Tuesday, June 15, 2010

Image Courtesy of:  condmat.physics.manchester.ac.uk/pictures/


Image Courtesy of: condmat.physics.manchester.ac.uk/pictures/

Graphene is Next

The Substance Found in Pencils will Speed Up our Computers One Thousand Fold


Graphene. If you’ve never heard about it, don’t worry, a lot of people haven’t, because it’s really only been “discovered” relatively recently, and most of the truly interesting news about it has been in the last year. The amazing thing is that we’ve actually been using it for centuries, in the form of the common pencil. Graphene is a form of carbon, much like carbon nanotubes and other fullerenes, with one major difference. While fullerenes are 3D structures of carbon atoms, graphene is a flat sheet. It’s a 2D lattice of carbon with bonds as strong as diamond. It’s this sheetlike nature that makes it so useful in a pencil. As you write, individual planes of graphite are sheared off the end and deposited on the paper. Those individual planes are pure graphene.

By now, most of you are familiar with carbon nanotubes, a.k.a. CNTs, and their potential for computers. Graphene has equally amazing properties, including some that might make it far more readily usable than CNTs. First, like CNTs, graphene is capable of conducting electricity with much less resistance than copper. That alone makes it useful, but graphene has even more interesting properties. As New Scientist reports bending graphene creates strains between atoms that can create isolated pathways which then act as nanoribbons — wires — within the still connected sheet. In other words, the morphology of graphene affects its electrical properties: change the flat sheet by bending parts of it, and you change how electricity flows through it.

But that isn’t all. The pattern of carbon bonds has effects as well. Graphene is a hexagonal grid of carbon, much like a roll of chicken wire. Remove one random atom from the pattern every so often, and graphene can exhibit magnetic behavior without needing the presence of magnetic metals. Adding hydrogen into the mix creates graphene’s non-conductive cousin, graphane. Taking precisely defined patterns of atoms out of the sheet can create well-defined circuits, creating wires that are almost superconducting.

Image Courtesy of: en.wikipedia.org/wiki/Graphene
Image Courtesy of: en.wikipedia.org/wiki/Graphene
All of these properties make graphene a very important material for the future of electronics. It has already been used to create field effect transistors, the primary component of a computer processor. When you combine this with the other features above, you have a single material that could be used for the majority of the components in every electronic device we currently have… with one major difference: speed. Current silicon based chips have a limited speed at which they can run at room temperature without overheating and malfunctioning. Go much over 3GHz without some major cooling and chips melt down. But replace those chips with graphene equivalents — without having made any other changes to the circuits -—and you can raise that limit much higher. Potentially 100 to 1000 times higher.

Let’s think about that for a moment. That’s 300GHz to 3000GHz or 3Terahertz.

That’s a jump of two or three orders of magnitude up the exponential curve, my friends, especially when you combine it with the advances in multi-core technology and parallel computing. We’re talking about that smartphone in your pocket having a thousand times the computing power of your desktop PC, but using no more power than it does right now. The resistance of graphene at room temperature is so much lower than copper and silicon that even though it’s running at 1000 times the speed, it’s not using any more current, or wasting any more energy as heat than an identical silicon device, and that’s without considering any other possible advances in the field of electronics design.

We’re talking about that smartphone in your pocket having a thousand times the computing power of your desktop PC, but using no more power.

That big a leap in processing speed will simplify a lot of extremely complex tasks that require extensive amounts of data. From SETI searches for extraterrestrial intelligence to the search for all the ways a protein can fold, scientists use millions of processors in parallel to speed up research. A thousand-fold increase in computer speed could cut months to years off the time needed for their projects. The same goes for DNA sequencing, data mining, and a host of other areas.

And science will not be the sole benefactor. Most smartphones these days have the ability to use their cameras to create virtual overlays on the images that they see, a technique called Augmented Reality. AR has advanced to the point that it’s possible to create virtual characters in photos on your phone using nothing more than a 2D patterned target on the ground, or to create interactive “virtual assistants” in projected video that are capable of interacting with real world objects. Ultrafast computers will be essential for ushering in the age of Virtual Reality.

Image Courtesy of:  en.wikipedia.org/wiki/Graphene
Image Courtesy of: en.wikipedia.org/wiki/Graphene
A massive increase in computer speeds is likely to benefit other complex computing tasks as well, such as real-time speech language translation. Right now, it is difficult to make these programs run quickly enough to be useful. A thousand-fold increase in computer speed could make brute force approaches a practical solution, enabling computers to crunch through entire dictionaries in milliseconds. It could make possible the elusive conversational interface that so many people believe will be the next step in operating systems. That speed will also be useful in the next generation of robotics, quite possibly bringing us a step closer to the kind of robots seen in movies like I, Robot or Star Wars. Ultrafast computers would enable a major reduction in the size of the computers needed to run some of the most complex robots we currently have, bringing the day of Rosie the Robot maid that much closer.

Obviously, ultrafast computers are going to have a very far-reaching effect on the way we do things, as well as how we interact with each other and our world, so the real questions are how practical is it to make graphene chips, and how soon can they be made? The answer is probably going to surprise you. Graphene has already been proven to be usable in current chip manufacturing processes with only minimal retooling needed. In fact, IBM has already created working 30GHz test devices using graphene transistors. In other words, graphene could begin making its way into computers as early as 2012 to 2015, and almost certainly by 2020.

Graphene, that same single-atom-thick layer of carbon that is a part of every pencil mark, is going to make all of this possible. Not bad for the humble Number 2, huh?

No comments:

Post a Comment