Topological track reconstruction in unsegmented, large-volume liquid scintillator detectors

Abstract

Unsegmented, large-volume liquid scintillator (LS) neutrino detectors have proven to be a key technology for low-energy neutrino physics. The efficient rejection of radionuclide background induced by cosmic muon interactions is of paramount importance for their success in high-precision MeV neutrino measurements. We present a novel technique to reconstruct GeV particle tracks in LS, whose main property, the resolution of topological features and changes in the differential energy loss $\mathrm{d}E/\mathrm{d}x$, allows for improved rejection strategies. Different to common track reconstruction approaches, our method does not rely on concrete track / topology hypotheses. Instead, based on a reference point in space and time, the observed distribution of photon arrival times at the photosensors and the detector's characteristics in terms of photon production, propagation and detection (optical model), it reconstructs the voxelized distribution of optical photon emissions. Techniques from three-dimensional data analysis can then be applied to extract parameters describing the topology, e.g., the direction of a track. We performed a first performance evaluation of our method using single muon events with up to $10\,\mathrm{GeV}$ from a Geant4 simulation of the LENA detector. The current results indicate that our approach is competitive with existing reconstruction methods -- although its full potential has not yet been exploited. We also remark on other detector technologies in astroparticle physics as well as applications in medical imaging that could benefit from the fundamental ideas of our method.