Assessment of evaluated (n,d) energy-angle elastic scattering distributions using MCNP simulations of critical measurements and simplified calculation benchmarks

Different evaluated (n,d) energy-angle elastic scattering distributions produce k-effective differences in MCNP5 simulations of critical experiments involving heavy water (D2O) of sufficient magnitude to suggest a need for new (n,d) scattering measurements and/or distributions derived from modern theoretical nuclear models, especially at neutron energies below a few MeV. The present work focuses on the small reactivity change of < 1 mk that is observed in the MCNP5 D2O coolant-void-reactivity calculation bias for simulations of two pairs of critical experiments performed in the ZED-2 reactor at the Chalk River Laboratories when different nuclear data libraries are used for deuterium. The deuterium data libraries tested include Endf/B-VII.0, Endf/B-VI.4, JENDL-3.3 and a new evaluation, labelled Bonn-B, which is based on recent theoretical nuclear-model calculations. Comparison calculations were also performed for a simplified, two-region, spherical model having an inner, 250-cm radius, homogeneous sphere of UO2, without and with deuterium, and an outer 20-cm-thick deuterium reflector. A notable observation from this work is the reduction of about 0.4 mk in the MCNP5 ZED-2 CVR calculation bias that is obtained when the O-in-UO2 thermal scattering data comes from Endf-B-VII.0. (author)