Recent advances in carbon nanotube materials have created a buzz in the IT industry of late. Skeptics say it's too soon, and the end product is too expensive. But there are aspects of this nanotechnology that might pay off sooner than you think.
Carbon nanotubes are a form of pure carbon that scientists grow in labs. They look like a powder or black soot, but they're actually hollow strands, thousands of times thinner than a human hair.
Researchers have predicted the material's use in manufacturing tiny, densely packed computer chips, yet that application is decades off on the horizon. But in the near future, scientists could add the materials to a polymer and use the highly conductive properties of carbon nanotubes to dissipate heat from a computer.
According to Danvers Johnston, researcher and a Ph.D. candidate at University of Pennsylvania in Philadelphia, this overlooked area is where the tiny fibers will have an impact on computer components in the short term.
"They have a tremendous ability to transfer heat. These physical properties are going to be what contributes to current technology," Johnston said. "The housing for a CPU has to conduct a certain amount of heat away from it, and it currently uses elaborate fans and cooling techniques. If the processor was sitting on a material that could whisk away the heat -- that would solve a major problem."
According to Johnston, the air outside of the box is going to be much cooler than the CPU, therefore the heat is going to move from inside to outside the server. With better materials based on carbon nanotubes, you could push computers faster without them overheating.
But not everyone is on board yet.
Tom Theis, director of physical science at IBM Research, agrees that the physical properties of carbon nanotube-based polymers may have an impact on server components in the future, though IBM has not heavily invested in it.
"To get any material that competes with metal for heat conductivity, the carbon nanotubes would have to be well aligned in one direction," Theis said.
But Theis said IBM has been researching thermal interface materials, and something non-carbon nanotube-based (he wouldn't specify what) is currently in product development.
The long view
What's more often discussed in the IT media is the potential for incredibly dense and powerful computer chips, based on carbon nanotubes. Johnston called harnessing properties of carbon nanotubes the Holy Grail of circuitry.
Carbon nanotubes grow in a random mixture of semiconductive and metallic strands. What that means for circuits is that they can act as both wires (metallic) and transistors (semiconductors) on a tiny computer chip.
But it's a long way off.
"In 10 years there could be an inkling of what applications are available," Johnston said. "Making circuits becomes challenging because we're trying to line up single molecules in the right place. The steps we've taken have allowed us to make a transistor, which can lead to more complex circuits that can actually compute. But to get to the point where we would have a useful device would take millions of transistors. It's a big gap and there are a lot of baby steps.
IBM started looking to replace silicon chips with carbon nanotubes in 1998, and it is still very early in the process. But Theis thinks the research will pay off.
"They have promised to outperform current technology and lower the operating power," Theis said. "Each new generation of electronics gets closer to the interruption of Moore's Law, and power dissipation has become the dominant issue."
The challenge to commercialization
The current technology for carbon nanotubes includes breakthroughs in the ability to create carbon nanotube-based semiconductors, as well as a spun-yarn material consisting of strips one-thousandth the thickness of a human hair.
Companies are even integrating small amounts of carbon nanotubes into plastic packaging to ground static electricity when shipping sensitive electronic components.
But according to Charles King, principal analyst with Hayward, Calif.-based Pund-IT Research, the commercial success for carbon nanotube products is going to depend on the capability to manufacture them reliably and cheaply.
"Carbon nanotubes will probably work in theory," King said. "But you can build the best technology in the world, and if it costs too much to produce it, people won't buy it."
The price of commercial computing is dropping, and King is skeptical that manufacturers will be able to produce carbon nanotube components cost efficiently enough to compete.
"It's still very early in the development cycle," King said. "They've just gotten to the point where they can construct them. Now they have to demonstrate that they can be produced reliably in commercial quantities. Theory is one thing, practice is another and commercial practice is yet another."
Let us know what you think about the story; e-mail: Matt Stansberry, News Editor