Figure 1. The micro-rod is designed to neatly abut with an incoming fiber, allowing the exiting beam to be expanded and collimated with low loss (or, conversely, focused and coupled into the fiber).
Researchers at Digital Optics Corporation (Charlotte, NC) have come up with a way to make optics and fibers work together. They have developed a new format for both diffractive and refractive optical elements -- the microrod -- that is the same width as the optical fiber itself. Not only should this allow fiber connectors to be significantly smaller than they have been in the past, but it should also make passive alignment easy. The new optics also have the advantage of being compatible with silicon processing technology. Digital Optics' engineers have exploited this, and found a way to pattern thousands of elements on a single wafer.
Researchers say they developed the microrod concept because existing optics for fiber coupling, such as gradient index rods and ball lenses, are both much larger than the fibers they "serve" and require that the passive alignment take place using two planes: one for the fiber and one for the optic. The basic concept for a diffractive microrod is incredibly simple (see Figure 1). The rod is made of fused silica and, unlike a fiber, is not structured as core and cladding.1,2 This alone provides an advantage in that it allows the beam exiting the fiber to expand freely without having to propagate through air. The light can then be collimated using a diffractive lens patterned onto the rod end.
Figure 2. Passive alignment of 2D and 3D arrays of micro-rods and fibers is possible using v-grooves micromachined into silicon.
The idea is to make the microrod a fixed element in a passively aligned fiber connector, such as one using v-grooves (shown in Figure 2). Then, all the user has to do is push the fiber in with a little index-matching fluid or epoxy. The fiber and the microrod will then form a single optical path, allowing light to be coupled in and out of the fiber easily.
Though the system is very elegant, fabrication -- at first glance, anyway -- is a messier problem. Manufacturing the individual rods is not particularly difficult, but patterning the end of each one can be troublesome (at least if you want to do it cheaply). Digital Optics decided the easiest way to manufacture a large number was to make an entire wafer filled with microrods standing on one end. This way, the other end was exposed and could be etched using standard silicon processes. However, these microrod wafers proved to be difficult in that they turned out a little differently each time. The inconsistency made it impossible to correctly pattern different batches using a single lithographic mask (or set of masks).
Figure 3. A 0.5-in. diameter substrate can carry more than 8000 tightly-packed micro-rods standing on their ends. Because the packing varies from wafer to wafer, a vision system is used to map the centers of each rod. This map is used to create the lithographic mask used for patterning.
Figure 4. The result of the map-mask-pattern process: the diffractive lenses are correctly positioned on the micro-rods. These rods are 125-µm in diameter.
Digital Optics' solution was to use a vision system to map the microrod wafer, and to use this map to produce the needed masks (see Figure 3). The result, shown in Figure 4, was a set of diffractive microrods with etched collimating lenses positioned with high accuracy. This technique can also be used with refractive lenses, which are created using etching and photoresist-reflowing to produce a smoothly curved shape.
1. Eric G. Johnson, Charles Koehler, Thomas J. Suleski, Jared Stack and Michael Feldman, Diffractive microrods for fiber optic applications, Proc. SPIE 3633, 1999.
2. H. Han, J. D. Stack, J. Mathews, C. S. Koehler, E. Johnson, A. D. Kathman, Integration of Silicon Bench with Micro Optics, Proc. SPIE 3631, 1999.