TCS Daily

No Speed Limits

By Charles Matthew Rousseaux - July 20, 2004 12:00 AM

Regardless of whether it is the leaden tidings of tragedy or the light gossip of Hollywood, most of this era's new currency now travels at astonishing speed. Information's near Einstein-ian speed imbues it with value (although the worth of news about Britney and J-Lo is certainly relative). But bits of information could travel even faster -- all of them at the speed of light -- if the developing field of silicon photonics produces its promise.

In silicon photonics, information travels through silicon at the speed of light, which is even faster than the speed of electrons. Fiber-optic lines already do so, but the information from silicon devices must be converted from the language of electrons into the language of light and back, a costly process since the materials used are expensive, the systems are bulky and the devices must be precisely aligned (sometimes by hand) so the data streams are not distorted.

In a true silicon-based optical system, there will be no meddling information middleman. Instead, streams of light will race through individual circuits and entire computing networks uninterrupted. In addition to their greater speed, optical systems also have the potential to carry far more information, than their electron-based counterparts. As subjects of two rules -- economies of scale and Moore's law -- the devices making up those networks will sprint past one another, each new model being cheaper and faster than its predecessor

Light is slowly catching up to electrons in silicon systems. Last February, Intel scientists announced -- through publication in the journal Nature -- that they had fabricated the first-ever optical silicon modulator. It transfers data at about 1 billion times per second (1 GHz), far faster than the 20 million cycles per second (20 MHz) of current devices. In June, Science and Nature published separate papers on similar developments speeding the field. In the Science paper, researchers reported that they had made a three-dimensional photonic crystal that suppressed light emission in all parts except for an artificially created flaw that allowed light through. In the Nature paper, scientists also announced they had created a three-dimensional photonic crystal with designed microcavities that guide light through it in a controlled, predictable way. As Graham Reed noted in a Perspective article in the same issue of Nature that published the breakthrough modulator, "The difficulties surrounding the optical performance of silicon seem to be being demolished one by one."

Still, additional technical and economic challenges remain. As Davis Janowski noted in PC magazine article a year ago on technologies to watch, innovators will have to develop an integrated optical circuit, which is, "a chip with a light source, optical filters, photodetectors, and optical wave guides." Such a chip will also have to be relatively cheap for economies of scale to tip the market scales in favor of photonics.

There are several ways that policymakers could help speed developments in that field and others, according to Paul Otellini, CEO of Intel. At a recent speech to the Western Governors Conference, Mr. Otellini described several ways to increase U.S. competitiveness including promoting innovation, producing a well-trained workforce and developing a better information infrastructure.

The last may be the easiest, since both President Bush and Sen. John Kerry have come out in favor of expanding broadband access. Still, the nation has a ways to go. Producing a well-trained workforce will be more difficult, since relatively few U.S. students are pursuing careers in the physical sciences. In the short-term, Mr. Otellini argued that foreign students pursuing advanced degrees in technical fields should be exempted from the H1B visa cap. He also said that federal spending on basic research should be doubled over the next seven years. Tax structure is another concern. Mr. Otellini said that his company had recently discovered that, over the course of a decade, it would cost his company $1 billion less to locate an advanced chip factory in Asia instead of the Western U.S. Surprisingly, neither labor nor materials costs accounted for the difference -- rather it was taxes and incentives. Mr. Otellini claimed that U.S. companies could reduce that offshore competitive edge by making research and development tax credits permanent and to looking into other measures such as manufacturing investment tax credits and free trade zones.

Those proposals merit serious consideration by policymakers. In a world in which even lite information travels at light speed, communications and computing will be combined, allowing computers to be composed of devices miles apart and consumers to view sporting events from almost any angle desired. A country with silicon bones and light-speed reflexes will pulse with information capital and have an autonomic ability to innovate.

It's a dream worth embracing; a hope worth chasing, if for no other reason than in a world of silicon photonics, consumers will have instant access to a high-definition version of the latest Hollywood blockbuster -- doubtless starring Britney and J-Lo.

Charles Rousseaux is an editorial writer for The Washington Times and a frequent TCS contributor. E-mail


TCS Daily Archives