Abstract
Correction of the chemical composition effect for soil samples in gamma spectrometry is very important to perform at low emission energy (<100 keV). This study presents an intensive Monte Carlo simulation using Geant4 code of CERN which investigates several items in this subject. First, it is confirmed that for soil samples, in presence of chemical composition effect, correction of the density effect at low energy is not sufficient to produce an accurate efficiency calibration of the detector. Chemical composition effect must also be corrected. Second, it was found that the slope of the linear fit to the decrease of the efficiency as a function of % of average chemical composition for a given emission energy is density independent. This is a new result which could be used to develop new efficiency correction methods against chemical composition effect. Finally, the usual method of correcting the efficiency using the simple ratio between the self absorption factor of the gamma ray inside the sample and inside the calibration standard, was found unable to correct the chemical composition effect for soil samples.