The term semiconductor covers a lot of territory, from big famous chips like Pentium microprocessors down to little tiny components that are almost too small to see. No single company makes or sells all these different kinds of semiconductors, and no engineer, programmer, or purchasing manager can keep them all straight. Everyone specializes in just one small aspect of semiconductors.
Just as with the medical profession, no one in the semiconductor industry can stay current with—or even fully understand—the whole industry at once. Doctors might specialize in podiatry, surgery, oncology, or pediatrics. Likewise, business managers and engineers might specialize in memory chips, semiconductors for radio and television broadcasting, or high-voltage components. There are entire trade groups and professional conferences devoted solely to tiny and arcane niches within the larger semiconductor industry. Market research firms sift and sort revenue and shipment data by categories that might make little sense to outsiders.
For our purposes, we’ll divide semiconductors into a few big categories in this chapter, and then delve into some of the more interesting ones in more detail, both here and in later chapters. These divisions are well accepted in the industry; we’re not being arbitrary. You’ll find that professionals working in one major division generally don’t understand (or care) very much about how semiconductors in the other divisions work, how they’re sold, or who uses them. Like doctors at a large medical convention, they might be able to discuss some basic ideas, but the detailed conversations are intelligible only to others in the same specialized field.
Semiconductor Family Tree
Figure 2.1 shows an overall “family tree” for all types of semiconductors. It starts with raw silicon (and other natural materials) on the left and moves to progressively more elaborate types of manufactured semiconductors on the right. Each box along the way represents a different type of semiconductor unit, or component, that you can buy today. (In this chapter we use the word component to refer to all different types of semiconductor products. Later on, we’ll fine-tune our terminology.)
Figure 2.1 is also a kind of evolutionary chart. Components on the left are more primitive and basic, whereas those on the right are more sophisticated or have been built up from the simpler components. The older components haven’t been replaced by any means. They’re all alive and well and selling millions or billions of units per year. (You can see a lot more detail on the actual sales volumes and revenue in Chapter 5, “Business and Markets.”) As a rule of thumb, the components on the left are the cheapest, but sell in the greatest numbers. As you move to the right, the components become more expensive but sell in lower volumes.
From Little Acorns to Great Oak Trees
Simple electronic components like resistors and capacitors are fundamental building blocks for all electronic products. They’re made from simple natural materials, such as silicon, carbon, or aluminum. These simple components are tiny, cost just pennies each, and are manufactured in huge volumes for customers around the world. A resistor, capacitor, or diode might be about as big as a grain of rice.
Transistors are also pretty elementary components, but they represented a major leap forward when they were invented in 1947. Two Bell Labs engineers, John Bardeen and Walter Brattain, and their boss, William Shockley, invented the first transistor, which opened the doors to further evolution in electronics. Transistors turned out to be the key component in a wave of innovation through the 1950s and 1960s. Transistors helped define “digital” electronics, ultimately ushering in the era of digital computers and other modern conveniences. Transistors also played a role in the invention of integrated circuits (ICs), which represented another leap forward in semiconductor manufacturing.
If you combine these basic electronic components—resistors, capacitors, diodes, inductors, and transistors—you can create just about any electronic or computer-controlled device in the world today. These humble components are the bricks and mortar of all modern electronics. Like bricks and mortar, it’s generally only a matter of combining these in the right order or the correct proportions that determines what grand structure you create. Electrical engineers are experts at combining these basic components to make wondrous devices.
Figure 2.1 shows that resistors, capacitors, diodes, inductors, and transistors can be combined to create more complex components such as logic gates, converters, and sensors. Mixing these basic components together into one larger component is called “integrating” them. This is where the term integrated circuit comes from. An IC is a mixture of simpler components combined to create a more complex component. A typical IC might sell for less than a dollar and be about the size of a postage stamp.
Building on these components, engineers create ever more highly integrated components. Some examples are memory chips, microprocessors, programmable logic, or network ICs—complex and sophisticated components that contain more than 1 million basic components within them. These ICs would typically sell for several dollars and be as big as a matchbook.
This sequence of integration on top of more integration can go on more or less forever. Electrical engineers constantly combine the components their predecessors created and integrate still more components into them. This leads to ever more complex ICs with more features and more power. This continual up ward spiral has given us components that have more than 50 million constituent components inside them. It would be impossible to re-create a Pentium 4 processor, for example, out of 50 million individual transistors, yet the real thing fits on a slice of silicon no bigger than your fingernail.
As you can see, components can be integrated to create new, bigger components, and so on. Each generation of this process has been given a name. When the four basic electronic components are integrated to make an IC, it’s sometimes called small-scale integration (SSI). When SSI components are integrated still further to create a second-generation IC, it’s called medium-scale integration (MSI). After MSI comes large-scale integration (LSI), and then very large-scale integration (VLSI). After this point, the terminology kind of runs out, so VLSI is a now blanket term covering any moderately large chip.
Integrated Circuits versus Discrete Components
Not all electronic components are ICs, as you can see from Figure 2.1. Any nonintegrated component is called a discrete, or separate, component. When people discuss semiconductors it’s usually ICs they’re talking about. In the popular imagination, all electronic components are ICs, when in reality only about 20 percent are ICs. The rest are discrete components.
ICs are also called chips. All three terms are synonymous and you can use them interchangeably. You might also hear people say “computer chip” or “microchip,” but the cognoscenti frown on these terms. Throughout this book you can transpose “chip” with “integrated circuit” or “IC” and no one will complain.
Discrete components tend to be very small, about the size of a grain of rice. ICs are generally bigger, rectangular, and black with small silver “legs” sticking out, as shown in Figure 2.2. Some ICs are fairly small (these are the first-generation SSI chips). Some ICs are comparatively huge, such as a 50-million-transistor microprocessor chip. If a chip looks big on the outside, it’s a safe bet there are millions of basic components integrated on the inside.
Jack Kilby and Robert Noyce, two engineers working independently and unbeknownst to each other, both developed the first ICs in 1958. (Noyce received the patent for it, but Kilby received a Nobel Prize in 2000.) Kilby created his IC during his 1958–1959 Christmas break from Texas Instruments. Robert Noyce, then at Fairchild Semiconductor, developed a similar chip using similar materials at the same time.
Digital versus Analog Components
You might have noticed from Figure 2.1 that some of the components are labeled as analog components, others are called digital components, and transistors seem to be a little of both. What does this mean? Not a lot, unless you’re planning a career in electronics. Analog and digital describe differences inside the component and how it handles electricity.
Digital components treat electricity as either on or off, with no in between. Digital components deal in black and white, true and false, and yes and no. They’re also sometimes called logic components because they see everything as neat and logical, with no gray areas. Computers are made up mostly of these digital, or logical, components.
Analog components treat electricity as a steadily flowing stream. They deal in smooth variations of gray, so to speak. Analog components smooth, shape, and modify the electricity passing through them. If you like, you can think of digital components as the ordered, left-brained, rules-oriented components and analog components as the changeable, artistic types. (Chapter 9, “Theory,” explains a lot more about the difference between digital and analog components.) Figure 2.3 represents how electricity passes through digital and analog components.
Digital and analog components are often used together in the same product, such as televisions. Computers, cell phones, and microwave ovens also include both digital and analog components. Like yin and yang or hardware and software, they are two fundamentally different things that complement each other.