is a semiconductor the same thing as a chip?

No, though the words are often mixed up. A semiconductor is a material, usually silicon. A chip is a finished device built from that material, holding transistors and wiring. People also say semiconductor when they mean the whole chip industry, which adds to the confusion.

why is silicon used instead of other materials?

Silicon is cheap, plentiful and easy to purify. Its biggest advantage is that its natural oxide is a very good insulator and grows straight on the surface, which makes transistor gates simple to build. Other materials beat silicon at specific jobs, such as radio signals or high voltage.

what does doping mean in electronics?

Doping means adding a tiny trace of another element to a silicon crystal to change how it carries current. Adding phosphorus gives extra free electrons and makes n-type silicon. Adding boron leaves gaps that act as positive charges and makes p-type silicon. Joining the two types creates useful devices.

what is a transistor?

A transistor is an electrical switch with no moving parts. A voltage on its gate opens or closes a channel, letting current pass or blocking it. Transistors are wired into logic gates, and logic gates are wired into processors and memory. Billions fit on a single chip.

What is a semiconductor?

Think about water in a pipe. A copper wire is an open pipe. Water runs through it whenever there is pressure behind it. A rubber sheet is a wall. Nothing gets through at all.

A semiconductor is a pipe with a tap on it. Sometimes it lets current through. Sometimes it blocks it. What makes it useful is that you get to decide which, and you can decide billions of times a second.

That is the whole reason the modern world runs on these materials. A semiconductor lets you build a switch with no moving parts.

What makes a material a semiconductor?

Materials are usually sorted by how easily electricity passes through them.

Metals such as copper and aluminium conduct well. Their electrons are loose and free to move. Insulators such as rubber, glass and dry wood hold their electrons tightly, so almost nothing moves.

Semiconductors sit in between. Silicon is the main one. In pure form, at room temperature, it barely conducts at all. Give it a little heat, a little light, or a small push of voltage, and it starts to conduct.

That in-between habit is the whole point. A material that always conducts cannot be switched off. A material that never conducts cannot be switched on. Only the middle ground gives you a switch.

How do you control it? By doping.

Pure silicon on its own is not much use. The useful step is to spoil it on purpose.

Doping means adding a tiny trace of another element into the silicon crystal. Add something with a spare electron, such as phosphorus, and the silicon gains extra loose electrons to carry current. That is called n-type, for negative.

Add something short of an electron, such as boron, and the crystal is left with gaps where electrons should sit. Those gaps shift about and behave like positive charges. That is called p-type.

The amounts are very small. A few added atoms per billion atoms of silicon can change how the material behaves. This is why purity rules this industry, and why the factories are kept so clean.

From material to switch

Put a piece of p-type silicon against a piece of n-type silicon. Something useful happens at the join.

Current crosses that join in one direction and is blocked in the other. You have made a diode, which is a one-way valve for electricity. A solar cell and a light-emitting diode are both built on the same join, working in opposite directions.

Now add a third region and a control terminal on top, kept apart by a very thin insulating layer. That control terminal is called the gate. Put a voltage on the gate and a channel opens underneath it, so current can flow between the other two terminals. Take the voltage away and the channel closes.

That is a transistor. It is a tap for electricity, turned by electricity. Everything else is built out of it.

How do switches become computing?

One transistor is a switch. Wire a few together and you get a logic gate, which is a small circuit that answers a yes or no question about its inputs.

One gate might say that its output is on only when both inputs are on. Another might say that its output is on when either input is on. Combine gates and you can add two numbers. Combine more and you can store a number, compare two numbers, or follow a list of instructions.

A modern processor holds billions of transistors on a piece of silicon smaller than a fingernail. They are all made at the same time, in the same run of steps. That is why chips are cheap even though they are fantastically complicated.

How is a chip actually made?

The short version has four stages.

Purify. Sand is mostly silica. It is refined until the silicon is about as pure as any material people make.
Grow and slice. The pure silicon is grown into one large single crystal, then sliced into thin round wafers.
Print. Layer by layer, patterns are projected onto the wafer with light, then etched, filled and doped. A leading chip goes through many hundreds of steps.
Test and cut. The finished wafer is tested, cut into separate chips, and packaged.

The printing stage is why these factories are so rare and so costly. Our explainer on the chip fab covers why only a handful of places on earth can do the hardest version of it.

Why silicon, and what else is used?

Silicon won for dull reasons, and dull reasons are the strongest kind.

It is everywhere, so it is cheap. It can be purified to an extreme degree. Best of all, its natural oxide is an excellent insulator that grows straight on the surface. That makes the gate easy to build, which matters more than almost anything else.

Other semiconductors handle jobs silicon does badly. Gallium arsenide and its relatives deal with radio signals well, so they turn up in phone and satellite parts. Silicon carbide and gallium nitride cope with high voltage and high heat. They appear in solar inverters, in electric vehicle chargers, and in the small fast phone chargers now on sale. Those two materials waste less energy as heat, which is why they keep spreading.

Where do you already meet them?

Everywhere with a battery or a plug.

The obvious places are phones, laptops and televisions. The less obvious ones matter more by sheer number. A solar charge controller, a motorbike ignition, a water pump controller, a mobile money terminal, a set-top box and an LED bulb all contain semiconductors.

Chips also come in kinds. A general processor runs any instruction you give it. A graphics chip does the same simple sum across huge amounts of data at once. A newer kind is tuned for machine learning, and our guide to the NPU explains how that differs from a graphics chip.

What is coming next, and what can you check?

The old pattern was steady shrinking. Transistors got smaller with each cycle, so chips got faster and cheaper. That pattern has slowed, and our explainer on Moore’s Law covers why.

The industry has three answers. Change the shape of the transistor so the gate wraps further around the channel. Stack chips on top of each other rather than only spreading them out. Build parts tuned for one job rather than making one part faster at everything.

Do one thing today. Pick up the charger for your phone and weigh it against an older charger of the same power. If the new one is much smaller and cooler, the reason is a change of semiconductor material inside it. That is a chip decision you can hold in your hand.

Just Out Tech explains new research in plain language. This article was drafted with AI assistance and checked by a human against the original source.