The most detailed history of the world's integrated circuit development

Preface

• Foreword The first transistor in history was born 60 years ago-December 16, 1947 in Bell Laboratories, New Jersey (Bell
  Laboratories. Inventor William Shockley, John
  Bardeert and Walter Brae Dayton (Walter Brattain do this to get the 1956 Nobel Prize in physics
  after the solid-state semiconductor (solid-state) of the invention makes possible the invention of the integrated circuit. this outstanding achievement has laid the foundation for the development of the world semiconductor industry. after In the past 60 years, the development of semiconductor technology has greatly increased labor productivity, promoted the development of the world economy, and improved people's living standards. George
  Scalise , President of the American Semiconductor Association (SIA ) once said:
  " The invention of the transistor 60 years ago brought tremendous changes to this ever-evolving world. The significance of this historic landmark invention cannot be underestimated. The quality tube is a key component of countless electronic products, and these electronic products It has brought revolutionary changes to almost every aspect of human life. In 2007, the world’s microelectronics industry produced 900 million quality tubes for every man, woman, and child on the planet—a total of 600000 billion 100 million industries have sales of more than 257 billion US dollars.”
  Looking back at the invention of the transistor and the development of the integrated circuit industry, we can see that
  the invention of the transistor 60 years ago was not an accident, it was the result of professional and technical talents in the world. Under the hard work, in an environment that encourages bold innovation, it is produced under the government's policy to encourage investment in research and development. At the same time, we can also see that the integrated circuit industry has grown from scratch and developed rapidly as a result of mutual cooperation and joint innovation among the entire industry. result.

Discover and study semiconductor effects

• In 1833, when British physicist Michael
  Faraday was studying the conductivity of silver sulfide crystals, he discovered that the conductivity of silver sulfide crystals increased with the increase of temperature, which was a "special phenomenon". This feature is just the opposite of the situation with copper and other metals.
  The discovery of Michael Faraday (Michael Faraday) made people begin to understand the semiconductor effect. In 1874, the German physicist Ferdinand Braun discovered that the current can only pass through the contact point of the metal probe and the galena crystal in one direction while studying the conductive properties of the crystal and electrolyte. Ferdinand
  Braun recorded and described the "contact rectification effect" of this semiconductor diode. Based on this discovery, Jagadis
  Chandra Bose, professor of physics at the Presidential College of the University of Calcutta in India , proposed the use of "semiconductor crystal rectifiers" as applications for detecting radio waves and applied for a patent (1901).
  When studying the characteristics of copper sulfide semiconductors, Julius Lilienfeld, an American physicist born in Poland, envisioned a three-pole semiconductor device "field effect transistor", and in 1926 submitted a patent for a three-pole amplifier based on the characteristics of copper sulfide semiconductors. . In the following decades, people have been trying to make such devices.
  The discovery of semiconductor physics has inspired people to study its theory. In 1931, Alan Wilson (Alan Wilson), a physicist at the University of Cambridge in the United Kingdom who was doing research in Germany at the time, published the idea of ​​using quantum mechanics to explain the basic properties of semiconductors and published "Semiconductor Electronics Theory". Seven years later, Boris (Boris
  Davydoy (Soviet Union), Nevill Mot, UK) and Walter Schottky
  (Germany) also independently explained the characteristics of semiconductor rectification. In the mid-1930s, Russell, an electrochemist at Bell Laboratories, USA
  (Oh) When studying the use of silicon rectifiers to detect radar signals, it was found that the ability of silicon rectifiers to detect signals increased with the improvement of silicon crystal purity. Moreover, in an experiment in February 1940, Russell
  Oh When testing a piece of silicon crystal. It was surprisingly found that the current increases when the silicon crystal is exposed to strong light. On the basis of this discovery, Russell
  Oh proposed the concept of pn junction and the theory of silicon photoelectric effect. This discovery led to the development of junction transistors and solar cells in the future.

