Reach into your pocket and pull out your cell phone. Even if you are rocking the most ancient, stripped-down flip phone from 2002, the object in your hands is still a marvel of modern electronics technology, packing billions of tiny switches into just a few cubic centimetres of space. By contrast, the Apollo Guidance Computer that sent humans to the moon contained only 17,000 switches, while the three-storey building-sized SAGE computer used for North American aerospace defence in the 1950s contained 50,000. This tremendous feat of miniaturization was made possible by advances in integrated circuit manufacturing, in which entire computers composed of parts just a few billionths of a meter across are etched into the surface of tiny silicon chips. But these developments would never have happened were it not for one key breakthrough made nearly 80 years ago – a discovery that forever changed the course of technology – and the world. This is the story of the transistor, one of the single most important inventions in modern history.

In the first half of the 20th Century, electronics design was dominated by one key piece of technology: the vacuum tube.

During Thomas Edison and his team’s thousands of experiments trying to create an economically viable incandescent lightbulb, there was a rather insanely revolutionary and far more unique device Edison accidentally invented in parallel with the lightbulb that was just one of his lightbulbs with a slight twist. But unfortunately for Edison, he did not realize the implication of what he’d just made in one of his thousands of tests, and how revolutionary it could be if refined a bit, and in the right applications. Because of his failure to realize any of this, nor be the one to perfect it for commercial use, despite his patent for the device, Edison is almost never given credit for his contribution on this world changing invention. Which is unsurprising as, as is a theme you’re probably picking up on, it’s the person who ultimately did the thing in its perfected commercial form, rather than was the first to come up with the thing, that usually gets credit in popular history.

On this one, enter English physicist John Ambrose Fleming, who was an advisor to Edison Electric Light. He would be inspired by Edison’s device to create his revolutionary Fleming valve vacuum tube in the early 20th century.

But going back to Edison’s original device, at one point during his experiments on the lightbulb, he and his staff were trying to figure out why carbon from the filament seemed to be jumping across the vacuum to the walls of the bulb. Clearly some current flow was involved. So in order to try to figure out what was going on here, Edison created a special bulb with a third electrode placed in between the legs of the filament, and then connected that to a galvanometer to measure the current. What he found was that if, relative to the filament, the plate was put at a negative potential, there would be no current between the plate and the filament. However, if the plate was at a positive potential, and the filament heated up enough, there would be a large current flow between the filament to the plate through the vacuum. Importantly in this, the electrons can only flow one way, from the hot element to the cold one, creating a rudimentary diode.

Edison ultimately patented the device for its potential use as a sort of voltage regulator, but seemingly did not understand the implications beyond that. Importantly, he did show it off at the International Electrical Exposition in Philadelphia in 1884, with one William Preece bringing several of these bulbs back to England and coining the term “Edison Effect,” also now known as “thermionic emission,” in a paper he published the following year on the phenomenon. And, of course, as noted, a couple decades later Fleming was inspired by all this and ultimately did his thing, as did others like Lee de Forest in the United States and the electronics age was born.

Vacuum tubes came in two basic varieties, which allowed electricity to be controlled in particular ways. The diode or thermionic valve, invented by Fleming in 1904, consisted of an evacuated glass bulb containing two basic components: a fine metal wire anode and a plate-shaped cathode. When current was run through the anode, as alluded to in the Edison test, the filament heated up red-hot and began giving off electrons via a process called thermionic emission. These electrons were then caught by the cathode, allowing the current to pass through the diode. If, however, the current was reversed, the lack of a filament on the cathode prevented it from heating up and emitting electrons – meaning the current could not flow in that direction. Diodes thus functioned like one-way valves – hence their alternative name – and were widely used as rectifiers for detecting radio signals, replacing the temperamental crystal detectors previously used in commercial radio sets.

