To inject energetic electrons into the Thomas Jefferson National Accelerator Facility's free electron laser, scientists at the facility in Newport News, Virginia, built a Photo-Emission Electron Gun designed to operate in vacuum at half a million volts. The gun's "barrel" is a pair of cylindrical ceramic insulators, known as accelerator columns, made of pure alumina and capable of withstanding the high voltage.
No matter how good an insulator may be in bulk, however, its surface is vulnerable to failure. In a strong electric field, free electrons can be accelerated onto its inner surface. A single electron impact may scatter one or two or more electrons, which impact the surface in turn.
As surface atoms lose more and more electrons, positive charge quickly develops and still more electrons are launched. In an instant, an avalanche of electrons is rolling over the surface in a catastrophic flashover.
"Flashovers not only bring operations to a halt, they can do expensive damage," says physicist Larry Phillips of Jefferson Lab. "When we built the electron gun we knew that a big problem would be to operate it at high electric field strength without breakdown."
One way to prevent flashover is to add some conductivity to the surface of the insulator, so charge can bleed away before it builds up. "In our first columns we fired on a thick coat of rare metal oxides," says Phillips, "but the coating behaved erratically and the columns still broke down a lot."
Then, at an accelerator conference held in Washington DC in 1993, Phillips heard a talk by Berkeley Lab's Simone Anders on controlling the surface resistivity of ceramics using metal ion implantation. "It sounded like it might be the answer to my problem," Phillips says, "which shows you the importance of going to conferences!"
Anders, then a scientist with the Plasma Applications Group in Berkeley Lab's Accelerator and Fusion Research Division (AFRD) headed by Ian Brown, described a vacuum-arc ion source of the group's invention which, says Brown, "had shown nice characteristics for certain kinds of implants. It is very good at putting metal into alumina, for example."
In the vacuum-arc implanter, a broad beam of ions formed from a cathode of the metal or alloy to be implanted propagates through a vacuum chamber toward the target. The device was originally built to produce beams of uranium ions for the Bevalac. After the Bevalac was decommissioned "our research evolved toward ion implantation. This kind of ion source and ion implantation works so well it has been copied at many laboratories around the world."
Brown, Phillips, and their colleagues at Berkeley Lab and Jefferson Lab joined forces to investigate the possibilities of using metal ion implantation to lick the flashover problem. They began by implanting small ceramic coupons with titanium, gold, and platinum.
"Ideally you want a good insulator with a surface that's just slightly resistive, so charge can drain off," Brown explains, "and you want the resistivity even over the length of the ceramic."
Tests established that indeed surface resistivity could be controlled by implanting metal ions; with energies at 135 keV (135 thousand electron volts), ions were implanted into the surface of the material to about 300 angstroms deep (30 billionths of a meter). The collaborators decided to use platinum, which does not oxidize and makes a good cathode material.
The remaining challenge was to scale up the process from samples the size of postage stamps to cylinders as big as cooking pots. The Jefferson Lab team built a special cradle to fit on the target end of the implanter, tilted to hold the cylinder at a 55-degree angle to the ion beam and fitted to slowly rotate it as the broad beam played over the entire inner surface. The device was immediately nicknamed the "rotisserie."
Implantation is periodically halted to check the surface, until resistance is lowered to the desired value. "We are used to dealing with targets a hundred times smaller and implantation times of a few minutes to a few hours," Brown says, "It takes several days to process a single accelerator column, running the ion source at maximum beam."
Brown credits physicist Efim Oks of the High-Current Electronics Institute in Tomsk, Russia, currently working with AFRD's Plasma Applications Group in a collaboration funded by the Department of Energy, for "getting the ion source operating at twice its former efficiency."
Three accelerator columns are presently being implanted with platinum ions, in addition to three implanted in an earlier joint effort. Two columns have already been assembled in the electron gun at Jefferson Lab.
"We've had no problems whatsoever," says Larry Phillips; indeed the gun quickly began operating at far higher efficiency than before. So encouraging are the results that a second electron gun is planned.
Jefferson Lab's free-electron laser, which has already achieved energies of 1.7 kilowatts, is being upgraded to far higher 10-kilowatt energies. Larry Phillips says a long line of potential users from universities and industries are already in line for the versatile machine, "a machine I'm very enthusiastic about."
Essential to the upgraded free-electron laser is a dependable high-voltage electron gun. By tailoring the surface resistivity of its accelerator columns through metal ion implantation, the collaborators have met a major challenge to operating the gun at high field strengths without breakdown.