Asymptotic-Preserving schemes for the Boltzmann mixture model with disparate mass

Zhen Hao, Ning Jiang, Liu Liu

Abstract

This paper presents a novel numerical framework for simulating gas mixtures with disparate molecular masses, a common scenario in plasma physics and aerospace engineering. By utilizing a truncated asymptotic expansion of the Boltzmann collision operator combined with a dual-grid approach, the proposed method achieves computational efficiency independent of the mass ratio. The scheme is asymptotic-preserving (AP), accurately capturing the multi-scale "epochal" relaxation of the system while allowing for time steps significantly larger than the collision frequency.

Introduction

In kinetic theory, the Boltzmann equation describes the evolution of gas particles through collisions. However, standard numerical methods face significant hurdles when dealing with mixtures containing species of vastly different masses, such as electrons and ions. Two primary challenges emerge: velocity disparity, where light particles move significantly faster than heavy ones, and time-scale stiffness, known as epochal relaxation. These mixtures do not reach equilibrium simultaneously; instead, they undergo distinct phases where the light species equilibrates first, followed by the heavy species, and finally a slow equalization of temperatures. Capturing these phases traditionally requires prohibitively fine grids and tiny time steps.

Methodology

1. The Asymptotic-Expansion (AE) Method

To bypass the high-dimensional complexity of the full Boltzmann collision integral, the authors employ a truncated asymptotic expansion. By focusing on the regime where the square root of the mass ratio ($\epsilon = \sqrt{m_l/m_h}$) is small, the inter-particle collision operators are simplified into a set of velocity space integrals and derivatives. This decoupling ensures that computational costs remain independent of the mass ratio, a significant improvement over traditional spectral methods.

2. Dual-Grid Procedure

A key innovation in evaluating these truncated operators is the dual-grid approach. The authors utilize a standard Cartesian grid for intra-species collisions while switching to a Polar grid for inter-species interactions. This allows for efficient angular integration using the trapezoidal rule before results are interpolated back to the Cartesian grid.

Figure 1: Illustration of the dual-grid approach in velocity space, showing the Cartesian grid (teal) and the Polar grid (red) used for inter-species collision calculations.

3. Asymptotic-Preserving (AP) Time Stepping

To manage stiff time scales, the scheme incorporates a BGK-penalization technique. This ensures stability and accuracy even when the time step is much larger than the collision frequency. The scheme is designed to automatically transition from a kinetic solver to a macroscopic fluid-like solver as the system approaches equilibrium.

Numerical Experiments and Results

The scheme was evaluated using synthetic initial conditions, such as "double-peak Maxwellians," and Maxwell molecules where explicit mathematical relations are known.

  • Computational Efficiency: In a benchmark test with an $800 \times 800$ grid, the traditional Spectral (SP) method required over 33,000 seconds. In contrast, the AE method completed the same simulation in just 46 seconds—a speedup of approximately 700x.
  • Uniform Accuracy: The AE method demonstrated consistent accuracy for small $\epsilon$, with errors remaining stable even as the mass ratio became more extreme.
  • Physical Fidelity: The numerical experiments successfully reproduced the three distinct time scales of epochal relaxation, matching theoretical predictions for temperature and velocity evolution.

Conclusion

The proposed AE method offers a robust solution for simulating disparate mass mixtures, particularly when $\epsilon < 0.1$. While currently optimized for Maxwell molecules and space-homogeneous flows, future work will focus on extending the framework to more realistic potentials like Hard Spheres and inhomogeneous spatial gradients. Additionally, further refinements are needed to ensure strict conservation of mass and energy during long-term simulations involving truncated operators.

Summary updated

BibTeX
@article{hao2025asymptotic,
  author = {Zhen Hao and Ning Jiang and Liu Liu},
  title = {{Asymptotic-Preserving schemes for the Boltzmann mixture model with disparate mass}},
  journal = {SIAM J. Sci. Comput.},
  year = {2025},
  doi = {10.1137/24m1715222},
}
Figure