Other articles in the series:
- The History of Relays
- The History of Electronic Computers
- The History of the Transistor
- The History of the Internet
For over a hundred years, the analog dog wagged its digital tail. Attempts to enhance our sensory faculties—sight, hearing, and even, in a sense, touch—led engineers and scientists on a quest for better components for the telegraph, telephone, radio, and radar. It was only by fortunate coincidence that these searches uncovered the path to creating new types of digital machines. And I have decided to tell the story of this ongoing , during which telecommunications engineers supplied the raw materials for the first digital computers, and sometimes even designed and built these computers themselves.
But by the 1960s, this fruitful collaboration came to an end, and so did my story. Manufacturers of digital equipment no longer needed to look into the worlds of telegraphy, telephony, and radio for new, improved switches, as the transistor itself provided an inexhaustible source of enhancements. Year after year, they dug deeper and deeper, always finding ways to exponentially increase speed and decrease costs.
However, none of this would have happened if the invention of the transistor had stopped at .
A Slow Start
There was no active enthusiasm in the popular press regarding Bell Labs' announcement of the invention of the transistor. On July 1, 1948, The New York Times allocated three paragraphs to this event at the bottom of its 'Radio News' summary. Notably, this news appeared after others that were clearly deemed more important: for example, the hour-long radio show 'Time of the Waltz,' which was set to air on NBC. With hindsight, we may want to laugh or even scold the unknown authors—how could they not recognize the world-changing event?
But looking back at the past distorts perception, amplifying those signals whose significance we are aware of, even though they were lost in a sea of noise at the time. The 1948 transistor was vastly different from the transistors in the computers on which you are reading this article (unless you've decided to print it out). They differed so much that, despite sharing the same name and a continuous line of inheritance connecting them, they must be considered different species, if not entirely different categories. They have different compositions, different structures, and different operating principles, not to mention the enormous differences in size. Only through constant reinvention was the clumsy device built by Bardeen and Brattain able to transform the world and our lives.
In fact, the germanium transistor with a single contact point did not deserve any more attention than it received. It had several defects inherited from the vacuum tube. It was, of course, much smaller than the most compact tubes. The absence of a heated filament meant it generated less heat, consumed less power, didn't burn out, and didn't require warm-up time before use.
However, the accumulation of dirt on the contact surface led to failures and negated the potential for a longer lifespan; it produced a noisier signal; worked only at low power and within a narrow frequency range; failed under heat, cold, or humidity; and could not be produced uniformly. Several transistors manufactured in the same way by the same people would exhibit strikingly different electrical characteristics. And all this came at a cost eight times greater than that of a standard tube.
Only by 1952 did Bell Labs (and other patent holders) solve the production problems sufficiently for single contact point transistors to become practical devices, and even then, they did not particularly spread beyond the hearing aid market, where price sensitivity was relatively low, and the benefits related to battery life outweighed the drawbacks.
However, the first attempts to improve the transistor into something better and more useful had already begun. In fact, they started long before the public became aware of its existence.
Shockley’s Ambitions
By the end of 1947, Bill Shockley was very excited as he traveled to Chicago. He had vague ideas about how to surpass the recently invented transistor by Bardeen and Brattain, but he had not yet had the chance to develop them. Therefore, instead of enjoying a break between stages of the work, he spent Christmas and New Year’s in a hotel, filling about 20 pages of his notebook with ideas. Among them was a proposal for a new transistor made of a semiconductor sandwich—a slice of p-type germanium between two pieces of n-type.
Encouraged by having such a trump card up his sleeve, Shockley confronted Bardeen and Brattain upon their return to Murray Hill, claiming all the glory for the invention of the transistor. Wasn't it his idea of the field effect that made Bardeen and Brattain lock themselves in the lab? Shouldn't all patent rights be transferred to him because of this? However, Shockley’s trick backfired: the patent lawyers at Bell Labs discovered that an unknown inventor, , had patented a semiconductor amplifier using the field effect almost 20 years earlier, in 1930. Lilienfeld, of course, never realized his idea given the state of materials at the time, but the risk of overlap was too great—it was better to completely avoid mentioning the field effect in the patent.
So, while Bell Labs did give Shockley a generous share of the inventor's credit, only Bardeen and Brattain were mentioned in the patent. However, what’s done is done: Shockley’s ambitions destroyed his relationships with the two subordinates. Bardeen ceased working on the transistor and focused on superconductivity. He left the labs in 1951. Brattain remained there but refused to work with Shockley again and insisted on a transfer to another group.
