Practical analysis of different neutral algorithms for particle simulation of Hall thruster

The modeling of neutral atoms is important for the full-particle simulations of Hall thrusters. In previous studies, researchers have developed various algorithms to model the neutral kinetics. The choice of those algorithms can influence significantly the computational speed, simulation convergence, and physical results. In this work, we perform a full-particle simulation of a typical 1 kW-class SPT-100 Hall thruster using four neutral algorithms, including the fixed-neutral algorithm (FNA), the algorithm of direct simulation of Monte Carlo (DSMC), the collisionless-neutral algorithm (CLNA), and the fluid algorithm (FA), to analyze the effects of different neutral iteration approaches on the simulation results. We found that FNA is sensitive to the initial number density of neutrals, and is difficult to converge properly, while the other algorithms not neglecting the atomic dynamics can get stable results. We count the parameters of the thruster, that is, thrust, specific impulse, and plasma density using different neutral algorithms. The time-averaged results match well with those of the experiment. However, the results differ in the time scale due to the low-frequency oscillations in Hall thrusters. We verify that the oscillations are due to the periodic change of neutrals and establish a zero-dimensional model to analyze the properties of the oscillations in the time scale. It indicates that the ratio of ion migration to neutral migration is the essential factor that significantly affects the calculation results. The model reveals that the direct neutral iteration methods, like DSMC and CLNA, can better simulate the characteristics of discharge fluctuations in Hall thrusters than the quasi-steady-state method, like FA. Finally, we proposed practical suggestions for the selection of the neutral algorithms for the SPT-100 thruster, which can also be generalized to other low- and medium-power Hall thrusters.