Oak Ridge National Laboratory


News Release

Media Contact: Ron Walli (wallira@ornl.gov)
Communications and External Relations


Optical switching could cause shakeup in electronics industry

OAK RIDGE, Tenn., Feb. 14, 2000 — Electronic transistors may one day be replaced by all-optical transistors, which are in early stages of development at the Department of Energy's Oak Ridge National Laboratory (ORNL).

"Our findings could have tremendous impact on the development of optical switching technology and could lead to applications in communications, sensing and computation," said Panos Datskos of ORNL's Engineering Technology Division. "We're talking about using photons instead of electrons and creating a new kind of transistor that's 100 times faster than today's transistors."

A transistor is an active component of an electronic circuit consisting of a tiny block of semiconducting material to which at least three electrical contacts are made. Transistors, which can be used as amplifiers, detectors or switches, are integral parts of computers, telephones and virtually all electronic components.

The work at ORNL builds on research that led to the development of uncooled micro-mechanical infrared detection. Datskos and colleague Slo Rajic's micromechanical quantum detector is a highly sensitive miniature photon - or light - detection device based on photo-induced stresses in semiconductors. It received an R&D 100 Award for being one of the most significant technological developments of 1999.

"Optical switching by photon activation can be thought of as a transistor with input photons controlling the signal photons," Rajic said. "It's analogous to the electronic transistor but it incorporates the benefits of photonics."

As speeds in fiber optic communications, sensing and computation processing increase, faster switching mechanisms are necessary.

The advantages of optical transistors include greater bandwidth and speed close to that of light. Developers worldwide, however, have been unable to overcome certain obstacles, Rajic said. They have had to make significant tradeoffs in at least one of four basic required areas: throughput vs. rejection efficiency, cost, speed and size.

"What we have is a novel approach for optical switching that might optimize all four parameters simultaneously," Datskos said. "Efficient, low-cost, fast and large-scale integration of an all-optical switch or transistor will revolutionize computing much as fiber optics have revolutionized the telecommunications industry."

While several approaches to developing optical switching mechanisms have been pursued for more than 20 years, they have not resulted in a practical all-optical transistor. The technique used by Datskos and Rajic uses a diode laser LED that causes optical absorption of waveguide material. If enough stress is created by absorption of the light, it puts a strain on the material. By making the waveguide move even a tiny amount, it can cause a switch to be turned on or off or can redirect light coming from the end onto two or more channels. It's equivalent to a switch or modulator used in conventional transistors.

The researchers have already seen this effect with frequencies in the 1 megahertz range. That's significant, Datskos said, because these devices can be driven thermally to just 1,000 hertz, thus the photon-driven transistor could be at least 100 times faster than a conventional transistor driven by electrons.

"The stress induced by the absorption of photons in a semiconductor occurs extremely quickly, making it far faster than any heat-generated stress," Rajic said.

As with conventional semiconductors using electrons, the devices would be tuned to achieve specific results by modeling physical and geometrical properties. Given mass production using micro-electronics fabrication methods, the price of these devices could be well below $100 with more capabilities and functionality than other devices that cost more than $20,000, Rajic said.

A significant amount of work remains, but Datskos and Rajic are encouraged by early results of their work, which is funded by the Laboratory Directed Research and Development program, provided by DOE.

ORNL is a DOE multiprogram research facility managed by Lockheed Martin Energy Research Corporation.