Environmental Stability of Crystals: Why Be Greedy When You Can Be Exact?

In a recent paper in this journal (Chem. Mater. 2022, 34, 2545-2552), Twyman et al. studied the environmental stability of crystals by introducing a greedy heuristic algorithm for determining possible oxidation reactions. We show how the problem can be solved exactly, with less code and comparable computational time by reformulating it as a linear optimization for the reaction enthalpy. The enthalpy of oxidation and the oxidation products, computed by Twyman et al.’s greedy heuristic is suboptimal in 22504 of 39634 cases (56.8%), with the error in enthalpy of oxidation as large as 1.480 eV/atom. Every one of these quantitative errors also results in qualitative differences in the oxide species formed. Using our exact approach to the problem, we re-evaluate the core results from the paper by Twyman et al. We also describe a minor modification enabling calculations of oxidation free energy minima.