Discrete Components – The Building Blocks of Circuits

All electronic devices are built using some combination of discrete components – the smallest building blocks of an electronic device.

These five major components are used to manipulate the flow of electricity. They can be either discrete (manufactured separately) or integrated (manufactured on the same substrate) and vary widely in size and dimension. We can see each component illustrated in Figure 2-1.

The schematic diagram of light-emitting diode, resistor, capacitor, transistor, and inductor.

Non-integrated discrete components are called “discrete” because they are manufactured separately from one another, as opposed to all on a single wafer. Integrated circuits, on the other hand, are made up of many “functional” components integrated onto a single substrate, the base semiconductor material on which integrated circuits are built, kind of like a house foundation, but at microscopic scale (Saint & Saint, 1999). Instead of manufacturing a given number of transistors, resistors, capacitors, and diodes separately and connecting them afterward, patterns can be etched together on the same chip, or die, using specialized manufacturing technologies like photolithography.

A single advanced die can literally have billions of individual functional components performing the same activities as discrete components manufactured separately from the IC. The important distinction here is that they are integrated and fabricated on the same substrate as the rest of the circuit. In addition, integrated circuits can be fabricated by the thousands, or even tens of thousands, on a single wafer. By tightly integrating these functional components on the same chip and manufacturing thousands at the same time, power is saved (smaller system), speed is increased (greater density), and area is reduced, which further lowers manufacturing unit costs (less materials and fewer process runs) (New World Encyclopedia, 2014).

The reason a smaller system requires less power is simple – it takes more power to drive current along a longer wire than a shorter one. The more tightly integrated a given chip, the closer the functional components are to one another, which means the interconnects and wires that tie everything together are closer to one another as well. Although all these discrete components are important, we will pay special attention to the transistor.

Transistors

When discussing the various kinds of components, transistors get a lot of attention, and for good reason – the transistor is one of the most important inventions of the modern era. Before transistors, computers were made using vacuum tubes that were large, inefficient, and fragile. The first functional digital computer, called the ENIAC, was made of thousands of these vacuum tubes, in addition to capacitors and resistors (Hashagen et al., 2002). It wasn’t until transistors were invented by William Shockley, John Bardeen, and Walter Brattain at Bell Labs in 1947 that computers began their rapid evolution to the microelectronics we see today. These scientists received the 1956 Nobel Prize in Physics for their work (Nobel Prize Outreach AB, 1956).

We can see the differences in vacuum tubes and transistors illustrated by the pictures in Figure 2-2. The vacuum tube (left) looks like a light bulb and is visibly fragile as compared to the transistor (middle) made of metal and enclosed in a protective plastic package. The ENIAC, invented by J. Presper Eckert and John Mauchly at the University of Pennsylvania in 1946, was basically a giant room full of vacuum tubes taking up almost 1,800 square feet and weighing nearly 50 tons (U.S. Army, 1947). We can see it pictured in Figure 2-3 (left) next to a picture of the ENIAC-on-a-Chip implemented (right) (1996). The chip which measures 7.44x5.29 square mm is placed next to a coin for reference, illustrating the progress made by the semiconductor industry in shrinking computer sizes over the course of several decades (Hashagen et al., 2002). Vacuum tubes are still used in select applications like microwave ovens and audio equipment, but transistors have taken over as the main building block of modern electronics. The transistor block diagram in Figure 2-2 (right) depicts the base, emitter, and collector that make up an early bipolar transistor. You can refer back to Figure 2-2 if you need help visualizing transistor structure, which we will discuss shortly.

Two photographs of an upright vacuum tube and a transistor, with a 252 M J E 1 3003 marking and labels of 3, 2, and 1 for its pins, on a textured surface. On the right is a block diagram with numbered transistor pins.

A photograph of a spacious room full of vacuum tubes while a man and a woman monitor the components. On the right is a photograph of its modern version in the size of an American coin.

A transistor is a semiconductor device that, in its simplest form, acts like a switch. Using voltage from a battery or other power source, transistors control something called a gate, allowing them to stop a current in its tracks or allow it to pass through. The pattern of “on” transistors with flowing current and “off” transistors with halted current are the basis for binary computer language used in digital electronics. Computers are able to interpret these patterns of 1’s and 0’s as information (called signals) which they can then manipulate, process, and store. Like a telegraph operator sending a message with different pulses of Morse code, transistors enable computers to process and store information by controlling the flow of individual electrons.

How Transistors Work

For the purposes of demonstration, imagine a PMOS transistor made of a p-type semiconductor sandwiched between two n-type semiconductor blocks. Though the source, gate, and drain are each lacking an electron or carrying an extra, the resulting combination is effectively neutral. Without power or voltage, the transistor is stagnant, and no electrons can flow through the system. Lucky for us, electromagnetic force causes opposites to attract, so if we apply positive voltage to the gate, the negative electrons from the source and drain are drawn to this voltage, while the positive charge in the gate is pushed away. This establishes an opening called a channel through which the electrons can flow, allowing current to pass through (Channel MOSFET Basics, 2018). We can see how this process works in Figure 2-4.

