At first sight, these new materials are simply odd: thin as a hair, transparent and full of holes. Like the optical fibers that are the mainstay of the telecommunications industry, they’re made of glass. But there the similarities with conventional materials come screeching to a halt.
The center of each of these novel fibers-which are made at the University of Bath, in England-is hollow. In existing optical fibers, light is transmitted through a glass core. In the fibers made at Bath, light travels unhindered through air. The light beam is confined to the hollow core by the holes in the surrounding glass material, which looks like a honeycomb in cross section and creates a strictly no-go region for light. The ability to confine light in air this way, says Philip Russell, a Bath physicist, “could completely revolutionize telecommunications.”
The reason for the excitement is that, in principle at least, sending light through air rather than through glass could greatly increase the efficiency and capacity of today’s high-speed telecom networks. These new materials, called photonic crystal fibers, should “leak” less light and carry more intense light pulses without distortion, reducing the need to constantly boost a signal-an expensive chore in today’s optical networks. Photonic crystal fibers should be able to convey much more information along fiber-optic networks while lowering installation and maintenance costs. They will be to existing fibers as a 10-lane freeway is to a country lane. Not only will they take more traffic, but the journey will be smoother and there will be less need for refueling.
It is still early in the development of this new generation of optical fibers. Even the most advanced of the new materials remain several years from widespread commercial use. But with so much at stake-optical telecommunications is a multibillion-dollar business-several industrial labs, including Corning and a handful of startups, are in hot pursuit of their own versions of photonic fibers. While it’s too soon to predict which will prevail, rival approaches developed at the University of Bath and at MIT are already competing head-to-head to become the optical fiber of tomorrow.
These efforts may bear fruit just in time for the telecommunications industry. The huge expansion of long-distance optical data transmission in recent years, fed by the growth of the Internet and its bandwidth-hogging applications, has led researchers to find ways to shoot more light and more complex signals through optical fibers (see “Wavelength Division Multiplexing,” TR March/April 1999). But many experts believe that in the coming decades it will become impossible to squeeze any more performance out of the current generation of glass fibers. Although it’s difficult to predict exactly when the roadblock will be reached, Jim West, a scientist at Corning’s research laboratories in New York, definitely believes “we’ll run into those limits.” And that’s when the next generation of fiber optics will become crucial in feeding the world’s apparently endless appetite for bandwidth.
Although photonic fibers are a next-generation technology in 2001, the history of conveying voice data using light extends back more than a century. After inventing the telephone in 1876, Alexander Graham Bell didn’t rest on his laurels. In 1880 he showed that light, rather than electricity, can carry a person’s words to a distant ear. Bell’s “photophone” used vibrating mirrors to transmit sound via sunlight. But it was an idea long before its time. Sending electrical signals down copper cables proved much more reliable, and the photophone was largely forgotten as telephone lines enmeshed the world.
After eight decades of the supremacy of copper wire, the invention of the ruby laser in 1960 put light back on the communications agenda. Here was a source bright enough to really put light to work. Just as the transistor ushered in the age of microelectronics, the laser sparked the age of photonics. In 1970 Corning proudly announced that it had sent a laser beam down a glass fiber and recovered as much as one percent of the light at the other end, a kilometer away (today’s glass fibers are so efficient that 80 percent of the light will survive that distance). By the 1980s, telephone companies began replacing copper cables with optical fibers.
An optical fiber can carry thousands of times more data than a copper cable: in principle, a single fiber can transmit up to 25 trillion bits per second. That’s enough capacity to carry all the telephone conversations taking place at any instant in the United States-with room to spare. Small wonder that the worldwide web of information technology is being woven from light-bearing glass.
In a conventional optical fiber, light is confined in a silica inner rod by a “cladding” of glass with a slightly different composition than that of the core. Typically, small amounts of germanium or phosphorus are added to the core (a process called “doping”), giving it a different refractive index from the cladding. Light striking the interface between core and cladding is reflected, so the signal bounces back and forth and remains within the core. Information is encoded in a series of pulses from electronically controlled lasers and fired down the fiber to a photodetector at the other end, which converts the signal back into electrical form for processing in a telephone, computer or routing device.
