Calibrating Thermodynamic Models to Experimental Conditions for Accurate Solid State Synthesis Prediction

Accurate, calculated formation energies are necessary for predictive modeling of solid state synthesis of energy materials. Previous studies have extensively benchmarked workflows and corrections to 0K enthalpies of formation generated from first principles calculations. However, little is reported on standards and errors of calculated free energies of formation, and associated critical reaction temperatures, for modeling solid state synthesis. Here, we present methods and considerations for constructing a “virtual furnace”. We fit solid phase enthalpies from density functional theory and common gaseous atmospheres for finite temperature models. Errors associated with critical thermodynamic stability changes are benchmarked for different adjustment schema. We validate our choice of thermodynamic models through comparison with solid state electrolyte synthesis results within our self-driving laboratory, A-Lab. Application of our models to carefully selected redox experiments allow us to explore the possibility of required underpotentials for achieving reduction. Our methods unite traditional Gibbs reaction energy, grand potential diagram, and Ellingham diagram formalism to (automated) experimental materials formation to enable a future of optimizable synthesis-by-design of battery materials.