Invention of the Integrated Circuit (1958): Kilby and Noyce

The invention of the integrated circuit in 1958–59 made it possible to build a complete electronic circuit on a single small piece of semiconductor, the device now known as the microchip. Jack Kilby of Texas Instruments demonstrated the first working example in Dallas on 12 September 1958, and Robert Noyce of Fairchild Semiconductor in California independently designed a more practical silicon version in 1959. Their work laid the foundation for modern computers, mobile phones and almost every electronic device in use today.

Background: the tyranny of numbers

The invention of the transistor at Bell Laboratories in 1947 had replaced fragile vacuum tubes with small solid-state switches. By the mid-1950s transistors were being made in large numbers, but they were still separate components. A complex circuit needed thousands of transistors, resistors and capacitors, all soldered together by hand.

Engineers called this problem the “tyranny of numbers”. Each extra part and joint added cost and weight and raised the chance that one bad connection would cause the whole system to fail. The problem mattered most to the military, which wanted lighter, more reliable electronics for missiles and aircraft, and to the computer industry, where early machines such as ENIAC had filled entire rooms.

Several solutions were tried. The US Army backed the “Micro-Module” programme, which stacked tiny standard components together. In Britain, the radar engineer Geoffrey Dummer had suggested in 1952 that electronic equipment might one day be made as a solid block with no connecting wires, but he did not build a working device.

Jack Kilby’s breakthrough at Texas Instruments

Jack Kilby, an electrical engineer who had grown up in Kansas, joined Texas Instruments (TI) in Dallas in May 1958. As a newcomer he had not yet earned holiday time, so he worked through the company’s summer shutdown in July with the laboratory nearly empty. He used the time to think about the miniaturisation problem.

Kilby’s key insight was that all the parts of a circuit, not only transistors, could be made from the same semiconductor material. Resistors and capacitors could be formed in a piece of germanium or silicon, even if they would not perform as well as conventional parts. If everything came from one block, the circuit could be made in one place.

On 12 September 1958 Kilby showed his managers a phase-shift oscillator built on a sliver of germanium about half an inch long. When he switched it on, an oscilloscope displayed a continuous wave, proving that the circuit worked. The device was crude: its components were linked by thin gold wires attached by hand. But it was the first working integrated circuit. TI filed a patent on Kilby’s “miniaturized electronic circuits” on 6 February 1959 and announced the “solid circuit” publicly at a trade show in New York in March 1959.

Robert Noyce and the planar chip

Meanwhile, in Mountain View, California, Fairchild Semiconductor had been founded in 1957 by eight engineers who had left the laboratory of William Shockley, one of the transistor’s inventors. One of them, the Swiss-born physicist Jean Hoerni, developed the “planar process” around 1957–59. It protected silicon transistors with a flat layer of silicon dioxide, so that devices could be made reliably on a flat wafer.

Robert Noyce, Fairchild’s research director, saw that the planar process allowed something more. On 23 January 1959 he wrote in his notebook that several components could be made in one piece of silicon and connected by a thin layer of metal laid down on the oxide surface, instead of loose wires. Fairchild filed his patent on 30 July 1959, and it was granted in April 1961.

To keep components on a single chip from interfering with each other, they had to be electrically isolated. A method using p-n junctions was patented by Kurt Lehovec of the Sprague Electric Company in 1959. A team at Fairchild led by Jay Last produced working planar integrated circuits in 1960, and Fairchild launched its “Micrologic” family of chips in 1961.

The patent battle

Because Kilby and Noyce had worked independently, TI and Fairchild fought a long legal battle over who owned the invention. Courts eventually decided several key claims in Noyce’s favour, but by then the two companies had already agreed in 1966 to cross-license their patents and share royalties. Both men were generous in later years, generally accepting each other as co-inventors.

The two approaches were complementary. Kilby proved the “monolithic” idea, that a whole circuit could be made from one piece of semiconductor. Noyce’s silicon planar design provided the practical method that the industry used to mass-produce chips.

Key people

• Jack Kilby (1923–2005), TI engineer who built the first working integrated circuit; he later co-invented the handheld electronic calculator.
• Robert Noyce (1927–1990), Fairchild co-founder who designed the planar integrated circuit; in 1968 he co-founded Intel.
• Jean Hoerni (1924–1997), physicist whose planar process made Noyce’s design possible.
• Kurt Lehovec (1918–2012), physicist who patented p-n junction isolation.
• Gordon Moore (1929–2023), Fairchild and Intel co-founder who in 1965 predicted the rapid growth in chip complexity now called Moore’s law.

Consequences

Early integrated circuits were expensive, and the first large buyers were the US government. The Air Force’s Minuteman II missile and NASA’s Apollo Guidance Computer, used in the Apollo 11 Moon landing, were among the most important early uses. These orders helped drive prices down dramatically during the 1960s.

In 1965 Gordon Moore observed that the number of components on a chip was doubling about every year, a rate he later revised to every two years. This trend, Moore’s law, held for decades. Chips grew from a handful of components to billions of transistors, making possible the microprocessor (1971), the personal computer and the smartphone.

Myths and facts

Myth: the first microchip was made of silicon. Kilby’s 1958 device was made of germanium, because that was the material TI had available for his experiments. Silicon became the standard material only with the planar chips developed at Fairchild, which is why the region around it later became known as Silicon Valley.

Myth: one person invented the microchip. The integrated circuit drew on many contributions: Kilby’s monolithic idea, Noyce’s metal interconnections, Hoerni’s planar process and Lehovec’s isolation method. Historians usually credit Kilby and Noyce jointly, while recognising the others.

Myth: the first chips were immediately cheap. Early integrated circuits cost far more than equivalent separate components. Government contracts for missiles and spacecraft, which valued small size and reliability over price, paid for the early production that later made chips affordable.

Timeline

• 1947, December: The point-contact transistor is demonstrated at Bell Laboratories.
• 1957, October: Fairchild Semiconductor is founded in California.
• 1958, July: Jack Kilby conceives the integrated circuit during TI’s summer shutdown.
• 1958, 12 September: Kilby demonstrates the first working integrated circuit.
• 1959, 23 January: Robert Noyce records his planar integrated circuit idea.
• 1959, 6 February: TI files Kilby’s patent application.
• 1959, 30 July: Fairchild files Noyce’s patent application.
• 1961, April: Noyce’s patent is granted; Fairchild begins selling Micrologic chips that year.
• 1966: TI and Fairchild agree to cross-license their integrated circuit patents.
• 2000, December: Kilby receives a share of the Nobel Prize in Physics.

Why it matters

The integrated circuit solved the “tyranny of numbers” and made electronics small, cheap and reliable enough to spread into every part of life. From spacecraft guidance to computers, phones and cars, nearly all modern technology depends on chips descended from the devices Kilby and Noyce designed in 1958 and 1959.