Self-Learning Method for Construction of Analytical Interatomic Potentials to Describe Laser-Excited Materials

Area of Science:

• Materials Science
• Computational Physics
• Chemistry

Background:

• Large-scale simulations offer insights into solid-state processes under external perturbations.
• Atomistic simulations of laser-induced ultrafast nonthermal phenomena are challenging due to limitations in ab initio methods and electronic temperature-dependent potentials.

Purpose of the Study:

• To develop a method for constructing accurate electronic temperature-dependent interatomic potentials.
• To enable ultralarge-scale atomistic simulations of systems in extreme nonthermal states.
• To investigate laser excitation effects on silicon nanoparticles.

Main Methods:

• A self-learning method was developed to construct electronic temperature- (T_e) dependent interatomic potentials.
• The method ensures global minimum identification in parameter space for high accuracy.
• Density-functional theory (DFT) accuracy was achieved for the potentials.
• An analytical T_e-dependent potential, Φ(T_e), was derived for silicon.

Main Results:

• The derived Φ(T_e) accurately describes laser-excited and unexcited silicon bulk and films.
• Ultralarge-scale atomistic simulations were performed on laser-excited silicon nanoparticles.
• Strong damping of breathing modes in silicon nanoparticles was observed.
• Nonthermal melting was identified as the cause of mode damping.

Conclusions:

• The developed method enables accurate, large-scale simulations of nonthermal phenomena.
• The study provides a detailed understanding of laser-matter interactions in silicon.
• Nonthermal melting significantly impacts the dynamics of excited silicon nanoparticles.