Hancock, N.H. - Photonic-bandgap materials research is turning a corner. Some designs that are suitable for manufacturing processes surfaced at the recent Optical Fiber Conference (OFC), introducing a level of control and refinement for optical-network component designers.
Photonic-bandgap materials, which are the optical equivalent of semiconductors, offer a degree of optical control not available with conventional materials used in such components as waveguides, mirrors and beam splitters.
One particularly promising manufacturing approach is a process called autocloning, developed jointly by Tohoku University (Sendai, Japan) and Photonic Lattice Inc. Here, a corrugated stack resembling cardboard is assembled from two materials with different refractive indexes: amorphous silicon and silicon dioxide, which are common materials used in semiconductor manufacturing. The tricky part is getting an accurate and reproducible pattern at 500-nanometer dimensions. But the small dimensions are critical, since the variation in refractive index has to be on the order of one-half the wavelength of the light with which the material is interacting.
The researchers reported that they can use their process to achieve that accuracy in a manufacturing line and that the approach has been proved out in practice. At the conference, they described four optical components that have been designed with autocloning.
The first is a simple polarization beam splitter built from the corrugated stack of amorphous silicon and silicon dioxide. When striking the stack from the top, light in one polarization direction-the TM mode-passes through unimpeded, but light in the TE mode is totally reflected. Since the material is operating on the light in a quantum mode, the operation has extremely high efficiency, a hallmark of optical-bandgap devices.
A more complicated device is an optical isolator designed to work with a diode laser. In this design, two of the photonic-crystals sandwich a Faraday rotator. The device features an ultrathin design and the ability to handle high optical power, since the photonic crystals do not absorb any of the light that passes through them.
Another device that has been assembled and characterized is a polarization beam combiner. In that design, a photonic crystal is placed between two lenses. On one side, a two-port fiber carries one input signal; on the other, a fiber carries a single-input port. The two input beams are focused on the photonic crystal, where they are multiplexed together. A reflected beam that carries both signals is reflected back into the output port. One advantage of using a photonic crystal is that it does not need to be polished. In a conventional design, a highly polished birefringent crystal is required to perform the reflection and mixing, driving up the cost. Once built, the photonic crystal needs no polishing at all to achieve a 0.6 insertion loss.
On the drawing board are some more complex designs: a novel three-port circulator and a polarization state monitor. The circulator has been assembled.
Big nano effort
A team at the California Institute of Technology (Pasadena, Calif.) is working with compound semiconductor processes to create a nanoscale photonic-circuit technology. With conventional optics, photonic circuits can no longer shrink to smaller dimensions, because VLSI technology has moved beyond the wavelength of light in terms of feature dimensions. One way around this is to fabricate subwavelength devices, which operate below the wavelength of light.
There have been a lot of research successes in the field of subwavelength design, but thus far researchers have not been able to produce a manufacturable and integrated photonic circuit. That is the hope of the Caltech group.
By fabricating photonic-bandgap microcavity structures in the indium-gallium-arsenic-phosphide (InGaAsP) system, researchers can use the new approach to concentrate optical signals into submicron "nano-optic" cavities. Because those features are built in a photonic-bandgap material, the photons at certain wavelengths are forbidden to travel through the material, creating a perfect container. Any other material would allow some leakage, reducing the possibility of efficient containment at subwavelength scales.
The basic system is made of alternating slabs of photonic crystals and high-index-of-refraction waveguide layers, which act as interconnects. The photonic crystal slabs are built by etching an array of submicron holes in a thin substrate, creating a periodic variation in index of refraction, which acts on photons in the same way that a periodic silicon lattice acts on electrons.
The approach has made it possible for researchers to build lasers, modulators, add/drop filters, polarizers and detectors. All of those devices have been built and characterized in the system. The technique makes it possible to reduce the feature size of photonic circuits, integrate them in large numbers on a single substrate and produce them on a manufacturing line in a way that is very similar to silicon integrated circuit technology.
Photonic-bandgap fibers are one version of the technology that has made it to the commercialization stage. Their construction is simple: concentric rings of air holes are introduced as the fiber is drawn so that the core is surrounded by a periodically varying photonic lattice. In normal fibers, the core and outside cladding layers have different refractive indexes, creating total internal reflection. In effect, the photons bounce back into the core and are thus transmitted along the fiber. The periodic array of air holes, on the other hand, produces a bandgap at the wavelength of the transmitted light so that the photons are unable to travel outside the core.
Tale of three waveguides
A group at OFS Laboratories (Murray Hill, N.J.) looking into ways to find an optimal geometry for fiber communications needs reported on comparison studies of three types of photonic-bandgap waveguide structures. A planar waveguide was compared with a fiber that had concentric rings of varying refractive index and one with an array of holes running along the fiber. The study was the experimental side of a project at the company to develop analytical tools that can predict the spectral properties of different photonic-crystal structures. The objective is to provide photonic-crystal device designers with better design tools that can be used to optimize performance.
Researchers at Brown University (Providence, R.I.) described their approach to a tunable photonic-bandgap technology based on liquid-crystal polymers. Certain holographic liquid-crystal systems have the periodic structure needed for photonic-bandgap systems. In particular, the ability of the liquid crystals to reorient themselves means that the lattice properties can be changed in real-time, opening up possibilities for optical-component design.
The process uses a liquid crystal mixed with a liquid-polymer solution. A photosensitive agent is added, and the mixture is prepared in a darkroom. That let the researchers use 3-D holography techniques to define regular patterns in the mixture. Light exposure causes the liquid crystal to separate out into a regular array of droplets and hardens the polymer, producing a tunable photonic lattice.
The Tohoku University/Photonic Lattice team also reported a photonic waveguide built with their autocloning process. The alternating corrugated layers were made from a tantalum-oxygen compound (Ta2O5) and silicon dioxide. The core waveguide region was formed from straight alternating layers, and the cladding region on both sides that confine the photons had the characteristic wave configuration. That resulted in a record low-loss transmission of 0.56 dB/mm.
One advantage of this configuration is the ease with which the output can be coupled to a fiber. The reason is that the waveguide structure shapes the optical beam into a round configuration, making it easy to couple the output to a fiber without much loss. This is always a difficult hurdle for photonic system design, since fiber and waveguides must be perfectly aligned to keep the coupling losses within acceptable bounds.
The photonic chip was fabricated by defining a series of wavy grooves in a silica substrate with electron-beam methods. The pattern was dry-etched and then alternating layers deposited using sputter deposition. The basic wavy pattern is reliably repeated in the process. The thickness of the layers was increased toward the middle of the stack by running each deposition step longer. The layers were then thinned out toward the top. The researchers could thus tune the refractive index of the cladding regions to maximize confinement.