Configuring an NGn Prototype Detector using Cosmic Muons for CNN Training and Testing

The MoNA Collaboration is developing a high-resolution Next Generation Neutron (NGn) detector array, for use with the MoNA-LISA array in future invariant-mass spectroscopy measurements at FRIB. Our current prototype NGn detector consists of a 24x24x2.5 cm plastic scintillator plate equipped with a 16 Silicon Photomultiplier (SiPM) array on one face. We are developing a CNN machine learning model for position reconstruction in the detectors, with a goal of reducing the position resolution to under 1 cm. We have configured the prototype detector and GEANT4 simulations to provide training data for these models. We have tested different SiPM gain-matching techniques using spectral photon peak separations, SiPM breakdown voltages and an on-board calibration pulser, different plate reflectivities, and coupled vs uncoupled SiPMs. We also determined the optimal bias and gain settings for the plate and tracking detectors for efficient detection of cosmic muons, which will provide training and testing data for our CNN. In Fall 2025 we will test our prototype detectors at the Los Alamos Neutron Science Center (LANSCE) using fast neutrons.