Engineers at a Scottish university have gained an important insight into the behavior of tiny wires, which could have vital implications for the way in which computers in the future are wired together.
Researchers from the University of Edinburgh's School of Engineering and Electronics carried out tests to show how tiny wires — 1,000 times thinner than a human hair — behave when manipulated. Wires on a nanoscale (measured in millionths of a millimeter) act quite differently than bigger wires.
Working with colleagues from the Karlsruhe Institute of Technology in Germany and the University of Rome, the researchers discovered that below certain thicknesses, wires can't be bent into a ring but instead look more like a misshapen paper clip.
The researchers' findings are published in the journal Science, and they have developed a computational tool to predict (and control) the behaviour of these tiny wires when they are bent.
While this may sound a relatively trivial discovery, it is actually pretty important to the development of computers in the future. Indeed, its practical relevance is when the linear scale of microelectronics devices goes down by another factor of 10 sometime in the future.
The problem is not designing a small wire but fabricating it and attaching it without breaking it or making it short-circuit. Getting this right could lead to supercomputers that fit into a device the size of a mobile phone.
The computational tool being developed by the researchers will allow chip makers to predict and control the wiring processes, even as the size of chips and wires falls below one micron (one thousandth of a millimeter). This is an important development for Moore's Law, with its ongoing challenge to keep developing much smaller (and hence faster) microchips and use thinner wires.
"This will help to make small devices much more powerful in the future," says Michael Zaiser of the University of Edinburgh team. "Holding a supercomputer in the palm of your hand will one day be possible — and we are going to make sure all the wires are in the right place."
However, Zaiser stresses that "our research is not about making tiny supercomputers.
"Rather, it is about the wiring that could be used in supercomputers."
"Maybe in 10 years from now, there will be an interesting problem as the conventional way of packaging semiconductors will run into novel material problems," he says.
"Very thin wires behave very differently from a thick wire, and the behaviour changes quite dramatically."