Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Development and transferability of neural-network models for plasma-surface interactions

Journal Article · · Journal of Vacuum Science and Technology A
DOI:https://doi.org/10.1116/6.0004843· OSTI ID:2999374

Plasma-surface interactions are increasingly critical to modern technologies; yet, accurate molecular dynamics simulations remain limited by the capabilities of interatomic potentials. Deep Potentials (DPs) promise to revolutionize the field by providing a systematic method for producing accurate interatomic potentials. The primary challenge of DP development is selecting a dataset, which efficiently spans the set of atomic environments one expects to encounter in the subsequent molecular dynamics simulations. The computational cost of density functional theory calculations, which are the typical basis for DP development, makes it impossible to directly verify the quality of a given DP. To address this challenge, we explore the development of a deep-learned interatomic potential, “DeepREBO,” trained to reproduce the behavior of the REBO2 empirical potential, enabling direct validation of training methodology and transferability. Using an active learning framework, we begin with a minimal dataset and iteratively expand it to train a Deep Potential-Smooth Edition model that faithfully reproduces REBO2 results for 25 eV hydrogen bombardment of diamond (001), a particularly challenging case. We show that small, carefully curated datasets can outperform large, unguided ones, with effective models requiring fewer than 15 000 snapshots. Subsequent transferability tests demonstrate that while DeepREBO generalizes well to diamond (111) surfaces, performance degrades for amorphous carbon or higher-energy impacts, highlighting the need for use-case-specific training data. We also evaluate methods to improve short-range repulsion. This study outlines best practices for training robust deep potentials and underscores the importance of dataset design for predictive plasma simulations.

Research Organization:
Princeton Plasma Physics Laboratory (PPPL), Princeton, NJ (United States)
Sponsoring Organization:
USDOE Office of Science (SC), Fusion Energy Sciences (FES)
Grant/Contract Number:
AC02-09CH11466
OSTI ID:
2999374
Journal Information:
Journal of Vacuum Science and Technology A, Journal Name: Journal of Vacuum Science and Technology A Journal Issue: 6 Vol. 43; ISSN 1520-8559; ISSN 0734-2101
Publisher:
American Vacuum Society / AIPCopyright Statement
Country of Publication:
United States
Language:
English

References (27)

DP-GEN: A concurrent learning platform for the generation of reliable deep learning based potential energy models journal February 2020
LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales journal February 2022
Reactive ion etching of single crystal diamond by inductively coupled plasma: State of the art and catalog of recipes journal October 2020
Isotope dependence of chemical erosion of carbon journal June 2010
Molecular Simulation of Lithium Carbonate Reactive Vapor–Liquid Equilibria Using a Deep Potential Model journal December 2023
DP Compress: A Model Compression Scheme for Generating Efficient Deep Potential Models journal August 2022
Atomistic Simulation of HF Etching Process of Amorphous Si3N4 Using Machine Learning Potential journal August 2024
ReaxFF:  A Reactive Force Field for Hydrocarbons journal October 2001
SpookyNet: Learning force fields with electronic degrees of freedom and nonlocal effects journal December 2021
Deep learning inter-atomic potential model for accurate irradiation damage simulations journal June 2019
DeePMD-kit v2: A software package for deep potential models journal August 2023
Simulation of lithium hydroxide decomposition using deep potential molecular dynamics journal October 2024
A molecular dynamics study of H radical bombardment of CH3 : Si(1 0 0)— comparison of simulation and experiment journal October 2009
A second-generation reactive empirical bond order (REBO) potential energy expression for hydrocarbons journal January 2002
Representing potential energy surfaces by high-dimensional neural network potentials journal April 2014
Atomic scale etching of diamond: insights from molecular dynamics simulations journal October 2024
A transient site balance model for atomic layer etching journal July 2024
keV particle bombardment of semiconductors: A molecular-dynamics simulation journal July 1989
Describing bond-breaking processes by reactive potentials: Importance of an environment-dependent interaction range journal October 2008
Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics, without the Electrons journal April 2010
Active learning of uniformly accurate interatomic potentials for materials simulation journal February 2019
Origins of Diamond Surface Noise Probed by Correlating Single-Spin Measurements with Surface Spectroscopy journal September 2019
Molecular dynamics simulations of Ar+ bombardment of Si with comparison to experiment
  • Humbird, David; Graves, David B.; Stevens, A. A. E.
  • Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, Vol. 25, Issue 6 https://doi.org/10.1116/1.2787713
journal November 2007
Future of plasma etching for microelectronics: Challenges and opportunities journal June 2024
Reactor wall effects in Si–Cl2–Ar atomic layer etching journal June 2024
Deep potential molecular dynamics simulations of low-temperature plasma-surface interactions
  • Kounis-Melas, Andreas; Vella, Joseph R.; Panagiotopoulos, Athanassios Z.
  • Journal of Vacuum Science & Technology A, Vol. 43, Issue 1 https://doi.org/10.1116/6.0004027
journal January 2025
Theorie der Streuung schneller geladener Teilchen I. Einzelstreuung am abgeschirmten Coulomb-Feld journal March 1947