Stanford Scientists Create First Carbon Nanotube Computer

(Norbert von der Groeben/ Stanford University ) - Max Shulaker, a grad student at Stanford University’s School of Engineering, talks about the new material used to make these computers.

By Meeri Kim, Published: September 25, 2013

Someday we may look back on today’s iPhones and laptops as huge, clunky devices with outdated chips made of silicon that was long ago replaced by carbon nanotubes. Tens of thousands of the tiny tubelike structures can fit inside a human hair, and now scientists have created the first carbon nanotube computer — a big step toward miniaturizing our electronics even further.

lthough it pales in comparison to today’s computers, the bare-bones machine works. It runs a basic operating system and can freely switch between two programs — one that counts in a loop and another that sorts numbers.

Graphic

The carbon computer

More health and science news

Obamacare: Understanding the Affordable Care Act

Washington Post staff SEP 27

Everything you need to know about navigating Obamacare.

“This is not a computer you would buy off the shelf at Best Buy,” said lead author and Stanford electrical engineering graduate student Max Shulaker. “But the functionality is still a complete computer.” The study was published online Wednesday in the journal Nature.

Shulaker gave the computer the pet name Cedric, a rough acronym for “carbon nanotube digital integrated circuit.”

The achievement by Stanford engineers marks the most complex electronic device ever built from carbon nanotubes, a man-made tube-like structure created from a rolled-up sheet of carbon one atom thick. The tubes’ remarkable electrical and mechanical properties have led researchers to explore their potential applications in bulletproof clothing, cancer therapy and electronics.

Currently, silicon is the standard material for manufacturing the chips used in computers and phones. But as our devices keep shrinking, silicon circuits are reaching their limit.

“Silicon is great. It’s very hard to beat,” senior author and Stanford electrical engineer Subhasish Mitra said. “But when everything gets so small, it’s not clear you can get high performance and energy efficiency from silicon transistors.”

Like nerve cells to the “brain” of a computer, transistors are the key active component in practically all modern electronics, and a single chip can contain billions of transistors. Using the minuscule carbon nanotube instead of silicon could mean smaller, faster and more efficient transistors.

Carbon nanotubes can be a single nanometer wide in diameter. In comparison, a strand of human DNA is 2.5 nanometers in diameter and your fingernail grows about a nanometer every second.

Ever since the first carbon nanotube transistor was built in 1998, “there was a dream in people’s minds that we would have a new era of digital electronics using these carbon nanotubes,” Mitra said.

But researchers ran into a brick wall. They found it extremely difficult to manufacture nanotubes without glaring imperfections that would render any transistor made from them dead.

A common way to make carbon nanotubes is to grow a forest of them on a semiconducting wafer baked in a high-temperature chamber.

“Fancy chemical reactions will happen and, as a result, carbon nanotubes will sprout,” Mitra said.

An ideal batch would sprout with the carbon nanotubes parallel to one another and without imperfections. But Mother Nature often intervenes, he said, and many a time they ended up with a bowl of nanotube spaghetti. They improved their technique, but still wound up with a few defective structures.

Mitra and Wong had their breakthrough when they stopped trying to make flawless batches of nanotubes and instead focused on designing computer circuits immune to imperfections. Using clever algorithms, they arranged their circuit in such a way that, even if a few nanotubes were misaligned, their computer would still function.

In total, the computer contains 178 transistors, each composed of 10 to 200 nanotubes. The device’s footprint is only 6.5 square millimeters, meaning you could fit about 40 on the head of a dime. But like modern-day computers, the memory is “off-chip” and not made up of nanotubes, Shulaker said.

He said that the computer could be faster, more powerful and smaller if it were built in an industrial facility rather than in an academic lab. Shulaker cited this as one reason they built a more primitive, no-frills computer.