Simulating Relativistic Magnetic Reconnection with a Pseudo-Spectral Maxwell Solver

Magnetic reconnection is an important cause of particle acceleration and heating in astrophysical settings like gamma-ray bursts, magnetars, and pulsars. Using the particle-in-cell code WarpX, we perform large first-principles 2D simulations of relativistic reconnection in plasmas representative of those found in astrophysical environments. These simulations produce particle spectra and plasmoid structures in agreement with results previously reported in the literature. Using this robust baseline case, we compare the accuracy and computational performance of three methods of solving Maxwell's equations, including the commonly-used second-order finite difference time domain (FDTD) method and an ultra-high-order pseudo-spectral analytical time domain (PSATD) method. This is the first time PSATD has been used in simulations of relativistic reconnection. We find that for the reconnection problem, FDTD and PSATD are comparably accurate, but that PSATD is >50% more computationally efficient. These performance gains will make 3D simulations of reconnection more tractable, and complement other efforts to improve simulation efficiency, such as the use of mesh refinement.