In recent years, non-destructive evaluation techniques which use either photon or neutron sources from accelerators followed by neutron counting signatures have been used in many national security and nuclear nonproliferation applications [4, 60]. Although the United States customs and border protection initiated and implemented a cargo security initiative to discover threats from others countries before they embark to the US, detectors with better sensitivities are more necessary than ever in view of the global threats faced by nations around the world. Photofission based applications which use delayed neutron signal ores as viable detection schemes for fissile material detection have been ongoing for many years. Applications of this technology to include cargo scanning applications are however lacking.

This work in this dissertation used the delayed neutron signature counting technique for fissile material detection in conjunction with new formulated Curries' expressions to establish the sensitivity (minimum detectable mass) limits. The fission reactions were induced in a uranyl nitrate solution containing 94.1 g of ²³⁸U using bremsstrahlung endpoint cue pies of 9 MeV to 21 MeV in 2 MeV steps. Preliminary data on the sensitvity measurement at bremsstrahlung end point energies of 9, 14, 18 and 22 MeV are also presented. We also present the effect of borated polyethylene and lead shielding on the sensitivity at 9 and 22 N1cV. The sensitivities were calculated for 5%u false positives and 5% fake negatives as well as for 1% false positives and 0.1% false negatives. A dose of 4 Gy, 5 mGy and 1 mGy were assumed to be delivered to Mutt cargo container. For a radiator and target-to-detector distance of 150 cm and 200 cup, the delayed neutron yield from calculation and experiment was also compared. Finally, feasibility studies was conducted to determine if the linac parameters used in this research was capable of detecting 1 mg, 1 g and 1 kg of ²³⁸U. This work was funded by a grant from FA8650-01-2-6541.