorphous forms of titanium dioxide minerals which can be found in Egypt usually associated with other minerals like calcite (CaCO3), quartz (SiO2), or hematite [4] or to a material related to the treatment of the strappo. The cross-section of sample 4P2 (second sample from a yellow area) was first investigated by SEM-EDS. This sample was obtained from the region where the stratigraphy reported in Fig. 2b was collected. The elemental composition of the first layer obtained by SEM-EDS was very similar to that found in sample 4P1. The elemental composition of the second layer was found to be dependent on the depth. In fact below the pigment layer a region rich in Ca (24%) and S (19%) was observed, and at a greater depth, a layer very rich in Si (61%), Al (14%), Fe (6%), whereas the amount of Ca (7%), and S (3%) decreased.
Fig. 3. Map of elements obtained by SEM-EDS of a) sample 1P (red pigment), b) 2P (black pigment), c) 3P (green pigment), and d) 4P1 (yellow pigment). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
723
Fig. 4. a) Micro-Raman spectrum of a red pigment (1P) showing the characteristic peaks of hematite. b) Infrared spectrum in reflection mode collected on the painting layer of the sample. c) Infrared spectrum in reflection mode collected on the back side of the sample. d) Raman spectrum of a black pigment (2P) showing the peaks of carbon black. e) Raman spectrum of a green pigment (3P) showing the peaks of atacamite.
It is worth to note that samples 4P1 and 4P2 showed a major difference in the composition of the region below the pigment layer. In the former case, a very low amount of sulfur and high amount of calcium was found, whereas in the latter case, a high amount of both S and Ca, indicating the presence of calcium sulfate, was detected. These results are also in a very good agreement with the NMR depth profiles collected on the two regions where samples 4P1 and 4P2 were obtained, see Fig. 2. In fact, the stratigraphy of the region of sample 4P1 displayed a thin pictorial layer whereas the stratigraphy of the region of sample 4P2 displayed a thick layer indicating that in this case the strappo removed both pigments and primer. In order to better clarify the molecular composition and the distribution of the different components along the original primer and the backing layers, a reflection infrared spectroscopy mapping on the cross-section [15] was carried out on sample 4P2. Fig. 5a and b show the optical images of the investigated cross-section, whereas Fig. 5c–f show the chemical maps obtained by using the integration of specific marker band of calcium sulfates, calcium carbonate, kaolin, and a synthetic resin. To obtain a well-defined and unambiguous color map distribution, bands neither affected by distortion due to reflection nor by overlap with other bands, were selected. It is possible to observe that the original primer layer was constituted
mainly by kaolin, calcium sulfates, and calcium carbonate distributed in different amounts on the entire primer layer. Calcium sulfates and kaolin signals completely disappeared in the region of the backing layer. The acrylic resin was found to be distributed along the entire sample; however, its amount decreased going from the backing toward the original primer. A high concentration of calcium carbonate was found at deeper depths beyond the primer, indicating the presence of a thick carbonate layer mixed with an acrylic resin used to level off the back of the detached painting during the strappo procedure, see Fig. 5d and f. 3.3. Inorganic materials of the primer investigated by solid-state NMR 27
Al and 29Si MAS NMR spectroscopy [16] was applied to characterize the alumino-silicate inorganic compounds present in the more superficial layers of original and non-original areas of the detached wall paintings. The 27Al MAS spectra of (a) sample 1A, non-original material obtained from a rim of a lacuna, (b) sample 2A, taken off from an original unpainted area, and (c) sample 3A, a sample of local clay, are all typical of a raw clay [17], with the resonance of Al in an octahedral environment (AlVI) apparently centered at 2.3 ppm, and the resonance of Al in
Fig. 5. a) and b) Optical images (10× and 16×) of the cross-section of the sample 4P2 where the blue grid indicates the mapped area by reflection infrared spectroscopy (single point measurement of dimensions 12.5 × 12.5 μm2). c) Color map of the calcium sulfates distribution integrating the band at 2080–2320 cm−1. d) Color map of calcite distribution integrating the peak at 2440–2630 cm−1. e) Color map of kaolin distribution integrating the peak at 3615 cm−1. f) Color map of synthetic resin distribution integrating the peak at 1740 cm−1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
724
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
a tetrahedral environment (AlIV) apparently centered at 56.3 ppm, see Fig. 6a, b, and c. The ratio between the intensity of the resonance of AlVI and that of AIIV was found to be definitely greater in sample 2A than in 1A, being 0.90 and 0.54, respectively. Therefore, these two samples differentiate in the AlIV content with respect to AlVI. On the other hand, the intensity ratio measured in the spectrum of the superficial layers of the original region (sample 2A) matches that of the sample of the local clay, being 0.90 and 0.93, respectively. Fig. 6c and d show the 29Si MAS spectra of samples 1A and 2A. In the spectrum of 1A, a broad unresolved resonance is observed, whereas sample 2A shows a better resolved spectrum where resonances of kaolin and quartz may be easily assigned at −91.4 and −107 ppm, respectively, by comparison with the spectra of standard compounds, see Fig. 6f (kaolin), and g (quartz). 4. Discussion As in the preceding work, 1H NMR depth profiling revealed the presence of six successive layers (L1–L6). L1 showed a variable thickness due to the different thickness of original material removed by the strappo, in some areas only the paint layer, with a small quantity of primer was removed, in other areas paint and primer were removed together. Micro-Raman and SEM-EDS analyses carried out on the surface layer of the detached paintings allowed the identification of original pigments to be carried out. Pigments were found to be red and yellow ochre, carbon black, calcium carbonate (white), and atacamite (green). These data are very similar to those obtained in an analogous parallel study recently carried out on the detached paintings of the Christian
Fig. 6. 27Al MAS NMR spectra of sample 1A obtained from the rim of a lacuna in a nonoriginal area, s/n = 100 (a), sample 2A obtained from an unpainted original region, s/n = 50 (b), and a sample of local clay, s/n = 100 (c). 29Si MAS NMR spectra of sample 1A (d), sample 2A (e), kaolin (f), and quartz (g).
