Ge Dedektörünün Zamana Bağlı Monte Carlo Yöntemi ve Genetik Algoritma ile Gerçekçi Modellenmesi

Abstract

Simulation of radiation detectors using the Monte Carlo method is extensively utilized for various purposes such as obtaining detector efficiencies for primary activity measurements, transferring values obtained from reference measurements to different counting geometries and gamma energies, and performing certain counting corrections in advanced gamma spectrometry. There are numerous studies in the literature on these topics. During periods when the production of large-volume high-purity germanium detectors was limited, these counting corrections could be neglected or easily reduced to a negligible level. However, in recent years, with the increase in the volumes of produced detectors, these counting corrections have become more significant. There are many articles and books available on one of the correction factors, which is the true coincidence gamma correction. On the other hand, there are very limited studies in the literature regarding the random-coincidence gamma correction, and almost all of them include corrections based on experimental results rather than simulations. An important issue arising from the increase in the detector volume is the difficulty in reproducing experimental results obtained with these detectors through simulation studies. In this study, firstly, a method is proposed for performing corrections for random summing effects in gamma spectrometry using time-dependent Monte Carlo simulation. In this method, the MCNPX code, commonly used for simulations, is iteratively run for short time intervals through an interface program, and the results are combined to make the simulation time-dependent. The time parameters used in the interface program are obtained experimentally, primarily by evaluating signals generated by a random pulse generator. Using the time-dependent simulation, the correction factor for chance coincidence was calculated for a 137Cs point source having 370 kBq activity, and the experimental comparison showed a relative error of 2%. The results of simulations performed with paralyzable and nonparalyzable models are also discussed in this part of the study.

In the rest of the study, it is proposed to define an equivalent geometry for problems in reproducing experimental results in the simulations of large-volume high-purity germanium detectors. For this purpose, a reference mixed-point gamma source emitting 12 gamma rays with high emission probabilities between 59.5-1836.1 keV was placed around the detector at 74 different positions. A genetic algorithm was used to reach an equivalent geometry that can reproduce the experimental efficiency results. The results indicated that smaller dimensions than the actual sizes should be used in the modeling for such large-volume detectors.

In both parts of the study, a method that uses the probability density and cumulative distribution functions obtained from the MCNPX code was proposed for correcting random and true summing effects of coincident gamma rays.