American Physicists William B. Shockley, Walter H. Brattain, and John Bardeen Produce the First Transistor, Initiating the Semiconductor Revolution

Overview

In 1947 Bell Laboratories scientists invented the transistor—a semiconductor device that could amplify electrical signals transmitted through it. Those in the "know" recognized the significance of the transistor as a compact, reliable replacement for the inefficient vacuum tube; but the development of what many now consider the twentieth century's most important invention, was not prominently reported. Not even the team responsible for the transistor, John Bardeen (1908-1987), Walter Houser Brattain (1902-1987), and William Bradford Shockley (1910-1989), were aware of the singular role their discovery was about to play in initiating the information age and making possible everything from miniature hearing aids to high-speed computers.

Background

In 1904 John A. Fleming (1849-1945) developed a vacuum tube diode—known as a "valve" because it forced current within the tube to flow in one direction. This was essential for converting alternating currents to direct current. In 1907 Lee De Forest (1873-1961) patented the Audion vacuum tube, which functioned as a valve as well as amplifying current. De Forest achieved amplification by inserting a metal grid into the tube. Varying grid input current allowed him to control the flow of a secondary current in the tube such that weak grid inputs resulted in strong secondary currents. De Forest conceived of cascading Audions to provide the necessary amplification for long-distance broadcasts and signal reception.

About this time Alexander Graham Bell's (1847-1922) telephone patents began expiring. AT&T sought an advantage over the competition. Their solution was to develop a transcontinental telephone service. They bought De Forest's Audion patent hoping it could be adapted for amplifying signals along telephone lines. AT&T's design improvements yielded an acceptable device that was quickly deployed and thus formed the basis not only for early long-distance telephony but also radio, radar, and computers.

AT&T's improved Audion was far from ideal, however. It proved extremely unreliable, consumed too much power, and generated a great deal of heat. By the 1930s Mervin Kelly, then Bell Labs' research director, believed a better solid-state device could be produced. He pushed research on semiconductors, which he thought might provide the answer.

Semiconductors possess the unusual property that their conductivity varies. They are composed mostly of atoms that do not conduct electricity. However, they do possess small numbers of conducting atoms. Depending on how they are handled, semiconductors can be made to conduct more or less electricity. It was hoped that this property could be exploited to produce an alternative to vacuum tubes.

After World War II Kelley assembled a team of Bell Lab scientists to develop a solid-state semiconductor switch to replace vacuum tubes. Shockley was the team leader. The team also included the experimentalist Walter Brattain and theorist John Bardeen.

In 1945 Shockley designed a semiconductor amplifier based on the field effect. The underlying principle here is that an electric field applied through a semiconductor surface should alter the charge density within and thus change the semiconductor conductivity. Shockley's field-effect mechanism, however, failed to amplify electric currents. He assigned Bardeen and Brattain the task of determining why.

Bardeen and Brattain developed a close working relationship. Bardeen suggested experiments and interpreted results while Brattain built the equipment and ran experiments. Their work went largely unsupervised, but by 1947 they were becoming increasingly frustrated by their inability to produce positive results. It was then that Bardeen had his historic insight that surface charges were hindering electric field penetration of semiconductor material.

Without notifying Shockley, they produced a new device having closely spaced contacts lightly touching the semiconductor surface. According to Bardeen's calculations, current coming through one contact should produce a stronger current at the other contact. On December 16, 1947, they successfully tested their device—the first point-contact transistor (from transfer resistor).

Chagrined at having had no direct role in the crucial breakthrough, Shockley conceived the junction transistor shortly thereafter. Special semiconductor crystals were necessary to make this device work. These were produced by Gordon Teal, and the first junction transistor was successfully tested on April 12, 1950. Shockley's device was more durable and practical than the point-contact transistor. It also proved much easier to manufacture and became the most commonly used transistor until the late 1960s.

Impact

On June 30, 1948, Bell Labs held a press conference to announce the invention, but the story gained only limited coverage and generated little excitement. This initial indifference quickly dissipated as appreciation of transistor applications grew. However, even after Bardeen, Brattain, and Shockley were jointly awarded the 1956 Noble Prize in physics for their work, no one fully grasped how profound and pervasive the transistor's influence would be.

Transistors were initially used as telephone amplifiers, in operator headsets, and for hearing aids. Widespread commercial use in the early 1950s was limited, though, because transistors were too expensive, unreliable, and difficult to manufacture. Considerable effort was devoted to developing more functional and efficient transistors that could be cheaply and reliably mass produced, but most U.S. companies concentrated on defense applications for which miniaturization was desirable and high costs were not prohibitive.