The triode or Audion, invented by de Forest in 1906, was similar to the diode but with an extra component: a metal grid between the anode and the cathode. Applying an electric charge to the grid repelled electrons coming from the anode, allowing the number that made it through to the cathode to be adjusted. This meant that a weaker current could be used to control a stronger one, allowing weak signals – like those from a radio receiver or telephone – to be effectively amplified. de Forest’s invention launched the modern era of electronics, making possible such breakthroughs as long-distance telephone and radio communications. Triodes were also widely used as electronic switches, being more reliable and less prone to wear than electromechanical relays. Indeed, the earliest electronic computers like the British Colossus – used to break the German Lorentz cipher – and the American ENIAC – used to generate ballistics tables for naval guns – used thousands of networked vacuum tubes to perform high-speed calculations.

However, vacuum tubes had a number of serious shortcomings. For one thing, their filaments needed to heat up in order to work, such that old electronic equipment like radios and television sets often took anywhere from a few seconds to a few minutes to fully power up. They were also fragile, consumed large amounts of power, and generated large amounts of heat, meaning early electronic computers required massive air conditioning plants to keep their processors cool. And while subminiature vacuum tubes just a few centimetres long were developed, these power and heat issues placed a lower limit on the size of electronic circuits. For such devices to be made truly compact and portable, a new, more compact and energy-efficient type of electronic switch was needed.

Ironically, the solution to this problem would ultimately be found in an older technology. As mentioned at the start of the video, early commercial radio sets used a device called a crystal detector to pick up radio signals. Also known as a cat’s whisker detector, this device comprised a crystal of lead sulphide or galena and a small spring called the cat’s whisker mounted on a pivoted handle. To use this type of radio, the user touched the cat’s whisker to various parts of the galena crystal until they found a spot that rectified the radio signal and allowed it to be heard over headphones.

As can be imagined, this device was finicky to use and took a great deal of practice to master. The crystal detector worked by forming a temporary metal-semiconductor junction, also known as a Schottky diode after its discover, German physicist Walter H. Schottky.

Galena, along with iron pyrite, carborundum, silicon, germanium, and several other substances, belongs to a class of materials known as semiconductors. Neither excellent conductors like most metals nor full-blown electrical insulators, semiconductors can have their electrical properties modified by treating or doping them with various impurities such as arsenic or phosphorus. Such doping produces either an N-type semiconductor, which has an excess of electrons in the outer shells of its atoms; or a P-type semiconductor, which has an excess of missing electrons – known as electron holes. Sandwiching a P and N semiconductor together produces a PN-junction. At the interface between the two semiconductors, the difference in electric charges causes a so-called diffusion current to flow, with electrons flowing from the N side to the P side and electron holes flowing from the P side to the N side. This in turn results in the formation of two adjacent layers of positive and negative change – known as the depletion region.

When an external current is applied from the N to the P side – that is, in the direction of the internal diffusion current – it will flow freely through the diode. If, however, the current is applied in the opposite direction, it will cause the depletion region to grow, forming a barrier through which the current cannot flow. A PN junction thus performs the same function as a vacuum tube diode, allowing current to only flow in one direction.

In a metal-semiconductor junction like a crystal detector, the semiconductor is N-type while the metal acts as the P-type semiconductor, with the interface between the two forming a depletion region or Schottky barrier like in a PN junction.

The PN junction diode was discovered in 1939 by Bell Labs researcher Russel Ohl when he accidentally cut a section of a silicon ingot across the PN junction and noted its rectifying qualities. During the Second World War, self-contained Schottky and PN diodes developed for use in military radars, as vacuum tubes could not operate on the required frequencies. These devices were the first truly solid-state miniaturized electronic components, and pointed the way toward the use of semiconductors to create a new, efficient analogue to the triode vacuum tube.

Interestingly, a design for a type of semiconductor-based electronic switch now known as a Field-Effect Transistor or FET was patented as early as 1925 by Austrian-American inventor Julius Lilienfeld. However, as sufficiently pure semiconductors were not available at the time, Lilienfeld was unable to construct a working prototype, and his design remained little more than a footnote in the history of electronics. It would not be until after the Second World War that his ideas would finally become a reality.

