Luke Silvestre

A numerical model of a dispersive, discrete-element nonlinear transmission line (NLTL) employing saturable inductors is reported and demonstrated to qualitatively reproduce the experimentally measured output of an NLTL. This numerical model is used with a genetic algorithm (GA) to optimize RF output power by varying the number of turns on each inductor at each stage and capacitance at each stage of the NLTL. Based on these results, an NLTL with the GA-optimized component values is fabricated and tested. In testing, the optimized NLTL produced $\sim$ 30% higher output power than the highest output power of a uniformly distributed NLTL. In selected cases, for identical bias conditions, the optimized NLTL generated nearly 80% higher output power. Notably, the optimized NLTL produced a nonnegligible higher output frequency while also generating higher output power.

This report employs a Vlasov–Poisson model to elucidate fundamental electron phase–space mechanics of a multipactor discharge from onset to saturation. At the onset of multipactor, the electron phase–space is primarily defined by sharp features in both the physical space and energy space. With increasing electron density, space-charge effects lead to debunching of the swarm in phase–space. The temporal evolution of the electron energy distribution is studied across a single impact cycle. The average and peak-to-peak saturation values for the entire first-order multipactor regime are presented. Comparisons between the third-and fifth-order multipactors highlight the nuanced similarities and differences in the energy distribution of the multipacting system. The Vlasov–Poisson approach, which neglects collisions, is well suited for such analysis since the multipactor phenomenon occurs under near-vacuum collisionless conditions. It also overcomes difficulties associated with kinetic schemes that require adequately sampling all of the electron phase–space, including sparely populated regions, or special treatments to model strong growths in carrier densities.

Multipactor is studied based on the coupled Vlasov–Poisson equation set and applied to a parallel plate geometry. This approach can be considered complementary to the particle-in-cell (PIC) methods that have provided excellent insight into multipactor behavior. However, PIC methods have limitations in terms of ‘particle noise,’ which can affect electron energy distribution functions and create re-scaling issues under conditions of charge growth. Utilizing our continuum Vlasov–Poisson approach yields susceptibility curves that are in line with reports in the literature, Spark3D PIC simulations, and experimental data. Playing to the strength of the Vlasov–Poisson approach, the differences between various multipactor orders are elucidated as they are observed in phase-space, revealing electron density dynamics without requiring increased computational resources due to electron growth. The method presented is general and can be extended to multi-input excitations and higher-dimensional phase-space.