Nubian churches of Faras and of the two close sites of Old Dongola and Baganarti [4]. These paintings were removed by a group of Polish archeologists during the same UNESCO rescue campaign of Nubian culture in 1967–70 and are currently in exhibition at the National Museum in Warsaw. The church of Faras was subject along the centuries to repeated rebuildings and its paintings are dated from the 7th to 14th century. The only difference in the pigments used in Sonqi Tino and those found in Faras is due to massicot (PbO), that is present, together with some ochre, in yellow areas of the paintings of Faras, but not in Sonqi Tino. Besides the analogy in the use of pigments, probably related to the availability of local minerals, strong analogies are also found in the materials which compose the primer below the paint layer. In our case, SEM-EDS revealed the presence of an aluminum-silicate matrix and calcium, while micro-FTIR analysis revealed the presence of calcium carbonate, kaolin, quartz, and calcium sulfate with different levels of hydration (gypsum, bassanite, and anhydrite). The presence of an aluminum-silicate matrix, possibly a clay, and kaolin and quartz was also confirmed by 27Al and 29Si MAS NMR analysis. 27Al MAS NMR also differentiated between an unpainted original plaster and the plaster used to fill up missing regions after the strappo. Furthermore, the 27Al MAS spectrum of the unpainted original plaster matched that of a local clay. FTIR mapping allowed the distribution of components in the original primer and backing layer to be visualized. Specifically, kaolin and calcium carbonate and sulfate were found to be distributed in the whole primer layer, whereas high concentration of calcium carbonate and a synthetic resin (Paraloid) were found mixed in the backing layer. This latter result is in agreement with the previous result obtained by 13C CPMAS NMR spectroscopy that detected the presence of an MA/EMA copolymer (Paraloid) in layer L2 [3]. It is worth noting that in all the examined paintings from the areas of Faras, a significant presence of titanium was recorded in a direct surface examination by XRF [4]. In the present case, titanium was significantly recorded only in a single case, showing the feature of a very thin layer of material at the interface between the first layer L1 (painting layer and primer) and the layer of calcium carbonate used to level off the detached materials (L2). Its location and layering suggest that the presence of Ti might be more related to the treatment during the strappo than to the minerals used in the original paintings. However, this topic deserves further deepening. Unfortunately no unambiguous identification of the binder was possible. A proteinaceous material was recorded in the paint layer, but this material could be a residue of glue used in the strappo. Limewash or kaolin could be also a possible binder, but no final conclusion can be achieved. Fig. 7 shows a sketch summarizing all layers constituting the detached mural paintings in relation to their 1H NMR depth profile. The original painting layer (L1) whose thickness varies between 100 and 500 μm, is made of pigments and primer. The original primer was only present in those regions where the strappo removed a sufficient thickness of material. All other layers (L2–L6) are due to the materials used in the strappo and adaptation of the painting to the new support. Paraloid embedded into calcite (L2) was used to consolidate and level the back of the paintings off. An adhesive (MA/EMA/PHEMA) was used to size a new canvas to the back of the detached painting (L3). The same adhesive was used to fix the back of the system paintingcanvas to plasticized PVC that constituted the new supporting panel (L5). L6 is the supporting panel (PVC). A remarkable note is that layer L4 is indicating a well evident loss of adhesion, in some areas, between L3 and L5. 5. Conclusions A multi-technique analytical study was carried out on the execution practices of Medieval painters working on the wall of the Christian
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
725
6. Ackowledgments The present research was supported by CNR (Italy) within the SM@ RTINFRA-SSHCH project, “Smart Integrated Infrastructures for Data Social Sciences, Humanities and Cultural Heritage Ecosystem” (2012–2015). We acknowledge Prof. L. Sist, B. Provinciali, A. Marinelli, and M. Necci for the useful discussion, and for their kind collaboration during measurements in situ and sampling.
References
Fig. 7. 1H NMR stratigraphy and sketch of all layers constituting the detached mural paintings. L1: original pigments and primer (kaolin, gypsum, and clays); L2: calcite and Paraloid; L3) MA/EMA/PEHMA resin and canvas; L4) interface between adhesive and support; L5) MA/EMA/PEHMA resin and PVC; L6) PVC.
Nubian church of Sonqi Tino, dated back to the 10th century. The study integrates a previous work dedicated to the characterization of the organic materials of the complex multilayered artifact obtained after the strappo of the paintings from the walls of the church. The investigation of the paint and primer lead to the identification of pigments as hematite (red), carbon (black), calcium carbonate (white), ochre (yellow), and atacamite (green). These pigments, when compared with those found in a recent parallel study on paintings detached from Nubian churches of the areas of Faras, Old Dongola, and Baganarti, confirm that these are the typical materials used by the painters of Nubian churches during Middle Ages (9th–12th century), with a recurrent use of the same materials probably related to the availability of local minerals. The use of local materials by Nubian painters was also confirmed by 27Al MAS NMR measurements on the alumino-silicate materials of the unpainted original primer's plaster that matched that of a local clay. The work demonstrates how an integrated multi-analytical approach with exploitation of adequate noninvasive and minimally invasive techniques, allows satisfactory results to be obtained even on complex systems as the detached mural paintings of Sonqi Tino. Crucial is the exploitation of the non-invasive 1H NMR depth profiling integrated by elemental and molecular spectroscopy mappings on crosssections and with 27Al and 29Si MAS NMR spectroscopy. The analytical protocol here proposed allows satisfactory answers to be given to questions regarding the practices of Nubian Medieval artists, a topic on which only very few technical and scientific information is available to date and on the current state of conservation of the detached paintings giving relevant information to museum's conservators.
