Zum Inhalt springen

Jack Kilby and the Integrated Circuit

Zusammenfassung

Jack Kilby was an electrical engineer at Texas Instruments who, during a summer when almost everyone else had gone on vacation, built the world’s first working integrated circuit in 1958. His invention — a few components on a single piece of germanium — launched the microelectronics revolution. It would take forty-two years, and the death of his chief rival, before Kilby received the Nobel Prize in Physics. Quiet, methodical, and constitutionally modest, he was almost the opposite of the Silicon Valley archetype his invention made possible.

The Man from Kansas

Jack St. Clair Kilby was born on November 8, 1923, in Jefferson City, Missouri, but grew up in Great Bend, Kansas, a small prairie town where his father ran an electric utility company. The story of how Kilby came to electronics is itself a kind of origin myth: as a teenager, he watched his father struggle to keep the rural power grid running during a catastrophic ice storm that knocked out telephone lines. The only way to reach isolated customers was shortwave radio, and the experience planted in Kilby a deep, practical fascination with electrical communication.

He had hoped to study at MIT, but narrowly failed its entrance exam in 1941. The rejection was a minor humiliation for a man who would later hold over sixty patents, but it was not a lasting detour — he enrolled instead at the University of Illinois at Urbana-Champaign (from which both his parents had graduated), and, after wartime military service, earned his bachelor’s degree in electrical engineering in 1947. Graduate school followed at the University of Wisconsin-Madison, where he completed a master’s degree in electrical engineering in 1950 while working at Centralab, a Milwaukee manufacturer of hearing aids and small electronic assemblies.

At Centralab, Kilby became expert at the painstaking hand-assembly that characterized electronics manufacturing of the era: discrete resistors, capacitors, and transistors wired together one connection at a time. It was skilled, meticulous work, and Kilby was good at it. But he could see clearly that it could not scale. The “tyranny of numbers,” as the industry called it, was becoming acute. The most capable electronic circuits required so many individual components that assembly became impossibly complex, prohibitively expensive, and catastrophically unreliable. Every new connection was a potential point of failure. If you wanted to build a missile guidance system with a hundred thousand transistors, you would need a factory floor of human solderers, and even then the failure rate would be unacceptable.

The Tyranny of Numbers

The problem was not restricted to military applications. The telephone system demanded ever more complex switching equipment. Business users wanted calculating machines that could handle payroll and inventory. The vision of a truly capable electronic computer — the kind that could actually solve differential equations in real time, or simulate physical processes — required circuit complexity that was, practically speaking, beyond reach.

The semiconductor industry’s response to this challenge was to make individual components smaller and more reliable, while improving assembly techniques. Kilby worked within this framework for years, but he was increasingly convinced that miniaturization alone could not solve the problem. What was needed was a different approach to the nature of a circuit itself.

His key insight was this: the different components in a circuit did not need to be made of their ideal materials. Resistors worked best in one material, capacitors in another, transistors in germanium or silicon. But if all of these could be fabricated from a single material — even if some of them were not quite as good as the optimal versions — then they could be manufactured simultaneously, on a single substrate, in a single process. The connections between them would not need to be wired by hand; they could be metal paths deposited on the surface.

This was the “monolithic idea”: one stone, one circuit, one manufacturing step.

The Summer of 1958

In May 1958, Kilby joined Texas Instruments in Dallas. TI was one of the most aggressive semiconductor companies in the country, and Kilby was hired for his expertise in miniaturization. Almost immediately, he ran into a circumstance that would define his career: as a new employee, he had not yet accrued vacation time. When the summer shutdown arrived in late June and nearly the entire TI workforce scattered to their families and beaches, Kilby stayed behind in an almost entirely empty laboratory.

What followed was one of the most consequential episodes of solitary invention in the history of technology. Kilby had access to the lab’s equipment, no meetings, no colleagues pulling him toward other problems. He worked steadily and deliberately toward the idea he had been refining since his time at Centralab.

On July 24, 1958, he sketched in his notebook the design for a circuit in which all components — transistors, resistors, capacitors — would be made from the same semiconductor material and connected by metal deposited on the surface. He estimated that such a circuit could function. He began to build it.

