Making Micrologic: The Development of the Planar IC at Fairchild Semiconductor, 1957–1963 David A. Laws Computer History Museum Michael Riordan Stanford University and University of California, Santa Cruz The modern integrated circuit evolved from developments at Fairchild Semiconductor during the late 1950s and early 1960s. Using infor- mation from laboratory notebooks, internal company memoranda, oral histories, and personal communications, this article reconstructs and analyzes the activities of the Fairchild development team, including the company’s efforts to improve and market ICs during the early 1960s. In 1949 Jack Morton, who headed Bell Labs transistor development, made a remarkably prescient prediction. ‘‘Imagine a technique in which not only the connecting leads and passive elements ... but also the active semi- conductor elements are ‘printed’ in one con- tinuous fabrication process,’’ he speculated in an internal report. 1 ‘‘Savings in life, size and initial cost per element ... seem great enough to warrant reduced maintenance costs on a large scale by replacement of whole assemblies.’’ This obscure report was the earliest adumbration we can find of the idea of an integrated circuit (IC), made at a time when the only working transistors were point-contact devices like that invented by John Bardeen and Walter Brattain, having two closely spaced metal leads impinging upon a germanium surface. 2 At the time, it was thus difficult, if not impossible, to imag- ine the active elements of a circuit (its vac- uum tubes or transistors) being included as part of the printing process Morton envi- sioned. But the transistor’s rapid evolution over the next decade—from point-contact to junction to mesa and especially to planar devices—is what made including them possible. 3 In the early 1960s, Fairchild Semiconduc- tor in Palo Alto, California, successfully incorporated both active and passive circuit elements as well as their metallic intercon- nections into a coherent lithographic pro- cess. This manufacturing process triumphed over the numerous other approaches then being attempted to fabricate monolithic ICs. 4 Afterward, chipmakers could effectively ‘‘print’’ these circuits using technologies such as diffusion, epitaxy, and vapor deposition that operated from a single side of a silicon wafer. Although it was technologically far more complex, this process resembled the lithographic techniques used to print alpha- numeric characters on paper in books, maga- zines, and newspapers, tapping into a deep, familiar technology pool that had been evolving for centuries. 5 This was key to the tremendous cost reductions that ensued. Today, a single transistor on the surface of a silicon microchip costs a tiny fraction of one cent, as compared with several dollars per discrete transistor in the late 1950s. As has been elaborated elsewhere, Fairchild Semiconductor was founded in September 1957 by eight dissident scientists and engi- neers from the Shockley Semiconductor Labo- ratory. 6 It was an auspicious moment to start a high-tech firm initially focused on manu- facturing double-diffused (base and emitter) silicon mesa transistors. Within weeks, the 20 IEEE Annals of the History of Computing 1058-6180/12/$31.00  c 2012 IEEE Published by the IEEE Computer Society