Optical components

Hollow Fiber: No Pipe Dream

Loss-free optical fiber may seem like a pipe dream, but it's one that just took a big step closer to becoming a reality.

At the OFC postdeadline paper session tomorrow, U.K. startup BlazePhotonics Ltd. will present results on a hollow-core fiber with an optical loss that's ten times better than what's been achieved before. What's more, BlazePhotonics says its results suggests that hollow core fiber could meet the target of having lower loss than standard transmission fiber.

"We're getting close to breakeven," says VP of research and development Dr. Henrick Sabert. "Now we can begin to understand the physics that take you there."

Being better than standard transmission fiber is no mean feat. After decades of development, the loss of optical fiber has settled at a figure of around 0.2 dB per kilometer (dB/km). The very best commercially available fiber is from Sumitomo Electric Industries Ltd., with a loss of just 0.151 dB/km.

Hollow-core Optical Fiber has the potential for virtually zero losses because light travels predominantly inside hollow spaces inside the fiber, rather than in glass. "Our results show that more than 99 percent of the light travels in the spaces," Sabert contends. BlazePhotonics' fiber has a honeycomb-like cross section, although other types have been proposed by other vendors (see Holey Fibers! for some images).

But achieving these low losses has turned out to be hard work, causing some folk to question if it is even achievable (see OmniGuide: Hollow Promises?).

Naturally, BlazePhotonics believes that ultra-low-loss hollow fiber is a viable proposition and says its latest results point the way. The new results that it will report tomorrow are losses of 1.7 dB/km, which is a significant improvement over the previous best of 13 dB/km, which was set by Corning Inc. (NYSE: GLW) in 2002.

Earlier fibers had so much loss that it was impossible to tell what the limiting factors were, according to Sabert. In the latest fibers, loss is caused predominantly by scattering of light at the surfaces between the spaces and the glass, and this mechanism appears to be fundamental -- it will set the limit on the lowest loss that can be achieved.

"For hollow core fiber to be commercially interesting, it would need to be a factor of two or four times better than standard fiber," says Sabert. He reckons a loss of 0.1 dB/km would be good enough. Lower-loss fiber would allow signals to travel farther between amplifier points, reducing or even eliminating the cost of using Erbium Doped-Fiber Amplifiers (EDFAs). Even small improvements in loss could result in significant cost savings.

Lower loss would be especially useful in two applications, he says: unrepeatered submarine links and quantum communications. "Folk in submarine battle to gain a fraction of a decibel." In quantum communications, which encode data using the physical properties of single photons, the signal cannot be amplified, because this would destroy the information it is carrying (see Quantum Crypto Gets a Chance).

BlazePhotonics also came to the conclusion that the optimum operating wavelength (the one with lowest loss) of a hollow core fiber is likely to be around 1,900 to 2,000 nanometers, rather than the usual 1,550 nm found in DWDM systems today. That's because the absorption by glass, which usually increases rapidly at wavelengths above 1,600 nm, is drastically reduced (most of the light travels in the spaces not the glass, remember). Since the surface scattering falls as wavelength increases, the new optimum point is at a higher wavelength.

This kind of argument isn't likely to make folk rip out their existing fiber and replace it. But in new-build scenarios, it could lead to some novel solutions.

— Pauline Rigby, Senior Editor, Light Reading

probably 12/5/2012 | 2:22:17 AM
re: Hollow Fiber: No Pipe Dream I have followed Blaze Photonics with interest for some time. I have a few questions/points and would be interested to hear other opinions:

1. Marketing question.

BP is a start-up company. As such, it has limited finances and resources and will be presumably feeling the pressure to make some sales soon. Historically the problem with many optical start-ups (to name a few!) is that they have over-hyped their technology and more importantly that they have been technology rather than market led.

BP therefore seems a little bit odd to me: they are pursuing a technology that hasn't been shown to equal the currently installed fibre base in terms of propagation loss. Perhaps with enough time and money they may do so, but they are doing this at a time when investor patience is at an all time low and the take up of unproven technology is not too impressive. In particular, the probably only have a best guess as to how long it will take to reduce the loss down to an acceptable level (it is always easy to convince yourself that you are almost there when you don't know how long the road is!).

