CERN Accelerating science

The “kladistron project” for high-efficiency klystrons
By Juliette Plouin, Antoine Mollard and Franck Peauger (CEA)
Figure 1: Schematic view of the standard 5 GHz TH2166 klystron (top) and of the 5 GHz kladistron (bottom)

Under the framework of EuCARD-2, the study of high efficiency - high frequency klystrons is in progress in close collaboration between CEA Saclay and Thales Electron Devices. The “Kladistron project” aims to design a high-power 12GHz klystron to be used in CLIC.  

The objective of the Innovative RF Technologies activity is to propose affordable and reliable solutions to increase efficiency and thus save high amounts of energy in accelerators. These will be essential for future testing capabilities for the CLIC accelerating structures. To achieve this, a new klystron concept is required. A new electron beam bunching technique has been proposed to increase the beam-to-RF interaction efficiency above the state-of-art limits (> 65%). This method is inspired by the Radio Frequency Quadrupole (RFQ) based injectors which bunch adiabatically a CW (continuous mode) hadron beam with more than 95% transmission. The new proposed device, called kladistron, includes a large number of bunching cavities (typically 15 to 20, which is twice as many as in a conventional klystron) separated by short drift tubes. This enables giving a “soft kick” to the beam in a large number of cavities instead of a “large kick” to the beam in a small number of cavities. The coupling between these cavities and the beam (r/Q) is lower than in a standard klystron. Preliminary calculations at this frequency, made with a 1D code, have given efficiencies around 60 % with 10 cavities and 70 % with 20 cavities.

The first experimental tests will be realized on a 5 GHz demonstrator, constituted of a standard klystron from Thales, the TH2166 (6 cavities, CW mode, 4.9 GHz, 26 kV, 4.3 A, efficiency 50%), in which the intermediate cavities are modified. All the other elements: gun, solenoid, collector, input and output cavities and window, are kept as is and the 4 intermediate cavities are replaced by 14 new cavities with lower r/Q. The overall length is kept the same, to fit in the existing solenoid, and the new cavities are very close to each other. The principle is shown in Figure 1, where the cavities pattern is placed on a schematic klystron for both standard klystron and kladistron. The calculations with Magic (2D Particule In Cell code) give an efficiency equal to 50% (as expected) for the standard klystron, and equal to 60 % for the kladistron. The beam shape given by Magic at the end of the simulation is also shown in Figure 1. One can see that the beam bunching is softer for the kladistron than for the standard klystron

The design of this kladistron demonstrator is nearly ready, and the fabrication and tests are planned in 2016. The aim is to verify that a higher number of low-coupled cavities lead to higher efficiency, all other elements being the same. Once this principle is being validated, it will be possible to reach higher efficiencies with thorough optimization of the electron bunching.