Abstract

The future prospects for both optical computing and photonic switching are clearly very dependent on advances in devices. This is especially true in the case of large scale applications requiring arrays of devices. Not only must the individual devices have good physical performance, they must also (i) operate at very low energies so that the array can be powered optically if required and have sufficiently low overall dissipation, (ii) be fabricatable in uniform arrays, and (iii) have sufficiently sophisticated functionality to allow efficient design of complex systems. Any one of these requirements suggests integration; taken together, we can see that a technology that offers straightforward integration of large numbers of flexible devices is essential for such array applications. Given that there are very few physical mechanisms that can offer sufficiently low operating energies for optical devices regardless of integration, we can see that this is a hard problem.

© 1989 Optical Society of America