A new optical photon transport model for application to high aspect ratio scintillation pillars

Abstract

A new Monte Carlo model of optical photon transport has been developed for application to scintillation pillars with side surfaces modeled using their measured 3D topography. The model addresses deficiencies of existing models specifically pertinent to high aspect ratio scintillator geometries. Inaccuracies with these models’ description of the reflectance of scintillation photons at media boundaries are exacerbated by the high number of surface interactions in such geometries. These models rely either on (a) user-specified parameters that are challenging to set, or (b) pre-computed look-up tables that assume geometrically unconstrained photon incidence on uniform surface finishes, an assumption that is invalid in high aspect ratio geometries. Alternatively, the new model reconstructs the side surfaces of a pillar from 3D topographic measurements and uses first-principles physics of reflection and transmission to transport photons over the surfaces’ microscopic features and throughout the entire pillar volume. It takes advantage of the recent availability of open-source software capable of complex surface modeling and ray-tracing, as well as high performance computing. The model was validated against measurements of the number of photoelectrons (PEs) generated in two silicon photomultipliers (SiPMs) located at the ends of 5 mm × 5 mm × 200 mm EJ-204 scintillation pillars with different surface finishes. It was also validated against the relative amplitudes and arrival time difference of the SiPMs’ signals. The model accurately predicts the measurements as function of the scintillation position along the pillars’ length with 2% mean-absolute-percentage error (MAPE) in the number of PEs, and 16% in the relative amplitudes. The root-mean-squared error (RMSE) in the arrival time difference is 0.114 ns. The model also predicts the experimentally observed higher scintillation light collection in pillars with smoother side surfaces. A proof-of-concept code of the model is available as open source.