nuclear.academy
Module 2

The Fuel Cycle, and Where Security Lives in It

Front end, reactor, back end: what each stage of the nuclear fuel cycle does, which stages handle material that could be misused, why safeguarding a bulk-handling plant is a statistics problem, and what the countries that reprocess have learned.

Learning objectives

This module is built from the author’s preparation for a public panel on the fuel cycle held at the University of Tennessee in August 2026. Everything that changes — inventories, prices, licensing dockets, the status of particular projects — lives on the situation page for the U.S. back end. This module keeps what does not change: what the stages are, where the material of concern is, and what the last fifty years have taught. The site takes no position on whether any country should recycle its fuel; it does try to state accurately what recycling does and does not do.

1. The cycle in one paragraph

Uranium is dug out of the ground, chemically cleaned up, converted to a gas, enriched so that a larger share of it is the isotope that fissions, converted back to a solid, and made into fuel. The fuel runs in a reactor for a few years, during which some of the uranium fissions, some becomes plutonium, and the fuel fills up with intensely radioactive fission products. Then it comes out. From there it goes one of two ways: it is stored and eventually buried as it is (the once-through or open cycle), or it is chemically taken apart so that the unused uranium and the plutonium can be made into new fuel, with the fission products separated out for burial (the closed cycle, or recycling). Either way, something ends up in a deep geological repository.

Everything that follows is that paragraph in detail, with attention to one question: at which stages is there material that could be turned into a weapon, and what does that imply for how the stage is guarded and measured?

2. The front end

Mining and milling. Uranium ore is mined, or leached from the ground in place, and processed into a concentrate called yellowcake (mostly U₃O₈). Natural uranium is 0.7 percent uranium-235; the rest is uranium-238, which does not sustain a chain reaction with slow neutrons (Foundations 4). Yellowcake is only mildly radioactive and is no use for a weapon without the steps that follow. Security here is about tracking bulk quantities and knowing where they go. The front end is global: a handful of countries mine most of the world’s uranium, and the largest of them, Kazakhstan, is not the one whose reactors consume it.

Conversion. Yellowcake is turned into uranium hexafluoride (UF₆), a compound that is a solid at room temperature and a gas when gently heated. It is the only form in which uranium can be enriched by the methods in use.

Enrichment. The share of uranium-235 is raised — to 3 to 5 percent for the reactors in service today, to just under 20 percent (HALEU, high-assay low-enriched uranium) for many of the new designs, and to 90 percent or more for a weapon. The commercial method is the gas centrifuge: thousands of spinning tubes in series, each separating the slightly lighter UF₆ molecules containing U-235 from the slightly heavier ones containing U-238. The work is measured in separative work units (SWU), and the important physical fact is this: most of the work of going from natural uranium to weapons-grade is done by the time you reach reactor grade. Going from 0.7 to 4 percent takes about two-thirds of the SWU that going all the way to 90 percent does. That is why an enrichment plant is inherently a dual-use facility, why the same centrifuge cascade can in principle be reconfigured, and why enrichment is one of the two stages the safeguards system watches most closely.

Fuel fabrication. Enriched UF₆ is converted to uranium dioxide powder, pressed into ceramic pellets, sealed into metal tubes, and bundled into assemblies. The product is countable: an assembly has a serial number, and a reactor operator knows exactly how many it has. Countability matters for what comes next.

3. The reactor

Inside the reactor, three things happen to the fuel at once. Uranium-235 fissions, releasing the energy the plant sells. Uranium-238 captures neutrons and becomes plutonium-239, some of which fissions in turn — by the end of a fuel’s life a third or more of the energy is coming from plutonium the reactor made itself. And fission products pile up: strontium-90, cesium-137, iodine-131, and hundreds of others, most of them intensely radioactive.

The result is spent fuel: still about 95 percent uranium by mass, about 1 percent plutonium, and the rest fission products and heavier elements. Freshly discharged, an assembly is dangerously radioactive — lethal within minutes at close range without shielding — and it stays hot, thermally and radiologically, for decades. That radioactivity is a hazard, but it is also, from a security point of view, a barrier: nobody steals a fresh spent-fuel assembly. The plutonium in it is unreachable without a heavily shielded chemical plant. This is what the field calls the spent-fuel standard, and it is the reason plutonium in spent fuel is treated so differently from plutonium separated from it.

Safeguards at a reactor are comparatively simple, because the material is in countable items. Inspectors count assemblies, verify serial numbers, seal the pool or cask, and watch the loading and unloading with cameras. If an item is missing, it is missing. Reactor designs that handle fuel as a liquid or as hundreds of thousands of pebbles lose that countability, which is why they need new methods (Module 1, section 7).

4. The back end, once-through

In the once-through cycle, spent fuel is cooled in a water pool for years, then moved into sealed steel-and-concrete dry casks on a pad, and eventually — in principle — moved to a deep geological repository, several hundred meters underground in stable rock, and left there.

