Neural Network Analysis of Germanium Detector Waveforms (2023)

Undergraduate: Ravi Pitelka

Faculty Advisor: Julieta Gruszko
Department: Physics and Astronomy

---

The LEGEND experiment is a next-generation neutrinoless double beta decay detector, for which background modeling and rejection is of the utmost importance. To facilitate the use of Germanium detectors as Compton cameras for low-energy backgrounds, a neural network was trained to extract positional information out of input waveforms. Data was collected at Los Alamos National Lab using a Co-60 source. The network achieved an average classification accuracy of 70% for the vertical position classes. Non-uniform classification results for the azimuthal position classes indicate that the detector may have been off-center within the cryostat. Future work will involve testing the network on low energy backgrounds to prepare for usage in LEGEND’s 200-kg phase.

Link to Poster