[1] F. Piqué, L. Tintori, E. Borsook, Wall painting conservation in Tuscany before the Florentine flood of 1966, Les anciennes en peiture murale: journées d’études de la SFIIC, Section française de l’IIC, Dijon, France 1993, pp. 91–105. [2] L. Mora, P. Philippot, P. Mora, Conservation of Wall Paintings, ICROOM, Rome, Italy, 1984 265–308. [3] V. Di Tullio, D. Capitani, F. Presciutti, G. Gentile, B.G. Brunetti, N. Proietti, Noninvasive NMR stratigraphy of a multi-layered artefact: an ancient detached mural painting, Anal. Bioanal. Chem. 405 (2013) 8669–8675. [4] O. Syta, K. Rozum, M. Choińska, D. Zielińska, G.Z. Żukowska, A. Kijowska, B. Wagner, Analytical procedure for characterization of medieval wall-paintings by X-ray fluorescence spectrometry, laser ablation inductively coupled plasma mass spectrometry and Raman spectroscopy, Spectrochim. Acta B At. Spectrosc. 101 (2014) 140–148. [5] J. Perlo, F. Casanova, B. Blümich, Profiles with microscopic resolution by single-sided NMR, J. Magn. Reson. 176 (2005) 64–70. [6] F. Presciutti, J. Perlo, F. Casanova, S. Glöggler, C. Miliani, B. Blümich, B.G. Brunetti, A. Sgamellotti, Non invasive nuclear magnetic resonance profiling of painting layers, Appl. Phys. Lett. 93 (2008) (033505-1-033505-3). [7] B. Blümich, F. Casanova, J. Perlo, F. Presciutti, C. Anselmi, B. Doherty, Noninvasive testing of art and cultural heritage by mobile NMR, Acc. Chem. Res. 43 (2010) 761–770. [8] V. Di Tullio, N. Proietti, D. Capitani, I. Nicolini, A.M. Mecchi, NMR depth profiles as a non-invasive analytical tool to probe the penetration depth of hydrophobic treatments and inhomogeneities in treated porous stones, Anal. Bioanal. Chem. 400 (2011) 3151–3164. [9] G.R. Fife, B. Stabik, A.E. Kelley, J.N. King, B. Blumich, R. Hoppenbrouwers, T. Meldrum, Characterization of aging and solvent treatments of painted surface using single sided NMR, Magn. Reson. Chem. 53 (2014) 58–63. [10] D. Bersani, P.P. Lottici, A. Montenero, Micro-Raman investigation of iron oxide films and powders produced by sol–gel syntheses, J. Raman Spectrosc. 30 (1999) 355–360. [11] C. Ricci, C. Miliani, B.G. Brunetti, A. Sgamellotti, Non-invasive identification of surface materials on marble artifacts with fiber optic mid-FTIR reflectance spectroscopy, Talanta 69 (2006) 1221–1226. [12] F. Rosi, A. Daveri, B. Doherty, S. Nazzareni, B.G. Brunetti, A. Sgamellotti, C. Miliani, On the use of second-order modes for the analysis of the CaSO4-H2O system by reflectance FTIR, Appl. Spectrosc. 64 (2010) 956–963. [13] F. Rosi, C. Miliani, C. Clementi, K. Kahrim, F. Presciutti, M. Vagnini, V. Manuali, A. Daveri, L. Cartechini, B.G. Brunetti, A. Sgamellotti, An integrated spectroscopic approach for the non invasive study of modern art materials and techniques, Appl. Phys. A 100 (2010) 613–624. [14] F. Tuinstra, J.L. Koenig, Raman spectrum of graphite, J. Chem. Phys. 53 (1970) 1126–1130. [15] F. Rosi, A. Federici, B.G. Brunetti, A. Sgamellotti, S. Clementi, C. Miliani, Multivariate chemical mapping of pigments and binders in easel painting cross-sections by micro IR reflection spectroscopy, Anal. Bioanal. Chem. 399 (2011) 3133–3145. [16] M.E. K. J. D MacKenzie, Multinuclear Solid-State NMR of Inorganic Materials, Pergamon Materials Series 6. , Elsevier Science Ltd, Oxford, UK, 2002. [17] F. Presciutti, D. Capitani, A. Sgamellotti, B.G. Brunetti, F. Costantino, S. Viel, A.L. Segre, Electron paramagnetic resonance, scanning electron microscopy with energy dispersion X-ray spectrometry, X-ray powder diffraction, and NMR characterization of iron-rich fired clays, J. Phys. Chem. B 109 (2005) 22147–22158.
Contents lists available at ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
A multi-analytical study of ancient Nubian detached mural paintings☆ Noemi Proietti a, Valeria Di Tullio a,⁎, Federica Presciutti b, Gennaro Gentile c, Brunetto Giovanni Brunetti d, Donatella Capitani a a Laboratorio di Risonanza Magnetica “Annalaura Segre”, Istituto di Metodologie Chimiche, Consiglio Nazionale delle Ricerche (IMC-CNR), Area della Ricerca di Roma 1, Via Salaria km 29,300, 00015 Monterotondo (Roma), Italy b Dipartimento di Scienze Farmaceutiche, Università di Perugia, via Fabbretti 48, 06123 Perugia, Italy c Istituto per i Polimeri, Compositi e Biomateriali, Consiglio Nazionale delle Ricerche (IPCB-CNR), Via Campi Flegrei, 34, 80078 Pozzuoli (Napoli), Italy d Dipartimento di Chimica e Centro SMAArt, Università di Perugia, via Elce di Sotto 8, 06123 Perugia, Italy
a r t i c l e
i n f o
Article history: Received 3 August 2015 Received in revised form 19 October 2015 Accepted 19 October 2015 Available online 24 October 2015 Keywords: NMR stratigraphy NMR spectroscopy SEM-EDS Raman and IR spectroscopy Detached mural paintings Egyptian pictorial technique
a b s t r a c t A multi-analytical study was carried out on materials, techniques, and state of conservation of a set of detached Nubian mural paintings, originally belonging to the 10th century pictorial cycle of the church of Sonqi Tino, Sudan (70 km south of the Egypt–Sudan border). These paintings represent one of the very few examples of Christian Nubian pictorial cycles preserved from the wide defacement experienced by these artifacts with the building up of the Aswan Dam. Non-invasive NMR depth profiling was used to preliminary study the hydrogen rich layers of the detached mural paintings. By this technique, it was possible to perform a virtual coring with a reconstruction of the complex series of layers that characterize the investigated artifacts. Then, micro-sampling and crosssection examination by SEM-EDX, micro-Raman, micro-FTIR, and 27Al and 29Si MAS NMR, allowed us to shed light on materials and technology exploited by the Medieval Nubian painters, with the identification of pigments and composition of the original primer, obtaining also information on layering of materials used by conservators during the detachment of the paintings and their transfer to a new support. The combination of preliminary noninvasive investigation with a set of micro-invasive analytical techniques allowed us to set up a protocol for a satisfactory decoding of the multilayered complex systems that characterize the detached paintings. © 2015 Elsevier B.V. All rights reserved.
