Nanohardness from First Principles with Active Learning on Atomic Environments

• DOI: 10.1021/acs.jctc.1c00783
• Publication Date: 2022-01-06T18:53:03Z
• Authors: Evgeny V. Podryab...; Alexander G. Kvas...; Milad Asgarpour; Igor I. Maslenikov; Danila A. Ovsyann...; Pavel B. Sorokin; Mikhail Yu Popov; Alexander V. Shapeev
• Keywords: Macro

We propose a methodology for the calculation of nanohardness by atomistic simulations nanoindentation. The is enabled machine-learning interatomic potentials fitted on fly to quantum-mechanical calculations local fragments large nanoindentation simulation. test our calculating nanohardness, as function load and crystallographic orientation surface, diamond, AlN, SiC, BC2N, Si comparing it calibrated values macro- microhardness. observed agreement between computational experimental results from literature provides evidence that method has sufficient predictive power open up possibility designing materials with exceptional hardness directly first principles. It will be especially valuable at nanoscale where measurements are difficult, while empirical models macrohardness are, rule, inapplicable.