The electron cyclotron drift instability: a comparison of particle-in-cell and continuum Vlasov simulations

TL;DR

This study compares particle-in-cell and continuum Vlasov simulations of the electron cyclotron drift instability, revealing differences in linear growth rates, nonlinear behavior, and electron heating effects related to simulation methods and azimuthal length.

Contribution

It provides a detailed comparison of PIC and Vlasov simulation results for ECDI, highlighting the impact of azimuthal length and simulation noise on nonlinear dynamics and electron heating.

Findings

Vlasov simulations match theoretical linear growth rates closely.

Inverse cascade observed in both methods with large azimuthal length.

PIC simulations show higher fluctuation amplitudes and electron heating.

Abstract

The linear and nonlinear characteristics of the electron cyclotron drift instability (ECDI) have been studied through the particle-in-cell (PIC) and continuum Vlasov simulation methods in connection with the effects of the azimuthal length (in the $E \times B$ direction) on the simulations. Simulation results for a long azimuthal length (17.82 cm $= 627\;v_d/\omega_{ce}$, where $\omega_{ce}$ is the electron cyclotron frequency and $v_d$ is the $E\times B$ drift of the electrons) are reported, for which a high resolution is achieved in Fourier space. For simulations with a long azimuthal length, the linear growth rates of the PIC simulations show a considerable discrepancy with the theory, whereas the linear growth rate of the Vlasov simulations remains close to the theory. In the nonlinear regime, the inverse cascade is shown in both PIC and Vlasov simulations with a sufficiently large azimuthal length. In simulations with a short azimuthal length, however, the inverse cascade is barely observed. Instead, the PIC simulations with a short azimuthal length (0.5625 cm $=19.8\;v_d/\omega_{ce}$) show an essentially continuous nonlinear dispersion, similar to what is predicted by the ion-sound turbulence theory. It is shown that, in the PIC and Vlasov simulations, the inverse cascade coincides with the formation and merging of electron structures in phase space. This process, however, terminates differently in the PIC simulations compared with the Vlasov simulations. Larger amplitudes of ECDI fluctuations are observed in the PIC simulations compared with the Vlasov simulations, leading to an intensified electron heating and anomalous current. This suggests that the statistical noise of PIC simulations might contribute to the extreme electron heating that has been observed in previous studies.