Due to his inability to work with others, Shockley never advanced in the labs and ultimately left. In 1956, he returned home to Palo Alto to found his own transistor manufacturing company, Shockley Semiconductor. Before leaving, he separated from his wife Jean, who was recovering from uterine cancer, and began a relationship with Emmy Lenning, whom he soon married. However, of the two halves of his Californian dream – a new company and a new wife – only one came to fruition. In 1957, his best engineers, angered by his management style and the direction he was leading the company, left him to establish a new firm, Fairchild Semiconductor.
Shockley in 1956
So Shockley abandoned the empty shell of his company and took a position in the electrical engineering department at Stanford. There, he continued to alienate his colleagues (and his oldest friend, physicist ) with his theories on racial degeneration and – topics that had fallen out of favor in the U.S. since the end of the last war, especially in academic circles. He took pleasure in stirring up debates, sensationalizing the media, and inciting protests. He passed away in 1989, estranged from his children and colleagues, only visited by his eternally devoted second wife, Emmy.
Although his pitiful attempts at entrepreneurship failed, Shockley sowed the seeds in fertile ground. The San Francisco Bay Area spawned numerous small electronics firms, sweetened by federal funding during the war. Fairchild Semiconductor, an accidental offshoot of Shockley, birthed dozens of new companies, a couple of which are still well-known today: Intel and Advanced Micro Devices (AMD). By the early 1970s, the area had earned the derisive nickname "Silicon Valley." But wait – Bardin and Breetain created the germanium transistor. Where did silicon come from?
This is what the abandoned site in Mountain View, where Shockley Semiconductor once stood, looked like in 2009. The building has since been demolished.
To the silicon intersection
The fate of the new type of transistor invented by Shockley in a Chicago hotel was much more fortunate than that of its creator. This was all thanks to one man's determination to grow pure semiconductor crystals. Gordon Teal, a physical chemist from Texas, who had studied the then useless germanium for his doctorate, got a job at Bell Labs in the 1930s. Upon learning about the transistor, he became convinced that its reliability and power could be significantly improved by creating it from pure single crystals rather than the polycrystalline mixtures used at the time. Shockley dismissed his attempts as a waste of resources.
However, Teal pushed on and succeeded, with the help of mechanical engineer John Little, in creating a device that extracted a tiny seed crystal from molten germanium. As it cooled around the seed, the germanium expanded its crystalline structure, creating a continuous and almost pure semiconductor lattice. By spring 1949, Teal and Little could produce crystals on demand, and tests showed they left their polycrystalline competitors far behind. In particular, the added non-essential carriers could survive inside for hundreds of microseconds or even longer (compared to no more than ten microseconds in other crystal samples).
Now, Teal could afford more resources and expanded his team to include another physical chemist, Morgan Sparks, who had come to Bell Labs from Texas. They began to change the melt for producing p-type or n-type germanium by adding beads of appropriate dopants. Within a year, they had refined the technology to the point where they could grow n-p-n germanium sandwiches directly in the melt. And it worked exactly as Shockley had predicted: the electric signal from the p-type material modulated the electric current between two conductors connected to surrounding pieces of n-type.
Morgan Sparks and Gordon Teal at the workbench in Bell Labs
This transistor with a grown junction surpassed its predecessor with a point contact in almost every aspect. Notably, it became more reliable and predictable, produced much less noise (and thus was more sensitive), and was extremely energy-efficient—consuming a million times less energy than a typical electronic lamp. In July 1951, Bell Labs held another press conference to announce the new invention. Even before the first transistor made it to market, it had essentially already become obsolete.
And yet, this was only the beginning. In 1952, General Electric (GE) announced the development of a new process for creating junction transistors, the alloy method. In this process, two indium balls (p-type donor) melted on either side of a thin slice of n-type germanium. This method was simpler and cheaper than growing junctions in alloys, and such transistors had lower resistance and supported higher frequencies.
Grown and alloyed transistors
The following year, Gordon Teal decided to return to his home state and took a job at Texas Instruments (TI) in Dallas. The company was originally founded under the name Geophysical Services, Inc., and initially produced equipment for oil exploration. TI opened an electronics division during the war and was now entering the transistor market under a license from Western Electric (the manufacturing division of Bell Labs).
