• Number 372  |
• September 24, 2012

Jefferson Lab prepares for new era of exploration

Hollywood’s finest tradition is to follow up a smash hit with a much-anticipated sequel, and so it will be with the Department of Energy’s Jefferson Lab and its Continuous Electron Beam Accelerator Facility (CEBAF). On May 18, CEBAF shut down after a long and highly successful 17-year run, during which scientists completed more than 175 experiments in the exploration of the nature of matter. The sequel will feature a return of CEBAF with double the energy and a host of other enhancements designed to delve even deeper into the structure of matter.

First proposed in detail in 2001, the upgrade is a $310 million project that will enhance the research capabilities of Jefferson Lab's CEBAF accelerator by doubling its energy from 6 to 12 billion electron volts, or GeV, along with other upgrades and additions. The CEBAF 12 GeV Upgrade will provide scientists with unprecedented precision and reach for studies of the particles and forces that build our visible universe.

In the wee hours of May 18, the last crew to run the 6 GeV CEBAF accelerator caught a glimpse of the future: Its newest section of hardware had operated for an hour at its full design energy for the first time. That meant while the machine was about to be shut down for well over a year for an upgrade, the core acceleration technologies that would make the upgrade a reality had just passed an ultimate test.

“That run was the culmination of a decade's worth of effort,” said Leigh Harwood, associate project manager for 12 GeV Upgrade accelerators.

The May test focused on new sections of accelerator, called cryomodules, and their associated systems.

“We did it literally years ahead of when we were supposed to have demonstrated it. So that's why it was a big deal,” said John Hogan, a staff engineer in the lab’s SRF Institute. Hogan is responsible for the final design, parts acquisition, manufacturing, initial testing and installation of the new cryomodules, called C100s.

Building a Module

At the core of a cryomodule are eight cavities: hollow metal cylinders with seven connected compartments that roughly resemble a stack of doughnuts. These cavities are designed to harness energy that is pumped into them, focusing it onto, and thus accelerating, a thin beam of electrons.

Before the cavities can be built into a cryomodule, their interior surfaces must be thoroughly cleaned and prepared using an acid wash. SRF Production Manager Tony Reilly said that exposing the surfaces to acid both smooths the surfaces and helps to strip away impurities. For the acid wash, there were two choices: buffered chemical polish, which had been used to prepare the cavities in the original CEBAF accelerator, or a relatively new procedure called electropolishing.

“At JLab, electropolishing has become a more mature surface treatment process. In our application as a final chemical polish, EP has shown to be stable and repeatable which, based on cavity performance, seems to have translated into reproducible cavity surface finishes,” said Reilly.

The system that was used for electropolishing the cavities was originally intended for use with cavities that Jefferson Lab processed for the Spallation Neutron Source at Oak Ridge National Lab. However, the technology and the process wasn’t ready to be used on those cavities. But since that time, Jefferson Lab staffers gained valuable experience and improved the process.

“We added a spray cooling system to the electropolish cabinet to control the temperature of the process,” Reilly said. “Now, the process for each cavity is more tightly controlled, which seems to give us a more consistent surface finish on the inside of these cavities.”

The cavities are made of a metal called niobium. At just a few degrees above absolute zero, niobium cavities become superconducting, allowing energy to flow through them without resistance. The cavities are encased in a helium vessel, where they are bathed in a constant stream of liquid helium to keep them cold. The vessels are cocooned with a variety of insulating material, including aluminized mylar, the stuff of shiny party balloons.

Tuners are attached to the ends of the cavities, which allow for adjustment of the cavities’ ability to harness the energy pumped into them. Other mechanical necessities, wires and parts are suspended in a space frame, and the entire assembly is then encased in the outer stainless steel shell, sometimes called the vacuum vessel.

“From a single cavity to a finished cryomodule is easily six months,” Hogan said. “A cavity alone takes two months to get through the whole qualification process. And once you have eight cavities qualified, you build them into a string. Once the string is complete, it takes nominally three months to build it out from a string to a complete cryomodule.”

Once complete, a cryomodule is three feet in diameter and 30 feet long. It weighs in at roughly 12,000 pounds and contains at least 1,000 major parts, many of which were designed at Jefferson Lab and are now patented technologies.

Testing and Installation

To make sure that a completed C100 cryomodule is in good working order, it is then tested by a team led by Mike Drury.

“Over in the Test Lab, we have a facility called the Cryomodule Test Facility - it looks like a big cave. When they finish building the cryomodule and they're satisfied with it and done all of their pressure tests and leak checks and stuff like that, it goes into the cave. We hook it up and run our series of tests,” Drury said.

Each cryomodule is first hooked up to systems analogous to those it will connect to in the accelerator. Then it is slowly cooled, its helium vessels filled with liquid helium for the first time. Next, the cavities are tuned, adjusting the cavities’ ability to harness the energy pumped into them. Finally, Drury and his colleagues will send energy into each cavity, a kind of dry run of its ability to harness energy. The full process takes about four weeks.

Once a cryomodule has cleared the tests in the cave, it is transported to the accelerator site, where it is installed in the CEBAF accelerator.