1. Introduction A multi-analytical study was carried out to investigate constitutive materials and state of conservation of a set of detached ancient Nubian mural paintings in permanent exhibition at Near East Museum of Sapienza University of Rome. The paintings were originally belonging to the 10th century pictorial cycle of the Nubian church of Sonqi Tino, Sudan (70 km south of the Egypt–Sudan border) and were detached during a UNESCO rescue campaign, just before the flooding of the area due to the rising of the Aswan Dam. The paintings were hastily removed and transferred to a new support using the so-called strappo (i.e. tear) procedure, a method of detachment that is practically never used in conservation, because considered dangerous, drastic, and irreversible [1,2]. Its application is only restricted to cases where the unique alternative is the complete loss of the artifact. The strappo consists of a number of successive steps: 1) areas with de-cohesions of the pictorial film are preliminary fixed; 2) the surface to be detached is prepared (facing), gluing a thin canvas to the portion ☆ Selected papers presented at TECHNART 2015 Conference, Catania (Italy), April 27–30, 2015. ⁎ Corresponding author. E-mail address: [email protected] (V. Di Tullio).
http://dx.doi.org/10.1016/j.microc.2015.10.022 0026-265X/© 2015 Elsevier B.V. All rights reserved.
of the wall painting to be removed; 3) the strappo is carried out, with a removal that involves the pictorial layer which remains adherent to the glued canvas, but also thin and irregular layers of the underlying preparation; 4) the back of the detached mural painting is leveled off and consolidated; 5) another canvas is glued to the back of the consolidated painting, which is thereafter fixed to a rigid support (backing); 6) finally, the canvas of the facing is removed. After the long procedure of the strappo, together with the pictorial film and preparation, all the inorganic and organic substances used during the transfer to a new support become parts of the artifact, which is finally composed by a complex multi-layered system of materials. This complexity renders any study of this type of detached paintings a real challenge. In a previous paper [3], we have exploited different NMR analytical techniques to evaluate the feasibility of a non-invasive or minimally invasive approach to understand the sequence of layers of the final detached system and their composition. First, through in situ non-invasive single-sided 1H NMR depth profiling, the whole stratigraphy has been investigated, leading to the identification of six distinguishable layers (L1–L6), then by 13C CPMAS NMR spectroscopy the organic materials mainly constituting these layers (up to the new support) have been investigated through a detailed analytical examination via selective sampling.
720
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
In this paper, we extended and integrated the previous study to better check the variability of the various layers in the whole examined paintings, characterized the execution technique of the paints, and finally obtained a full overview of the exhibited artifacts in their current status. For this purpose, further areas were non-invasively examined by single-sided 1H NMR, some new regions were sampled, and in order to extensively identify the original inorganic and organic materials constituting the paintings, an integrated multi-technique analytical approach was followed exploiting scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDS), micro-Raman, micro-FTIR, and 27Al and 29Si MAS NMR. In particular, to investigate the composition of the original pictorial film (L1) and the distribution of its elemental and molecular components, reflection-FTIR and EDS mappings on cross-sections were carried out. In spite of the complexity of the systems investigated, the obtained results finally allowed us to shed light on the technique and materials used by Christian Nubian painters in Sonqi Tino. The data, compared with those recently obtained on other detached paintings from the area of Faras, Old Dongola, and Baganarti [4], greatly contribute in defining material and techniques practiced by Medieval painters of Nubian art.
the composition of paintings detached from churches of the Nubian areas of Faras and Dongola. The paintings of the church of Sonqi Tino represent a chronological cornerstone in the history of Nubian painting. In fact, they are precisely dated due to the wishing inscription “Long life to King George” who reigned between 970 and 1003. They show relevant examples of the Christian iconography typical of this and other relevant Nubian archeological sites, such as those of Faras, Abdallah Nirqi, Abdel-Qadir, and Tamit. The representations include saints, kings, theological icons, and episodes from Old and New Testament. During the removal work, the mural paintings appeared in a very compromised state of conservation, since they were found to be fragmented in many areas, at risk of collapse, and partly detached from the wall or pushed out by plant roots. Because of the forthcoming flooding, they were detached from the masonry and brought elsewhere. Nowadays, most of the detached Sonqi Tino mural paintings are in exhibition at the Sudan National Museum in Khartoum, one piece is at the Vatican Museums, and some others are at the Near East Museum of Sapienza University of Rome, where the preliminary in situ measurements of this work were carried out.
2. Materials and methods
1 H NMR stratigraphies were collected at 13.62 MHz with a unilateral NMR instrument from Bruker Biospin interfaced with a single-sided sensor by RWTH Aachen University, Aachen, Germany [5]. Experiments were carried out by repositioning the single-sided sensor in steps of 50 μm to cover the desired spatial range, from the outermost surface of the detached mural painting to a depth of 0.5 cm with a resolution of 57 μm, the magnetic field gradient was 14.4 T/m, 100 experimental points corresponding to 100 different depths were collected to obtain the profile; the number of scans was 256, the total scan time was 3.5 hours, the recycle delay was optimized at 0.5 s, the π/2 pulse width was 11 μs. The spatial resolution was obtained from the pulse width to the intensity of the magnetic field gradient ratio. The area of measurement was 2.5 × 2.5 cm2.
2.1. The case study In 1966, UNESCO promoted the rescue of several Nubian archeological sites which would have soon been flooded by the river Nile due to the Aswan Dam raising. Within this framework, an archeological mission was organized to recover artifacts from Sonqi Tino, a village on the west bank of Nile river, at the south of the Egypt–Sudan present border. Four excavation campaigns took place between 1967 and 1970 with the aim of recovering the ancient Christian Nubian vestiges from the whole area, which included the church, the adjacent burying ground, and the Diff, a fortress built in an area at the south of Sonqi. Fig. 1a shows a detail of the cob made church, photographed during the excavation campaign. According to the report of the archeological mission, the church was found to be partially sanded up, the masonry covered with a local mud for leveling off the surface and painted over a white thin preparation layer of lime. Apart the archeological reports based exclusively on visual observation, no written sources are available on the composition and technology of the Nubian wall paintings. To this lack of knowledge strongly contributed the wide, if not total, defacement experienced by most of these pictorial cycles during the time. Only very recently, a study parallel to this one has been carried out on
2.2. Unilateral NMR
2.3. Solid-state NMR Samples were packed in 4-mm zirconia rotors with an available volume reduced to 12 μl, and sealed with Kel-F caps. The spin rate was 12 kHz. Solid-state 13C CPMAS NMR spectra were measured at 100.63 MHz on a Bruker Avance III spectrometer. The contact time for the crosspolarization was 1.5 ms, the recycle delay was 3 s, and the 1H π/2 pulse width was 3.5 μs. Spectra were collected with a time domain of 1024 data points, and Fourier transformed with a size of 2048 data
b) a)
Fig. 1. a) A detail of Sonqi Tino church photographed during the excavation campaign. b) A mural painting found in Sonqi Tino Church. We acknowledge Prof. L.Sist and M.Necci for the pictures.