Sounds great. So, where’s the catch? It’s a matter of limits. As communications networks get bigger, busier and more ambitious, the drawbacks of conventional glass fibers are becoming evident, and existing optical-fiber networks will eventually be unable to cope. One factor that limits performance is the fading of the light signal over distance. A certain amount of the light is “scattered”-impurities in the silica disrupt the transmission of some of the signal-as it travels through the glass core; other light simply escapes from the fiber altogether, because the interface between glass core and cladding is not a perfect mirror.
Unremedied, these losses would cripple long-distance fiber-optic communications: eighty percent transmission over a kilometer would leave less than a ghost of a signal at the far end of a transatlantic cable. The answer is to amplify the light every 70 kilometers or so. But amplifiers are expensive, and they require their own power sources (see “5 Patents to Watch: Booster Shots”). Each amplifier typically adds a million dollars to the price of a long-distance transmission line. For a cable thousands of kilometers long, that begins to add up to real money. And when an amplifier breaks down mid-Atlantic, there’s no option but to send out a ship to dredge up the cable. “It costs a fortune to fix them at the bottom of the ocean,” says Bath’s Russell.
This daunting economic reality is the spur for developing the new generation of fibers. Cambridge, MA-based OmniGuide Communications, founded last year by several MIT professors, claims its new fibers will be able to squeeze losses so low there would be no need for any amplification. What’s more, the company says, the usable bandwidth will be substantially larger than in existing optical fibers. The trick is to strip out the fiber’s glass core and replace it with-well, nothing at all.
It sounds so obvious. Light travels through air with little scattering. So why not just send laser light down a hollow glass tube? The answer lies in physics. To achieve the internal reflection necessary to keep light confined in the center of a conventional optical fiber, the cladding has to have a lower refractive index than the inner medium. But all known materials have a higher refractive index than air. So the conventional arrangement doesn’t work in making a hollow fiber.
Which means an unconventional approach is needed. Enter photonic crystal fibers. Researchers worldwide are busy making materials that act as “light insulators,” which are impassable to light just as most plastics are impassable to electrical currents. In the jargon of physics, these light insulators have a “photonic band gap” corresponding to specific wavelengths of light; those wavelengths simply cannot enter the material. If made correctly, these materials-unlike the cladding in glass fibers-should permit virtually no light to escape from an empty core wrapped in them.
Of course, many substances will stop light from passing through; but this is generally because the materials simply absorb the light rather than reflecting it. And while you might think of metallic mirrors-silvered glass-as good light reflectors, the truth is that they are not nearly reflective enough to work in fiber optics; they absorb and dissipate a small but significant part of an incoming beam. A light signal traveling down a silver-lined glass tube would travel only a short distance before dispersing entirely. Photonic-band-gap materials, on the other hand, block all photons of particular wavelengths; the oncoming light is reflected almost perfectly. In other words, they are just the thing for confining light inside a hollow tube.
In 1998, Yoel Fink, then an MIT graduate student, fabricated a “perfect mirror” out of a photonic-band-gap material. Others had previously made specialized mirrors from thin layers of dielectric materials (materials that contain electrically charged particles but have insulating properties). These mirrors have photonic band gaps, and can be extremely efficient reflectors, but they have a major flaw: they work only with light striking absolutely face-on, limiting their use to specialized applications. Fink figured out how to make a version of a dielectric mirror that reflects light coming at it from all angles, as the material would have to in the core of a fiber-optic thread.
Once you have such a mirror, seeing the commercial potential is (for photonics researchers, at least) obvious. Fink and a pair of his MIT professors, physicist John Joannopoulos and materials scientist Edwin Thomas, along with Uri Kolodny, cofounded OmniGuide. The company’s goal is to use the perfect mirror as cladding for an optical fiber. Imagine taking a flat mirror and bending it around the inside of a tube, and you have a crude picture of an OmniGuide fiber.
So just how small are the losses of light in such a next-generation fiber? Because the company is still in its early stages, the founders are keeping that information close to their chests. “All I am free to say at this stage,” says Joannopoulos, “is that with a hollow-tube OmniGuide [fiber] we could in principle achieve losses less than optical fiber.” But for a telecom industry looking to push more and more light through optical networks-and eventually facing the limits of current-generation fibers-even such carefully worded pronouncements are tantalizing.
The company is developing a series of fiber products based on the OmniGuide concept. These fibers are, in theory, far more efficient in transmitting light than a standard optical fiber. Indeed, they should be able to overcome the current limitations of glass fibers, achieving, among other things, less signal loss as the light travels down the fiber. Such heightened performance is possible, says Fink, now an assistant professor of materials science at MIT, “because we can achieve an unrivaled degree of confinement.”
