BOSTON A systems strategy to make nanoscale devices as reproducible and controllable as silicon systems-on-chip is being urged by a pioneer in this growing world of shrinking sizes.
Charles Lieber, a Harvard University chemistry professor and co-founder of Nanosys Inc., said that while extensive work has gone into building nanoscale devices, researchers seem far from anything as practical and serviceable as current electronic fabrication methods.
"We have to go beyond the idea of testing single devices," he told this month's Nano Science and Technology Institute's annual Nanotech conference here. "It's a lot of fun but one needs to learn how to organize these ideas into increasingly complex systems." And, he said, that must be done with a design that both takes advantage of the state-of-the-art's strengths and compensates for its weaknesses.
"Making things small is probably one of the least unique things about nanoscience. It certainly enables massive integration, but making small things has been around for literally billions of years," said Lieber, referring to biological systems.
What sets the new wave of nanoscale devices apart from previous technologies is the creation of fundamentally new physical properties that derive from size scales. And interactions across different size scales will be the key to a systems approach to building practical systems.
"Nanoscale structures can exhibit new properties which really depend on size, morphology and composition. More important is the fact that building blocks can be utilized and combined in ways that simply aren't possible with conventional manufacturing," he said. "If you couple that with the idea that you can organize things on multiple length scales, that will enable the creation of complex functional systems with few limitations."
Although the basic building block might be as simple as the field-effect transistors that form the basis of current integrated circuits, the resulting systems could address a much broader range of applications. Computers, medical instruments, photonic devices, information storage systems and displays could all be built from the same basic component, unlike electronic systems, which are mostly confined to silicon chips that process digital information.
For Lieber, the "key issue" is how to decide what that fundamental component of nanotechnology should be. "One needs a synthetic approach that is general and predictable-control chemical composition, control size and control structure on different length scales," he said.
Lieber enumerated the properties that will be required of a building block for nanotechnology. The basic component must support a wide variety of functions that can be controlled by varying basic physical parameters. The processes used to fabricate them must be low-temperature-near room temperature, not the higher temperatures encountered in semiconductor fabrication. Finally, the basic components must support interconnection and integration easily as part of their structure.
While a lot of work has been done on carbon-nanotube-based devices, that ap-proach will only lead to a dead end because they do not meet those basic criteria. Carbon nanotubes can be grown in a variety of forms that have both semiconducting and conducting properties. But they are in a sense too perfect: Carbon nanotubes cannot be doped or have varying physical properties along their length. Moreover, they can be formed only at high temperature, which makes them difficult to integrate into a general fabrication process.
Nanowires formed from semiconductors such as silicon or gallium nitride will meet the criteria Lieber believes are essential for a nanoscale building block. He cited recent work at his lab and results from other groups showing that semiconducting nanowires with a variety of sizes and dopants can be fabricated in a low-temperature process.
The basic idea was discovered a few years ago by a student at the California Institute of Technology, who found a nanoscale catalyst that could synthesize semiconducting nanowires while controlling both their size and composition. Also, the catalyst operates at room temperature, meeting the requirement for low-temperature synthesis.
Since then researchers have found that single-crystal silicon nanowires can be controllably grown with any predetermined diameter. The basic process can be applied to virtually any material, Lieber said.
At Nanosys, researchers have been able to grow nanowires with a wide variety of electronic properties all the way down to molecular scales. "The beauty of this intellectual idea is that it can be applied to grow any technologically important material in single-crystalline form at any diameter down to molecular dimensions," he said.
In addition to being able to build long structures with a variety of functions, Lieber and his colleagues have found ways to distribute arrays of them on a substrate and sequence the process to build up multiple layers of functional nanowires that intersect. That will promote the fabrication of complex systems based on catalyzed nanowires.
Finally, the problem of interconnecting these systems with macroscopic electronics is being tackled with multiplexing schemes and methods for contacting nanowires with metal electrodes.