How fission makes heat

A nuclear power plant is a steam plant. It boils water, the steam spins a turbine, the turbine turns a generator. The only unusual part is where the heat comes from. A coal plant burns carbon. A reactor splits uranium atoms. Everything else on the site, from the switchyard to the cooling towers, would look familiar to anyone who has toured a coal plant.

So the whole trick of this lesson is the heat source. Get that right and most nuclear news makes more sense.

Atoms, briefly

Every atom has a nucleus made of protons and neutrons, with electrons around it. The number of protons decides the element: 92 protons means uranium, always. The number of neutrons can vary. Those variants are called isotopes.

Natural uranium dug out of the ground is about 99.3 percent uranium-238 and about 0.7 percent uranium-235, according to the World Nuclear Association. The numbers are the total count of protons plus neutrons. That tiny 0.7 percent is the part that matters, because U-235 splits easily when a slow neutron hits it. U-238 mostly does not.

What happens when U-235 splits

A neutron strikes a U-235 nucleus. The nucleus wobbles, stretches and breaks into two smaller atoms called fission products (barium and krypton are a classic pair, though many combinations happen). Two or three new neutrons fly out, about 2.5 on average. And a lot of energy is released, about 200 million electron volts per split. Most of that energy shows up as the fission fragments slamming into the surrounding fuel, which becomes heat.

Two hundred million electron volts sounds tiny, and for one atom it is. But a kilogram of uranium holds an absurd number of atoms.

The chain reaction, and keeping it at one

Each fission releases about 2.5 neutrons. If every one of them caused another fission, the reaction would grow fast. In a reactor most of them are lost: some leak out of the core, some are absorbed by U-238, the water, the steel or the control rods. The operators and the physics of the core keep the balance so that, on average, exactly one neutron from each fission causes one more fission.

That steady state is called criticality. Plenty of people trip on this word. "The reactor went critical" means it started working normally. That is the goal of running a reactor, and nobody should hear alarm bells in it. Above one new fission per fission the reactor is supercritical and power climbs. Below one it is subcritical and power falls.

Moderators: slowing the neutrons down

Neutrons leave a fission moving very fast. U-235 is far more likely to split when hit by a slow, or thermal, neutron. So most reactors surround the fuel with a moderator, a material full of light atoms that neutrons bounce off and lose speed, like a cue ball hitting other balls.

• Ordinary (light) water: cheap and does double duty as coolant. It absorbs some neutrons, so it only works with enriched fuel.
• Heavy water: water made with the hydrogen isotope deuterium. It absorbs very few neutrons, so it can run on natural uranium. Canada's CANDU reactors use it.
• Graphite: carbon. Used in British gas-cooled reactors, the Soviet RBMK design at Chernobyl and new high-temperature gas reactors.
• None: fast reactors skip the moderator on purpose. Lesson 2 covers why.

Water has a useful safety property. If the water in a water-moderated core boils away or heats up, there is less moderator, fewer neutrons get slowed, and the chain reaction weakens on its own. Engineers call that a negative void coefficient. Chernobyl's RBMK design had a positive void coefficient under some conditions: more steam meant more power. That difference shows up again in lesson 6.

Control rods and delayed neutrons

Control rods are made of materials that soak up neutrons, such as boron, cadmium or hafnium. Push them into the core and the chain reaction slows. Pull them out and it speeds up. Drop all of them in at once (a "scram") and the reactor shuts down in seconds. Pressurized water reactors also dissolve boron in the cooling water to fine-tune power over months as the fuel is used up.

Control would be nearly impossible if all neutrons appeared instantly. They do not. About 0.66 percent of the neutrons from U-235 fission come out seconds later, released by certain fission products as they decay. That small delayed fraction stretches the reactor's response time from thousandths of a second to seconds or minutes, which is long enough for control systems to act.

If a reactor is pushed so far that it goes critical on prompt neutrons alone, the delay no longer helps. That is what happened at the SL-1 test reactor in Idaho in 1961, when one control rod was pulled about 20 inches by hand. Lesson 5 tells that story. The design rule it produced, that the reactor must shut down even if the single most powerful rod is stuck out, is still in use.

Why a power reactor cannot explode like a bomb

This is the most common fear, so be precise about it.

A nuclear weapon needs material that is very highly enriched, typically around 90 percent U-235 or weapons-grade plutonium, slammed together so fast (in millionths of a second) that a huge number of fast-neutron generations happen before the material blows itself apart. Power reactor fuel is 3 to 5 percent U-235. At that dilution there is simply no way to get the runaway fast-neutron reaction a bomb requires. The physics is not close.

That does not mean reactors are harmless. A reactor that loses cooling can overheat and melt its fuel, even after shutdown, because fission products keep giving off decay heat (about 6 to 7 percent of full power right after shutdown, falling over hours and days). Overheated fuel cladding reacts with steam and makes hydrogen, which can explode chemically. Chernobyl had a steam explosion and a fire. Fukushima had hydrogen explosions. Those were serious accidents with real releases of radioactive material. None of them was a nuclear explosion, and calling one that gets the facts wrong.

From heat to the grid

The heat from fission is carried away by coolant (usually water) and used to make steam. Like any steam plant, a nuclear plant loses most of its heat to the environment. A typical large reactor produces about 3,000 to 3,400 megawatts of heat and about 1,000 to 1,150 megawatts of electricity, roughly 33 percent efficiency. That rejected heat is why plants sit near rivers, lakes or the ocean, or build cooling towers.

Where nuclear shines is staying on. The EIA reports that U.S. nuclear plants ran at an average capacity factor of 91 percent in 2025, meaning they produced 91 percent of the energy they would make running flat out all year. As of March 2026 the EIA counted 94 operating reactors at 54 plants in 28 states, providing about 17 to 18 percent of U.S. electricity in 2025, depending on which EIA table you use (the utility-scale generation figure is about 18 percent).