Report phase I
Micro-PIXE measurements for the Pietroasa Hoard
The present phase has had as main goal the performance of a study on gold objects from the Romanian heritage, namely the Pietroasa Hoard (Pietroasele, Buzau county). The study, based on PIXE (Particle Induced X-ray Emission) micro-sample (micro-beam) method, indented to determine the elemental composition and the possible presence of metal inclusions, with the scope of provenience determination of the metal used for the casting of these objects, which have an important historical and artistic value.
We have to mention here that the origin of the materials used for the metallic objects from the cultural heritage represent an important issue for historians, as well as archaeologists. The study of trace-elements compositions from the archaeological metallic objects may offer important evidence regarding the manufacturing process and/or the commercial changes from the respective epoch. The pieces made of noble metals play a significant role in archaeometry, not only due to the objects' value, but as well due to the fact that these are corrosion highly stable. For this reason, these objects are very well preserved, despite their age; at the same time, they can be easy analyzed using surface non-destructive nuclear and atomic techniques, such as: PIXE (Particle-Induced X-ray Emission), PIGE (Proton Induced Gamma-ray Emission), CPAA (Charged Particle Activation Analysis), NRA (Nuclear Reaction Analysis) and/or XRF (X-Ray Fluorescence).
Micro-PIXE is a variant of the PIXE method, using special limited beam, with micrometers spot dimensions, presenting the advantage of providing simultaneously details regarding the superficial distribution-at micrometer scale-of the elements contained in the sample; the technique allows the scanning of a limited sample surface (areas of hundreds of microns, or even square millimetres). Consequently, relevant results may be obtained out of these experiments, regarding the composition and homogeneity of the analysed sample, namely if the sample contains or not inclusions (micrometer areas with elemental composition different from the rest of the sample). The results of the micro-sample experiments are generally presented as sets of elemental maps, obtaining one map (plan view of the scanned surface) for each of the interested metallic element. Another acquisition method is that of the spectrum in a certain point (pixel), which has a measurable dimension (at micrometers or decimal micrometers levels) through experiments. Within the undertaken testing from this phase by the IFIN-HH researchers, the micro-PIXE method has been used in order to study the distribution of trace-elements from archaeological golden samples, as to investigate their provenience. Particularly, the analyse of the golden objects from the Pietroasa Hoard had the aim to determine on the one hand, the possible metallic inclusions of platinic elements (Ru, Rh, Pd, Os, Ir, Pt) and, on the other hand, the elements with high-temperature melting point (Ta, Nb), these elements being essential for determining the provenience of the metal used for manufacturing the objects.
Moreover, within this phase have also been analysed geologic golden samples with known provenience (particularly from the area of Apuseni Mountains) in order to determine a possible connexion between their composition and the composition of the archaeological samples, as well as to establish a detailed characterisation (minor trace-elements) of the golden sources from the Apuseni Mountains.
The studied objects have been represented by fragments of objects coming from the Pietroasa Hoard ("The Golden Boord Hen with its Chicken"); these fragments, having square millimetres dimensions, had been assayed during foregoing restorations. Native golden samples coming from the Apuseni Mountains (Brad-Rosia Montana region) have been also analysed during this phase.
The measurements executed within this phase have been undertaken at the AN 2000 accelerator from Laboratori Nazionali di Legnaro (LNL), Italy. The experiments from LNL have beneficiated by financial support from the 6th Frame Programme of the European Union-action Transnational Access to Research Infrastructure (TARI).
We have to mention here that the interest regarding the provenience study of the gold out of which have been manufactured the objects from the Pietroasa Hoard (Figure 1), has been manifested even before the beginning of this project. Therefore, micro-PIXE measurements at the AGLAE accelerator from the Louvre Museum, Paris, have been undertaken in June 2005, with the financial support from the Eu-ARTECH European programme. For the AGLAE experiment it has been used a beam of 3.2 Me V protons extracted in helium atmosphere, beam currents of nA and decimal of nA level and dimensions of the spot of 30, respectively 100 microns. The acquisition system has been multi-parametric, based on two detectors of Si(Li). The experimental data implied the rolling of the GUPIX code calculus, especially developed for the PIXE spectrum analyse. On the data obtained from this preliminary experiment, it was possible to determine the best working conditions for the detecting of different platinic elements-using absorbing filter made of different materials (Cu, Al) in order to grow the detection limit for the interested trace-elements. Part of these quantitative results is presented in the table with concentrations, comprised by the present phase report.
Within the November 2005 experiment, undertaken at Laboratori Nazionali di Legnaro (LNL), it has been used a proton beam of 2 Me V generated by the Van de GraaffAN AN-2000 accelerator. The beam has been focused at dimensions of 3 ' 3 m m 2. The maximum amperage of the beam has been around 1 nA. In order to reduce the intensity of the pics in the low energy spectral region (under 4 keV), a funny filter made of aluminium with a thickness of 100 microns (orifice of 0,4%) has been used. The maps have been scanned on areas of 255 ' 255 pixels. The used detector has been made of hyper-pure germanium (HP Ge) and had a FWHM resolution of 150 eV at 5.9 keV. The acquisition times have varied from one spectrum to another, being generally of a couple of hours, due to the requirements for obtaining a satisfactory statistic for every area or point analysed.
In all of the cases, the golden samples-archaeological or mineralogical- have been placed on an adhesive tape. Precautions have been taken as not to analyse the rifty areas from the samples or the areas affected by anterior manipulations. Preliminary to the micro-PIXE experiment, has been performed an optical examination of the samples with a microscope.
