Carbon Nanotubes Could Solve Overheating Problem for Next-Generation Computer Chips

Carbon nanotubes could remove the speed limit on new computing devices

2 min read
Carbon Nanotubes Could Solve Overheating Problem for Next-Generation Computer Chips
Image: Chip, Ice: iStockphoto; Nanotubes: Arnero/Wikipedia

Computer chips used in next-generation smartphones and supercomputers can't get much faster without overheating. That's why engineers hope carbon nanotubes offer a possible cooling solution that could enable processing speeds to continue accelerating.

The overheating problem has become steadily worse as engineers cram more power-hungry transistors into the same microchip space, because much of the electricity that powers the transistors is wasted as heat. Carbon nanotubes have high thermal conductivity that could carry the excess heat away from the microchips, but only if engineers can figure out how to improve the heat transfer at the point of contact between the nanotubes and microchips.

"The thermal conductivity of carbon nanotubes exceeds that of diamond or any other natural material but because carbon nanotubes are so chemically stable, their chemical interactions with most other materials are relatively weak, which makes for  high thermal interface resistance," said Frank Ogletree, a physicist with the Lawrence Berkeley National Laboratory’s Materials Sciences Division, in a news release.

Ogletree and his colleagues worked with two former Intel researchers to figure out how to make a six-fold improvement in the heat flow between metal and carbon nanotubes. Their work is detailed in the 22 January issue of the journal Nature Communications.

The new study's success rests upon using organic molecules as a bridge between the carbon nanotubes and metal—a method that greatly reduces the interface resistance that would otherwise prevent heat from flowing more efficiently between the materials. The organic molecules, including aminopropyl-trialkoxy-silane (APS) and cysteamine, create strong covalent bonds between the carbon nanotubes and the metal used in microchips. (The same bonding technique pioneered by the researchers can also work with graphene—a promising material for complementing silicon transistors.)

"With carbon nanotubes, thermal interface resistance adds something like 40 [micrometers] of distance on each side of the actual carbon nanotube layer," said Sumanjeet Kaur, lead author of the Nature Communications paper and an industrial postdoctoral scientist at Porifera. "With our technique, we’re able to decrease the interface resistance so that the extra distance is around seven microns at each interface."

This success will help pave the way for carbon nanotubes' use in this application. But there is still a ways to go before we see them in commercially-available gadgets. One problem is that most nanotubes, grown in vertically-aligned arrays on silicon wafers, don't make contact with the metal surfaces. But the Berkeley team hopes to improve the density of the contacts between the nanotubes and metal over time.

Image: Chip, Ice: iStockphoto; Nanotubes: Arnero/Wikipedia
The Conversation (0)

The Future of Deep Learning Is Photonic

Computing with light could slash the energy needs of neural networks

10 min read
Image of a computer rendering.

This computer rendering depicts the pattern on a photonic chip that the author and his colleagues have devised for performing neural-network calculations using light.

Alexander Sludds
DarkBlue1

Think of the many tasks to which computers are being applied that in the not-so-distant past required human intuition. Computers routinely identify objects in images, transcribe speech, translate between languages, diagnose medical conditions, play complex games, and drive cars.

The technique that has empowered these stunning developments is called deep learning, a term that refers to mathematical models known as artificial neural networks. Deep learning is a subfield of machine learning, a branch of computer science based on fitting complex models to data.

Keep Reading ↓Show less