How Thorium Reactors Would Work: A Technical Guide 2026

Thorium reactors would work by converting thorium-232, which cannot split on its own, into uranium-233, a fissile isotope that does. A neutron does the converting, the new uranium-233 fissions and releases heat, and some of those neutrons breed more fuel in the same pass.

The rest of this guide walks through that sequence step by step, then covers what actually happens inside a molten-salt core, what the fuel cycle involves, and why the interesting problems are cost, corrosion and licensing rather than physics. A thorium reactor is still a nuclear reactor. It uses the same fission reaction as a light-water plant; the fuel is just different.

Last updated October 2026. Figures are attributed to the IAEA, the IAEA-NEA and the World Nuclear Association where possible.

What Is a Thorium Reactor?

A thorium reactor is any fission reactor that uses thorium-232 as its fuel source. Because thorium-232 will not split when a neutron hits it, the reactor first manufactures a different fuel, uranium-233, and burns that instead.

The distinction that trips most readers up is between fertile and fissile. A fissile isotope splits when it absorbs a neutron. A fertile one cannot, but it can absorb a neutron and decay into something that can. Thorium-232 is fertile. Uranium-233, uranium-235 and plutonium-239 are fissile.

Not all thorium concepts are molten salt. The molten-salt designs get the attention, but thorium fuel has also been burned, or proposed, in heavy water reactors, high-temperature gas-cooled reactors, fast reactors and accelerator-driven systems. They differ mainly in how the fuel is carried and how the heat leaves the core.

How Thorium Reactors Would Work

Here is the whole mechanism in four steps. Once it is running, steps 1 and 4 happen continuously and at the same time.

Step	What happens	Isotope	Timescale
1. Neutron capture	A free neutron is absorbed by a thorium-232 nucleus, making it thorium-233	Th-232 to Th-233	Essentially instant
2. First beta decay	Thorium-233 sheds a beta particle and becomes protactinium-233	Th-233 to Pa-233	About 22 minutes
3. Second beta decay	Protactinium-233 sheds another beta particle and becomes uranium-233	Pa-233 to U-233	About 1.2 days
4. Fission	The uranium-233 absorbs a neutron and splits, releasing heat and roughly two more neutrons	U-233 fission	Continuous

Those extra neutrons are the whole point. Each fission gives back more than one neutron, and each spare neutron has two possible jobs: keep the chain reaction going, or get captured by more thorium-232 in a surrounding blanket to make more uranium-233. If the second job happens often enough, the reactor produces more fissile material than it consumes. That is what a breeder does, and a thorium reactor designed for it is called a fuel-cycle breeder.

The catch readers usually miss is start-up. A thorium core is not self-sustaining on day one, because there is no uranium-233 in it yet. A small amount of fissile material, typically enriched uranium or plutonium, is needed to reach criticality and get the reaction started. Once the core is running, the breeding blankets supply the new uranium-233 and the external seed is no longer required in any meaningful quantity.

This is the single most misunderstood point in the debate. Thorium reactors do not need uranium or plutonium forever. They need a starter, the way a fire needs a match, and after that the chain reaction feeds itself from thorium.

What Happens Inside a Molten-Salt Reactor?

In a molten-salt design, the fuel is not a solid pellet in a cladding. It is dissolved in a hot fluoride salt, which acts as both the fuel and the coolant, and the whole thing circulates by pump and heat exchanger.

Because the salt is liquid, the normal instruction to keep water away from the core is much less fragile. There is no water, so a loss of coolant pressure cannot produce the steam explosion or hydrogen risk that shapes light-water safety analysis. The IAEA notes the flip side plainly: the salts are hot and chemically corrosive, which turns the materials problem into a different, harder problem.

Salt chemistry also changes the physics. Fluoride salts carry only light elements like fluorine, lithium, beryllium or thorium, so there is very little neutron absorption in the coolant itself. That gives the design a more generous neutron budget, and the moderator can be added in a separate part of the circuit or omitted entirely, since a liquid-fueled fast reactor is not slowed down much by its own coolant.

Control and shutdown work by draining. A section of pipe called the freeze plug is kept just below the salt’s melting point. Insert control rods or let the temperature rise and the plug melts, and the hot salt runs by gravity into dump tanks underground. The core is then empty, and cooling is passive because the heat is leaving with the salt, not being generated.

Electricity is generated conventionally at the back end. A heat exchanger passes the hot salt to a secondary loop, that heats water into steam, and the steam spins a turbine. The interesting proposals sit upstream of that step: some designs use electrolysis to split water and make hydrogen, others drive desalination or chemical synthesis directly.

What Is the Thorium-to-Uranium-233 Fuel Cycle?

The fuel cycle is everything between taking material out of the ground and putting spent material back. For thorium it is the deciding part of the story, because almost every advantage and every hard problem lives here.

Where the uranium-233 comes from

In a solid-fuel design, thorium oxide is blended with uranium or plutonium oxide, pressed into pellets and loaded into the core. During operation, the pellets breed U-233 the same way any other thorium core does, and the material is later separated chemically into its components.