The diagram exhibits 3 neutral semiconductor blocks which are then doped according to their electron number. A set of 3 blocks is composed of a negative source, a positive gate, and a negative drain.

How Transistors Work – A Water Analogy

A little confused? Don’t worry. Let’s use something we’re all familiar with to help explain how this works. Think of charge as water, and the current, or flow, of electricity as the movement of water down a pipe. In theory, we could control the amount of electricity running through a gate by varying the degree of voltage, like twisting a knob to open or shut a valve. This is done in some analog circuits, where voltages and currents are tightly controlled. But in more common digital applications, the transistor simply acts like a switch, blocking electrons or allowing them to pass. This is where we get the binary computer language with which most of us are familiar. Each bit consists of an open gate (1) or a closed gate (0).

We can see this analogy played out in Figure 2-5. Individual “valve transistors” can be combined to make higher level logic gates like “AND” and “OR,” which can then be used to program more sophisticated logic (more on logic gates soon). By opening or closing thousands, millions, or billions of these pipes, the flow of water can be used to create elaborate patterns and sequences. The “brains” of the computer can then read different sequences and use these instructions to save a file, send an email, or take a selfie. Swap out valves for transistors and water for electrons, and you have a complex electronic system!

A diagram illustrated on = 1, off = 0, 0, 1 O R gate, and 1, 1 A N D gate. Below is a network of these components where a section in the western region is magnified.

FinFET vs. MOSFET Transistors

From the hulking 1960s transistors at Bell Labs to the microscopic transistors in modern electronics, transistor structures have been on a path of continuous evolution. There are two primary transistor types – Bipolar Junction Transistors (BJT) and Field Effect Transistors (FET). For illustration's sake, we described a simple bipolar transistor for our water analogy. In reality, bipolar transistors are primarily used for a limited set of applications like power management and signal amplification for wireless and audio devices (Electronics Tutorials, 2021).

Outside of this limited set of use cases, most modern computing devices are built using FETs. Of the FET family, the most popular transistor type is the MOSFET (Metal Oxide Semiconductor Field Effect Transistor) (Teja, 2021). Developed at Bell Labs in the 1970s, it has served as the bedrock for microelectronic design and manufacturing for decades. Without getting into too much detail, this means that a special kind of material called a metal oxide is used to separate the gate from the channel, and that an electric field (via voltage applied to the gate) is used to create the channel between the source and drain. You don’t need to understand the physics behind why this is important, just that there are structural differences that make MOSFETs different from other types of transistors.

As semiconductor technology continues to evolve, engineers have devised new and creative ways to make them more efficient. A new generation of devices called FinFET transistors have helped mitigate performance challenges as transistors reach their physical limitations. While traditional MOSFET transistors are 2-D structures where the gate covers only the top of the channel, FinFET transistors raise the channel through which current can travel, allowing the gate to surround it on three sides (Cross, 2016).

We can see these structural differences illustrated in Figure 2-6. On the left, we have a traditional 2-D planar MOSFET transistor. On the right, we see a more advanced 3D FinFET transistor. By raising the source and drain to surround the gate on three sides, FinFET transistors allow for more efficient control of current through the transistor. The term FinFET isn’t a technical term, it just refers to the fact that the gate is flipped on its side and looks like a “fin.”

Though FinFET transistors are more difficult to manufacture, they allow for greater control over the flow of current, consume less power, and reduce the amount of current leakage (Cross, 2016). Currently, FinFET and MOSFET transistors are the primary transistors in production, although there are new developments on the horizon. Two of these in particular – Gate All Around (GAA) and Nanosheet transistors – will enable greater control and significant performance advantages. We discuss these in our final chapter on the Future of Semiconductors and Electronic Systems.

A diagram illustrates the arrangements of standard M O S F E T and F i n F E T components. The main difference is observed in the alignment of the source, channel, and drain to the gate in the latter.

CMOS

Making high-performance ICs at scale is challenging and expensive, especially as transistors shrink to ever smaller geometries. Most chips today use advanced CMOS (complementary metal-oxide semiconductor) technology to get the job done. CMOS may be used to refer to the circuitry itself, but can also refer to the design methodology and processes that are used to manufacture Integrated Circuits. The “complementary” part of CMOS just means that both p-channel and n-channel transistors are used – believe it or not, early technologies used only n-channel or p-channel transistors, so CMOS was a major development. CMOS has long been the dominant IC design and fabrication technology and enjoys a competitive advantage in power consumption, area requirements, and cost over more specialized alternatives like bipolar semiconductor manufacturing.

