Nanotechnology: Replicating the Snowflake in the Fab

By Jeff Chappell -- Electronic News, 2/23/2004

SANTA CLARA, Calif.--Containing the cost of fabrication may be the key to realizing the future of nanotechnology and successfully implementing it in the world of information technology business.

That was the conclusion today of long-time IBM Corp. research veteran Thomas Theis, director of physical sciences for IBM Research. Theis delivered a keynote address here at the SPIE Microlithography Conference.

Theis noted that the semiconductor industry has been working on nanotechnology for some time, given the standard definition of nanotechnology: atomic, molecular and macromolecular R&D involving dimensions between 1nm and 100nm. Nanotechnology as it is discussed today is a natural culmination in many ways of what the chip industry has been doing for decades, Theis suggested.

So why all the fuss?

Theis theorized that the perceived limits of silicon microelectronics is on the horizon here in the age of deep submicron lithography, combined with the rise of molecular biology, supramolecular chemistry, combined with recent materials discoveries, has led to the buzz. It has also led, consequently, to large increases in government funding.

But it is semiconductor technology that is actually at the vanguard of nanotechnology, Theis said; the transistor itself is evolving into a nano device. He presented some R&D findings from IBM that showed an experimental 6nm gate, and discussed work in Japan that involves a gate of less than 1nm -- only three atoms of silicon in thickness -- that achieved reasonable mobility.

The results of this type of work suggest that viable transistors could eventually be built that small, as processes are refined. "Silicon will go to the molecular level," Theis said. "But is there something beyond silicon?"

There is no definitive answer to that question, and even if silicon does reach its physical limits in terms of semiconductor fabrication, it is still quite possible that silicon will become like iron and steel, Theis suggested. At some point it may not be at the cutting edge, but it will remain prevalent and pervasive.

In the meantime, there is promising work being done in the field of nanotechnology fabrication that could lead to a silicon successor, such as carbon nanotubes and nanomechanical data storage, a potential replacement for today's hard disk storage technology that utilizes MEMS technology to achieve much greater speeds.

Theis also showed results of IBM research on molecular cascade logic, which involved a three-input sorter, essentially a series of and/or gates, fabricated of individual carbon monoxide molecules.

Making Self Assembly Cost Effective

The key to implementing these technologies in real-world information technology applications will be the same as it always has been -- containing manufacturing costs and making the process cost effective, Theis said.

Many technologists tend to think of nanotechnology in old vs. new manufacturing concepts. Current chip manufacturing utilizes a top-down model using lithographic techniques that utilize digital information and depend upon low data error rates. Some suggest that a new manufacturing paradigm for nanotechnology will involve a bottom-up approach using chemical synthesis processes that utilize analog information and that can tolerate high error rates.

"This is an absolutely false dichotomy," Theis said. No manufacturing technique is purely digital or analog, he explained, adding that even today's processes contain aspects of both.

He offered the rather mundane example of a snowflake -- water vapor condensing on a dust mote in a cloud under the right conditions -- as means of naturally occurring self-assembly, a term often associated with nanotechnology and the new model of manufacturing. Nanotechnologists are not doing anything more mysterious in their R&D labs than trying to find a technologically useful implementation of the snowflake, Theis suggested.

"What we really need to do is figure out how to implement these in the highly hierarchical field of information technology," Theis concluded.

But it may not be as big a stretch for today's tech industry as it sounds.

As for IBM, it is working on what techniques such as directed self-assembly and lithographically directed polymer self assembly, which involves both digital (lithography) and analog (self assembly) manufacturing, Theis noted. But then directed self-assembly is not a new concept in chip manufacturing, he added. In processes today, fab engineers set conditions in process chambers, introducing the relevant substances, and the molecules involved assemble themselves accordingly on a wafer.