corundum, sapphire, heat treatments.6, gems

The variable response of Sri Lankan “ottu” geuda corundum to heat treatment is investigated through electron probe microanalysis, with particular attention to unsuccessful blue colour development and the formation of an opaque white surface layer known in the trade as “dead milk”. Approximately 50 initially colourless or light blue stones bearing dark blue surface patches were heated for 1 hour at 1800–1850 °C under reducing conditions. Three outcomes were observed: clear stones with satisfactory blue colour, clear stones with insufficient blue colour, and stones with a thick white coating. More than half developed the coating. Subsequent heating under oxidizing conditions at 1600–1800 °C reduced the blue colour of the first two groups but produced no observable change in coated specimens; the coating persisted even after further heating at higher temperatures. Representative samples were examined using wavelength-dispersive electron probe microanalysis, including polished cross sections comparing blue interiors with white outer regions. Iron and titanium were estimated semiquantitatively against standards, with a reported detection limit of 0.03%, while other impurities were identified qualitatively. Well-coloured and poorly coloured clear stones contained similar iron levels, but the well-coloured group generally had lower titanium concentrations and fewer additional impurities. Sodium, potassium and manganese were the only additional elements detected in that group, and some specimens showed none. Poorly coloured stones contained a broader range, including sodium, potassium, calcium, zinc and silicon. White regions had higher iron and titanium concentrations than their blue interiors, with titanium exceeding iron, and contained a wider range of minor impurities. Vanadium was detected in every examined specimen with “dead milk”, although its concentration was not quantified. Analyses of specimens without the special surface preparation argued against preparation-induced contamination. Interpreting blue colour in terms of iron–titanium pairs, the authors suggest that additional impurities may hinder formation of the colour-producing centres despite the presence of both elements. They associate white-layer formation with elevated impurity concentrations, particularly iron and titanium, and propose that vanadium and other minor constituents may assist the process.