For all the analysed samples, it has been obtained series of elemental maps for each scanned area; moreover, spectrums have been also acquisitioned in a certain point (pixel), with dimensions of decimal or hundreds of microns. Their localisation has been chosen by the team of experimentalists, depending on the interest areas presented in each sample (i.e. points where a certain trace-element was present in concentrations higher than in the surrounding areas).
As it has been mentioned, based on these experiments, several sets of elemental maps have been established for each sample. The interpretation of the micro-sample results is by far a complicated operation. A correlation between the elements is the key-answer, namely the presence of an element (or more elements) and the absence of others in certain areas, from their super-positioning finding as a result the presence of inclusions. The spectrums acquisitioned in point (on a reduced dimension area) have permitted the clarification of the inclusions hypothesis. The GUPIX programme has been used for the analysis of X-ray spectrums obtained at LNL, calculus code especially developed for the filtering of X-ray spectrums, from its progress being obtained the qualitative results (the type of elements contained by the sample), as well as the quantitative ones (concentrations).
As regarding the samples from the Pietroasa Hoard, the results (the mediated elemental composition, maps) obtained for each object have been completely different, confirming the fact that the gold sources and the manufacturing workshops have been distinct for each piece-see Table 1, detailing the concentrations in point/points for every analysed sample. The fact that the results indicate recipes (elemental compositions) completely different for the analysed pieces is accordingly to the object prior classification established on stylistically criteria. Therefore, the patera and the oenoche cup are made in Greco-Roman style; the dodecagonal basket combines characteristics of the sassanid (Persian) style and influences from the art of the steppe; while the fibulae belong to the steppe tradition (the pray bird decor). From the performed analyses result that this hoard is heteroclite, comprising pieces manufactured in different places and epochs. Evidently, only an extensive analyse of the entire lot of pieces from this hoard may lead to conclusive results.
In several samples-as the one taken from the patera from the Pietroasa Hoard-it has been emphasized the iridium, element characteristic to east-Mediterranean gold, most probable coming from Anatolia, from the gold field of the Pactolus river, today exhausted of such auriferous filed.
As concerning the dodecagonal basket, the most important result of these experiments is the presence of palladium (Pd) traces. The only accessible sources of gold with Pd during that time were the ones from Nubia (Sudan) and Anatolia (Turkey), frequently used in the workshops from Egypt (Alexandria) and Syria (Antioch); the palladium has been assayed as trace-element in the coins emitted by Alexander the Great after the conquest of the Persian Empire. The high amount of gold comprised by this sample, also suggest the presence of re-fusion procedure, probably using Roman imperial coins issued in the Oriental provinces.
From the analyse of the elemental maps, in one of the samples, part of the large eagle-headed fibulae, have been found Ta-Nb (tantalum and columbium, see Figure 2, presenting a PIXE spectrum in point with Ta high concentration, where the Nb presence is also visible; Figure 3 represent a map with the Ta distribution on a fragment from the small fibulae). These are hard-fusible metals, resisting thus to the melting process. These elements are specific to the auriferous ores from the Ural Mountains-samarskite/columbite ores type. We may conclude thus that the gold, out of which this piece has been made, has as origin the ores from the Ural Mountains area.
As regarding the Romanian native golden samples, it has been studied especially the presence of specific trace-elements, such as tin, antimony, tellurium, mercury, lead. In the Transylvanian gold, the tellurium and the antimony have been found as trace-elements in all the samples; the lead and the mercury have appeared as trace-element in the samples from Brad area (besides tellurium and antimony)-in very small quantities. We have to remark the fact that it has not been detected in any sample the presence of tin, trace-element specific to the Mediterranean gold, which can be found in several pieces from the Pietroasa Hoard, manufactured in Greco-Roman or Greco-Oriental style. The native Transylvanian gold comprise silver from 5% to 25%, high percentage comparing the analysed archaeological gold pieces, which have a low silver percentage. Regarding these results, it is very clear that the Transylvanian gold does not match the one comprised by the Pietroasa Hoard.
Synthesising the obtained results in the present phase, we may conclude that for the analysed objects, belonging to the Pietroasa Hoard, there are four possible gold sources: the southern region of the Ural Mountains, the deposits from Nubia (Sudan), the micro-Asian source (the most probably the Pactolus river from Central Anatolia) and/or different Roman imperial coin emissions.
It may also be emphasized the fact that the measurements undertaken up to the present have not affected the so-called "germanic" pieces, such as the big pan and the simple necklace. The performed experiments within this phase may represent an example of detailed study of the presence of trace-elements and micro-inclusions in golden patrimony objects, by using the micro-PIXE technique in order to obtain historical conclusions regarding the origin of the studied objects.
The obtained elemental compositions and maps have allowed the establishing of several historical conclusions and the developing of hypothesis regarding the provenience and the techniques of manufacturing of the used metal. It has been also intended a characterisation of the Transylvanian native gold, concluding that the golden objects from the Pietroasa Hoard, analysed so far, have not been manufactured using gold from this geographical area. The performed experiments represent only an initial phase of this extensive project regarding the study of precious metal objects belonging to Romanian cultural heritage.
The described results have been obtained by the IFIN-HH physicists, their historical interpretation being ascertained in collaboration with the historians and specialists from the Romanian National History Museum (MNIR) and the Faculty of History, University of Bucharest (FI-UB).