The effort which resulted in the development of the first practical transistor was spearheaded by Mervin Kelly, director of research at Bell Telephone Laboratories in Murray Hill, New Jersey. Dissatisfied with the poor efficiency and reliability of vacuum tubes, in the late 1930s Kelly assembled a solid-state physics research team to come up with a semiconductor-based alternative. This work was interrupted by the Second World War, but resumed soon after. Strangely, this project was of relatively low priority for Bell, for while the triode or Audion had originally been developed for long-distance telephony, by the late 1940s the Bell Telephone System was based not on vacuum tubes, but complex yet reliable electromechanical devices known as Strowger Switches. A solid-state switch, if practical, was only anticipated to have limited, specialized applications, such as military radio and radar equipment.

Kelly assembled a diverse team of theoreticians, experimentalists, and engineers, including John Bardeen, Walter Brattain, Robert Gibney, Bert Moore, John Pearson, and the aptly-named William Shockley. Of these, it was the trio of Bardeen, Brattain, under the supervision of Shockley, who would ultimately make the vital breakthrough. While the often difficult Shockley preferred to work alone at home, Brattain and Bardeen formed a productive partnership, embracing the free-wheeling, anything-goes research culture of Bell Labs by working unsupervised late into the night.

The first design the team investigated was proposed by Shockley, and worked similarly to Julius Lilienfeld’s 1925 concept. Built around a block of silicon, like a vacuum tube the device had an anode and cathode – now named the source and drain – at either end, but instead of a grid used a third electrode called a gate to control the flow of electricity through the device. In theory, when current was applied to the gate, the electric field generated would impede electrons from flowing between the source and drain. In practice, however, the design failed to work. Nevertheless, Shockley was convinced his design was workable, and pushed Bell Labs to file a patent with himself named as sole inventor. To Shockley’s dismay, however, Bell had recently unearthed Lilienfeld’s original patents and informed Shockley that his idea was not original.

After much experimentation, Walter Brattain determined that the failure of Shockley’s design was due to a buildup of electrons on the surface of the silicon blocking the gate’s electric field. At the suggestion of Robert Gibnet, he and Bardeen tried getting around this problem by dunking the prototype in distilled water, filling in the air gap between the gate and the silicon and enhancing the strength of the electric field. Incredibly, this actually worked – though nowhere near as efficiently as the team had hoped. As Shockley later noted:

“This new finding was electrifying…at long last, Brattain and Gibney had overcome the blocking effect.”

Replacing the water with a chemical called glycol borate produced better results, but the device still had a slow response time and could not handle high frequencies – a key requirement for use in radio and radar equipment. Eventually, the team abandoned silicon as the substrate and focused instead on germanium, whose manufacture had already been perfected for use in diodes. But this material exhibited the same barrier effect as silicon, and though the team tried countless remedies like freezing the germanium with liquid nitrogen, full-scale amplification still continued to elude them.

It was at this point that a pair of serendipitous accidents nudged the team in the right direction. For their newest prototype, Brattain grew a thin layer of oxide on the surface of the germanium crystal and deposited an even thinner layer of gold onto this, hoping that the oxide would insulate the gold from the germanium. At first this seemed to work, but Brattain soon realized that the oxide layer had actually been washed away, meaning the gold was in direct contact with the germanium. This indicated that the device was not operating according to the field effect as Shockley had predicted, but some other, still unknown phenomenon.

On another occasion, while measuring the amplification or gain in a prototype, Brattain accidentally shorted out and ruined one of the gate electrodes by touching it with the emitter electrode. But when he placed the emitter close to the gate electrode, he suddenly observed the gain the team had been searching for.

Based on this, Bardeen suggested placing the emitter and gate electrodes extremely close to each other – within 50 micrometers – to enhance the effect. To accomplish this, Brattain wrapped a piece of thin gold foil around the point of a plastic triangle, cut a thin slit in the foil with a razor blade, and forced this pair of closely-spaced contacts into a crystal of germanium with a spring. Two electrodes known as the emitter and collector were connected to both halves of the gold foil, while a third base lead was connected to the germanium crystal, which had been specially prepared so that it consisted of two layers: an upper P-type layer full of electron holes and a lower N-type layer with excess electrons. In this configuration, the current flowing from the collector to the base was modulated by applying a current to the emitter.

