2 HIGH-GRADIENT TESTING AND RF BREAKDOWN PHENOMENOLOGY

Accelerating gradients in excess of 100 MV/m are essential for future multi-TeV linear colliders in order to limit their overall length and cost. Experience has shown that to obtain such high gradients with normal-conducting (nc) structures requires operation at a relatively high frequency. A brief review of the very high gradient test work being carried out around the world will be presented with special emphasis being placed on the effect of frequency on the maximum achievable gradient. In the framework of the CLIC study at CERN, evidence has been found during high gradient tests at 30 GHz that part of the limitation on gradient is due to the choice of copper as the standard cavity fabrication material. To support this hypothesis, a breakdown mechanism leading to material destruction is presented together with a review of results and experience obtained with alternative materials. High-gradient test cavities are built covering a wide range of designs from single-cell standing-wave (SW) geometries to multi-cell travelling wave (TW) geometries. High-gradient tests are performed at different frequencies using, as a consequence, different types of power sources and different criteria for the conditioning procedure. These differences make it difficult to compare results, draw conclusions or even to extract scaling laws. It is therefore useful to adopt a phenomenological approach, and to ask the following fundamental questions. • What is a high-gradient test? A typical high-gradient test consists of a cavity equipped with directional couplers on the input and output RF transmission lines to facilitate the calculation of the power balance. Missing energy in the power balance is a very good indicator of RF breakdown. Traditionally a huge input reflection was an indicator for breakdown but measurements with short pulses at 30 GHz have as well as long pulses at 11 GHz showed that missing energies of up to 50 % are possible without an increase in the reflected signal. Fig. 1. shows an example of a 16 ns RF pulse at 30 GHz measured in the CLIC Test Facility (CTF II) [3].