From cars to medicine, nanotubes may be miracle material
By Michael Kanellos
Staff Writer, CNET News.com
October 20, 2003, 4:00 AM PT
They are stronger than steel and as flexible as plastic, conduct energy better than almost any material ever discovered and can be made from unexotic raw materials such as methane gas.
Now the question is whether they can live up to their promise.
In a relatively short time, carbon nanotubes--thin tubes of carbon atoms that have unusual characteristics because of their unique structure--have emerged as a miracle material that could revolutionize a number of industries.
Single-walled nanotubes are expected to debut this year in polymers as a way to strengthen plastic parts in cars or get them to conduct electricity through normally nonconducive materials. Paint that can deflect radar is also anticipated in the not-too-distant future. Computer and TV manufacturers plan to use them to drastically reduce the cost of screens in an estimated two years.
"Any major industrial corporation that has an interest in advanced materials, from plastics companies to semiconductor companies, is buying from us," said Tom Pitstick, vice president of business development at Carbon Nanotechnologies Inc. (CNI), a Houston company founded by Rick Smalley, a 1996 Nobel Prize winner and Rice University professor.
Within a decade, nanotubes could replace silicon as the transistors inside processors and memory chips. Tubes could also be used to convey light through optical fibers and, further out, to deliver medicines to specific cells inside a body or even restructure the nation's power grid.
Mass production of nanotubes, however, remains a challenge. CNI plans to increase its manufacturing capacity to the point where the company can make 1,000 pounds of nanotubes a day by 2005. Right now, it can make only about a pound or two daily.
This cumbersome process makes the technology too costly for wide use. The going price on the company's Web site is $500 a gram.
Other researchers also say that silicon nanowires--solid microscopic strands of silicon--could prove to be easier for semiconductor makers to graft onto existing manufacturing processes.
"Silicon nanowires might be less perfect, but they may be easier to integrate into chips," said K.J. Cho, a professor of mechanical engineering at Stanford University.
Reducing the dimensions
A carbon nanotube is essentially a sheet of carbon atoms--arranged in hexagons--that curls up into a tube. It comes in two basic varieties: a single-walled nanotube, which is a single coil of carbon hexagons; and a multiwalled version, wherein a single tube is encased in a wider tube, which itself is inside other tubes. Most of today's research is concentrated on single-walled tubes.
The tubes' properties are significant because of two factors: their size, which allows them to function as one-dimensional objects, and the intrinsic nature of carbon.
From a purely Euclidean perspective, physical objects on this planet, including nanotubes, all exist in three-dimensional space, which can be measured through X (horizontal), Y (vertical) and Z (depth) coordinates.
Scientists, however, assert that dimensions can become irrelevant. A film negative, for instance, functions more like a two-dimensional object. Negatives technically have height, measured by the Z coordinate, but it can't readily be used.
Because one-dimensional nanotubes have no height or width, they are the atomic equivalent of a bowling-ball return. As a result, electrons can travel ballistically on them--that is, barring obstacles or flaws in the material, electrons don't get scattered or lost.
"If you have a ballistic conductor, your charge can go completely unimpeded," said Joerg Appenzeller, a carbon nanotube researcher at IBM Research. "The electronic properties are outstanding."
Such confined dimensionality means that nanotubes can conduct heat better than any other material ever discovered, including diamonds, and could even be used to transfer energy in homes or between power stations. Tubes can also be used to carry light, enhancing or replacing optical fiber.
In chips, nanotubes could lead to transistors that switch off and on much faster than today's silicon variety.
Appenzeller said it is impractical to compare their performance to silicon transistors because researchers have only tested how single nanotubes work. Still, the early results are very promising, he noted, and the same basic transistor structure can be used.
"You just replace, ideally, hypothetically, the access device with a nanotube. The source, the drain, the architecture is the same," Appenzeller said. One-dimensional objects can be formed from other materials, such as boron nitride, but carbon has been studied the most so far.
While nearly everyone agrees that carbon won't likely appear in chips or fiber for several years, other products in the near term will likely take advantage of nanotubes' electrical properties.
Several companies are looking at ways to use nanotubes in TVs, liquid crystal display monitors and plasma screens for 2005. In traditional TV sets, electron guns shoot electrons at the screen, which must be 18 inches away. LCDs and plasma screens don't require electron guns, but the manufacturing process required to implant the glass with circuitry costs billions.
Nanotube monitors would be thinner than LCDs and far cheaper to make. The tubes can be mixed into a paste and printed onto glass. Hyperspecialized facilities wouldn't be needed.
"It is amazingly simple," Pitstick said. "You put nanotubes in ink and print them down."
