This article explores the development of a machine learning interatomic potential (MLIP) with density functional theory (DFT) level accuracy for modeling grain boundaries (GBs) in α-Fe, focusing on general grain boundaries (GGBs) that typically present complex atomic structures. Conventional methods using symmetric tilt GBs often fall short of capturing the diverse environments in GGBs, leading to inaccuracies in energy and structural predictions. The authors used an innovative approach, employing a diverse dataset generated by crystal space groups and moment tensor potential (MTP) techniques, to model GGB structures without needing extensive DFT computations. The study demonstrated that the MTP model accurately estimates average GB energy and captures the nuanced atomic and energy variations in polycrystals, allowing for more reliable predictions in polycrystalline materials, which is critical for advanced materials design aimed at carbon neutrality. This work paves the way for MLIP applications in simulating polycrystalline properties and mitigating GB-related issues like hydrogen embrittlement in high-strength materials.
For more details, please continue reading the full article under the following link:
https://www.nature.com/articles/s41524-024-01451-y
Please consult also the Quantum Server Marketplace platform for the outsourcing of computational science R&D projects to external expert consultants through remote collaborations:
#materials #materialsscience #materialsengineering #computationalchemistry #modelling #chemistry #researchanddevelopment #research #MaterialsSquare #ComputationalChemistry #Tutorial #DFT #simulationsoftware #simulation