26 Mar 2004
A new type of optical fibre that rewrites the rules for guiding light is now coming of age. Rob van den Berg talks to three start-ups that are busy exploring the commercial potential and applications of photonic crystal fibre.
From Opto & Laser Europe April 2004
In the last five years three new start-up companies have been founded to commercialize the technology and the fibre giants Corning and Mitsubishi Cable are also active in the field. The focus of all this attention is photonic crystal fibre (PCF), a waveguide that resembles standard silica fibre but with one very important difference - its cladding contains an array of micron-sized holes running along its length. The big attraction is that by varying the size and location of the holes, the fibre's mode shape, nonlinearity, dispersion and birefringence can reach values that are not achievable in conventional fibres. PCFs come in two varieties: solid core fibres (also known as index-guiding fibres) and hollow core (bandgap guiding).
One of the first pioneers of PCF is Philip Russell of the University of Bath, UK, who in 1995 succeeded in making fibres with a photonic crystal structure, and then in 1999 demonstrated the first hollow-core PCFs that guided light in an air core. He is now chief technology officer of BlazePhotonics, a company in Bath which was established to commercialize PCF and is now selling dozens of different types for a wide range of uses.
One application is the generation of a supercontinuum (broadband "white light") that can be guided down the fibre. At the start of the year Blaze launched a new PCF that is specially optimized for converting the output of Nd3+ microchip lasers into such an optical supercontinuum. The supercontinuum can span an octave or more in wavelength, providing a broadband output in a singlemode fibre with a spectral brightness more than 10,000 times that of the Sun.
Blaze says this combination of microchip laser with PCF provides a high-performance alternative to conventional broadband sources such as lamps and superluminescent LEDs. Its applications include testing the spectral response of optical communications equipment, optical coherence tomography (OCT), multi-photon spectromicroscopy and chemical sensing, for which a high brightness over a large spectral range is paramount.
"PCFs have now clearly moved from being a scientific curiosity to becoming a technology that is the preferred solution in a rapidly growing range of applications," said Hendrik Sabert, vice-president of R&D at Blaze.
He added: "In hollow-core PCF we have now demonstrated that more than 99% of the light can propagate inside the hollow core and cladding holes, and theoretical results show that this may be improved to more than 99.8%. Moreover, we can now fabricate these fibres for the entire wavelength range from the visible to the near-infrared, say, from 440 to 2000 nm."
"Although there is still much work to be done we are confident that hollow-core PCF can deliver much higher power levels than conventional fibre," said Sabert. "With nonlinearity and material dispersion virtually absent, PCF can, for example, guide 100 fs pulses of some hundreds of kilowatts of peak power at 800 nm. This is ideal for the delivery of ultra-short pulses from Ti:sapphire or Nd3+:glass lasers."
The Bath-based firm is not the only enterprise trying to commercialize PCF. In 1998, Yoel Fink - then a graduate student at the Massachusetts Institute of Technology (MIT) - discovered a way to fabricate a perfect dielectric mirror. His mirrors use dielectric layers of varying thicknesses to reflect light at a range of wavelengths coming from all angles of incidence.
Fink realized that such an omnidirectional "perfect mirror" could be used as a cladding for an optical fibre that potentially has a very low transmission loss. Excited by this prospect, he established OmniGuide Communications of Cambridge, Massachusetts, to commercialize his work. The main breakthrough came at the end of 2002 when Fink and his MIT collaborators showed that it is possible to make a hollow-core fibre for guiding 10.6 µm infrared light from a CO2 laser - something that simply isn't possible with standard silica fibre.
According to OmniGuide's business development and product manager Gregor Dellemann there is a big potential market for fibre delivery of CO2 laser light, especially in the medical and materials processing fields. One potential application is the treatment of patients suffering from lung cancer: "The CO2 laser is a very precise cutting tool. With the help of our fibre, doctors would be able to get the laser light into the upper part of a patient's lung to selectively remove cancerous tissue that is blocking the larger airways," explained Dellemann. "Because of its precision there would be substantially reduced risk of penetrating the lung tissue itself while cutting away the tumour."
A third start-up that is busy in the field is Crystal Fibre of Lyngby in Denmark, which claims to have been the first commercial supplier of PCF. Its sales manager Rene Kristiansen now sees a growing demand for solid-core PCF, especially for double-clad fibres which convert light from multimode diodes into a high-quality singlemode beam.
"It is a way of converting cheap photons into high-quality photons," explained Kristiansen. "Traditionally, double-clad fibres have been made as step-index fibres of different materials - most typically with a polymer outer cladding. With the PCF technology, the fibres can be made of all silica."
As for its most recent success, at the Optical Fiber Conference (OFC) in Los Angeles in February, a partnership between Crystal Fibre and the California Institute of Technology demonstrated an air-silica Bragg fibre. The design comprises three concentric air-silica rings that confine the light in a hollow core. According to Crystal Fibre, the fibre has the potential to break the current loss limits for guiding singlemode light without polarization mode dispersion - a troublesome effect that degrades optical communication systems.
In fact, in spite of all the current activity, the real moneyspinner for PCF will probably be in telecoms, simply because of the lengths of fibre involved. However, to be successful the loss level of PCF must be reduced to less than that of conventional fibre. Although this is a tough challenge, steady progress is being made. Index-guiding PCFs with an attenuation of less than 0.3 dB/km at 1550 nm have already been demonstrated by NTT of Japan at last year's European Conference on Optical Communication. But since most of the light travels in the solid core, loss is ultimately limited to the loss of the bulk material (0.2 dB/km), which is the same situation as for standard silica fibre.
As hollow-core fibres do not suffer from the same limitations there is optimism that they may provide a lower loss solution, but there is still much work to be done. Currently, the lowest light loss for a hollow core fibre - 1.7 dB/km, presented by BlazePhotonics at OFC 2004 - is still an order of magnitude higher than that of the best conventional fibres.
"It is quite possible that hollow-core fibres will ultimately achieve losses well below that of conventional fibres," said Sabert. "But at this point, it is certainly fair to say that no-one knows where the limits really are."
How PCF is made
For both BlazePhotonics and Crystal Fibre the starting point for making PCF is to create a macroscopic glass preform with the same structure as that required of the final fibre, albeit on a larger scale. One way to do this is to stack hundreds of silica capillary tubes and rods by hand into a structure with the appropriate pattern of holes. The hollow core is made by replacing one or more of the capillaries by a hollow tube. This preform is then introduced into the furnace of a fibre drawing tower.
At the temperature at which silica softens - about 2000 °C - the preform is fused together and drawn down to a size of 1-10 mm with air holes of 0.05-0.5 mm. After an additional sleeve tube has been added this "cane" is drawn down to its final dimensions in a second draw step. Typical hole sizes in the final fibre range from 0.5 to 5 µm, with hole spacing in roughly the same range.
OmniGuide's dielectric fibres are made in a similar way. The fibre preforms are produced by evaporating a chalcogenide glass (As2Se3 ) onto a polymer sheet (polyethersulphone), wrapping the film around a glass mandrel and consolidating the layers under heat. After the glass mandrel has been etched out the pre-form is ready to be drawn.