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

Analysis - GEGUS Ernő, BORSZÉKI János: Investigation of archaeological metal findings by a laser-microspectral analysis method and characterization of results using pattern recognition methods

Valuable results were achieved by using laser^nicrospectral investigations to elucidate the origin of various gold and silver findings, in revealing fakes [9], or in detecting different layers on metal objects (e.g. silver coating on sarmata bronze findings) [10]. The composition of ancient alloys significantly differs from that of todays alloys, according to the development status of the techniques; this circumstance makes easier in general the decision of originality or fake, but at the same time it causes problems in the calibration of relative methods of instrumental analysis — including that of the lasermicrospectral analysis procedure - since it is not always possible to, collect reference samples of suitable known composition. In order to solve this problem, we followed two ways [11]: 1/a small amount of sample was taken from selected findings similar to the sample to be investigated, suitable for reference samples, which could be destroyed and analysed by a spectrochemical method in solution; 2/ using logarithmic intensity data of analysis lines selected in the laser spectra we classified the findings by mean's of mathematical-statistical pattern recognition methods without determining the chemical composition [12, 13]. Possibilities of spectral analysis investigations of archaeological findings, in general, have been dealt with in respect of archaeometry in a comparative study [14]. On the basis of our experience gained in the cited works, the experimental conditions and efficiency of a practically nondestructive laser-microspectral analysis procedure are given as follows, suitable for identifying the chemical composition of archaeological metal findings. Later, pattern recognition methods are systematized which may be suggested for mathematical statistical data processing of the experimental results and for the classification of the findings. 2. Application of laser-microspectral analysis in archaeology Lasernrnicrospectral analysis [15] is especially suitable for determining the elemental composition of metal objects in which the low energy (several tenths of a joule) but high power (in the order of MW, according to its short period) laser beam, focused onto the selected place, will be absorbed in the target, and a small quantity of the sample (a fraction of a microgramme) evaporates in a burst. The investigation is practically nondestructive, the sample will not be damaged noticeably, and it requires no surface preparation, removal of corroded layer, cleaning or placement in a vacuum chamber. In the rapid process of evaporation selective vaporization cannot take place therefore the composition of the emitted vapour cloud will be identical with that of the sample. The absorption of laser energy in metals of high reflection power and high heat conductivity (Cu, Ag, Au) can be accomplished to a sufficient degree only by using a Q-switched laser of suitable power density, while an uncontrolled laser is inconvenient for the reproducible vaporization of samples of this type [5, 7]. Furthermore, by controlling the resonator of a Q-switched laser, one can vary the number and the power density of pulses emitted with a single laser shot. 50—100 pulses, produced at a low threshold value of the resonator, create a crater about 0.1 mm in diameter and 0.102 mm deep on the selected point of the sample, while 1—2 "giant pulses" of high power density, produced at a high threshold value, do not penetrate deep into the sample, thus one can acquire information about the composition of a surface layer or coating, having evaporated a material portion on a surface of about 03 mm diameter, but only at a depth of 1-5 urn. The emission of a vapour cloud, created using a laser shot from a sample, is not intense enough by itself to produce a spectrum of sufficient intensity, therefore we