The decay path of thorium is well understood. If bombarded with neutrons, isotopically pure 232Th forms 233Th, which has a half-life of 22 minutes and β-decays to 233Pa. That isotope has a half-life of 27 days and β-decays to 233U, which can undergo fission. The International Atomic Energy Agency (IAEA) considers 8 kilograms of 233U to be enough to construct a nuclear weapon1. Thus, 233U poses proliferation risks.
Although 233U is not used today in commercial reactors, the United States accumulated two tonnes of it during the cold war. Plans to dispose of much of it by burial are controversial and pose security and safety risks, according to a 2012 report2.
The chemical reprocessing needed to separate 233U from spent nuclear fuel requires major infrastructure, such as large reprocessing plants, which are difficult to hide. With thorium fuel, the presence of highly radiotoxic 232U means that the spent fuel must be handled using remote techniques in heavily-shielded containment chambers.
After irradiating thorium with neutrons for around one month, chemical separation of 233Pa could yield minimal 232U contamination, making the 233U-rich product easier to handle. If pure 233Pa can be extracted, then it merely needs to be left to decay to produce pure 233U. The problem is that neutron irradiation of 232Th could take place in a small facility, such as a research reactor, of which around 500 exist worldwide. The 232Th need not be part of a nuclear-fuel assembly nor be involved in energy generation.
It has been demonstrated that around 200 g of thorium metal could produce 1 g of 233Pa — and hence 1 g of 233U — if exposed to neutrons at levels typically found in power reactors and some research reactors for a month, followed by protactinium separation3. Thus, only 1.6 tonnes of thorium metal would be required to produce the 8 kg of 233U required for a weapon. This amount of 233U could feasibly be obtained by this process in less than a year.