Analysis of Paint Layers by Light Microscopy, Scanning Electron Microscopy and Synchrotron induced X-Ray Micro-Diffraction

Light microscopy (LM) and scanning electron microscopy (SEM) in combination with energy dispersive x-ray microanalysis (EDX) were used for the characterization of the structure of the paint layers of a specimen taken from a mural painting. The sample consisted of 7 layers in total, whereby a thin layer of pure gold was suspected to be the uppermost layer. The sequence of the various paint layers as well as the distribution of the elements present in the pigments could be obtained from the cross-sectioned specimen. Additionally, synchrotron induced x-ray micro-diffraction analysis (XRD) enabled the identification of the crystalline structure of the pigments used for the painting. Traversing the sectioned sample through a focused x-ray beam with a size of 2 μm allows microscopic resolved analysis of the crystalline constituents within the diverse paint layers. By this, it is possible to attribute the usage of various pigment minerals within the paint layers, even including a 2 μm thick gold layer at the surface. XRD it is rather difficult to carry out those investigations of specific pigments, as the thickness of the paint layers is in the range of several tens micrometers or even below. Therefore, synchrotron induced x-ray micro-diffraction was used in the present work, where the step scan resolution can be much smaller than the thickness of the paint layers. The combination of SEM/EDX analysis and x-ray micro-diffraction has proved to be suitable for the identification and characterization of the composition of this cross-section. 2 EXPERIMENTAL METHODS A specimen taken from a mural painting of the Baroque periods could be employed for the investigations. The specimen was embedded with a particular orientation in a transparent resin, ground and polished with SiC-paper up to 4000 mesh perpendicular to the paint layers. Highly polished layers are required for an informative microscopic investigation and flawless photographs. For analyzing this sample three different methods were applied. 2.1 Light Microscopy Analysis with the light microscope (Leitz, Orthoplan) was performed by using polarized light and the dark field reflectance technique. With this method it is possible to gain information about the thickness of the cross-section and the structure of the paint layers. Furthermore, this method yields information about the grain size and grain size distribution of the used pigments. This information is a useful indication for the way of manufacturing the painting, since modern pigments have uniform distribution of their grain size whereas hand ground colouring matters are strongly heterodisperse. 2.2 SEM/EDX Analysis In addition to the light microscopic images SEM analysis of the same cross-section was carried out with a Philips XL 30 ESEM microscope without coating the specimen, although the materials in the paint layers as well as the embedding material are electrically non-conducting. Energy dispersive x-ray microanalysis in the SEM (EDAX Phoenix) was used for qualitative analysis of the elements present in the pigments as well as for x-ray mappings yielding to the distribution of the elements in the different paint layers. 2.3 X-Ray Micro-Diffraction For this purpose a micro-diffraction facility installed at a beamline at the European Synchrotron Radiation Facility (ESRF) in Grenoble, France was used. The central quality of the synchrotron induced x-ray radiation is the minute source from which the radiation is emitted, and the small solid angle into which the radiation is confined. At a distance of several ten meters from the source an intensive x-ray beam displaying a cross-section of around 1 mm is attained. This gives the possibility to collect the majority of the beam with the x-ray optical elements, which typically have apertures of a few hundred micrometer diameter. In particular, compound refractive x-ray lenses (CRL) (Snigirev 1996, Lengeler 1999) are used, which focus x-rays in a similar manner as glass-lenses would do with visible light. The focal spot is a demagnified image of the x-ray source, imaged by the CRL. The value of demagnification can be directly derived from the geometrical properties of the imaging set-up, by ray-tracing methods usually applied in geometrical optics. The demagnification value is the ratio between the source-lens distance and the lens-sample distance. This setup benefits from the large distance of the experimental station from the storage ring, being 42 m, which is made possible because of the slowly diverging x-ray beam. With a typical lens-sample distance of one meter it is possible to get a demagnification ratio in the order of 40. The effective source size is of 50 μm vertical and 600 μm horizontal, and consequently a focal spot of 1-2 microns in the vertical direction and 15 microns in the horizontal direction is achieved at sample location. Thus, it is possible to perform x-ray scanning microscopy with micrometer resolution, if the scanning direction is chosen to be vertical. The use of a monochromator device located in the beam path between source and sample helps to select the desired x-ray wavelength by means of x-ray diffraction on Si single-crystals. Here a wavelength of 0.620 Å is selected, with a relative bandwidth of 1.4 ·10. The focusing of x-rays with CRL is particularly advantageous for the recording of diffraction signals. Diffraction is measured by observing the angular distribution of x-ray radiation scattered by the sample. Thus, the angular divergence of the incoming beam will account for the resolution of the experiment. Because the aperture of the CRL is small in comparison to the focal length, the lens provides a small divergence in the incoming beam of only 0.2 mrad. This value translates into a relative accuracy in the determination of crystalline lattice spacing in the order of 10. This value