Nuclear Forensics—Fission Track Analysis—Simulation and Image Analysis

To answer the nuclear forensics questions, we are developing new techniques and new approaches to make this analysis more reliable and more accurate.<br/>In nuclear forensics we usually deal with very low quantities of sample, in the range of picograms.<br/>From one side we need very sensitive equipment to deal with those very low quantities of sample, from the other side we need a selective technique so we will not need to measure every particle in the sample.<br/>In one swipe we can get up to 10 million particles, but only a few of them are interesting from a nuclear forensic point of view. As an example, in one swipe of natural soil, only few tens of particles are uranium. Measuring the ratio between isotopes can indicate the enrichment level. For this reason, we developed the fission track analysis (FTA). In that technique we place our particles in between two solid state plastic detectors (Lexan).<br/>The Lexan detectors with the embedded particles between them are radiated in a reactor with high thermal neutrons flux for several minutes. The flux of neutrons causes a nuclear fission in the material particles. The fission products are indenting marks (fission tracks) on the Lexan detectors.<br/>Since this reaction occurs randomly and uniformly according to the half-life of the specific isotope, we get our fission tracks that look like a rose or a sea urchin having a 3D character.<br/>High resolution pictures (taken by scanning stage microscope) of the given detector are taken to identify the locations of the fission particles (2D).<br/>In this work, we show the image analysis of the pictures to find the location of the fission materials.<br/>To have the ability to simulate those fission stars and have better understanding of the nature of the fission of the tracks we developed suitable software.<br/>The fission tracks were simulated by Monte-Carlo software, GEANT4, which uses all the physics behind the nuclear fission tracks - thermal neutrons flux, fission cross-section, radiation time, particle size, enrichment, etc.<br/>Identify the location of the particles of a real or simulated roses found in the pictures can be done with that new software.<br/>A new approach of identifying the different stars was developed by utilizing AI (Artificial Intelligence).<br/>After being properly located, the particle can be separated from the rest of the sample and used for a mass spectrometry technique, like ICP-MS or TOF-SIMS.<br/>Two examples of fission material measured will be shown, natural sand and a soil sample from Trinity test (first big-scale test in New Mexico, July 16, 1945).