On muon energy group structure based on deflection angle for application in muon scattering tomography: A Monte Carlo study through GEANT4 simulations

TL;DR

This study investigates muon energy classification using deflection angles in muon scattering tomography, exploring a four-group structure and assessing misclassification probabilities through GEANT4 simulations to improve energy estimation accuracy.

Contribution

It introduces a four-group energy classification scheme based on deflection angles and evaluates misclassification probabilities, enhancing muon energy estimation methods in tomography.

Findings

Misclassification probabilities increase with more energy groups.

Four-group classification allows rough energy estimation up to 4 GeV.

Classification accuracy diminishes beyond four groups.

Abstract

The average deflection angle of the tracked muons in the muon scattering tomography exponentially declines in function of the initial kinetic energy, the angular dependence of which provides an opportunity to set out a binary relation between the initial kinetic energy and the average deflection angle, thereby leading to a coarse energy prediction founded on the mean deflection angle in the case of experimental incapabilities or limitations. In this study, we address the problem of the muon energy classification for a tomographic system consisting of 0.4-cm plastic scintillators manufactured from polyvinyl toluene and we explore a four-group structure besides a ternary partitioning between 0.25 and 8 GeV. In the first instance, we determine the deflection angles by tracking the hit locations in the detector layers on the sub-divided uniform energy intervals through the GEANT4 simulations. In the latter step, we express two misclassification probabilities where the first approach assumes a symmetrical linear propagation bounded by one standard deviation in one dimension, whereas the second procedure employs a positively defined modified Gaussian distribution that governs the overlapping area in two dimensions. In the final stage, we compare qualitatively and quantitatively the adjacent energy groups by using the computed misclassification probabilities. In the absence of any further data manipulation, we explicitly show that the misclassification probabilities increase when the number of energy groups augments. Furthermore, we also conclude that it is feasible to benefit from the mean deflection angle to roughly estimate the muon energies up to four energy groups by taking the misclassification probabilities into consideration, while the classification viability significantly diminishes when the partition number exceeds four on the basis of standard deviation.