In my part of Africa, when people are faced with a large, daunting problem, someone often asks, "How do we eat this elephant?" The typical response is: "Piece by piece!" Minimizing the use of highly enriched uranium (HEU) can be seen as an elephant that, indeed, can only be eaten piece by piece. But it may also be the case that certain parts of this animal are simply inedible.
My colleagues Pablo Cristini and Alexandr Vurim have expressed somewhat different views about when it is reasonable to expect that reactors using HEU should convert to low-enriched uranium (LEU). But consensus has emerged on one point: At reactors that produce molybdenum 99, converting fissionable targets to LEU is a feasible project, though possibly a painful one. Much the same determination was reached in a 2012 report by the Nuclear Energy Agency, which concluded that, despite the investments of money and time that conversion entails, "conversion is important and will occur." In any event, as long as only a few nations supply HEU, and these countries continue to exert pressure on molybdenum-producing reactors to convert, these facilities will find that their alternatives to conversion are very limited. Reactors that want to be involved in commercial production of medical isotopes will eventually have to eat their piece of the elephant -- or go hungry.
Another portion of the meal involves certain research reactors that use HEU and are engaged in activities beyond isotope production. New research reactors, of course, are generally expected to be designed in such a way that they can function well with low-enriched uranium. But older reactors may find their elephant flesh too tough to swallow. This is mainly because possibilities for reconfiguring reactor cores are often limited at research reactors with older designs. In these cases -- for instance at Vurim's IGR reactor in Kazakhstan -- conversion in many cases implies reduced performance.
At these facilities, the developed world can provide a sort of meat tenderizer by furnishing technical and financial support for conversion efforts. But even support of this type won't always be sufficient -- not as long as the cut-off between LEU and HEU remains defined as 20 percent uranium 235. So in a few instances, the meal simply cannot be served, and any international minimization strategy that aims to make all research reactors completely safe from a proliferation perspective is unlikely to succeed. The International Atomic Energy Agency's database of research reactors gives an indication of the scope of the problem: Many dozens of the reactors listed there have little financial or technical ability to convert on their own. More to the point, they have no real desire to convert. When it comes to eating the elephant, they just aren't hungry enough.
Even if conversion can't be accomplished in all cases, developed countries can contribute to nonproliferation by continuing to carry out risk evaluations of HEU inventories in developing nations, and by helping to improve security and material management programs where appropriate. Inventories that pose the highest risks -- for example, stocks of fresh HEU fuel in developing countries with insufficient programs for nuclear material management and security -- should be given special priority. Spent fuel, in view of its high radioactivity, to a certain extent provides its own deterrent to theft or diversion.
Nuclear-powered submarines represent perhaps a more alarming problem than research reactors. Submarines are by definition physically isolated much of the time. They will inevitably make ports of call in foreign countries, whether planned or not, and it is not really possible to apply proper security to facilities that are so mobile. And as Vurim indicated in his second essay, HEU for naval applications is unlikely to be eliminated any time soon.
But when it comes to research reactors, conversion to LEU at facilities that produce medical radioisotopes is a meal already on the table, and one that will be consumed fairly quickly. Conversion of all research reactors to LEU, however -- this animal may be even harder than an elephant to eat.