Invention of the Transistor (1947): Bell Labs and the Chip Age

The invention of the transistor took place at Bell Telephone Laboratories in Murray Hill, New Jersey, in December 1947, when physicists John Bardeen and Walter Brattain made a small germanium device that could amplify an electrical signal. Their point-contact transistor, soon followed by William Shockley’s junction transistor, replaced the bulky, power-hungry vacuum tube. It became the basic building block of radios, computers, phones and nearly every modern electronic device.

Background: the limits of the vacuum tube

In the first half of the 20th century, electronics depended on the vacuum tube, or valve. A tube could amplify weak signals and switch currents, which made long-distance telephone calls, radio, radar and the first electronic computers possible. But tubes were fragile glass bulbs with hot filaments. They were large, consumed a lot of power, generated heat, needed time to warm up and burned out regularly.

These weaknesses were a serious problem for the American Telephone and Telegraph Company (AT&T), whose national network relied on thousands of amplifiers and on mechanical relays in telephone exchanges. Its research arm, Bell Labs, was one of the best-funded industrial laboratories in the world. Mervin Kelly, Bell Labs’ director of research, wanted a solid-state replacement for both tubes and relays.

Scientists had long known that certain crystals, the semiconductors, could conduct electricity in unusual ways. Crystal detectors had been used in early radios, and during the Second World War germanium and silicon diodes were developed for radar. Wartime research greatly improved the purity of these materials, and quantum physics had begun to explain how electrons moved through solids.

The solid-state physics group

In 1945 Kelly set up a solid-state physics group at Bell Labs, led by William Shockley, a brilliant and ambitious theoretical physicist. Its members included the experimentalist Walter Brattain, who had worked on semiconductor surfaces since the 1930s, and the theorist John Bardeen, who joined Bell Labs in late 1945.

Shockley’s first idea was a “field-effect” amplifier, in which an electric field applied from outside would control the flow of current through a thin layer of semiconductor. In 1945 he tried to build such a device, but it did not work: the effect was far weaker than his theory predicted.

The breakthrough of December 1947

Bardeen proposed an explanation in 1946. Electrons trapped in “surface states” at the semiconductor’s surface, he argued, were screening the interior from the external field. Over the following year Bardeen and Brattain carried out a series of experiments to understand and overcome these surface effects, using electrolytes, metal contacts and different arrangements of germanium and silicon.

In November and December 1947 their experiments moved quickly. Brattain wrapped a strip of gold foil around the edge of a plastic triangle, slit the gold at the point with a razor blade to create two contacts very close together, about 50 micrometres apart, and pressed the point onto the surface of a slab of germanium. A spring held it in place.

On 16 December 1947 the device amplified a signal. A small current flowing through one contact controlled a much larger current through the other, which is the essence of a transistor. On 23 December Bardeen and Brattain demonstrated it to Bell Labs managers, including Harvey Fletcher and Ralph Bown, by using it to amplify speech. That date is generally given as the transistor’s birthday.

The junction transistor and the public announcement

Shockley was proud of the group’s success but frustrated that he had not been directly involved in the decisive experiment. Over the next weeks, working largely on his own, he worked out the theory of a different and more robust device. In January 1948 he conceived the junction transistor, in which layers of differently treated semiconductor, called p-type and n-type, are stacked like a sandwich.

The junction transistor was easier to understand and manufacture reliably than the delicate point-contact design. The first working junction transistors were made at Bell Labs in 1950 and 1951, with important contributions from Morgan Sparks and Gordon Teal. Junction transistors soon became the standard.

Tensions over credit split the group. Shockley’s patent claims were limited, Bell Labs’ lawyers built the point-contact patent around Bardeen and Brattain, and relations between the three men soured. Bardeen left Bell Labs in 1951.

Bell Labs kept the invention secret for several months while patent applications were prepared. The name “transistor”, suggested by the engineer John R. Pierce, was chosen in May 1948. On 30 June 1948 Bell Labs announced the transistor at a press conference in New York.

The public response was muted at first; The New York Times reported it in a short item. Early point-contact transistors were noisy, varied from one unit to the next and were expensive. Engineers needed several years to turn the laboratory device into a reliable industrial product.

