Electron Inertia and Magnetic Reconnection
Allen H Boozer
TL;DR
This work proves that when electron inertia is the sole non-ideality, the three-dimensionally evolving modified magnetic field $\mathcal{B} = \vec{B} + \nabla\times\left( \frac{m_e}{n e^2}\vec{j} \right)$ cannot reconnect, even though $\vec{B}$ itself can; the evolution follows a Faraday-like law with a compact form for $\mathcal{E}$, and a Clebsch-based argument shows field lines of $\mathcal{B}$ are advected without topological change when $\mathcal{E}_r$ vanishes. The paper further shows that electron-inertia–driven reconnection effects on guiding-center trajectories are second-order in $(c/\omega_{pe})^2$ and typically smaller than gyroradius effects, except in extremely large current sheets. In 3D, however, the natural chaotic flow of field lines implies any nonzero diffusive effect leads to large-scale reconnection on a timescale $\tau_u\ln(\tau_d/\tau_u)$, rendering the nonreconnection result less practically relevant for realistic plasmas. The discussion also connects $\mathcal{B}$ to Voigt-regularized field behavior and clarifies how particle trajectories should be analyzed relative to the modified field rather than $\vec{B}$ alone.
Abstract
When electron inertia is the only non-ideal effect in the evolution of a magnetic field $\vec{B}$, the field lines of $\vec{B}$ reconnect, but the lines of a related field $\vec{\mathcal{B}}$ do not. $\vec{\mathcal{B}} \equiv \vec{B} + \vec{\nabla}\times \left( (c/ω_{pe})^2μ_0\vec{j} \right)$ with $ω_{pe}$ the plasma frequency and $\vec{j}$ the current density. Although a full four-dimensional relativistic calculation of $\vec{\mathcal{B}}$ has been made, studies of $\vec{\mathcal{B}}$ have been focused on systems that depend on only two spatial coordinates. Three results are given: (1) A relatively simple demonstration in three dimensional space that the lines of $\vec{\mathcal{B}}$ do not reconnect when electron inertia is the only non-ideal effect. (2) The guiding center motion of charged particles is modified by a term that is proportional to $(c/ω_{pe})^2$, which is smaller than the drifts proportional to the gyroradius unless the current density is extremely large. (3) In three dimensional space, the evolution velocity of $\vec{\mathcal{B}}$ is characteristically chaotic, which means neighboring streamlines separate exponentially on a timescale $τ_u$. $\vec{\mathcal{B}}$ undergoes large scale reconnection on a timescale that is only an order of magnitude or two longer than $τ_u$ unless all diffusive non-ideal effects, such as resistivity, are absolutely zero.
