Pulse-shape discrimination with machine learning for CZT detectors at the DAΦNE beam test facility
Abstract
Cadmium zinc telluride (CZT) detectors offer versatility, operational simplicity, and room-temperature X- and gamma-ray spectroscopy, making them attractive for collider applications, yet their use under high-flux conditions remains limited. Here, we present a preliminary feature-based pulse-shape analysis employing machine learning, on data acquired with a quasi-hemispherical CZT detector at the DAΦNE beam test facility of the National Laboratory of Frascati of INFN. A 300-MeV electron beam impinging on a lead target produced characteristic Pb X-rays together with a broad background extending up to the electron-positron annihilation region. Physically motivated temporal and morphological features were extracted from the recorded waveforms and used to distinguish nominal photon-like pulses from anomalous events. An XGBoost classifier trained and validated on 10,000 labeled waveforms achieved an accuracy of approximately 97%, with most of its classification performance reached using only a few hundred labeled examples. The trained model was applied to more than 700,000 events, substantially reducing the spectral continuum and coincidence peaks, while preserving the characteristic Pb X-ray lines up to the 511-keV annihilation peak. These preliminary results demonstrate the potential of machine-learning-assisted pulse-shape discrimination for improving CZT spectroscopy in collider environments.
Related
- Pulse shape discrimination for event rejection in BEGe-type high-purity germanium detectors
- Harnessing Machine Learning for Ultrafast Science in Facility Settings: Opportunities and Challenges
- A comparative study of variational quantum classifiers and classical machine learning for tW−tt̅ discrimination at the LHC
- Enhancing electromagnetic calorimeter signal reconstruction with machine learning-based noise discrimination
- Lead-free piezoelectric perovskites for arterial-pulse e-skin: from configurational complexity to equivariant machine-learning potentials
BibTeX
@misc{manti2026pulseshape,
title = {{Pulse-shape discrimination with machine learning for CZT detectors at the DA$Φ$NE beam test facility}},
author = {Simone Manti and Francesco Artibani and Leonardo Abbene and Massimiliano Bazzi and Manuele Bettelli and Giacomo Borghi and Damir Bosnar and Mario Bragadireanu and Antonino Buttacavoli and Mario Carminati and Alberto Clozza and Francesco Clozza and Luca De Paolis and Raffaele Del Grande and Kamil Dulski and Carlo Fiorini and Ivica Friščić and Gaetano Gerardi and Carlo Guaraldo and Mihai Iliescu and Masa Iwasaki and Alexander Khreptak and Johan Marton and Pawel Moskal and Hiroaki Ohnishi and Kristian Piscicchia and Fabio Principato and Alessandro Scordo and Francesco Sgaramella and Michał Silarski and Diana Sirghi and Florin Sirghi and Magdalena Skurzok and Antonio Spallone and Kairo Toho and Oton Vazquez Doce and Andrea Zappettini and Johann Zmeskal and Catalina Curceanu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.25156},
eprint = {2609.25156},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.25156},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Pulse-shape discrimination with machine learning for CZT detectors at the DAΦNE beam test facility
AU - Simone Manti
AU - Francesco Artibani
AU - Leonardo Abbene
AU - Massimiliano Bazzi
AU - Manuele Bettelli
AU - Giacomo Borghi
AU - Damir Bosnar
AU - Mario Bragadireanu
AU - Antonino Buttacavoli
AU - Mario Carminati
AU - Alberto Clozza
AU - Francesco Clozza
AU - Luca De Paolis
AU - Raffaele Del Grande
AU - Kamil Dulski
AU - Carlo Fiorini
AU - Ivica Friščić
AU - Gaetano Gerardi
AU - Carlo Guaraldo
AU - Mihai Iliescu
AU - Masa Iwasaki
AU - Alexander Khreptak
AU - Johan Marton
AU - Pawel Moskal
AU - Hiroaki Ohnishi
AU - Kristian Piscicchia
AU - Fabio Principato
AU - Alessandro Scordo
AU - Francesco Sgaramella
AU - Michał Silarski
AU - Diana Sirghi
AU - Florin Sirghi
AU - Magdalena Skurzok
AU - Antonio Spallone
AU - Kairo Toho
AU - Oton Vazquez Doce
AU - Andrea Zappettini
AU - Johann Zmeskal
AU - Catalina Curceanu
PY - 2026
DA - 2026/09/21
PB - arXiv
DO - 10.48550/arxiv.2609.25156
UR - https://arxiv.org/abs/2609.25156
AB - Cadmium zinc telluride (CZT) detectors offer versatility, operational simplicity, and room-temperature X- and gamma-ray spectroscopy, making them attractive for collider applications, yet their use under high-flux conditions remains limited. Here, we present a preliminary feature-based pulse-shape analysis employing machine learning, on data acquired with a quasi-hemispherical CZT detector at the DAΦNE beam test facility of the National Laboratory of Frascati of INFN. A 300-MeV electron beam impinging on a lead target produced characteristic Pb X-rays together with a broad background extending up to the electron-positron annihilation region. Physically motivated temporal and morphological features were extracted from the recorded waveforms and used to distinguish nominal photon-like pulses from anomalous events. An XGBoost classifier trained and validated on 10,000 labeled waveforms achieved an accuracy of approximately 97%, with most of its classification performance reached using only a few hundred labeled examples. The trained model was applied to more than 700,000 events, substantially reducing the spectral continuum and coincidence peaks, while preserving the characteristic Pb X-ray lines up to the 511-keV annihilation peak. These preliminary results demonstrate the potential of machine-learning-assisted pulse-shape discrimination for improving CZT spectroscopy in collider environments.
ER -