Using air as the core of optical fiber is not a new idea, but manufacturing problems have long stymied its commercial development. Now OmniGuide Communications Inc., a company formed out of technology from the Massachusetts Institute of Technology (MIT), hopes to make air-based fiber a reality.

OmniGuide is the third joint venture for Ray Stata and Mukesh Chatter. The partnership began with Nexabit Networks, a terabit router startup that was acquired by Lucent Technologies Inc. (NYSE: LU). Then earlier this year, the pair cofounded another optical networking company called Axiowave Networks Inc. (see Chatter's New Box). Now, they have stumped up $4 million to develop a weird kind of optical fiber. Yoel Fink, an MIT graduate student, is cofounder and CTO.

OmniGuide is making some pretty wild claims. According to director of marketing Uri Kolodny, the OmniGuide fiber will increase the available bandwidth by 1000 times compared to the state-of-the-art technologies that we have today. What’s more, it will eliminate the need for optical amplification and regeneration. "It’s mind boggling how expensive that is," he says.

How much money could be saved, exactly? Kolodny quotes the TAT-14 transatlantic cable as an example. With only four fiber pairs, it’s not a high-end system. Being a ring, the total distance covered is 10,000 kilometers. Each fiber requires an amplifier every 50km, which adds up to a grand total of 200 amplification points with four amplifiers at each one. Given that each amplification point costs upwards of a million dollars, you’re looking at total direct equipment costs in excess of $200 million. This doesn't consider the costs involved in installing, maintaining or powering up these devices.

To achieve its goal, the fiber will need to have both low loss and linear properties. What is the ultimate low-loss and linear medium? In a word, air.

That's right, OmniGuide’s fiber is designed to guide light through air. Ordinary fibers can’t do this because they need an optically dense (high refractive index) material at the core to prevent light from leaking out.

To visualize the structure of the fiber, imagine a layered structure like plywood that has been bent round to form a long, hollow tube. (In fact it’s a dielectric mirror made up of alternating layers of non-metallic materials.)

Kolodny claims that it is possible to pump a lot more power into the OmniGuide fiber without incurring transmission penalties due to non-linear effects. However, on the issue of loss he refuses to comment -- “probably because the fibers have high losses,” says one source who did not want to be named.

Loss aside, the startup faces a huge challenge in manufacturing. Reportedly, the fiber is “hellishly difficult to make," and some people are skeptical that it can be made at all. OmniGuide reckons it will be 12 to 18 months before it has a prototype, lending some credence to the suspicions.

Surprising as it might sound, the idea of guiding light through air is not new. Far from it, in fact.

“We have a different structure that does the same thing,” says Jonathan Knight from the University of Bath, in the U.K. He is busy setting up a spinoff called Blaze Photonics , to commercialize so-called “holey” fibers, which contain lots of air-filled holes that run through the length of the fiber.

There is also a venture in Denmark, called Crystal Fibre, with similar aims. The big component and fiber manufacturers like Corning Inc. (NYSE: GLW) and Lucent are also doing R&D in this area.

In fact, as far back as 1978 people had proposed using a hollow fiber to guide light in air, as Yoel Fink and his coworkers point out in the small print of a recent paper published in the leading academic journal Science.

The paper describes what OmniGuide calls the optical equivalent of a coaxial cable -- it’s got a solid rod in the center of the hollow tube. This is a product that the startup plans to develop later on, though manufacturing would appear to be even more of a problem.

-- Pauline Rigby, special to Light Reading http://www.lightreading.com