As transistors became faster, more reliable, and cheaper, the commercial market began to look more attractive. Texas Instruments decided to probe this market by producing a hand-held radio. Released in October 1954, the 5-inch (12.7-cm) Regency was the first transistorized radio and popular culture's first highly visible exposure to the new technology. When the Regency was discontinued the next year, Sony Electronics was poised to capture the market. Their mass-produced miniature, transistorized radios made it possible for information to be quickly disseminated across the globe. Sony also developed the first transistorized TV and went onto manufacture VCRs, stereo equipment, and computer games.

In 1955 Shockley left Bell Labs to found Shockley Transistor Corporation in Palo Alto, California. This was the first semiconductor company established in what is now called Silicon Valley. The research team Shockley assembled consisted of some of the brightest people working in the field, but within a year Shockley's domineering managerial style drove eight of his best employees away. Known as the "traitorous eight," they went on to found Fairchild Semiconductor (1957).

In 1959 Fairchild Semiconductor and Texas Instruments independently developed integrated circuits—single "chips" of semiconducting material that combine the functions of transistors and other electronic components. Instead of making separate bulky components and inefficiently wiring them together, it was now possible to produce complete miniaturized circuits from single crystals. Texas Instruments exploited these chips to produce the first hand-held calculator in 1967, and Intel was founded in 1968 to produce computer memory chips. Today centimeter-by-centimeter chips contain millions of transistors.

Integrated circuits were a significant advance but limited precisely because they were wired to perform specific functions. This was overcome by Intel's microprocessor.

In 1969 the Japanese calculator firm Busicom contracted Intel to produce chips for a new calculator. Intel assigned Ted Hoff (1937- ) to work on the contract. After reviewing Busicom's 12-chip calculator design, Hoff proposed an alternate single-chip architecture that combined the separate functions into a 2300-transistor CPU. The resulting 4004-chip was announced to the industry in November 1971.

The 4004 was more than a complicated integrated circuit; it was programmable. Calculator chips perform specific functions that are hard-wired in during the manufacturing stage. Such chips can perform these and only these functions. The functions wired into the 4004 were different. They could be combined and controlled in various ways by user-implemented programs, thus making the microprocessor more versatile.

Microprocessors were quickly incorporated into computers. Measuring only 1/8 inch by 1/16 inch (0.3 × 0.16 cm), the 4004 was as powerful as the 30-ton ENIAC computer built in 1946. Such a reduction of size made possible super computers and initiated the PC revolution. Today Intel is the world's leading manufacturer of computer chips, with 1999 revenues over $21 billion.

Often referred to as the "invisible technology," transistors are at the core of every major information age innovation. Computer networks and the internet as well as satellite communication would all be impossible without transistors. Miniaturization has also made possible pacemakers, cellular phones, remote controls, fuel injection, compact disc players, video games, smoke detectors, photocopiers, global positioning, and literally thousands more devices and applications.

As miniaturization continues further, applications will be found, but miniaturization cannot continue indefinitely. Fundamental physical principles—the size of the atom and electron—place ultimate constraints on the size of transistors. Nevertheless, much progress is to be expected, and it is likely that by 2010 transistors will be twice as small as they are now, opening up new applications possibilities.

STEPHEN D. NORTON

Further Reading

Books

Eckert, Michael, and Helmut Schubert. Crystals, Electrons, Transistors: From Scholar's Study to Industrial Research. New York: American Institute of Physics, 1990.

Hoddeson, Lillian, E. Braun, J. Teichmann, and S. Weart. Out of the Crystal Maze: Chapters from the History of Solid State Physics. New York: Oxford University Press, 1992.

Queisser, H. The Conquest of the Microchip: Science and Business in the Silicon Age. Cambridge: Harvard University Press, 1988.

Riordan, Michael, and Lillian Hoddeson. Crystal Fire: The Birth of the Information Age. New York: W. W. Norton & Company, 1997.

Seitz, Frederick, and Norman G. Einspruch. Electronic Genie: The Tangled History of Silicon. Urbana, IL: University of Illinois Press, 1998.

Periodical Articles

Bardeen, John, and Walter H. Brattain. "The Transistor: A Semiconductor Triode." Physical Review 74 (1948): 230-231.

Bardeen, John, and Walter H. Brattain. "The Physical Principles Involved in Transistor Action." Physical Review 13 (1949): 1208-1225.

Brattain, Walter H., and John Bardeen. "Nature of the Forward Current in Germanium Point Contacts." Physical Review 74 (1948): 231-232.

Kilby, J. "Invention of the Integrated Circuit." IEEE Transactions on Electron Devices ED-23 (1976): 648-654.

Shockley, William. "The Theory of P-N Junctions in Semiconductors and P-N Junction Transistors." Bell System Technical Journal 28 (1949): 435-489.