Alternative propellant performance in Hall effect thrusters

Munro-O'Brien, Thomas Francis (2026) Alternative propellant performance in Hall effect thrusters. University of Southampton, Doctoral Thesis, 326pp.

Record type: Thesis (Doctoral)

Abstract

This thesis investigates the scaling, design, and performance of a novel modular high-voltage Hall effect thruster operating on alternative propellants, with the aim of enhancing understanding of alternative propellant performance and reducing mission costs compared to conventional xenon-propelled systems. As the electric propulsion community seeks to reduce reliance on xenon, a propellant marred by scarcity and volatile price, lighter alternatives such as krypton and argon offer potential cost-saving and advantages in specific impulse without the xenon-associated supply chain constraints. However, in traditional Hall thrusters this potential increase in specific impulse is often offset by lower propellant utilisation, stemming from their smaller ionisation cross-sections and higher ionisation energies. Furthermore, while prior work has established empirical scaling methods for xenon-fed thrusters, the lack of comprehensive performance data for alternative propellants has hindered their optimisation.

To address this gap, two key contributions are presented: the development of an extended semi-empirical scaling methodology that preserves propellant-specific properties, and the experimental characterisation of a modular discharge channel thruster designed to investigate the influence of channel geometry, namely mean channel diameter, and channel width, on alternative propellant performance.

The scaling framework builds upon semi-empirical methodologies by Dannenmayer et al. and Kim et al., extending previous work through preservation of propellant-dependent metrics and enabling direct scaling for krypton Hall effect thrusters. To validate this scaling methodology, a novel modular Hall thruster was designed, featuring a discharge channel with adjustable width and mean diameter while maintaining constant magnetic, electrical, diagnostic, and fluidic configurations. This design enabled a systematic comparison of xenon, krypton, and argon across nine distinct channel geometries at discharge voltages up to 600 V. The final thruster, scaled using this new method for krypton at 2.5 kW and 600 V, was designated the Southampton High-voltage Anode-layer Research Krypton 600 V “SHARK-600V”. An optional thruster-with-anode-layer configuration was also tested as a potential method for enhancing alternative propellant performance, though full characterisation was not possible.

The experimental results from the University of Southampton campaign reveal clear propellant-specific trends: xenon exhibits efficiency degradation at high voltages, krypton shows similar degradation only for certain geometries, and argon demonstrates reduced performance sensitivity to geometry, with greater stability sensitivity to geometry across the range of tested flow rates and voltages. At optimised conditions, krypton achieved a specific impulse of 2809 s and an anode efficiency of 35.8%, exceeding that of xenon of (2422 s, 34.7%) but with slightly reduced thrust (109.3 mN vs. 115 mN). Argon attained comparable specific impulse (2537 s) but with reduced thrust (81.7 mN) and anode efficiency (24.8%).

To further utilise this dataset, a Gaussian process regression model was trained on this experimental dataset, creating a continuous surrogate-model of thruster performance across the input parameter space. The model enabled interpolation between tested geometries and extrapolation to untested conditions, all with quantified uncertainty. Crucially, the optimal geometries identified by the GPR model showed meaningful agreement with the semi-empirical scaling predictions, providing experimental validation of the scaling approach developed for this thesis. For argon, where conventional scaling methods cannot be applied due to insufficient literature data, the GPR model provided a novel experimentally derived optimal channel geometry, offering preliminary design guidance for future argon thruster development.

Cross-facility validation at the University of Michigan on krypton further strengthened these findings through direct thrust measurements and far-field plume diagnostics using a retarding potential analyser, ExB probe, and a Faraday probe. Direct thrust measurements for krypton identified a peak performance at 50 sccm and 600 V, with a specific impulse of 3511.2 s, thrust of 107.3 mN, and an anode efficiency of 58.1%. At 60 sccm and 500 V, direct thrust measurements yielded 2901.5 s, 106.4 mN, and 50.6% (+/- 2.24%) efficiency, while probe-derived results closely matched, reporting 50.7% anode efficiency comprised of 97.1% charge efficiency, 70.2% divergence efficiency, 88.4% current efficiency, 95.0% voltage efficiency, and 88.5% mass utilisation efficiency.

Testing at the University of Michigan revealed consistently higher performance in terms of anode efficiency, thrust, and specific impulse. These results from direct thrust measurements are corroborated by the far-field probe measurements, showing good agreement between the direct thrust anode efficiency and the probe-derived anode efficiency. These cross-facility results both validate the core geometry and highlight the high-performance capability of high-voltage Hall thrusters operating on krypton. Importantly, although absolute performance at Michigan exceeded that at Southampton for the same configuration, both facilities reached peak anode efficiency at the same operating point: 50 sccm, 600 V. The enhanced performance is attributed to the thruster operating at a consistently lower discharge current with a corresponding slight increase in thrust output. This results in a lower power draw for a slightly enhanced performance at the Michigan test site compared to the University of Southampton test site.