Fig. 2: Validation of the here-developed MLIPs (MLIP-[0001], MLIP-[10\(\overline{1}0\)], and MLIP-[\(\overline{1}2\overline{1}0\)]) against atomic scale room-temperature AIMD tensile tests. | npj Computational Materials

Fig. 2: Validation of the here-developed MLIPs (MLIP-[0001], MLIP-[10\(\overline{1}0\)], and MLIP-[\(\overline{1}2\overline{1}0\)]) against atomic scale room-temperature AIMD tensile tests.

From: Machine-learning potentials for nanoscale simulations of tensile deformation and fracture in ceramics

Fig. 2

a Comparison of AIMD (dash-dotted line) and ML-MD (solid line) stress/strain curves for TiB2 subject to \([0001],[10\overline{1}0]\), and \([\overline{1}2\overline{1}0]\) tensile deformation at 300 K, using a 720-atom supercell with dimensions of ≈ (1.52 × 1.58 × 2.57) nm3. Only the stress component in the loaded direction is plotted. b, c Snapshots of the fracture point in AIMD (b-1, b-2, b-3) and ML-MD (c-1, c-2, c-3). d Illustration of fracture surfaces (see also ref. 100).

Back to article page