We observed a problem concerning branching ratios in some radioactive decays. They were ok with the Geant4 7.0, but they changed with Geant4 8.2.p01. The same version of the databases was used, so it must be something in the RadioactiveDecay code which was changed from 7.0 to 8.2.p01. The primary particle is a Ba-133 nucleus at rest, and Geant4 RDM takes care of its decay. The branching ratios are ok with Geant4 7.0 (usually within a few %) but wrong with Geant4 8.2.p01 for some lines. The effect is relevant for low-energy lines (<100 keV) while no difference is seen at higher energy. For instance, this is the comparison with the Table of Isotopes intensities: 79.6 keV ToI: 2.62% G4 7.0: 2.53% G4 8.2: 2.96% 81.99 keV Toi:34.06% G4 7.0:33.71% G4 8.2:42.79% (!) 302 keV ToI:18.33% G4 7.0:18.01% G4 8.2:18.37% 356 keV ToI:62.05% G4 7.0:59.96% G4 8.2:59.86% K x-rays ToI:121.4% G4 7.0:120.1% G4 8.2: 85.6% (!) K x-rays are around 30 keV. The very same effect has been observed for low-energy gamma lines of Am-241. Some diagnostics: the values are obtained with the SAME source code and with the SAME databases (in particular G4RADIOACTIVEDATA, G4LEVELGAMMADATA and G4LEDATA). Versions 3.0 and 3.1 of G4RADIOACTIVEDATA have been tested. The differences between the two runs are: 1) G4INSTALL points to geant4 8.2.p01 and to geant4 7.0 2) NeutronHPCrossSections database is 3.10 and 3.7, respectively 3) The CLHEP version is 1.9.2.2 and 1.8.1.0, respectively I cannot see how NeutronHPCrossSections and CLHEP could affect the gamma-ray yield after a nucleus decay. Furthermore, the effect does not depend on the cuts. Luciano Pandola, Dusan Budjas
The assignment of this problem report to the lowenergy electromagnetic package is incorrect: in fact, the text mentions the relation to RadioactiveDecay changes between releases 7.0 to 8.2.p01 as the source of appearance of this problem. Maria Grazia Pia
I think I traced back the problem to line #279 of processes/hadronic/models/de_excitation/photon_evaporation/src/G4VGammaDeexcitation.cc, which was changed between the versions 7.2 and 8.0: if ( _vSN != -1) (dynamic_cast <G4DiscreteGammaTransition*> (_transition))->SetICM(false); This line prevents more than one electron conversion in the same de-excitation cascade. This yields to the over-estimation of the branching of some gammas and to the under-estimation of fluorescence lines. This is visible in those nuclei (as Ba-133) in whose cascades the probability of having more than one electron conversion is not negligible. I don't know the reason why this restriction had been implemented in the past (maybe for fixing some other problems). If line #279 of G4VGammaDeexcitation.cc is commented, branching ratios for gamma rays emitted in Ba-133 decays are consistent with version 7.0 and with the Table of Isotopes. I don't know if commenting the line has some unwanted side-effects in other parts of the code. Luciano Pandola - INFN Gran Sasso
The changed proposed below fixes the problem for the gamma-ray branchings (i.e. the line at 81 keV). The branching ratios of K x-rays are still different with respect to Geant4 7.0.
The bug has been fixed (see patch described below) in version 9.1.
