Wave-Driven Torques to Drive Current and Rotation

TL;DR

This paper explains how electromagnetic energy and momentum fluxes in plasma waves enable net current and rotation drive by transferring torque from antenna to plasma, resolving a longstanding paradox in plasma physics.

Contribution

It introduces a steady-state model showing electromagnetic flux transfer allows rotation drive, unlike traditional recoil-based explanations, bridging initial value and boundary value problem perspectives.

Findings

Electromagnetic flux transfers torque to resonant particles.

Steady-state model explains local momentum conservation.

Nonresonant recoil vanishes in steady state, enabling rotation drive.

Abstract

In the classic Landau damping initial value problem, where a planar electrostatic wave transfers energy and momentum to resonant electrons, a recoil reaction occurs in the nonresonant particles to ensure momentum conservation. To explain how net current can be driven in spite of this conservation, the literature often appeals to mechanisms that transfer this nonresonant recoil momentum to ions, which carry negligible current. However, this explanation does not allow the transport of net charge across magnetic field lines, precluding ExB rotation drive. Here, we show that in steady state, this picture of current drive is incomplete. Using a simple Fresnel model of the plasma, we show that for lower hybrid waves, the electromagnetic energy flux (Poynting vector) and momentum flux (Maxwell stress tensor) associated with the evanescent vacuum wave, become the Minkowski energy flux and momentum flux in the plasma, and are ultimately transferred to resonant particles. Thus, the torque delivered to the resonant particles is ultimately supplied by the electromagnetic torque from the antenna, allowing the nonresonant recoil response to vanish and rotation to be driven. We present a warm fluid model that explains how this momentum conservation works out locally, via a Reynolds stress that does not appear in the 1D initial value problem. This model is the simplest that can capture both the nonresonant recoil reaction in the initial-value problem, and the absence of a nonresonant recoil in the steady-state boundary value problem, thus forbidding rotation drive in the former while allowing it in the latter.