HENDRIK SABERT
Of all the applications that optical physicists are finding for photonic-crystal materials, optical fiber is probably the most advanced. Photonic-crystal fiber (PCF) is a new breed of optical waveguide with properties very different from those of standard fibers. These new fibers come in two basic varieties—index guiding and bandgap guiding. Due to the many degrees of freedom in the cross-sectional index profile, index-guiding PCFs can be designed with highly unusual dispersion, nonlinear, and birefringent properties, but ultimately most of the light in these fibers still travels in the glass. Bandgap-guiding or hollow-core fibers are arguably the more revolutionary aspect of PCF technology because in this type of PCF light can be trapped in a central hollow core by generating a photonic bandgap in the fiber cladding.
The advantage of using a hollow core—instead of a conventional core made of doped high-purity silica—is that the performance of the final fiber is no longer limited by the properties of the core material. In conventional fibers, the damage threshold, attenuation, nonlinear response, and group-velocity dispersion are all dominated by the corresponding values for bulk silica. Properly designed, hollow-core fiber can guide more than 99% of the light outside the glass, greatly reducing the impact of the bulk properties of the fiber material on the optical properties and performance of the fiber. Hollow-core PCFs (HC-PCFs) can therefore be expected to outperform their conventional counterparts in many important applications.
Unlike conventional fibers, HC-PCFs do not guide light by total internal reflection. Instead, the underlying operating principle is closer in spirit to a multilayer mirror. Multilayer mirrors work by in-phase reflection of light from a multitude of interfaces. In HC-PCF, a two-dimensional (2-D) array of tiny air holes running down the fiber length acts as the functional equivalent of these interfaces. To trap light in the core, the holes surrounding the hollow core must be arranged very uniformly in a regular lattice and they must be closely spaced so that they almost touch. The resulting cladding resembles a honeycomb in cross section, consisting of a web of fine strands of silica, sometimes as thin as 100 nm across. This lattice can act as an almost perfect reflector, keeping the light in the core, but its operation is limited to a specific range of propagation constants. As a result, HC-PCFs have a spectral response very different from conventional fibers, guiding light over a finite range of frequencies, typically around 20% of the central frequency. Despite this, the mode-field shape can be surprisingly similar to that of a conventional single-mode fiber (see Fig. 1).
Fabrication
Hollow-core PCFs can be fabricated using standard optical-fiber drawing equipment. First, a preform is built by stacking hundreds of thin-walled capillaries. This preform is then jacketed, drawn down to the final diameter and coated with a polymer to form a fiber with dimensions and mechanical characteristics very similar to standard single-mode fiber. The fabrication process is now sufficiently well developed that essentially unlimited lengths of fiber with uniform optical properties can be made—at least for HC-PCFs made from fused silica glass.
Although the guiding bandwidth is largely determined by the photonic bandgap of the cladding, core size and shape as well as minute variations in the distribution of solid material around the hollow core can significantly alter the optical properties of the fiber. It is not surprising, therefore, that a large part of the current research effort is directed toward improving the fiber design and related fabrication techniques.
Attenuation
In a practical example of a HC-PCF designed to operate in the telecom wavelength range, the low-loss window spans some 150 nm, centered on 1570 nm (see Fig. 2). The attenuation rises rapidly outside of this range. The minimum attenuation is 1.7 dB/km, which is the lowest value demonstrated for any hollow core waveguide to date (see Fig. 3). In this particular example there are clearly defined regions of higher attenuation within the low-loss window. These are due to "surface" modes (resonances located on, or close to, the glass-air interface of the core), which become degenerate with the fundamental mode at certain wavelengths. For wavelengths at which such a degeneracy occurs, the amount of light that interacts with the surfaces increases sharply, leading not only to an increase in attenuation but also changes to the dispersion characteristics of the waveguide. For practical applications such features may not be desirable. It is likely, however, that they can be removed by careful design of the core and surrounding cladding.
Beam delivery
Although HC-PCFs are not yet ready to challenge the supremacy of conventional optical fibers in the field of long-haul telecommunications, they are able to outperform them in several important applications, most significantly, perhaps, in laser-beam delivery. One important advantage of HC-PCFs over conventional fibers is their increased damage threshold. Because so little of the light actually travels in the glass, the power-carrying capacity of HC-PCFs is potentially far superior to that of conventional fibers.
Another difference is their low level of optical nonlinearity, again a direct consequence of the small overlap of the mode with the glass—typically, the gas in the core has a nonlinear refractive index roughly 1000 times less than that of solid silica, reducing the nonlinear response of the fiber by up to three orders of magnitude relative to that of a comparable conventional fiber. This makes it possible to transmit light at very high power through a hollow-core fiber without spectral distortion, either continuous-wave or as a train of short pulses. In practice, fibers can be designed such that either the nonlinearity of the gas in the core, or that of the glass dominates the overall fiber nonlinearity. Furthermore, the fiber can be filled with gases other than air, giving unprecedented control over the nonlinear properties of the fiber as a whole.
For pulses significantly shorter than 1 ps, new constraints become important. The intrinsic bandwidth of the pulses themselves starts to become comparable to the width of low-loss windows in the HC-PCF. In addition, the group-velocity dispersion in HC-PCF means that pulses shorter than 1 ps will be significantly dispersed over fiber lengths of just a few meters. Crucially, however, the low nonlinearity of HC-PCF means that such dispersion is not accompanied by significant spectral distortion, even for pulses down to 100 fs at powers typical of modelocked laser oscillators.
In conventional fibers, the combined effects of nonlinearity and dispersion very rapidly tear such short pulses apart after just a few millimeters of propagation. The low level of nonlinearity in HC-PCF means that pulses can indeed be delivered through a few meters of HC-PCF, as long as the linear dispersion of the fiber is properly compensated, for example, by using a piece of glass to prechirp the pulses before they are coupled into the fiber. Another possibility is to deliver the pulses as solitons, balancing the linear dispersion with the small nonlinear response of the HC-PCF. Optical-fiber solitons have previously been observed in the 1500-nm wavelength band at relatively low powers, using conventional fibers, but HC-PCFs enable the delivery of higher-power pulses—up to a few megawatts peak power—over a far wider range of wavelengths.
Outlook
Future research will be aimed at further broadening and optimizing fiber designs, material properties, and fabrication techniques. Reducing loss is certainly one key objective. While 1.7 dB/km is an important milestone, it is quite possible that hollow-core fibers made from silica could eventually outperform even the best conventional fibers, where loss is concerned.
Another exciting possibility is that low-loss fibers can be made from relatively high-loss materials, given that very little of the light actually "sees" the glass. This is particularly interesting for longer IR wavelengths, where highly developed glasses are available, and could, for instance, lead to the development of high-power fibers for the 10.6-µm wavelength. Progress has also been made at the other end of the optical spectrum and recently the first HC-PCFs for visible and UV light have become commercially available. Despite the rapid advances that have been made since the first HC-PCF was reported in 1999, PCF technology is still at an early stage of development so hollow core PCFs will likely continue to improve dramatically.
Hendrik Sabert is chief technology officer at Blaze Photonics, University of Bath Campus, Bath, BA2 7AY England; e-mail: [email protected].