| Abstract: | SUMMARIES. Several traditional paint samples were prepared for infrared spectroscopic analysis in a number of different ways. The methods applied include the formation of KBr pellets, squeezing of dissected or multi-layer samples in a diamond cell, embedding in a modern resin, and a new method in which a paint cross-section is embedded in KBr and polished from both sides to obtain a thin layer. Microtoming of a paint cross-section embedded in a modern resin was not successful. The various samples were analysed by infrared spectroscopy to assess the usefulness of these sample preparation techniques in the field of paintings research. The amount of information that can be derived from the infrared spectra obtained appeared optimal for sample preparation methods that allow the application of transmission techniques. Tire results obtained for specular reflectance techniques were of lesser quality, while no diffuse reflectance spectra could be obtained. The results can best be traced to a specific layer or structure in an iiihomogeneous paint system if the layer structure of the sample is left intact, i.e., using specular reflectance of an embedded cross-section or transmission of a thin section obtained by doubly polishing a cross-section embedded in KBr.
CONCLUSIONS. The quality of the transmission spectra presented (Figures 5, 6, 9, 10) is high, and clearly indicates that transmission would currently be the preferred analytical mode. In fact, the spectra recorded with the different transmission techniques are very similar, and only differences in relative peak intensities have been found. These peak ratios can provide a further characterization of the sample preparation methods. The silicate peak (1050 cm-1) is the most intense peak in the spectrum of layer 1 (red ground layer). It is truncated in diamond cell and specular reflectance spectra (Figures 5—8) to prevent overlap with other spectra. The intensity ratio of this silicate peak and the carbonyl band (I(- 1050 cm-1)/ I(-1710 cm-1) ) is 19 for specular reflectance, 5.2 for the diamond cell, and only 2.4 for the KBr polishing technique. The carbonate peaks (1420 cm-1) in the spectra of layers 4 and 5 relate in a similar way to the carbonyl (1720 cm-1) peak. In fact, the carbonyl band is hardly present in the specular reflectance measurement. The high relative intensity of the inorganic peaks in the specular reflectance spectra is explained by the higher reflective index of the inorganic materials and the presence of subsurface reflections. However, measurements using the diamond cell and KBr polishing method are both performed in transmission mode and should have a constant path length. The clear differences between peak ratios are therefore to be interpreted as changes in the sample. The most probable explanation is the loss of inorganic pigment particles due to the polishing of the KBr-embedded sample. In fact, the partial loss of pigment particles upon polishing is not surprising, as the size of the thin section is in the same order of magnitude as the size of the pigment particles. This indicates that the KBr polishing method changes the quantities of the materials in a sample and should be considered as a qualitative rather than a quantitative method. This is an advantage rather than a problem, as the analysis of pigments is well established, while the analysis of organic binding medium is normally disturbed by the presence of these pigments. As the quality of the spectra is sufficient for the diamond cell and KBr polishing and, if enough sample material is available, KBr pellets, other characteristics of the techniques will be decisive for the choice of a specific technique in a specific situation. Bulk techniques (KBr and diamond cell) will be the methods of choice when a spatially resolved measurement is impossible. The methods are fast and reliable, but accuracy of tracing the analytical results to a specific spot in the painting is limited by the sampling procedure. The analysis of a squeezed cross-section in a diamond cell improves the spatial resolution, and can be performed very rapidly. Disadvantages are the possible mixing of materials, especially for larger, multi-layer samples. Microtomy is a fast method, but it is not expected to work with every paint system. It is expected that microtomy will be of value when the samples investigated are small, as the support due to the embedding medium increases with decreasing sample size. In addition, microtomy of modern paints might be more straightforward, as these are normally more flexible and contain smaller pigment particles than traditional seventeenth-century paints. A disadvantage might be that the embedding medium disturbs the visual (using AgCl 18) or infrared (using modern resin [10]) observation. The KBr polishing procedure proposed in this paper provides a very high spatial resolution. This analysed thin section was obtained with a minimum of equipment, and appeared to be straightforward even for the rather complex layer structure of the painting investigated. This sample preparation might therefore prove a valuable tool in the investigation of multi-layer paint samples. |
|---|
