Engineers from Stanford and the U. of Southern California have devised a method to design circuits containing carbon nanotubes (CNTs) that should work even when many of the nanotubes are twisted and misaligned.
The work is based on the assumption that with transistors made out of CNTs, invariably some of them will be askew and not work properly. “You want to create transistors out of these things, and hook up these transistors and make them turn on and off independently. But if twisted carbon nanotubes, for example, short out the circuit, you lose the opportunity to do that,” noted Subhasish Mitra, assistant prof. of electrical engineering and computer science at Stanford, in a statement.
Instead, the scientist say they can show how to design chips that will work regardless of the orientation of its CNTs. Starting with a NAND gate (a single circuit element in which nanotube layout is insignificant, they came up with an algorithm that they claim “guarantees” a working design for any circuit element even when “a large number” of CNTs are misaligned — and moreover, with insignificant tradeoffs in cost, speed, or power consumption, they note, adding that their next step is to build and test real circuit elements based on the algorithm.
The key, they say, is modeling the circuit into a fine grid, so that engineers can determine which grid squares the CNTs must pass through and which they shouldn’t, in order to make a design work correctly. CNTs in the so-called “illegal” regions can be etched away or otherwise rendered electrically irrelevant. The Stanford algorithm then automatically determines where these regions should be laid out in the design of a circuit element with a particular function, essentially making them immune to these “illegal” effects. The algorithm can’t guarantee, however, whether a CNT will make a desired connection, the researchers note, adding that CNTs also are hard to predict whether they will conduct electricity instead of switching on and off.