A new look at old fission mysteries

Much has changed in the field of high-performance computing and modeling since Möller did that calculation. Punched cards are obsolete, computing has become ubiquitous and Laboratory physicists like Möller now use computers to explore their theoretical models in previ- ously unimaginable detail. Faster, more powerful computers mean that the number of grid points or data sets (in this case the number of nuclear shapes) that can be investigated can now run in the hundreds of thousands or even millions.

Recently Möller, together with David Madland and Arnold Sierk at Los Alamos and Akira Iwamoto of the Japan Atomic Energy Research Institute, had yet another chance to push the limits of computational power at Los Alamos and at the same time help to further unravel one of the great mysteries in nuclear science – the process of nuclear fission.

Since its discovery in 1938, the phenomenon of fission has frequently been explained in terms of a liquid drop. In such a depiction, when a nucleus starts to deform the energy increases, caused by the surface tension of the drop. If the nucleus deforms, but is stopped early in the deformation process, it snaps back to its original shape just like a rubber band that is pulled out and released. But if the nucleus is deformed beyond a certain configuration – beyond a point of no return – it snaps, and like the rubber band, the two fragments fly apart.

Möller's computer model is based upon a similar analogy of a ball being pushed up toward a mountain pass. The pass itself represents the point of no return:after being crossed the ball will roll down into another mountain valley. The height of the mountain pass corresponds to the threshold energy of the fissioning nucleus. Since a five-dimensional energy landscape cannot be visualized on a two-dimensional sheet of paper, unlike that of a geographical map, a challenge in the group's research was to establish which of the many passes in the five- dimensional energy landscape represented the relevant fission threshold. This problem was solved by considering, in the computer model, imaginary water flowing in five dimensions.

Möller's model used nearly three million physical grid points to define critical shape coordinates related to various aspects of elongation, neck diameter, emerging fragment deformation and mass division in the fission of radium and fermium. Because several million grid points and five shape dimensions are required to reach a sufficient level of physical detail to adequately describe fission, structures such as those revealed in the calculation by Möller and his collaborators had never before been seen or identified in nuclear structure calculations.

The results of this groundbreaking research have allowed a number of fundamental conclusions to be drawn about the fission process. First,there are several fission paths possible for most heavier nuclei, which means the fission process is more complex than is accounted for in most existing models. Second, for lighter actinide elements like radium and thorium, two paths dominate: one mass-asymmetric, with division into unequal fragment masses, and another mass-symmetric with equal fragment masses. Finally, the calculations are in agree- ment with experimental observations that for elements lighter than fermium – that agreement being that the average kinetic energy is higher for the asymmetric mode than for the symmetric mode. The calculations also reproduce, for the first time, the average fragment masses observed in fission.

The net result of this research is a greater and more comprehen- sive understanding of nuclear structure and the underlying mechanisms behind nuclear fission. The new insight into fission obtained from the computer studies by Möller and his colleagues are expected to lead to improvements in related models associated with science-based stockpile stewardship, the safe storage of nuclear waste and even the synthesis of elements in supernovae. The group's most recent calculations required about 2,000 CPU days of computer time to process and were performed on the Avalon cluster at Los Alamos – a group of 144 interconnected computers running at 500- MHz each. Funding from the Department of Energy's Offices of Defense Programs and Science supported Möller's work.