Creating computers that use 10,000 times less energy
A team of scientists in Edmonton, Canada has done just that.
University of Alberta PhD student Taleana Huff teamed up with her supervisor Robert Wolkow to channel a technique called atomic force microscopy - or AFM - to pattern and image electronic circuits at the atomic level.
This is the first time the powerful technique has been applied to atom-scale fabrication and imaging of a silicon surface, notoriously difficult because the act of applying the technique risks damaging the silicon.
However, the reward is worth the risk, because this level of control could stimulate the revolution of the technology industry.
"It's kind of like braille," explains Huff. "You bring the atomically sharp tip really close to the sample surface to simply feel the atoms by using the forces that naturally exist among all materials."
One of the problems with working at the atomic scale is the risk of perturbing the thing you are measuring by the act of measuring it.
Huff, Wolkow, and their research collaborators have largely overcome those problems and as a result can now build by moving individual atoms around: most importantly, those atomically defined structures result in a new level of control over single electrons.
This is the first time that the powerful AFM technique has been shown to see not only the silicon atoms but also the electronic bonds between those atoms.
Central to the technique is a powerful new computational approach that analyzes and verifies the identity of the atoms and bonds seen in the images. "We couldn't have performed these new and demanding computations without the support of Compute Canada. This combined computation and measurement approach succeeds in creating a foundation for a whole new generation of both classical and quantum computing architectures," says Wolkow.
He has his long-term sights set on making ultra-fast and ultra-low-power silicon-based circuits, potentially consuming ten thousand times less power than what is on the market.
"Imagine instead of your phone battery lasting a day that it could last weeks at a time, because you're only using a couple of electrons per computational pattern," says Huff, who explains that the precision of the work will allow the group and potential industry investors to geometrically pattern atoms to make just about any kind of logic structure imaginable. ■
What to read next