Wikipedia Deep Dive

CANDU reactor

14 min read

Here's a puzzle that stumped Canadian engineers in the 1950s: How do you build a nuclear reactor when you can't make the enormous steel pressure vessels that every other country uses, and you don't have access to uranium enrichment technology?

The answer they came up with was so clever that it's still generating electricity today, from Ontario to Romania to Argentina. They called it CANDU—Canada Deuterium Uranium—and it works on a principle that sounds almost like cheating.

The Heavy Water Trick

To understand what makes CANDU special, you first need to understand what happens inside any nuclear reactor. Uranium atoms split apart, releasing enormous energy and also spraying out fast-moving neutrons. Those neutrons need to hit other uranium atoms to keep the reaction going. The problem is that neutrons fresh from fission are moving too fast—they zoom right past the uranium atoms they're supposed to hit.

So you need to slow them down. You do this with a moderator, a substance that neutrons bounce off of like pinballs, losing a bit of speed with each collision until they're moving slowly enough to trigger more fission reactions.

Most reactors use regular water as their moderator. Water works beautifully for this purpose—hydrogen atoms are almost exactly the same mass as neutrons, so they absorb a lot of energy in each collision, like when a moving billiard ball hits a stationary one and nearly stops dead. But regular water has a nasty habit. It doesn't just slow neutrons down; it also absorbs some of them entirely. And once a neutron gets absorbed by a water molecule, it can't split any more uranium atoms.

This absorption creates a problem. Natural uranium contains only about 0.7 percent of the fissile isotope uranium-235, the kind that actually wants to split apart. The rest is uranium-238, which is much harder to fission. When regular water swallows too many neutrons, there aren't enough left to sustain a chain reaction with such dilute fuel. The reactor just fizzles out.

The standard solution is enrichment. You take natural uranium and process it through massive, expensive facilities that increase the concentration of uranium-235 from 0.7 percent up to somewhere between 2 and 5 percent. Now you've got enough fissile material that the chain reaction can sustain itself even though regular water is eating some of your neutrons.

But Canada in the 1950s didn't have access to enrichment technology. That was closely guarded military technology, controlled by the United States, which wasn't sharing. Canadian engineers needed another solution.

When Hydrogen Gains Weight

Water—regular water, the kind you drink—is made of two hydrogen atoms bonded to one oxygen atom. Each hydrogen atom has a single proton in its nucleus and nothing else. But there's a variant of hydrogen, called deuterium, that carries an extra neutron alongside that proton. When you make water using deuterium instead of regular hydrogen, you get "heavy water"—chemically almost identical to regular water, but about 10 percent denser.

Heavy water has a remarkable property. Because its hydrogen atoms already have that extra neutron, they're much less hungry to capture more. Neutrons passing through heavy water mostly just bounce off, losing energy with each collision but rarely getting absorbed. This means you can sustain a nuclear chain reaction using nothing but natural, unenriched uranium.

There's a trade-off, though. Deuterium is twice as heavy as regular hydrogen, so each collision transfers less energy from the neutron—imagine a ping-pong ball bouncing off a golf ball rather than another ping-pong ball. You need more collisions to slow the neutrons down, which means you need a thicker layer of moderator between your fuel rods. Your reactor gets bigger.

Much bigger.

The Pressure Problem

In a typical nuclear power plant, the reactor core sits inside a massive steel vessel that can be pressurized to about 150 times atmospheric pressure. This pressurization keeps the water from boiling even at temperatures over 300 degrees Celsius. Hotter water carries more heat, which means you can extract more energy from a smaller reactor core.

Building one of these pressure vessels is an extraordinary industrial feat. You're casting and machining a single piece of steel the size of a small building, with walls thick enough to contain immense pressures and resist constant bombardment by neutrons. In the 1950s, only a handful of foundries in the world could do it, and none of them were in Canada.

When engineers tried to order a relatively small pressure vessel for an early experimental reactor called NPD, they couldn't find a Canadian manufacturer capable of making it. They had to import it from Scotland.

For commercial-scale heavy water reactors, the problem was even worse. Because heavy water requires more moderator volume, a CANDU-style reactor needs a core much larger than an equivalent enriched-uranium reactor. The pressure vessel would have to be enormous—and manufacturing such a thing in Canada seemed essentially impossible.

So the Canadians did something clever. They got rid of the giant pressure vessel entirely.

Tubes Instead of Tanks

Instead of placing all the fuel inside one giant pressurized container, CANDU spreads it across hundreds of small pressure tubes, each about 10 centimeters in diameter. These tubes are much easier to manufacture than a single massive vessel—you can make them in a regular factory rather than a specialized heavy-engineering facility.

The pressure tubes run horizontally through a large tank called a calandria, which is filled with heavy water serving as the moderator. But here's the key insight: the calandria itself doesn't need to be pressurized. It sits at near-atmospheric pressure and relatively cool temperatures, meaning it can be built from ordinary steel using ordinary manufacturing techniques.