1833-The semiconductor effect was recorded for the first time
1874-the discovery of the semiconductor point-contact rectification effect
1901-Semiconductor rectifier applied for "whisker" detector patent
1926-Field-effect semiconductor device concept applied for patent
1931-published "Semiconductor Electronics Theory"
1940- p-皓Discovery

The invention of the "contact" transistor-December 16, 1947

• In early 1945, William Shockley
  organized a solid-state physics research group at Bell Labs in the United States. In addition to other research, this research group is also conducting research on trying to replace the less rugged vacuum tubes with semiconductors and electromechanical switches used in Bell telephone systems. In April of the same year, William Shockley conceived a "field effect" amplifier and switch based on the fault and silicon technology developed a few years ago, but the experiment was not successful. A year later, the theoretical physicist John
  Bardeen pointed out that electrons on the surface of the semiconductor may prevent the electric field from penetrating into the material, thereby cancelling any effects. John
  Bardeen (ohn Bardeen) and experimental physicist Walter
  Brattain (Walter Brattain) began to study the properties of these "surface states". On December 16, 1947, their research led to the success of the first semiconductor amplifier. John Barding and Walter Bratton used a plastic wedge to fix two gold contact points very close to a small high-purity surface. The voltage of one contact point modulates the current flow to the other point. The input signal is amplified to 100 times.

The origin of the transistor name

• Every new invention needs a name. Bell Labs originally envisioned several, including "semiconductor triode vacuum tube", "solid state triode vacuum tube", "surface state triode vacuum tube",
  "crystal triode vacuum tube", " lotatron" etc. But in the end, the
  term "transistor" proposed by John Pierce was adopted . John Pierce (John Pierce) recalled:
  "The reason why I proposed this name focused on what the device is for. At that time, it should have been a copy of the tube. The tube has transconductance, and the transistor should have transimpedance. In addition, the name of this device should match the names of other devices such as varistors, thermistors, and so on, so I suggest using the name "transistor".
  William Shockey, John Bardeer and Walter
  Brattain (Walter Brattain for the invention of the transistor) jointly won the Nobel Prize in Physics in 1956. Later, John Bardeen was hired as Erie. Professor at the University
  of
  llinois; Walter Brattain at Whitman College in Walla Walla, Washington, Harvard
  University, University of Minnesoa and University of
  Washington served as a lecturer; while William Hockley served as
  a professor in the School of Electrical Engineering at Stanford University, he also established the Shockley Semiconductor Laboratory (Beckman Instruments Co., Ltd.) Part), made outstanding contributions to the birth of Silicon Valley and the electronics industry, the inventor of integrated circuits,

  

The birth of junction transistors-1948

• Due to the fragile mechanical structure of "contact" transistors, production and application are restricted. Willim
  Shockley of Bell Labs in the United States began to envisage the development of a new transistor structure using the pn junction effect. Based on the understanding of the pn junction effect theory, William
  Shockley proposed that positively charged "holes" not only move on the surface, they should also be able to penetrate and pass through the wrong crystal, which is called "minority carrier injection" "This idea is the key to the realization of junction transistors. Based on this concept, junction transistors can be composed of materials in three regions:
  n-type/pn junction/p-type. The experiment was successful in 1948, William Shockley applied for a patent and published the result in 1949.

  The production of junction transistors requires the use of large single crystals. Gordon Teal, a chemist at Bell Labs in the United States, used the wrong "seeds" to stretch the single crystals from the melted faults and used simple equipment to grow large single crystals. Crystal fault. In this process, Gordon
  Teal and Morgan
  Sparks discovered that the pn junction can be obtained by "doping" in the melting fault, and they collaborated to fabricate an npn junction transistor. This "grown pn junction transistor" has excellent characteristics.

  

Bell Labs Licensed Transistor Technology-1952

• From the late 1940s to the 1950s, in order to promote the development of transistors and other solid-state devices, Bell Labs of the United States carried out a "basic research and development" project for semiconductor technology. Under
  the leadership of electrical engineer Jack Morton , this project developed the technology for "regional purification" and the growth of bulk monocrystalline threads and monocrystalline silicon. The laboratory has also developed manufacturing technologies for forming Pn junctions, semiconductor surface treatment, and fixing metal connections. At the same time, the laboratory has also developed logic circuits and equipment using transistors.
  Jack Morton (ack
  Morton sang that Bell Labs should share this transistor technology with other researchers and companies, because Bell Labs and its parent company AT&T can benefit from the technological advances of others. Therefore, in the 20th century They hosted three seminars in the 1950s, allowing other scientists and engineers to visit Bell Labs to learn first-hand information about this new semiconductor technology. Among them, the first meeting held in 1951 was dedicated to defense-oriented applications.

