Influence of the magnetic field configuration on the plasma flow in Hall thrusters

Abstract

In Hall propulsion, the thrust is provided by the acceleration of ions in a plasma generated in a cross-field configuration. Standard thruster configurations have annular channels with an almost radial magnetic field at the channel exit. A potential difference is imposed in the axial direction and the intensity of the magnetic field is calibrated in order to hinder the electron motion, while leaving the ions non-magnetised. Magnetic field lines can be assumed, as a first approximation, as lines of constant electron temperature and of thermalized potential. In typical thruster configurations, the discharge occurs inside a ceramic channel and, due to plasma-wall interactions, the electron temperature is typically low, less than few tens of eV. Hence, the magnetic field lines can be effectively used to tailor the distribution of the electrostatic potential. However, the erosion of the ceramic walls caused by the ion bombardment represents the main limiting factor of the thruster lifetime and new thruster configurations are currently under development. For these configurations, classical first order models of the plasma dynamics fail to grasp the influence of the magnetic topology on the plasma flow. In the present paper, a novel approach to investigate the correlation between magnetic field topology and thruster performance is presented. Due to the anisotropy induced by the magnetic field, the gradients of the plasma properties are assumed to be mainly in the direction orthogonal to the local magnetic field, thus enabling a quasi-one-dimensional description in magnetic coordinates. Theoretical and experimental investigations performed on a 5 kW class Hall thruster with different magnetic field configurations are then presented and discussed.

Latest progress in Hall thrusters plasma modelling

“…Developed models are fully fluid or hybrid when electrons are treated as a fluid and heavy species (ions and neutrals) with a kinetic description. Morozov and Savelyev (2000a) 1D (z) hybrid No Yes QN Channel Keidar et al (2001) 1D (r) fluid Yes No Poisson Channel 1D (z) hybrid No No QN Channel/ Near field Roy and Pandey (2002) 1D (z) fluid Yes Yes QN Channel Ahedo (2002) 1D (r) fluid Yes No Poisson Channel Ahedo et al (2003) 1D (z) hybrid Yes No QN Channel/ Near field 1D (z) fluid Yes Yes QN Channel Hara et al (2012) 1D (z) hybrid No No QN Channel Komurasaki and Arakawa (1995) 2D (z,r) hybrid No No QN Channel Fife (1998) 2D (z,r) hybrid Yes Yes QN Channel/ Near field Hagelaar et al (2002) 2D (z,r) hybrid No Yes QN Channel/ Near field Koo and Boyd (2004) 2D (z,r) hybrid No Yes QN Channel/ Near field Keidar et al (2004) 2D (z,r) fluid No No QN Channel Parra et al (2006) 2D (z,r) hybrid Yes Yes QN Channel/ Near field Garrigues et al (2006) 2D (z,r) hybrid Yes Yes QN Channel/ Near field Mikellides and Katz (2012) 2D (z,r) fluid Yes Yes QN Channel/ Near field Lam et al (2015) 2D (z,q) hybrid Yes Yes QN Channel/ Near field Andreussi et al (2018) 2D (z,r) fluid Yes Yes QN Channel/ Near field Tab. I 1D and 2D fluid/hybrid models of HTs by chronological order.…”

“…(7). Finally, the pressure term is either neglected considering cold ions (Andreussi et al 2018) either simplified considering a specified ion temperature (Mikellides and Katz 2012).…”

Triple Langmuir Probes Measurements of LaB6 Hollow Cathodes Plume

“…The electron temperature, plasma potential and plasma density along the axis of the cathode were measured using a triple Langmuir probe mounted on an articulated arm. The moving system and the arm were the same used to measure the plasma parameters inside the SITAEL Hall thruster channel [39], whereas the Langmuir probe was modified according to the expected plasma parameters in the cathode plume. Following the approach described in Andreussi et al [39], the arm rapidly inserts and extracts the probe from the cathode plume, with a maximum residence time in the dense plasma region of less than 0.2 s. It is important to notice that, even though the residence time of the probe is less than 0.2 s, such measurements are intrusive, since the spatial gradients as well as part of the plasma dynamics may be affected by the probes [40].…”

“…The data will be further analyzed implementing the parameterization described in Mausbach [41] of the Laframboise solution for cylindrical Langmuir probes. In addition, a Bayesian integrated data analysis will be implemented, following the approach of Andreussi et al [39].…”

Three‐dimensional immersed finite‐element method for anisotropic magnetostatic/electrostatic interface problems with nonhomogeneous flux jump

“…In most existing models, there are usually two ways to deal with the anisotropic problem. Characteristic line mesh method: In this method, the simulation domain will be divided into elements by the characteristic lines as mesh lines, such as the magnetic field lines . In this way, the material properties along the mesh lines will be consistent and thus easy to be described. Noncharacteristic line mesh method: In this method, the mesh lines are not the characteristic lines any more.…”