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
721
points. The ppm scale was externally referenced to tetramethylsilane (TMS). 27 Al MAS NMR spectra were recorded at 104.26 MHz. The π/2 pulse width was 1.5 μs, and the recycle delay was 3 s. The ppm scale was externally referenced to a 1 M aluminum nitrate aqueous solution. 29 Si MAS NMR spectra were recorded at 79.49 MHz. The π/2 pulse width was 4 μs, 512 data points were collected with a recycle delay of 60 s. The ppm scale was externally referenced to TMS. 27Al and 29Si MAS spectra were collected with a time domain of 512 data points, and Fourier transformed with a size of 1024 data points. 2.4. Scanning electron microscopy Morphological and stratigraphic investigations were carried out by using optical microscopy (OM) and scanning electron microscopy (SEM) coupled with energy dispersive X-ray spectroscopy (EDS). OM images were taken with a Leica DMR microscope equipped with a Leica DC300 digital camera. SEM analysis by secondary electrons (SE) and EDS analysis were carried out with a FEI Quanta 200 FEG SEM (Eindhoven, The Netherlands) equipped with an Oxford Inca Energy System 250 and an Inca-X-act LN2free analytical silicon drift detector. Sections were then mounted onto SEM stubs by means of carbon adhesive disks and analyzed at 20 kV acceleration voltage in low vacuum mode (PH2O = 0.6 ÷ 0.9 torr). 2.5. Infrared spectroscopy Micro-FTIR investigation was carried out with a Jasco FTIR 4100 equipped with an IMV-400 optical microscope. In particular, the conventional spectra collected on the surface and on the back of samples were obtained using a single nitrogen-cooled mercury cadmium telluride (MCT) detector. Reflection measurements were recorded with a 16× Cassegrain objective collecting 800 scans with a spectral resolution of 4 cm− 1 on a spot of 625 × 625 μm2. The mapping measurements were carried out by a nitrogen-cooled linear array detector composed of 16 MCT detector elements. An x–y computer-controlled stage enabled scanning of the sample line by line according to the selected area. Reflection measurements were performed with a 16× Cassegrain objective collecting 10,000 scans with a spectral resolution of 8 cm−1 and a lateral resolution of 12.5 μm. 2.6. Micro-Raman spectroscopy Micro-Raman spectra were recorded by a Jasco NRS-3100 spectrometer coupled to an optical Olympus microscope (100× objective) and equipped with an argon laser source at 514 nm. The laser power at the sample was 1–2 mW. The instrument was equipped with a 1200 lines/mm grating, providing approximately a resolution of 3–5 cm−1, and a CCD detector Peltier cooled to − 50 °C. Spectra were acquired with 10 s and 5 scans. 3. Results 3.1. 1H NMR stratigraphy Following the 1H NMR method already described in our previous paper [3], stratigraphic determinations were carried out in situ in a fully non-invasive way. The method encodes the amplitude of the 1H NMR signal as a function of the depth scanned [6–9] allowing us to discriminate and visualize different layers made of different hydrogencontaining materials. The trend of the stratigraphy was found similar in all the scanned regions of the investigated paintings. As example of the numerous results obtained on different areas, two cases are shown in Fig. 2, chosen as representative of analogy and variability of the recorded sequences of layers. The stratigraphy showed generally a flat region (L1) with a
Fig. 2. 1H NMR stratigraphy of a wall painting belonging to Sonqi Tino pictorial cycle and exhibited at the Near East Museum of Sapienza University of Rome. The labeling of each layer (L1–L6) is reported. a)1H NMR stratigraphy where thin L1 and L4 layers are observed. b)1H NMR stratigraphy where thick L1 and L4 layers are observed.
very weak NMR signal, followed by a broad shoulder (L2) and a sharp peak (L3). After this maximum, a region (L4) characterized by a net decrease of the NMR signal was generally recorded, while at a deeper level, another sharp peak (L5) appeared, followed by a region (L6) with a very weak NMR signal. While this layers' sequence was substantially confirmed everywhere, the thickness of the various layers was found to be largely inhomogeneous. In particular, L1 and L4 were those showing the most variable thickness. Layer L1 was made of the inorganic materials constituting the pictorial layer together with the underlying preparation (primer). The large thickness' variability of L1, as will be demonstrated by the data of the following paragraph, is related to the different quantity of original primer removed by the strappo (Fig. 2). As example, in 2a, the thickness of layer L1 is about 100 μm whereas in 2b, it is about 500 μm, indicating that in the former case only the paint layer was detached from the original site while in the latter the detached original material is composed by the paint and the primer (or at least a fraction of it). On the opposite, layer L2 which is the first layer of material used by conservators for the adaptation of the removed painting to a new support [3], is thicker in 2a than 2b, indicating that a higher amount of consolidating material had been necessary to level off and consolidate the regions where the strappo removed a thinner layer of material. The thickness of layer L4 gives information on the level of adhesion between layers L3 and L5. Basically, the thick layer L4 observed in Fig. 2b is characterized by the absence of NMR signal (no organic materials) indicating a detachment at the interface between L3 and L5. On the opposite, the thin layer L4 in 2a, where an NMR signal is present, indicates that L3 and L5 are still joined together.