  

The invention of the "integrated circuit"

• Jack Kilby Demonstrates "Solid State Circuits"-1958

  When electronic devices become complex, people begin to find relatively simple ways to connect thousands of transistors, resistors, capacitors, etc. together. From 1952 to 1957, scientists in Britain, Japan, and the United States all tried differently. These early "integrations" did not provide a "connection" method that can be widely used.
  In December 1958, Texas Instruments' ack
  Kilby used etching to form transistor, capacitor, and resistor regions on a pnp transistor wafer with a staggered mesa, and used thin gold wires to connect these regions to Demonstrate the function of an oscillator point. -After a week, he made another amplifier in the same way. Texas Instruments announced the concept of "solid-state circuits" in March 1959 and introduced its first commercial devices, including binary flip-flops, in March 1960. However,
  "fine gold wire" is not a practical main production method, nor can it solve the problem of identifying "thread ends". This method was later replaced by
  the "metal vapor deposition connection" method invented by Robert Noyce.

  The invention of planar technology led to the invention of monolithic integrated circuits-1959

  In 1959, Lean
  Hoerni , a physicist at Fairchild Semiconductor, invented the planar process in order to solve the reliability problem of mesa transistors. The key to this process is to use an oxide layer to protect the surface of the pn junction. Infected. This invention revolutionized semiconductor production. The devices manufactured by the planar process not only show better electrical performance-the use of an oxide protective layer can significantly reduce the leakage current, which is extremely critical for the logic design of the computer. It also makes it possible to manufacture all the components of an integrated circuit from only one side of the wafer.
  In order to further develop other uses for the planar process, Robert
  Noyce , the co-founder of Fairchild Semiconductor, conceived the idea of ​​making monolithic integrated circuits. The electrodes, transistors, resistors, and capacitors scattered on the silicon surface are connected to each other by evaporating aluminum metal wires on the protective oxide layer. In this way, people can make complete circuits on a single silicon chip. The use of "evaporated aluminum metal wires" to replace "thin gold wires"
  provided a practical method for Jack Kiby's solid-state circuits. Robert
  Noyce applied for it in 1959. A patent for a mass-produced monolithic integrated circuit structure. Subsequently, one of the founders of Fairchild Semiconductor, Jay (lay
  Last), succeeded in 1960 based on Holney's planar process and the method of Noyce monolithic integrated circuit structure. Developed the first commercial integrated circuit-a bistable logic circuit consisting of 4 transistors and 5 resistors.

Manufactured the first metal oxide semiconductor (MOS) insulated gate field effect transistor-1960

• In 1959, John
  Atalla and Dawon Kahng of Bell Laboratories in the United States developed the first insulated gate field effect transistor (FET). Their success factor was to control the influence of the "surface state" Allows the electric field to penetrate the semiconductor material.

  In the process of studying the thermally grown silicon oxide layer, they found that in the structure of the metal layer (M-gate), oxide layer (0-insulation) and silicon layer (S-semiconductor), these "surface states" will be in the silicon and The junction of oxides is greatly reduced. In this way, the electric field applied to the gate can affect the silicon layer through the oxide layer, which is the origin of the MOS name. Due to the slow speed of the original MOS devices and the inability to solve the problems faced by telephone equipment, this research was stopped.
  But researchers at Fairchild and RCA recognized the advantages of MOS devices. In the 1960s, Karl
  Zaininger and Charles (CharlesMeuller) manufactured metal oxide semiconductor transistors at Radio America (RCA). Fairchild’s CT
  Sa has manufactured a MOS tetrode with a control electrode. Subsequently, MOS transistors began to be used in the development of integrated circuit devices. In 1962, Fred
  Heimar and Steven
  Hofstein made an experimental single-chip integrated circuit device composed of 16 transistors at Radio America (RCA). Inventing the complementary MOS circuit structure-
  In 1963, a paper by the author for the Fairchild R&D laboratory CT
  Sah and Frank Wanlass (Frank Wanlass) showed that when the P-channel and the channel MOS are connected in a complementary symmetrical circuit configuration When transistors form a logic circuit (called CMOS today), the power consumption of this circuit is almost zero. Frank Harrison applied for a patent for this invention. CMOS technology has laid the foundation for low-power integrated circuits and has become today's mainstream digital integrated circuit production technology.