is exceedingly sufficient to distinguish diffraction patterns arising from different mineral species which are envisioned for analysis. Figure 1. Experimental set-up for micro-diffraction, at beamline ID22, ESRF, Grenoble. The monochromatic x-ray beam enters from the left. (a) is the set-up in diffraction mode, (b) shows the alignment mode. A scheme of the instrumental set-up used is shown in Figure 1. Basically, two modes come into operation. The micro-diffraction mode (Fig. 1a) has the CRL inserted, and the diffraction pattern is then collected by means of a two-dimensional x-ray camera behind the sample. At the chosen wavelength the diffracted intensities from the relevant crystalline lattice spacing values are confined within an angular cone of 90° opening angle. For recording the diffraction patterns, the diffraction-camera with an active diameter of 100 mm is located at a few centimeter distance from the sample (Fig. 2). The diffraction spectrum analyzed at a later stage is obtained by summing up all intensities belonging to the same diffraction angle and plotting the so obtained intensity distribution versus diffraction angle. Figure 2. X-ray micro-diffraction. Experimentally obtained powder diffraction patterns from different positions across the paint layers. Position number 20, 60, and 90 (from left to right, referring to numeration scheme of Figure 7). For the second operation mode the lens and diffraction camera are removed and instead a high-resolution x-ray camera is placed directly behind the sample (Fig. 1b). Illuminated by the plane x-ray beam, the radiogram of the paint layers is obtained (Fig. 7b). Primarily, this high resolution imaging mode is used to align the sample in respect to the focused beam position, which can be made with an accuracy of 2 microns. However, the different attenuation within diverse paint layers can yield some approximate knowledge about their mean density. As both modes are performed in transmission geometry, the sample was sectioned into a 300 μm thick slice, prior the x-ray experiment. The complete micro-diffraction analysis was accomplished by step-wise moving the section across the focused beam in steps of 4 μm. The paint layers were oriented horizontally, parallel to the larger beam-dimension, thus allowing to traverse the layers with a beam-size limited resolution of around 2 μm. In-between each step a diffraction-pattern is recorded, with an acquisition time of 2s. The diffraction signal arises from a sample volume which is defined by the lateral beam-dimensions, i.e. 2 x 15μm2, and the sample thickness, which was 300 μm. After 160 steps, the beam passed through all seven layers and 160 diffraction spectra were collected, which are representative for the crystalline composition of the sample. Finally, the diffraction spectra were processed in order to eliminate instrumental parameters like wavelength λ or diffraction angle 2Θ. By applying Bragg’s law for 1 order diffraction ( θ λ sin 2 ⋅ = d ) the abscissa was converted into the lattice spacing d. This makes the data comparable to the tabulated values from similar compounds. In practical terms, the measurements were evaluated by analyzing crystalline phases by means of a search/match program. In order to identify unknown sample constituents, it was necessary to compare the measured diffraction spectra with a collection of tabulated spectra of known compounds. For this purpose, the collection by the International Centre for Diffraction Data (ICDD 2000), the Powder Diffraction File (PDF 2000), was available. Using this method it is possible to identify pigments with a crystalline structure. However, amorphous dyes and binding media can not be identified. 3 RESULTS AND DISCUSSION In general paintings are made by mixing pigments with a liquid binding medium (e.g. animal glue or linseed oil). After applying a layer of paint, it is dried and other layers are added as needed (Mantler & Schreiner 2000). A final coating mainly made of natural resin normally protects the painting and is responsible for the visual quality of the optical depth. 3.1 Light Microscopic Examination The gained information by light microscopy and UV-fluorescence microscopy concerns the sequence and the thickness of paint layers and distribution of the pigments (Wülfert 1999). Figure 3 depicts a light-optical micrograph of the cross-section of the specimen investigated. The thickness of all paint layers is about 500 μm. The different layers can be seen clearly, although the gold layer, which should be on top (Fig. 3), can hardly be detected at this magnification (approximately 100x). With a higher magnification of about 400x the gold layer (layer 1) can be seen in Figure 4 as a thin dark layer on the left side of the layer 2 (red/orange layer). Both, the light microscopic (Fig. 4a) as well as the UV-microscopic image (Fig. 4b) show the surface layer as a noncontinuous thin layer clearly due to its high metallic reflectance of the visible light and the total absorption of the UV radiation. The light microscopic images in Figure 3 and Figure 4a yield to the structure of the different paint layers, where 7 layers can be determined: The surface layer (layer 1) is the gold layer followed by a red/orange layer (layer 2) clearly visible in Figure 4a. Layer 3 is a thick layer transparent in the visible light (Fig. 4a) and showing a bright bluish fluorescence in the UVfluorescence microscope, which indicates the presence of natural resin. Underneath layer 3 is a thick yellow paint layer (layer 4), followed by a thin white layer (layer 5) as shown in Figure 3. Layer six and seven are of white color with a light touch to yellow and gray, where the transition between these two layers seems to be very unsteady. Figure 3. Cross-section of the paint layers of a specimen taken from a golden part in a mural painting. (a) light microscopic image Figure 4. Detail of the paint layers in Figure 3. (b) UV-fluorescence microscopic image