CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Making mountains out of molecules

At IBM Research, Phaedon Avouris spends his days delving into the basic behavior of matter and takes a no-nonsense approach to nanotechnology.

An average day for Phaedon Avouris involves delving into questions about the basic behavior of matter.

As the manager of nanoscale science and technology at IBM Research, Avouris is in charge of Big Blue's research into designing chips, storage devices and other systems out of strands of molecules. Among other breakthroughs, the group has concocted carbon nanotube transistors as well as techniques for creating nanotubes. Avouris, who has published more than 300 papers, also recently received the Irving Langmuir Prize from the American Physical Society.

Avouris recently spoke with News.com's Michael Kanellos about the nature of nanotechnology, the problems ahead in the field and, of course, the hype.

Q: With the growth of this field, are you finding a lot of bad science coming into it?
A: (Laughs). There is a lot of hype. It is very hard to tell what is going on. It is the time of the nice pictures, claims that nobody is going to referee and disprove. All of this is going on because everyone believes that there is a lot of money to be made. I don't like this mixture of investment and science. There are a lot of people with money without a product. Scientists becoming entrepreneurs can lead to conflicts.

There is an Israeli firm that sent me a map of nanotechnology. It is a people map of who does what in nanotechnology. So who are the people doing the best work, according to this map? In nanoelectronics, it was Newt Gingrich (a former Speaker of the House of Representatives). In molecular electronics, it is Leo Esaki. You probably don't know Leo. He was an early Nobel Prize winner for the Esaki diode, but he has not worked on molecular electronics. He used to be at IBM. God knows what they did to come up with these names. They probably looked at the Web and saw who was listed in articles. Newt Gingrich is the authority on nanoelectronics!

If you mix profit with science, it can lead to conflict. At IBM, I have no conflict. I am not selling anything. I only have to please my management. I have to be careful. If I exaggerate, it is going to hit me hard. I am going to pay for it, while at a university in order to get the grants and so on, you have to promise them the world, and I don't like that.

That's interesting. Most people would assume that academics have more freedom to speak their mind.
But this is the problem. This is the distortion that nanotechnology has brought. According to a nanotechnology business letter, there are 670 or so nano companies in the U.S., and almost all of them originated from an academic. The usual trend is that a student in a lab makes an interesting observation, and before you know it they are forming a company to exploit it, but technologically they don't realize what is involved.

In science, you can have a breakthrough overnight. It is a conceptual breakthrough. Technology requires a lot of hard work and ironing the details to the point where you have something that is reproducible and durable and all of this stuff. And that is where they fail.

I don't like this mixture of investment and science. There are a lot of people with money without a product.
When will nanotechnology progress to something practical?
I think we need at least three more years of science to see if the materials look promising, and by materials I mean the devices, the transistors, the memory and so on. It has to be by far better than silicon to make any kind of development worth it. Then there are all of the (manufacturing) issues. We are talking at least 10 or 15 years, and I don't expect any other technology to take less than that.

Look at simple things like switching to copper wiring (in semiconductors). You look at it in principle and say "what is the problem?" But it took a lot of work to figure out how to work out the contacts and the interfaces. Imagine starting with a completely new material.

I was talking with a European guy who said, "We are looking toward a new, multivalue logic that would be inside the molecule." And I was thinking "Oh my God." Imagine throwing everything we know--Boolean algebra, device application, device materials, software--into a product. We cannot replace Windows because Bill Gates had a head start, and he's saying "Revolutionize everything.'"

Eventually, where will nanotubes be used?
We may not be able to ever displace silicon as the global technology, but there are niches everywhere. You look at NASA. They are spending a lot of technology on nanotube research. If you think about it, for them, size, weight and radiation hardness are very critical components. For them, price is not an issue. If they are going to put a vehicle on Mars, they don't mind if they spend $100 million for that.

How is the U.S. doing in funding nano research?
I think we have probably the best funding, but everywhere it is picking up. The Japanese are putting up a lot of money. The Europeans are. Everybody is trying to get involved because of the hype.

What makes nanotechnology different from the materials used in the past?
The interesting and novel part of the nanometer scale is that materials develop different properties. A micrometer piece of a material has the same properties as bulk material. By going to the nanometer scale, the quantum nature of the material takes over. It is not a quantitative transformation. It is a qualitative one. A metal can become an insulator at very small sizes. Correspondingly, an insulator can become metallic. (Editors' note: A nanometer measures one-billionth of meter; a micrometer, also known as a micron, is one-millionth of a meter.)