On December 16, 1947, Brattain and Bardeen tested their new design for the first time. To their delight, it worked perfectly, exhibiting a 30 percent gain in power and voltage gain of 15% at a frequency of 1,000 Hertz. Carpooling home that night, Brattain exclaimed to his colleagues that they had just conducted the most important experiment of their lives and swore them to secrecy until Bell Labs officially announced their discovery. Bardeen, however, could not help sharing the news, telling his wife at dinner that “We discovered something today.” His wife, distracted by the couple’s children, reportedly replied: “That’s nice, dear.”

By contrast, William Shockley, on sabbatical in Europe at the time, was enraged to discover that not only had he not been directly involved in the team’s breakthrough – but that they had strayed so far from his original field-effect concept. It was a bitterness which was to prove surprisingly productive.

On June 30, 1948, Bell Labs officially announced Brattain and Bardeen’s discovery, which by now had acquired a new name: transistor. The term had been coined by fellow Bell engineer and part-time science fiction writer John Pierce as a contraction of “trans-resistor”. Unfortunately, however, the announcement of the transistor received little attention in either the popular or scientific press. Not only were there few apparent applications for the device, but it was fragile, temperamental, and difficult to manufacture. Furthermore, even its inventors didn’t understand exactly how it worked.

Meanwhile, Shockley, fuelled by jealousy and indignation, doggedly pursued his quest to one-up his colleagues. While attending a meeting of the Physical Society in Chicago in late 1947, he began filling his notebook with page after page of detailed notes describing a new type of transistor, consisting of one layer of P-type semiconductor sandwiched between two layers of N-type semiconductor. By January 23, 1948, Shockley had come up with a workable design, which worked similarly to a PN diode but with three terminals: the emitter, the collector, and the base. When a positive current was applied to the base, it disrupted the depletion region between the semiconductor layers by draining away excess electrons, allowing current to flow between the emitter and the collector. Bardeen and Brattain’s transistor worked in a similar fashion, only the currents travelled through a thin layer at the top of the germanium crystal. One month after Shockley perfected his theoretical design, Bell Labs filed four patents for semiconductor amplifiers – both Brattain and Bardeen’s original point contact design and Shockley’s bipolar junction or NPN transistor.

Though Shockley’s design was successfully demonstrated on April 2, 1950, the first commercial transistors, produced by Western Electric in 1951, were of the point-contact type. But while these saw limited use in long-distance telephone switching gear and military equipment, it soon became clear that the junction transistor was far more robust and easy to manufacture, and this became the standard design going forward.

Still, for several years the transistor remained a solution looking for a problem. It was not until 1952 that New York-based firm Sonotone introduced the lightweight transistorized hearing aid – the first consumer product to make use of the new technology. Two years later, researcher Gordon Teal at Texas Instruments figured out how to replace germanium – which was unreliable and sensitive to heat fluctuations – with silicon, producing an even more reliable and robust transistor. That same year, Texas Instruments and Industrial Development Engineering Associates unveiled a groundbreaking product: the Regency T-1, the world’s first portable, fully-transistorized radio. Though plagued by technical problems, the radio was an instant hit, selling over 150,000 units over its brief production run.

It is difficult to overstate the cultural impact of the TR-1 and its descendants. Previously, consumer radios were heavy, bulky devices restricted to the home living room. With transistor radios, however, consumers – particularly teenagers – could take their music wherever they wanted – an ability that profoundly shaped the development of youth culture.

The transistor also helped reshape the global economic landscape. As American manufacturers began increasingly focusing on Cold War military contracts, foreign entrepreneurs saw an opportunity to cash in on the emerging consumer electronics market. Among these were Japanese engineers Masaru Ibuka and Akio Morita, who in 1946 founded an electronics company called Tokyo Teletech. In 1958, the company changed its name to Sony. Soon, inexpensive Sony transistor radios and television sets began flooding the global market, establishing Japan as a global leader in consumer electronics and finally bringing the era of the vacuum tube to an end.

