Design principles for high power linear accelerators

The demand for high power proton beams is at an all time high. The global community has identified many applications ranging from spallation sources, material irradiation facilities and secondary beams factories to accelerator driven systems for energy production, transmutation of waste or production of tritium. The typical path to high power beams involves the use of a linac at least in the lower energy stages. For high intensity, high power operation, significant developments are needed particularly in the linac section and the front end of the machine. Consequently, this thesis brings original contributions in tackling several limiting aspects to do with two major pillars of high power operation in linacs: energy and intensity. One of the major decisions in any linac design concerns the choice of normal accelerating structures. The general aim of every designer is to find the optimal path to the final energy without compromising the beam quality or increasing excessively the structure complexity. In the absence of a much needed comparative assessment of accelerating cavities, this choice was often made based on available local expertise rather than solid reasoning. This problem is tackled at length in two chapters of this thesis in which a framework is created for a methodical examination of available structures. The result is the first systematic analysis of normal conducting structures for proton acceleration which includes beam dynamics, electromagnetic, mechanical, thermal and vacuum aspects of cavity design. On the intensity side, several innovative developments have arisen through involvement in the Front End Test Stand Project (FETS) and the ISIS linac upgrade efforts. Beam dynamics studies for the Medium Energy Beam Transport line (MEBT) of FETS, as well as an analysis of several other MEBT designs for existing international projects, highlighted the difficulty in reducing beam loss and emittance growth in high current MEBT lines incorporating beam choppers. Through end-to-end tracking studies, it was shown that the initial beam quality produced in the MEBT will heavily influence the emittance evolution, halo development and beam loss in subsequent structures. As a result, a novel distributed MEBT design is proposed as an alternative, allowing better matching and chopping as well as higher intensity, lossless operation. Further downstream, the beam quality is not only affected by initial mismatch and MEBT beam quality, but by the choice of operating points as well. Theoretical work developed over the last two decades indicate that safe tunes outside conventional equipartitioning limits can be found, but the lack of experimental verification remains a problem. This problem was tackled through an experimental campaign at J-PARC, where for the first time emittance exchange driven by the kz/kt = 2 resonance was measured in a linac with emittance ratios close to 1. Finally, these principles are applied to the design of a future upgrade of the ISIS linac. Three linac options have been developed accelerating the beam to 100, 180 and 800 MeV opening the possibility of MW-level operation in ISIS.