The OmniGuide fibers should be able to convey much more intense signals than normal optical fibers. High-intensity light traveling in glass fibers suffers from distortions that can disrupt the transmission of signals at different wavelengths, causing cross talk between channels unless they are widely separated in frequency. This effect limits the number of different wavelengths you can stuff in a conventional glass fiber, and also how bright they can be. Because signals in air don’t suffer these effects, Fink explains, the OmniGuide fiber can convey signals at higher powers, with channels spaced closer together. That is great news for telecom companies, since stronger signals travel farther before losses begin to compromise them, and closer channels mean that more data can be packed within a given wavelength range.
The MIT approach, however, is only one way to make a photonic fiber. Other researchers have produced photonic-band-gap materials that, in cross section, are like a honeycomb in which the holes form structures that refuse entry to light of certain wavelengths. These kinds of photonic crystals, first made in the late 1980s, also nearly totally block out light. The glass fibers made at Bath, for example, are penetrated by an orderly array of holes running parallel to the thread along its entire length; at the center is an empty core in which light can be nearly perfectly confined. To give some indication of the precision involved in making the fibers, if the long, parallel holes were the diameter of the Chunnel connecting England and France, the experimental fibers made at Bath would reach Jupiter. How does one drill such perfect tunnels through a glass strand thinner than a human hair?
Fortunately, the holes don’t have to be drilled at all. They are ingeniously constructed by drawing the glass fibers from a bundle of hollow capillary tubes. The tubes are packed together in a hexagonal array a few centimeters in width, and the bundle is heated to soften the glass. As the array is pulled out into a fine fiber, its cross section gets shrunk by a factor of a thousand or so but remains laced with holes.
Initially, the Bath physicists made a light-conducting channel at the core of the fiber by substituting a solid glass rod for the central glass capillary. But still better than carrying the light in a solid core would be to send it through a hollow core-through air, with the very low losses and absence of distortion that entails. In collaboration with Douglas Allan, a researcher at Corning, the Bath team succeeded in achieving light confinement in a hollow-core photonic crystal fiber in 1999. Recently they have formed optical fibers many meters long out of their novel materials.
Taking on existing optical fibers will be a tall order. Conventional glass fibers have been optimized over several decades and are made using well-entrenched technology. In contrast, the new photonic fibers represent a manufacturing unknown. For one thing, their structure must be exact. “The existing [fabrication] systems are simply not up to it,” admits Russell.
Still, companies are lining up to meet the commercialization challenges. Fink says OmniGuide is working on a series of products based on different-length fibers. Projects include the development of active fiber-based devices for optical switching, as well as the development of fibers for light transmission over 10 to 100 meters, which could be useful for tasks such as connecting servers over short distances. Long-haul fibers for telecom networks will have the biggest impact, says Fink, but these “will take a little time.”
Researchers from the Bath group have launched their own spinoff, BlazePhotonics, and have secured funding from venture capital firms in the United Kingdom and United States. In Denmark a company called Crystal Fibre, started by scientists at the Technical University of Denmark in Lyngby who were early collaborators with the Bath group, is making photonic fibers with a solid glass core. While its initial products might serve such purposes as confining light in high-precision lasers, no one is losing sight of the big prize. “Telecommunications is definitely the medium-term target,” says CEO Michael Kjaer.
Like the founders of Denmark’s Crystal Fibre, scientists at Corning have worked closely with the Bath researchers in the past, but they are now racing to the marketplace on their own. Jim West reports the company can now make photonic fibers up to a hundred meters long. But he reserves judgment about whether the new materials will eventually transform the information superhighway. Conventional optical fibers, he points out, are a difficult act to top. “It’s only when you start working with the state-of-the-art versions that you realize how remarkable they are.”
Although sending light through air may solve many of the limitations of today’s fibers, it poses its own problems. For one thing, the composition of air is not uniform; as a result, light may be transmitted differently in different parts of the world. “Air in the U.K. is very different from air in the Sahara,” explains West.
“It’s a fascinating technology,” says West of the new generation of photonic crystal fibers, “but there is a long way to go.”
Still, if these new materials eventually fulfill their potential of transforming long-distance transmission in the telecommunications industry, it will be a journey well worth taking.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Driving companywide efficiencies with AI
Advanced AI and ML capabilities revolutionize how administrative and operations tasks are done.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.