Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
HK: Fachverband Physik der Hadronen und Kerne
HK 27: Poster Session
HK 27.20: Poster
Mittwoch, 18. März 2026, 16:15–18:30, Redoutensaal
PID using unsupervised clustering methods and pulse shape discimination in a Phoswich detector — •Simon Glennemeier-Marke1,2, Kai-Thomas Brinkmann1,2 und Lara Dippel1,2 — 12nd Physics Institute, Justus-Liebig-University, Gießen — 2LOEWE Research Cluster for Advanced Medical Phzsics in Imaging and Therapy (ADMIT)
On this poster, we present a workflow for particle identification (PID) in a Phoswich detector, which consists of a fast plastic scintillator optically coupled to a thicker barium fluoride (BaF) scintillator. The dissimilarity in the scintillation characteristics of both materials enables pulse shape discrimination (PSD) between different particle species. We use a dataset obtained at Marburg Ion Beam Therapy Center (MIT) using 300 MeV/u carbon ions impinging on a water target. Feature engineering and unsupervised clustering methods are applied to categorize new observations, unseen during training, even in high rate environments where signal pile-ups occur. Classified detections can then be used to determine the relative abundances. This project is part of the ADMIT consortium under Project Part A, which focuses on estimating spectral neutron fluxes for flash therapy in tumor treatment applications. This project is financed with funds of LOEWE - Landes-Offensive zur Entwicklung Wissenschaftlich-ökonomischer Exzellenz, Förderlinie 2: LOEWE-Schwerpunkte.
Keywords: Particle Identification (PID); Phoswich Detector; Machine Learning; Pulse Shape Discrimination