X-Ray Fluorescence analysis (XRF) at the Curt-Engelhorn-Zentrum Archäometrie
All samples extracted from the blades are measured by energy dispersive X-ray fluorescence spectroscopy (ED-XRF) with a Spectro XEPOS HE spectrometer at the Curt-Engelhorn-Zentrum Archäometrie. Prior to analysis, each sample is checked under a microscope and freed for pieces of corrosion that may negatively influence the analytical results. Afterwards, the samples are measured for their chemical composition along with copper-alloy reference materials for quality control.
The Sögel-Wohlde blades are made of bronze, a metal alloy consisting predominantly of copper (Cu) and tin (Sn). Beyond these main elements, prehistoric metal objects often contain additional minor and trace elements that are related to the ores, which were used to produce the metal. Copper ores often contain impurities of various elements such as nickel (Ni), cobalt (Co), arsenic (As), silver (Ag), lead (Pb), antimony (Sb) and bismuth (Bi), which can be used to narrow down the origin and to identify metal groups with chemical similarities. Therefore, the chemical information can be instrumental to identify blades with similar compositions and as such, make statements about potential metallurgical traditions, the identification of workshops, metal trade and the ores used to produce an artefact. The chemical composition of the artefacts is obtained by measuring metallic samples by energy EDXRF, which is based on the measurement of the so-called characteristic X-radiation emitted by a material that is exposed to X-rays.
The principle of XRF is based on the atomic structure of elements, with each element having a unique atomic number (Z) which represents the number of protons in a nucleus. These protons, together with neutrons make up most of the atomic mass in a nucleus. Furthermore, the nucleus is surrounded by negatively charged electrons equal to the number of protons, which are at specific distances from the nucleus. According to Bohr atom model, these electrons are present in shells at specific distances from the nucleus in a unique electron shell configuration. The shell closest to the nucleus (K-shell) holds electrons that are most strongly bound to the nucleus due to its proximity, while further shells (L- and M-shells) have a weaker bond to the nucleus. This structure is of particular importance for X-ray fluorescence as the emittance of electrons by a primary X-ray beam from the K, L etc. shells by X-rays creates vacancies in the electron shell configurations. As a result, the vacancies in the electron shell configuration are filled by electrons from outer shells to ensure electro neutrality. The difference in energy between the shells is released as X-ray photons of a characteristic energy or wavelength. These emitted X-ray photons, which can be measured, have a specific energy signature, which is related to the element it derives from whereas the intensity of the emission is related to the abundance of that element. By comparing the sample with reference materials of known chemical composition, the composition of the sample can be determined quantitatively.
Limits of XRF-analysis on archaeological bronzes
The analysis of archaeological bronze artefacts, however, has the added difficulty that most of these artefacts have become corroded or patinated over time, meaning that the surface is degraded or contaminated with other elements from the surrounding environment and, more severely, that elements have been enriched or depleted due to the different solubilities of the resulting mineral compounds. For instance, corrosion products of copper are much easier dissolved under burial conditions than those of tin leading to a relative enrichment of tin and a depletion of copper compared to the uncorroded metal. The corrosion complicates in particular in situ XRF analyses (without sampling) because the measured surface is rarely representative for the original uncorroded bulk material. Additionally, the method has a limited depth of penetration, meaning that only the upper micrometres of the artefact are measured. For copper alloys, the spectral information for copper and tin come from the upper 20 micrometres. Corrosion crust on bronzes, however, may easily reach several tens and hundreds of micrometres in thickness and as such limit the accuracy of XRF analysis. Thus, drilled samples are preferred for archaeometallurgical investigations as the chemical but also the tin and copper isotope composition of patinas is often not representative for the original metal. Without sampling, no useable results concerning the project’s aims are obtained.