Acoustic propulsion of a small bottom-heavy sphere

TL;DR

This paper derives how a small, bottom-heavy sphere can be propelled by a transverse acoustic field, highlighting the role of density inhomogeneities and frequency-dependent propulsion direction, with results matching experimental observations.

Contribution

The study provides a comprehensive theoretical derivation of acoustic propulsion for a bottom-heavy sphere, emphasizing the effects of density inhomogeneity and frequency on propulsion direction and speed.

Findings

Propulsion speed scales inversely with viscosity.

Speed scales with the cube of acoustic amplitude.

Propulsion direction can reverse at certain frequencies.

Abstract

We present here a comprehensive derivation for the speed of a small bottom-heavy sphere forced by a transverse acoustic field and thereby establish how density inhomogeneities may play a critical role in acoustic propulsion. The sphere is trapped at the pressure node of a standing wave whose wavelength is much larger than the sphere diameter. Due to its inhomogeneous density, the sphere oscillates in translation and rotation relative to the surrounding fluid. The perturbative flows induced by the sphere's rotation and translation are shown to generate a rectified inertial flow responsible for a net mean force on the sphere that is able to propel the particle within the zero-pressure plane. To avoid an explicit derivation of the streaming flow, the propulsion speed is computed exactly using a suitable version of the Lorentz reciprocal theorem. The propulsion speed is shown to scale as the inverse of the viscosity, the cube of the amplitude of the acoustic field and is a non trivial function of the acoustic frequency. Interestingly, for some combinations of the constitutive parameters (fluid to solid density ratio, moment of inertia and centroid to center of mass distance), the direction of propulsion is reversed as soon as the frequency of the forcing acoustic field becomes larger than a certain threshold. The results produced by the model are compatible with both the observed phenomenology and the orders of magnitude of the measured velocities.