Portland, Ore. Two research groups in Holland have joined worldwide efforts to apply atomic lithography to nanoscale integration of semiconductors. Researchers using this method tackle the usual process in reverse: Instead of forcing light through a physical mask, they focus a physical beam of atoms with a mask made from standing waves of light.
First demonstrated in 1997, atomic lithography has since been applied in labs in the United States, Japan and Europe for beams of matter as diverse as chromium, sodium and aluminum, as well as to indium and gallium. To that list, the Dutch researchers have added iron.
"We have used iron, because we want to make magnetic nanostructures. Our substrate at the moment is simply a silicon wafer, but it is important to note that in principle any substrate material can be used glass, metal, ceramic, even organic substrates," said Eindhoven University of Technology professor Ton van Leeuwen. Researchers in another Dutch group, led by professor Theo Rasing at Radboud University Nijmegen, report similar results.
No exotic materials
Unlike other research groups that are attempting to create coherent beams of matter from exotic Bose-Einstein condensates, a coherent form of matter, the two Dutch groups are working with conventional laser and holographic techniques to create an industrial process that can achieve results impossible for traditional lithography first by building microelectromechanical systems, then by scaling them down to the nanoscale.
As lithographic line widths descend into the netherworld of nanoscale patterning, the search for new approaches has intensified. At these levels, traditional masks transmit less and less light, making it difficult to define features in traditional photoresists. Switching to shorter-wavelength light such as extreme ultraviolet or even X-rays may extend the usefulness of traditional masks, but not enough, according to the two Dutch groups.
"The light field acts like an array of microlenses for the particle beam, spaced at half the wavelength of the light used," said van Leeuwen. "We just position a substrate in the focal plane of these microlenses, and arrays of nanostructures are directly deposited. For a simple standing light wave, we get an array of very fine lines."
Using the technique, van Leeuwen's group directly deposited lines as narrow as 50 nanometers spaced at a pitch that is half the wavelength of the light used, which was 186 nm for the experiment. Rasing's group at Radboud University has similarly laid down 95-nm lines at a pitch of 186 nm by using the same 372-nm light as Eindhoven University of Technology. Both groups are now cooperating in the hopes of slimming down the lines to the 10-nm range.
The groups hope that their choice of iron will allow them to use a combination of magnetic materials to produce novel ferroelectric and spintronic effects in future versions. "Eventually we hope to simultaneously deposit a second material in a homogeneous layer along with the iron. With this co-deposition technique we can produce, for instance, iron-chromium layers where the iron-doping concentration is patterned on a nanometer scale. This is something that can simply not be achieved with the conventional lithographic production of nanostructures," said van Leeuwen.
He described how the technique works: "We start with a beam of atoms produced by heating iron in a high-temperature oven to a temperature of 1,600°C. The iron vapor escapes through a small hole in the oven and expands in a vacuum vessel. This gives us a beam of iron atoms in the vacuum, but it is too divergent to use directly. We then use laser cooling to collimate the beam: We shine laser light from four sides onto the atom beam.
"By choosing the right frequency, the laser light can literally push the atoms in the beam such that they are all traveling in a collimated beam afterward."
The second step is to pattern the collimated beam with a standing wave, or even a holograph, so that the atoms are diverted into the shape of the desired lines on the substrate.
"The iron beam crosses an intense standing light wave, formed by a laser beam that is reflected back on itself. The light that is used for this is almost resonant with an atomic transition. The energy of the atom is somewhat shifted by the intense light. As the standing light wave has nodes [where there is no light and thus no energy shift] and antinodes [very intense light and thus a large energy shift], the energy shift of the atoms depends on position.
"Whenever energy depends on position, forces occur that push the atoms to the positions where the energy is smallest. In our case, these are the nodes of the light field. Therefore, the atoms are focused toward the nodes," said van Leeuwen.
Using this approach, the researchers predict that they will be able to achieve 10-nm line widths. They also want to create more elaborate patterns by going to holography, instead of just simple standing waves. By beaming atoms through a holographic pattern, very complicated circuit diagrams defining, say, tiny magnetic domains and their interconnections, could theoretically be patterned across a wafer.
Ultimately, both groups seek a laser-like particle beam similar to what others have produced from Bose-Einstein condensates what van Leeuwen calls a "single-point writer." With the ability to focus a particle beam on any part of the substrate, theoretically entire circuits could be directly written onto wafers.