Each successive generation of CMOS technology has accomplished these advantages by shrinking transistors and other components through a process called geometric scaling. The key metric in geometric scaling is the gate length – effectively the distance between source and drain. The smaller that length, the smaller your overall circuit and the less distance current needs to travel between components. When you hear people talking about “seven nanometer technology,” seven nanometers is referring to the gate length. When the founder of Intel, Gordon Moore, made his famous prediction, formerly known as Moore’s Law, that computer processing power will double every two years as a result of shrinking transistor sizes, he was referring to geometric scaling. We can see this dynamic at play in Figure 2-7, which charts the transistor counts of major processors released from 1970 to 2020.

A scatterplot represents the transistor counts from 1000 to 50,000,000,000 from 1970 to 2020. A positive correlation is observed from the plotted data.

As transistors shrink, they require less electricity (power), take up less space (area and cost), and enable faster signal processing (performance) (Schafer & Buchalter, 2017). For decades, geometric scaling of transistors has driven functional scaling, which measures meaningful, real-world performance improvements. Because geometric scaling has continued unabated for so long, the engineering community has not had to squeeze as much efficiency out of their designs at each process node. By the time they were ready to roll out a next generation design at a given node or gate length, the next generation of smaller and more powerful transistors were ready for production. In recent years, however, the pace of geometric innovation has slowed as researchers have approached the physical and practical boundaries of transistor sizes. The smaller transistors get, the more expensive it becomes to manufacture them and the harder it is to accurately etch circuit patterns on chip substrates. After all, you can’t build things out of atoms that are smaller than the atoms themselves!

We can see the differences between geometric and functional scaling in Figure 2-8 across generations of transistor technology. Each successive generation of semiconductor manufacturing technology is called a technology node, or process node. These technologies are made of a mix of improved equipment, new materials, and process improvements that enable chip makers to make chips with smaller transistors (measured in nm). The smaller the node, the smaller the transistors and the more powerful the chip. Geometric scaling aims to increase performance by shrinking transistor sizes across the board – the smaller we make transistors, the greater the performance we will receive for next generation chips. An IC made with 3nm transistor technology, for example, can operate much faster, consume less power, and take up less area than one made with 90nm transistor technology. Functional scaling, on the other hand, aims to increase performance by maximizing utilization at existing transistor sizes. It accomplishes this through things like application-specific design, tighter system integration, and developing new packaging and interconnect technologies. We’ll discuss each of these developments in later chapters.

A curve graph represents nanometers per micrometer versus frequency or performance per hertz or clock cycles per second. The data declines from more than 10 micrometers before 1972 and to below 3 nanometers after 2021.

Logic Gates

Logic gates are simple circuits that use Boolean logic to enable simple computations (Fox, n.d.). They are built from as few as two transistors and perform Boolean operators like “and,” “or,” and “not.” In Boolean logic, values can be only true or false, or in the case of transistor-based digital electronics, ON (1) or OFF (0). Logic gates can receive multiple input data signals, which they can compare to one another before spitting out an output signal to the next gate in the system.

You can think of logic gates like a bouncer who only lets people enter the bar if they show a valid ID. Taking our analogy a step further, let’s say the bouncer is acting like a logic gate who has been given strict instructions that when a group comes to the door, everyone must show a valid ID. In digital logic, we call this an “AND gate,” because the gate outputs a 1 if the first input AND the second input are 1 (or true). If you and a friend come to the door, you only gain access if both of you have a valid ID. In this case, the AND conditions have been met, the output would be 1, and you can party the night away. We can see this scenario played out in Figure 2-9.

Other logic gates like OR and NOT work in a similar way. Using logic gates as a lowest common denominator or functional unit, hardware engineers can build complex systems that perform important base functions like addition, subtraction, multiplication, and division.

A diagram of three different combinations of the two inputs and one output of a logic gate. Only a result that equals 1 lead to a photo of two friends getting into a bar, while those with 0 remain at the gate.

The information era would not be possible without transistors. Moore’s rosy 1965 prediction that computing power would double every two years has made it a lengthy 55 years, though there are signs that this pace of innovation is slowing. Arguably the most important invention of the 20th century, the transistor is responsible for the world as we know it today.

We’ve covered a lot of ground in the last two chapters. Don’t worry if you feel a knot in your stomach – silicon engineering is complicated, but we’ve got your back. Everything in semiconductors is strongly interrelated – the further you read, the more the things we’ve covered will make sense. In the next chapter, we’ll cover Step 2 in the semiconductor value chain – Design – and begin to tie all these various elements together.

Summary

In this chapter, we explored the major kinds of discrete components and functional building blocks used to construct electronic systems. We learned what makes discrete and integrated circuits different and the advantages of tighter integration. Using a water analogy, we took a deeper dive into a special type of component – transistors – examining their structure and how they function. We analyzed CMOS technology and two of the most popular transistors in production today – MOSFET and FinFET. Finally, we broke down how transistors fit together to form logic gates and build a more complex system.

Integrated circuits are really just collections of functional components all fit together on a single piece of silicon. Transistors – the most important building block – combine to form the most basic operational unit – logic gates – that run software and keep our computers cranking day after day.