Deep-Learning Enabled Atomistic Understanding of Thermomechanical Behaviors and Fracture Mechanisms of High-Entropy...

High-entropy transition-metal diborides represent a promising class of ultra-high temperature ceramics. However, atomic insights into their high-temperature elastic response, anisotropic deformation, and fracture mechanisms remain elusive. Herein, we perform molecular dynamic simulations to study the thermomechanical behaviors of (Hf_0.2Zr_0.2Ta_0.2Ti_0.2Nb_0.2)B₂ from 900 to 3300 K by developing an ab initio accuracy deep-learning potential. The proposed potential accurately reproduces lattice parameters, equations of state, and elastic constants, in excellent agreement with density functional theory calculations and available experiments, and remains transferable under thermally expanded and compressed states. The simulations reveal anisotropic thermal expansion, with the out-of-plane expansion exceeding the in-plane expansion, together with progressive elastic softening while preserving C₁₁ > C₃₃ due to the dominant in-plane B-B bonding network. Furthermore, strain-rate- and temperature-dependent tensile and compressive responses show marked crystallographic anisotropy, tension-compression asymmetry, and severe thermomechanical degradation. Atomic structural evolution demonstrates that tensile fracture is dominated by bond stretching and progressive damage accumulation, whereas compressive failure is attributed to densification- and shear-mediated structural instability. These findings provide an atomistic understanding of the thermomechanical behavior and fracture mechanisms of the prototypical single-phase (Hf_0.2Zr_0.2Ta_0.2Ti_0.2Nb_0.2)B₂ high-entropy diboride, offering valuable insights into the design of ultra-high temperature ceramics under extreme service environments.