8.1.4 Non-proliferation

The International Fuel Cycle Evaluation (INFCE) study (1978-1980) summarized thorium fuel activities world wide and considered particular issues related to the technical barriers to proliferation. It was shown that the technical characteristics that would inhibit proliferation for thorium cycles with up to 20% of fissile material were similar to those of uranium-plutonium cycles.

Depending on the design, it takes between 5 and 15 kg of U-233 to make a nuclear weapon, which is not very different from plutonium. Thus, the U-233 bare sphere critical mass is 16 kg, compared to 10 kg for Pu-239 and 48 kg for U-235. Moreover, like U-235, a simple bomb made of U-233 is easier to fabricate than one made of plutonium because there are very few spontaneous neutrons emitted (only 1 neutron/sec/kg). It is therefore possible to design and fabricate a ‘gun — type’ weapon (in which the assembly comes together with the speed of a rifle bullet as opposed to an order-of-magnitude greater speed using high explosive). This is not possible with plutonium, because neutrons emitted by its even mass number isotopes (Pu-238, Pu-240 and Pu-242), always present at some quantity, require the manufacture of a more sophisticated implosion device. In this regard, it must be remembered that ‘civil’ plutonium contains a large proportion of these isotopes, making the manufacture of such a weapon very difficult. Another important feature of U-233 regarding proliferation is that it generates less heat than the even mass number isotopes of plutonium (but more than U-235). This property makes U-233 potentially less troublesome when fabricating a nuclear weapon. In fact, according to some experts and unclassified documents, the USA conducted a test of a U-233 bomb core in 1957 (the ‘Teapot test’) and has since conducted a number of other tests using this isotope.

Nevertheless, a specific technical hurdle does exist in the case of U-233. This is due to the small quantities of U-232 always mixed with U-233 and its associated strong gamma emitters (See section 8.1.2), which create a substantial difficulty in handling purified U-233 during weapon fabrication. In fact, after U-233 containing U-232 is processed, over a few years Th-228 ingrows to a nearly constant level, balanced by its own decay so that the gamma emissions increase and then stabilize. A 10 kg sphere of weapons grade U-233 (with U-232 as low as 5 ppm) could be expected to reach 0.11 mSv/hr at 1 metre after one month, 1.1 mSv/hr after one year, and 2 mSv/hr after two years. Because weapons are usually assembled and disassembled in unshielded glove-boxes, the build-up of the U-232 daughters would quickly create difficulties in complying with limits on the radiation exposure of workers. Terrorist groups, of course, may be less scrupulous about observing such limits.

To some extent, these radiation problems can be overcome by a ‘quick’ processing of U-233 after its separation and/or by the use of appropriate remote handling equipment. Alternatively, it is possible to reduce the concentration of U-232 by taking advantage of the fact that only very energetic neutrons (E > 6 MeV) can bring about the (n,2n) nuclear reaction responsible for its production. According to the French CEA calculations, a concentration of U-232 as low as 5 ppm can be reached if thorium is irradiated in the blanket of a fast reactor where the number of high energy neutrons is relatively low.

The presence of gamma emitters in a U-233 device would be useful to the extent that they provide a radioactive ‘tag’, which can help in the detection and prevention of covert diversion attempts. Once a U-233 weapon is fully assembled, of course, the various neutron-absorbing materials surrounding the fissile core such as neutron reflectors, would reduce the level of external radiation although enough would still penetrate to provide a distinctive signature that can be used to detect and track the weapons from a distance.

Another deterrent to the diversion of U-233 for weapons usage may be obtained by its dilution with U-238. This may be easily performed by mixing thorium with natural or depleted uranium in the fresh fuel (this is the so-called ‘denatured thorium cycle’, mentioned above). However, this option would lead to plutonium production (through U-238) and, therefore, would also raise proliferation concerns (because plutonium can be separated chemically). Another option would be isotopic dilution, mixing U-233 with uranium (natural or depleted) in the course of reprocessing thorium fuel. This option would be ineffective with plutonium because, unlike uranium, all of its isotopes have sufficiently small bare-sphere critical masses to potentially permit their use in nuclear explosives. The drawback with isotopic mixing of U-238 and U-233 is that recycling of the latter would be much less attractive.

Another potential difficulty in using U-233 to make a nuclear weapon results from the high alpha activity of U-232. Indeed because of (alpha, n) nuclear reactions on light element contaminants in the fissile material, neutron emissions would also occur. However, this process produces much fewer neutrons in uranium metal than spontaneous fission of Pu-240 contaminant in plutonium. Furthermore, a high degree of purification would allow the virtual elimination of this potentially disturbing neutron source.

To sum up, U-233 is clearly a material that can be used to make a nuclear weapon but several routes can be implemented to ‘denature’ this material easily enough. Thus, should a uranium-thorium cycle be developed, it would likely offer a degree of proliferation resistance equivalent to that of the LEU cycle, provided that uranium mixed with thorium is not used in conjunction with HEU (enrichment > 90%12).