After tooling up, production of the CEBAF cryomodules ran at a rate of two per month for most of 1991 and 1992, ending in 1993. 41 eight-cavity cryomodules plus the short injector “cryounit” were in place and all operational at 2 K in 1993. The cryomodules easily met the design requirements for a 5-pass 4 GeV, 200 microamp beam. As often occurs, there was soon serious interest in ascertaining and exploiting all available capacity to extend the physics reach of the machine.
An effort was launched to more carefully assess the performance limitations of each installed cavity and RF system. This effort yielded a useful capacity to support ~5.2 GeV operation at less than the design current. The chief limitation was found to be periodic arcing trips associated with the cold (2 K) ceramic RF windows attached to the cavities. When cavity gradients are pushed up into a regime in which residual contaminants produce electron field emission, an effect is charging of the very high resistivity ceramic material in the windows. This charging process eventually relaxes through flashover or arcing. In this event, the supplied RF power - and thus the beam delivery - must be interrupted to avoid sustaining damage. The periodicity of the induced arcing then rises steeply with operating gradient of the affected cavity.
To combat this operational limitation, most of the cavities received in-situ processing to increase the field-emission-limited gradient. This occurred during 1996-1998. In the end, the same phenomenon limits the integrated cavity performance, but the top physics reach then exceeded 5.7 GeV.
During the more than 11 years of operation, the cryomodules in CEBAF have been one of the most reliable systems, accounting for very little lost beam time. The two components known to age are the rotary feedthroughs associated with the tuners and the warm polymeric RF windows, which slowly yield with time and can fail. Both of these components are serviceable in-situ and are part of regular maintenance.
The original CEBAF design includes seven indium wire seals per cavity between beam vacuum and superfluid helium, and one between the input waveguide vacuum and helium. Over time the mechanical flexing associated with thermal cycling presents some risk of creating a leak across these joints. One such leak to beam vacuum is present in the cryomodule in zone NL11. The condition is managed operationally by “burping” the adsorbed helium at 15 K roughly every six weeks. Two or three cavities have been lost to use due to helium leaks into their input waveguide spaces. Servicing these leaks will require removal from the tunnel and disassembly of the cryomodules. All CEBAF upgrade cryomodules designs have eliminated all helium-to-vacuum seals.
SRF performance of the cavities in CEBAF has been quite stable. No long-term degradation of the cavity Qs (reciprocal to dissipated heat) has been observed. The principal challenges remain the preservation of particulate free conditions in the linac beamline and vacuum integrity.