Solid State New Magic Word.
Bell Labs gets credit for the transistor, Texas Instruments for the integrated circuit (1958), and Signetics for the linear integrated circuit (1964) . . . all of which added up to the magic of "solid state" during these two dynamic decades.
Solid-state, simply stated, describes a device, circuit or system whose operation in dependent upon any combination of optical, electrical or magnetics phenomena within a solid.
Inventions are sometimes thought to result from unexpected discoveries on the part of isolated researchers. Not so with the transistor. The transistor was the culmination of a directed interdisciplinary research effort at Bell Laboratories to find a solid-state amplifier and switch.
Prior to World War II, Dr. Mervin Kelly, director of research at Bell Labs, became aware of the constraints vacuum-tube technology would impose on the telephone industry's requirements 10 to 15 years hence.
The industry, Kelly felt, would be in trouble because of the inherent physical limitations of tubes and mechanical relays. Tubes were inefficient consumers of huge amounts of power. They were delicate, expensive to manufacturer, and had limited life. Relays, though inexpensive and reliable, operated slowly. And tubes and relays occupied a lot of space.
Clearly, new techniques and components had to be developed to satisfy the communications needs of the future. The most promising field to explore was research on semiconductors--materials whose electrical conductivity is between that of a conductor and an insulator. Bell Labs decided to explore the behavior of electrons in such solid materials.
World War II interrupted most of this work. But the war stimulated research at Bell Labs, Purdue University, and in England to satisfy the need for good silicon and germanium detectors for radar. Understanding these two semiconductors contributed greatly to the invention of the transistor.
In 1945, Dr. Mervin J. Kelly authorized a major them effort in solid-state physics, covering the fundamental investigation of conductors, semiconductors, dielectrics, insulators, piezoelectric, and magnetic materials. John Bardeen, Walter Brattain, and William Shockley, and many other Bell Labs scientists resumed full-time semiconductor research. Under the direction of William Shockley, the team's goal was to produce a device the would perform better in a solid what the electron tube does in a vacuum . . . conduct, modulate, and amplify electrical signals.
Research was centered on germanium and silicon with enphasis on understanding both chemical and physical properties. Experiments led to new theories. Shockley, for example, proposed a field-effect structure for a semiconductor amplifier. Performance was less than that anticipated in the theory. Bardeen then suggested extending the theory to include surface effects to explain why the device didn't work as expected.
Further advances required a better understanding of the basic physics--particularly of the regions near the surfaces of the materials. In this investigation of surface properties, the phenomenon later called the "transistor effect" was observed. (In naming the transistor, John Pierce took into consideration that the resistance of one point contact was found to depend on the current flowing through the other point contact. In other words, the resistance effect was transferable from one point to the other. The device was a "transfer resistor"--telescoped into "transistor".)
On December 23, 1947, John Bardeen, who had worked on theory, and Walter Brattain, who had worked on surface properties, demonstrated a device that amplified a speech signal 40 times. The device was called a point-contact transistor because it consisted of two pointed gold contacts, less than two thousandths of an inch apart, on one side of a piece germanium wafer.
In 1956, Bardeen, Brattain, and Shockley were awarded the Nobel prize for discovering the transistor effect.
Specificfally excluded from the marvelous world of "solid states" are devices, circuits or systems dependent on even macroscopic physical movement, rotation, contact or non-contact of any combination of solids, liquids, gases or plasmas.
Other solid-state developments followed the invention of the point-contact transistor in 1947.
The junction transistor was proposed by W. Shockley in a broad patent application in mid 1948. The junction transistor eventually became the dominant form of bipolar transistor. It employs a single crystal of a semiconductor, such as germanium or silicon, in which two closely spaced pn junctions are formed. These junctions are provided with an emitter, base, and collection electrode. The earliest practical form was the grown junction transistor in 1950, which was followed in 1952 by the alloy junction transistor, the diffused-base transistor (1955), and the epitaxial transistor (1960).
Into the mainstream of solid-state development came . . . in 1958 . . . the ferreed switch and the Read diode. The ferreed switch is a sealed electromechanical device used in electronic switching systems. It uses bistable magnetic materials to switch states rapidly and to maintain switch contacts in either state indefinitely. The name "ferreed" came from early models in which a ferrite was used as the bistable magnetic material, with a magnetic reed switch, sealed in glass, serving as the output contacts. The Read diode is a member of a larger class of oscillators, referred to as avalanche diodes, ATT (avalanche transit time) diodes, or IMPATT diodes which were introduced in 1965. The Read diode, proposed in 1958, was based on the knowledge that a sufficiently high voltage applied across a diode causes the diode to break down electrically in one region producing a charge multiplication. This effect, together with the time it takes the charge to move through the rest of the diode, results in a net negative resistance at high frequencies.