To keep the hot, pressurized coolant in the fuel tubes from heating up the surrounding moderator, each pressure tube is wrapped in a second tube called a calandria tube. Between the two tubes sits a thin layer of carbon dioxide gas, acting as insulation.

This design solved Canada's industrial problem beautifully. You could build the whole reactor domestically, using Canadian factories and Canadian workers, without needing to import exotic pressure vessels from overseas.

But it also created an unexpected bonus: you could refuel the reactor without ever shutting it down.

The Robot Ballet

In a conventional pressurized water reactor, refueling is a major operation. You have to shut down the reactor, wait for it to cool, open up the pressure vessel, and swap out roughly a third of the fuel assemblies. This takes weeks and means the plant isn't generating any electricity during that time.

CANDU's tube design made possible something elegant. Two robotic machines, one on each side of the reactor, can latch onto any pressure tube while the reactor is running at full power. One robot pushes in fresh fuel bundles while the other catches the spent fuel that gets pushed out the far end. The whole operation takes about an hour per tube, and the reactor keeps humming along the entire time.

On-line refueling, as it's called, offers several advantages beyond just avoiding shutdowns. If a fuel bundle develops a leak and starts releasing radioactive fission products into the coolant, operators can locate which tube contains the problem and remove that specific bundle. In a conventional reactor, you'd have to wait until the next scheduled shutdown to fix such a problem, living with elevated radiation levels in the cooling system the whole time.

What the Fuel Looks Like

Each CANDU fuel bundle is a surprisingly compact object—about 10 centimeters in diameter and 50 centimeters long, weighing roughly 20 kilograms. It's made of thin tubes filled with ceramic pellets of uranium oxide, all bundled together and held in place by end plates.

The tubes themselves are made of a zirconium alloy called zircaloy, which has a useful property: neutrons pass through it almost as if it weren't there. This "neutron transparency" is essential because you want neutrons moving freely between fuel bundles, not getting absorbed by structural materials.

Older CANDU designs used bundles with 28 or 37 fuel elements. Newer designs use something called CANFLEX, which packs 43 elements of two different sizes into the same bundle. The mixed sizes allow the reactor to run at higher power without the hottest fuel elements overheating—a clever bit of engineering that shows how the design has matured over decades.

A typical CANDU pressure tube contains 12 or 13 fuel bundles lying end to end, like a string of sausages. The whole reactor might have 380 or more pressure tubes, giving you something on the order of 5,000 fuel bundles in the core at any given moment.

Burning Fuel More Efficiently

One might assume that using unenriched uranium is simply a cost-saving measure, trading performance for economy. The reality is more interesting. CANDU reactors actually extract more energy from each ton of mined uranium than light-water reactors do—somewhere between 30 and 40 percent more.

How can this be? The answer involves some subtle nuclear physics.

In any reactor, uranium-238 (the common isotope) occasionally captures neutrons and transforms into plutonium-239, which is itself fissile. This plutonium contributes to the chain reaction, partially compensating for the uranium-235 that gets used up. But eventually, fission products accumulate—the leftover fragments from split atoms—and many of these absorb neutrons. When there are too many neutron-absorbing fission products relative to fissile material, the reaction sputters out and you need fresh fuel.

In a light-water reactor, you start with enriched uranium containing maybe 4 percent uranium-235. By the time the fuel is spent, you've burned through most of that 235, but a lot of energy potential remains locked in the remaining uranium and the plutonium that formed during operation.

CANDU's heavy water moderator runs at a lower temperature than light-water moderators, which means the neutrons settle to a lower average energy. Lower-energy neutrons are actually better at causing fission. This lets CANDU squeeze more fissions out of the same amount of fissile material.

There's another factor at work too. Heavy water moderators are physically larger, which seems like it should waste neutrons by letting more of them escape the core. But the benefit of reduced neutron absorption outweighs this leakage penalty. The net result is that CANDU gets more useful fissions per unit of uranium than reactors using enriched fuel.

The Stability of Slowness

Reactor engineers care a great deal about something called reactivity feedback—how the reactor responds when conditions change. A stable reactor is one where any increase in reaction rate triggers mechanisms that push the rate back down. An unstable reactor amplifies small changes into dangerous runaway conditions.

CANDU has some interesting characteristics here. It has what's called a positive void coefficient, meaning that if the coolant starts boiling and forming steam bubbles (voids), the reaction rate actually increases. This sounds alarming—it's one of the factors that contributed to the Chernobyl disaster in a different reactor type.

But CANDU has features that compensate. The huge mass of relatively cool moderator in the calandria acts as an enormous thermal buffer. Even if something goes wrong in the fuel tubes, that mass of heavy water soaks up heat and slows any temperature excursion.

There's also an exotic effect involving deuterium nuclei. The binding energy holding a deuterium nucleus together is relatively low—only 2.2 million electron volts. Some of the gamma rays and energetic neutrons produced by fission carry enough energy to break deuterium nuclei apart, releasing additional neutrons.

