Why is nuclide identification so difficult? Lessons from the RASE program

Correct nuclide identification is critical for first responders. It is the final result of the complex chain occurring in a detector: a spectrum with poor statistics in a variable environment is recorded then analyzed. Industries are facing a complex challenge, as they have to design instruments that can identify materials that for the most part are difficult to obtain. Modelling a source is a viable option to skirt this issue, or participating in data collection events organized by governmental agencies such as DNDO with their RASE program. Modelling a source means spending a lot of processing time for a result that is not likely to be an exact replica of the real world, especially when it comes down to backscattering. Data collection events are only a small excerpt of the real world: performance evaluation based on only that may not reflect the true performances of the instrument. Nuclide identification can fail due to poor instrument performance, a template or peak database that is not optimized. The conditions of the experiments strongly influence the final spectrum and can throw the identification off. Some examples of source modelling (infinite volume sources, planar sources), experiments and database building considerations will be presented in this paper.