M. Járó - L. Költő szerk.: Archaeometrical research in Hungary (Budapest, 1988)

Analysis - ZIMMER Károly: Spectrochemical investigation and classification of Hungarian glass finds

meter. The operating parameters are given in the original paper [21]. Calibration lines were prepared using synthetic standards for both the minor and the trace elements. The mean relative standard deviations of the standard solutions were in the range of one and a few % for all elements [22]. The detection limits for the elements in glass are practically identical with the detection limits obtained for the trace elements in silicates by means of the ICP method, and completely satisfy the requirements for the characterization of archaeological glasses. Analysis by GDS. The matrix composition of standard samples was chosen to agree with the major components in the glass samples to be investigated. The glasses were ground to a size of < 50 urn and heated to avoid water contamination, then mixed with copper powder (or zinc powder for the determination of copper) and pressed to pellets. The operating parameters are given in the original paper [23]. For non-metal powder pellets constant power operation has been found exceptionally good [18]. For high concentrations (Fe, Mn, Al, Cu and Zn), a linear relationship between concentration and intensity was used to construct the analytical curves whereas for the low concentration range (Ti, V, Cr, Ni and Ag) a logarithmic relationship was employed. The limits of detection and relative standard deviations met the requirements for archaeological glass analysis. By comparing the two spectrometric methods used in the detemiination of microcomponents it can be stated that both methods are adequate, but for ICP—AES sample preparation and standardization are simpler. Therefore, for the determination of the microcomponents of mediaeval glasses, particularly in the case of investigation of elements in a concentration range below 10 ppm, the use of the latter method is preferable. EPMA-investigation. In the course of the investigations of colouring components of archaeological glass fragments [21] on the mediaeval glasses of the Palace of Buda Castle only sample inspection of the glass fragments indicated unanimously that the quality, colour and state of the surface of several fragments are significantly different from the properties of the bulk. The idea emerged that the concentration distribution of elements should be investigated by electron probe microanalysis (EPMA) on the surface and in the bulk of glasses, in order to draw conclusions on inhomogeneity, eventually on manufacturing technology, on interactions with the components of soil in the environment and on the diffusion processes taking place in the glass. The accompanying and trace elements of 67 glass fragments were determined with d. c. arc excitation on a PGS—

2 plane grating spectrograph. The experimental conditions are given elsewhere [24—27]. Based on the spectrochemical investigations, the elementary composition of various glass fragments fluctuated significantly: Al: 0.72-1.8%, Ti: 0.18-0.42%, Fe: 039-0.77%, Mn: 0.37-5%, Cu: 0.003-0.0037%. Ag: 0.0002-0.002%, Sn: 0.11-0.42%, V: 0.005-0.036%, Zn: 0.02-0.18%. The thicker glass samples (4-5 mm in width) were cut with a diamond saw • into plates 1 mm in thickness, and the concentrations of AI, Ba, Ca, Cu, Fe, K, Mg, Mn, Na, S and Si were determined on both surfaces of the plates obtained. Thin glasses were cut into half perpendicularly to their planes, and the new surface was scanned. The element distribution in micro dimensions was investigated with an ORTEC energy dispersive analytical system combined with a Cambridge Stereoscan 150 B electron microscope, and connected to a PDP 11 data system. Gold and carbon conductive layers, respectively, were evaporated onto the samples. Analysis of the samples taken from different parts of the glass fragments indicated significant inhomogeneity in the bulk of the fragment, and the concentrations of various elements showed a distinct, increasing or decreasing tendency when going from the bulk toward the surface. These conclusions have been confirmed by measurements performed in longitudinal and transversal directions in the inner layers of the fragment.