722
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
3.2. Pigments and inorganic materials characterized by SEM-EDS, micro-Raman and micro-IR spectroscopy In order to investigate the execution technique of the detached wall paintings, a few micro-samples were collected from red (1P), black (2P), green (3P), and yellow (4P1 and 4P2) painted areas and analyzed by SEM-EDS, micro-Raman, and micro-IR spectroscopy. Samples were examined directly and in cross-section, with the aim to identify pigments (and possibly binder), the original primer, and the backing. Fig. 3a shows the map of elements of sample 1P obtained by SEMEDS. The pigment layer was constituted mostly by aluminum, silicon, and iron. Iron-rich particles were found from the outermost surface to a depth of about 20 μm, although some aggregates containing iron were also found at higher depth. Specific attention was focused on the characterization of the primer whose thickness in sample 1P was found to be about 200 μm by SEM-EDS analysis. The elemental imaging shows that S, Si, Al, and Ca are distributed in the whole primer layer, see Fig. 3a. Below this layer, at a deeper level and beyond the primer, a white thick layer very rich in calcium (96 wt%) was observed. This latter layer is composed by the calcium carbonate used to level off and consolidate the detached original paint. The micro-Raman measurements performed on the painting layer showed the characteristic peaks of hematite, see Fig. 4a, typical of a red ochre. The bands at 298, 410, and 610 cm−1 are ascribed to the hematite Eg symmetry vibration modes, while the band at 660 cm−1 is attributed to the relaxation of the selection rules induced by disorder effects in the crystal structures of hematite for the possible substitution of Fe3+ by Al3+ cations [10]. The IR spectrum of the sample acquired in reflection mode from the surface of the painting layer, without any treatment, is shown in Fig. 4b. The spectrum revealed the presence of aluminum silicates, mostly kaolin and quartz [10], calcium carbonate [11], and calcium sulfate with different levels of hydration (gypsum CaSO4•2H2O, bassanite CaSO4•0.5H2O, anhydrite CaSO4) [12]. A proteinaceous substance was also clearly revealed on the surface of the painting layer. This latter material might be ascribed either to the original binder or, most probably, a residue of the glue used in the first steps of the strappo. The FTIR spectrum acquired in reflection mode from the back side of the sample is shown in Fig. 4c. In this case, calcium carbonate and an acrylic resin [13], used as consolidant after the strappo, were identified. The SEM-EDS analysis of sample 2P (sample with a black pigment), revealed the presence of a layer about 150 μm thick, constituted mainly by Al and Si, whereas Ca was the main element of the inner part of the
sample, see Fig. 3b. The surface of the sample was quantitatively analyzed by EDS without embedding it in resin. Apart from C and O with relative amounts about 26.5 and 44.4 wt% respectively, Si (15.7 wt%) and Al (7.6 wt%) were the main constituents of the surface. Ca (1.8 wt%) and Fe (1.4 wt%) were also present on the external surface, together with a low amount (b1 wt%) of potassium, sulfur, and sodium. These data suggested the presence of carbon black, as effectively found by micro-Raman spectroscopy. In fact, the Raman spectrum of Fig. 3d shows two bands at 1585 and 1350 cm−1 known as G (graphite) and D (disorder) bands due respectively to E2g and A1g modes [14]. The SEM-EDS analysis of sample 3P (sample with a green pigment) revealed the presence of an external layer about 20 μm thick, constituted by Cu, Cl, and Mg, see Fig. 3c. Again Ca was the main element of the inner part of the sample along with a very small amount of Si, S, Al, and Fe. The Raman spectrum showed bands at 976, 914, and 830 cm−1 ascribable to hydroxyl deformation bands, and bands at 511 and 150 cm−1 due to copper oxygen vibrations typical of atacamite, see Fig. 4e. Therefore, Raman spectra along with SEM analysis of the external layer clearly indicate the use of atacamite as green pigment. The cross-section of sample 4P1 (first sample of the two with a yellow pigment) was investigated by SEM-EDS, see Fig. 3d. This sample was obtained from the region where the stratigraphy reported in Fig. 2a was collected. A first layer about 40 μm thick was found to be mainly constituted of Al, Si, and Fe. A second layer about 250 μm thick mostly constituted of Ca was observed. A very thin layer (b5 μm) of Ti was present at the interface between the first and the second layer. The Raman spectrum did not allow for the identification of the pigment because of a very intense florescence which covered any other signal. However, from the elemental composition obtained by SEM-EDS, it was reasonable to hypothesize that the pigment employed was a yellow ochre. The presence of the thin Ti layer just below the paint might be ascribed to a polymorphous forms of titanium dioxide minerals which can be found in Egypt usually associated with other minerals like calcite (CaCO3), quartz (SiO2), or hematite [4] or to a material related to the treatment of the strappo. The cross-section of sample 4P2 (second sample from a yellow area) was first investigated by SEM-EDS. This sample was obtained from the region where the stratigraphy reported in Fig. 2b was collected. The elemental composition of the first layer obtained by SEM-EDS was very similar to that found in sample 4P1. The elemental composition of the second layer was found to be dependent on the depth. In fact below the pigment layer a region rich in Ca (24%) and S (19%) was observed, and at a greater depth, a layer very rich in Si (61%), Al (14%), Fe (6%), whereas the amount of Ca (7%), and S (3%) decreased.
Fig. 3. Map of elements obtained by SEM-EDS of a) sample 1P (red pigment), b) 2P (black pigment), c) 3P (green pigment), and d) 4P1 (yellow pigment). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
723
Fig. 4. a) Micro-Raman spectrum of a red pigment (1P) showing the characteristic peaks of hematite. b) Infrared spectrum in reflection mode collected on the painting layer of the sample. c) Infrared spectrum in reflection mode collected on the back side of the sample. d) Raman spectrum of a black pigment (2P) showing the peaks of carbon black. e) Raman spectrum of a green pigment (3P) showing the peaks of atacamite.
It is worth to note that samples 4P1 and 4P2 showed a major difference in the composition of the region below the pigment layer. In the former case, a very low amount of sulfur and high amount of calcium was found, whereas in the latter case, a high amount of both S and Ca, indicating the presence of calcium sulfate, was detected. These results are also in a very good agreement with the NMR depth profiles collected on the two regions where samples 4P1 and 4P2 were obtained, see Fig. 2. In fact, the stratigraphy of the region of sample 4P1 displayed a thin pictorial layer whereas the stratigraphy of the region of sample 4P2 displayed a thick layer indicating that in this case the strappo removed both pigments and primer. In order to better clarify the molecular composition and the distribution of the different components along the original primer and the backing layers, a reflection infrared spectroscopy mapping on the cross-section [15] was carried out on sample 4P2. Fig. 5a and b show the optical images of the investigated cross-section, whereas Fig. 5c–f show the chemical maps obtained by using the integration of specific marker band of calcium sulfates, calcium carbonate, kaolin, and a synthetic resin. To obtain a well-defined and unambiguous color map distribution, bands neither affected by distortion due to reflection nor by overlap with other bands, were selected. It is possible to observe that the original primer layer was constituted