Info

Kilby’s chosen material was germanium — not because it was optimal, but because TI had germanium in the lab and he understood it. Silicon was newer and harder to work with. Noyce’s subsequent planar silicon process was technically superior for manufacturing at scale, but Kilby’s germanium prototype was sufficient to prove the concept. Both inventions were genuine; neither required the other.

The prototype he assembled was crude. A sliver of germanium, roughly the size of a pencil eraser, with a transistor and other components etched into its surface and connected by fine gold wires attached by hand. On September 12, 1958 — the day Kilby chose for his demonstration to TI management — he connected the device to an oscilloscope. It produced a continuous sine wave. The circuit worked.

TI’s vice president Mark Shepherd watched the needle move on the oscilloscope and understood immediately that he was looking at something transformative. Kilby filed a patent application on February 6, 1959, describing his “miniaturized electronic circuits.”

Robert Noyce and the Patent War

While Kilby was working alone in Dallas, Robert Noyce at Fairchild Semiconductor in California was pursuing a parallel path toward the same destination. Noyce’s version, filed in July 1959, used Jean Hoerni’s planar process — depositing all components and their metal interconnections on a flat silicon surface through photolithographic masking. It was, in manufacturing terms, decisively superior to Kilby’s approach. Where Kilby’s gold wire bonds still required individual attachment by hand, Noyce’s metal connections were printed.

The patent dispute that followed consumed enormous legal resources and generated lasting industry bitterness. Texas Instruments held the earlier filing and the broader conceptual patent; Fairchild held the manufacturing method. Both sides pursued the litigation aggressively, and the fight lasted nearly a decade. In 1966, U.S. Court of Claims ruled on some of the core patents. In 1966, the two companies reached a cross-licensing agreement that acknowledged both Kilby and Noyce as independent co-inventors. The industry moved forward.

The historical consensus that emerged is generous to both men: Kilby first demonstrated a working IC and filed first; Noyce invented the more manufacturable and ultimately more commercially dominant version. Their contributions were genuinely complementary, not competitive. Kilby proved the concept; Noyce proved it could be manufactured. The modern semiconductor industry descends from both.

See also: Robert Noyce and Fairchild Semiconductor and The Integrated Circuit Revolution.

Character and Contrast

The contrast between Kilby and Noyce was almost total. Noyce was charismatic, articulate, socially gifted — the archetypal Silicon Valley founder before Silicon Valley existed as a concept. He wore open-collared shirts to meetings with Pentagon generals, gave interviews easily, and inspired intense loyalty. He became what colleagues called the “Thomas Jefferson of Silicon Valley”: the man who set the cultural tone for an entire industry.

Kilby was the opposite. Tall — six feet six inches — soft-spoken, and constitutionally modest, he deflected praise with the patient persistence of a man who simply did not believe in making claims he could not prove. He spent his entire career at Texas Instruments, never founding a company, never becoming a public figure, never seeking the limelight that Noyce seemed to attract effortlessly.

When journalists found Kilby in his later years, he tended to emphasize that he and Noyce had both contributed essential things, and that the invention of the integrated circuit was inseparable from the work of hundreds of other engineers and scientists. He genuinely believed the IC was an inevitable invention — that someone would have arrived at it within months had he and Noyce not done so. Whether or not that is true, it is a revealing attitude. Kilby was not interested in historical priority; he was interested in the work itself.

This quality of character — the preference for problems over recognition — was consistent throughout his life. He did not attend conferences to network; he attended to learn. He did not write articles for prestige; he wrote when he had something to say. He was, by all accounts, a genuinely good engineer and a genuinely decent man, and these qualities made him less famous than his invention deserved.

Beyond the IC: Calculator and Printer

Despite his low public profile, Kilby’s inventive work continued for decades after 1958. In the late 1960s, he led the TI team that developed the first handheld electronic calculator. The prototype — called the “Cal Tech” — appeared in 1967. It was battery-powered, could perform the four arithmetic operations, and fit in a jacket pocket. Nothing like it had existed before. The commercial product, the TI-2500 Datamath, went on sale in 1972 for $149.95 — roughly $1,000 in 2024 dollars.

