"The waste that is generated cannot be used for weapons, but does contain elements that are valuable for use as medical isotopes."

While it is very difficult to use nuclear materials from any operating commercial reactor to make weapons, as Thorium advocates, we sometime oversell the idea that U-233, the fissile component of the Thorium fuel cycle, cannot be used for building a nuclear explosive device.

The limit of enrichment for U-233 to be considered LEU is only 12%, while the limit of enrichment for U-235 is 20%. U-233 is in fact in many respects a better material to begin with when starting to design a nuclear explosive than U-235. Reactor grade U-233 taken from an epi-thermal neutron spectrum LFTR would probably be about 60% enrichment U-233 (reactor grade U-233 would be mixed with other Uranium isotopes like U-232, U-234, and U-236). 60% U-233 would be considered "weapons usable" material which, while not ideal for making weapons, is possible to use to make weapons after addressing some design difficulties.

U-232 contamination has frequently been pointed to as a reason why U-233 weaponization cannot be accomplished and is impractical. While I agree that high levels of U-232 makes weaponization of U-233 difficult, it unfortunately does not make weapons production of highly U-232 contaminated U-233 impossible. ORNL worked out a scheme where the U-232 decay daughters responsible for most of the activity experienced from U-232 contamination can be temporarily be removed for about ~3 weeks. Three weeks is enough time to machine U-232 contaminated U-233 safely and produce the basic components for a device. After three weeks, treated U-232 contaminated U-233 would gradually build up in activity, but as long as shielding is provided (11 inch thickness of lead to knock down the 2.6 MeV hard gamma); such devices could still be delivered to a target.

ORNL-4731 Removing U-232 decay daughters from U-233 - http://www.ornl.gov/info/reports/1972/3445605995418.pdf

It is better to back off from the claim that U-233 (even reactor grade) could not be used to make a weapon.

Note: The Indian government fielded a small successful pure U-233 device in the Pohkran II Shakti-5 nuclear test of 1998.

Lawrence Livermore National Lab Senior Weapons designer's opinion on U-233 for weapons -

W.K. Woods "LLNL interest in U-233"

http://1.usa.gov/g5XVno

Robert, I think all will agree that U233 is weaponizable.

However, the reality of any MSR/LFTR is that even a 1GWe machine will process about 1/2kg of U233 per hour (whether breeding or fissioning), and that it's dilutetd about 30:1 in the reactor salt mass.

This means that to steal U233 from even the largest reactor would mean tapping the flow of UF4 to the core at a rate that's tiny, compared to the desired bomb-making amount. The pipe for the breeding system is very small. Many hours of subtle diversion, undetected by power reduction of the plant, would be necessary.

It seems there are many more alternatives, with less risk, to those demanding several kg of fissiles to make a bomb.

Smaller reactors would raise even larger barriers to thieves.

AlexC - I agree.