The bonds that bind
Bonding is another key property that makes nanotubes attractive. Carbon atoms bond tightly to each other and gravitate toward the stable, hexagonal rings. Nanotubes "heal" themselves by shifting to replace atoms that get removed.
"Silicon is very finicky about defects," said David Tomanek, a professor of physics at Michigan State University. "We have concluded that carbon nanotubes are relatively defect-tolerant."
That has the potential to relieve huge headaches. Chipmaking facilities cost $3 billion today and will likely cost $6 billion by 2007. The lion's share of those funds goes toward equipment that's needed to draw circuits.
Self-assembling, correcting tubes eliminate the need for many of these machines. Most of the equipment required "is all pretty standard chemical industry stuff," Pitstick said.
Other applications benefit from bonding as well. Single-walled nanotubes, which are incredibly resilient to physical twisting or pulling, can be kinked to a 120-degree angle and bounce back to original form undamaged, said Hongjie Dai, an associate professor of chemistry at Stanford.
They can be long, too. Researchers have created defect-free nanotubes as long as four microns, which is 40 times the length of the average size of features on regular silicon chips. Some nanotubes with less-than-perfect ballistic features have been made as long as 120 microns.
Hypothetically, this could allow engineers to replace wires in airplanes with tubes, strengthening parts while reducing weight.
Carbon is also good for exploiting van der Waals forces, which cause different types of atoms to bond spontaneously. In experiments, researchers have noted that nanotubes will adhere to silicon posts that stick up from a wafer. As a result, they can be arranged in a useful array. Nantero, a start-up that had its beginnings at Harvard University, is aiming to exploit van der Waals forces to make a new type of memory chip.
"You get a wafer of tubes with a reasonable orientation," Dai said. "They really like to land on the post to enjoy the van der Waals contact."
To the drawing board
While the benefits seem infinite, researchers are quick to point out that such results have been limited so far because mass manufacturing has yet to take place.
Today, carbon nanotubes are made in two ways. The first, known as the laser ablation method, was pioneered by CNI and involves blasting graphite with a laser. The second, the modified gas method, involves spraying a hydrocarbon gas like methane or CO2 over a molten metal catalyst.
Removing impurities, such as metallic catalyst particles, is a challenge in both. IBM and others are experimenting with new fabrication techniques, such as building silicon-carbon crystals and then evaporating the silicon, but no one has an answer yet.
Another major problem lies in controlling something called "chirality," a measure of the arrangement of the hexagons on the surface of a tube. If the carbon hexagons run in parallel vertical lines on the surface of the tube, they will act like a metal and can't be used in electronics. If the rows of tubes are slightly swirled (think of the cardboard on a paper towel roll), they will act like semiconductors and can be used as transistors.
Unfortunately, the factors behind tube formation remain something of a mystery.
"You leave the world of classic physical mechanics, and you enter the world of quantum mechanics," Appenzeller said. "The graphene sheet is the same. That's why it is so difficult to predict the chirality." Graphene sheets are made up of carbon hexagons.
Smalley, Dai and others are hoping to control both of these characteristics through "selective catalysts."
"If you can control the seed (catalyst) well, you should be able to control the nanotube," Dai said. "Over the years, we have found that the catalyst controls everything."
The next challenge is arranging the nanotubes in products. Placing tubes in exact locations in products such as chemical sensors or flat panels isn't a problem, because they are painted in. Chips, however, will require that individual nanotubes be placed between specific contacts.
Scientists hope to grow the nanotubes on a wafer. Researchers at Duke University and Stanford have shown that it is technically possible to grow and position tubes, but many hurdles still need to be cleared.
Silicon strikes back
In the end, silicon compatibility could tip the balance toward silicon nanowires.
Silicon nanowires are made by siphoning molecules of SiH4 (a single silicon atom surrounded by four hydrogen atoms) through a gold particle, said Andre DeHon, a professor at the California Institute of Technology. The gold strips off the hydrogen atoms and allows the naked silicon atoms to form into a wire.
"Our goal is to build interesting-size memories out of these things," he said at the Hot Chips industry conference in August. "This is something that could come through in single-digit years--three to five years, if someone really wanted to push it."
As futuristic as it sounds, the technique was first described by researchers at Bell Labs in 1964.
Although nanowires may not exhibit the same electrical properties as nanotubes, silicon nanowires may be easier to grow on the wafer itself, DeHon added. Nonetheless, the process cannot be done overnight.
"It will be a number of years before we see a change," said Pat Gelsinger, chief technology officer at Intel, which is working with university researchers on both approaches. "It is preliminary to say it is either one."
Despite the challenges, researchers and companies are optimistic about nanotubes, buoyed by positive experimental results that are occurring at a fairly rapid pace.
"We have made enormous progress," Appenzeller said. "Everything is working out so far fine."