Abstract

Guided-wave optical interconnections offer many potential benefits for high performance integrated electronic systems¹. These include higher packing density, lower noise, lower propagation delay, and the easing of many problems encountered in the design of high-frequency on-and off-module connections. Their major impediment to practical use, apart from being a new, unproven technology with little reliability data, has been the requirement to perform two electrical-optical conversions, one each at the source and receiver ends of the optical line. Most previously proposed schemes for guided-wave interconnects have required one laser or LED per signal. In effect, a multichip carrier, monolithic chip, or fully integrated wafer having thousands of interconnects would necessarily have thousands of lasers as well. For direct laser modulation, the output drivers of a given IC must supply sufficient current to achieve and surpass threshold to drive the line with the required temporal edge. For external modulation, the output drivers must supply a sufficient voltage, usually half-wave, to provide sufficient contrast ratio, again with the required temporal edge.

© 1991 Optical Society of America