2026/06/24 by Lauren N. Walters, Anubhav Jain, Gerbrand Ceder · 1 voice
Earth and Planetary Sciences · Materials Science · #Electronic and Structural Properties of Oxides #High-pressure geophysics and materials #Machine Learning in Materials Science
paper · doi:10.1021/acs.jpcc.6c01086
openalex publication_date 2026/06/24 · openalex created_date 2026/06/25 · openalex updated_date 2026/07/30
Connecting 0 K density functional theory (DFT) energies to finite-temperature, finite-pressure synthesis conditions is a well-established thermodynamic formalism, yet quantified guidance on when and how to accurately calibrate these results to experimental chemical potentials in practice remains sparse. In this work, we systematically benchmark a complete workflow─the virtual furnace─that constructs effective oxygen chemical potentials (μ O2 ) for common synthesis atmospheres (air, Ar, H 2, CO) as functions of temperature and partial pressure and propagates quantified errors from formation enthalpies through reaction energies to critical chemical potentials. Applying this workflow to 11 binary oxides, we find that “gas-only” thermal corrections are sufficient at low temperatures and for Group II oxides across all temperatures, while solid-phase vibrational contributions become critical at elevated temperatures for transition metal oxides. We quantify this threshold through the ratio |Δ S solid /Δ S total | as a practical guide for when phonon calculations are warranted. Predicted critical reduction temperatures and oxygen chemical potentials show strong agreement with industrial practice and experimental data, with typical errors of 150–250 °C for most oxides. This benchmarked workflow provides practitioners with explicit, quantified guidance for predictive modeling of phase stability and synthesis condition design.