Within a few years, competition had driven the price of such devices below $20. The slide rule — which had been the engineer’s essential tool since the seventeenth century — became a museum piece almost overnight. Mechanical calculators, which required maintenance contracts and occupied half a desk, were replaced by devices small enough to fit in a shirt pocket. The calculator was the first consumer product to mass-produce integrated circuits, and in doing so it drove down IC manufacturing costs for everything that followed.

Kilby also held the patent on the thermal printer — a technology that used heat rather than ink to produce marks on specially coated paper. Thermal printing became ubiquitous in receipt printers, fax machines, and label makers, and it remains common in point-of-sale terminals worldwide.

He was never wealthy from these inventions in the manner that Noyce and Gordon Moore became wealthy from Intel. Texas Instruments paid him a salary and eventually royalties, but the enormous fortunes of the semiconductor era were made by founders of companies, not by employees — however transformative their work.

The Nobel Prize, Forty-Two Years Later

In 2000, the Nobel Committee awarded Jack Kilby the Nobel Prize in Physics “for his part in the invention of the integrated circuit.” The announcement acknowledged a debt the committee had been extraordinarily slow to pay: by that point, the integrated circuit had been the foundation of the global electronics industry for four decades. Personal computers, mobile phones, the internet, medical imaging, industrial automation — all of it depended on the principle Kilby had demonstrated in a nearly empty laboratory in Dallas in the summer of 1958.

The committee noted that Robert Noyce would have shared the prize had he survived. Noyce had died in June 1990, of a sudden heart attack at sixty-two. The Nobel Prize is not awarded posthumously. This limitation of the committee’s rules — and the decades-long delay in recognizing the IC — generated considerable comment. Some argued that the committee should have acted while both men were alive. The transistor had been recognized in 1956, only nine years after its invention. The integrated circuit waited forty-two years.

Kilby, in his Nobel lecture, was characteristically restrained. He acknowledged Noyce’s contribution directly and spent considerable time discussing the “social” impact of the integrated circuit — the way cheap computation had changed medicine, education, and communication. He noted that the cost of a single transistor had fallen from roughly a dollar in 1958 to less than a millionth of a cent by 2000 — a price reduction without precedent in the history of manufactured goods. He did not dwell on the patent disputes or his place in history. He was seventy-six years old, in imperfect health, and he seemed genuinely surprised and genuinely pleased to be in Stockholm.

He returned to Dallas and continued working at TI as a consultant. He died on June 20, 2005, at the age of eighty-one, of non-Hodgkin’s lymphoma.

Dead End: The Hybrid Circuit Approach

The integrated circuit was not the only proposed solution to the tyranny of numbers. An alternative, the hybrid circuit, produced separate miniaturized components and mounted them on a ceramic substrate with printed connections. Hybrid circuits were smaller than discrete-component circuits and larger than ICs, and they allowed each component to be made from its optimal material.

Warnung

The military, particularly the U.S. Army Signal Corps, initially preferred hybrid circuits because they could be assembled from proven components and inspected for individual failures. The Army sponsored competing hybrid miniaturization programs through the late 1950s in parallel with TI’s IC work. When the Apollo guidance computer program chose the integrated circuit in the early 1960s — requiring unprecedented quantities of ICs and driving manufacturing costs down — the economic argument for hybrids collapsed. Hybrid circuits survive today in specialized applications where power density, RF performance, or radiation hardness make monolithic ICs impractical, but they never became the dominant platform for any volume market.

Legacy

The integrated circuit is among the most consequential inventions in human history. The density of transistors on a chip has roughly doubled every two years for six decades — a trajectory captured in Gordon Moore’s Law — moving from Kilby’s handful of components in 1958 to the tens of billions of transistors in a contemporary processor. The cumulative economic value created by this trajectory is impossible to quantify; it underlies essentially all of the information technology that has transformed global society since the 1970s.

Kilby’s specific contribution was to prove the monolithic idea was physically realizable. He did so from a position of institutional isolation — new to his job, alone in a lab — by taking a problem that had been obvious to everyone in the industry and actually building the solution. In the history of invention, that last step — from insight to demonstration — is often the hardest, and the most overlooked.

📚 Sources