The first Transistor
The MOSFET
The dawn of Silicon Valley
The integrated circuit and the dawn of the Information Age
To the Moon and back
The first electronic Gadget

Semiconductors

Electronics was still a new technology. But the demands for sophisticated electronics increased day by day. Not just for building mobile telephones. One example for the need of effective electronics was for example Computing. The basic functions of a computer were clear and prototypes had already been built with electromechanical relays by the German pioneer Konrad Zuse. The first computers with tubes were built and the limits were immediately reached. The tubes were sensitive, did not last long and consumed a lot of electricity. They also had to be cooled and tubes could not be placed next to each other to avoid overheating. As a result, complex electronics became large and expensive. Without the invention of an alternative, electronics would certainly have come to a standstill. This alternative was the transistor.

How semiconductors work

The basis of the transistor are the so-called semiconductors. Since the beginning of the 20th century, it was known that certain crystals had the property of allowing electricity to pass in one direction only. As already described, Ferdinand Braun used crystals for diodes that could be used to demodulate radio receivers. In the 1920s there was a lot of research going into the properties of semiconductors, especially germanium and silicon.

One of the researchers was Julius Lilienfeld. He was born in Austria-Hungary, studied in Germany and fled to America in the 1930s. Lilienfeld started with crystalline silicon and described how an electronic element could be built based on such a crystal that had the same properties as a triode. Today this is called a field effect transistor (FET). He even filed a patent for this element, although he never built it. It was not possible to build such a transistor at that time.

At this point we should perhaps briefly explain what makes semiconductors, especially silicon and germanium, so interesting. Normally they are non-conductive. This is because all electrons are bound in the crystal lattice. However, if you “contaminate” the crystals with foreign atoms, free electrons or electron holes can form in the crystal. This is called doping (Latin dotare: administer) of the crystal. Especially of n-doping and p-doping. This makes the crystal conductive. This is not such a big deal but things get interesting when you bring n-doping and p-doping together. A non-conductive layer forms at the boundary through which no current can flow. This is because n-layer electrons diffuse into p-layer holes. This creates a zone that is negative on the p-side and positive on the n-side. If a voltage is applied, the non-conductive barrier layer can either be increased or completely removed after a certain threshold, depending on the direction of the voltage. This creates the effect of a diode. Current can only flow in one direction.

The first transistor that Lilienfeld wanted to build was a so-called field effect transistor (FET). This consists of a conductive n layer (it can also be a p layer). A p layer is added to this layer on one or two sides. Like a diode, this initially creates a barrier layer, but a gap remains in which current can flow. If you now apply a voltage to the p layers, you can enlarge the barrier layers as seen with the diode and thereby reduce the current through the n layer until it comes to a complete standstill. This means that on the one hand you have a switch with which the current can be switched on and off, but on the other hand you also have an amplifier with which the current through the n layer can be regulated. With this idea of a transistor you can see that the tube/triode was the inspiration. The gate corresponds to the gate of the tube which controls the flow of electrons.

The problem with building transistors was that they had to be tiny. Before the war it was practically impossible to build an FET or any other transistor. Lilienfeld’s patent remained a vision for a long time.

The first Transistor

The functionality of the FET was known from the Lilienfeld patent. So two groups of engineers tried to build such a transistor. Two Germans, Herbert F. Mataré and Heinrich Welker, conducted research at Telefunken at the FET in the 1940s with only moderate success. In the USA, William Shockley, Walter H. Brattain and John Bardeen worked to solve this problem. They didn’t make any progress either. It was technically extremely demanding to work on tiny doped metal surfaces. Both parties wanted to understand the processes better and conducted research with point-contact diodes that they already knew well. They wanted to make measurements in the immediate vicinity of the diode contacts. As early as 1943, Mataré discovered that two point-contact diodes working in close proximity did not work independently of each other. One diode influenced the conductivity of the other diode. Similar experiments were carried out by Bell Labs. There the “double point contact” was made with gold foil, which was placed over a very sharp wedge and cut through at the tip. This made it possible to achieve very short distances. In fact, the trio Shockley, Brattain and John Bardeen managed to create a “transistor effect”. They were able to use one point-contact to control and amplify the current of the other point-contact. This resulted in the first realized transistor. This wasn’t a FET that they actually wanted to build, but a simple so-called bipolar transistor, but it worked reproducibly. A patent was filed in 1947 and all three researchers received the Nobel Prize in 1956 for this development. Mataré and Welker managed to build the point contact transistor completely independently of Bell Labs just a few months later.

