The Promise of the Optical Chip

By Jim Lorick

Even in the post dot-bomb economy, in which investors regard communications technology in particular with a fondness usually reserved for an adulterous spouse, hundreds of millions of dollars are being raised on the promise of the next big thing - the "optical chip." Before you jump on the optical chip bandwagon, however, it might make sense to pause for a short history lesson. Who knows, maybe this time knowing a little something about your bride before you head to the altar wouldn't be such a bad idea.

Once upon a time, long, long ago, something called the "electronic age" was born of technology that harnessed electrical energy. As technology became more sophisticated, the control of the flow of electrical energy was so precise that individual electrons could be directed down the microscopic "wires" in the surface of silicon chips. Now as we know, the modern world has become completely dependent on silicon chips and high-speed electronic micro processing. We have come to take this remarkable technology for granted, despite the fact that our offices, homes, cars and in some cases even our hearts, could not function without the power of the microchip.

With each generation of chip, the processing speed and capacity have increased tremendously. These staggering increases have fueled the growth of increasingly complex applications that continue to push the limits of chip innovation. In the quest to keep a quantum leap ahead of the demands for speed and capacity, science and industry may now be poised on the edge of a new technological revolution built upon the optical chip. Like the "wires" that guide electrons in a silicon chip, the technology that will usher in the "photonic" age uses optical channels to direct photons. In the broadest terms, the advantage of using photons over electrons is in basic physics. Photons are faster than electrons and, theoretically, can move without significant loss of energy.

The scientific discipline of "photonics" developed in the aftermath of the creation of the first Laser. Scientists began to investigate the way in which light could be channeled by creating different materials through which the light could be transmitted. The development of fiber-optic communications networks was a significant contribution from the first phase of the photonic revolution.

The current developments in photonics represent attempts to cope with the problems created in converting electronic information into light, for swift transmission, and then back into electronic information for the end-user. While electronic equipment has become smaller, cheaper and mass-produced, the filters, routers and switching gear that support the electronic-photonic conversion are comparatively bulky, expensive and essentially hand-built. Development of optical chips for this environment is seen as a natural extension of photonics, and as a proving ground where the tantalizing promise of the optical chip will be tested.

The "photonic integrated chip" built by LNL Technologies Inc., is billed as an important step on the road to creating an advanced microprocessor of light signals. While the current chip is light years less sophisticated than a Pentium chip, with its millions of processors, LNL hopes to create devices to replace bulky telecommunications gear by combining lasers, routers, receivers and other devices commonly used in fiber-optic networks into a single device that would be comparable in size to a Pentium chip.

However, before you decide that LNL is in fact the optimal solution, let's not forget about last year's big announcement from Cupertino start-up Infinera. Once again tackling the problem of telecommunications gear, the "integrated photonic circuit" developed by Infinera promised to shrink telecommunications components down to microscopic levels and combine them on a slice of indium phosphide (InP) instead of silicon. Like LNL, Infinera said its chip controlled photons at micron levels - roughly 1/50 the width of a human hair - and could be modified easily to fit into any piece of optical-networking equipment.

LNL announced in January that it had raised $7.1 million in seed funding from investors that include members of the management team; Adam Chowaniec, chairman and founder of Tundra Semiconductor; Larry Mohr, founder of Mohr, Davidow Ventures; and Sandy Robertson, founder of Robertson Stephens and chairman of Francisco Partners. However, Infinera seems to be winning the funding war, having raised $86 million through Accel Partners, Benchmark Capital, and Kleiner Perkins Caufield & Byers. Infinera is succeeding in the publicity department as well, having been featured in May in the "Red Herring 100: Future contenders" as one of ten young companies likely to create new markets or disrupt existing ones.

Silicon chip giant Intel has been watching the development of optical circuits with great interest and has recently acquired several companies at home and abroad who are poised to make significant advances in optical networking and other opto-electronic and photonic technology sectors.

As with advances in electronics, innovation in photonics depends on innovations in design, materials and manufacturing. And as the technology emerges, the interdependence of each area of innovation on the others is clear. A possible means of creating the optical channels is to use silicon itself. Microscopic layers of the material are etched to create channels or "wave guides." Because of the unique behavioral characteristics of different types of light, and the importance of not creating a distortion or shift in the wave-length, researchers are experimenting with ways of coating and constructing the layers, and developing new tools to cut the wave guides.

A potentially significant innovation in materials may offer a non-silicon based approach to solving the problem of creating stable broad-spectrum wave guides. Scientists at the University of Toronto recently announced a breakthrough development. Now this gets a bit esoteric for most, but if you've stayed with us so far this shouldn't slow you down - and I'm afraid this information will be on the "are you a smart technology investor?" quiz.

Canadian researchers claim to have developed a hybrid plastic that can be electrically stimulated to produce light at wavelengths used for fiber-optic communication. The colors of light the researchers generated, ranging from 1.3 microns to 1.6 microns in wavelength, spanned the full range of colors used to communicate information using light. If this material proves to be more than another fantasy of cold-fusion, it will go a long way in advancing the development of the next generation optical computer chip.

That material, developed by a joint team of engineers and chemists, is a plastic embedded with quantum dots - crystals just five billionths of a meter in size - that convert electrons into photons (aka "nano-crystals"). The first obvious use of such a material would be in applications to create direct links between high-speed electronic computers and the high-speed photonic networks that transmit information using light. Another promising development involves the way in which this new material was created. The team created nanocrystals of lead sulphide using a cost-effective technique that allowed them to work at room pressure and at temperatures of less than 150 degrees Celsius. Traditionally, creating the crystals used in generating light for fiber-optic communications means working in a vacuum at temperatures approaching 600 to 800 degrees Celsius.

Like the Canadians, Australian scientists have been part of a significant national effort to make sure that their homeland is not left out of the next technological revolution. Researchers at the Australian National University's Research School of Physical Sciences and Engineering are working on a solution to the problem of distortion caused by creating wave-guides using the same technology currently used to create electronic chips. The Australians are using a system called "laser direct write". Instead of using a mask to draw the patterns for the wave guides, the group uses a scanned ultra-violet laser beam to write the pattern into the optical chip. In the past this technique has failed to realize its full potential because of the "Gaussian," or bell shaped, intensity profile characteristic of laser beams. The Australians have overcome this difficulty with a novel doughnut profile beam that gives much more uniform intensity and makes better wave-guides.

Now that we've suffered through all that science stuff, the real question now is whether the development curve of the optical chip is going to mirror that of the silicon chip. Will the speed and capacity of this new technology also grow geometrically? Are these really going to supplant electronic chips - or will the market find the real power of these optical chips is in a complementary application?

As the technology advances, it seems a good bet that mass-produced optical chips will become ubiquitous in voice and data networks and may become part of new systems designed to deliver internet access, high-bandwidth movies, music and games to consumers. But as optical chip technology continues to evolve along competing paths, it is hard to predict which avenues will ultimately prove fruitful, and which will prove to be dead-ends. Investors and industry are hoping that the next Intel is in sight and perhaps in their portfolio.

Return to this week's issue of VOX