The word “principle” is doing work. As of 2026 no country has yet placed commercial spent fuel in an operating deep repository, though Finland is at the threshold and Sweden and France have sites and designs. In the United States the repository program has been stalled since 2010, for reasons that were legal and political rather than technical, and the fuel sits where it was made: on the order of 95,000 metric tons at roughly 75 sites, growing by about 2,000 tons a year, now mostly in dry casks. The figures, the money, and the litigation are on the situation page.

From a security standpoint the once-through cycle has one large virtue: the plutonium never leaves the spent fuel. Dry casks are a physical-protection problem — a large, heavy, well-marked target that cannot be moved quickly — but they are not a diversion problem in the safeguards sense. The material of concern stays behind its own radiation barrier for as long as the fuel remains intact, which is centuries.

5. The back end, closed: reprocessing

In the closed cycle, spent fuel is dissolved in acid and chemically separated into three streams: uranium (to be re-enriched or stored), plutonium (to be made into mixed-oxide fuel, MOX, or stored), and everything else (to be immobilized in glass and buried). The process in use at industrial scale everywhere it is done is PUREX, developed in the United States in the 1950s for weapons production and adapted for civil fuel. Variants that keep the plutonium mixed with uranium or other elements have been demonstrated at laboratory and pilot scale; none has run at commercial scale.

Three facts about reprocessing are durable enough to belong here rather than on the dated page.

It does not eliminate the repository. Every state that reprocesses is also building or siting a deep repository. The fission products still have to go somewhere; losses and residues from the process itself are long-lived waste; and the recycled fuel, when it is spent in turn, is generally not reprocessed again. The U.S. National Academies put it in one sentence in 2023: “Regardless of the fuel cycle, a geologic repository will be required.” Recycling changes the form and the volume of the waste. Whether that is worth doing is the question countries answer differently; that they still need a repository is not in dispute, and the leading advocates of recycling say so themselves.

Volume is the wrong number to argue about. Reprocessing shrinks the volume of high-level waste a great deal. But the size of a repository is set by heat, not volume: the waste has to be spaced so the rock does not overheat, and most of the heat for the first century comes from cesium-137 and strontium-90, which reprocessing does not remove. Separating out the plutonium and uranium reduces the heat load only slightly. So “recycling cuts the waste by ninety percent” and “recycling barely changes the repository” can both be true, because they are about different things.

It produces separated plutonium. This is the security fact. Plutonium in spent fuel is behind the radiation barrier; plutonium coming out of a reprocessing plant is a powder or a solution that can be handled by a person in a glove box. Eight kilograms of it is what the IAEA defines as a significant quantity — the amount from which a first weapon could be made. A commercial reprocessing plant separates several tons of plutonium a year. The world’s stock of separated civil plutonium is now measured in the hundreds of tons and has grown for decades, because separation has run ahead of the fabrication of MOX fuel to use it. The United Kingdom, which holds the largest civil stockpile, decided in 2025 to immobilize most of it rather than recycle it at all.

6. Why a bulk plant cannot be exactly accounted for

Here is the point on which everything else in this module turns, and it is a matter of arithmetic rather than policy.

At a reactor, material is in items, and you count them. In an enrichment or reprocessing plant, material is in bulk — gas in pipes, solution in tanks, powder in hoppers — and you cannot count it. You measure it: weigh the inputs, sample and analyze the outputs, and compare. The comparison is called the material balance, and the number that comes out of it is the material unaccounted for, or MUF: inputs minus outputs minus what is still inside. In a perfect world MUF is zero. In the real world every measurement has an uncertainty, and the uncertainties add up.

Do the sum for a plant that separates 8,000 kilograms of plutonium a year. Suppose every measurement is good to 1 percent — which is very good for chemical analysis of a hot, dirty solution. The uncertainty on the year’s balance is then on the order of 80 kilograms. That is ten significant quantities. The plant could be missing ten weapons’ worth of plutonium and the material balance could not tell you. This is not a criticism of any particular plant or instrument. It is a property of measuring large flows: the uncertainty scales with the throughput, and the significant quantity does not.

Safeguards at bulk plants therefore do not rely on the material balance alone. They add containment and surveillance — seals, cameras, and process monitors that watch for material leaving by any path other than the measured ones — and near-real-time accountancy, which closes the balance every few days instead of once a year so that a discrepancy shows up before it can grow. Those measures are real and they work. What has never been solved, in the thirty-odd years since the question was first posed, is how to combine them with the material balance into a single number that says how confident you are. The honest statement is that a large reprocessing plant is safeguarded by a layered system whose overall effectiveness is judged, not computed.

Two consequences follow. First, safeguards have to be designed into a bulk plant from the flowsheet up, not added at licensing: where the measurement points are, how the tanks are instrumented, where the cameras can see. The field’s term is safeguards by design. Second, the question “who is the accountancy regulator for this plant?” has to have an answer before the plant is built. In the United States, as of 2026, for a commercial reprocessing plant it did not — see the situation page.

