Semiconductor Device-and-Lead Structure. U.S. patent 2,981,877, filed on July 30, 1959, granted on April 25, 1961.
[Washington D.C. United States Patent Office, 1961]. First edition — an original first-issue printing, produced by the United States Patent Office for the grant of 25 April 1961, the earliest form in which the invention entered print — of the first patent ever issued for the monolithic integrated circuit, and the only original copy of that patent which has so far been traced anywhere. Its importance lies not simply in priority, but in the specific technical solution it secures. Jack Kilby at Texas Instruments had already demonstrated a working solid circuit, but Kilby’s device, built in germanium and interconnected by hand-attached gold wires, did not provide a manufacturing architecture suitable for large-scale production — in the judgment of Leslie Berlin, Noyce’s biographer, it could not have been manufactured in any quantity, genuine integrated circuit though it undoubtedly was. Robert Noyce’s invention, developed independently at Fairchild Semiconductor in Mountain View, California, and built on Jean Hoerni’s planar process, described a complete circuit formed within a single piece of silicon, electrically isolated by p–n junctions and interconnected by metal deposited over an insulating oxide layer. That was the crucial step. It made the integrated circuit flat, protected, reproducible, and scalable. In practical and industrial terms, this is the patent that made the integrated circuit manufacturable, and it is for that reason that it may justly be called the blueprint of Silicon Valley. Provenance: the present copy descends through Austrian patent-office channels rather than through any American institutional or commercial route. From its founding in 1899 the Austrian Patent Office maintained one of the most complete collections of full foreign patent specifications anywhere in Europe — the American series among them from the very first year, when full runs of United States specifications came over from the Technische Hochschule in Vienna, and some 130,000 American specifications stood in the founding examiners’ stock. Where most institutions outside the United States economised with condensed abstracts, summaries, or later photographic formats, Vienna kept the complete paper record, under the routine document-exchange between patent offices later codified in 35 U.S.C. § 12: by the end of 1957 the office library alone held over 90,000 bound volumes of patent specifications, beside a separate examiners’ working stock — the Vorprüfermaterial — of some three million loose specifications kept in numerical working files rather than bound library runs. It is to that Austrian retention, the fullest documentary holding of American patents outside the United States, that the survival of the present copy is owed: original first printings of important U.S. patents have surfaced from Austrian holdings, and the present example belongs to that route of survival. The institutional history is set out by Hans Jancik, the Patent Office’s own library director, in the office’s sixty-year Festschrift (1959), and in Wirth and Pinwinkler’s history of the Patentamt (2024). The technical background remains essential to understanding what this document is. Noyce, the son of a Congregationalist minister, grew up in Grinnell, Iowa, and was introduced to solid-state physics at Grinnell College by Grant Gale, a friend of John Bardeen. In the summer of 1948, months after the transistor had been announced, Gale obtained samples from Bardeen for his students, and Noyce was therefore among the first young Americans to work directly with the device. He decided to pursue solid-state physics in graduate school, took his Ph.D. at MIT in 1953 — the field so new that formal courses in it barely existed — and went into industry rather than the academy, because industry was then where the frontier of semiconductor research actually ran. After three unsatisfying years working on germanium transistors at Philco in Philadelphia he drove west in April 1956 to Mountain View, California, where he and Jean Hoerni joined William Shockley’s newly founded Shockley Semiconductor Laboratory. Shockley’s laboratory was, in institutional terms, one of the starting points of Silicon Valley. It was small, ambitious, and technically distinguished, but it was also unstable: Shockley, though a brilliant physicist, was a notably poor manager, imposed intrusive internal procedures, and remained committed to a four-layer germanium diode that his leading researchers regarded as commercially unpromising. Noyce, Hoerni, and six colleagues believed the future belonged to silicon. On 18 September 1957 the eight resigned together — industry legend remembers them as the traitorous eight — and on 1 October Fairchild Semiconductor opened with the backing of the Fairchild Camera & Instrument Corporation of New York, in a financing brokered by the young Arthur Rock of Hayden, Stone & Co. that became a template for the venture-funded start-up. Noyce, for his technical authority and his ability to lead people, became the company’s general manager in 1959. At