|
Scientific analysis of historic glass objects is needed to enhance our understanding of glass evolutions regarding sourcing and production technology and to distinguish fake from authentic. Furthermore, glass researchers often face the challenge of performing non-destructive analytical methods on site, using portable devices. A recent focus of our research concerned the identification of optical signatures to distinguish various post-medieval glass families. UV-Vis-NIR absorption spectroscopy has proven to be extremely useful as a non-destructive technique for discerning two subgroups of Ca-rich glass (High Lime-Low Alkali or HLLA glass). However, the period of production of Ca-rich glass coincides with other glass composition groups, a pre-requisite is to first exclude non-HLLA materials. To accomplish this, we developed a methodology utilising portable X-ray fluorescence spectrometry (p-XRF). Despite p-XRF’s limited detection capabilities for elements with low atomic numbers, we successfully demonstrated its usefulness as a screening technique for distinguishing different glass types, including Roman Na-rich glass, Industrial Na-rich, K-rich glass and Ca-rich glass. The main rationale behind clustering historic glasses into different chemical groups, is to use heavy elements as proxies for lighter ones. After designing an instrument-specific 3D-printed attachment, we calibrated the recorded p-XRF data against multiple glass standards. Subsequently, we verified the reliability of the empirical calibration for relevant glass major and trace elements by testing it on a sample set whose composition was determined using Laser Ablation Inductively Coupled Mass Spectroscopy (LA-ICP-MS). The defined workflow showed a strong correlation between the corrected p-XRF and the LA-ICP-MS data for the most important light elements (K, Ca, Mn, and Fe), and the heavier glass trace elements (Rb, Sr, and Zr) with R2 values exceeding 0.95. Finally, we defined a flowchart based on the calibrated elements to identify the glass composition groups.
|
