Towards better reconstruction of smelting temperatures: Methodological review and the case of historical K-rich Cu-slags from the Old Copper Basin, Poland.

Temperature is a crucial factor influencing the effectiveness of metal smelting process. When studied in historical slags it provides information on the evolution of smelting technologies over time and how humankind approached metal recovery. Methodological approaches to temperature reconstruction using pyrometallurgical slags vary and this study serves as a critical review of the methods used so far with the complex historical smelting site used as an example of the application of the methods. The example includes silica-undersaturated, potassium-rich Cu slags from four sites in the Old Copper Basin (Stare Zagłębie Miedziowe), Poland, and is of particular interest because (a) K-rich slag compositions are not widely found and this provides additional challenge to temperature reconstruction approach and (b) the site provides material of different ages (18th and 19th century) and, therefore, gives insight into potentially different pyrometallurgical processes. Several methods are compared and include phase diagrams, geothermometry (Cpx-WR), thermodynamic modelling with MELTS software and temperature XRD analyses with the main implication that most of the methods have their merits, but modelling by MELTS software returns the most reliable estimates of liquidus temperature. In fact, MELTS is the only method that returns differences of 200 °C in reconstructed temperatures between the studied sites, which signifies a change in smelting technique probably related to the change in input material. More importantly, MELTS may also be used to estimate crystallization interval (temperature difference between solidus and liquidus) that together with liquidus temperature gives insight into metal extraction efficiency for a given site and provides information on society advancement in metallurgy. The software can also be used cost-effectively for a large number of samples and its combination with experimental approach (temperature XRD analyses) provides the most promising tool to reconstruct differences in smelting technologies of ancient and historical metallurgical sites. • Liquidus temperatures are best estimated by thermodynamic modelling. • Modelling shows variable efficiency in achieving eutectic composition. • Geothermometers often return cooling rather than liquidus temperatures. • Estimated temperatures for the slags studied range from 1215 to 1400 °C. • Earliest smelting had low metal extraction because of less effective ore-flux mixing. [ABSTRACT FROM AUTHOR]