Why a molten-salt design removes a step

A liquid fuel can be cleaned while the reactor runs. Fission products accumulate in the salt as neutron poisons, so a thorium MSR design siphops off a small stream of salt, removes noble gases and fission products, and returns the cleaned portion. The breeding and cleanup happen in the same plant, at the same time.

Why that is also the hard part

On-line chemistry is the feature and the headache at the same time. You need fluorination equipment, materials that survive molten fluoride at high temperature, and a facility that regulators are willing to license as a chemical plant attached to a nuclear plant. Removing the need for a separate large reprocessing plant does not remove the need for chemistry; it relocates it.

Recycling and resource claims

Because uranium-233 is chemically identical to other uranium isotopes, standard uranium processing chemistry can separate it from fission products, and in principle from plutonium, if the safeguards are not tight. Recycle and you extract a large share of the energy still sitting in the material, which is where claims about stretching known uranium resources by an order of magnitude come from.

Thorium itself is three to four times more abundant in the Earth’s crust than uranium, and the IAEA-NEA estimated roughly 6.2 million tonnes of identified and estimated thorium worldwide in its 2014 Uranium assessment. Abundance is real, and it is also not the constraint people assume. No current reactor burns enough thorium to make the resource figure urgent.

Why Proponents Say Thorium Reactors Could Be Safer

The safety argument is about heat and pressure, not radiation. Remove the water and much of the accident sequence a light-water plant is designed around goes with it.

• No high-pressure primary system. A molten-salt core operates near atmospheric pressure, so the pressure-driven failure modes that dominate light-water accident analysis are largely absent.
• Drain and freeze. Loss of power, loss of coolant or a temperature rise can all end the reaction passively by melting the freeze plug and emptying the core into tanks.
• Negative temperature feedback. In a thermal-spectrum molten-salt design, the salt expands and slows neutrons down more as it heats, which pulls the reaction back on its own.
• Low pressure in the decay heat path. After shutdown, decay heat still has to go somewhere, and circulating salt carries it without a pressurised boundary.

None of these make a thorium reactor risk-free. A drained core still has decay heat, still has a fuel salt that is chemically hot, and still has fission products that need containment. Passive mechanisms are passive only if the geometry, the materials and the decay heat load work out, and each of those has to be demonstrated on a real machine rather than argued on paper.

It is fair to say that a well-designed molten-salt thorium reactor would remove whole categories of accident that a light-water reactor cannot remove without abandoning its water coolant. It is not fair to say the design is incapable of an accident.

What About Radioactive Waste and Byproducts?

Thorium reactors produce fission products, exactly like other reactors. The claim worth examining is about which types, how much, and how long they last.

The argument runs like this: burn uranium-233 instead of plutonium or uranium-235 and you produce proportionally less transuranic waste, the heavy elements above uranium that are the hardest and longest-lived part of spent fuel. A thorium cycle is not free of transuranics, because reactor physics unavoidably makes some plutonium, but the amount and the heat load per unit of waste are lower. The transuranics that do form also tend to be shorter-lived than the plutonium in a conventional spent fuel assembly.

Fission products do not go away. Strontium-90, cesium-137, iodine-131, technetium-99 and the rest appear in thorium spent fuel, salt cleanup streams and waste volumes that a light-water plant of the same output would also produce.

So the honest summary is: smaller quantities of one category, similar quantities of another, different chemistry to manage. That is a real difference and it is a research question. It is not the elimination of long-lived waste, and any article claiming otherwise is selling something.

Where Would Thorium Reactors Fit in the Electricity Grid?

Different thorium concepts suit different grid jobs, and the ones that fit best are not always the ones with the best physics.

• Baseload electricity. A 500 MWe unit would behave like any other firm generator, sitting at high output around the clock. Designs optimised for electricity are the easiest to model against existing nuclear fleets.
• Load following. Molten-salt designs can run at partial power without the control rod and xenon problems that constrain a light-water reactor, which matters as grids add variable wind and solar.
• Industrial process heat. High-temperature versions could supply process steam, hydrogen electrolysis, or ammonia and synthetic fuel production. This is the fastest-growing commercial interest among private developers.
• Remote and distributed power. Small units that need no on-site reprocessing and move less fissile material would suit mines, ice routes and military installations.
• Waste transmutation. Accelerator-driven thorium systems are being studied for burning long-lived transuranics from existing spent fuel, which is a different goal from cheap electricity.
• Retiring coal sites. Several developers pitch thorium designs as drop-in capacity on coal plant sites, where transmission, water and permits already exist.

Each use case carries its own infrastructure requirement. Process heat needs hydrogen customers. Load following needs a grid willing to pay for flexibility. Retiring coal sites needs a licensing pathway that treats it as a repowering rather than a new nuclear plant.

What Are the Main Technical and Commercial Barriers?

This is where the honest thorium advocate and the honest sceptic mostly agree on the diagnosis, though not on the prognosis.

