by James Schultz

If they were tops for pans, at $3,000 apiece superconducting niobium cavities would be among the most expensive kitchenware ever created. But the 10 fat, lid-like discs welded together in five pairs serve a different purpose. Were it not for these five-cell cavities, Jefferson Laboratory’s accelerator operations would come to an abrupt halt.

"The cavities are at the heart of the accelerator," says John Mammosser, staff electrical engineer in the Accelerator Division. "They accelerate the electron beam. But the cavities themselves can’t do anything without a bunch of other equipment needed to drive them."

A pair of cavities comprise the innermost components of a three-part system. Encasing each cavity pair is a tank containing the liquid helium needed to cool them. The tank is in turn enclosed by a Thermos-bottle-like structure known as a cryostat. The cryostat provides insulation to guarantee that the cells remain cooled to two degrees Kelvin, nearly absolute zero. Four of the cryostats are assembled together to form one cryomodule; the accelerator contains 42 cryomodules. Each completed cryomodule weighs in at six tons and measures 8.3 meters (approx. 27 feet) in length.

A new seven-cell cavity is being designed by the Accelerator Development Group, headed by Jean Delayen. This copper prototype incorporates design changes that should help boost JLab beam energies to significantly higher levels.

The radio-frequency (RF) power used to accelerate the electron beam within the cavities is provided by a microwave-oven-like device known as a klystron. Normal-conducting cavities would require large and prohibitively expensive amounts of energy expenditure because most of the RF power would be absorbed into the cavity surface, rather than in actual beam acceleration. The solution: supercooled, superconducting cavities made from niobium, which lose nearly all resistance to the flow of electric current, enabling nearly the entirety of RF power to be applied to acceleration of the electron beam.

"A good analogy for superconductivity is a bell," Mammosser explains. "For a non-superconducting cavity, you ring the bell and the sound dies down quickly. For a superconducting cavity, when you ring the bell it resonates for a long time."

Supercooling Wins Out
The decision to commit to supercooling was not a foregone conclusion. Shortly before the Laboratory’s original design was finalized, management learned of niobium-cavity technology developed at Cornell University. Although superconducting technology had not been used on a large scale, plans were made to incorporate the innovation since long-term energy consumption would prove relatively modest.

In all, 360 of the five-cell cavities were ordered from a specialty manufacturer. Fabrication began in 1990, ending by December 1993 with delivery of the final 12-unit lot. Not every one of the 30 dozen cavities were used; 338 ended up in the accelerator proper, with the remainder set aside as replacements. Some have been used in the Free Electron Laser accelerator.

Niobium is ductile and strong, properties needed for ease of cavity fabrication. Despite the material’s advantages, however, manufacturing inevitably leaves behind a thin layer of dust and debris, which degrades performance. Prior to installation, cavities are therefore routinely cleaned and polished chemically, in an acid bath. Nevertheless, this process isn’t able to remove all surface contaminants. So Laboratory specialists have developed a technique known as helium processing.

Helium processing involves injection of helium gas directly into the cavities. When RF power is applied, contaminant particles are heated and spray out electrons, ionizing the helium. Helium ions then backscatter into the contaminant particles, destroying them. "It’s like getting shot by a machine gun," Mammosser says. "Eventually the particles burn themselves out."

The New Generation
For all the improvements to cavity performance in the last six years – the five-cell units are operating 50 percent more efficiently than their original design parameters – the current configuration cannot accommodate the Lab’s long-term plan to boost beam energy to 12 billion electron volts, or GeV, by the middle of the coming decade and then to 24 GeV soon after. Jean Delayen, head of the Accelerator Development Department, says the goal is to improve future cavity performance three-fold, but with no increase in the cost per module.

Peter Kneisel, senior staff scientist (left), and John Mammosser, staff electrical engineer, hook the newly designed seven-cell niobium cavity up to the Network Analyzer. The analyzer will send radio frequency waves through the cavity as part of the cavity tuning process.

"We have squeezed as much performance from the current generation cryomodule as possible," he says. "Now it’s a matter of making new cryomodules with cavities which are longer and then driving them harder."

So Lab designers are developing a seven-cell model cavity intended to operate even more efficiently and at higher energies. The new cavities will be slightly longer than their predecessors. In addition, the basic configuration of the cryomodules will change; the cavities will be assembled in groups of eight instead of in pairs.

For the moment, the Lab is building prototypes in-house. The remainder of calendar year 1999 will be spent testing individual components. In two years, Lab scientists and engineers will begin testing a half-length cryomodule – four seven-cavity units. Depending on funding, seven-cell cavity upgrades could begin in 2004 and conclude by end of calendar year 2007. "We already have a site. We have a machine," says Delayen. "We’re planning the next generation of cryomodules. So it makes sense that our upgrades improve on what we know."

maintained by webmaster@jlab.org