Chemical Composition, Thermodynamics, and Recycling: The Beginnings of Predictive Behavioral Modeling for Ancient Copper-based Systems

In their attempts to understand the unwritten past of human technology and progression, archaeologists have borrowed aspects of the natural sciences to answer big questions. In one such pursuit, fundamental aspects of the sciences have been employed towards the chemical compositional analysis of copper-based artifacts, often to simply classify which is bronze, brass, or pure copper, and to explain why they are significant in limited space and time. This thesis takes the variety of identified metal types and compositions from these analyses and builds the beginnings of an ambitious thermodynamic model based on the accepted premise of consistent and widespread recycling of ancient metals over time. Following the laws of thermodynamics, in systems at equilibrium, the model predicts the outcome of metal losses over the course of ancient pyrometallurgical processes from molten systems through both volatilization and oxidation using rigorous and established mathematics and theory. Elemental loss likelihoods are modeled for all binary copper-based metals, using activity coefficients, and ternary copper and zinc-based systems, with the excess Gibbs free energy, respectively. The calculations are performed using custom-written software designed to account for hundreds of thousands of compositional permutations after the method described by Redlich and Kister (1948). The results of these calculations are given as activity (binary) and isoactivity (ternary) contour lines. Quantified tables for the oxidation and volatilization of elements from a copper melt at 1200 ºC and 1 atm are also given as rough indicators of element loss in ancient pyrometallurgical systems. A proof of concept of the models viability is also provided for binary Cu-M and ternary Cu-M-Zn (M = Ag, As, Au, Bi, Co, Fe, Ni, Pb, Sb, Sn, Zn), Cu-Sn-Pb, and Cu-Sb-As systems from the Late Bronze Age to post-medieval periods in Britain, which is based on several substantial artifact chemical datasets. For each ternary system, the interaction parameters used for higher-order calculations from the fitted behavior of each contributing binary systems are provided. Comparison of the calculated models to available experimental system assessments, and to published archaeological chemical datasets, show that in both respects the proposed modeling of ancient copper-based metal losses works. And given the near ubiquity of ancient metal use around the world, the consistency in metal production and recycling technology, and the chemical analyses available, this preliminary model can be applied virtually anywhere the technology for smelting and recycling existed. In addition to loss modeling, this thesis has the additional offshoots of predicting ancient furnace conditions based on the calculated behavior of interacting metals, and of the controlling thermodynamic factors in the ancient calamine process