In the months and years that followed, the transistors continued to improve. The sensitive point-contact transistors were quickly replaced by so-called junction transistors, which were cheaper to produce.

No invention of the twentieth century had as great an impact on history as the invention of the transistor. However, when this intentioned was announced by Bell Labs, it was only mentioned on page 46 in a small article in the New York Times.

The MOSFET

Although the first components were now available with the bipolar transistors that could be used as amplifiers as a replacement for tubes, research continued into the development of a field-effect transistor, especially the metal oxide silicon field-effect transistor (MOSFET). This seemed to be the most suitable candidate, especially as a simple switch.

A MOSFET consists of a p-type substrate. Two n-dopings are introduced into the p-substrate at a small distance. The processes described above create insulation at the p-n junctions, which prevents current flow between the n-doped areas.

In the area between the n-doped material, a metal surface is attached, which is separated from the substrate by an oxide layer. This means that no current can flow but an electric field can be generated in the substrate. When a voltage is applied to the metal surface (the gate), an electric field is created. Any free electrons still present from the p-substrate are now drawn through the field between the n-doped areas and remove the insulation there. A current can flow between the source and the drain.

This device is a simple electrical switch or an amplifier. Such a switch is inherently very small and can theoretically be as small as desired. But is it possible to create a MOSFET?

The Dawn of Silicon Valley

According to Bell Labs, the invention of the transistor has three fathers: William Shockley, Walter Brattain and John Bardeen. Bell Laboratories insisted this was a team effort. All three received the Nobel Prize for this. But the three weren’t a team and that was probably due to William Shockley. There were arguments between him and his colleagues, not just Brattain and Bardeen. Ultimately, this resulted in Shockley leaving Bell Laboratories in Murray Hill. He first went to Caltech University in California in 1953. In 1956, his mother, who lived in Palo Alto, a place on the San Francisco Bay also known as the “Bay Area”, fell ill. Apart from the famous Stanford University, there was only a large air force base and a few institutes in the city of San Jose. In the small town of Mountain View Shockley founded his own company, the Shockley Semiconductor Laboratory. Shockley had a very good reputation; he was considered the father of the transistor. This is how he managed to bring the best engineers into his company. Shockley was a great scientist but he was just a bad manager. He was domineering, autocratic, difficult to please and increasingly paranoid. Within a year, his team had had enough. Eight of his best scientists left Shockley and founded their own company called Fairchild Semiconductors in 1957. These employees were later referred to as the “treacherous eight.” Fairchild Semiconductors was the beginning of a new industry, the semiconductor industry. Many other companies emerged from or alongside Fairchild. It used silicon as a semiconductor material.

Later, in 1971, a journalist named Don Hoefler wrote a series about the Santa Clara Valley industry in the Electronic News magazine. He called the series “Silicon Valley, USA”. Since then the area is called the Silicon Valley until today.

Fairchild was the leading company in the early years of the semiconductor industry. Gordon Moore (famous because of Moore’s Law) and Robert Noyce (the inventor of the integrated circuit) founded Intel (Integrated Electronics) in 1968, for a long time the largest semiconductor company in the world. Intel’s competitor AMD was also founded a year later by employees of Fairchild.

The integrated circuit and the dawn of the Information Age

Fairchild focused on a manufacturing process for transistors and diodes. What was needed to build a MOSFET was:

• application of insulation
• Targeted doping
• Application of metallization

A first important step came from a Bell Labs employee Mohamed Atalla. He worked with silicon oxide, which could be applied to a silicone substrate using a chemical process and which has an insulating effect. This provided a tool to create a insulation layer e.g. for a metal gate. In fact, Atalla, together with Dawon Kahng, a Korean American scientist, managed to create the first working MOSFET, 24 years after it was patented by Lilienfeld. 1959 was the birth of the most successful transistor ever.

