Finite-temperature disorder physics in high-entropy carbides: collective response and carbon diffusion.

Abstract

Chemical disorder in rocksalt (Hf,Zr,Ta,Nb,Ti)C produces heterogeneous local environments that can affect collective thermomechanical response and activated carbon transport differently. We examine this contrast using the unmodified MACE-MP-0 potential, with explicit limits on its transferability. Binary optical phonons are 4-16 meV softer than density-functional perturbation theory, multicomponent lattice parameters exceed density functional theory (DFT) values by 0.14-0.31%, and binary melting estimates deviate from experiment by -21% to +35%. Thermal expansion of nineteen compositions is nearly additive (α = 5.4-6.7 × 10-6K-1, within ±5% of the binary rule of mixtures), and elastic softening to 2000 K is mode-selective (C11decreases by 18-37%, C44by 3-7%). Along a group-IV-rich to group-V-rich composition path, the nominal two-phase melting estimates rise by about 1000 K. A linear composition weighting of the binary errors leaves about 460 K in the same direction, but this sensitivity estimate does not establish robustness to unmeasured multicomponent bias. Twenty static single-vacancy carbon-migration paths have a mean barrier of 2.53 ± 0.40 eV. At 12.5% carbon vacancies, fitted MD Arrhenius barriers are 1.36-1.54 eV, below the sampled ensemble minimum of 1.89 eV. These point estimates do not distinguish local-environment effects from vacancy concentration, correlated motion, finite cell size, or limited diffusion statistics. A matched DFT comparison finds a MACE barrier underestimate of 1.22 eV. The results identify near-additive expansion and mode-selective collective softening, while showing why a static barrier ensemble alone cannot determine the interpretation of finite-concentration carbon transport.

Abstract as supplied by the indexing source. © the authors / publisher.