Numerical sensitivity analysis of facility effects on CO₂ Hall thruster performance

Abstract CO₂ is a promising propellant candidate for Hall thrusters due to its cost, storability, and potential for in-situ resource utilization (ISRU). In this study, fully kinetic particle-in-cell (Full-PIC) simulations were conducted to elucidate the influence of background pressure in ground test facilities, known as the facility effect, on the performance of a CO₂ Hall thruster. Comparing a no-backpressure condition with dynamic backflow conditions from the chamber demonstrated that utilizing a dynamic inflow model with velocities ranging from tens to hundreds of m/s reasonably reproduces the increasing trend of thrust performance observed in experiments. The resulting thrust increments significantly exceeded theoretical values on the assumption that the background gas acts merely as an additional propellant, suggesting that the ingested background gas chain-reactively promotes the ionization of the main propellant via dissociation and multi-step ionization processes. Furthermore, comparisons with simulations where an equivalent gas flow rate was added from upstream indicated that the disproportionate increase in wall ion flux relative to the thrust increment under high backpressure is not unique to downstream backflow. Conversely, structural changes unique to backpressure inflow were indicated. Specifically, the main ionization region is essentially fixed due to the competing ionization effects of the gas inside and outside the channel, and the high background gas density accumulating in the plume region enhances electron mobility from the channel exit to the cathode, relaxing the axial potential gradient. Although the inclusion of dissociation reactions successfully reproduced the decreasing trend in the CO₂⁺ fraction with increasing backpressure, the underestimation of certain ion beam currents suggests that further model refinement is required. In particular, incorporating charge-exchange (CEX) collisions and the backflow of dissociation products (CO and O), together with future experiments using varied probe installation distances to evaluate long-path collision effects, will be essential for improving predictive accuracy and validating the model.