MassHighTech December 29, 2003

 

Taking on the Titans: Life after the capacitor: Nantero plans to be there
By Jeff Miller

Nantero’s technology may be tiny, but the market it hopes to upend is as large as they come: the $25 billion RAM industry.

“Having even just a percentage of all the markets we’re going after would allow us to become a big company,” said Greg Schmergel, co-founder and chief executive of Nantero.

Their innovation? Replace the current method of storing data with carbon nanotubes, microscopic pipes of carbon that are roughly 1/10,000th the diameter of a human hair.

Currently, there are three basic types of RAM (random access memory), all of which use electrostatic means to store data. There are no moving parts.

Dynamic RAM, which is used principally in personal computers, stores data via an electric charge. It’s cheap and fast, but it also requires a constant flow of power to retain the data. Once the power is cut off, all data is lost.

Static RAM, used mainly in servers and high-speed applications, is even faster than DRAM and uses less power, but it is also much more expensive. Like DRAM, once the power is turned off, the data disappears.

Flash memory, which is usually found in portable consumer electronics, doesn’t need a constant supply of power to retain data. But it’s expensive, it’s slow and it has a more limited life than DRAM or SRAM.

Using electrostatic forces, a nanotube can be bent so that the top touches the bottom, representing a one. Likewise, a bent nanotube can be straightened to represent a zero. Billions of nanotubes can fit into one four-inch wafer.

Because this process moves only a few carbon atoms in each tube, Schmergel claims Nantero can run its so-called NRAM (nano RAM) at 2 GHz. If the power is turned off, the nanotubes’ states don’t change, so no data is lost. That would allow personal computer users, for instance, to enjoy the same kind of “instant boot” enabled by Flash memory, but without the disadvantages of high cost and slow speeds.

What’s more, Schmergel claims his NRAM has a virtually unlimited lifetime, uses less power than conventional RAM and has an inherent resistance to radiation.

This year, Nantero claims to have built a working, 10 Gb array, and in such a way that the nanotubes can be integrated into any conventional complimentary metal oxide semiconductor (CMOS) fabrication facility. Nantero aims to license its NRAM intellectual property and process techniques to established RAM manufacturers.

Some may question whether nanotube technology is ready for integration into commercial products. After all, researchers still haven’t developed a method for making nanotubes that possess consistent properties.

“That’s a problem if you want to take advantage of the nanotubes’ electrochemical properties,” said Juan Sanchez, an analyst at Punk Ziegel and Co., a New York investment bank. “But you don’t need that level of consistency in manufacturing for the mechanical properties that Nantero is using.”

All that Nantero needs to ensure is that the nanotubes are free of dust and long enough to bend, Sanchez said. And he believes these standards are attainable.

Unlike most pre-revenue tech startups, Nantero has an in-house attorney who helps the company manage its portfolio of 40 patents and applications, six of which have already been granted.

But developing a prototype that works in the lab is entirely different from building a new device in a conventional fabrication facility.

“There’s uncertainty about how (nanotubes) can be incorporated into a CMOS process,” said Craig Hampel, technical director for Rambus, a California company that makes high-speed interfaces for DRAM. “That’s a big hurdle.”

Schmergel claims that Nantero has designed its NRAM so that it can be manufactured in a conventional fab. After all, though he hopes someday to bring an NRAM product to market made completely of nanotubes, in the current design, nanotubes overlay silicon.

Since the first part of 2003, Nantero has been testing the compatibility of its RAM design with ASML Holding NV, a large Dutch company that makes lithography equipment for the semiconductor manufacturing industry.

“One of the biggest challenges is optimizing the yield of the devices,” Schmergel said. “That translates into cost. Whenever you introduce a new material into a semiconductor fab, it requires some learning.”

It’s an audacious plan, but one that has garnered investment from technology luminaries such as Alex d’Arbeloff, founder of Teradyne Corp. and former chairman of the MIT Corp., and Ray Stata, founder of Analog Devices.

Recently, Charles River Ventures led Nantero’s $10.5 million B round, which also included Harris & Harris Group and Draper Fisher Jurvetson.

But nanotubes aren’t the only technology researchers are grooming for the next generation of RAM. Minnesota-based NVE Corp., for example, is developing magnetic RAM, which is made by sandwiching layers of magnetic metal compounds between two rows of wires. When electricity travels over wires that are crossed, the metal underneath becomes magnetized.

It’s fast and it’s nonvolatile, but like NRAM, it’s a new technology using new materials. But unlike Nantero, several large companies have already licensed the technology from NVE, and the company managed to go public on the Nasdaq SmallCap Market in January 2003.

A third company, California-based Ovonyx Inc., is developing a method for storing data using phase-change technology. The company uses electrical current to change points in a thin film from amorphous to crystalline, which each represent ones and zeros. Again, the technology is raw but nevertheless has financial backing from Intel Corp.

Which will win out? Maybe none of the above, said Rambus’ Hampel.

“It’s such a huge, momentum-driven industry, it’s hard to disrupt,” Hampel said. “Most problems with current technologies don’t become an issue until you get down to the 35 nanometers, and that won’t be in volume until 2010 by most roadmaps.

“On the other hand,” he added, “if you can find a technology that has a significant advantage, this is an industry that’s forecast to go to $38 billion in the next five years. There just might be something after the capacitor.”