A $θ$-implicit hybrid particle-in-cell algorithm for kinetic simulations of magnetized plasmas

Abstract

Hybrid particle-in-cell (PIC) models with kinetic ions and an inertialess electron fluid are widely used to simulate magnetized plasmas at ion scales, but explicit time advances are constrained by the whistler-wave CFL condition $\dt \propto Δx^2$, and the sub-cycling and artificial dissipation used to mitigate it carry accumulating costs. We present a $θ$-implicit hybrid PIC algorithm, implemented in the open-source GPU-capable code WarpX, in which fields and particles are advanced as a single nonlinear system solved by a particle-suppressed Jacobian-free Newton--Krylov method: the particles are nonlinearly eliminated, the electric field of the generalized Ohm's law is the sole unknown, and mass matrices remove all particle operations from the linear iterations. The electron pressure is advanced inside the nonlinear solve, and the scheme satisfies a discrete energy law whose single dissipation term vanishes at $θ= 1/2$ and damps only temporally under-resolved modes for $θ> 1/2$. The advance reproduces parallel and perpendicular wave dispersion including the ion Bernstein modes, conserves total energy to machine roundoff where the explicit advance heats secularly, does not develop the cold-beam finite-grid instability, and reproduces the $m=0$ instability of a Bennett pinch in RZ geometry. We conclude with 3D simulations of a field-reversed configuration initialized from a Grad--Shafranov equilibrium, comparing the explicit and implicit advances on the growth of the tilt instability and the preservation of trapped magnetic flux.