In semiconductor manufacturing and basic science alike, higher-resolution imaging begets progress, whether it’s by yielding clearer views of integrated circuits, magnetic materials, or the innards of biological cells. The highest-resolution microscopes are found in a few powerful synchrotron facilities around the world, where X-ray beams can be precisely manipulated. Now, with the recent demonstration of a high-resolution ultraviolet microscope that fits on a tabletop, industry and researchers may soon have a far easier time getting the images they need.
The resolution of an image ultimately depends on the wavelength of light producing it—the shorter, the better. Ultraviolet radiation can probe smaller nooks and crannies than visible light, for example, which has a longer wavelength. Makers of next-generation microprocessors want to harness the resolution of 10- to 100-nanometer-wavelength UV radiation, called extreme UV, to control the quality of finely etched lithographic masks, the templates used in producing integrated circuits. Such masks would contain features as small as 32 nm, compared with around 65 nm in today’s masks.
Inconveniently, conventional lenses absorb extreme UV light. At synchrotron facilities, where looping electron beams shed intense light of many wavelengths, radiation in the extreme UV and shorter soft X-ray spectrum is focused by passing it through a set of concentric, finely spaced circular ridges called a diffraction zone plate. The circular grating creates dim rings around a bright bull’s-eye.
Synchrotrons have focused X-rays down to 15 nm, but smaller, tabletop lasers operating in the UV band have resolved features only a bit smaller than 100 nm. To push further, researchers from Colorado State University, in Fort Collins, and Lawrence Berkeley National Laboratory’s Center for Xâ''ray Optics, in California, shot pairs of infrared laser pulses from a conventional titanium-sapphire laser at a small silver-cadmium pellet. Each pulse pair boils off a layer of metal and excites the resulting puff into high-energy plasma, which radiates extreme UV laser light at a wavelength of 13 nm.
The laser shines onto a sample and then through a diffraction zone plate into a detector, creating an image similar to those familiar from synchrotrons [see picture, "Test Pattern"]. No single laser burst has enough photons to create a good image, but by flickering light onto the sample for up to 30 seconds to build up an image, the team resolved lines separated by as little as 38 nm.
”This is something that’s been around, but it’s been kind of a physics experiment,” says Terrence Jach, a research physicist at the National Institute of Standards and Technology, in Gaithersburg, Md. ”To actually make it into a practical piece of equipment is impressive. It shows that you don’t need some kind of giant laser laboratory to carry this out.”
The trick to getting enough resolving power from a relatively small setup, says team member and Colorado State graduate student Courtney Brewer, is to get a better alignment of the titanium-sapphire laser, the silver-cadmium pellet, and the diffraction plates. By incorporating finer diffraction plates and a finer initiating laser beam, the system could ultimately resolve details down to 20 nm, she says.
With this result, tabletop UV lasers are finally proving to be capable of serving as candidate inspection systems for future lithographic masks, says Stefan Wurm, program manager for extreme UV strategy at Sematech, an Austin, Texas�based consortium of semiconductor manufacturers. ”Three years ago there were a lot of doubters whether these power levels could be reached,” says Wurm.