A major development was the epitaxial film technique for fabricating transistors, a thin layer of a semiconductor material, such as silicon, is deposited on a single-crystal wafer. As the deposited layer grows, it develops the same crystal arrangement as the base wafer; hence, the term "epitaxial," which is derived from two Greek words meaning "on" and "arrangement." Both the base and emitter are formed by introducing impurities into the film, which itself serves as the collector. By varying the thickness and resistivity of the grown layer, the desired characteristics can be introduced into the transistor or other device. With this process, the concentration of impurities in the substrate crystal and in the grown film can be closely controlled. Silicon epitaxy is vital for bipolar silicon integrated-circuit fabrication.
Light-emitting diodes were developed in the 1960s at Bell Labs and elsewhere as a result of studies of the physics of the electron-hole recombination in semiconductor devices. In this process, portion of the recombination energy is converted to light. By the late 1960s, development work was aimed at incorporating LEDs in a wide variety of indicator and illuminator applications, such as in pushbuttons on key telephones.
In 1963 "solid state" began to take on real significance with the development of beam-lead, sealed-junction integrated circuits which were integrated mechanically and electrically and could be bonded to a thin-film substrate in a single operation that attaches all external electrical contacts. A metal-insulator-silicon system of materials provides the intraconnections between circuit elements, the leads for external connections, and a junction-seal encapsulation that replaces the vacuum-tight encapsulation used in previous integrated circuits. At about the same time supercurrent junctions were found to exhibit the Josephson effect, namely, the flow of an electron-pair tunneling current through an insultor separating two superconductors.
In 1966 silicon gate technology saw development of the metal-oxide semiconductor integrated circuits making it possible to fabricate a self-aligned MOS transistor structure. The polysilicon gate is used as a diffusion mask for both the source and drain regions; therefore these regions are self-aligned. The polysilicon used for the gate electrode is also used for the interconnection between devices, leading to a two-level interconnect capability. The results are devices that have improved performance and increased packing density. Silicon gate technology, along with an earlier development, oxide masking, contributed to the ability of industry to design large-scale integration circuits on relatively small chips.
Solid-state crosspoint arrays were first fabricated in 1967 although the use of solid-state four-layer devices (a pnpn transistor switch, for talking-path switching) had been suggested at Bell Labs in 1956. In the period following the initial suggestion, batch fabrication of an active switching array became possible, and a shorted emitter thyristor design facilitated controlling device parameters. In the late 1960s, a 32-by-32 line-pair matrix was built and met the technical requirements for both voiceband and "Picturephone" switching. It was not until 1976, h owever, that economical application of solid-state crosspoint arrays as switching devices for electronic switching systems was achieved. In 1977 a small-size pnpn integrated circuit array with 32 switches was packaged and incorporated in Bell's "Horizon" business communications system.
Solid-state devices have had animpact on a multitude of markets including military, commercial equipment and systems, medicine, education, broadcasting, musical instruments, automobiles, navigation, test and measuring instruments, industrial control and automation, aerospace, and telephone communications.
And, sensational tho the solid-state achievements of the past two decades have been, even greater things are still to come.
Dr. Frank Sewell, director of the Sperry Semiconductor Laboratory, says: "High speed, micro computers of the future will be developed by using superconductors instead of semiconductors. Some recent developments and innovations at Sperry relating to Josephson, gallium arsenide, and silicon devices are of interest because of their capability for logic and memory for high speed computers."
According to the latest issue of the "Henderson Electronic Market Forecast," worldwide production of semiconductor products will exceed $32 billion in 1984. This will represent a 36 percent increase over the $23.9 billion output chalked up during 1983. The forecast calls for another strong 21.9 percent advance during 1985.
The forecast, published by Henderson Ventures, a Los Altos, California consulting and market research firm, points out that North America and Western Europe will continuously lose production share to Far Eastern semiconductor producers. During 1984, American production will grow by 34.1 percent, while Western Europe advances by only 23.9 percent. In contrast, Japan's output will soar by 41.8 percent his year and the rest of the world will leap ahead by 64.4 percent.
Says Ed Henderson, president of Henderson Ventures: "Hong Kong, Taiwan and Korea are dramatically increasing their financial commitment to semiconductor technology. And South Korea has made it a strategic economic investment for their country." As a result, ROW production will advance from a slight $329 million in 1982 to four times that size in 1985, or $1.4 billion.
|Printer friendly Cite/link Email Feedback|
|Date:||Sep 1, 1984|
|Previous Article:||Satellites Get the Top Billing.|
|Next Article:||Electronic Switching Is 'In'.|
|Vermont's small breweries are big business.|
|Magic words to obtain confessions.|
|Magic of legal language.|