Key people

• John Bardeen (1908–1991): theoretical physicist whose idea of surface states guided the experiments. He later shared a second Nobel Prize in Physics, in 1972, for the theory of superconductivity, the only person to win the physics prize twice.
• Walter Brattain (1902–1987): experimental physicist who built the first working point-contact device.
• William Shockley (1910–1989): head of the group and inventor of the junction transistor. He later founded Shockley Semiconductor Laboratory in California, and in later life became notorious for promoting discredited racial theories.
• Mervin Kelly (1894–1971): Bell Labs research director who set up the solid-state programme.

Consequences: from hearing aids to computers

AT&T, which was under antitrust pressure from the US government, chose to license the transistor widely. In 1952 Bell Labs held a symposium for licensees, who paid $25,000 each to learn how to make transistors. Among the companies that took licences were Texas Instruments and a small Japanese firm, Tokyo Tsushin Kogyo, later renamed Sony.

The first commercial products using transistors included hearing aids in 1952–1953. The Regency TR-1, the first commercial transistor radio, went on sale in the United States in late 1954, and Sony’s transistor radios followed. In 1954 Texas Instruments, with Gordon Teal, produced the first commercial silicon transistors, which coped better with heat than germanium.

Computers changed too. Experimental transistorised computers appeared in the early 1950s, including one built at the University of Manchester in 1953, and by the end of the decade companies such as IBM were building transistor computers that were smaller, cooler and far more reliable than tube machines like ENIAC.

In 1956 Bardeen, Brattain and Shockley shared the Nobel Prize in Physics “for their researches on semiconductors and their discovery of the transistor effect”. In the same year Shockley moved to California and opened Shockley Semiconductor Laboratory in Mountain View. Eight of his researchers left in 1957 to found Fairchild Semiconductor, a key moment in the rise of Silicon Valley.

Legacy: the integrated circuit and beyond

The next great step was to put many transistors on a single piece of semiconductor. Jack Kilby at Texas Instruments and Robert Noyce at Fairchild independently developed the integrated circuit in 1958–1959. Another type, the metal-oxide-semiconductor field-effect transistor (MOSFET), demonstrated at Bell Labs in 1959–1960 by Mohamed Atalla and Dawon Kahng, finally realised the field-effect idea Shockley had pursued in 1945. The MOSFET became the most widely manufactured device in history.

Transistor counts have risen enormously. Gordon Moore observed in 1965 that the number of components on a chip was doubling at a regular rate, an observation later called Moore’s law. Modern processors contain billions of transistors, each far smaller than a living cell, and the total number manufactured is estimated in the sextillions.

Myths and facts

• Myth: Shockley invented the first transistor. The first working device was built by Bardeen and Brattain. Shockley led the group and invented the junction transistor shortly afterwards. Bell Labs publicity photos, such as the one that shows Shockley at the microscope, helped blur this distinction.
• Myth: The transistor was an instant sensation. The 1948 announcement attracted little attention, and it took most of the 1950s for transistors to become cheap and reliable.
• Myth: Nobody else was close. In Paris, the German physicists Herbert Mataré and Heinrich Welker, working for a Westinghouse subsidiary, independently developed a similar point-contact amplifier, the “transistron”, in 1948. The two efforts were unconnected, a sign that semiconductor science was ripe for the discovery.

Timeline

• 1945: Bell Labs creates a solid-state physics group under William Shockley.
• 1946: John Bardeen proposes the theory of surface states to explain failed experiments.
• 1947, 16 December: Bardeen and Brattain’s point-contact device amplifies a signal.
• 1947, 23 December: The transistor is demonstrated to Bell Labs managers.
• 1948, January: Shockley conceives the junction transistor.
• 1948, 30 June: Bell Labs announces the transistor to the press.
• 1952: Bell Labs holds a symposium for transistor licensees.
• 1954: The Regency TR-1 transistor radio goes on sale; Texas Instruments makes silicon transistors.
• 1956: Bardeen, Brattain and Shockley receive the Nobel Prize in Physics.

Why it matters

The transistor is arguably the most important invention of the 20th century for everyday life. By replacing the vacuum tube with a tiny, efficient solid-state switch, it made possible portable radios, modern computers, mobile phones and the internet, and it launched the global semiconductor industry.