An intercomparison of Monte Carlo codes used for in-situ gamma-ray spectrometry

In-situ gamma-ray spectrometry is widely used for monitoring of natural as well as man-made radionuclides and corresponding gamma fields in the environment or working places. It finds effective application in the operational and accidental monitoring of nuclear facilities and their vicinity, waste depositories, radioactive contamination measurements and environmental mapping or geological prospecting. In order to determine accurate radionuclide concentrations in these research fields, Monte Carlo codes have recently been used to obtain the efficiency calibration of in-situ gamma-ray detectors. This work presents an inter-comparison between two Monte Carlo codes applied to in-situ gamma-ray spectrometry. On the commercial market, Canberra has its LABSOCS/ISOCS software which is relatively inexpensive. The ISOCS mathematical efficiency calibration software uses a combination of Monte Carlo calculations and discrete ordinate attenuation computations. Efficiencies can be generated in a few minutes in the field and can be modified easily if needed. However, it has been reported in the literature that ISOCS computation method is accurate on average only within 5%, and additionally in order to use LABSOCS/ISOCS it is necessary a previous characterization of the detector by Canberra, which is an expensive process. On the other hand, the multipurpose and open source GEANT4 takes significant computer time and presents a non-friendly but powerful toolkit, independent of the manufacturer of the detector. Different experimental measurements of calibrated sources were performed with a Canberra portable HPGe detector and compared to the results obtained using both Monte Carlo codes. Furthermore, a variety of efficiency calibrations for different radioactive source distributions were calculated and tested, like plane shapes or containers filled with different materials such as soil, water, etc. LabSOCS simulated efficiencies for medium and high energies were given within an accuracy of below 5% for point sources and 10% maximum accuracy for more complicated geometries. Accuracy for LabSOCS decreased for low energy emissions. Finally an analysis and discussion of the results and the main differences found between LabSOCS and GEANT4 were performed.