Simulation of a radio frequency plasma device via drift-diffusion models

This thesis investigates the numerical simulation of a Radio Frequency (RF) plasma device using Drift-Diffusion (DD) models, focusing on the challenges posed by very low-pressure regimes. These operating conditions are of growing interest in electric propulsion, particularly for next-generation RF plasma thrusters. An existing iterative simulation framework, coupling a DD fluid solver implemented in OpenFOAM with an electromagnetic code, is adapted and improved to extend its applicability to these low-pressure regimes. Numerical stability is enhanced through the formulation of a new set of implicit and semi-implicit boundary conditions. The modified framework is applied to study the source region of Moa, a RF plasma device located at the University of Auckland. Initial plasma parameters are estimated via a Global Model and refined through successive iterations of coupled RF power deposition and fluid simulations. The results indicate that the framework captures key trends in plasma behavior, however, quantitative agreement with experimental data remains limited. Significant discrepancies are observed since non-local effects are believed to dominate plasma dynamics with respect to classic collisional transport. To partially account for these effects, heuristic corrections, such as anomalous diffusion and modified Bohm velocity coefficients, are introduced in the model. These lead to modest improvements, but they do not eliminate the fundamental limitations of the fluid approach in this regime. Overall, the results underscore both the strengths and constraints of the DD approach in modeling RF plasmas at low pressures. While the model proves useful for identifying trends and informing preliminary design, its predictive capability outside validated regimes is limited. Moreover, due to the necessity of calibrating empirical correction coefficients, decoupling simulations from prior experimental knowledge remains a significant challenge.