The integrated circuit industry has entered a period of development

1963-Development of standard logic integrated circuit series
Speed, cost and density advantages established the transistor-transistor logic (TTL) integrated circuit to become the most popular standard logic building block in the 1960s and 1970s

1964-Hybrid microcircuit reaches its peak
production. The multi-chip SLT packaging technology developed for the IBM system/360 computer series enters mass production.

1964—The birth of the first commercial MOS integrated circuit
General Microelectronics used metal oxide semiconductor (MOS) technology to achieve a higher level of integration than bipolar integrated circuits and used this technology to manufacture the first calculator chipset.

1964—The birth of the first widely used analog integrated circuit
David Talber and Robert Widlar of Fairchild Semiconductor successfully developed commercial analog integrated circuits, which opened up an important application area.

1965-Package design suitable for system integration The
dual in-line housing (DIP) format greatly facilitates printed circuit board wiring and reduces computer assembly costs.

1965—Read-only storage was the first dedicated storage IC storage
Programmable read-only memory (ROMs) produced the first random-read storage integrated circuit.

1966-The semiconductor RAMs
16-bit bipolar transistor device developed for high-speed storage is the first dedicated integrated circuit for high-speed reading and writing storage applications.

1968-Current source integrated circuits
with integrated data conversion functions. High-precision manufacturing processes integrate analog and digital circuits into a single chip.

1968 -
Federico Faggin and Tom Klein , the silicon gate technology developed for integrated circuits , used silicon gate structures (replaced metal gates) to improve the reliability, speed and Package integration level. Fagor designed the first commercial silicon gate integrated circuit (Fairchild 3708)

1969-Schottky barrier diode doubles the speed of TTL memory.
The innovation of design method improves the speed and low power consumption of standard 64-bit TTL random access memory. It was quickly applied to new bipolar logic and memory designs.

1970-MOS dynamic random access memory (DRAM) is similar in price to magnetic core memory
Intel's dynamic random access memory i103 opened up the challenge of semiconductors to magnetic core memory used for computer storage.

1971-Microprocessor condenses CPU functions into a single chip.
In order to reduce the number of chips required for arithmetic unit design, Intel engineers created the first single-chip microprocessor (CPU), i4004

1974-Digital display watch is the first system-on-chip (SoC) integrated circuit The integrated circuit
used for liquid crystal display digital watch is the first product to integrate the entire electronic system onto a single silicon chip (SoC)

1978-(Programmable
Arrangement Logic) User Programmable Logic Device was born . John Birkner and HT Chua of Monolithic Memory Company developed an easy-to-use programmable array to allow customers to quickly define logic functions Logic (PAL) devices and software tools.

1979-Single-chip digital signal processor was born
Bell Labs' single-chip DSP-1 digital signal processor device structure made the electronic switch system more perfect. Texas Instruments developed a programmable DSP

"Moore's Law" predicts the development of integrated circuits in the future-1965

• In 1965, Gordon
  Moore, director of research and development of Fairchild Semiconductor, wrote an internal document. He compiled 5 sets of products developed from 1959 to 1964 and charted the integration of chips and the lowest cost of individual devices. , And
  then draw a line through these points. From this graph, Gordon Moore found that each new chip roughly contains the capacity of its predecessor, and each new chip is generated after the previous chip is generated. within 18-24 months. If you press this trend continues, computing power with respect to the rise time period exponentially style. Gordon
  observations Moore (Moore), is now called Moore's Law. he was predicted in The number of devices on the chip will double every year in the next ten years, and will reach 65,000 in 1975.

  "For integrated circuits, cost reduction is quite attractive. As the development of technology enables it to integrate more and more circuit functions on a single chip, the cost advantage will continue to grow." Electronics, Volume 38 No. 8, April 19, 1965. In
  1975, Gordon
  Moore, who had joined Intel, made a modification to his own "Moore's Law" and pointed out that the number of transistors integrated on the chip will be every two years. double.