Criticality Unlikely at Yucca Mountain

I recently found myself in a discussion about the likelihood of a criticality (BOOM!) in the proposed repository at Yucca Mountain.

The argument for its being a concern went roughly as follows: Each giant steel canister of spent fuel would contain a large amount of U-235 and Pu-239. The reactor in Oklo contained less fissile (BOOM!able) material than that. Therefore, criticality is a real concern. The presence of water or hydrogen (e.g., associated with any steel corrosion products), or collapse of the shelves holding up the spent fuel would increase that possibility.

I’m not convinced a nuclear reaction is so easy to start. For one thing, we’re talking about spent nuclear fuel (SNF). That means it’s already undergone about as many fission reactions as our nuclear engineers could get out of it in the power plant. If it were so easy to cause the stuff to fission more, wouldn’t we still be using it for fuel? The fissile isotopes may be in there, but in order to for them to undergo fission again the fuel would need to be reprocessed, so that the fissile isotopes are separated from the non-fissile material. Even if some of the SNF is suitable for going critical a second time, there are neutron absorbers (I think mostly boron, but I’m not sure) in the package that should prevent this.

The license application for the proposed Yucca Mountain repository specifically addresses this issue in a 2042-page mega-document called “Features, Events, and Processes for the Total System Performance Assessment: Analyses”. According to this analysis, “for a configuration to have potential for criticality, all of the following conditions must occur: (1) sufficient mechanical or corrosive damage to the waste package outer corrosion barrier to cause a breach, (2) presence of a moderator (i.e., water), (3) separation of fissionable material from the neutron absorber material or an absorber material selection error during the canister fabrication process, and (4) the accumulation (external) or presence of a critical mass of fissionable material in a critical geometric configuration.” After analyzing the likelihood of each of these four factors, the report concludes that “the probability of criticality for the in-package location is much less than 1 chance in 10,000 of occurrence within 10,000 years after disposal.” There are a couple of revisions in later documents, but the ultimate conclusion is that a criticality event is so unlikely that there's no point in worrying about it.