Würzburg 2018 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 65: Detektorsysteme II

T 65.2: Vortrag

Mittwoch, 21. März 2018, 16:45–17:00, Z6 - SR 2.007

Pulse shape discrimination using machine learning techniques for neutron detectors — Oliver Pooth, •Christian Teichrib, Simon Weingarten, and Christian Wysotzki — III. Physikalisches Institut B, RWTH Aachen University

The Physics Insitute III B is working on the development of a detector for fast neutrons based on stilbene sctintillator crystals and silicon photomultipliers. It is to be used for neutron tomography in future probe analyses in extractive industries. For this purpose the proper distinction of neutrons from photons originating from a radioactive neutron source is critical. Classical separation techniques utilise a pulse shape discrimination variable relying on a longer decay time component in neutron induced signals in order to distinguish between the neutron signal and photon background. One major weakness of this method lies in the difficulty of separating low energetic events.

An alternative approach using machine learning can extract more complex features out of pulse shapes allowing a potentially better discrimination between neutrons and photons. The performance of this method is discussed with regard to its advantages and disadvantages compared to the classical approach.