Meanwhile, the importance of Brattain, Bardeen, and Shockley’s discoveries were finally recognized when, in 1956, the trio shared the Nobel Prize in Physics for “…their researches on semiconductors and their discovery of the transistor effect.” But their elation was short-lived, for by then Shockley’s ruthless pursuit of sole credit for the invention of the transistor had broken the team apart. Shortly after receiving the Nobel Prize, Shockley moved to Palo Alto California and founded Shockley Semiconductor Laboratory, the first tech company in what would come to be known as Silicon Valley. But while Shockley’s clout initially attracted the best and brightest to his company, his difficult personality and tyrannical management style soon drove them away.

One group of exiles known as the “traitorous eight” went on to found Fairchild Semiconductor, which in 1959 developed the world’s first practical integrated circuit or microchip. Two of the eight, Bob Noyce and Gordon Moore, later founded the Intel Corporation, today one of the world’s largest manufacturers of microprocessors.

After losing his company, in 1963 Shockley accepted a position at Stanford University as a professor of Engineering and Applied Science. And it is here that his career took a dark turn. Despite holding no degree in genetics or related disciplines, Shockley began vocally promoting pseudoscientific theories about race, intelligence, and eugenics, declaring, for example, that:

“My research leads me inescapably to the opinion that the major cause of the American Negro’s intellectual and social deficits is hereditary and racially genetic in origin and, thus, not remediable to a major degree by practical improvements in the environment.”

Such was Shockley’s conviction that miscegenation – AKA race mixing – posed an existential threat to the United States that he ran as a Republican candidate in the 1982 Senate Election on the single-issue platform of opposing the, to quote, “dysgenic threat” posed by African-Americans and other minority groups. He came in eighth place in the primary, receiving a paltry 0.37% of the vote. By the time Shockley died in 1989 at the age of 79, he had become a pariah, with his obituary in the Los Angeles Times stating:

“He went from being a physicist with impeccable academic credentials to amateur geneticist, becoming a lightning rod whose views sparked campus demonstrations and a cascade of calumny.”

Meanwhile, the co-discoverers of the transistor fared somewhat better. In 1951, John Bardeen left Bell Labs for the University of Illinois, where he began investigating the phenomenon of superconductivity – the ability of certain materials to attain zero electrical resistance when cooled to extremely low temperatures. This pioneering work earned him the 1972 Nobel Prize in Physics, making him the only person in history to win this award twice. He died in 1991 at the age of 82.

Walter Brattain continued to work at Bell Labs until 1967 before joining the faculty at Whitman College in Walla Walla, Washington, where he remained until his retirement in 1976. He died in 1987 at the age of 85. Thus, while the transistor launched a multi-billion-dollar global industry, beyond their Nobel Prizes none of its three inventors significantly benefited financially from their discovery.

In any piece discussing the origins of the transistor, we would be remiss in not pointing out that Bardeen, Brattain, and Shockley were not the sole people working on the transistor when they came up with it. At around the same time as semiconductor research was ramping up at Bell Labs, Herbert Mataré and Heinrich Welker, German physicists working at the Compagnie de Friens et Signaux in Paris, were investigating similar germanium-based modulation devices. In June 1948, they succeeded in building a working point-contact transistor remarkably similar to Bardeen, Brattain, and Shockley’s 1947 prototype. Shortly thereafter, however, Mataré and Welker were dismayed to learn that Bell Labs had already beaten them to the punch. Nevertheless, in 1949 their employer became the first company in Europe to commercially produce transistors.

It should also be mentioned that less than a decade later a number of inventors including Ian Ross, John Wallmark, and Mohammed Atalla developed workable Field-Effect Transistors or FETs. Today, FETS – in particular Metal Oxide or MOSFETs – are the most widely used transistor type in the world, being particularly well-suited to miniaturization. Indeed, while the earliest commercial transistors were on the order of one centimetre in size, modern integrated circuit transistors are so inconceivably tiny that the world’s most powerful single computer chip at the time of the writing of this piece – the Cerebras Wafer Scale Engine 2 – contains an unfathomable 2.6 trillion of them.