Silicon is approaching its limits in our electronics, so researchers are looking for alternative semiconductors that can go smaller, faster, and cooler. One possibility is carbon nanotubes, long chains of carbon molecules just nanometers in width, but they have some challenges to overcome. Researchers at the Stanford School of Engineering have overcome those challenges and created the first carbon nanotube computer.
There are different kinds of carbon nanotubes, including semiconducting and metallic. For use in computers, we want the semiconducting kind, but we are not yet able to grow exclusively the semiconducting kind. As manually sorting through and removing the metallic nanotubes would take too long, the researchers developed a shortcut method. After switching off the semiconducting nanotubes, the researchers pumped enough electricity into the circuit to cause the metallic nanotubes to burn away in puffs of carbon dioxide. This solves the problem of metallic nanotubes, but leaves the problem of misaligned nanotubes, which could cause a short circuit. To work around those, the researchers developed an algorithm that can map out circuits that will work regardless of where any askew nanotubes are.
Armed with those two techniques, the researchers were able to build a computer of 178 transistors, using the university's limited fabrication facilities. This computer is able to perform some basic tasks, such as counting and number sorting, and can run the MIPS instruction set, developed in the 1980s. It may not seem like much now, but this is a great step toward a future of carbon nanotube technologies.
Source: Stanford School of Engineering