Year
2003
Abstract
Neutron multiplicity measurements of small and large plutonium materials performed recently in the short-die-away-time Neutron-Capture Counter for Residues (NCCR) were analyzed by a numerical procedure that is new to multiplicity analysis. The technique calculates singles doubles and triples rates for samples with any combination of 240Pu effective mass, multiplication and “alpha” (m, M and a) measured in the NCCR. The procedure uses Monte-Carlo simulation of the NCCR that tracks fission and a,n neutrons emitted by a plutonium item (specified by its m, M and a) in time. The procedure also includes a software emulator of the multiplicity shift register to determine the multiplicity distributions of tracked events. A numerical alternative to the analytical point model was required to analyze the experimental data. This arose from the experimental need to eliminate same-channel correlated events from the prompt (real plus accidental) multiplicity distribution for the short-die-away-time NCCR. The hardware change alters the valid accidental multiplicity distribution so that accidentals can no longer be determined analytically from the data. The numerical analysis executes uniquely for multiplicity measurements of each measured item. Analytical point-model assumptions guide iterative adjustments of m, M and a for the item. Multiple iterations converge rapidly to reproduce the measured singles, doubles and triples for a given m, M and a. Multiplicity measurements of small metal and oxide samples, large oxides with low a, and small samples with high a performed with the NCCR and analyzed by the new numerical approaches show good results in a wide range of 240Pu effective mass. Energetic proton recoils account for detection of excess doubles in measurements of the highest-a materials. This paper describes the numerical approach, which offers improvements to multiplicity analysis in general. Results of multiplicity measurements with the NCCR and experimental evidence for understanding the behavior at high a are presented. A solution to the excess doubles is implicit in these experimental NCCR results.