NEUTRON CORRELATIONS IN SPECIAL NUCLEAR MATERIALS, EXPERIMENTS AND SIMULATIONS1

Fissile materials emit neutrons with an unmistakable signature that can reveal characteristics of the material. Fission chains in particular in special nuclear material (SNM) give rise to a distinctive, measurable time correlation signal. An analysis was developed to infer the presence of SNM in a system. Using this analysis, it is possible to infer the presence of SNM in an assembly containing a few kilograms of highly-enriched uranium (HEU) for instance, purely passively, using helium detectors of modest size, in a reasonable time. In order to illustrate our capability to identify SNM, we show examples of experimental data, and compare it to Monte Carlo simulations. These examples span a range of possible sources. Three examples are presented: a) a sample of 252Cf, a nonmultiplying spontaneous fission source, b) an unclassified multiplying ball of plutonium moderated by 3” thick polyethylene, and c) an unclassified bare shell of HEU, a very weak low-multiplying neutron source. For the three examples, we analyze the neutron counts for time correlations. We compare random time gate neutron count distributions to Poisson distributions of the same count rate, which gives a visual signature of fission. Given enough measurement time, even a weak and low-multiplying source such as HEU occasionally generates very long chains and a distinctive detectable fission signal. We also examine the time-dependent pair and triple cumulant moments, which give a visual signature of moderated fission. Monte Carlo simulations using COG are compared with the experimental results. Statistically, the simulated count distributions and cumulant moments are essentially the same as the measured data in all three cases. For nonmultiplying sources like californium, the mass of fissile material can be determined by the neutron count rate and the Feynman pair moment. A theory exists for multiplying systems, but will not be presented here. Key lessons from the LLNL measurements and COG simulations are that even low multiplying systems with low spontaneous fission activity can be detected, and that simulations can faithfully model the physical features of fissioning systems. Based on the results one can judge the capability of simulations to faithfully reproduce as yet unmeasured objects.