Silicon chip tech? Bah humbug. For the future of computer power, you may have to look no further than the graphite in the pencil in your desk drawer.
There's some really very clever innovation going on in semiconductor design, and for the next couple of generations of silicon chips this means we'll be able to jam still more transistors into an ever smaller space on their shiny gray surfaces. Our computers will get faster, and more powerful as a result. But no matter how clever we are at this task, there's a looming stop sign for silicon chips, based on the immutable laws of physics. Simultaneously we're beginning to realize that these simple, boring old boolean algebra-based chips are pretty limiting in their own way. So where next? University of California researchers have just served up an innovation of their own to help answer this, and it's all about wonder material graphene.
Graphene is remarkable in so many ways, from the randomness of its discovery (literally by peeling atom-thick layers of the stuff from a common lump of carbon graphite via some Scotch tape) to the extraordinary number of industries it could revolutionize.
But when it comes to semiconductor chip designs, graphene has a problem: It doesn't have an energy band gap. This is the property of a material that determines, for example, if it's a conductor (with a thin band gap), or a semiconductor (with a thicker and highly tuneable band gap). The band gap determines all sorts of things about how semiconductors work, including the way electricity flows through a transistor. There have been efforts to force graphene to have a band gap, and thus behave like a semiconductor, but these are said to have failed to produce a workable material in different ways.
What the new team, from the University of California, Riverside Bourn's College of Engineering has found, is a way to use the non-band gap properties of graphene to process information in a different way. They've discovered that if you let go of the need to have wholly on or wholly off transistors driving chips that use math based on ones and zeros, you can create "non-Boolean" chips from graphene.
The team looked at what's called the negative differential resistance behavior of field-effect transistors made of graphene. Resistance in an electronic device like a resistor typically means that when you increase the current running through the device, the voltage goes up too. The opposite happens in the graphene devices, which demonstrate dropped voltage as current goes up. The weird behavior actually lets you design strange new transistors that can process different voltage values instead of merely flipping between current and no current for the zeros and ones in a silicon chip.
Considering that graphene works like this on a nanoscopic scale, it could allow future engineers to design extremely tiny non-linear circuits that consume very low electrical power and which flip through the various states needed to perform calculations incredibly fast. No "ones" or "zeros" in sight. This is a lot like the way your brain works, remember. Good job we're working on new programming languages for this tech already, eh?
[Image: Flickr user Chad Kainz]