Silicon chips have been the backbone of the computer industry for decades. To date, nobody has found a compelling alternative for most machines. But silicon chips are reaching their physical limits; they can only get so small before the current in the transistors starts to leak, leaving switches in the on position when they should be off.
Now for the first time engineers at Stanford University have demonstrated a working computer made of carbon nanotubes. It isn’t very powerful, running a billion times slower than a typical desktop or laptop, and it hasn’t got that many transistors — only 178, where even a mobile phone has millions. But it was able to do simple mathematical operations and demonstrate the technology can work. The new kind of computer was described in the journal Nature.
Carbon nanotubes are made of a single layer of carbon atoms rolled into a tube. They’ve been on the radar of computer scientists for years because they can conduct electricity, be turned on and off and are small, nanometer-sized, which makes them ideal with packing into tighter spaces. Essentially they could lead to super-small, super powerful computer chips.
But building a carbon nanotube computer chip has proved difficult. The problem is that some nanotubes — not all — can conduct electricity like metal. That makes them essentially a wire and not a semiconducting material, and wires can short-circuit a transistor. Another problem is that although scientists chemically “grow” carbon nanotubes on surfaces and try to keep them in parallel lines, a few don’t always line up. Those stray tubes can cause short circuits too.
The Stanford group solved the first problem by building a layer of nanotubes and running a strong current through it. The nanotubes that conducted electricity like metal wires vaporized. That left only the semiconducting layer. They then etched away parts of the semiconducting layer to make circuits, just as one would do with silicon. And the scientists were able to get around the fact that some of the tubes didn’t line by using a sophisticated algorithm that compensated for any malfunctions caused by stray out-of-line nanotubes. The result was a working computer chip.
Future research needs to be done to show that the thickness of the carbon nanotube layer can be kept uniform. The circuits also have to be made smaller and it isn’t clear that can be done efficiently. Most computers today have circuits tens of nanometers across; the Stanford team built one 100 times larger than that.
Even so, it could spell the end of the half-century reign of silicon — and computers will keep on shrinking.
Credit: Norbert von der Groeben / Stanford University