Materials and corrosion

Hot fluoride salt eats through steel. The IAEA lists corrosion as a core challenge, and the engineering response has been exotic alloys, graphite liners, or separating the fuel from structural metal entirely. Corrosion is the most-cited objection from physics-literate critics, and it is not a solved problem at commercial scale.

Fuel fabrication

Thorium dioxide melts at roughly 3300 degrees Celsius, far above the sintering temperatures conventional pellet fabrication relies on, so it cannot be pressed the same way. There is no thorium fuel supply chain, no enrichment services and no established fabrication capacity anywhere in the world.

Licensing and regulation

Industry insiders repeatedly point at regulation as the binding constraint rather than physics. Regulators such as the US NRC have historically required substantially complete designs before opening a review, which forces a developer to absorb years of engineering and capital cost before there is a decision to appeal to. A molten-salt plant with on-site chemistry compounds this, because it must be licensed as a chemical facility as well.

Tritium, blanket removal and maintenance

Reactor and blanket structures in a breeding core are heavily irradiated and hard to replace. Tritium management from the lithium-bearing salts is an ongoing research area. Neither has a long commercial operating record.

Money and schedule

Conventional nuclear is already expensive and slow, and thorium adds a fuel cycle with no precedent at scale. Sceptics on nuclear-focused forums tend to land on economics rather than physics, and the cost scepticism is not easy to dismiss. There is no thorium plant anywhere that has produced electricity at commercial scale, so cost-of-electricity comparisons are model outputs rather than measured results.

What has actually been built is worth knowing. The Molten-Salt Reactor Experiment at Oak Ridge reached criticality in about five years and then ran for roughly five more, on a budget that looks small only against later nuclear projects. Fort St. Vrain in the United States ran a helium-cooled graphite reactor on thorium in the 1970s. India’s 500 MWe PFBR at Kalpakkam and its AHWR series at Kakrapara are real programmes built around thorium and U-233. China has operated the TMSR-LF1 molten-salt test reactor since 2021 and has claimed thorium breeding at burnup levels above what earlier designs achieved, a claim that has not been independently published in full.

So the fair summary is demonstrated in principle, partially demonstrated in practice, not yet commercial.

Thorium Reactors vs. Conventional Nuclear Reactors

Here is how the two fuel cycles compare on the points that actually drive cost, safety and waste decisions.

Point of comparison	Thorium cycle	Conventional light-water cycle
Primary fuel behaviour	Thorium-232 is fertile and must be converted to U-233	Uranium-235 is fissile and splits directly
Can it sustain a chain reaction alone	No, a small fissile seed is required for start-up	Yes, once enriched fuel is loaded
Enrichment	Far less or none, since the fuel is not gaseous and is not enriched for U-235	Enrichment to 3 to 5 percent U-235 is standard
Chain reaction and moderation	Thermal or fast spectrum; molten-salt designs carry little moderator in the salt	Thermal spectrum with water as coolant and moderator
Operating pressure	Near atmospheric in molten-salt designs	High, roughly 155 bar in a pressurized water reactor
Waste profile	Lower quantity of long-lived transuranics; similar fission products	Larger transuranic inventory with longer-lived plutonium
Breeding potential	Designed to breed U-233 from a thorium blanket	Not a breeder; a fuel and plutonium producer
Proliferation sensitivity	U-232 co-production gives a strong gamma field that makes material harder to handle and easier to trace	Enriched uranium and separated plutonium are the sensitive materials
Chemical processing	Fluoride chemistry, often on-line in molten-salt designs	Established PUREX reprocessing, mostly off-site
Deployment status	No commercial plant; test and prototype reactors only	About 440 reactors operating worldwide, mostly in the same handful of designs

Read the table honestly and the pattern is consistent. Thorium’s advantages are real physics and chemistry benefits that mostly require new equipment and new supply chains to deliver. The light-water fleet’s advantage is the opposite: five decades of operating experience, a mature fuel and licensing base, and dozens of identical designs.

That is why thorium has never displaced light water. It is not that thorium is unworkable. It is that a better fuel is worth very little without a factory, a regulator and a balance sheet that can build the rest.

Frequently Asked Questions

Conclusion

How thorium reactors would work comes down to one line. Thorium-232 is not fuel; it is feedstock. Catch a neutron, wait out two beta decays, and you have uranium-233, which behaves as a perfectly ordinary fissile fuel and hands back enough extra neutrons to keep breeding more of it.

The physics is settled and has been since the 1940s. Everything still open sits downstream of it, and anyone evaluating a thorium proposal should ask five questions in order. What is the specific reactor design, not the fuel in general? What evidence exists for the fuel cycle, especially the processing step? What does the safety analysis assume, and has any of it been demonstrated on hardware? Where is the project in licensing? And what does a real cost and construction schedule look like rather than a modelled one?

Thorium is a real option that no one has yet built at commercial scale. That is a more interesting sentence than either the booster club version or the dismissal version.

Source: https://www.pgm-blog.com/how-thorium-reactors-would-work/