An interview with Phillip Wong, senior manager of Nanoscale Material Process and Devices, IBM Research Laboratory, Yorktown Heights, N.Y.
The recorded interview can be heard at http://AmpCast.com/RColinJohnson
EET: What do think the future holds for nanotechnology over the next five years?
IBM: The next five years are going to be very exciting for nanotechnology, not that we will be manufacturing nanotechnologies in a big way. The reason they will be very exciting is that we will see a lot of very fundamental developments coming up in the next five years that will turn an interesting nanoscale science into a real technology. You can see a hint of it now in the area of carbon nanotubes, where many research labs, ourselves included, are going from the basic scientific research into engineering and benchmarking with conventional technologies, and trying to find the right niche in the application spectrum.
EET: Do you think that nanotechnology can be smoothly integrated with existing technology, or do you subscribe to the disruptive thesis?
IBM: I think that certainly it will be smoothly integrated with existing technologies. For instance, today in microelectronics silicon is king. If you look at how CMOS devices are built today, we are already, without making any big announcement, moving into the nanoscale regime. For example, everybody is announcing new 90-nanometer technologies-in those devices the gate length is about 50 nm, so we are already doing nanotechnology without even realizing; we are heralding in nanotechnology, even in conventional regions.
EET: Carbon nanotubes seem to be one of the most promising ways of integrating the new nanotechnologies, at least with silicon.
IBM: Yes, that's right. And there are a couple of reasons for the excitement about carbon nanotube technologies. The first is that they do not require any change in the circuit, or system design and -architecture. Because [nanotube transistors] continue to look like, feel like and smell like an FET [field-effect transistor], and that is a desirable attribute that you cannot underestimate, because there is a huge amount of industry infrastructure in the design community as opposed to a technology that would require a new circuit architecture, carbon nanotubes and similar technologies that look, feel and smell like an FET have a tremendous advantage. You can basically leverage all the infrastructure of the incumbent technology--that is a huge advantage.
EET: What you seem to be saying is that we don't have to go to this complete bottom-up approach that the founders of nanotechnology spoke of, with nanoscale assemblers working from molecular building blocks, but you seem to think that studying those kinds of things is already helping to solve some of your sub-100-nm problems with silicon?
IBM: Absolutely, you don't have to conceive of building a whole Pentium chip by shaking a beaker. You can use 99 percent of your existing technology and just add a few of these self-assembling, and recognizing techniques to help you do certain processes better, offering a huge leverage.
EET: Do you think carbon-nanotube technology is the most promising approach to nanotech?
IBM: There are many different kinds of devices and processes across a broad spectrum of technologies going from building logic chips or memory chips, to magnetic storage, plus there is a broad spectrum of applications. For instance, we have been doing some very interesting work on self-assembled magnetic nanoparticles, which have applications outside electronics.
EET: Yes, they have been proposed for use as medical tracers.
IBM: That's right, but if you are just talking about logic and processor chips, then right now, in this moment of time, the carbon nanotubes seem to have a slight edge in the degree of understanding and maturity that they have, but that doesn't mean it's the only thing. That's why at IBM we are hedging our bets with many other potential candidates, because a temporary lead does not mean you will be the final winner.
EET: Do you expect any surprises in the next five years?
IBM: I think the memory area is the most promising. Something [unexpected] could happen in the memory area; I don't know what, but something interesting. In the logic area, I think the major advances in the next five years will be a better understanding of what each candidate has to offer, and a pruning out of less promising candidates, moving forward in the next five years. And in the conventional chip arena, the [silicon] technologies continue to surprise everybody at how far they can go.
EET: Yes, there are new laser patterning techniques that seem to go well beyond the limits of even e-beam [lithography].
IBM: That's right. Traditional devices is one of the games that IBM plays very well, and we continue to play that because we think there will be very nice surprises coming up.
EET: You know the ultimate vision of nanotechnology, with the auto-assemblers and molecular building blocks: Its almost like MEMS but at a smaller level, where you take the idea of a factory and you scale it down to nanoscale parts. Do you think we should just forget about that whole thing? Is it really relevant?
IBM: I think we should keep it [the ultimate vision] in our minds, so we can have some fun in the science-fiction area, but I think by and large that vision is a little bit too grand. Certain parts of it will happen: I think self-assembled processes will be used and in fact we are seeing it a little bit already without people knowing it. Let me give you an example. If you look at an FET transistor today, almost the entire gate area uses only one lithographic step. All the rest are using some kind of self-aligned or self-limiting processes to make the rest of the device. That's how we make them so small. If we had to depend on lithographic tolerances to make it so small it would be hopeless. So self-recognizing features have been used for several years and we will continue to use those kinds of things. I don't expect a big shift in going from micro- to technologies-technologies. There will be a smooth transition from today's technology to tomorrow's. Initially, tomorrow's technology will be a hybrid addition of nanoscale fabrication techniques with conventional devices together, then a gradual shifting to more and more nanotechnology.