Carving New Frontiers for Ion-Beam Technology

Focused ion beams are important in the semiconductor industry, where they are used to carve structures with dimensions measured in billionths of a meter, repair defects in masks used for photolithography, isolate and analyze elements of integrated circuits, "dope" semiconductors with specific atomic species and perform other tasks.

Focused ion beams have also been used to create images of surfaces, pattern thin films for dense magnetic storage, analyze the chemical content of samples, and investigate biological systems. And because ion beams can shape materials with microscopic precision, they can micromachine miniature medical implants, such as cardiac stents that hold weak blood vessels open.

Complicating these applications, however, is the fact that "problems arise when positive ions are used for imaging or micromachining insulating materials," says Qing Ji, who authored the report with her colleagues Lili Ji, Ye Chen and Ka-Ngo Leung.

Qing Ji is a postdoctoral fellow in the Center for Imaging and Mesoscale Structures at Harvard University and a guest at Berkeley Lab. Coauthors Lili Ji and Chen are with Berkeley Lab's Accelerator and Fusion Research Division (AFRD) and the Department of Nuclear Engineering at the University of California at Berkeley, as is Leung, who heads the Plasma and Ion Source Technology Group in AFRD. The trouble with using positive-ion beams on insulating samples, Ji explains, is that "the target material is charged by the positive ions; as the positive charge builds up on the sample it repels the ions and defocuses the beam."

Traditionally, two methods have been used to keep a nonmetallic sample from acquiring charge from a positive-ion beam, she says. "One method is to pass the beam through a gas cell, where it is partially neutralized before it reaches the sample by acquiring electrons from the gas. The other is to train a separate beam of electrons on the sample."

Both have significant disadvantages. A gas cell may require too much distance between the beam accelerator and the sample, which can interfere with beam focusing. And a separate electron beam requires a separate accelerator, which must be precisely aligned with the ion beam at all times. If the ion beam is scanning the sample, this can be difficult; if multiple ion beamlets are being directed at the sample simultaneously, it's virtually impossible.

"In fact, our new beam system was inspired by one of our group's previous inventions, a multiple-ion-beam system that can steer hundreds of beamlets simultaneously," says Ji. "The device had no room for a neutralizing gas cell, and there was no way to use a separate electron beam to neutralize the sample." A new way to neutralize ion beams
The group came up with a novel solution: Instead of a liquid-metal ion source, standard in many focused ion beam devices, the new system uses two chambers in which plasma is generated by radio-frequency electromagnetic fields, which separate gas molecules into their component electrons and positive ions. The two chambers are divided by an arrangement of electrodes that allows only high-energy electrons to exit the first chamber and at the same time keeps positive ions out.

In the second chamber, an ion beam is formed and accelerated by a lower voltage, which does not impede the high-energy electron beam. Both beams combine in a single mixed beam and are extracted by the accelerator column. The self-neutralizing, mixed beam stays tight on its way to the target, because with electrons present there is little "space charge" -- the positive ions do not push one another apart -- nor does it charge the sample upon striking it.

The combined-beam system can accelerate numerous species of ions, including noble gases like argon, metals like manganese and even molecular ions like carbon-60 "buckyballs," useful in biological studies because of their stability.

In proof-of-principle experiments, the researchers used perforated stencil masks as the forward electrode of the accelerator, causing the beam to transfer the stencil's distinct shapes to the sample.

The dimensions of the shapes could be altered dramatically by establishing an electrostatic field between the mask and the sample. The researchers used argon ions to sputter stainless steel foils with an arc shape, the same length but more than twice as narrow as the aperture in the mask. In another experiment with an oxygen-ion beam, they cut trenches into a graphite sample three times narrower than the mask aperture.

The same technique can be used with 3-D masks, for example a cylindrical mask that accelerates surrounding electron and ion plasma to carve out features in a cardiac stent. Ka-Ngo Leung points out the advantages: "There's no need for scanning, no need to rotate the target. Unlike the way cardiac stents are manufactured now — one at a time, machined by a laser -- ion-beam imprinting would allow hundreds or thousands of stents to be produced with just one shot."

Other current industrial applications that could benefit from imprinting with electron/positive-ion beams include, says Leung, "sound suppressors for jet engines that require millions of holes, which could be produced in one shot. Or cutting the many trenches needed to increase surface area in hydrogen fuel-cell electrodes, which could also be done in one shot."

In these and many other industrial applications involving micromachining, most of which currently employ laser systems, combined electron/positive-ion beams offer an economical way to greatly increase efficiency and throughput.

For more information, visit: www.lbl.gov