It’s a big deal in the realm of the very small.
Scientists at the University of Pennsylvania, in Philadelphia, have invented a microscopy technique to carve tiny gold, silver, and aluminum structures a few nanometers across. Using a special electron microscope, assistant professor of physics Marija Drndi' and her graduate student Michael D. Fischbein have made minuscule structures less than 10 nanometers across.
Besides searching for ways to squeeze more transistors onto chips, researchers are seeking to understand nanoscale physics, which is different from the physics of our everyday world. Because the size scales are so small, quantum effects can dominate.
”Many different approaches have been undertaken to fabricate the small structures needed to probe nanoscale phenomena, but the most widely used and versatile techniques are limited to tens of nanometers,” Drndi' said. ”Reliably and consistently fabricating devices at the sub-10-nanometer scale from the top down is generally still challenging, but our technique offers a route to this regime.”
To prove the versatility of their method, Drndi' and Fischbein have made nanodisks, nanorings, nanowires, nanoholes, and multiterminal nanotransistors.
The roots of the new technique lie in a method that was devised in the 1980s—which has since fallen into disuse—that made it possible to drill holes in crystalline solids. In her lab, Drndi' uses a transmission electron microscope as a tool for nanofabrication, working with thin films of metals. Fischbein was mastering the use of a transmission electron microscope as part of his doctoral research. As he inspected some devices, he noticed that the metals start melting at a beam energy of around 200 kiloelectron volts. When he pointed that out to Drndi', the two of them started thinking of how to put the phenomenon to use and came up with their method, which goes by the formal name of transmission electron beam ablation lithography (TEBAL).
The scientists use the beam of electrons produced inside a transmission electron microscope as a scalpel to carve various metal nanostructures from films of gold, silver, and aluminum. The films of metal they worked with were cut into strips 80 nm wide, with thicknesses ranging from 10 to 50 nm. They managed to make devices with near atom-scale detail.
Compare that to 50 nm or so for conventional approaches. The added advantage is that TEBAL overcomes the problem of how to reliably attach leads to very small devices.
Until now, ”you could do small stuff but couldn’t connect it,” says David Muller of Cornell University, in Ithaca, N.Y., an expert in electron microscopy.
A high-resolution image of a silver lattice that has had a circular region ablated. The crystallinity is well preserved even at the edges of the ablated spot.
As devices on a chip have shrunk, the problem of attaching wires to them has only grown worse. Devices such as transistors made from the more common ”bottom-up techniques”—devices that are assembled from even smaller components—need to have leads attached to them, and that can prove troublesome because the thinner you make the wires that connect the devices, the more their resistance goes up, bringing down circuit efficiency. Because with the TEBAL technique the entire device is sculpted, no leads need to be attached. Thus no additional resistance is introduced in the circuit, and TEBAL-produced devices tend to be more efficient.
Drndi' and Fischbein described their technique in a paper published recently in the journal Nano Letters .
In the past, scientists explored using e-beam techniques to pattern chips at the nanometer level, but those techniques were different from the TEBAL technique. They involved making complex masks and ended up with a small throughput. The e-beam approach lost favor to the extreme-ultraviolet-lithography technique, partly because e-beams were harder to scale to the demands of commercial processes.
A drawback is that TEBAL tends to be slow, as each device is fashioned by a person wielding the electron beam. Drndi' acknowledges the issue but says that the time required to make a device is only seconds.
Because it is dependent on someone steering the electron beam by hand to create new devices, TEBAL is unlikely to become an industrial process like those used to make semiconductor chips. Drndi' says that one could computerize the process to get some more efficiency; however, it is unlikely that it will ever be used on a large scale. But it could prove invaluable in creating test devices and prototypes for study.
Drndi', who trained as a condensed-matter physicist, sees a host of uses for TEBAL, including creating nanofluidic channels and using it to probe the physics of tiny superconducting wires. She even envisions creating small atom traps using magnetic fields—which could open up a new window to understanding Bose-Einstein physics. And, of course, she intends to use it to fashion nanoelectronic devices.
”This is a technique whose time has come,” Cornell’s Muller says.