By Peter Weiss
After years of building individual electronic components from single molecules and submicroscopic rods, researchers have now linked the tiny devices into prototype circuits.
Separate teams at Harvard University and Delft University of Technology in the Netherlands have built the circuits out of transistors, wires, and other components as narrow as a few atoms across. The circuits can carry out simple computations, such as adding two bits together.
By creating such circuits, teams led by Charles M. Lieber at Harvard and Cees Dekker at Delft may be opening the door to electronic chips far more powerful and compact than those that can be made by current means, the scientists say. Both research groups describe their novel circuits in the Nov. 9 Science.
These studies represent “dramatic steps toward the realization of electronic nanocomputers,” say Greg Y. Tseng of Stanford University and James C. Ellenbogen of Mitre Corp. in McLean, Va. in a commentary accompanying the research articles.
In today’s chips, even the smallest transistors span thousands of atoms, or hundreds of nanometers. Chip makers build such components using a process in which they apply semiconducting, metallic, and insulating layers to a semiconductor wafer to create microscopic circuitry. Manufacturers orchestrate the procedure using light for imprinting patterns onto the wafer.
For some 40 years, engineers have steadily shrunk those patterns, roughly doubling the number of transistors on a chip every 18 months. Many electronics specialists predict that this miniaturization may end within another decade or two, as such devices reach their physical limits. The new prototype circuits, which are slower than today’s circuitry, may be forerunners of the devices that will supersede conventional microelectronics.
In earlier studies, Lieber’s group showed that crossing two semiconductor nanowires, each thinner than a virus, can create a transistor or some other electron-controlling component (SN: 5/5/01, p. 286: Getting Nanowired). The scientists have now fabricated multiple transistors on a silicon wafer by flowing liquids containing suspended nanowires over it. Flows in perpendicular orientations direct many filaments at a time to lay down, making a criss-cross arrangement. Using metal contacts deposited by conventional means, the researchers then link those components into circuits.
Dekker and his colleagues make their transistors out of so-called carbon nanotubes (SN: 5/9/98, p. 294), which are much thinner than nanowires. Each tube is composed of a one-atom-thick sheet of carbon rolled so that the sheet’s long edges cinch together. The resulting tubular molecules are roughly a nanometer in diameter and can be many micrometers long. The Delft researchers also rely on conventional metal contacts to make working circuitry.
Other research teams have recently made rudimentary circuits from nanotubes and from organic molecules, but the Harvard and Delft circuits are more complex, note Tseng and Ellenbogen. The new designs also afford scientists more control over individual transistors.
Despite the transistors’ minuscule dimensions, their conventional-size contacts require that the circuits span relatively large swathes of wasted space, notes Paul S. Weiss of Pennsylvania State University in State College.
Ways to avoid that unwanted sprawl are already under development, Lieber’s team reports. It’s working with novel electrode arrays that may eventually permit densities of nearly a trillion transistors per square centimeter, which would be like packing the circuitry of tens of thousands of Pentium chips into the space that one chip now occupies.