Transistor: 60 years of great achievements of a small element

Stages of the long journey

Televisions, cars, radios, medical and household appliances, computers, space shuttles and even programmable door locks in hotels – it is probably difficult to imagine even one more or less complex electronic device that surrounds us that does not use transistors . The invention of the transistor 60 years ago by employees of the research center Bell Labshas been a critical factor in driving many remarkable innovations and technology developments. In fact, without the transistor, the existence of almost the entire modern electronic and digital industry would be impossible. It was thetransistor – a tiny device, an element of a microcircuit that acts like a miniature switch and thereby allows the implementation of information processing algorithms – that ensured the phenomenal triumph of computers.

What is the secret of success? Microelectronics is steadily developing, constantly enriching the scientific and technical community with innovations. With each new generation of technological processes for their production, transistors become more and more compact, faster, and more economical in energy consumption. In November 2007, Intel engineers – for the first time in its long history semiconductor integrated circuits – broke the silicon “monopoly” in the production of transistors and introduced new materials into the structure of semiconductor components.

This made it possible to create microprocessors based on the microarchitectureIntel< span style=”font-family: Arial;”>® Core™, using a revolutionary 45-nanometer production technology using a gate insulator (dielectric) based on hafnium with a high dielectric constant – high–kspan>, as well as a metal shutter, which ensures record performance and efficient power consumption.

What’s next? Intel Corporation intends to continue to push the boundaries of what is possible through technological innovation, to create new types of products that can qualitatively change our lives – the way we work, relax, and exchange information.

“On” / “Off”

The official date of the birth of the first transistor is considered to be December 23 (according to other sources – December 16), 1947. The authors of this wonderful invention were American physicists William Shockley, John Bardeen and Walter Brattain. True, the scientific community initially greeted this invention rather coolly, but already in 1956, all three Americans were awarded the Nobel Prize in physics. Moreover, subsequently John Bardeen became the only one in the history of the Nobel Prize to win twice in the same category: the second prize in the field of physics was awarded to him in 1972 for the creation of the theory of superconductivity.

Well, the name itself – “transistor” was invented by their colleague John R. Pierce. In May 1948, he won an internal competition organized among laboratory employees for the most successful name for the invention, which at that time was only a few months old. The word “transistor” is formed by combining two terms: “transconductance” (active interelectrode conductivity) and “variable resistor” or “varistor” (variable resistance, varistor).

The first to actively use transistors were radio amateurs who used these elementary devices to amplify the signal. This is why the first portable wireless radios of the fifties were called transistor radios, or even simply “transistors.” Over time, however, they came to be used primarily as elements of integrated circuits, giving the transistor a critical role in human technological advances over the past forty years.

It’s interesting to note that a transistor essentially does what a regular switch does: turn current on and off. The on position for the transistor means “1”, the off position means “0”. A huge number of transistors on an integrated circuit generate ones and zeros, which add up to a computer-understandable binary code – a “language” that computers use in the process of calculations, word processing, playing movies and audio, displaying images…

Transistors and… rock’n’roll

The first transistor, in which a current flowed along the surface of a semiconductor, was used to amplify the electrical signal passing through it – transistors coped with this task more efficiently than the popular at that time, but more bulky and fragile vacuum tubes.

In order to speed up the popularization of transistors as much as possible, the Bell Labs research center decided to license transistor technology. Twenty-six companies purchased the license, costing US$25,000. However, for transistor technology to be commercially successful, it was necessary to attract the attention of a mass audience. This became possible thanks to transistor radios.

The first model of such a device, containing as many as four transistors, was presented in October 1954. With the advent of a portable radio, radio fans gained the opportunity to listen to music and receive information anywhere – this was immediately taken advantage of by young people who had the opportunity to escape from parental care and assert themselves with the help of a new subculture. Thus, portable radio stimulated a new revolution… and in music – rock and roll sounded on the air everywhere!

Integrated circuit

By the end of the 50s, the transistor “settled” in radios, telephones and computers, and although its dimensions were much smaller than those of vacuum tubes, this was clearly not enough to create a new generation of electronic devices. To realize the enormous computing potential of transistors, make them mass-produced, and reduce costs, another invention was needed.

In 1958, Jack Kilby of Texas Instruments and Robert Noyce of Fairchild Semiconductor, who later became one of the founding fathers of Intel Corporation, invented a way to combine large numbers of semiconductor transistors into a single integrated circuit, or chip. This was a giant step forward, since previously the individual components of an electrical circuit had to be connected manually.

