Using a lens made of atoms to focus light is old. But using light as a lens to focus atoms is new--and may be a way to put very fine circuits on a chip.
Most of the advances in electronics--the ever-smaller chips and faster computers--have been made possible by a process called photolithography. The intricate electronic circuit designs on microchips are etched by light passing through a stencillike mask cut into the shape of the circuit. As small as those circuits are, electronic engineers would like to shrink them still further. But to do so they have to overcome a fundamental limitation of photolithography--the fact that light always diffracts, or fans out, as it shines through the thin incisions of the patterned mask. This spreading of the light beam limits the narrowness of the circuit lines that can be etched onto a microchip.
Harvard physicist Mara Prentiss and her colleagues at AT&T; Bell Laboratories think they’ve found a novel way to overcome this problem and to make electronic circuits a tenth the size of what is possible with photolithography. Instead of etching a circuit with light, they say, it may be possible to build a circuit on a chip with a beam of atoms, using light as a lens to focus the beam. Indeed, the researchers have already used a light lens to deposit micro-size sodium wires onto silicon.
How did they do it? With mirrors--but there’s no deception involved. By reflecting laser light back upon itself with a series of mirrors, the researchers create a standing wave of light just above a piece of silicon. Like a plucked guitar string, standing waves vibrate up and down--crests become troughs, and vice versa, from one moment to the next--but they don’t move forward or backward.
Next, Prentiss and her colleagues position a small, long-necked ampoule of sodium above the standing light wave and heat the ampoule until it releases a stream of sodium atoms. The sodium atoms fall on the standing wave like rain falling on a mountain range. The peaks and valleys in this case are those of the vibrating electromagnetic field that constitutes the light wave--in particular, of the electric component of that field. In the peaks the light is at its most intense and the electric field is strongest.
As the sodium atoms approach the standing wave, the electric field exerts a force on them because the atoms are made of electrically charged particles. The force pushes the atoms away from the peaks and toward the valleys. Basically the atoms fall into the valleys, says Prentiss. Following the path of least resistance, like rain washing down the side of a mountain, the atoms skid into the parallel valleys in the standing wave and come to rest on the silicon surface, forming distinct parallel lines.
With this technique Prentiss and her colleagues have deposited wires less than .00002 inch apart on a silicon chip. That’s about the same scale as the best photolithography, but Prentiss says there is nothing that should prevent her from reaching much smaller realms. Ultimately, she thinks, she can create wires separated by just a few atoms.
We’ve done calculations showing that we should be able to deposit lines ten times smaller than what photolithography does, she says. Her technique has other advantages as well: it’s faster and less complicated than photolithography, which requires a lot of chemical processing to create a photosensitive surface on a chip that can then be etched away by light. You don’t have to do extra chemistry with our technique--you just deposit what you want and leave it there, she says.
There are still two hurdles Prentiss must overcome before her research can find practical application. Although sodium is easy to work with, she needs to experiment with materials used in real electronic circuits, such as gallium, chromium, and indium. And while the light lens nicely focuses atoms into parallel lines, real circuits have much more complicated shapes. Prentiss thinks she can get around this problem by using a more elaborate array of lasers, creating in effect a hologram of the circuit to lens the atoms.
If her light lenses do indeed fulfill their promise, will they represent the ultimate limit in circuit size? From a scientific point of view I’d be loath to say there’s a limit, says Prentiss, and from a deep personal point of view I’d like to think we can always do a little bit better.