that date Fairchild and Shockley were the only true semiconductor firms in the Santa Clara Valley; the later history of Silicon Valley begins, in no small part, with this break. Fairchild concentrated on silicon. In the late 1950s the state of the art was the silicon mesa transistor, which was promising but structurally vulnerable: the mesa projected above the wafer surface, exposed to contamination and mechanical instability, and its attached wires could slip under even minor disturbance. In late 1958 and early 1959 Hoerni solved the problem with the planar process — he diffused the active regions down into the silicon wafer and covered the surface with a protective layer of silicon dioxide, leaving only selected windows open for contact. The resulting structure was flat, protected, and far more reliable: Hoerni had entered the idea in his notebook as early as December 1957 — Noyce countersigned the entry — walked it through with Fairchild’s patent attorney on 14 January 1959, and had a working planar transistor by March. Hoerni’s advance did not by itself create the integrated circuit, but it provided the necessary platform for one. A second element came from Kurt Lehovec at Sprague Electric, who recognised that p–n junctions could electrically isolate components formed on a common substrate. Noyce’s achievement was to combine the available elements into a single manufacturable system. On 23 January 1959 — nine days after Hoerni’s planar disclosure — he entered the idea in his Fairchild patent notebook under the heading ‘Methods of isolating multiple devices’: in many applications, he wrote, it would be desirable to make multiple devices on a single piece of silicon, in order to make the interconnections between them part of the manufacturing process. He took Hoerni’s planar structure, used reverse-biased junctions for isolation, incorporated resistive elements, and formed the interconnection wiring by evaporating metal across the protective oxide through a mask — eliminating the protruding wires and fragile geometries that limited every earlier design. Noyce later described the invention as a progressive synthesis rather than a single sudden insight, and the description is convincing; what matters is that the synthesis worked, and that industry could use it. Fairchild filed the application on 30 July 1959, its patent attorney having asked Noyce and his key engineers to think about the process as broadly as possible, so that the patent would cover the greatest range of potential applications. The patent itself runs to six pages of specification with three leaves of diagrams. The drawings show, in plan and cross-section, a silicon body with active regions formed by diffusion, the whole surmounted by a continuous insulating oxide film through which selected windows have been opened, with metallic conductors deposited on top of the oxide and running between the contacts to interconnect the regions into a complete circuit. The claims secure exactly that geometry — in the language of the first claim, an insulating layer of the semiconductor’s own oxide, and an electrical connection comprising a conductor adherent to that layer, extending from contact to contact across the junctions beneath. The wording is at once specific enough to define a manufacturable structure and broad enough to cover the architecture into which integrated circuits would in fact develop; that combination of specificity and generality is part of what made this the decisive document of the field, and the single word ‘adherent’ would later decide the eight-year contest over priority in the invention of the integrated circuit itself. The patent was granted on 25 April 1961, the first patent issued anywhere in the integrated-circuit field: Hoerni’s planar patent followed in March 1962, Lehovec’s isolation patent in April 1962, and Kilby’s principal application, though filed nearly six months before Noyce’s, did not issue until June 1964. Weeks before the grant, at the IRE Show in New York in March 1961, Fairchild had unveiled Micrologic, the first practical, mass-manufacturable family of planar integrated circuits — Texas Instruments had offered a germanium solid circuit commercially a year earlier, but shipped only a few dozen hand-wired evaluation units, and its own planar line followed only in October 1961. The product launch made the press; the grant of the patent drew no contemporary notice that has been traced in the accessible 1961 runs of the general and trade press. Nor was there ever a companion paper: Bardeen and Brattain had followed their 1948 transistor patent within weeks with a disclosure in the Physical Review, but no contemporary scientific paper by Noyce announced the integrated circuit — its manufacturable form entered the printed record here and nowhere else. An issued patent was a working document — any copy could be had from the Patent Office for twenty-five cents — and nobody in 1961 treated the document as an object to be kept. John Bardeen — whose friend Grant Gale had put one of the first transistors into Noyce’s hands at Grinnell in 1948 — lived, in Berlin’s account, to call the integrated circuit an invention “as important as the wheel”. The legal history bears directly on the document. The Patent Office declared an interference between Noyce’s issued patent and Kilby’s pending application, and the Board of Patent Interferences initially awarded the four principal interconnection counts to Kilby. On cross-appeals, the Court of Customs and Patent Appeals ruled in November 1969 that Kilby’s application had never disclosed leads that adhere to the insulating oxide — his gold wiring was merely ‘laid down’ — and every disputed count ended with Noyce; the United States Supreme Court declined review the following year. By then the commercial question had long been settled privately: in the summer and early autumn of 1966 Fairchild and Texas Instruments negotiated a cross-licence under which each acknowledged the other’s claim and every third party building integrated circuits thenceforth needed licences from both — and in October 1966 the Franklin Institute awarded Noyce and Kilby its Ballantine Medal jointly. History has settled on the same verdict: Kilby built the first working circuit; Noyce patented the circuit in the form the world could manufacture. When Kilby received the Nobel Prize in Physics in 2000 — Noyce had died of a heart attack in June 1990, at sixty-two, and the prize is not awarded posthumously — the Royal Swedish Academy’s published account described the two men as jointly considered the inventors of the integrated circuit, and identified Noyce’s aluminium-over-oxide interconnection, against Kilby’s gold, as the determining factor in the patent contest. In his Nobel lecture Kilby credited Noyce with showing “the desirability of using the planar process with metal leads over the oxide”; in the autobiographical essay written for the Nobel Foundation he went further: “If he were still living, I have no doubt we would have shared this prize”. Noyce’s standing among his peers had long been of that order — more than a dozen patents, the IEEE Medal of Honor, the National Medals of Science and of Technology, and, with Kilby, the first Charles Stark Draper Prize of the National Academy of Engineering in 1989, for what the Academy’s citation calls their independent development of the monolithic integrated circuit. And when the National Inventors Hall of Fame inducted Noyce in 1983, a year after Kilby, the patent its record names for the induction is the present one — U.S. Patent 2,981,877. The wider consequences followed quickly. President Kennedy’s commitment to a lunar landing created an immediate market for miniaturised, reliable electronics: in 1962 the MIT Instrumentation Laboratory chose integrated circuits for the Apollo guidance computer, each machine carrying some four thousand Fairchild Micrologic logic gates — the direct commercial embodiment of this patent’s architecture — and through 1965 the Apollo programme was the largest consumer of integrated circuits on earth, until the Minuteman II missile-guidance programme overtook it. In 1965 Gordon Moore, Noyce’s director of research at Fairchild, plotted the growth in the number of components per chip and extrapolated the doubling curve now known as Moore’s law — an observation inseparable from the manufacturing logic secured in the present patent, since only the flat, photolithographically printed circuit could be shrunk on that schedule. In 1968 Noyce and Moore left Fairchild to found the company that became Intel, and the memory chips and microprocessors that followed belong to the same unbroken technical lineage. By 1961 the two foundational mathematical ideas of digital computing had been available for a generation — Turing’s universal machine of 1936, and Shannon’s demonstration in 1938 that Boolean logic could be realised in switching circuits — and their roots reached back through Boole’s Laws of Thought (1854) to Leibniz’s Dissertatio de arte combinatoria (1666). Each generation of hardware had made the switches smaller and faster: relays, then vacuum tubes, then discrete transistors. Noyce’s patent is the step that made the switches small enough, numerous enough, and cheap enough to print Turing’s universal machine onto a single piece of silicon. Sixty-five years on, that architecture is the substrate of the digital world, and of the artificial-intelligence era being built on it now. The first planar integrated circuit Fairchild fabricated to this design, in 1960, integrated four transistors; the processor die NVIDIA introduced for the AI data centre in 2022 carries eighty billion, patterned on silicon beneath protective dielectric with metal interconnection above — the same structure this patent claims, iterated across six decades of Moore’s law. The transistor built this way has become, by an industry estimate reported by the Computer History Museum, the most frequently manufactured object in human history, some thirteen sextillion by 2018; the industry that manufactures it recorded 791.7 billion dollars in sales in 2025 and is projected to reach roughly a trillion in 2026; and the museum’s own summary of the field stands as the plainest statement of the patent’s durability: the planar method remains the fundamental approach by which