nuclear.academy
Module 8

Reprocessing Options: What the Choices Are and What Each One Costs in Plutonium

The ways spent fuel can be taken apart — PUREX and its variants, pyroprocessing, and the processes that keep plutonium mixed with something else — what each has actually been demonstrated to do, the history of the U.S. decision not to reprocess, and what each option means for safeguards.

Learning objectives

This module follows Module 2, which explains why reprocessing is a security question, and takes no position on whether any country should do it. It describes processes at the level of the open technical literature and the regulators’ documents: what each does, not how to run it.

1. What reprocessing is for

Spent fuel comes out of a reactor as a mixture: about 95 percent uranium, about 1 percent plutonium, and a few percent of fission products and minor actinides, all locked in a ceramic inside metal cladding and intensely radioactive. Reprocessing takes that mixture apart chemically so that the uranium and plutonium can be reused and the rest can be packaged for disposal. Every option below is a different way of doing the taking-apart, and the options differ mainly in what comes out in the plutonium stream and how clean it is. That is why a security site cares which one a country chooses.

2. PUREX: the process that exists

PUREX (plutonium–uranium reduction extraction) was developed in the United States in the early 1950s for weapons plutonium production and has been the basis of every commercial reprocessing plant since. In outline:

  1. Chop and dissolve. Fuel assemblies are sheared into pieces and the ceramic fuel is dissolved in hot nitric acid, leaving the cladding hulls as waste.
  2. Extract. The acid solution is contacted with an organic solvent (tributyl phosphate in a kerosene-like diluent) that takes up uranium and plutonium and leaves the fission products behind. Repeated contacting in columns or mixer-settlers gives high purity.
  3. Partition. Plutonium is chemically shifted to a form the solvent will not hold, and stripped out separately from the uranium. Each is purified further and converted to an oxide.

The fission-product raffinate is concentrated and, in modern plants, vitrified — melted into borosilicate glass in steel canisters — for eventual deep disposal. Cladding hulls, solvents and process wastes are intermediate-level waste, and in volume they are most of what a reprocessing plant sends to a repository.

PUREX works: it is reliable, its chemistry is understood in detail, and it has been run at scales up to more than a thousand tons of fuel a year at La Hague in France. Its defining feature for this site is that step 3 produces pure separated plutonium — a powder that can be handled in a glove box, from which a significant quantity is eight kilograms (Module 3). Every other process on this page was invented because of that feature.

3. The alternatives, and how far each has got

The OECD Nuclear Energy Agency’s expert group rated the readiness of the options on the standard nine-level scale, and the ratings are the most useful single fact about them.

Modified aqueous processes — COEX, UREX and relatives (TRL about 6). Variants of PUREX in which the plutonium is never fully separated: it is kept with uranium (COEX, the process the French developed for their MOX line), or with uranium and neptunium, or with other actinides (the UREX family developed in the U.S. national laboratories). The chemistry is close to PUREX and the equipment is similar. What has been demonstrated is pilot-scale operation on real spent fuel; in the United States, the largest integrated campaign on actual commercial fuel totals about 4.5 kilograms, at Oak Ridge. No plant using any of them has been built.

Pyroprocessing (TRL below 6). Not aqueous at all: the fuel is converted to metal and dissolved in a molten salt at several hundred degrees, and uranium and a mixture of plutonium with other actinides are recovered on electrodes. Developed for the metallic fuel of fast reactors, it has one operating example: the treatment of the fast-reactor fuel from the EBR-II program at Idaho, which has processed about 5.5 of some 25 tons over 27 years. The national laboratory’s own 2023 report to the NRC says the process “has not been developed past the engineering scale.” South Korea has a long-running research program.

Advanced partitioning. Processes to pull minor actinides (americium, curium) out of the waste for burning in fast reactors, reducing the long-lived heat and radiotoxicity of what is buried. Laboratory to pilot scale, in France, Japan and elsewhere; a research field rather than an option a utility can buy.

The pattern: the processes that avoid separated plutonium are precisely the ones that have not been run at scale, and the one that has been run at scale is the one that produces it.

