1958: Tunnel Diode Promises a High-Speed Semiconductor Switch

Leo Esaki's novel device is an example of many celebrated semiconductor breakthroughs that do not sustain their early promise as they are overtaken by competing technologies.

 

 

 

At a June 1958 conference in Brussels, Leo Esaki reported on a new diode he had developed at Sony in which the current decreased as the voltage increased, effectively exhibiting "negative resistance." The Esaki or tunnel diode, named for the quantum-mechanical tunneling effect it exploited, offered fast switching speed with very low power consumption. Although he had discouraged a young Robert Noyce from pursing a similar idea recorded in his lab note book in August 1956, conference chairman William Shockley hailed its development and predicted its extensive use in computers. In 1973 Esaki, by then with IBM, shared the Nobel Prize for his pioneering work on electron tunneling in solids. While the Esaki diode found useful niche applications, it never achieved its early promise. As a discrete device, it could not compete with integrated circuits that were fast enough and offered significant advantages in cost, reliability, and packing density.

Many other semiconductor developments have been hailed with great expectations but were later eclipsed by lower-cost solutions. They include magnetic-bubble memories and charge-coupled devices (CCDs). Invented at Bell Labs and widely used today as image sensors, CCDs were initially targeted by Fairchild, Intel and others for high-density serial-memory applications. Another Bell Labs invention, magnetic bubbles used semiconductor-processing techniques to create magnetic domains capable of non-volatile data storage in garnet material. In the late 1970s Intel, Rockwell, and Texas Instruments, plus Fujitsu and Hitachi in Japan, made bubble memories with capacities up to 1 megabit. Both technologies found early applications but were unable to keep pace with cost and density advances in rotating-disk memories.