As copper conduits in silicon chips are fast approaching their physical limitations, chipmakers have suggested using carbon nanotubes to create thinner wiring that will work on conventional silicon chips. Thinner wiring will allow packing an even larger amount of transistors on a given surface. Carbon nanotubes have potential with a lot of prospect applications but the concept, however, remained unproven. That was until electrical engineers at Stanford University and Toshiba have unveiled a 1GHz CMOS circuit that uses carbon nanotubes as an interconnect.
The 11,000-transistor chip has 256 ring oscillators and was fabbed at Taiwan Semiconductor Manufacturing Company (TSMC). According to electrical engineering professor Philip Wong, this particular chip is the first to demonstrate a commercial speed of 1GHz – not desktop speeds to be certain, but still promising.
When designing the chip, the team of researchers purposely left one wire of each oscillator unconnected so the circuit is not completely wired up. After the semiconductor foundry TSMC made the chip, a few more fabrication steps at the Stanford Nanofabrication Facility was done to complete the missing connections with the nanotubes. Each nanotube measured between 50 and 100 nanometers (billionths of a meter) in diameter and about 5 millionths of a meter in length.
The nanotubes, purchased from a commercial vendor, were “metallic” in that they were synthesized for maximum electrical conductivity.
The quality of the nanotubes and their connections varied widely, but in the end 19 of the ring oscillators were successfully connected. The nanotubes rested directly above the transistors they were connecting, minimizing electrical capacitance and allowing for the transmission of zeroes and ones at 1.02GHz, or billions of times a second, in the best case. In 16 of the 19 good connections, the oscillators ran at speeds better than 800MHz, or millions of cycles a second.
Tubes rather large
Although the accomplishment is impressive, some industry observers note that the chip in question is not a ”true” carbon nanotubes design. The core is an array of ring oscillators. Each oscillator (all built on standard copper) is missing a connection. The nanotubes in question were inserted to bridge that gap and complete the chip’s various circuits. The tubes in question are rather large at 5 microns long and 50-100nm in diameter – researchers eventually hope to build single-wall nanotubes as small as 1nm in diameter.
One of the challenges of working with nanotubes has been assembling them into appropriate structures. With some ingenuity, the team got around the issue by designing a circuit with multiplexing circuitry that allowed each individual oscillator to be charged separately. A solution of nanotubes was then floated over the oscillators, and an AC signal was applied. Once a nanotube snapped into place to bridge a particular oscillator gap, the AC charge was shut off for that particular oscillator until all the individual circuits were completed with their nanotube components.
Conclusion
The team of engineers says that while the concrete demonstration bodes well for the future, no direct application is to be expected soon. Many improvements are needed for nanotubes wiring to enter commercial use. Nevertheless, this kind of demonstration – mostly proof of concept level – is expected to push forward in the years to come.
Courtesy of David Orenstein, Communications and Public Relations Manager, Stanford School of Engineering
|