DEVELOPMENT OF THE INTEGRATED CIRCUIT (IC)

by R. J. Reiman, Historian

Robert Noyce, son of a Congregationalist minister, as a physics major at Grinnell College, Iowa, was introduced to solid-state physics by the school's physics professor, Grant Gale, who was also a friend of John Bardeen of Bell Labs, one of the inventors of the transistor. Gale asked Bardeen to send him a sample of the transistor in 1948. Noyce experimented with it at Grinnell and he decided to specialize in solid-state physics at graduate school.

He went to MIT, but found no courses in solid-state physics and the transistor was unknown to them. Noyce and Gale continued to compare notes, and when Noyce received his Ph.D., he met up with Swiss-born physics Jean Hoerni when both arrived in MountainView, California to work for Schockley Semiconductor Laboratory, a small company owned by William Schockley, the transistor's co-inventor. Noyce, Hoerni, and 6 other scientists and engineers left the Lab to found Fairchild Semiconductors and develop advances in silicon transistors. The state of the art was the SILICON MESA TRANSISTOR which had been invented by Bell Labs and had been used by Jack Kilby in his IC's at Texas Instruments. Hoerni came up with a solution to the problem of the MESA, which was susceptible to contamination, and whose connection wires tended to slip. Hoerni diffused the MESA into a wafer, i.e., chemically-embedded the transistor into a piece of silicon. The result was a completely flat transistor, which he coated with a thin layer of silicon dioxide which insulated and protected the transistor. Spots were left for contact points. The technique was called the PLANOR PROCESS.

Sprague Electric Company's Kurt Lehovec, a Czech-born physicist, had been working on a better way to make ALLOY JUNCTION TRANSISTORS, which Schockley had also invented. He successfully devised and improved manufacturing process. He then took the next step: isolating components electrically in an integrated circuit. His solution was similar to one that Jack Kilby used, i.e., employ a PN JUNCTION as a diode so that electricity could only flow in one direction. This was also successful.

Noyce was also thinking about making IC's in January 1959, and he succeeded in developing an IC combining Hoerni's PLANOR PROCESS and Lehovec's PN JUNCTION technique. Noyce's resulting device made manufacturing practical. For a resistor, he used a diode-isolated piece of silicon, and he built logic circuits by evaporating metal on top of the insulating layer through a mask. Horizontal connecting wires were an improvement over vertical ones.

Commercial development of the IC followed with all participating companies freely licensing their patents. In 1961, Fairchild and TI introduced the first chips that performed Boolean logic such as OR and NOR. In the mid-1960's, prices of the chips became reasonable, and NASA selected the IC's for the GEMINI capsule on-board computer.

Both difficult and expensive was the redesign of large computers both from tubes to transistors, and transistors to IC's.

The next step: why not put a central processor on a chip and call it a MICROPROCESSOR? To follow would be the personal computer.