A minimal model for acoustic forces on Brownian particles

TL;DR

This paper introduces a generalized inertial coupling method to accurately simulate acoustic radiation forces on small particles with arbitrary compressibility, incorporating thermal fluctuations and enabling efficient GPU-based computations.

Contribution

The authors extend the inertial coupling method to handle particles with arbitrary compressibility, improving accuracy in simulating acoustic forces on Brownian particles.

Findings

The generalized method accurately reproduces monopolar and dipolar acoustic force contributions.

Thermal fluctuations are incorporated, enhancing realism at microscopic scales.

The approach is computationally efficient, suitable for large particle simulations on GPUs.

Abstract

We present a generalization of the inertial coupling (IC) [Usabiaga et al. J. Comp. Phys. 2013] which permits the resolution of radiation forces on small particles with arbitrary acoustic contrast factor. The IC method is based on a Eulerian-Lagrangian approach: particles move in continuum space while the fluid equations are solved in a regular mesh (here we use the finite volume method). Thermal fluctuations in the fluid stress, important below the micron scale, are also taken into account following the Landau-Lifshitz fluid description. Each particle is described by a minimal cost resolution which consists on a single small kernel (bell-shaped function) concomitant to the particle. The main role of the particle kernel is to interpolate fluid properties and spread particle forces. Here, we extend the kernel functionality to allow for an arbitrary particle compressibility. The particle-fluid force is obtained from an imposed no-slip constraint which enforces similar particle and kernel fluid velocities. This coupling is instantaneous and permits to capture the fast, non-linear effects underlying the radiation forces on particles. Acoustic forces arise either because an excess in particle compressibility (monopolar term) or in mass (dipolar contribution) over the fluid values. Comparison with theoretical expressions show that the present generalization of the IC method correctly reproduces both contributions. Due to its low computational cost, the present method allows for simulations with many particles using a standard Graphical Processor Unit (GPU).