X-ray fluorescence spectroscopy is a non-destructive technique employed for elemental analysis of a wide range of materials. Its advantages are especially valued in archaeometry, where portable instruments are available. Considering ancient glass, such instruments allow for the detection of some major, minor, and trace elements linked to the deliberate addition of specific components or to impurities in the raw materials of the glass batch. Besides some undoubted advantages, portable XRF (p-XRF) has some limitations that are addressed in this study. The performance assessment of four different p-XRF units and the reconciling of their output were conducted. The results show the limitations in cross-referencing the data obtained from each unit and suggest procedures to overcome the issues. The p-XRF units were tested on the set of Corning reference glasses and on a small set of archaeological glasses with known composition. The compatibility of the output was assessed using multivariate statistical tools. Such a workflow allows us to consider data from multiple sources in the same frame of reference.
Keywords: ancient glass; composition; portable-XRF; reproducibility.