It's a Small, Small World: The Chemistry of Microchip Fabrication

Your beloved pooch has a tendency to run off. It’s time to pay a visit to the vet to get her microchipped.

On the morning of your appointment, the alarm on your smartphone wakes you up. The house is frigid, so before getting out of bed you use your phone to crank up the thermostat. A steaming hot cup of coffee—perfectly brewed to your specifications—awaits you when you go downstairs. You scoop up your pup and head to the car, inserting a fat-headed key—a chip key—into the ignition. A GPS satellite provides directions.

The veterinarian uses a hypodermic needle to painlessly implant a microchip—the size of a grain of rice—just below your dog’s skin, usually between the shoulder blades. You pay using a credit card, embedded, of course, with a chip. Now, if your dog runs off again, a simple scan of the microchip will reveal your contact info.

We live in the age of the "Internet of Things." More and more of our electronic devices are digitally connected. At the heart of it all is the microchip. To understand how microchips work, we must first take a quick tour of the technology that made such a thing possible.

Transitors

In 1947, researchers at Bell labs had a breakthrough: They invented the transistor, which replaced the bulky and unreliable vacuum tubes of first-generation computers. The transistor is widely considered to be the greatest invention of the 20th century.

Transistors are the building block of all computers. Just like the human body contains billions of cells, a modern computer contains billions of transistors.

The word transistor is an amalgamation of the words transmitter and resistor—they can either transmit current or resist it. They are often referred to as electronic switches. Transistors not only control the flow of current, they also amplify it.

By using a transistor, a small input current can create a much larger output current. A transistor is like putting your finger over the end of a water hose where the water represents the electrons in the transistor. If you completely cover the end of the hose, no water comes out, but if you cover it partially, the water comes out with a much stronger flow.

Transistors act like tiny on-off switches and can be used to generate binary data—an open gate on a transistor might be “1” and a closed gate “0.” As these tiny switches oscillate on and off, they can perform a multitude of operations.

Computers use binary code, representing every bit of data as either a 0 or a 1. Any number or letter can be represented by a string of 0s and 1s. The article you are reading is stored on a computer in just this way—even the pictures!

Semiconductors

Transistors are made from semiconductors, which can have properties similar to conductors or insulators. Conductivity is best explained by band theory. In a solid material, orbitals of similar energy can overlap and form a band.

Overlap of orbitals that form bonds between atoms create an energy band known as the valence band. Electrons in the outermost shell of an atom are known as valence electrons, and comprise the valence band. Because this band represents bonding orbitals, it is lower in energy. Overlap of orbitals that do not form bonds, antibonding orbitals, create the energy band called the conduction band, which is higher in energy.

Electrons need to move between orbitals without losing too much energy to conduct electricity. In metals, the valence and conduction bands overlap, allowing for the free movement of electrons between the bands. 

Most metals are conductors, while most nonmetals are insulators. Insulators have a large gap between the top of the valence band and the bottom of the conduction band such that electrons cannot move between the bands easily and, therefore, do not conduct electricity.

In semiconductors, though, the gap between the valence and conduction bands is small—small enough in some semiconductors that when heated, valence electrons can be jarred loose to allow electrons to move between the valence and conduction bands. 

Customizing the Gap: Doping

The first transistor used germanium, but today silicon is the semiconductor of choice. It is no accident that the birthplace of the modern computer—the San Francisco bay area in California—was dubbed Silicon Valley.

If voltage is applied to pure silicon, however, not much happens—at room temperature, pure silicon is an insulator. Before it can conduct electricity, it must be doped. Doping means adding an impurity to the silicon lattice.

Adding an element to silicon that has more valence electrons than silicon, such as phosphorus or arsenic, adds additional electrons to the silicon lattice. This type of doping is called n-type doping.

Conversely, doping with an element with fewer than four valence electrons, such as boron or aluminum, means there is one fewer electron in the valence band; this electron vacancy is called a “hole.” This type of doping is called p-type doping.

Microchips

Wires, a power source, and a device constitute a circuit. Circuits enable electrons to flow in a controlled path so they can do useful work. If you shrink down a whole bunch of circuits onto a very thin wafer of silicon, you have a microchip. A key piece of this electronic circuitry is the transistor.

More transistors means greater processing speed. Intel’s best chip in 1985 contained 250,000 transistors and took more than four hours to do what today’s chip—containing billions of transistors—can do in a second.

Cramming this staggeringly large number of transistors into such a tiny space—most microchips would fit comfortably on your fingernail—seems like the stuff of science fiction. The technological wizardry needed to build a chip is as wondrous as it is arduous, requiring hundreds of steps over a three-month period.

But the first step is rather simple: Get some sand, and lots of it. Sand (SiO₂) is the source of silicon in electronics. The sand is physically and chemically processed until it is refined into silicon ingots that have a purity of 99.9999999% silicon—“nine nines” in industry speak. Even the smallest impurity can wreak havoc on a system.

The Making of a Microchip

The making of a microchip is a multistep process. The silicon ingot is first sliced into wafers, which are then cleaned and heated to create a durable silicon dioxide coating. The thin wafers are then coated with a light-sensitive polymer.

Photolithography is used to create a pattern that can be chemically etched into the silicon dioxide coating to expose the pure silicon, so it can be n- or p-doped, as needed.

This same process is repeated over and over on the wafer, which will eventually be cut up into hundreds or even thousands of individual chips. Chips typically range in size from a square millimeter to a square centimeter. A central processing unit (CPU), which is essentially one giant microchip, will be larger.

Although transistors are an integral part of a microchip, a host of other electronic components, such as resistors, diodes, and capacitors must also be incorporated into the wafer. Each of these structures is etched into the silicon substrate.

Transistors, though, is where the magic happens: A single microchip can contain 160 billion transistors!

Maximizing Space

Even with nanosized transistors, there is only so much room on a silicon wafer. Consider what builders in large cities do when they run out of room: They build upward. Today’s microchips are more like three-dimensional skyscrapers than two-dimensional circuit boards: A single chip might contain more than a hundred layers!

All of these layers need to be wired together often using nm-thick copper wires, but even though copper is an excellent conductor, electrons still encounter resistance. The thinner the wire, the greater the resistance. Resistance produces heat, and heat draws power.

Today, research is shifting from making chips smaller to making them more energy efficient. The use of carbon nanotubes instead of copper has shown promise; they are lighter and offer less resistance.

A robust supply of microchips presents a national security concern, with any number of crucial technologies—think missiles, jets, and radar, to name a few—relying on them. From farms to factories, microchips are firmly embedded in our world. Without them, life as we know it would grind to a halt—and your dog might never be found.