2. Physics question.

Surely it is obvious that the *limiting* loss factor in hollow fibre is set by surface roughness resulting in scattering? The physics that the CTO talks about therefore comes down to reducing this through
a)A physics point. Reducing the field magnitude at the boundary, i.e. by more tightly confining the field in the voids.
b)A manufacturing point. Reducing the surface roughness of the interface by clever manufacturing and control of the fibre pre-form. (I was under the impression that the pre-form holes are drilled...perhaps this impressive loss reduction has come through chemically smoothing the preform holes with HF or some other technique?)
c)A pointless point. Convincing everyone to change wavelength to nearer 2um!

I would imagine that what they have been struggling against so far is not "physics" but a lack of understanding as to whether poor manufacturing or inaccurate modelling is resulting in the large loss.

The normal way of doing this is to model the propagation and to predict the loss given:

a) An ideal structure.
b) An ideal structure with some errors that they can only estimate (surface roughness, departure from symmetry, etc.). This is where the modelling inaccuracy mainly creeps in.
c) Manufacture it as best as possible.
d) Measure it in the lab and get something horrible compared to what is expected.
e) Modelling guys blame process guys and vice versa.
f) Back to b) and repeat until funding runs out.

I wonder:
1. How close the measured loss is now to the predicted loss? That is to say, with a *measured* surface roughness how close the model comes to predicting the loss.
2. What a histogram of loss per km would look like for each km made from the same preform. This would test intra batch consistency.
3. What a histogram of loss per km would look like batch to batch to test manufacturability.

Sorry about the length of this post...I find them to be an interesting case!

Peter Heywood 12/5/2012 | 2:22:13 AM
re: Hollow Fiber: No Pipe Dream Good stuff! I would very much like to hear Blaze Photonic's response. Are you there, BP?
opticalweenie 12/5/2012 | 2:22:03 AM
re: Hollow Fiber: No Pipe Dream Probably,
Nice points you make, but I would still argue that this is very important research that BP is doing. My reason for this is simple. You point out that the telecom community is not going to want to switch to 2 microns. I agree with that. But where BP could really make a mark is creating a low loss fiber for the 5 - 7 and 8 - 12 micron ranges, which is where homeland security sensing applications want to target today and where there is a great need for integrated, yeah even miniature, components.

Allow me one rant. It seems like everyone who posts on LR thinks that telecom is the only industry in the world and that all matters scientific, economic, politic should be tailored to fit only what this wallowing industry with wallowing unemployed whiners needs. I disagree, there is a lot more revenue and job generation in a host of other industries.

Well that should ruffle a few feathers!

probably 12/5/2012 | 2:21:45 AM
re: Hollow Fiber: No Pipe Dream OpticalWeenie,

I agree with your point about there being interesting markets other than telecom: my post was not meant to imply that Blaze Photonics should or are restricting themselves to telecom. I am particularly interested how this announcement from BP relates to their current position in terms of market, technology, funding, and the interplay between them.

In fact, out of all the optical start-ups I know well, the ones that seem to have the best chance of survival are those that have managed to diversify away from telecoms, sometimes dumping their core technology en-route. Old chestnuts like "defence" and "automotive" and new chestnuts like "homeland security" are appearing more freqently in the press releases.

Just as an aside, I don't think the phrase "Homeland Security" existed when BP was founded! This doesn't however reflect on the validity of OpticalWeenie's point: there are other markets to be investigated, even when these markets are not the basis on which the company was founded.

marc_goofy 12/5/2012 | 2:19:39 AM
re: Hollow Fiber: No Pipe Dream Hi There,

Absolutely right, Folks : there's a world out there, i mean outside the Telecoms. I always have in mind that very simple fact : SMA connectors, that were once *the* standard for the so-called datacom fibers (well, at this time, 100/140 and the like multimode fibers ;-) are widely used in Medical apps since years, for the benefits of their manufacturers (once you burn a patchcord, you need a new one...).

Just have to find that kind of "mass" market, right ?

Sign In