An event-triggered coincidence algorithm for fast-neutron multiplicity assay corrected for cross-talk and photon breakthrough

A model quantifying detector cross-talk and the misidentification of events in fast neutron coincidence distributions is described. This is demonstrated for two experimental arrangements comprising rings of 8 and 15 organic liquid scintillation detectors. Correction terms developed as part of this model are tested with  252Cf and a relationship is developed between the  235U enrichment of U3O8 and the order of correlated, fast neutron multiplets induced by an americium-lithium source. The model is also supported by Geant4 simulations. The results suggest that a typical assay, for experimental arrangements that are similar to the examples investigated in this research, will exhibit cross-talk for less than 1% of all detected fast neutrons but, if not accounted for, this can bias the numerical analysis by a margin of 10% and 35% in second- and third-order coincidences (i.e. couplet and triplet counts), respectively. Further, for the case of  252Cf, it is shown that a relatively low proportion of 4% breakthrough by γ rays (that is, photons misidentified as neutrons by the pulse-shape discrimination process) can lead to an erroneous increase of 20% in total neutron counts in the assay of a mixed-field, in this case of  252Cf. These findings will help direct the developments needed to enable organic scintillation detectors with pulse shape discriminators to be applied reliably to nuclear safeguards and non-proliferation verification tasks.