Suppression of $^{14}\mathrm{C}$ photon hits in large liquid scintillator detectors via spatiotemporal deep learning

View PDF HTML (experimental)

Abstract:Liquid scintillator detectors are widely used in neutrino experiments due to their low energy threshold and high energy resolution. Despite the tiny abundance of $^{14}$C in LS, the photons induced by the $\beta$ decay of the $^{14}$C isotope inevitably contaminate the signal, degrading the energy resolution. In this work, we propose three models to tag $^{14}$C photon hits in $e^+$ events with $^{14}$C pile-up, thereby suppressing its impact on the energy resolution at the hit level: a gated spatiotemporal graph neural network and two Transformer-based models with scalar and vector charge encoding. For a simulation dataset in which each event contains one $^{14}$C and one $e^+$ with kinetic energy below 5 MeV, the models achieve $^{14}$C recall rates of 25%-48% while maintaining $e^+$ to $^{14}$C misidentification below 1%, leading to a large improvement in the resolution of total charge for events where $e^+$ and $^{14}$C photon hits strongly overlap in space and time.

Comments: 14 pages, 11 figures

Subjects:

Instrumentation and Detectors (physics.ins-det); Artificial Intelligence (cs.AI); High Energy Physics - Experiment (hep-ex)

Cite as: arXiv:2603.27727 [physics.ins-det]

(or arXiv:2603.27727v1 [physics.ins-det] for this version)

https://doi.org/10.48550/arXiv.2603.27727

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Junle Li [view email] [v1] Sun, 29 Mar 2026 15:03:42 UTC (15,870 KB)