Transistors

A valve for electricity

A diode is a one-way valve for current. A transistor goes further: it is a valve that can be carefully opened and closed to control the current passing through it.^[1] A small current at one terminal controls a much larger current through the other two.

Inside a junction transistor

The classic junction transistor is a sandwich of three layers of doped semiconductor:^[1]

• an n-type layer called the emitter;
• a thin p-type layer in the middle, called the base;
• another n-type layer, called the collector.

When the base is connected to a positive terminal, a small current of electrons, the base current, flows through it. This causes a large current, the collector current, to flow through the transistor. Adjusting the small base current controls the large collector current.^[1]

The math (optional)

The two currents are linked by the transistor’s current gain, β (beta):

I_C = β × I_B^[1]

For example, if a transistor had a gain of 100, a base current of 1 milliamp would allow a collector current of 100 milliamps.

Invention at Bell Labs

On 16 December 1947, at Bell Labs, John Bardeen and Walter Brattain achieved transistor action with the point-contact transistor, and a week later, on 23 December, they demonstrated it to lab officials. It was a small slab of high-purity germanium touched by two closely spaced gold contacts.^[2] The goal was to replace the unreliable vacuum tubes and electromechanical switches used in the Bell Telephone System, and a Bell Labs spokesman said the device “may have far-reaching significance in electronics and electrical communication”.^[2]

William Shockley, John Bardeen and Walter Brattain are credited together with inventing the first working transistor.^[1] The p-n junction, discovered in 1940, was the foundation for the junction transistor that Shockley conceived in 1948.^[4]

The MOS transistor

In 1959, M. M. (John) Atalla and Dawon Kahng at Bell Labs made the first successful transistor of a different design, the insulated-gate field-effect transistor, better known as the metal-oxide-semiconductor (MOS) transistor.^[3] MOS transistors are smaller and consume less power than the earlier bipolar kind, and they reached the market in 1964. Today over 99 percent of microchips use them.^[3] MOS technology is what made it possible to build dense integrated circuits.^[1]

From one transistor to vast numbers

Transistors can be miniaturized and packed into high-density integrated circuits, with vast numbers of them on a single piece of silicon.^[1]

• 1958: Jack Kilby at Texas Instruments etched transistor, capacitor and resistor elements from germanium slices and connected them with fine gold wires into a working oscillator, an early “solid circuit”.^[5]
• 1965: Gordon Moore predicted that the number of components per chip would double every year. In 1975 he revised this to every two years.^[6]
• 1971: Intel’s 4004 microprocessor held 2,300 transistors.^[7] It was designed by Ted Hoff, Federico Faggin and Stan Mazor at Intel, with Masatoshi Shima of the Japanese calculator company Busicom.^[8]

Common misconceptions

“The transistor was invented for computers.” It was developed at Bell Labs to replace vacuum tubes and electromechanical switches in the telephone system.^[2] Computers came later.

“A transistor is just an on/off switch.” It can switch, but a small base current can also be adjusted to control the collector current smoothly.^[1]

Going further

• The Computer History Museum’s Silicon Engine timeline, listed below, tells the story year by year.
• Microcontrollers shows what happens when a whole computer’s worth of transistors fits on one chip.