Quantification of Ra-226 and U-235 in soil using HPGe gamma spectra and linear regression

Abstract

This study aims to develop a linear regression model using the Scikit-Learn libraries for direct quantification of Ra-226 and U-235 in soil using HPGe gamma spectrometers. The proposed approach eliminates the need for radioactive equilibrium with Rn-222 and its decay products, as well as bypasses conventional spectral correction procedures. The model was trained on a dataset comprising 20,000 synthetic 8192-channel gamma spectra and achieved quantification accuracies within 15% for both radionuclides. The findings highlight the model’s simplicity and reliability, underscoring its potential as a rapid and practical tool for routine environmental radioactivity assessment, particularly for overcoming challenges posed by spectral overlap and equilibrium constraints.