Quantitative analysis of copper alloys by means of portable X-ray fluorescence: A comparison between analysis of shavings and surfaces

The investigation of the alloy composition is a key point for studying archaeological and historical metal artefacts. Portable X-ray fluorescence is the most commonly used technique to perform non-invasive alloy analysis. Nevertheless, quantitative results are usually not reliable when metal surface is covered by either corrosion or patination layers. In such cases, micro samples, representative of the bulk alloy, have to be taken. Shavings samples are a good compromise between representativeness and degree of invasiveness when the purpose of the analysis is to obtain accurate quantitative data. This work aims at evaluating and comparing accuracy and precision of X-ray Fluorescence quantitative data obtained by performing analysis on the surface and on shavings of certified reference materials. Results point out that quantitative data obtained from surface and shavings samples have comparable accuracy and precision if a suitable calibration model using certified reference materials in the same form, i.e. surface or shavings, is applied. Repeatability tests were also carried out using the same certified reference materials. Results proved repeatable within 5% for all the detected elements. Uncertainties are slightly higher for shavings with respect to surface measurements. Two case studies concerning an archaeological and a Renaissance bronze artefacts are also discussed.

Automated summary

The investigation of alloy composition is crucial for studying archaeological and historical metal artifacts. Portable X-ray fluorescence is commonly used for non-invasive alloy analysis, but its quantitative results are not reliable when the metal surface is covered by corrosion or patination layers. In such cases, micro samples representative of the bulk alloy need to be taken. This study evaluates and compares the accuracy and precision of X-ray fluorescence quantitative data obtained from surface and shavings samples, and suggests that a suitable calibration model using certified reference materials should be applied. Repeatability tests show consistent results within 5%, with slightly higher uncertainties for shavings compared to surface measurements. Two case studies of archaeological and Renaissance bronze artifacts are also discussed.