Programme

O.5 -- Oral, VCTP-51

Date: Monday, 3 August 2026

Time: 15:00 - 15:20

Helium Transport in Silicon Channels: A Coupled Study Using DFT, Molecular Dynamics, Machine Learning and Monte Carlo Simulations

Quy-Dong To, Céline Léonard, Christian Soize

Laboratoire MSME, Université Gustave Eiffel, France

Helium and silicon are two important materials in nanotechnology. Semiconductors are primarily made of silicon, and helium is known for its excellent cooling capacity. Modeling and characterizing He–Si interactions at the atomic scale are crucial for understanding helium transport in silicon micro- and nanochannels. It is well known that fluid flows in extremely confined channels are complex and strongly influenced by adsorption/desorption processes and surface diffusion mechanisms at the boundary walls. In this work, density functional theory (DFT) calculations are performed for a system composed of a helium atom and a silicon cluster, from which the interatomic pair potential for this material system is derived. Next, molecular dynamics (MD) simulations of helium atoms impacting a silicon substrate at different temperatures are carried out. Based on the resulting collision data—including velocity, residence time, and surface displacement—a surrogate stochastic wall model is constructed using probabilistic Machine Learning (ML) approaches [1]. Since the data are generated under equilibrium conditions, special attention is paid to preserving the equilibrium distribution of classical particle velocities and ensuring compliance with the principle of time reversibility [2]. The stochastic model is then used in Monte Carlo (MC) simulations of Knudsen diffusion for gas particles traveling between two parallel walls. It is also employed to determine interfacial coefficients, such as velocity slip and temperature jump coefficients, for fluid flows in the continuum regime based on the Navier–Stokes–Fourier equations [3]. [1] Soize C., To Q.D. (2024) Polynomial-chaos-based conditional statistics for probabilistic learning with heterogeneous data applied to atomic collisions of Helium on graphite substrate. Journal of Computational Physics, 496, pp.112582. [2] To Q.D., Soize C., (2025) Size effects of rarefied gas flows in nanochannels by simulations based on statistical surrogate model for atomistic collisions. Physical Review E, 005300. [3] M. Liao, Q.-D. To, C. Léonard, and V. Monchiet (2018) Nonparametric wall model and methods of identifying boundary conditions for moments in gas flow equations, Phys. Fluids 30, 032008.

Presenter: To Quy Dong