By Burt Helm As chipmakers look beyond silicon for the Next Big Thing, single-wall carbon nanotubes seem like a good bet. Never heard of them? Well here's a brief introduction to what may just be the most multitalented natural structures in existence -- and some of the most cantankerous.
Measuring between one and five nanometers in diameter -- that's roughly 1/50,000 the width of a hair -- they're nearly 100 times stronger, one-sixth as heavy, and 20% more flexible than steel. They far exceed copper's heat-retaining capacity, accounting for almost no thermal leakage, and they can carry an electrical charge at twice the speed of circuits embedded in silicon. Together, these attributes suggest an enormous potential for smaller, faster, cooler-running chips.
HOIST TO THE HEAVENS. Indeed, some scientists believe nanotubes could lead to three-dimensional integrated circuits that would see transistors stacked not only side-by-side, but also atop and beneath each other -- a configuration researchers have never been able to achieve with silicon.
Beyond chips, the proposed applications are even more spectacular, such as superefficient power cables and all-but-impenetrable armor. There's even talk of building a "space elevator" -- in effect, a giant ladder that would dispense with rockets and allow payloads to be hoisted into low orbit. Currently, no material boasts the strength or flexibility to support a structure soaring to such heights. A raft of small startups is working to commercialize the technology, such as SouthWest NanoTechnologies in Norman, Okla., Carbon Nanotechnologies in Houston, and Liftport Group in Bremerton, Wash., which hopes to build a hoist to the heavens by 2018.
Trouble is, researchers haven't been able to figure out how to make nanotubes in consistent, commercially viable quantities. And while nanotubes are darlings of technology buffs, they're about to meet a tougher crowd. The San Jose-based Semiconductor Industry Assn. (SIA), a trade group whose members include the major U.S. chip outfits, is working with federal agencies to start a program aimed at evaluating the potential of various nanoelectronic technologies and fostering government and private research funding.
CONSISTENTLY INCONSISTENT. The SIA hopes to narrow the field to just two or three technologies, with the most encouraging results getting the most funding. "It would be a horse race to see [which technology] gets to that point," says Bob Doering of Texas Instruments (TXN), one of the three U.S. representatives on the International Technology Roadmap for Semiconductors (ITRS), another organization that helps set international industry standards.
Among semiconductor engineers, dependability, cost, and the potential for unlocking new design capabilities are paramount. And that's where nanotubes run into trouble. If they were human, we might say they have an attitude problem, insisting on doing things in their own, largely unpredictable way.
When being grown, each nanotube seems to decide its own size, length, and semiconductor and conductor properties. Scientists have little idea how to grow specific and consistent types and sizes. Further compounding the problem is that the end product is a sticky, tangled, microscopic mess. To the naked eye, a million nanotubes look just like a speck of black soot.
BABY STEPS. Much progress recently has been made in growing the nanotubes cleanly and predictably. Scientists at DuPont (DD), Rice University, and elsewhere have been able to separate nanotubes by bonding nonattracting compounds to their walls. "Imagine nanotubes are spaghetti and we've just invented butter," says Rick Smalley, a professor at Rice who won a Nobel Prize for discovering the buckyball, a cousin of the nanotube. Researchers at Southwest Nanotechnologies have succeeded in growing tubes to extremely consistent diameters, and last month collaborators at Duke University and Los Alamos National Laboratory grew the longest tube to date, four centimeters.
But it's precisely because scientists have had to grapple with these simple structural details that many observers regard nanotubes as poor candidates for mainstream electronic products. Tom Jackson, a physicist at Pennsylvania State University, sees nanotubes adding little to the future of computing, despite their many impressive attributes.
Jackson concedes that CMOS, the type of silicon-based transistor that now dominates the chip business, "looked like a long shot" when first developed. But he also points out that CMOS "provided a higher level of architectural complexity" than anything that had gone before, enabling new capabilities. The transistor beat the vacuum tube in the 50s for the same reason: It enabled new and elegant possibilities in manufacturing and design.
NANOWIRE RIVALS. Right now, the nanotube is less an innovator than a show-off. On a good day, it outperforms present materials. But barring a major breakthrough, the commercial fabrication methods that seem most likely to be used for making nanotubes -- some combination of purifying, untangling, sorting, and straightening -- remain far more complicated than growing a large, relatively simple crystal of silicon. Recreating current chip designs with nanotubes would be a daunting struggle, the experts say.
The nanotube also has some fierce competitors. Silicon nanowires, which are much easier to work with, represent a major, but incremental, step up from existing silicon technologies. Like tubes, nanowires can form supersmall transistors and wires. And while they lack tubes' spectacular strength, but they can still be counted on to produce complicated configurations.
Charles Lieber, a Harvard physicist, gave up on nanotubes to work with nanowires, citing the tubes' unreliability. "The advantage of nanowires is that we can make the systems much more complex, and we can think of new architectures and combining things that we can't do [with the tubes]," he says.
HIGH-TECH PAINT JOBS. Paulo Gargini, another U.S. representative to the ITRS and a fellow at Intel (INTC), where scientists have evaluated both possibilities, says nanowires seem to be a stronger choice for manufacturing. "If you look at the relative benefits...silicon nanowires are a natural extension of [our present] technology.... You can still control structure and still get performance." Gargini makes sure to emphasize, however, that it's still far too early to make a final decision on the future of nanoelectronics.
Despite the disadvantages and daunting technical challenges, nanotube research will likely continue, albeit with the liklihood being that greater research efforts and funding will go to other high-performance materials, such as nanowires or more typical semiconductor compounds like gallium arsenide. "Nanotubes are still a dark horse" in the race to become a big deal in semiconductor materials, says Michael Fuhrer, a physicist who builds nanotube transistors at the University of Maryland.
It's entirely possible that nanotubes will be used mainly in relatively low-tech applications. Mixing a paste of them with paint can make it electrostatic. That could allow cars on the production line to be plated with pigment, much as watches are covered with a thin layer of gold, thus trimming manufacturing costs. Nanotube pastes also could enhance liquid-crystal and flexible displays (a technology still in the research phase), providing sharper images than the low-grade silicon and carbon-based films that are currently used.
But beyond the more basic applications, nanotubes' future appears bleaker. Science-fiction enthusiasts shouldn't expect to take that elevator ride into space anytime soon. Helm is a reporter for BusinessWeek Online in New York