Sparking rates measured on the CRITS RFQ

During the test of the LEDA injector on the CRITS RFQ, an automatic data acquisition system has been implemented. The purpose was to measure the sparking rate of this CW RFQ. The RF level has some influences on vacuum, but there is no evidence of any reciprocal effect. The raw sparking rate is very difficult to interpret, since burst of sparks bias the statistics. A more convenient and useful interpretation is the number of sparking seconds. At the nominal field level (1.75 Kilp), the sparking-second rate is 0.5 per minute without beam. It strongly depends on the field, with a logarithmic law: 4.5 decade/Kilp. With beam, the sparking rate jumps to 3.0 per minute. As far as tested, it depends neither on the beam current (20 to 80 mA) nor on the field (1.5 to 1.7 Kilp tested). With sparking rates as measured here, one could not hope to build an RFQ that would be free of sparks over a several months continuous operation. Such a requirement, based on an extrapolation of the curves presented here, would lead to a maximal electric field much lower than the Kilpatrick value, an unreasonable requirement for a functional RFQ. A conclusion is that a sparkless RFQ is hopeless, even with a very carefully conditioned cavity. It will probably be necessary to deal with a few sparks per day, and the linac must be able to restart automatically after a short beam interruption