integrated circuits are produced today. If the historical importance of the patent is clear, its physical rarity requires explanation. Original U.S. utility patents of this date were not commercial publications intended for bibliographical survival. They were official working documents: the complete patent, with full text and diagram leaves, was printed for those who needed it, while routine public notice ran through the Official Gazette, which in no era printed the entire patent. In practice engineers and attorneys relied on the Gazette, on abstracts, on library access, or on ordering individual copies at twenty-five cents when a particular patent required study; the roughly twenty depository libraries then operating filed their copies loose, in numerical order, as reference stock rather than as library books. Such documents were used heavily and preserved poorly — and by the later 1960s the Patent Office itself was reproducing thousands of copies a day from microfilm masters rather than maintaining paper. That publication system explains why original paper copies of the major twentieth-century patents are now far scarcer than their significance would suggest. Once searching moved to microfilm and then to databases, the paper files lost their working value: in March 2002 the Patent Office formally declared the paper patent copies in its own search rooms temporary records, to be discarded as wastepaper or donated, and the National Archives preserves the prosecution files of American patent history — Edison’s light-bulb application among its milestone documents — not the patents as issued. The depository libraries went the same way: the Patent Office had supplied them on paper only until 1982, when the programme converted to microfilm — its own programme literature records paper back-files being reclaimed from depository libraries during the transition — and no library has been identified that still holds a paper run covering 1961. For the present patent the census is now specific: the Robert Noyce papers at Stanford contain no copy of it (the patent series was checked at first hand in 2026, and the public finding aid names no folder for it); the Smithsonian’s only holding is a curator’s exhibition file whose complete contents proved to be reproduction prints; the Computer History Museum’s holding is catalogued as photocopies; and no original has been traced in WorldCat, in the National Archives’ patent holdings, or in any auction record. No other original first-issue copy is documented anywhere. Because individual patent specifications were distributed into uncatalogued numerical files and are almost never recorded item by item, that negative is a limit on tracing rather than a proven survival count, and an uncatalogued institutional or corporate file may yet yield another; but every repository where a copy would first be expected has been examined, and none holds one. This copy should therefore be understood in three ways at once: as the first issued patent of the integrated-circuit age, the document in which the manufacturable form of the chip enters the public record; as an original first-issue Patent Office printing rather than any later reproduction; and as the only documented survivor of a class of documents that was read to death and then systematically discarded. Gordon Moore’s Electronics paper of 19 April 1965 projected the future of the architecture his colleague had claimed here, and six decades of that projection have now been redeemed in silicon, most recently by the processors of the artificial-intelligence era. The blueprint of Silicon Valley is not a metaphor: it is this document. References: Berlin, The Man Behind the Microchip: Robert Noyce and the Invention of Silicon Valley (Oxford, 2005) — Reid, The Chip: How Two Americans Invented the Microchip and Launched a Revolution (1984; revised 2001) — Lojek, History of Semiconductor Engineering (2007) — Lécuyer, Making Silicon Valley: Innovation and the Growth of High Tech, 1930–1970 (2006) — Noyce v. Kilby, 416 F.2d 1391 (C.C.P.A. 1969), certiorari denied 400 U.S. 818 (1970) — Kilby, ‘Turning Potential into Realities’, Nobel Lecture, 8 December 2000, with the autobiographical essay, in Les Prix Nobel 2000 — Moore, ‘Cramming more components onto integrated circuits’, Electronics, 19 April 1965 — Jancik, ‘Zur Geschichte der Bibliothek des Österreichischen Patentamtes’, in 60 Jahre Österreichisches Patentamt 1899–1959 (Vienna, 1959), pp. 99–101 — Wirth & Pinwinkler, Behörde, Wissensspeicher, Serviceeinrichtung: Das Österreichische Patentamt 1899–2024 (2024) — 35 U.S.C. §§ 11–12 — Furman, Nagler & Watzinger, ‘Disclosure and Subsequent Innovation: Evidence from the Patent Depository Library Program’, American Economic Journal: Economic Policy 13 (2021) — Computer History Museum, The Silicon Engine — National Inventors Hall of Fame, inductee record for Robert N. Noyce (1983) — Semiconductor Industry Association, annual sales data, February 2026. Folio (277 × 195 mm), pp. [6, the last blank], with three leaves of diagrams printed on rectos only; wire-stitched as issued. An exceptionally well-preserved original pamphlet.
Item #6003
Price: $250,000.00
