azu_td_1349475_sip1_... - Arizona Campus Repository

11
architectures. 3,4 5 8 In all of these schemes, the optical interconnect generally
consists of a source, detector, and a channel which provides the actual
interconnection between the source and detector. Sources and modulators
such as laser diodes, LEDs, and spatial light modulators exist in array form.
Likewise, high speed detectors exist in array formats. It is therefore necessary
to develop interconnect channels that can exploit the connectivity of multiple
sources and detectors if the massive parallelism of optics is to be fully utilized.
An interconnection component that has the ability to exploit the parallel
nature of two dimensional sources and detectors is the microlens array.
Many types of refractive microlens arrays have been demonstrated for this
purpose. For example, Iga et aP proposed a method of making arrays of
distributed index lenses by diffusing dopants into a substrate through a mask.
These lenses where then employed in a concept called stacked planar optics 8
which utilized two dimensional arrays of microlenses stacked in series with
other two dimensional components to build compact, optically interconnected
systems. Another type of microlens array, introduced by Borrelli et al 9 , used a
photothermal technique to produce microlenses in photosensitive glass. The
glass was first exposed with UV radiation through a photomask and then
heated. The exposed regions swelled during heating to form spherical surfaces
in the glass. Refractive microlens arrays have also been produced using a
plasma-activated chemical vapor deposition (PCVD) process. 10 In this method,
holes, which are etched into a glass substrate, are covered and filled with a film
that is deposited onto the glass substrate by the PCVD process. The film layer
is then ground down to the surface of the glass plate to form lenses within the