This might seem like it would make the reactor less stable, adding extra neutrons when the reaction rate goes up. But the fission fragments that emit these gamma rays have half-lives ranging from seconds to hours or even years. This creates a delayed feedback mechanism that smooths out the reactor's response and gives operators more time to react in abnormal situations.

The large size of CANDU cores also provides what engineers call spatial decoupling. If something causes the reaction rate to spike in one part of the reactor, it takes time for that change to propagate to other areas—the neutrons have to traverse significant distances through the moderator. This gives automatic safety systems and human operators more time to intervene before a local anomaly becomes a core-wide problem.

Fuel Flexibility

One of CANDU's most remarkable properties is its willingness to burn almost anything fissile. Because the neutron energy spectrum in the reactor depends on the moderator rather than the fuel, you can swap in different fuel types without fundamentally changing how the reactor behaves.

This has practical implications. CANDU can burn recovered uranium from reprocessed spent fuel, mixed oxide fuel containing plutonium, or even thorium-based fuels. It can accept slightly enriched uranium for higher power density when that makes economic sense. Some CANDU operators have experimented with burning spent fuel from light-water reactors, essentially using CANDU as a way to extract additional energy from material that other reactors consider waste.

This flexibility was built into the design from the beginning. The Canadian engineers who created CANDU didn't know exactly what the nuclear fuel market would look like decades into the future. By creating a reactor that could accept multiple fuel types, they built in adaptability that's still paying dividends today.

A Global Reactor

CANDU reactors operate on four continents. Within Canada, you'll find them in Ontario, Quebec, and New Brunswick. The CANDU 6 model—a standardized 600-megawatt design intended for export—was sold to Pakistan, Argentina, South Korea, Romania, and China. A variant went to India.

The Ontario installations grew larger over time as engineers became more confident in the design. The units at Darlington Nuclear Generating Station near Toronto produce nearly 900 megawatts each, making them among the most powerful CANDU reactors ever built.

For decades, Atomic Energy of Canada Limited (AECL) developed and marketed CANDU technology as a crown corporation, essentially a government-owned company. But by the 2000s, the market was shifting. Other reactor designs were competing effectively, and CANDU sales were slowing.

AECL tried to respond with new designs. The CANDU 9 was supposed to be an optimized version of the larger Ontario-style reactors. When that failed to attract buyers, they proposed the Advanced CANDU Reactor (ACR), which would have used slightly enriched uranium and light water as coolant while keeping heavy water as the moderator. The ACR was intended for an expansion at Darlington, but that project was cancelled in 2009.

In 2011, the Canadian government effectively privatized CANDU technology, licensing the design to a company called Candu Energy, a subsidiary of the engineering giant SNC-Lavalin (now part of AtkinsRéalis). The AECL team that had developed and marketed reactors for decades moved to the new company.

The Small Modular Future

The nuclear industry's latest obsession is small modular reactors, or SMRs—factory-built units small enough to truck to a site and install relatively quickly. The idea is that mass production of smaller units might bring costs down in ways that custom-built giant reactors cannot.

In 2017, the Canadian government established an "SMR Roadmap" to explore this technology. SNC-Lavalin responded by developing a 300-megawatt CANDU SMR, about half the size of the standard CANDU 6. Though this design wasn't selected for Canada's initial demonstration project in 2020, the company continues to market it internationally, particularly to countries looking to reduce carbon emissions without building massive traditional nuclear plants.

Whether CANDU technology has a future in this new landscape remains uncertain. The design's fundamental advantages—fuel flexibility and no need for enrichment—matter less now that enriched uranium is widely available on the global market. But the decades of operating experience and the proven reliability of the technology count for something.

What CANDU Tells Us About Engineering

The CANDU reactor is a monument to constrained creativity. Canadian engineers couldn't access enrichment technology, so they figured out how to make natural uranium work. They couldn't build giant pressure vessels, so they invented a modular tube system instead. Each limitation forced an innovation that turned out to have benefits beyond simply solving the immediate problem.

On-line refueling wasn't designed because someone thought it would be convenient to avoid shutdowns—it emerged naturally from the tube-based architecture that solved the pressure vessel problem. The ability to burn diverse fuel types wasn't a primary goal—it fell out of the physics of heavy water moderation. These are what engineers sometimes call emergent properties, benefits that arise from fundamental design choices rather than being explicitly designed in.

CANDU also shows how engineering decisions compound over decades. The choice of heavy water in the 1950s meant that every CANDU plant built since then needs a supply of heavy water, needs maintenance procedures designed around heavy water's properties, and needs operators trained in heavy water safety. Design decisions made seventy years ago continue to shape how people work today.

Whether you view this as technological lock-in or as the power of standardization depends on your perspective. But either way, CANDU stands as proof that there's usually more than one way to split an atom.