A class of integrated devices with high electronic circuit complexity and a multiplicity of optical inputs and outputs can appropriately be called optoelectronic integrated systems. Although free-space propagation implies that the complete system extend physically beyond the integrated device, the term is suitable if a primary system-level processing task is performed on the optoelectronic substrate itself. Examples of such devices are the VLSI-based spatial light modulator, the optical-in/optical-out silicon retina device, and monolithic, optically interconnected, digital processor arrays. These devices must match systems requirements; whereas conventional (VLSI) integrated systems must comply with constraints on speed, power, and area, optoelectronic integrated devices must satisfy additional requirements founded in optical physics and established by the imaging configuration in the overall system. The interplay of device constraints with characteristics of various technologies for electronic-optical transduction will be examined in different regimes of granularity, concurrency, and speed. The holistic consideration of optical devices as parts of systems leads to a number of non-obvious conclusions. For example, in optoelectronic systems that are optimal in a certain sense, light modulators need not always dissipate less power than one or even 100 transistors, nor need they always switch in less than 1 ns, or even 100 ns. Examination of the liberal, but nonetheless realistic, constraints arising from this analysis will reveal that a number of existing light modulator technologies can offer satisfactory performance in interesting and important applications.

Date Published: 1 December 1991
PDF: 10 pages
Proc. SPIE 1562, Devices for Optical Processing, (1 December 1991); doi: 10.1117/12.50778

Published in SPIE Proceedings Vol. 1562:
Devices for Optical Processing
Debra M. Gookin, Editor(s)