OM - SEM/EDS - µFT-IR Synergy
Rapid Palette Identification of Paintings Cross-Sections
- Fig. 1: Gianfrancesco da Tolmezzo´s "Curcifixion" (ca. 1505) during restoration (Nikolina Oštarijaš, Croatian Conservation Institute, Department of Documentation).
- Fig. 2: Sample cross-section observed by optical microscope in visible light (A), visible polarized reflected light (B), UV reflected light (C) and BSE image (D).
- Fig. 3: Sample cross-section observed in visible light (A), UV reflected light (B), BSE image (C) and EDS elemental maps (D).
- Fig. 4: Sample cross-section in visible light (A), µFT-IR ATR spectrum of the reddish lacquer (B).
Microscopy represents a fundamental method for cultural heritage materials research and it is essential for paintings cross-sections investigation. Because of the smallness of sample, microscopy and micro-spectroscopy analysis techniques have taken a leading role in the past decades, improving and expanding research possibilities. The increasing accuracy and versatility of the instrumentation enable immediate and precise results. Furthermore, the combination of different non-destructive micro-analytical techniques, allows a more complete characterization of the sample.
We regularly perform analyzes related to cultural heritage materials using a very little amounts of sample on which we carry out several non-destructive examinations in order to respond to the questions that art historians or conservators may have. The technologies today available allow us to "extract" a lot of information enclosed in a sample. The spectra and images obtained are subsequently interpreted and correlated with historical data. These results sometimes don't coincide with the information already existing about an artifact: the detection of a modern, synthetic material can change its dating, undermine its authenticity or reveal past technological achievements unknown so far. As what concern paintings analyzes, up to a few milligrams of painted layers, are embedded in transparent resin and polished to obtain a cross-section. The first step is observation by optical microscopy to define the number of layers, their function (preparation, underpaint, pigment, repainting...), thickness, structure and color. At this phase, some pigments are immediately recognizable for their distinctive color and particles size and morphology (figs. 2 and 3). Furthermore, the observation with UV or IR light allows detecting the presence of varnishes or pigments that absorb in the UV region. For chemical composition characterization, micro-spectroscopic non-destructive techniques are indispensable, like micro-Fourier transform infrared spectroscopy (µFT-IR) or energy dispersive spectroscopy in electron scanning microscopy (SEM/EDS).
Together, these methods allow the determination of both inorganic and organic pigments, elemental composition as well as characterization of binding media. Observation with back scattered electrons enables accurate layer definition as well as a detailed particles analyses.
What represents a difficulty is the characterization of mixtures of pigments, filling materials and/or varnishes. The identification of each compound becomes increasingly complicate as the number of compounds inside a layer rises. Painting materials were rarely used as pure substances, much more often admixtures, adulterations and concoctions can be encountered. Their preparation was well known among the old masters and their use was so widespread and established that they had names which indicated univocally the substances included and their ratio . For this reason, a multidisciplinary approach is required. Knowing the period of creation and the author, the presence of some synthetic pigment(s) could be excluded a priori.
Microscopy Techniques Arrangement on the Example of the Painting "The Crucifixion" of Gianfarncesco da Tolmezzo
The restoration of the painting is a project currently underway at the Croatian Conservation Institute and coordinated by Pavao Lerotić and Višnja Bralić (Department for Easel painting - Movable Heritage). The analyzes presented in this paper were conducted in the Chemical Department of the Materials Research Centre of Istrian County - METRIS, as collaboration in the mentioned project. Sampling and resin embedding were performed at the Natural Science Laboratory of the Croatian Conservation Institute. On an altarpiece which painted surface exceed two square meters, six centuries old, like the "Crucifixion" (ca. 1505) of Gianfrancesco da Tolmezzo (1450-1510) commissioned by Bishop Lucas for the Zagreb Cathedral  which interior was restored many time due to earthquakes, raids or fires, it is expected to find many retouches, beside the original layers (fig. 1). X-rays revealed more than one retouch and during the XX century the artifact was restored many times, the last one, about which no documentation is available, was done within an exhibition in the eighties (information provided by project coordinators).
Owing to its dimensions and history, the definition and characterization of original layers could be complex in the first place. Twenty samples were taken: some of them have already been analyzed. Parts of the results are presented in this paper. The examination of the rest is still in process. The examination by optical microscopy (AXIO Imager M2m Carl Zeiss, BX51 Olympus), in visible, UV and polarized light, was the first step which allowed choosing samples being subject to chemical composition analysis and indicating which painting strata (more than 30) to focus the chemical characterization on. Chemical composition of the examined layers was carried out by Fourier transformed infrared microscope analysis (FT-IR microscope Hyperion 1000, ATR objective - Tensor 27, Bruker) and a field emission scanning electron microscope with energy dispersive spectroscopy device (FEI Quanta 250 FEG). SEM images and micro-analyzes were carried out in low vacuum mode (80- 100 Pa), at 15-20 kV acceleration voltage and backscattered electron detection. These techniques revealed presence of azurite, lead white, malachite, zinc white, chrome green and cinnabar. The gilding, detected by EDS as silver and considered an original layer, was first observed by optical microscopy with polarized light which allows metallic layers detection (fig. 2B).
