Monte Carlo parametric studies of neutron interrogation with the Associated Particle Technique for cargo container inspections

Abstract

The detection of Special Nuclear Materials (SNM) by neutron interrogation is extensively studied by Monte Carlo simulation at the Nuclear Measurement Laboratory of CEA Cadarache (French Alternative Energies and Atomic Energy Commission). The active inspection system is based on the Associated Particle Technique (APT). Fissions induced by tagged neutrons (i.e. correlated to an alpha particle in the DT neutron generator) in SNM produce high multiplicity coincidences which are detected with fast plastic scintillators. At least three particles are detected in a short time window following the alpha detection, whereas nonnuclear materials mainly produce single events, or pairs due to (n,2n) and (n,n’γ) reactions. To study the performances of an industrial cargo container inspection system, Monte Carlo simulations are performed with the MCNP-PoliMi transport code, which records for each neutron history the relevant information: reaction types, position and time of interactions, energy deposits, secondary particles, etc. The output files are post-processed with a specific tool developed with ROOT data analysis software. Particles not correlated with an alpha particle (random background), counting statistics, and time-energy resolutions of the data acquisition system are taken into account in the numerical model. Various matrix compositions, suspicious items, SNM shielding and positions inside the container, are simulated to assess the performances and limitations of an industrial system.

Sea container inspection with tagged neutrons

“…Output files are processed with dedicated software based on ROOT [77]. In addition to the useful signal induced by tagged neutrons, simulation also takes into account untagged neutrons emitted isotopically by the generator that create accidental coincidences and reduce the selectivity between nuclear and non-nuclear materials [78]. The simulation of an isotropic neutron emission with MCNP-PoliMi gives the main characteristics of random background events, that is, their neutron and gamma energy spectra, and their spatial distribution on detectors as illustrated in Figure 34.…”

Detection of special nuclear materials with tagged neutrons

“…The detection of explosives and other illicit materials with the APT has been simulated with MCNP dedicated data processing tools [7], which allowed to model realistic measurements taking into account the counting statistics, random background, time and energy resolutions of the data acquisition system. Since the application of APT for the detection of SNM involves multiple coincidence analysis, MCNPX PoliMi [8] has been used here to track the events taking place in the plastic scintillators for each tagged neutron history, and specific post-processing tools have been developed with ROOT framework [9] to take into account the random background and to implement cross-talk rejection algorithms [10]. To test the validity of the simulation tools, the experimental setup shown in Fig.2 has been built at CEA DEN Cadarache.…”

“…The system proposed in Fig.1 includes three large panels of plastic scintillators (300 cm length, 240 cm sides, and 10 cm thickness), each panel being made of 48 individual 37.5 cm × 40 cm × 10 cm detectors [10]. As indicated in this reference, a hole in the panel facing the neutron generator allows limiting the detection of tagged neutrons which did not interact in the container.…”

“…along the tagged neutron beam, and z is height. The borders of the fission neutron time window have been adapted to each (y,z) position [10], according to neutron TOF. The evolution of the three-fold coincidence ratio between the item (HEU, iron, or lead) and the bare iron matrix can be explained by a combination of different effects.…”

“…when y tends to -110 cm or +110 cm, fission neutrons can be detected inside the same plastic scintillator, or in neighboring detectors. In such instances, only one particle is counted by cross-talk rejection algorithms [10], and the coincidence multiplicity is therefore underestimated. Therefore, the coincidence ratio decreases, but one can note that it remains larger than 1.5 regardless of (y,z) position.…”