“It's about a half-day operation to get it out there and get it installed on the stand. And then you have another group of guys that make sure everything's lined up right. The vacuum guys come in, and they hook up the actual beamline. Then we hook up the waveguides. Then we have to connect and check out all of the instrumentation, make sure everything works, and then we'll do the cool down,” Drury explained.

Powering Up the System

The first two 12 GeV-style, C100 cryomodules were installed during the six-month down in 2011. The down preceded the last months of running of the 6 GeV accelerator. (The final 6 GeV run was January - May.)

“The plan, back in 2005, was to have them all put in during the Long Shutdown. But we were making good-enough progress, that we were confident that we could try to get them in for this run and try to get a little experience with them,” said Hogan.

For some of the run, the new cryomodules were handled by the accelerator operators. But operating the cryomodules during the tests fell primarily to Curt Hovater, a senior staff engineer in electrical engineering, and his group.

Hovater’s group designed the new radiofrequency system that powers the cavities. They completely re-worked the original system, taking it from analog to a mostly digital system. The digital system includes new klystrons, which provide the energy that is sent into the cavities, a new digital interface to control the system and a new cooling system for the amped-up equipment.

“When we are running our tests with beam, we're there to mainly operate the C100 cryomodules. We need experts there to track what's going on and try to minimize any types of trips or incidents,” Hovater said.

According to Arne Freyberger, head of Accelerator Operations, the first real tests of the cryomodules in the CEBAF accelerator began in November 2011, when the accelerator was not scheduled to deliver beam for experiments. “We had a dedicated week in November for testing of the modules without a user. And those tests went reasonably well. But then when we first turned them on to send beam to the user, the beam wasn't quite good enough, and we had to turn them off,” he said.

Through the winter and spring, Hovater’s group worked to integrate the new RF system with the new cryomodules. The group also had to make sure the new equipment would work well with the older control system used to run the rest of the accelerator. The group calibrated measurements of the cavities’ performance, improved the modules’ up time and shortened the time it takes to recover from an incident that forces them to shut down - from a half hour to just six minutes.

“A whole team of individuals since September have been working on procedures and algorithms to make the cryomodules useable for operations,” Hovater said.

Last-Minute Test Success

All the hard work finally paid off on the morning of May 18, Hovater’s group staffed the controls of the new cryomodule dubbed SL25. Approaching an hour after midnight, they tentatively ramped the module up to its full design specification and ran it there for a minute. It successfully ran at its full specification, imparting 108 Megavolts of power to the electron beam (C100 refers to this ~100 MV design specification). In comparison, the average original CEBAF cryomodules impart just 20 MV of power to the electron, with the best original module reaching just 32 MV.

Then came the real test: could the cryomodule run at full specification, while also delivering the most demanding beam ever required of CEBAF for experiments in two halls simultaneously?

The team attempted the one-hour test three times in quick succession. The third proved the charm. At 2:55 a.m., the cryomodule had achieved an hour of stable running at full specification.

"This test has demonstrated that the integrated cryomodule plus microwave-power system can successfully deliver the performance that was envisioned in 2001, and which is needed for the planned nuclear physics research program in the 12 GeV era," said Leigh Harwood, associate project manager for the 12 GeV Upgrade project. "We've demonstrated that there are no fundamental design problems that we overlooked. We now know we can rely on this critical new technology."

Freyberger agrees.“This was a significant accomplishment, which I've already said, and it took a lot of work. It also took a lot of cooperation with the users as well, the scientists. And so, I thank them for their patience. And I think it shows that we can do more than one thing at once here, and do it all exceptionally well,” he said.

The Upgrade Continues

Now, plans for the debut of CEBAF's much-anticipated sequel are are underway. The CEBAF accelerator was shut down at 8:18 a.m. on May 18 for the Long Shutdown. Over the next year-plus, Jefferson Lab staffers will be busy preparing CEBAF - the accelerator, the experimental halls, the cryogenics system and all other related systems - for operations at 12 GeV.

The racetrack-shaped CEBAF originally operated with about 20 cryomodules in each of its two straight sections. For the upgrade, each straight section will get an additional five cryomodules, for a total of 10 new modules. All of the components of the new radiofrequency system, along with the many other support components, will also be installed. Many of the magnets in the accelerator, used to control and steer the electron beam, are being refurbished so they can handle electrons at higher energies. The cryogenic system, needed to keep the cavities superconducting, is also undergoing a major upgrade and maintenance.

So far, six cryomodules have been completed, and the rest are in various stages of construction. This fall, Harwood expects that half of the new cryomodules will be installed. By next fall, the machine and all its systems will be gearing up for the first, weeks-long commissioning run of all of the new modules and systems.

“The effort that was put in from November to May in understanding all the control issues to get the cryomodules working correctly, that effort was very productive and resulted in these modules reaching their design energy,” Freyberger said. “In November 2013, when we come to commission this machine, we're going to be so much further up the learning curve in how to run this machine. The C100s are the major new component to the CEBAF accelerator. There will be new magnets and there will be new power supplies for magnets, but those are not as challenging to commission as a new cryomodule. So to get this far up the learning curve is a tremendous advance.”

[Kandice Carter, 757.269.7263,
kcarter@jlab.org]