Engineers are near the outer limits of what can be done with optical lithography, the process by which light shone through a patterned mask defines the fine structures of microprocessors and memory chips. Now three teams of optics experts have independently hit upon what could turn out to be a way to extend optical lithography’s use—and, what’s even more critical, to do it cheaply.
All three methods are inspired by the seminal work of Stefan Hell at the Max Planck Institute for Biophysical Chemistry, in Göttingen, Germany. In 2005, Hell managed to push the resolution in an optical fluorescence microscope, used mostly in biology, well beyond its expected limits.
In fluorescence microscopy, the resolution is determined by the size of the spot a laser pulse makes on a material. Hell and his colleagues greatly reduced the size of the spot by following the first laser pulse with a second pulse of a longer wavelength. Tuned properly, the second pulse can create a ring of light instead of a spot. The interaction of the electromagnetic fields of the two pulses compresses the first laser’s spot. Hell called the technique stimulated emission depletion (STED) microscopy.
Inspired by STED, Rajesh Menon and his colleagues at MIT have developed a photolithography method in which they place a thin film of material on top of a photoresist—a liquid that coats the silicon wafer and hardens where it is exposed to light. This film has the unusual property of being transparent to ultraviolet light but opaque to visible light. The researchers shoot lasers of both colors at the film simultaneously, which creates a small, temporary transparency in the film and sets the resist below. Menon used the process to make features 35 nm across. He says it should ultimately be possible to etch 10-nm features.
Two other groups engineered the photoresist rather than add a thin film. The resist contained molecules that promoted hardening when exposed to blue light and other molecules that inhibited hardening when exposed to UV light. Using both colors of laser at once is like drawing a line with a thick pencil and erasing the edges simultaneously. ”We are basically getting the material to respond to the difference of the two light beams,” says Robert McLeod, the University of Colorado assistant professor of electrical and computer engineering who led one of the research groups.
John Fourkas and his colleagues at the University of Maryland used a similar approach, except that both of their laser beams were of the same color. The laser used to cause polymerization produced short bursts of light, while the laser used to inhibit polymerization was beamed continuously.
The cost of two-laser lithography should be a fraction of that for extreme ultraviolet lithography, to which chipmakers plan to shift in the coming decade, says David Back, who managed advanced semiconductor technology development programs at the Albany Nanotech R&D center, in New York. ”So far the economics look good.”