Monte Carlo calculations of 14-MeV neutron multiplication in thick beryllium assemblies and comparison with experimental results

Integral experiments performed at the Institute for Reactor Development at Juelich, West Germany, have raised doubts about the adequacy of current nuclear data and calculational methods in predicting neutron multiplication and leakage from beryllium assemblies bombarded with 14-MeV neutrons. Experimental values of inferred neutron multiplication were reported to be less than calculated values by approx. 25%. We have performed calculations of the experiments using the TART Monte Carlo code. The ratio of measured leakage multiplication to our calculated leakage multiplication is 0.85 for 8-cm-thick beryllium and 0.99 for 12-cm-thick beryllium. However, much uncertainty exists in the procedure leading to the stated experiment values of apparent multiplication. We also performed calculations for a series of experiments done at LLNL from 1955 to 1956. A 14-MeV neutron source was placed in the center of a cylindrical beryllium assembly 8 in. in radius and 24 in. in height. The beryllium assembly was encased in an aluminum can surrounded by a manganese bath. The ratio of experimental to calculated neutron leakage multiplication is 0.88. Much uncertainty exists in both series of experiments