Corning, Inc. (GLW)

[rapoole]

Progress In Optical Chips

Maybe the fiber market does have a future, for those with patience.

http://online.wsj.com/article_print/0,,SB1042067428547244144,00.html

Can New Optical Chips
Replace Telecom Gear?

By ANTONIO REGALADO
Staff Reporter of THE WALL STREET JOURNAL


CAMBRIDGE, Mass. -- How many engineers does it take to send light through silicon?

At a recent technical demonstration here, about a dozen stuffed themselves into a dark, closet-sized room to watch a light beam curve around a microscopic racecourse etched into a silicon wafer.

On display was a "photonic integrated chip" built by LNL Technologies Inc., a device that could one day become the Pentium processor of light signals. LNL (www.lnltech.com1) says its goal is to replace bulky telecommunications gear with devices the size of a quarter that combine lasers, routers, receivers and other devices commonly used in fiber-optic networks.

Simple optical chips already are on the market -- but they aren't nearly as sophisticated as the chips inside a personal computer, whose processors contain billions of transistors. Still, the technology is advancing quickly, and in five years, mass-produced versions could be slashing costs of voice and data networks, and could eventually become a new type of gateway to deliver high-bandwidth movies, music and games directly to the home.

The possibility of miniature optical circuits has drawn interest from computer-chip manufacturers such as Intel Corp., which calls the concept "silicon photonics" internally. Eric Mentzer, chief technology officer of Intel's communications group of 6,000 workers, says that if telecommunications components can be crafted from silicon, they may start to follow their own version of Moore's law, the rule of thumb that predicts how quickly computer-processor performance will double. "If we can figure out how to make that work, it will make optics very inexpensive, and therefore ubiquitous," he says.

BUILDING BANDWIDTH



• Page One: After Internet's Big Bust, Broadband Shift Went On2
01/08/03

• Wildly Optimistic Data Drove Telecoms to Build Fiber Glut3
09/26/02




At LNL, it's still more Murphy's law than Moore's. "Don't be distracted by a little dirt," says Christian Hoepfner, director of technology development, as light pouring out of the chip's edge appears on an overhead video monitor. The chips, made in Singapore, had arrived two days earlier and a hastily glued-on cover had trapped speckles of dust.

LNL faces a tough market. The telecommunications sector is reeling from a glut of capacity following years of massive infrastructure investments, and with capital costs and work forces being cut, research and development has taken a back seat. The companies that sell the lasers, routers, switches and other devices that shuttle photons along fiber-optic lines have been among the hardest hit.

However, developments in optical chips are being closely tracked, in part because they may speed the goal of "adding intelligence and reducing cost" in networks, according to Patrick Nettles, chairman and founder of telecom-equipment maker Ciena Corp. Instead of bulky, custom-built systems built from racks full of lasers and switches, photonic chips could permit cheap, mass-manufactured devices.

Optical chips aren't a new idea. In 1969, Bell Laboratories engineer Stewart E. Miller envisioned "guiding laser beams on miniature transmission lines" made using the same manufacturing methods then blossoming in integrated circuits. He cautioned that, due to technical challenges, his ideas were "goals rather than accomplishments."

Today, engineers have solved many of those issues. Demonstration devices have combined lasers, light detectors and "demultiplexers" that carve out data from different wavelengths of light on semiconductor chips just a centimeter across.

But moving from the lab to real products remains challenging, and rocky markets have thinned out the ranks of start-up companies. LNL's founders, who formed the company in March 2001, bought out patents held by two failed ventures, including former highflier Nanovation Technologies Inc., which once promised $90 million to fund research at the Massachusetts Institute of Technology. Nanovation delivered only $4.2 million before filing for Chapter 11 bankruptcy protection. Now former MIT graduate students Desmond Lim and Kevin Lee head LNL's research efforts.

Some start-ups in the field say the slowdown is helping. "The competitors we are looking at have had a difficult financial time over the last couple of years, and have had to focus on incremental product development," says Dave Welch, chief technical officer at Infinera Corp., of Sunnyvale, Calif. (www.infinera.com4)

Infinera has positioned itself as the photonic-chip sector's hottest property, and has raised more than $100 million in post-bubble financings led by top-tier venture capitalists. The company has been secretive so far, but indicates it is developing a cheap photonic chip that can switch light signals into electronic ones, and vice versa.

Optical fibers carry data cross-country on "long-haul" routes and link to metropolitan networks, in some cases directly to large businesses. But equipment costs and other factors still prevent fiber from being more widely used. Right now, it would cost around $10,000 per home to switch light into an electronic signal for a home PC or television. "As the technology creeps out from the core, obviously the cost sensitivities are much, much higher," says Intel's Mr. Mentzer, who nevertheless believes optical chips could one day provide "liquid bandwidth" to the home.

The notion of killer consumer applications continues to motivate many investors, such as Warren Zide, the Hollywood producer responsible for the "American Pie" films. Mr. Zide, along with Sandy Robertson, a founder of investment bank Robertson Stephens, and others have so far provided LNL with around $7.1 million in capital.