4. “Proliferation-resistant” reprocessing

The modified processes are described as proliferation-resistant because the plutonium never appears alone: a thief would get a uranium–plutonium mixture, or a mixture with highly radioactive actinides, and would need further chemistry to make it usable. That is a real benefit against theft.

It is a much smaller benefit against the state that owns the plant. In 2009 six U.S. national laboratories assessed the question and concluded that the alternatives offer “only a modest improvement in reducing proliferation risk over existing PUREX technologies, and these modest improvements apply primarily for non-state actors.” A state can add the extra chemistry; the plant’s owner already has the expertise, the facility and the material. The claim to keep in view, then, is that process choice can raise the bar for a thief, and that against a determined state the bar is set by safeguards, not by chemistry.

5. The safeguards problem is the same for all of them

Whatever the process, a reprocessing plant is a bulk-handling facility: tons of plutonium a year in solution or salt, measured rather than counted. The material-balance uncertainty scales with throughput and comes to many significant quantities a year at commercial scale (Module 2, section 6). The mitigations — solution monitoring on every tank, near-real-time accountancy that closes the books every few days, containment and surveillance on every path — are the same for PUREX and for its variants; the pyroprocess is harder, because molten salt is not easily sampled and its composition is not easily measured.

Two implications hold for every option. Safeguards by design: the measurement points, tank instrumentation and camera lines of sight have to be in the flowsheet, not added later. And a defined regulator: somebody has to own the material accountancy rules for the plant before it is built. In the United States in 2026, for a commercial plant, that was not settled — the NRC’s material-accounting rule for Category I material excludes reprocessing plants by name, and the 2026 proposed rules on reprocessing licensing and physical protection did not change that (situation page).

6. The U.S. history in six dates

7. Elsewhere

France runs the world’s largest civil reprocessing operation at La Hague and recycles the plutonium as MOX in its own reactors, once. The United Kingdom reprocessed at Sellafield for fifty years, closed the Magnox plant in 2022 and THORP in 2018, and now holds the largest civil plutonium stockpile, which it has decided to immobilize. Japan’s Rokkasho plant, begun in the 1990s, has been postponed repeatedly and is not yet in full operation. Russia reprocesses at Mayak, a weapons-program site. China has a pilot plant and is building larger ones. India reprocesses for its fast-reactor program. Every one of those states is also siting or building a deep repository, because reprocessing does not remove the need for one.

Explore. (a) Write PUREX’s three steps in your own words and mark the step at which the security concern arises. (b) A country announces it will build a “proliferation-resistant” recycling plant. What three questions from this module would you ask first? (c) Why is a pyroprocessing plant harder to safeguard than an aqueous one? (d) Using the six dates, explain why the United States has spent fifty years neither reprocessing nor deciding not to.


Sources

  1. OECD Nuclear Energy Agency, Expert Group on Fuel Recycling Chemistry, technology-readiness assessment (reported in Progress in Nuclear Energy, 2019, INL/JOU-18-44641).
  2. Idaho National Laboratory, INL/RPT-23-03365 (NRC TLR-RES/DE/REB-2023-10, December 2023) — status of pyroprocessing.
  3. Oak Ridge National Laboratory, ORNL/TM-2020/1751 — coupled end-to-end demonstration on commercial spent fuel.
  4. U.S. Department of Energy, Report on Proliferation Resistance of Advanced Fuel Cycles (six national laboratories, 2009).
  5. U.S. Nuclear Regulatory Commission, 91 FR 38124 (24 June 2026) and 91 FR 38928 (26 June 2026); 10 CFR 74.51(a); history of the reprocessing rulemaking (2013, 2016, 2021 Commission papers).
  6. Presidential statement on nuclear power policy, 7 April 1977; statement of 8 October 1981; Executive Order 14299 (2025).
  7. U.S. Department of Energy, West Valley Demonstration Project history; NRC Barnwell docket history.
  8. International Atomic Energy Agency, Safeguards Techniques and Equipment (International Nuclear Verification Series No. 1, latest revision) — solution monitoring and near-real-time accountancy at reprocessing plants.
  9. National Academies (2023), OECD-NEA No. 7469 (2021), and the country sources listed in the U.S. back-end situation page.

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.