Validation of GATE optical transportation with experimental single scintillator dual-end readout data

GATE is open source software for emission tomography Monte Carlo simulation. Since the version 3.0, GATE simulates not only the high energy physics process inside phantoms and detectors, but also the generating and tracking of optical photons as well, which provides a potential useful tool for the block detector design for PET/SPECT. However, while GATE is using standard libraries and constants for high energy interactions provided by Geant4, it lacks of standard models and constants for common materials in the optical photon range, which makes the optical simulation difficult due to the sensitiveness of the output dependence on the optical properties of the materials, such as the refraction index, attenuation length, surface roughness, etc. Users must provide correct optical parameters in order to get meaningful and trustable results. In this study, with the model of a single scintillator coupled to two SiPMs at both ends, we simulate the dual-end DOI decoding using GATE optical transportation function. Different optical parameters are tested and the results are evaluated by published experimental data. The LSO scintillator of 20, 30 and 40 mm length with saw-cut surface roughness are simulated. The results of the DOI decoding ratio, the DOI resolution, the energy resolution and the energy peak shifting are used to evaluate the accuracy of the simulation modeling. Critical parameters for the optical model are found. The Lambertian component is the critical and dominant factor in the surface finish. We also find that additional noises are needed to match the simulations to the real data in terms of DOI and energy resolution. The whole sets of optical parameters that match the simulations to the experiments are obtained. The parameters and the finding in this study will be the important references for future block design using GATE optical transportation simulations.