mainly by kaolin, calcium sulfates, and calcium carbonate distributed in different amounts on the entire primer layer. Calcium sulfates and kaolin signals completely disappeared in the region of the backing layer. The acrylic resin was found to be distributed along the entire sample; however, its amount decreased going from the backing toward the original primer. A high concentration of calcium carbonate was found at deeper depths beyond the primer, indicating the presence of a thick carbonate layer mixed with an acrylic resin used to level off the back of the detached painting during the strappo procedure, see Fig. 5d and f. 3.3. Inorganic materials of the primer investigated by solid-state NMR 27
Al and 29Si MAS NMR spectroscopy [16] was applied to characterize the alumino-silicate inorganic compounds present in the more superficial layers of original and non-original areas of the detached wall paintings. The 27Al MAS spectra of (a) sample 1A, non-original material obtained from a rim of a lacuna, (b) sample 2A, taken off from an original unpainted area, and (c) sample 3A, a sample of local clay, are all typical of a raw clay [17], with the resonance of Al in an octahedral environment (AlVI) apparently centered at 2.3 ppm, and the resonance of Al in
Fig. 5. a) and b) Optical images (10× and 16×) of the cross-section of the sample 4P2 where the blue grid indicates the mapped area by reflection infrared spectroscopy (single point measurement of dimensions 12.5 × 12.5 μm2). c) Color map of the calcium sulfates distribution integrating the band at 2080–2320 cm−1. d) Color map of calcite distribution integrating the peak at 2440–2630 cm−1. e) Color map of kaolin distribution integrating the peak at 3615 cm−1. f) Color map of synthetic resin distribution integrating the peak at 1740 cm−1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
724
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
a tetrahedral environment (AlIV) apparently centered at 56.3 ppm, see Fig. 6a, b, and c. The ratio between the intensity of the resonance of AlVI and that of AIIV was found to be definitely greater in sample 2A than in 1A, being 0.90 and 0.54, respectively. Therefore, these two samples differentiate in the AlIV content with respect to AlVI. On the other hand, the intensity ratio measured in the spectrum of the superficial layers of the original region (sample 2A) matches that of the sample of the local clay, being 0.90 and 0.93, respectively. Fig. 6c and d show the 29Si MAS spectra of samples 1A and 2A. In the spectrum of 1A, a broad unresolved resonance is observed, whereas sample 2A shows a better resolved spectrum where resonances of kaolin and quartz may be easily assigned at −91.4 and −107 ppm, respectively, by comparison with the spectra of standard compounds, see Fig. 6f (kaolin), and g (quartz). 4. Discussion As in the preceding work, 1H NMR depth profiling revealed the presence of six successive layers (L1–L6). L1 showed a variable thickness due to the different thickness of original material removed by the strappo, in some areas only the paint layer, with a small quantity of primer was removed, in other areas paint and primer were removed together. Micro-Raman and SEM-EDS analyses carried out on the surface layer of the detached paintings allowed the identification of original pigments to be carried out. Pigments were found to be red and yellow ochre, carbon black, calcium carbonate (white), and atacamite (green). These data are very similar to those obtained in an analogous parallel study recently carried out on the detached paintings of the Christian
Fig. 6. 27Al MAS NMR spectra of sample 1A obtained from the rim of a lacuna in a nonoriginal area, s/n = 100 (a), sample 2A obtained from an unpainted original region, s/n = 50 (b), and a sample of local clay, s/n = 100 (c). 29Si MAS NMR spectra of sample 1A (d), sample 2A (e), kaolin (f), and quartz (g).
Nubian churches of Faras and of the two close sites of Old Dongola and Baganarti [4]. These paintings were removed by a group of Polish archeologists during the same UNESCO rescue campaign of Nubian culture in 1967–70 and are currently in exhibition at the National Museum in Warsaw. The church of Faras was subject along the centuries to repeated rebuildings and its paintings are dated from the 7th to 14th century. The only difference in the pigments used in Sonqi Tino and those found in Faras is due to massicot (PbO), that is present, together with some ochre, in yellow areas of the paintings of Faras, but not in Sonqi Tino. Besides the analogy in the use of pigments, probably related to the availability of local minerals, strong analogies are also found in the materials which compose the primer below the paint layer. In our case, SEM-EDS revealed the presence of an aluminum-silicate matrix and calcium, while micro-FTIR analysis revealed the presence of calcium carbonate, kaolin, quartz, and calcium sulfate with different levels of hydration (gypsum, bassanite, and anhydrite). The presence of an aluminum-silicate matrix, possibly a clay, and kaolin and quartz was also confirmed by 27Al and 29Si MAS NMR analysis. 27Al MAS NMR also differentiated between an unpainted original plaster and the plaster used to fill up missing regions after the strappo. Furthermore, the 27Al MAS spectrum of the unpainted original plaster matched that of a local clay. FTIR mapping allowed the distribution of components in the original primer and backing layer to be visualized. Specifically, kaolin and calcium carbonate and sulfate were found to be distributed in the whole primer layer, whereas high concentration of calcium carbonate and a synthetic resin (Paraloid) were found mixed in the backing layer. This latter result is in agreement with the previous result obtained by 13C CPMAS NMR spectroscopy that detected the presence of an MA/EMA copolymer (Paraloid) in layer L2 [3]. It is worth noting that in all the examined paintings from the areas of Faras, a significant presence of titanium was recorded in a direct surface examination by XRF [4]. In the present case, titanium was significantly recorded only in a single case, showing the feature of a very thin layer of material at the interface between the first layer L1 (painting layer and primer) and the layer of calcium carbonate used to level off the detached materials (L2). Its location and layering suggest that the presence of Ti might be more related to the treatment during the strappo than to the minerals used in the original paintings. However, this topic deserves further deepening. Unfortunately no unambiguous identification of the binder was possible. A proteinaceous material was recorded in the paint layer, but this material could be a residue of glue used in the strappo. Limewash or kaolin could be also a possible binder, but no final conclusion can be achieved. Fig. 7 shows a sketch summarizing all layers constituting the detached mural paintings in relation to their 1H NMR depth profile. The original painting layer (L1) whose thickness varies between 100 and 500 μm, is made of pigments and primer. The original primer was only present in those regions where the strappo removed a sufficient thickness of material. All other layers (L2–L6) are due to the materials used in the strappo and adaptation of the painting to the new support. Paraloid embedded into calcite (L2) was used to consolidate and level the back of the paintings off. An adhesive (MA/EMA/PHEMA) was used to size a new canvas to the back of the detached painting (L3). The same adhesive was used to fix the back of the system paintingcanvas to plasticized PVC that constituted the new supporting panel (L5). L6 is the supporting panel (PVC). A remarkable note is that layer L4 is indicating a well evident loss of adhesion, in some areas, between L3 and L5. 5. Conclusions A multi-technique analytical study was carried out on the execution practices of Medieval painters working on the wall of the Christian
N. Proietti et al. / Microchemical Journal 124 (2016) 719–725
725
6. Ackowledgments The present research was supported by CNR (Italy) within the SM@ RTINFRA-SSHCH project, “Smart Integrated Infrastructures for Data Social Sciences, Humanities and Cultural Heritage Ecosystem” (2012–2015). We acknowledge Prof. L. Sist, B. Provinciali, A. Marinelli, and M. Necci for the useful discussion, and for their kind collaboration during measurements in situ and sampling.
