Machine-learning interatomic potential for temperature-dependent properties of Nb2AlC MAX phase as a bond coat

초록

Nb2AlC, a 211-type MAX phase, has attracted attention as a potential bond coat (BC) material for thermal barrier coating (TBC) systems. A BC must buffer the thermal expansion mismatch between the ceramic topcoat and the metallic substrate, making its temperature-dependent thermomechanical properties the key criterion for assessing BC suitability. However, quantitative thermomechanical data for Nb2AlC across the relevant service temperature range remains scarce. In this work, we developed a moment tensor potential (MTP) for Nb2AlC and used it to perform large-scale molecular dynamics (MD) simulations to quantify its thermomechanical behavior over a temperature range of 300-2000 K. The trained potential reproduces density functional theory (DFT) energies and forces with near-DFT accuracy and yields lattice and elastic constants in close agreement with literature values. The Young's modulus decreases monotonically from 242 GPa at 300 K to 179 GPa at 2000 with tensile failure propagating preferentially along the Al-containing layers. The average linear coefficient thermal expansion (8.3-9.4 & times; 10-6 K-1) and its pronounced axial anisotropy are quantitatively captured. These results provide a quantitative basis for assessing the thermomechanical suitability of Nb2AlC as a potential bond coat candidate.

키워드