A two-dimensional version of the fiber Bragg grating—a planar holographic optical processor (planar HOP)—should allow signals to be selectively routed by wavelength and temporally manipulated. According to inventor Tom Mossberg of LightSmyth Technology Inc. and the University of Oregon (both in Eugene), the new device should be robust, cost-effective, and fully integrated, due to the relative ease of replicating holograms and the fact that small defects are not critical to fabrication. Mossberg sees applications particularly in packet decoding and demultiplexing.
The planar HOP is based around three ideas.1 The first is geometric: given initial constraints, rays of light reflected by concentric circular reflectors will arrive at the same point regardless of the reflector diameter (see figure). The second is that, by layering many inefficient reflectors (high/low refractive-index interfaces) on top of each other, a highly efficient mirror can be built up. Finally, by choosing the correct period for these reflectors, only a narrow-band design wavelength will be reflected: all the others will destructively interfere. These three concepts describe a curved Bragg reflector. Importantly, it is a very thick Bragg reflector, which makes it very efficient.
The concept described by Mossberg is more complex than that, however. For instance, if a square pulse is reflected by the device, then the energy returned by the first index-change interface travels a much shorter optical path than that reflected by the outer rings. This spreads the pulse temporally. In addition, if the reflectivity of the rings is manipulated, the temporal shape of the pulse can also be controlled.
In communications systems, these properties can be used in several ways. First, different holograms can be recorded within the same planar waveguide. Each one can be tuned to reflect a different color of light, thus diverting light from a single input port into different wavelength channels (demultiplexing), or vice versa. Second, each hologram can be recorded to implement some optical coding or pulse manipulation, allowing multiple functions to be performed simultaneously. Third, just the fact that the inputs and outputs are separate means the devices can be used without the need for optical circulators. With fiber Bragg gratings, circulators are necessary because light has to be fed in and out via the same channel and then sent out in the right direction.
Mossberg says that one of the crucial aspects of the technology is the cost. In particular, he says, the cost to manufacture embossed structures (like those used in display holography) would make their price "unthinkably low." Even polymer-based devices, he says, would be inexpensive enough to "open entirely new strategies."
Another fabrication option—making the hologram using semiconductors—would also make it possible to integrate the optics and electronics on a single chip. In addition, given the right materials, the planar HOP could be used dynamically. Switchable holograms, generally fabricated using liquid crystals, have already been demonstrated for other applications. It is possible that some evolution of this technology could be used with the new waveguides—allowing, for instance, the pulse shaping performed by a particular device to be altered on the fly.
REFERENCE
- T. W. Mossberg, Opt. Lett. 26, 414 (April 1, 2001).