Using handheld pXRF to study medieval stained glass: A methodology using trace elements

The surfaces of 30 pieces of glass from panel 3b of the Great East Window of York Minster (1405-1408 CE) were analyzed by handheld portable X-ray fluorescence (pXRF) and small samples from the same pieces were analyzed by electron microprobe (EPMA). Comparison of the two methods reveals significant divergences which are not systematic, particularly for elements lighter than Ti. Rather than a problem with pXRF calibration or correction software, the non-systematic error is attributable to the presence of a thin surface layer of weathered glass. Analysis of the depths of X-ray generation indicate that virtually all X-rays characteristic of Ca and K are generated within the top 50 mu m of the glass. However, for heavier elements such as Rb, Sr and Zr, most emitted X-rays are generated below 100 mu m. Using pXRF data for the heavier elements, it is possible to replicate the compositional groupings identified by quantitative EPMA. White glass in the window is likely to have originated in England, while colored glasses were probably obtained from the Continent. The alkali contents of the green and yellow glasses appear to have been manipulated to generate their colors. Glass which is medieval in technology but not original to the panel was identified. In particular, zirconium proved a useful indicator of glassmaking regions, and rubidium and strontium were more sensitive to differences between batches, which has interesting implications for future work.