Executives at Bell Labs had recognized that semiconductors might lead to solid-state alternatives to the electron-tube amplifiers and electromechanical switches employed throughout the nationwide Bell telephone system. In 1936 the new director of research at Bell Labs, Mervin Kelly, began recruiting solid-state physicists. Among his first recruits was William B. Shockley, who proposed a few amplifier designs based on copper-oxide semiconductor materials then used to make diodes. With the help of Walter H. Brattain, an experimental physicist already working at Bell Labs, he even tried to fabricate a prototype device in 1939, but it failed completely. Semiconductor theory could not yet explain exactly what was happening to electrons inside these devices, especially at the interface between copper and its oxide. Compounding the difficulty of any theoretical understanding was the problem of controlling the exact composition of these early semiconductor materials, which were binary combinations of different chemical elements (such as copper and oxygen).

With the close of World War II, Kelly reorganized Bell Labs and created a new solid-state research group headed by Shockley. The postwar search for a solid-state amplifier began in April 1945 with Shockley’s suggestion that silicon and germanium semiconductors could be used to make a field-effect amplifier (see integrated circuit: Field-effect transistors). He reasoned that an electric field from a third electrode could increase the conductivity of a sliver of semiconductor material just beneath it and thereby allow usable current to flow through the sliver. But attempts to fabricate such a device by Brattain and others in Shockley’s group again failed. The following March, John Bardeen, a theoretical physicist whom Shockley had hired for his group, offered a possible explanation. Perhaps electrons drawn to the semiconductor surface by the electric field were blocking the penetration of this field into the bulk material, thereby preventing it from influencing the conductivity.

Bardeen’s conjecture spurred a basic research program at Bell Labs into the behaviour of these “surface-state” electrons. While studying this phenomenon in November 1947, Brattain stumbled upon a way to neutralize their blocking effect and permit the applied field to penetrate deep into the semiconductor material. Working closely together over the next month, Bardeen and Brattain invented the first successful semiconductor amplifier, called the point-contact transistor, on December 16, 1947. Similar to the World War II crystal rectifiers, this weird-looking device had not one but two closely spaced metal wires jabbing into the surface of a semiconductor—in this case, germanium. The input signal on one of these wires (the emitter) boosted the conductivity of the germanium beneath both of them, thus modulating the output signal on the other wire (the collector). Observers present at a demonstration of this device the following week could hear amplified voices in the earphones that it powered. Shockley later called this invention a “magnificent Christmas present” for the farsighted company, which had supported the research program that made this breakthrough.

Not to be outdone by members of his own group, Shockley conceived yet another way to fabricate a semiconductor amplifier the very next month, on January 23, 1948. His junction transistor was basically a three-layer sandwich of germanium or silicon in which the adjacent layers would be doped with different impurities to induce distinct electrical characteristics. An input signal entering the middle layer—the “meat” of the semiconductor sandwich—determined how much current flowed from one end of the device to the other under the influence of an applied voltage. Shockley’s device is often called the bipolar junction transistor because its operation requires that the negatively charged electrons and their positively charged counterparts (the holes corresponding to an absence of electrons in the crystal lattice) coexist briefly in the presence of one another.

The name transistor, a combination of transfer and resistor, was coined for these devices in May 1948 by Bell Labs electrical engineer John Robinson Pierce, who was also a science-fiction author in his spare time. A month later Bell Labs announced the revolutionary invention in a press conference held at its New York City headquarters, heralding Bardeen, Brattain, and Shockley as the three coinventors of the transistor. The three were eventually awarded the Nobel Prize for Physics for their invention.

Although the point-contact transistor was the first transistor invented, it faced a difficult gestation period and was eventually used only in a switch made for the Bell telephone system. Manufacturing them reliably and with uniform operating characteristics proved a daunting problem, largely because of hard-to-control variations in the metal-to-semiconductor point contacts.

Shockley had foreseen these difficulties in the process of conceiving the junction transistor, which he figured would be much easier to manufacture. But it still required more than three years, until mid-1951, to resolve its own development problems. Bell Labs scientists, engineers, and technicians first had to find ways to make ultrapure germanium and silicon, form large crystals of these elements, dope them with narrow layers of the required impurities, and attach delicate wires to these layers to serve as electrodes. In July 1951 Bell Labs announced the successful invention and development of the junction transistor, this time with only Shockley in the spotlight.

Commercialization

Commercial transistors began to roll off production lines during the 1950s, after Bell Labs licensed the technology of their production to other companies, including General Electric, Raytheon, RCA, Sylvania, and Transitron Electronics. Transistors found ready applications in lightweight devices such as hearing aids and portable radios. Texas Instruments Inc., working with the Regency Division of Industrial Development Engineering Associates, manufactured the first transistor radio in late 1954. Selling for $49.95, the Regency TR-1 employed four germanium junction transistors in a multistage amplifier of radio signals. The very next year a new Japanese company, Sony, introduced its own transistor radio and began to corner the market for this and other transistorized consumer electronics.

Transistors also began replacing vacuum tubes in the digital computers manufactured by IBM, Control Data, and other companies. “It seems to me that in these robot brains the transistor is the ideal nerve cell,” Shockley had observed in a 1949 radio interview. “The advantage of the transistor is that it is inherently a small-size and low-power device,” noted Bell Labs circuit engineer Robert Wallace early in the 1950s. “This means you can pack a large number of them in a small space without excessive heat generation and achieve low propagation delays. And that’s what you need for logic applications. The significance of the transistor is not that it can replace the tube but that it can do things the vacuum tube could never do!” After 1955 IBM started purchasing germanium transistors from Texas Instruments to employ in its computer circuits. By the end of the 1950s, bipolar junction transistors had almost completely replaced electron tubes in computer applications.