Beyond the Moore's Law: Nanocomputing using nanowire tiles

Jan 27, 2014
The nanowire finite-state machine (in false color) fabricated from assembled nanowires occupies the small central region of the chip with all other features corresponding to metal lines used for testing the integrated system. Credit: Jun Yao and Charles Lieber, Harvard University.

An interdisciplinary team of scientists and engineers from The MITRE Corporation and Harvard University have taken key steps toward ultra-small electronic computer systems that push beyond the imminent end of Moore's Law, which states that the device density and overall processing power for computers will double every two to three years. In a paper that will appear this week in the Proceedings of the National Academy of Sciences, the team describes how they designed and assembled, from the bottom up, a functioning, ultra-tiny control computer that is the densest nanoelectronic system ever built.

The ultra-small, ultra-low-power control processor—termed a nanoelectronic finite-state machine or "nanoFSM"—is smaller than a human nerve cell. It is composed of hundreds of nanowire transistors, each of which is a switch about ten-thousand times thinner than a human hair. The nanowire transistors use very little power because they are "nonvolatile." That is, the switches remember whether they are on or off, even when no power is supplied to them.

In the nanoFSM, these nanoswitches are assembled and organized into circuits on several "tiles." Together, the tiles route small electronic signals around the computer, enabling it to perform calculations and process signals that could be used to control tiny systems, such as miniscule medical therapeutic devices, other tiny sensors and actuators, or even insect-sized robots.

In 2011, the MITRE-Harvard team demonstrated a single such tiny tile capable of performing simple logic operations. In their recent collaboration they combined several tiles on a single chip to produce a first-of-its-kind complex, programmable nanocomputer.

"It was a challenge to develop a system architecture and nanocircuit designs that would pack the control functions we wanted into such a very tiny system," according to Shamik Das, chief architect of the nanocomputer, who is also principal engineer and group leader of MITRE's Nanosystems Group. "Once we had those designs, though, our Harvard collaborators did a brilliant job innovating to be able to realize them."

Construction of this nanocomputer was made possible by significant advances in processes that assemble with extreme precision dense arrays of the many nanodevices required. These advances also made it possible to manufacture multiple copies of the nanoFSM, using a groundbreaking approach in which, for the first time, complex nanosystems can be economically assembled from the bottom up in close conformity to a preexisting design. Until now, this could be done using the industry's expensive, top-down lithographic manufacturing methods, but not with bottom-up assembly.

For this reason, the nanoFSM and the means by which it was made represent a step toward extending the very economically important five-decade-long trend in miniaturization according to Moore's Law, which has powered the . Because of limitations on its conventional lithographic fabrication methods and on conventional transistors, many industry experts have suggested that the Moore's Law trend soon may come to an end. Some assert that this might occur in as little as five years and have negative economic consequences, unless there are innovations in both device and fabrication technologies, such as those demonstrated by the nanoFSM.

James Ellenbogen, chief scientist for nanotechnology at MITRE and an expert in the development of computers integrated on the nanometer scale, said, "The nanoFSM and the new methods that were invented to build it are not the whole answer for the industry. However, I believe that they do incorporate important steps forward in two of the key areas the electronics industry has been focused upon in order to extend Moore's Law."

In addition to Das and Ellenbogen, the development team at MITRE included James Klemic, the corporation's nanotechnology laboratory director. The researchers from MITRE—a pioneer in the nanotechnology field since 1992—collaborated with a three-person team at Harvard, led by Charles Lieber, a world-leading investigator.

Explore further: Nanodevices for a 'More than Moore' world

More information: "Nanowire nanocomputer as a finite-state machine," by Jun Yao et al.

Related Stories

Nanodevices for a 'More than Moore' world

Oct 04, 2013

Moore's Law - which holds that the number of transistors on an integrated circuit, and hence its processing power, doubles every 18 months - has been the guiding principal of chip design for almost half a ...

New nanowire transistors may help keep Moore's Law alive

May 02, 2013

( —Two French researchers, Guilhem Larrieu and Xiang‑Lei Han, may have succeeded in possibly setting back the date to which Moore's Law would no longer apply by creating a new kind of nanowire ...

Plasma etching pushes the limits of a shrinking world

Nov 10, 2011

Plasma etching (using an ionized gas to carve tiny components on silicon wafers) has long enabled the perpetuation of Moore's Law -- the observation that the number of transistors that can be squeezed into an integrated circuit ...

Scientists seek silicon's successor

Aug 08, 2013

In the hunt for a sequel to silicon, scientists at the SLAC National Accelerator Lab have flipped an "on-off" switch in the mineral magnetite that is far faster than today's transistors.

Recommended for you

First direct observations of excitons in motion achieved

19 hours ago

A quasiparticle called an exciton—responsible for the transfer of energy within devices such as solar cells, LEDs, and semiconductor circuits—has been understood theoretically for decades. But exciton ...

User comments : 2

Adjust slider to filter visible comments by rank

Display comments: newest first

5 / 5 (3) Jan 28, 2014
How much faster? Smaller? Less power? Are they less susceptible to other problems like quantum tunneling? Or is the point that these nano-switches form a new architecture?
Peter Trypsteen
not rated yet Jan 31, 2014
Very interesting, maybe these nonvolatile nanoswitches could be used as memory.
Maybe an alternative for flash in SSD's.

The low-power and miniaturization characteristics of these nanoelectronic devices would be good for medical microrobots.

Last these nanoelectronic devices might be very good for building blocks for FPGA's.

More news stories