Speaker
Description
One of the most extensively studied characteristics of vacuum breakdown (VBD) is the conditioning process and the VBD occurrence statistics, in various systems, including Radio-Frequency (RF) accelerators and pulsed-DC large electrode systems. Despite abundant data on VBD statistics, drawing useful conclusions regarding the physical processes that determine various patterns within those data is extremely challenging. The existing VBD models focus on low-level physical mechanisms and thus cannot produce direct quantitative predictions that are comparable to the aforementioned data.
Here we attempt to bridge this gap between theory and experiment with a Monte-Carlo model that simulates the occurrence of VBD based on various general assumptions. We model VBD occurrence as a Markov chain process. The metal surface is separated into small elements, each of which is described by a local field E, a power coupling impedance parameter Z, and a surface state parameter β. On each pulse, each surface element is randomly tested for the occurrence of thermal runaway. The probability of thermal runaway (TR) is a sharply increasing function of βΕ. The occurrence statistics of TR are fitted to the experimental measurements of dark current fluctuations interpreted here as TR events that did not lead to full VBD.
Then elements where TR occurred are tested for development into full VBD, depending on E and the power coupling parameter Z of each point. The surface state parameter β of a point is updated after each pulse in a different manner, depending on whether nothing, TR, or VBD occurred.
Finally, we test the above method for various distributions of the simulation parameters, fitting them to reproduce well-known experimental conditioning curves and other VBD statistics.