The analysis of the green colored sample reveals numerous strata. UV reflected light image showed a layer containing particles with strong ultraviolet absorption (fig. 3B). No composition differences in the same layer were revealed by the BSE image (fig. 3C) where instead cracks among painting strata were evidenced. EDS mapping discovered zinc and chrome in the same layer. Zinc is present as zinc white (blue phosphorescent particles in figure 3B), a pigment introduced by the end of the XVIII century . The above green pigment was identified as chrome green (C 42,04 %, O 27,67 %, Cr 21,96 %, Zn 6,15, Ca 1,46 %, S 0,72 %) a synthetic pigment in use from the half of the XIX century . Due to the absence of lead in the yellowish pigment showing phosphorescence (fig. 3B), usage of chrome yellow (lead chromate), historically the most important pigment among the chromates, was excluded . Silicon, though detected, is not actually present in any of the strata; its occurrence is due to particles of the silicon carbide grinding paper, used to polish the sections, trapped in the spaces among the strata.
The ATR spectrum of the red colored sample shows the presence of lead carbonate (fig. 4), based on C-O stretching and CO32- bending vibrational bands at 1405 and 679 cm-1, present in the layer below, mixed with red pigment. Peaks at 1647 and 1537 cm-1 correspond to the amide bands (C=O stretching and C-N-H bending vibrations), that with the intense bands at 2923 and 2853 cm-1 (C-H bonds) indicate the usage of a proteinaceous binder. Animal glue, another proteinaceous binder which spectra presents very similar bands, may have also been applied to consolidate the painted layers during restoration in the past. Considering the lack of the C=O stretching band at 1740 cm-1, typical for oils , it is possible to exclude the presence of a lipid binder, which supports the assumption that tempera painting technique have been used. The bands in the range of 1140-991 cm-1 could be assigned to glucose in the molecule of the natural dye carmine (cochineal) . The explained spectrum illustrates well the possibilities and the limitations of FT-IR micro-spectroscopy in distinguishing analogous binders and in pigments characterization. Due to the presence of several substances and band overlapping, the interpretation could be complex and wrong explanations even of a few peaks, can be misleading. To proceed with the appropriate restoration approach, it is of great importance for the restorer to differentiate original layers from interventions that occurred later. Thus, he can decide about removing a layer or not, choosing compatible restoration materials to treat the artifact, fearless of causing some damage or even removing parts of the original material.
In the presented research, the appropriate combination of microscopic and micro-spectroscopic techniques allowed a fast and inexpensive determination of most of the materials today present on the artifact. Beside of a logic chronological interpretation of the layers sequence, a conservation state assessment was possible at the same time, simultaneously providing very valuable information from the conservation as well from the historical point of view. The restoration and ongoing investigations on the imposing altarpiece, have not an easy goal, that is narrate its vicissitudes.
PhD Višnja Bralić and Pavao Lerotić of the Croatian Conservation Institute Zagreb are acknowledged for providing the samples, materials and information about the altarpiece and for their support. Prof. PhD Snežana Miljanić, Division of Analytical Chemistry, Faculty of Science in Zagreb, is thanked for her advices and assistance.
 Church A.H.: The chemistry of paints and painting, 4th edition, London, 148 (1915)
 Lina Slavica Plukavec: Križni put u Zagrebu i Zagrebačkoj Katedrali (Station of the cross in Zagreb and in the Zagreb Cathedral) in Naša Katedrala, 12, Zagreb (2008)
 Kuhn H.: Zinc white, in Artists' pigments: A handbook of their history and characteristics, Vol. 1., ed. Feller R., Washington and Cambridge (1986)
 Kuhn H. et al.: Chrome yellow and other chromate pigments, in Artists' pigments: A handbook of their history and characteristics, Vol. 1., ed. Feller R., Washington and Cambridge (1986)
 Vlachos N. et al.: Analytica Chimica Acta 461, 573-574 (2006)
 Rosu M.C. et al.: Materials Science in Semiconductor Processing (16) 1554 (2013)
 Kirby J. et al.: The Technology Of Eighteenth- And Nineteenth- Century Red Lake Pigments, National Gallery Technical Bulletin, Volume 28 (2007)
Dott.ssa Tea Zubin Ferri (Corresponding author via e-mail request)
METRIS - Material Research Centre