Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Efficient molecular dynamics simulation of 2D penta-silicene materials using machine learning potentials

Le Huu Nghia, Pham Thi Bich Thao, Truong Do Anh Kha, Vo Khuong Dien, Nguyen Thanh Tien

TL;DR

This paper addresses the challenge of achieving quantum-mechanical accuracy in large-scale MD simulations of 2D penta silicene by developing a DeepMD-kit MLIP trained on ab initio data and benchmarking it against the Tersoff.SiC potential. The authors demonstrate that the MLIP reproduces phonon spectra, structural stability, and the melting transition with $T_g$ around 606–632 K, closely matching AIMD/DFT references, while the classical potential overestimates stability. They further validate an on-the-fly ML force field during AIMD, supporting efficient, accurate MD at elevated temperatures. The results highlight the potential of MLIPs to enable large-scale, high-precision simulations of 2D materials, with implications for experimental synthesis and future applications of penta silicene and related systems.

Abstract

Machine Learning Interatomic Potentials (MLIPs) are a modern computational method that allows achieving near-quantum mechanical accuracy (DFT) while still describing large-scale systems in molecular dynamics (MD) simulations. In this work, we use MLIP from DeepMD package and the classical Tersoff potential for SiC (Tersoff.SiC potential) to fully and accurately describe atomic interactions and apply them to molecular dynamics simulations of penta silicene sheet. The results show that the melting points (T$_g$) temperatures of the system in the canonical NVT and isobaric NPT sets are 632 K and 606 K, while the Tersoff.SiC potential have the high melting points, respectively. In addition, the radial distribution function exhibits characteristic peaks at interatomic distances of 2.275 Å\text{} and 2.375 Å, while the Tersoff.SiC potential only describe distance of 2.375 Å. Furthermore, penta silicene was also simulated using on-the-fly machine learning for 10 ps to evaluate the structural stability of the system. This study investigates the thermodynamic properties of two-dimensional penta silicene sheets with pentagonal structures using a high-precision, cost-effective method, contributing further evidence to support experimental synthesis and opening up potential future applications of this material.

Efficient molecular dynamics simulation of 2D penta-silicene materials using machine learning potentials

TL;DR

This paper addresses the challenge of achieving quantum-mechanical accuracy in large-scale MD simulations of 2D penta silicene by developing a DeepMD-kit MLIP trained on ab initio data and benchmarking it against the Tersoff.SiC potential. The authors demonstrate that the MLIP reproduces phonon spectra, structural stability, and the melting transition with around 606–632 K, closely matching AIMD/DFT references, while the classical potential overestimates stability. They further validate an on-the-fly ML force field during AIMD, supporting efficient, accurate MD at elevated temperatures. The results highlight the potential of MLIPs to enable large-scale, high-precision simulations of 2D materials, with implications for experimental synthesis and future applications of penta silicene and related systems.

Abstract

Machine Learning Interatomic Potentials (MLIPs) are a modern computational method that allows achieving near-quantum mechanical accuracy (DFT) while still describing large-scale systems in molecular dynamics (MD) simulations. In this work, we use MLIP from DeepMD package and the classical Tersoff potential for SiC (Tersoff.SiC potential) to fully and accurately describe atomic interactions and apply them to molecular dynamics simulations of penta silicene sheet. The results show that the melting points (T) temperatures of the system in the canonical NVT and isobaric NPT sets are 632 K and 606 K, while the Tersoff.SiC potential have the high melting points, respectively. In addition, the radial distribution function exhibits characteristic peaks at interatomic distances of 2.275 Å\text{} and 2.375 Å, while the Tersoff.SiC potential only describe distance of 2.375 Å. Furthermore, penta silicene was also simulated using on-the-fly machine learning for 10 ps to evaluate the structural stability of the system. This study investigates the thermodynamic properties of two-dimensional penta silicene sheets with pentagonal structures using a high-precision, cost-effective method, contributing further evidence to support experimental synthesis and opening up potential future applications of this material.
Paper Structure (14 sections, 5 equations, 17 figures, 1 table)

This paper contains 14 sections, 5 equations, 17 figures, 1 table.

Table of Contents

  1. Introduction
  2. Computational details
  3. Machine Learning Interatomic Potential
  4. MLIP in molecular dynamics simulations
  5. On-the-fly machine learning force field
  6. Results and discussions
  7. Structure and electronic properties
  8. Thermodynamic properties
  9. Training data for MLIPs
  10. Phonon dispersion
  11. Relaxation and system stability
  12. The phase transition of penta silicene
  13. On-the-fly machine learning force fields
  14. Conclusions

Figures (17)

  • Figure 1: a) The optimized structure of penta silicene monolayer is shown in both top and side view. b) The first Brillouin zone. c) The phonon dispersion of the penta silicene monolayer is presented based on DFT calculations. d) Electronic band structure calculated using the PBE functional.
  • Figure 2: The phononLAMMPS calculations by two potentials: a) Tersoff.SiC potential. b) MLIP with a 4x4x1 supercel.
  • Figure 3: a) System temperature under NVT and NVE ensembles. b) Radial distribution function at 10 ps. c) Radial distribution function at 20 ps. d) Bond length between sp$^2$-sp$^2$ and sp$^2$-sp$^3$ hybridization.
  • Figure 4: The phase transition (T$_g$) of penta silicene with MLIP.
  • Figure 5: The compared phonon dispersion DFT and MLFF.
  • ...and 12 more figures