Invite Michael Kanellos into your in-box
Get senior department editor Michael Kanellos delivered right to your in-box with a full column and commentary on chips, servers and all the hardware critical to your business. Enterprise Hardware, every Wednesday.




In semiconductors, I've read that air can change the behavior of nanotubes.
That is one of the most important aspects. It is going to be true for just about every small device that we have. We are used to planar structures where silicon bonds to a metal surface and there are zillions of bonds being formed. But when you bond a molecule or a nanotube to a metal, you just have a few dipoles (pairs of equal and opposite electric charges separated by a small distance). But you can have around it all kinds of other things: oxygen, hydrocarbons, dirt. They form dipoles, and these dipole fields can dominate over any of the intrinsic dipoles made by the bonding.

So in effect, you can't really say that the electrical properties of this molecule are this. It depends on what else is around it. That is something that people have to take into account.

Being able to control the transistor through the environment seems kind of handy, though.
It is both good and bad in the sense that you can change the properties of the device, but also you have to control them so they don't change with the environment again. The good news is that we were worried about what would happen if we, say, covered the devices with silicon dioxide. Would the nanotube be affected? It isn't, so it is nice you can encapsulate them and get rid of these problems.

What are some of the other major preliminary challenges in nanotechnology that we have to deal with in the next three to five years?
The main issue is materials. We need to have a way of making reproducibly one type of nanotube. It is not a big problem for us because we do scientific studies. We always use a single nanotube to do our devices. But for any kind of technological development, you have to have a supply of one type of nanotube, and unfortunately people have not done that.

All published work on the synthesis of single nanotubes involves some kind of metal catalyst, usually a transition metal. If you think about it, it is terrible. At the end, you end up with almost every nanotube ending up with a metal particle. In fact, when we get the nanotubes from Rice (University), if you take a magnet you see the tubes move back and forth. For certain applications you don't care, but for electronics you certainly do. You don't want to have nickel and cobalt and iron in your devices.

The suggested process is to boil this stuff in nitric acid. That dissolves the metals. But in the process, of course, you damage the tubes.

Last year, IBM came out with a paper on making carbon nanotubes by growing them in crystals. I thought this eliminated the problem of catalysts. How does that work?
If you look at the structure of the silicon carbide crystal that we used, it is sort of layered. If you heat it, the silicon goes away and leaves the carbon behind. But the carbon has dangling bonds where it was bonded to

I am convinced that there is not a single example in science of a discovery that was not accidental.
silicon. So adjacent carbons tend to start bonding. Imagine it roughly like a moussaka of carbon and silicon, and suddenly you remove the silicon. At the temperature that we do this, which is 1,500 degrees, the things curl up and satisfy these bonds.

I must tell you this is not proven. This is inferred from our data. We have not tried to make quantities of this thing. Silicon carbide is not a cheap material, so I wouldn't propose it as a way of making large quantities of nanotubes, but more as a way to tell the materials people, "Hey, don't set your mind only on catalysts. There are other ways. Use your imagination."...I am convinced that there is not a single example in science of a discovery that was not accidental, no matter what we say afterward.

The synthetics people seem to be more interested in making yet another exotic form of carbon because that makes you a little letter in nature. But if you go and say, "I improved the yield of one type of nanotube" there is no glory.

You mentioned once before that boron nitride is also being touted as a nanotube material.
Yes. That is another nanotube material that is interesting, but there hasn't been that much of it around. We were the first ones to look at the electrical properties. We got our samples from France, and there is only one group in the states makes boron nitride (tubes) and they make it in a different form. It is a double-walled (tube). The ones we got from France are single-walled.

You hear a little about nanowires. What is going on there?
Nanowires are a different kind of beast. They are solid, so in effect they have the same properties as bulk silicon. But it is an interesting way, a simple way, of making these nanostructures without using expensive lithography.

Nanowires are made with a gold catalyst, and gold and silicon are supposed to be a no-no because gold forms deep traps in silicon and kills the transport. So the question is, how does that work? I don't have the answers, but in a couple of weeks I am going to Harvard and will interrogate the people doing it.