The Spallation Neutron Source (SNS) is currently in a planned one and one-half month maintenance outage. During the last beam period, we continued to ramp up operating power, performed extensive beam physics studies, and provided stable production beam to the three commissioned experiments. Production runs were at the 15 to 30 kW power levels on target, with a half-shift run at 60 kW. I am told by the beam physicists that SNS set two records during the last run – most protons stored in an accumulator (1014 protons per pulse, or about 15 microcoulombs), and most intense pulses on a target for neutron production (6.6 microcoulombs).
During this outage one of the primary aims is to improve the reliability of accelerator components that have already indicated they have limited lifetimes or will require high-maintenance attention. It is dose-effective to perform as many in-tunnel improvements as possible while power levels are ramping up and activation levels are low. For example, signal leads for some in-tunnel beam diagnostics are being rerouted into conduits in order to reduce noise pick-up and so improve sensitivity. At the same time additional diagnostic detectors are being added in an area that has proved to be a tuning challenge. Estimated in-tunnel time for the workers performing this task is in the hundreds of man-hours, and so even one mrem/h average field is significant. Early job tracking indicates the workers are being very sensitive to ALARA practices and doses are running below predictions.
The mercury target group is working on repairs to the main mercury pump (a leaking seal and grease oozing from a fitting – pumping a lot of mercury fast is off the performance charts!) and the original manufacturer is reworking the heat exchanger that serves the cold moderators (heat capacity and time between warm-ups need improvement). The main radiological job-planning concern in this case is that the mercury pump is located in a hot-cell-like service bay and remote-handling plans did not include these sorts of tasks. The work area has become significantly contaminated with radioactive mercury and with by-product impurities left behind when the mercury evaporates. In an odd and unexpected development, dose rates are lower in the "service bay" when the activated mercury is left in the circulation loop instead of being drained into its shielded storage tank. It appears that the mercury self-shielding is more helpful than the contribution of the added activation products in the bulk mercury. One contributing factor appears to be that the radioactive by-product impurities float on the surface of the mercury as it is drained and coat the inside of the piping, leaving behind a disproportionate fraction of the radioactive material in the mercury. Some things are just too strange to predict.
A quick factoid from the "we hope we never test this one" file: if you stop a 1-GeV beam with an operating power of 1 MW in a relatively localized area in the beam tunnel by smashing it into a piece of stainless steel (for instance), the resulting local radiation levels are predicted to be in the 10 to 20 megaRad/h range. It should be of some comfort (although we take no credit for it) that the Laws of Physics do not allow that situation to exist for long – you will burn a hole through a vacuum chamber VERY quickly at those power levels. In a future issue I will describe the measures we use to protect personnel against least-incredible accidents such as point loss of the full beam.