The chips had two advantages: lower cost and higher performance. Both advantages resulted from miniaturization, which allowed for exponential reduction in device size and extraordinary dynamism in the production process. Gordon Moore, who co-founded processor giant Intel in 1968 with Noyce, formulated a prediction in a journal article that was published in 1965 and became known as “Moore’s Law.”

According to this law, the number of transistors in a microcircuit was supposed to double every one and a half to two years, which in turn would provide an increase in computing power and a reduction in the final cost of the product during its mass production. The ability to fit many compact elements onto a small surface has proven to be a critical factor in the successful advancement of chips.

Chip manufacturers have been able to maintain this exponential increase in transistor density on a chip for decades. Intel® Corporation’s first computer microprocessor, the 4004, released in 1971, contained 2,300 transistors. In 1989, the Intel® 486 processor already had 1,200,000, and in 2000, the Intel® Pentium® 4 processor surpassed 42 million. The new quad-core Intel® Core™ 2 Extreme processor is built on 45nm manufacturing technology and contains 820 million transistors.

Atom Games

Moore’s Law is constantly predicted to die. Infinite exponential growth is impossible by definition – and yet processor manufacturers still manage to bypass the limitation. Last September, Gordon Moore said that the law named after him has every chance of remaining in force for at least another 10-15 years, but then new fundamental barriers to its implementation may arise. One way or another, difficult times awaited this most famous law of the computer world of the 21st century.

The miniaturization battle has exhausted one of the transistor’s most critical components: the silicon dioxide (SiO2) layer that serves as the insulating layer between the transistor’s gate and the channel through which current flows when the transistor is turned on. With each new generation of processors, this insulating layer became thinner – until two generations ago its thickness reached 1.2 nm, or 5 atoms. Intel engineers were no longer able to make this layer thinner by at least one more atom.

As the thickness of the insulating layer decreased, the leakage current increased. The insulating layer began to pass current inside the transistor, the behavior of the device changed, it began to dissipate an increasing amount of energy. As a result, the current consumption of the processor increased, and additional heat was generated during its operation.

Fundamental limit

Transistor current leakage has become a major problem in the semiconductor industry: without a breakthrough in this area, the long-predicted fundamental limit has become insurmountable. Moreover, this would not only mean the end of Moore’s Law – the digital revolution of recent decades would suddenly stop. Computer processors, which were practically doubling their performance every 24 months, could sink into oblivion.

To solve the crisis, it was necessary to increase the thickness of the insulating layer, but make this thicker layer from a different dielectric material with a higher dielectric constant (high-k) to preserve the nature of the gate-channel interaction. In January 2007, Intel announced that for the first time in forty years, the insulating layer would not be made of silicon oxide, but of a material based on hafnium, a silver-gray metal that has superior electrical properties to silicon and can reduce leakage current by a factor of ten. Gordon Moore himself called this major achievement “the most important change in transistor technology since the late sixties.”

However, this breakthrough solved the problem only half. The new material turned out to be incompatible with an important component of the transistor – the gate. Even worse, transistors with the new insulating material worked less efficiently than with the old one. Then it was proposed to replace the shutter material: Intel Corporation discovered a unique combination of metals, the composition of which is kept strictly secret.

On November 12, 2007, Intel introduced a new generation of processors based on these materials and 45-nanometer production technology. The new manufacturing technology, thinner than the previous 65-nanometer one, allowed Intel to nearly double the number of transistors placed on the same die area – now it was possible to choose between increasing the total number of transistors or reducing the size of the processor. The new transistors require 30% less energy to turn on and off. As a result, the new generation of Intel processors, manufactured using 45-nanometer manufacturing technology, not only demonstrates record-breaking performance, but also marks a breakthrough in the field of energy consumption.

The increase in computing power, which is a consequence of Moore’s law, allows humanity to more effectively calculate ways to solve the most important problems facing it: climate change, hereditary diseases, the secrets of genetics, etc. Modern ways and rates of solving such problems were difficult to even imagine five years ago. New applications help change our lives and make it even safer…

Number of transistors in the processor:

Intel® 4004                   1971     2300

Intel® 8086                   1978     29 000

Intel® 486                     1989     1,200 000

Intel® Pentium® III        1999     9 500 000

Intel® Pentium® 4 2000     42 000 000

Intel® Core™ 2 Duo     2007     410,000,000

Processors and production technology:

1993     Intel®Pentium®                             800 nm

1999     Intel®Pentium® III                      250 nm

2002     Intel®Pentium® 4                      130 nm

2005     Intel®Pentium® D                      90 nm

2006     Intel® Core™ 2 Duo                        65 nm

2007     Intel® Core™ 2 Duo                 45 nm