Integrated Experimental–GEANT4–ROOT Determination and Discrepancy Analysis of the Half-Lives of Cs-137, Sr-90 and Eu-152

Reliable half-life determination requires more than fitting an exponential curve: the result depends on source characterization, detector response, background correction, calibration, statistical estimation, and the reference nuclear-data set used for comparison. This study develops an integrated experimental–computational framework for determining and evaluating the half-lives of Caesium-137 (Cs-137), Strontium-90 (Sr-90) and Europium-152 (Eu-152), three radionuclides selected to represent contrasting radiation-measurement problems. Cs-137 was measured by background-corrected radiation counting, Sr-90 by Geiger-Müller beta counting with explicit consideration of the Sr-90/Y-90 parent-daughter system, and Eu-152 by high-purity germanium (HPGe) gamma-ray spectrometry. GEANT4 11.3.2 was used to model radioactive decay and radiation transport, while ROOT was used for Gaussian photopeak analysis, nonlinear exponential fitting, residual analysis, and goodness-of-fit evaluation. The measured/simulated half-lives were compared with adopted evaluated values using absolute difference, percentage deviation, and uncertainty-based standardized differences. Cs-137 gave 29.96 ± 0.12 years experimentally and 30.045 ± 0.020 years from GEANT4, against 30.018 ± 0.022 years as the adopted reference. Sr-90 gave 28.98 years experimentally and 27.95 ± 1.35 years by GEANT4, against 28.80 ± 0.07 years. For Eu-152, the experimental and GEANT4 estimates were 14.83 ± 0.21 and 14.43 ± 0.22 years, respectively. Using the consistently documented Eu-152 reference of 13.517 ± 0.009 years, the corresponding deviations are 9.71% and 6.75%. The results demonstrate that the integrated workflow can reproduce expected decay behaviour and detector response while also revealing radionuclide-specific systematic sensitivities. The largest residual discrepancy occurs for Eu-152, where HPGe efficiency, source geometry, self-attenuation, coincidence summing, and calibration assumptions require further control. The study therefore supports uncertainty-aware integration of measurement, Monte Carlo transport, and statistical analysis for nuclear metrology and radiation-detection research.