Newtonpedia ES
Home › Technology and communications › Field-effect transistor
Technology and communications · 1925

Field-effect transistor

A transistor that controls the current through a semiconductor channel using an electric field applied by a gate, without injecting charge carriers as bipolar transistors do.

Year
1925
People involved
Julius Edgar Lilienfeld
Section
Technology and communications

History

The field-effect transistor, or FET, is an electronic component that regulates the flow of current through a semiconductor, a material that conducts electricity only under certain conditions. It works like an electrical tap: a voltage applied to its control terminal, called the gate, changes the width of the path that carriers travel through, and with it the amount of current that flows. Its importance is hard to overstate, because today's computer chips, phones and most modern electronic devices rest on this idea.

The concept predates any working device. The Austro-Hungarian physicist Julius Edgar Lilienfeld applied for a Canadian patent in 1925 for a method and apparatus for controlling electric currents, and it is regarded as the ancestor of today's FETs. He later filed US patents in 1926 and 1928, but he published no research and his patents describe no working prototype. The main obstacle was material quality: high-grade semiconductors were not yet available, so his amplifiers found no practical use in the 1920s and 1930s. Oskar Heil proposed a similar concept in 1934, with equally little practical result.

In the 1940s, the Bell Labs team in the United States was looking for a solid-state amplifier that could replace vacuum tubes. William Shockley envisioned a FET in 1945 but could not build one. John Bardeen blamed the failure on surface states, trapped charges at the semiconductor's surface that block the external field. While trying to get around this problem, Walter Brattain and Bardeen stumbled upon the point-contact transistor in 1947, the first working transistor, which led to Shockley's bipolar junction transistor in 1948. That same year Bardeen patented the theoretical basis of what would become the MOSFET, though he did not build one. Meanwhile, Heinrich Welker had patented a junction FET in 1945, and in Europe Herbert Mataré and Welker developed a point-contact device in 1948 while working for a French subsidiary of Westinghouse, which put it into production as the transistron for the French telephone network.

The first FET to work in practice was the junction FET, or JFET. Shockley published a theoretical treatment of it in 1952, and in 1953 George Clement Dacey and Ian Ross at Bell Labs built a working one. Junction transistors, however, were bulky and hard to mass-produce, and the insulated-gate design that could have solved this was still held back by surface states. By the mid-1950s many researchers had dropped the FET idea in favour of bipolar transistors.

The breakthrough came from work on silicon. In 1955, Carl Frosch and Lincoln Derrick, also at Bell Labs, accidentally coated a silicon wafer with a layer of silicon dioxide. They found that the layer protected the surface and blocked some impurities from entering while letting others through. This oxide masking technique allowed engineers to etch tiny openings and introduce impurities only in chosen spots. In 1959 Mohamed Atalla and Dawon Kahng proposed a silicon MOS transistor, and in 1960 their team demonstrated a working device: the first MOSFET, the oxide-insulated FET type that now dominates the market.

The MOSFET won out because it was far smaller, used less power and allowed many devices on a single chip. In 1963, Chih-Tang Sah and Frank Wanlass at Fairchild Semiconductor developed CMOS, which pairs p-channel and n-channel transistors so that, at any moment, one is on while the other is off. This arrangement cuts power consumption and became the foundation of digital integrated circuits. Since then, manufacturing has shrunk transistors to a few nanometres, allowing billions on one chip.

Today the FET, and above all the MOSFET, appears in processors, memory, sensors, audio amplifiers and power systems. Its ability to switch signals quickly while drawing very little control current makes it central to digital electronics and communications. What began as a patent without a prototype has become the most manufactured component in history.

Timeline

  • 1925Lilienfeld files a Canadian patent for a method of controlling electric currents, the ancestor of the FET.
  • 1947Bardeen and Brattain discover the point-contact transistor while seeking a FET.
  • 1953Dacey and Ross build a working JFET at Bell Labs.
  • 1955Frosch and Derrick accidentally discover the silicon dioxide protective layer.
  • 1960Atalla and Kahng build the first working MOSFET at Bell Labs.
  • 1963Sah and Wanlass develop the CMOS process at Fairchild Semiconductor.

Trivia and anecdotes

A transistor without a prototype

Lilienfeld's patents describe no specific example of a device that actually worked.

A discovery by accident

Brattain and Bardeen found the point-contact transistor while failing to build a FET with germanium.

People involved

Other milestones in Technology and communications