Small Scale, Big Business
In the wake of the dot-bomb crash, tech investors are turning to nanotechnology as the next big thing, with huge advances in materials science, manufacturing, and medical treatments on the horizon. But while the headlines predict a sci-fi future of microscopic robots clearing clogged arteries or wind-blown “smart dust” surveying enemy battlefields, the semiconductor industry is building nanotech devices right now, and vendors like Intel and IBM are positioned to lead this fledgling industry.
Indeed, the chip industry has been racing toward nanotechnology for years; any manufacturer will tell you that micro-electrical machines (MEMs) will inevitably be replaced — or at least enhanced — by nanoelectrical machines (NEMs). There’s no hard and fast rule for where micro ends and nano begins, but the consensus seems to peg it at around 100 nanometers. (A nanometer is one-billionth of a meter; a micrometer, better known as a micron, is one-millionth of a meter.) It’s at this point that conventional physics give way to the unique properties of nanoparticles.
“Sub-micron MEMs are a reality and have been a reality for three or four years. There has just been no market for these devices,” says Om Nalamasu, CTO of the New Jersey Nanotechnology Consortium. Based on Lucent’s corporate campus in Murray Hill, N.J., the Consortium has a long history working with integrated circuits and currently works mostly on nanoscale production.
Speaking at this week’s NanoBusiness 2003 conference in New York City, Nalamasu conceded that the wind was sucked out of the nano-hype sails in 2001, when the optical networking market crashed: There was so much oversupply that no one was investing in new equipment or technologies. Now, however, he sees a convergence of electronics, photonics, and biology as driving a new interest in nanotech.
That means the CPU industry no longer has the technology to itself: “Fifteen years ago, it was just nanolithography for making processors, because [vendors] knew they were going to hit the wall when it came to performance,” Dr. Richard Colton, head of the Surface Chemistry Branch of the Naval Research Laboratory, told the crowd at NanoBusiness 2003. “Now, if you read the journals, nanotechnology is everywhere from materials science to engineering, manufacturing, bioscience, and so on.”
Follow the Money
President Bush’s proposed 2004 budget provides $847 million — nearly a 10-percent increase over this year — for the multiagency National Nanotechnology Initiative to advance the understanding of nanoscale phenomena. Congress has gone even further down the nanopath, authorizing $2.135 billion in federal research money over the next three years.
Nanotechnology revenues are already at about $45.5 billion globally and are expected to reach $1 trillion in 10 years, according to estimates from the National Science Foundation. Numbers like those attract funds from the venture capital market despite today’s economic slump; venture investment in nanotechnology this year will be over $1 billion, up from under $100 million in 1998.
“Given these times of fiscal austerity, it is amazing that nanotechnology can maintain this support,” says Phil Bond, chief advisor on science and technology for the U.S. Secretary of Commerce. Part of this support comes because nanotech plays in so many sectors, including the medical and chemical as well as manufacturing markets. Still, it is the chip industry that has invested the most time and money so far.
Government support pales in comparison to the chip vendors’ investment in nanoscale production facilities. IBM’s chip plant in East Fishkill, N.Y., for instance, will cost $2.5 billion before it comes online in 2004. The plant will enable Big Blue to produce chips with little connections — 100 nanometers.
Exotic Memories and Ever-Faster Processors
Such production facilities will be invaluable in the mass production of products like magnetic RAM (MRAM), a nonvolatile memory that stores bits by altering the spin of electrons. MRAM promises write speeds far faster than today’s SRAM and DRAM — according to IBM, as fast as 2.3 nanoseconds, with power consumption in the 2-milliamps range. If it could be produced cheaply enough, MRAM could be used in everything from PCs to cell phones.
IBM has also been able to use nanotechnology to build the world’s smallest working transistor. The transistor gate is just 6 nanometers (roughly 25 atoms) wide. By contrast, today’s 0.13-micron PC processors have transistor gates measuring something under 70 nanometers, which Intel says its 90-nanometer-process “Prescott” CPU will shrink to under 50 nanometers.
Researchers at Woburn, Mass.-based Nantero Inc. say they’ve developed a way to use carbon nanotubes to store data. By using buckminsterfullerenes, a.k.a. “buckyballs,” to store bits, the tubes could be switched on and off in less than a nanosecond and retain their status even when no electrical charge is applied, making what the company calls NRAM a solid-state technology that could be used like a hard drive.
Threading the Needle
Hewlett-Packard is working on stringing nanoparticles together to create conducting wires. The company has teamed up with the University of California to build grids of nanowires and place molecules at their intersections; by using these molecules to block or conduct electrical current, the company has shown the viability of a nanomicroprocessor.
Although not particularly expensive to make, these nanowire grids tend to have imperfections. This requires a lot of reengineering after the chip has been made. HP is currently investigating how to use fault-tolerant software to bypass these imperfections and take advantage of the chips’ processing power.
Although nanowires are small — between 3 and 5 nanometers thick — they have the same properties as bulk silicon. To take advantage of quantum properties, the size needs to be below 3nm. Scientists at the City University of Hong Kong have fabricated silicon wires with diameters approaching 1 nanometer, and claim to see quantum effects at work in their prototype tests.
But before the topic turns back toward science fiction, it’s important to remember that the first uses of nanotechnology will be to help the chip industry maintain its performance curves and lower costs. Intel is already working at the nanoscale level with Prescott, and has announced plans to go further with its TeraHertz Transistor project for the second half of this decade. AMD has developed a transistor with a gate length of just 15 nanometers, which could be the first step toward 30-nanometer process CPU manufacture (perhaps by 2009 or 2010).
The real question is whether other industries can duplicate the chip industry’s success with nanotechnology. As Phil Bond put it at this week’s conference, “With nanotech, the entire manufacturing process starts over again.”