Teal brought with him new skills acquired in the labs: the ability to grow and silicon monocrystals. The most obvious weakness of germanium was its sensitivity to temperature. When exposed to heat, the germanium atoms in the crystal quickly lost free electrons, making it increasingly behave like a conductor. At a temperature of 77 °C, it completely ceased to function as a transistor. The primary target market for transistors was the armed forces—a potential consumer with low price sensitivity and a huge need for stable, reliable, and compact electronic components. However, the temperature-sensitive germanium would not be suitable in many military applications, particularly in aerospace.
Silicon was much more stable, but it came with a significantly higher melting point, comparable to that of steel. This posed enormous challenges, considering that high-quality transistors required very pure crystals. Hot molten silicon would absorb impurities from any crucible it was in. Thil and his team at TI managed to overcome these difficulties by using ultra-pure silicon samples from DuPont. In May 1954, at the Institute of Radio Engineers conference in Dayton, Ohio, Thil demonstrated that new silicon devices produced in his lab continued to function even when immersed in hot oil.
Successful upstarts
Finally, about seven years after the transistor was first invented, it could be made from the material that became synonymous with it. And it would take about the same amount of time before transistors that roughly resembled the form used in our microprocessors and memory chips emerged.
In 1955, scientists at Bell Labs successfully learned to make silicon transistors using a new doping technique – instead of adding solid impurity beads into the liquid melt, they introduced gaseous additives into the solid surface of the semiconductor (). By carefully controlling the temperature, pressure, and duration of the process, they achieved precisely the necessary depth and degree of doping. Enhanced control over the manufacturing process led to improved control over the electrical properties of the final product. Importantly, thermal diffusion allowed for batch production – large wafers of silicon could be doped and then cut into transistors. The military funded Bell Labs as the organization of production required significant upfront costs. They needed a new product for the ultra-high-frequency line of early radar detection (In the chain of Arctic radar stations designed to detect Soviet bombers flying from the North Pole, they were willing to pay $100 for a transistor (this was a time when a new car could be bought for $2000).
Alloying together with , which managed the placement of impurities, opened up the possibility of etching the entire pattern at once on a single semiconductor substrate – this was simultaneously conceived at Fairchild Semiconductor and Texas Instruments in 1959.from Fairchild used chemical deposition of metal films to connect the transistor's electrical contacts. It eliminated the need for manual wiring, reduced production costs, and increased reliability.
Finally, in 1960, two engineers from Bell Labs (John Atalla and Devon Kan) implemented Shockley’s original concept of the field-effect transistor. A thin oxide layer on the semiconductor surface effectively suppressed surface states, allowing the electric field from the aluminum gate to penetrate into the silicon. This led to the birth of the MOSFET [metal-oxide semiconductor field-effect transistor], which turned out to be easy to miniaturize and is still used in almost all modern computers (interestingly, Atalla was from Egypt, and Kan from South Korea, and practically these two engineers are the only ones in our entire history without European roots).
Finally, thirteen years after the invention of the first transistor, something resembling your computer's transistor emerged. It was easier to produce, used less energy than the planar transistor, but was relatively slow to respond to signals. Only after the widespread adoption of large integrated circuits with hundreds or thousands of components located on a single chip did the advantages of field transistors come to the forefront.
Illustration from the patent for the field-effect transistor
The field effect was the last major contribution of Bell Labs to the development of the transistor. Major electronics manufacturers, such as Bell Labs (with their Western Electric), General Electric, Sylvania, and Westinghouse, amassed an impressive volume of semiconductor research. From 1952 to 1965, Bell Labs alone filed over two hundred patents on this subject. Yet, the commercial market quickly fell into the hands of new players like Texas Instruments, Transitron, and Fairchild.
The early transistor market was too small to attract the attention of major players: about $18 million a year in the mid-1950s, compared to a total electronics market of $2 billion. However, the research laboratories of these giants served as unintended training camps where young scientists could absorb knowledge about semiconductors before moving on to sell their services to smaller firms. When the market for vacuum tube electronics began to shrink seriously in the mid-1960s, it was too late for Bell Labs, Westinghouse, and others to compete with the upstarts.
The Transition of Computers to Transistors
In the 1950s, transistors invaded the world of electronics in four significant areas. The first two were hearing aids and portable radios, where low power consumption, and consequently long battery life, outweighed other considerations. The third was military applications. The U.S. Army had high hopes for transistors as reliable and compact components that could be used everywhere from field radios to ballistic missiles. However, initially, their spending on transistors resembled a gamble on the future of technology rather than a confirmation of their value at the time. Finally, there were also digital computing.
