Statistics to Identify and Quantify Pure Radionuclides by Their Neutron Emission Patterns

Perfect Knowledge of emission time for every neutron from a radionuclide allows identification of the material and estimation of the quantity present via nondestructive assay. In practice, we lack this perfect knowledge. Detector efficiency is less than unity, thermalized neutron arrival is delayed randomly, and neutron showers triggered by cosmic rays in our atmosphere ass noise to the process. We compare two classes of statistical estimates capable of recovering characteristic parameters for radionuclides from this imperfect informations method of moments (the current standard) and maximum likelihood, which holds the advantage as the number of parameters to be estimated simultaneously increases.