In 2000, startup OmniGuide Communications Inc. boasted that its hollow-core fiber would revolutionize telecom transmissions, by eliminating the need for optical amplification (see A Fiber Filled With Air).

Three years on, despite never having launched a product, the startup has managed to secure another $15 million in funding (see OmniGuide Scoops Up $15M). What's more, this round was obtained at a higher valuation than previous rounds, according to the company. So what's going on?

Not surprisingly, it's a story of diversification. "Fortunately, we have a very fundamental technology," says Uri Kolodny, OmniGuide's director of marketing. "Our fibers can be manufactured at other wavelengths and have applications in medical, military, and industrial markets."

OmniGuide is close to getting its first product to market -- a fiber for delivering laser light to human tissue in medical procedures -- and has shipped samples for evaluation to a couple of carefully-selected institutions. Silicon is opaque to the laser wavelengths used in surgery (for example 10.6 microns), so no fiber currently exists for this application. This, rather than telecom, has likely enticed the investors.

But what about the telecom applications? On the one hand, Kolodny says the startup is continuing "full force" with development of telecom transmission fibers. On the other hand, he says these products definitely won't be ready this year.

The delay, he says, is due to the difficulty in achieving the incredibly low loss level required in telecom transmissions. Given that hollow fiber was originally hyped for its low loss qualities, this sounds like a pretty fundamental problem. Kolodny declined to elaborate, but competitor BlazePhotonics Ltd. was more forthcoming.

Telecom applications are particularly demanding for several reasons, says Dr. Henrick Sabert, VP of research and development at BlazePhotonics. For starters, the lengths of fiber used in telecom typically run into hundreds of kilometers, so loss effects are multiplied. Medical laser delivery would require only ten or so meters of fiber, so a loss of even 1 dB per meter (dB/m) would be acceptable. The best commercially available optical fiber has a loss of around 0.2 dB per kilometer (dB/km).

Another challenge is scaling down OmniGuide's fiber technology by a factor of ten, from 10.6 microns used in surgery to 1.5 microns for telecom. The losses in conventional fiber are due to material scattering and absorption. In hollow fiber, the center is made of air, so these two mechanisms disappear but are replaced with a new source of loss -- surface roughness at the interfaces that make up the fiber (see Holey Fibers!, page 3, for some examples of fiber structures). If the surfaces are not perfectly smooth, they will reflect and refract light at odd angles. Surface roughness matters more at shorter wavelengths, according to Sabert. So losses are likely to increase when a structure is scaled down.

Finally, there are issues of Chromatic Dispersion and Polarization Mode Dispersion (PMD). "All dispersion compensating technology will have to be redesigned to work with hollow fiber," notes Sabert.

BlazePhotonics has already announced a range of hollow core fibers, although not at telecom wavelengths (see BlazePhotonics Debuts Hollow Fiber). "It will take another year, and then we will know one way or the other" whether Blaze can achieve low-loss hollow fibers, Sabert claims. "Without bullshitting, it is tantalizingly within our grasp."

BlazePhotonics hopes to announce a new best for optical losses in hollow fiber at the upcoming Optical Fiber Conference. The best result to date was announced by Corning Inc. (NYSE: GLW) at ECOC 2002. Corning achieved 13 dB/km. Remember that commercial transmission fibers are as low as 0.2 dB/km, so there's clearly a long way to go.

— Pauline Rigby, Senior Editor, Light Reading