7. What other countries have done

The durable lessons, country by country; the numbers are on the situation page.

The pattern: reprocessing is a choice some countries have made for reasons of energy security, industrial policy, or waste form, at a cost that has consistently exceeded the once-through alternative at prevailing uranium prices; it has not, anywhere, removed the need for a repository; and it has, everywhere, created stocks of separated plutonium that have to be guarded, accounted for, and eventually disposed of.

8. Four numbers to ask for

When you read a claim about the fuel cycle, these are the questions that separate the durable from the promotional.

  1. “Reduces waste by X percent” — volume, or heat? If volume, ask what happens to the cesium and strontium. If heat, ask over what period.
  2. “Recycling supplies fuel for advanced reactors” — at what enrichment? Uranium recovered from ordinary spent fuel is barely above natural, about 0.8 to 0.9 percent U-235, and carries contaminants that make it harder to enrich. It is not HALEU and does not become HALEU without going back through an enrichment plant.
  3. “Demonstrated technology” — at what scale, on what fuel? Kilograms in a hot cell and thousands of tons a year in a plant are different engineering problems. Ask for the largest quantity of real commercial spent fuel the process has ever handled.
  4. “Proliferation-resistant” — compared with what, against whom? Processes that keep plutonium mixed with uranium raise the bar for a thief; they do little against a state that owns the plant. Six U.S. national laboratories said so in 2009, and nothing since has changed the conclusion.

Explore. (a) Trace one kilogram of natural uranium through the once-through cycle and say at which stage it is countable and at which it is bulk. (b) Why does the spent-fuel standard make dry-cask storage a physical-protection problem rather than a safeguards problem? (c) Redo the material-balance arithmetic in section 6 for a plant handling 20,000 kilograms of plutonium a year at 0.5 percent uncertainty. How many significant quantities is the uncertainty? (d) Two countries both say they have “solved” their waste problem, one by recycling and one by opening a repository. What would you check in each case?


Sources

  1. International Atomic Energy Agency, IAEA Safeguards Glossary, 2022 edition — significant quantity (8 kg plutonium; 25 kg U-235 in HEU), material balance, MUF, containment and surveillance.
  2. National Academies of Sciences, Engineering, and Medicine, Merits and Viability of Different Nuclear Fuel Cycles and Technology Options and the Waste Aspects of Advanced Nuclear Reactors (2023): “Regardless of the fuel cycle, a geologic repository will be required to dispose of spent nuclear fuel, high-level wastes, and other wastes that contain long-lived radionuclides.”
  3. OECD Nuclear Energy Agency, Strategies and Considerations for the Back End of the Fuel Cycle, NEA No. 7469 (2021) — high-level waste volume and decay-heat comparison for mono-recycling.
  4. OECD-NEA Expert Group on Fuel Recycling Chemistry, technology-readiness assessment, reported in Progress in Nuclear Energy (November 2019; INL/JOU-18-44641) — PUREX at TRL 9; alternatives at TRL 6 and below.
  5. Idaho National Laboratory, Pyroprocessing status report to the NRC, INL/RPT-23-03365 (TLR-RES/DE/REB-2023-10, December 2023): “Pyroprocessing has not been developed past the engineering scale.”
  6. Oak Ridge National Laboratory, ORNL/TM-2020/1751 — coupled end-to-end campaigns on Dresden-1 and North Anna spent fuel, about 4.5 kg total.
  7. M. Bunn, S. Fetter, J. P. Holdren and B. van der Zwaan, The Economics of Reprocessing versus Direct Disposal of Spent Nuclear Fuel (Harvard, 2003); Congressional Research Service, R48364 (January 2025); Congressional Budget Office (2007); GAO-08-483.
  8. Nuclear Decommissioning Authority (UK), decision of 24 January 2025 to immobilize the UK-owned plutonium inventory.
  9. Kim et al., Values of Recovered Uranium from HALEU Used Nuclear Fuels (Argonne, Brookhaven, Oak Ridge, March 2024) — recovered uranium preferable only above about 7 percent residual U-235.
  10. U.S. Department of Energy, Report on Proliferation Resistance of Advanced Fuel Cycles (six national laboratories, 2009) — “only a modest improvement in reducing proliferation risk over existing PUREX technologies.”
  11. Andra, Cigéo key figures (cigeo.gouv.fr) — vitrified HLW about 10,000 of 83,000 m³ of deep disposal volume.
  12. H. L. Hall, Fuel Cycle Panel — Speaking Notes and Data Annex, Baker School Policy Forum, University of Tennessee, 12 August 2026 (the author’s own compilation; every figure above traces to a primary source listed there).

This material is based in part upon work supported by the Department of Energy National Nuclear Security Administration through Defense Nuclear Nonproliferation’s Enabling Capabilities in Technology Consortium under Award Number DE-NA0004197. The views and opinions expressed are the author’s own and do not necessarily state or reflect those of the United States Government or any agency thereof.

Last reviewed 2026-09-22.