Good post on reddit from a nuclear engineer -
There are a lot of misconceptions flying around the internet about Thorium-fueled reactors. They are not some magical energy cure-all that will make all our problems disappear. When you add it up, do all the math and compare all the pros and cons, they wind up being comparable to conventional breeder reactors.
Yes, Thorium is far more abundant that Uranium and thus cheaper to mine. Because it is non-fissile, much like U-238 is in a conventional reactor, it requires an initial seed reaction from U-235 or Pu-239 to create neutrons. These are then absorbed by a Thorium (Th-232) blanket to eventually make U-233. Just like in a regular breeder reactor, you have to reprocess the irradiated Thorium to extract the U-233 to make into fresh fuel to continue the cycle. In a Molten Fluoride reactor, this can technically be done on-line by diverting the flowing fuel. While it's easy to draw this on-line refueling scheme in a block diagram on paper, a prototype of such a system has yet to be developed.
The catch in all this is that U-232 is also produced during the U-233 breeding process as a byproduct. This is nasty shit whose daughter products emit a highly penetrating 2.6 MeV gamma ray during its decay, and it has a relatively long half life of 60 years or so. This intense gamma radiation is what makes it proliferation resistant (by virtue of the fact you'd cook yourself trying to build a bomb), but it also makes fuel reprocessing much more difficult (re: expensive) than that in conventional breeders. Any reprocessing scheme would have to be done by robots in a heavily shielded hot cell.
It's a misconception that this system produces "less waste" than conventional reactor. The fission product distribution isn't that different than U-235 or U-238, so you'll still get many of the same product nuclides. What is true is that the need for reprocessing in a Thorium reactor isolates the fission products right away, making the waste easier to deal with than the spent fuel we have on our hands today. With the exception of I-129, Tc-99 and a few others, most of the fission products are gone after 1000 years or so, a much more manageable timespan than unreprocessed fuel.
While there were a few lab-scale tests a long time ago, to this day there exists no full-scale prototype of a Thorium reactor. Getting the necessary licensing for a new reactor takes years, sometimes decades, and requires reams of lab data in order to conduct a proper safety analysis. With no modern operational examples, how can we get the necessary data needed for licensing? It's a tricky problem.
I could go on, but I think you get the idea. It's an interesting technology, but there are a lot of obstacles to overcome in order to implement it. Economics will be the driving factor in any decision to pursue Thorium reactors both in the US and internationally.
http://www.reddit.com/r/askscience/comments/qvfr0/if_thorium_is_safe_efficient_abundant_and_in/