In the field of computing, the drawbacks of vacuum tube switches were well known, and some skeptics even believed before the war that an electronic computer would not be made into a practical device. When thousands of tubes were assembled in one device, they consumed a great deal of electricity while generating immense heat, and in terms of reliability, one could only count on their regular burnout. Thus, the low-power, cool, and filament-free transistor became the salvation of computer manufacturers. Its disadvantages as an amplifier (for example, a noisier output signal) were not much of a problem when used as a switch. The only obstacle was cost, which would eventually start to drop sharply.
All early American experiments with transistor computers occurred at the intersection of the military's desire to explore the potential of a promising new technology and engineers' eagerness to switch to improved switches.
At Bell Labs in 1954, the TRADIC was built for the U.S. Air Force to see if transistors could enable the installation of a digital computer aboard a bomber, replacing analog navigation and target search assistance. The Lincoln Laboratory at MIT developed the TX-0 computer as part of a comprehensive air defense project in 1956. The machine used another variant of the transistor, the surface-barrier type, which was well-suited for high-speed computations. Philco built its SOLO computer under contract with the Navy (though in reality, at the request of the NSA), completing it in 1958 using another version of the surface-barrier transistor.
In Western Europe, which was not as well-endowed with resources during the Cold War, the story was quite different. Machines like the Manchester Transistor Computer, (another name inspired by the ENIAC project, and encoded by writing it backward), and the Austrian were side projects that utilized the resources their creators could scrape together—including first-generation point-contact transistors.
There are many debates about which computer can be considered the first to use transistors. Ultimately, it all depends on the selection of appropriate definitions for terms like 'first,' 'transistor,' and 'computer.' In any case, we know where history concludes. The commercialization of transistor computers began almost immediately. Year after year, computers became more powerful for the same price, while computers of the same power became cheaper, and this process seemed so relentless that it was elevated to the level of a law, alongside gravity and the conservation of energy. Should we really argue about which pebble started the avalanche?
Where did Moore's Law come from?
As we approach the end of the history of the switch, it's worth asking: what led to this avalanche? Why does Moore's Law exist (or did exist – we can debate that another time)? There is no Moore's Law for airplanes or vacuum cleaners, just as there is none for vacuum tubes or relays.
The answer consists of two parts:
- The logical properties of the switch as an artifact category.
- The ability to use purely chemical processes to manufacture transistors.
First, let's discuss the essence of the switch. The properties of most artifacts must satisfy a broad range of inexorable physical constraints. A passenger plane must withstand the total weight of many people. A vacuum cleaner must be able to suction a certain amount of dirt within a specific time over a defined physical area. Airplanes and vacuum cleaners would be useless if scaled down to the nanoscale.
However, for the switch – an automatic switch that has never been touched by human hands – the physical constraints are much fewer. It must have two distinct states and must be able to communicate changes in its states to other similar switches. In essence, all it needs to do is turn on and off. So, what is so special about transistors? Why have other types of digital switches not experienced such exponential improvements?
Here we come to the second fact. Transistors can be made using chemical processes without mechanical intervention. From the very beginning, the key element in transistor production was the use of chemical impurities. Then the planar process emerged, eliminating the last mechanical step in production – the connection of wires. As a result, it removed the final physical limitation on miniaturization. Transistors no longer needed to be large enough for human fingers or any mechanical device. Everything was achieved through simple chemistry on an unimaginably small scale: acid for etching, light to control which parts of the surface would resist etching, and vapors for doping impurities and metal films onto the etched paths.
But why is miniaturization even necessary? Reducing size brought a whole range of pleasant side effects: increased switching speed, reduced energy consumption, and lower costs per unit. These powerful incentives motivated everyone to search for ways to further reduce switch sizes. And the semiconductor industry went from producing switches the size of a fingernail to packing tens of millions of switches into a square millimeter during a person's lifetime. From charging eight dollars for a single switch to offering twenty million switches for one dollar.
The Intel 1103 memory chip from 1971. Individual transistors, measuring just a few micrometers, are already indistinguishable by the eye. And since then, they have shrunk another thousand times.
What else to read:
- Ernest Bruan and Stuart MacDonald, Revolution in Miniature (1978)
- Michael Riordan and Lillian Hoddeson, Crystal Fire (1997)
- Joel Shurkin, Broken Genius (1997)
Source: habr.com
