How Semiconductors Work: The Physics Behind Computer Chips

Every device that computes anything — a phone, a car's engine controller, a pacemaker, a data center server — ultimately relies on the same basic building block: a transistor made from a semiconductor material, almost always silicon. A transistor is a switch with no moving parts, able to flip between an "on" state that lets current flow and an "off" state that blocks it, controlled entirely by a small voltage applied to one of its terminals. The physics that makes this possible starts with a material property most people never think about: how easily different substances conduct electricity.

Conductors, Insulators, and the Space Between

Metals like copper conduct electricity easily because their outer electrons move freely between atoms. Materials like rubber or glass are insulators because their electrons are tightly bound and essentially never move. Silicon sits in an unusual middle ground: in its pure form it barely conducts at room temperature, but its conductivity can be adjusted, deliberately and precisely, by changing the material at the atomic level. That adjustability, not any property silicon has on its own, is the entire reason the electronics industry is built on it rather than on a naturally better conductor.

Doping: Adding Impurities on Purpose

Manufacturers introduce trace amounts of other elements into a silicon crystal, a process called doping, and the specific element chosen determines what kind of semiconductor region results. Adding phosphorus, which has one more outer electron than silicon, creates "n-type" silicon with extra free electrons available to carry current. Adding boron, which has one fewer outer electron, creates "p-type" silicon with extra "holes," effectively an absence of an electron that behaves like a positive charge carrier moving through the material. Neither n-type nor p-type silicon does much interesting on its own. The useful behavior appears where the two types meet.

The P-N Junction: A One-Way Gate

Where a region of p-type silicon touches a region of n-type silicon, electrons from the n-side and holes from the p-side diffuse across the boundary and combine, creating a narrow zone depleted of free charge carriers. This depletion zone acts like a one-way gate: current flows easily when voltage pushes carriers toward the junction in one direction, but the depletion zone widens and blocks current when voltage is applied the other way. A device built around a single p-n junction is a diode, and it was the earliest practical semiconductor component, used to convert alternating current to direct current long before anyone built a transistor out of the same principle.

From Diode to Transistor to Logic Gate

A transistor extends the same idea with three regions instead of two, typically arranged n-p-n or p-n-p, where a small voltage or current applied to the middle region controls whether a much larger current can flow between the two outer regions. This is the switch: apply the control signal and current flows (on); remove it and current stops (off). Wire enough transistors together in the right pattern and they form logic gates — AND, OR, NOT, and combinations of these — which are the physical implementation of Boolean logic. Every calculation a computer performs, no matter how complex, ultimately reduces to enormous numbers of these simple on-off switches flipping in coordinated sequence, timed by a clock signal that can pulse billions of times per second in a modern processor.

How Small These Switches Actually Are

Chip manufacturers now fabricate transistors with features measured in single-digit nanometers, a scale where a few dozen silicon atoms span the narrowest part of the switch. Building at that scale requires photolithography: a chip-sized pattern is projected through a lens system onto a silicon wafer coated with a light-sensitive material, and hundreds of such patterning steps, alternating with etching, doping, and material deposition, are layered on top of each other to build a finished chip. A single modern processor can contain tens of billions of transistors, all functioning correctly, which explains why fabrication plants cost billions of dollars to build and why only a handful of facilities worldwide can manufacture the most advanced chips. The National Institute of Standards and Technology's CHIPS program office tracks the measurement and standards challenges involved in pushing this manufacturing process further.

Why the Same Physics Powers Solar Panels

The p-n junction that makes a transistor switch also does something else useful: when sunlight strikes a p-n junction, photons can knock electrons loose and the junction's built-in electric field sweeps them in a consistent direction, generating current instead of merely switching it. This is exactly the mechanism behind how solar panels convert sunlight into usable electricity, and it is one of the clearest illustrations that a transistor and a solar cell are built from the same underlying semiconductor physics, just wired and doped for different jobs. The precise, controllable behavior of doped silicon junctions is also what carries out the logical operations behind the encryption that protects data moving across a network, since every encryption algorithm ultimately runs as a sequence of transistor switches.