A Scalable Time-Based Molecular Dynamics Approach for Simulating Single-Bubble Sonoluminescence

Abstract

We present a scalable time-based molecular dynamics (TBMD) framework for simulating single-bubble sonoluminescence within a hybrid continuum-MD formulation. Unlike prior event-based approaches, which model gas dynamics through instantaneous hard-sphere collisions, the present method integrates continuous Lennard-Jones and damped shifted force Coulomb interactions at each timestep, enabling self-consistent tracking of ionization state and long-range electrostatics throughout the collapse. To bridge the gap between the physical particle count ($N_\mathrm{real}\sim 10^{10}$) and computationally tractable ensemble sizes, we introduce an ensemble particle (EP) scaling formalism that preserves temperature, pressure, and ionization statistics while reducing the simulated particle count by up to four orders of magnitude. Applying the framework to argon under standard single-bubble sonoluminescence driving conditions, we perform a systematic sweep over the ionization model and thermal accommodation coefficient $α_t$, with ensemble sizes up to $N_\mathrm{ensem} = 10^8$ particles. The results establish that ionization is the dominant regulator of peak temperature, reducing $T_\mathrm{max}$ by approximately a factor of two relative to the non-ionizing baseline, while $α_t$ primarily controls the spatially averaged temperature at the collapse minimum. Scalar observables at $N_\mathrm{ensem} = 10^8$, including peak temperature, minimum bubble radius, and maximum wall velocity, are assessed against prior studies to help validate the EP scaling formalism and our hybrid continuum-MD framework.