The transistor was invented 50 years ago. Joanna Bawa reports on the significance of the device
There were only nine days to go to Christmas but the two physicists had other things on their minds. Huddled secretively in a corner of their laboratory, John Bardeen and Walter Brattain were building a primitive device whose impact on the world not even these engineering geniuses appreciated.
Cautiously they glued a fragment of gold foil to a wedge of plastic, pressed the wedge on to a sliver of germanium and secured the contact with a paperclip ”spring”. Then they clamped their clumsy handiwork into a plastic vice, and wired up an electrical circuit; copper wire linked the device, on one side, to a battery to supply power and, on the other, to an oscilloscope to record the device’s response. When all was ready, Brattain switched on the battery, sending a trickle of power through the circuit. As the men watched in silence, the faint signal leapt to 100 times its strength, a sudden powerful glow on the oscilloscope.
A week later, on December 23, the experiment was performed in front of their supervisor, William Shockley, and a handful of ”top brass” at their research centre, Bell Laboratories of Murray Hill, New Jersey. Shockley, already an eminent physicist, was quick to explain the significance of the work to his bosses.
But in a world in which slow electro- mechanical relays or unreliable vacuum tubes (or ”valves”) controlled electrical circuits, he had his work cut out. Besides, no one fully understood then how the crystalline structure of germanium provided different levels of resistance to electrical current that allowed small changes in input to cause large changes in output. This constituted the ”transistor” effect – the power leap the team had just witnessed.
For the first time, said Shockley, you could control and amplify an electrical signal without the need of a vacuum tube. By eliminating the laboriously produced glass vacuum tube, you were eliminating a fragile, costly and bulky limit to the scope and power of electronics.
Transistors exert control in two ways: by switching a current on and off (equivalent to ”0” and ”1” in a microchip) and by boosting the strength of the incoming current, or signal. A modern transistor can control many others through switching, and combinations of switching can be used to create complex instructions.
Anxious to understand the mechanism and secure a patent, Bell Labs waited until June 1948 before formally announcing the ”point-contact” transistor. Less than 10 years later, in a world buzzing with new electronic devices and networks, Shockley, Bardeen and Brattain jointly shared the 1956 Nobel Prize for Physics – although to Shockley’s endless chagrin, only Bardeen and Brattain were included in the patent and ultimately credited with the invention.
The peculiar gadget that clicked into life in Bell Labs back in 1947 was, even then, the culmination of years of research and of a decision to investigate a poorly understood group of materials, the semiconductors. Bardeen and Brattain were able to show that semiconductors offered a far smaller, faster and cheaper alternative to relays and valves. In fact, the demonstration established solid state physics (the study of the behaviour of electrons in solid structures) as a promising area of research, now one of the most important areas of science.
Following its launch in 1948, Bell Labs concentrated on moulding the transistor into something that would completely blow away relays and valves. Among the young scientists recruited to Bell Labs in the spring of 1952 was Ian Ross, fresh from Cambridge University with a PhD in electrical and electronic engineering.
”I was faced with a paradox when I began work on the transistor,” recalls Ross, who went on to become the president of Bell Labs between 1979 and 1991 before retiring to pursue his own research interests. ”The first transistors were small and slow, and they lacked power. To make them faster, they had to be smaller; to make them more powerful, they had to be bigger.”
Still bitter about his exclusion from the original find, Shockley made the design of a better transistor his personal mission. Determined to regain the starring role, he treated his colleagues with increasing hostility, driving them away from the key research projects.
The first step to producing a reliable transistor was to develop a more robust design by replacing the fragile surface contact points with three wires which formed a kind of junction deep within the semiconductor material itself – the junction transistor. The second step was to produce better quality semiconductor material – principally silicon, a faster conductor than germanium – more cheaply. It was becoming clearer that the curious behaviour of semiconductors was due in part to naturally occurring impurities that disrupted the crystalline structure of the material, freeing some electrons to move around between atoms. The freed electrons left ”holes” through which current could flow, an effect that could be enhanced by temperature – or by artificially controlling the level of impurity. By producing pure silicon, it became possible to introduce precise levels of impurity in a process called ”doping”, thus producing semiconductors with predictable and consistent abilities.
Once these techniques were perfected, the transistor shrank in size and grew in significance. Its early appearance as a thumbnail-sized gadget in the telephone network was soon overshadowed by IBM’s decision in 1954 to drop vacuum tubes in favour of Shockley’s junction transistors as the core component of its (then) vast computers. The device became a cultural icon as the heart of the transistor radio, soon followed by TVs, cameras, hi-fi equipment and clocks. In 1967 the transistor was barely visible to the naked eye, and today, millions of transistors fit on a thumbnail.
In 1955, Shockley left Bell Labs to set up his own company, Shockley Semiconductor Laboratory, in northern California. But his fascination for technology gave way to a growing obsession with eugenics, the practice of selective human breeding, and he published papers that expounded aggressively racist views. Shockley Labs faded.
By 1958, development was thriving, and miniaturisation joined mass production as the twin goals of transistor research. With a new vision and new techniques, it was not long before the tiny devices were being combined into a larger, more powerful unit: the integrated circuit, and much later in 1971, the microprocessor. If the transistor was the nerve cell of the information age, the microprocessor was the brain.
Desperate to secure a footing in the lucrative calculator business, a young engineer called Gordon Moore, a former employee at Shockley’s laboratory, approached a Japanese company, Busicom, with an idea. In a burst of effort over nine months, nine people directed by Moore compressed the functions of 13 different integrated circuits on to a single silicon slice, creating the world’s first microprocessor – and Intel Corporation.
The microchip, as it quickly became known, combined the efficiency of the transistor with the power of many integrated circuits. The density of transistors on a microchip, even then, made it possible to programme so many combinations of instructions that a microchip could do any information-handling task.
”Nothing compares with the scale and rate of development of the transistor,” says Ross, ”except maybe an epidemic.” The fantastic rate of progress was summarised in 1965 by Moore, who observed that each new microprocessor has twice the capacity of its predecessor and is developed within 18-24 months of the last – put more simply, transistor density on microprocessors doubles every two years.
Moore’s Law remained true throughout the 1980s, but Ross and others know it cannot define progress forever. ”Moore’s Law will hold true until around 2020,” says Ross, ”at which point it will simply collapse.” By then, he says, miniaturisation techniques will have reduced the transistor to less than one tenth of a micron across – less than 100 atoms of matter.
Another shift of vision is enabling the transistor to grow in size and handle more, rather than less, power. These super- transistors (known as IGBTs, or insulated gate bipolar transistors), about the size of a postage stamp, can work in partnership with their microscopic cousins to produce a sort of electronic muscle to complement the microchip brain. ”And of course,” adds Ross, ”biotechnology may lead to partially organic transistors, or maybe quantum or particle physics will reveal more about the nature of atoms and let us progress in that direction.”
None of the transistor’s three inventors has survived to celebrate its half- centenary. But their brainchild has earned the right to be called the most significant scientific development of the century.
Mobile computing, the Internet and wireless communication will one day be joined by electronic business cards, watches that control our central heating, inexhaustible organ replacements and hand-held videoconferencing gear. And all this because of a wedge of plastic, a paperclip and three wise men.