| Abstract Scope |
Phase diagrams are essential tools to guide the synthesis of functional materials, but they are almost never available in the frontier chemical spaces where we perform materials discovery and design. Experimental phase diagram assessment can require years of tedious measurement, whereas computing the high-temperature regions of phase diagrams computationally can be extremely costly. Here, we present a strategy to combine the best of both worlds—using computation to assess the energies of solids, and experiments to assess their phase transformations at high-temperature. By using an appropriate thermodynamic referencing scheme to connect their energies, we can rapidly assess phase diagrams and non-equilibrium Gibbs free-energies in complex, multi-component phase spaces. We conclude with a vision for a unified computational and robotic experimental platform that consolidates materials synthesis and processing, robotic calorimetry, ab initio thermodynamics, CALPHAD thermodynamic modeling, and artificial intelligence/machine learning (AI/ML), to generate phase diagrams on demand. |