24 years after Lilienthal patented the MOSFET, the first MOSFET was produced in 1959.

In addition to the use of an oxide layer, one of Fairchild’s employees Robert Norton Noyce also used photolithography to produce semiconductor components. In this process, light-sensitive photoresist is applied. This can be exposed using a mask similar to a photo negative. Where it is exposed to light, it hardens. It is then possible to wash out the unexposed parts. This creates targeted gaps in the photoresist through which underlying structures can be etched away. In this way, it is possible to build structures of oxide layers and metal and also to create windows through which targeted doping is possible. A detailed description how e.g. a MOSFED is build can be found here: Photolithography.

With Photolithography it was not only possible single elements on a silicon wafer. It was possible build several elements and interconnect them. In 1960, Noyce and his colleagues at Fairchild succeeded in producing the first monolithic integrated circuit. The first IC. This technology, which was first achieved in 1960, would change the world forever.

Since that day, when it was first possible to produce and connect 5 transistors on a silicon area, it was possible make transistors smaller every year and place more transisors on an integrated Circuit. Today, more than 60 years later, we can connect billions of transistors on a chip and do incredible things with it, things that were never dreamed of back then. So in 1960, the third industrial revolution and a new age began that we call the information age or digital age.

To the Moon and Back

When the transistor became available in the early 1950s, people in the consumer industry were initially reluctant to replace the established tubes with transistors. However, there was a lot of interest in the military sector. Small, robust electronics were needed to control rockets. The military was therefore the first customer of the integrated circuits, which were initially very expensive to produce.

There was also another area that was interested in the new microelectronics. In 1957, the USA experienced a shock when the Soviet Union sent the Sputnik satellite into orbit. Since then, they have been lagging behind in space travel. So they created the Apollo lunar flight program to send people to the moon and back in the 1960s. For this ambitious goal, the newly founded NASA needed a computer that was powerful enough to navigate the spacecraft but was also small and energy efficient so that it could fit on board the spacecraft. This secured lucrative contracts for Fairchild in particular to miniaturize the necessary circuits. In addition, they were able to improve their manufacturing processes and also became cheaper for future civil applications of the ICs.

The Apollo Guidance Computer (AGC) was one of the first computers to use integrated circuits. „During 1963, the MIT Instrumentation Lab consumed 60 percent of the integrated circuit production in the United States. By 1964, more than 100,000 IC’s had been used in the Apollo program. Approximately 2000 man-years of engineering were consumed in the development of the Apollo computer hardware.“ (Source: IEEE Milestone)

Without the AGC the flight to the moon would not have been possible. It was probably more this Computer than the „rocket science“ of Werner von Braun that the USA won the Space Race since the Soviet Union was not so advanced in electronics at that time.

The first electronic gadget

The market leader in the production of early transistors was Texas Instruments. In 1954 TI commissioned a company called Industrial Development Engineering Associates to build a small portable radio. The result was the Regency TR-1, an AM receiver (AM = amplitude modulation) built with only 5 transistors. It had a small 22.5 V battery, which was enough for almost 20 hours of reception. The size was just 7.62 cm × 12.7 cm × 3.2 cm including a speaker. It was possible to slide it in your shirt pocket. It was released in 1954. Nobody believed that it would have been possible to build a radio that small. It became perhaps the first “must-have” device in existence. Today it would be called a live style product. Everybody wants it just because it was cool. The price was still very high at the time, $49. The quality was poor. It was noisy and unstable. But it was cool and futuristic.

There was one country that was very fond of this type of radio: Japan. There was a small company there called Tokyo Tsushin Kogyo (Telecommunication Engineering Cooperation). This company received a license for transistors from Bell Systems early on. After Texas Instrument stopped producing transistor radios, the Japanese saw the opportunity to enter the American market with their own product. This was in 1955. The radio was a little smaller than the Regecncy TR-1. The company was looking for a good name for this product that would sound good on the US market. They come up with the name Sony. The product was successful and finally the company renamed itself to Sony. This was the beginning of Japanese companies entering and finally dominating the American consumer electronics market. For Sony it was the start of a big rise to the top.

The transistor radio anticipated what would happen to the cell phone 40 years later. It was cool, fit in your pocket, and connected the user to news.