RF Gun optimization study

Abstract

Injector gun design is an iterative process where the designer optimizes a few nonlinearly interdependent beam parameters to achieve the required beam quality for a particle accelerator. Few tools exist to automate the optimization process and thoroughly explore the parameter space. The challenging beam requirements of new accelerator applications such as light sources and electron cooling devices drive the development of RF and SRF photo injectors. RF and SRF gun design is further complicated because the beam is space charge dominated. A genetic algorithm (GA) has been successfully used to optimize DC photo injector designs at Cornell University [1] and Jefferson Lab [2]. We propose studying how GA techniques can be applied to the design of RF and SRF gun injectors. In this paper, we report on the initial phase of the study where we model and optimize a system that has been benchmarked with beam measurements and simulation.

Innovative applications of genetic algorithms to problems in accelerator physics

“…(2) are f 2 ðx 1 ; x 2 Þ < 20 and ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi f 2 ðx 1 ; x 2 Þ p < 10. We developed a user-friendly, script-based hybrid of PISA and APISA and use it for all optimizations presented, except the rf gun optimization that builds directly on APISA [24][25][26].…”

“…Here we describe an optimization system that can vary the shape of the field profile of an rf gun in response to the performance of the beam dynamics [24][25][26]54]. Often rf guns are used in linac-based light sources.…”

“…For a multicell cavity, the resonance frequency and the field profile shape are very sensitive to small changes in the cell radii [25]. The resonance frequency for the -mode alone is not sufficient to determine if a cavity geometry produces a reasonable field profile because for very small changes in cavity dimensions, while the frequency may remain the same, the relative field amplitudes in the field profile can change significantly.…”

“…The resonance frequency for the -mode alone is not sufficient to determine if a cavity geometry produces a reasonable field profile because for very small changes in cavity dimensions, while the frequency may remain the same, the relative field amplitudes in the field profile can change significantly. A figure of merit, such as field flatness, related to the field profile characteristics is needed to differentiate between these cases [25]. Field flatness is a gross characterization of the relative differences in peak field amplitude across a cavity.…”

“…Note that for n cells ¼ 2, 100% field flatness means jE peak j max ¼ 3jE peak j min [25]. Thus, the signed field flatness constraints are not overly restrictive and allow the optimization to consider a variety of relative peak field amplitude configurations.…”