Monte Carlo simulations and characterization of hazardous substances detection using the SABAT neutron-based sensor integrated with an underwater drone

The aquatic environment contains numerous hazards, particularly those resulting from human activities, such as chemical and explosive remnants deposited on the seafloor. Detection and identification of these hazardous substances remains a major challenge due to limited accessibility and the risks associated with safe sampling and analysis. A promising alternative to commonly used methods is the use of neutrons, which enable non-invasive identification of substances using the neutron activation analysis technique. Additional separation based on the type of process by which the neutron interacts with matter through time measurement can increase the efficiency of material identification. Monte Carlo simulations in the Geant4 environment were performed by generating $$10^9$$ events for the SABAT detector mounted on a remotely operated vehicle, for five hazardous materials that may be found in the sea, such as: TNT, Mustard Gas, Clark1, Clark2, and Adamsite. The analysis of the generated energy spectra was performed by estimating the quality of a given characteristic line based on the Line Classifier parameter for identification using a dedicated spectral analysis program. Four characteristic lines with the highest intensity were selected as candidates for signal normalization: the annihilation line (0.511 MeV), the hydrogen line (2.223 MeV), the oxygen line (6.110 MeV), and the chlorine line (1.950 MeV). Water and hydrogen lines (2.223 MeV) were used as references for the material and the characteristic line, respectively. The annihilation line (0.511 MeV) proved to be at least twice as good (according to the line classifier) at identifying hazardous materials as the next-best lines, chlorine (0.79 and 1.17 MeV) and carbon (4.44 MeV). Additionally, time separation significantly improved only the quality of the oxygen (6.11 MeV) and chlorine (7.79 MeV) lines.