End-to-End Geant4 Waveform Modeling and Geometry Selection of EJ-276/SiPM Detectors for Neutron-Gamma Discrimination

An end-to-end simulation and experimental validation framework is presented for compact neutron detectors based on EJ-276 plastic scintillators coupled to silicon photomultipliers. The workflow combines Geant4 optical-photon transport with a physics-based signal-formation model to predict preamplifier-level voltage waveforms. Benchmarking with a cobalt-60 source demonstrates close waveform agreement after baseline subtraction and peak-time alignment, with high similarity over 100-500 ns (correlationr= 0.997 and normalized root mean square error nRMSE= 2.8%). The model preserves key pulse observables, including peak amplitude (0.5% difference), peak timing (about 1 ns), and pulse integral (0.34% difference), while the residual mismatch is mainly confined to fast/tail dynamics (slightly faster rise and a shorter effective decay constant in simulation). The validated model is then applied to eight geometries to quantify trade-offs between neutron detection efficiency and neutron–gamma separation. Using figure of merit, intrinsic detection efficiency, and misclassification rate, we identify that designs near Ø50 mm × 50 mm balance discrimination and efficiency for compact, field-deployable instruments. The proposed approach enables predictive optimization of pulse-shape discrimination capable plastic-scintillator detectors without iterative prototyping, and is readily extensible to other silicon photomultipliers based systems.