High-fidelity MCNP modeling of a D–T neutron generator for active interrogation of special nuclear material

Abstract

Abstract Fast and robust methods for interrogation of special nuclear material (SNM) are of interest to many agencies and institutions in the United States. It is well known that passive interrogation methods are typically sufficient for plutonium identification because of a relatively high neutron production rate from 240Pu [1] . On the other hand, identification of shielded uranium requires active methods using neutron or photon sources [2] . Deuterium–deuterium (2.45 MeV) and deuterium–tritium (14.1 MeV) neutron-generator sources have been previously tested and proven to be relatively reliable instruments for active interrogation of nuclear materials [3] , [4] . In addition, the newest generators of this type are small enough for applications requiring portable interrogation systems. Active interrogation techniques using high-energy neutrons are being investigated as a method to detect hidden SNM in shielded containers [4] , [5] . Due to the thickness of some containers, penetrating radiation such as high-energy neutrons can provide a potential means of probing shielded SNM. In an effort to develop the capability to assess the signal seen from various forms of shielded nuclear materials, the University of Michigan Neutron Science Laboratory’s D–T neutron generator and its shielding were accurately modeled in MCNP. The generator, while operating at nominal power, produces approximately 1×1010 neutrons/s, a source intensity which requires a large amount of shielding to minimize the dose rates around the generator. For this reason, the existing shielding completely encompasses the generator and does not include beam ports. Therefore, several MCNP simulations were performed to estimate the yield of uncollided 14.1-MeV neutrons from the generator for active interrogation experiments. Beam port diameters of 5, 10, 15, 20, and 25 cm were modeled to assess the resulting neutron fluxes. The neutron flux outside the beam ports was estimated to be approximately 2×104 n/cm2 s.