Superkicks and the photon angular and linear momentum density

TL;DR

This paper investigates the proper definition of electromagnetic momentum density in structured light fields, revealing significant differences between canonical and symmetric energy-momentum tensors with implications for light-matter interactions.

Contribution

It demonstrates that using the canonical energy-momentum tensor leads to different predictions for momentum and angular momentum than the symmetric tensor, impacting the understanding of light-induced forces.

Findings

Canonical and symmetric tensors produce drastically different recoil momenta.

Numerical estimates show significant effects on small objects and atomic rotors.

Implications for the interpretation of light's momentum in structured electromagnetic fields.

Abstract

We address a problem of proper definition of momentum density for spatially structured electromagnetic fields. We show that the expressions for the momentum and angular momentum obtained locally are not the same when one uses the canonical energy-momentum tensor instead of the symmetric Belinfante energy-momentum tensor in electrodynamics. This has important consequences for interaction of matter with structured light, for example, twisted photons; and would give drastically different results for forces and angular momenta induced on small test objects. We show, with numerical estimates of the size of the effects, situations where the canonical and symmetrized forms induce very different torques or (superkick) recoil momenta on small objects or atomic rotors, over a broad range of circumstances.