References
Fig. 7. 1H NMR stratigraphy and sketch of all layers constituting the detached mural paintings. L1: original pigments and primer (kaolin, gypsum, and clays); L2: calcite and Paraloid; L3) MA/EMA/PEHMA resin and canvas; L4) interface between adhesive and support; L5) MA/EMA/PEHMA resin and PVC; L6) PVC.
Nubian church of Sonqi Tino, dated back to the 10th century. The study integrates a previous work dedicated to the characterization of the organic materials of the complex multilayered artifact obtained after the strappo of the paintings from the walls of the church. The investigation of the paint and primer lead to the identification of pigments as hematite (red), carbon (black), calcium carbonate (white), ochre (yellow), and atacamite (green). These pigments, when compared with those found in a recent parallel study on paintings detached from Nubian churches of the areas of Faras, Old Dongola, and Baganarti, confirm that these are the typical materials used by the painters of Nubian churches during Middle Ages (9th–12th century), with a recurrent use of the same materials probably related to the availability of local minerals. The use of local materials by Nubian painters was also confirmed by 27Al MAS NMR measurements on the alumino-silicate materials of the unpainted original primer's plaster that matched that of a local clay. The work demonstrates how an integrated multi-analytical approach with exploitation of adequate noninvasive and minimally invasive techniques, allows satisfactory results to be obtained even on complex systems as the detached mural paintings of Sonqi Tino. Crucial is the exploitation of the non-invasive 1H NMR depth profiling integrated by elemental and molecular spectroscopy mappings on crosssections and with 27Al and 29Si MAS NMR spectroscopy. The analytical protocol here proposed allows satisfactory answers to be given to questions regarding the practices of Nubian Medieval artists, a topic on which only very few technical and scientific information is available to date and on the current state of conservation of the detached paintings giving relevant information to museum's conservators.
[1] F. Piqué, L. Tintori, E. Borsook, Wall painting conservation in Tuscany before the Florentine flood of 1966, Les anciennes en peiture murale: journées d’études de la SFIIC, Section française de l’IIC, Dijon, France 1993, pp. 91–105. [2] L. Mora, P. Philippot, P. Mora, Conservation of Wall Paintings, ICROOM, Rome, Italy, 1984 265–308. [3] V. Di Tullio, D. Capitani, F. Presciutti, G. Gentile, B.G. Brunetti, N. Proietti, Noninvasive NMR stratigraphy of a multi-layered artefact: an ancient detached mural painting, Anal. Bioanal. Chem. 405 (2013) 8669–8675. [4] O. Syta, K. Rozum, M. Choińska, D. Zielińska, G.Z. Żukowska, A. Kijowska, B. Wagner, Analytical procedure for characterization of medieval wall-paintings by X-ray fluorescence spectrometry, laser ablation inductively coupled plasma mass spectrometry and Raman spectroscopy, Spectrochim. Acta B At. Spectrosc. 101 (2014) 140–148. [5] J. Perlo, F. Casanova, B. Blümich, Profiles with microscopic resolution by single-sided NMR, J. Magn. Reson. 176 (2005) 64–70. [6] F. Presciutti, J. Perlo, F. Casanova, S. Glöggler, C. Miliani, B. Blümich, B.G. Brunetti, A. Sgamellotti, Non invasive nuclear magnetic resonance profiling of painting layers, Appl. Phys. Lett. 93 (2008) (033505-1-033505-3). [7] B. Blümich, F. Casanova, J. Perlo, F. Presciutti, C. Anselmi, B. Doherty, Noninvasive testing of art and cultural heritage by mobile NMR, Acc. Chem. Res. 43 (2010) 761–770. [8] V. Di Tullio, N. Proietti, D. Capitani, I. Nicolini, A.M. Mecchi, NMR depth profiles as a non-invasive analytical tool to probe the penetration depth of hydrophobic treatments and inhomogeneities in treated porous stones, Anal. Bioanal. Chem. 400 (2011) 3151–3164. [9] G.R. Fife, B. Stabik, A.E. Kelley, J.N. King, B. Blumich, R. Hoppenbrouwers, T. Meldrum, Characterization of aging and solvent treatments of painted surface using single sided NMR, Magn. Reson. Chem. 53 (2014) 58–63. [10] D. Bersani, P.P. Lottici, A. Montenero, Micro-Raman investigation of iron oxide films and powders produced by sol–gel syntheses, J. Raman Spectrosc. 30 (1999) 355–360. [11] C. Ricci, C. Miliani, B.G. Brunetti, A. Sgamellotti, Non-invasive identification of surface materials on marble artifacts with fiber optic mid-FTIR reflectance spectroscopy, Talanta 69 (2006) 1221–1226. [12] F. Rosi, A. Daveri, B. Doherty, S. Nazzareni, B.G. Brunetti, A. Sgamellotti, C. Miliani, On the use of second-order modes for the analysis of the CaSO4-H2O system by reflectance FTIR, Appl. Spectrosc. 64 (2010) 956–963. [13] F. Rosi, C. Miliani, C. Clementi, K. Kahrim, F. Presciutti, M. Vagnini, V. Manuali, A. Daveri, L. Cartechini, B.G. Brunetti, A. Sgamellotti, An integrated spectroscopic approach for the non invasive study of modern art materials and techniques, Appl. Phys. A 100 (2010) 613–624. [14] F. Tuinstra, J.L. Koenig, Raman spectrum of graphite, J. Chem. Phys. 53 (1970) 1126–1130. [15] F. Rosi, A. Federici, B.G. Brunetti, A. Sgamellotti, S. Clementi, C. Miliani, Multivariate chemical mapping of pigments and binders in easel painting cross-sections by micro IR reflection spectroscopy, Anal. Bioanal. Chem. 399 (2011) 3133–3145. [16] M.E. K. J. D MacKenzie, Multinuclear Solid-State NMR of Inorganic Materials, Pergamon Materials Series 6. , Elsevier Science Ltd, Oxford, UK, 2002. [17] F. Presciutti, D. Capitani, A. Sgamellotti, B.G. Brunetti, F. Costantino, S. Viel, A.L. Segre, Electron paramagnetic resonance, scanning electron microscopy with energy dispersion X-ray spectrometry, X-ray powder diffraction, and NMR characterization of iron-rich fired clays, J. Phys. Chem. B 109 (2005) 22147–22158.