UCLA

This dissertation describes three computational models developed to simulate important aspects of low-temperature plasma devices, most notably ring-cusp ion discharges and thrusters. The main findings of this dissertation are related to (1) the mechanisms of cusp confinement for micro-scale plasmas, (2) the implementation and merits of magnetic field aligned meshes, and (3) an improved method for describing heavy species interactions.

The Single Cusp (SC) model focuses on the near-cusp region of the discharge chamber to investigate the near surface cusp confinement of a micro-scale plasma. The model employs the multi-species iterative Monte Carlo method and uses various advanced methods such as electric field calculation and particle weighting algorithm that are compatible with a non-uniform mesh in cylindrical coordinates. Three different plasma conditions are simulated with the SC model, including an electron plasma, a sparse plasma, and a weakly ionized plasma. It is found that the scaling of plasma loss to the cusp for a sparse plasma can be similar to that for a weakly ionized plasma, while the loss mechanism is significantly different; the primary electrons strongly influence the loss structure of the sparse plasma. The model is also used, along with experimental results, to describe the importance of the local magnetic field on the primary electron loss behavior at the cusp.

Many components of the 2D/3D hybrid fluid/particle model (DC-ION) are improved from the original version. The DC-ION code looks at the macroscopic structure of the discharge plasma and can be used to address the design and optimization challenges of miniature to micro discharges on the order of 3 cm to 1 cm in diameter. Among the work done for DC-ION, detailed steps for the magnetic field aligned (MFA) mesh are provided. Solving the plasma diffusion equation in the ring-cusp configuration, the benefit of the MFA mesh has been fully investigated by comparing the solution with a uniform mesh. It is found that the MFA mesh can still produce a relatively large error due to the misalignment at the domain boundary but still provides a significant improvement in the bulk plasma region. The mesh generation routine can be further improved by enhancing the smoothness of the near-boundary grid elements.

The Ion Beam (IB) model is similar to the SC model but primarily focuses on heavy species. The model implements detailed calculation of the heavy species collisions, solving the classical scattering equation with higher order interaction potentials for different collision pairs. A parametric study has been conducted, and the simulation results have shown very good agreement with the experimental results when using an appropriate value for the secondary electron yield. Further study on the elastic collision has shown that the initial atom velocity should not be neglected in order to accurately compute the post-collision CEX ion velocity. The knowledge gained has led to an improvement on the current method by defining an effective elastic collision cross-section and significantly reducing the frequency of the collision calculation. The same method can be applied to collisions between fast and slow atoms.