Created attachment 102 [details] Spectrum obtained for 227Ac decay (see text) Dear all, We are interested in the simulation of radioactive decays of isotopes inside a detector, that is, all the emissions generated are relevant for us. We encountered problems in the past using the Radioactive Decay Model mainly due to the missing energy in beta- and EC emitters, as reported in problem report #1001( http://bugzilla-geant4.kek.jp/show_bug.cgi?id=1001). When using the last available version 4.9.4.p01, many of these problems have been solved, although some remain and also some new issue has appeared. Our particular doubts, using the last version, are the following: 1) From the sequence of isotopes in the natural radioactive chains, there are still problems for 228Ra (from 232Th chain) and 227Ac (from 235U chain), both beta- emitters with Q around 45 keV. In an ideal detector registering all the emissions, we expect a superposition of several beta spectra starting at the energies of the excited levels of daughters, but this expectation is not acomplished. Beta shapes start in some cases below the level energies, so it seems that there is still some missing energy (I attach the spectrum obtained for 227Ac). 2) Concerning the electron energy spectrum for a beta- decay, I would like to know if as stated in the forum post (http://hypernews.slac.stanford.edu/HyperNews/geant4/get/hadronprocess/766.html) and as it seems having a glance at the code in G4RadioactiveDecay.cc, LoadDecayTable(), case BetaMinus:, the considered shape is still always that of allowed transitions, and if the correction implemented in version 4.9.4 at G4RadioactiveDecay.cc suggested in the problem report #1068 (http://bugzilla-geant4.kek.jp/show_bug.cgi?id=1068) is exclusively for 40K isotope. Our problem here is the following. The beta spectrum shape obtained for 40K with previous version (9.3.p02) agreed very well with the spectral information obtained from the book "Table of Isotopes" (Wiley, 1999) coming apparently from calculation having into account the order of forbiddenness, while after the specific correction for 40K implemented in the new version (9.4.p01) the agreement is much worse. The following table shows the comparisons: energy range (keV) % Simulation (geant4.9.3.p02) % Tables % Simulation (geant4.9.4.p01) 0 a 10 0.49 0.495 0.403 10 a 20 0.51 0.51 0.41 20 a 40 1.1 1.1 0.87 40 a 100 3.9 3.9 3.02 100 a 300 18.0 17.9 13.8 300 a 600 32.0 32 27.4 600 a 1300 33.3 33.4 43.4 >1300 0.001 0.000326 0.0016 <E> 453.9 455 525.8 Mean energy <E> is calculated over 40K decays (not over electrons emitted in beta decays). We have checked that for other isotope (210Bi, first-forbidden transition) the agreement of the simulated beta spectrum with the information on Tables is also very good with the previous version. Are we misunderstanding anythig? 3) It seems that the decay of 234Th (from 238U chain) is not properly reproduced in version 4.9.4.p01. * According to tables of isotopes, the beta decay ends almost always to the metastable state of 234Pa at 73.92 keV, either directly (70.3%, according to /RadioactiveDecay3.3/z90.a234 file) or through excited states with higher energy. The 234mPa decays beta- (mainly to the ground state of 234U, 98.2% from /RadioactiveDecay3.3/z91.a234 file) having a negligible probability of IT to the ground state of 234Pa (0.16%). * From the simulation, however, I encounter that in many cases (roughly, 70%) 234Pa[0.0], that is the ground state, is produced. Its beta- decay is very different from that of the metastable state, giving rise to the emission of a lot of gamma lines from 234U deexcitation. This problem was not present in previous version (I checked 9.3.p02) where 234Pa[73.9] state was correctly produced. Are you aware of this problem? Can be avoided in some way? Many thanks for your help in any of these questions, Susana Cebrian PS: I posted this message at Hypernews http://hypernews.slac.stanford.edu/HyperNews/geant4/get/hadronprocess/1147.html but following recommendations, I have created this problem report.
Created attachment 103 [details] Beta minus spectrum from Ac227 The plot compares the e- spectrum (probability density function) of the Beta minus decay of Ac227 as obatined from Geant4 simulation and as given by the FERMI theory of beta decays as implemented in Geant4. The theoretical curve are also reproduced for the different decay channel. An excellent agreement is observed between the theory and the simulations.
Dear Susana, I have work on some of the problems that you refer to in your bug report. Below are my "at the moment" answers to some of your points. 1)On the beta minus decay of 227Ac I think that your expectation of how the spectrum should look like is not correct. Or may be I do not understand you point? Top my knowledge the different beta - spectrum should start at 0 energy and go to Qi. Where Qi represents the Q of the different decay channels and is given by Qi=~45keV-E_exi where E_exi is the energy of the excited levels of the daughter nuclei. After that the excited elevle can decay either by gamma deexittaion or by internal conversion. In Geant4 the Fermi beta decay theory is used to sample the e- energy with an a approximated Coulomb-Fermi Factor. I have done a program that plots the e- energy spectrum that should be expected from the the 227Ac decay (following the theoretical formula used in G4 ) and compared that to the output of a G4 simulation of the 227Ac decay. The results is presented in the attachment "Beta Minus Spectrum from Ac227" Agreement between the theory and the G4 computation is excellent. So at the moment I do not see a bug there. Would you expect another e- spectrum? 2) 40K isotope. You are correct the special spectrum is applied only to 40K isotope following the request of the people who do write the bug 1068. Could you please put a picture of the expected spectrum such that I check the G4 output against experiment? 3) decay of 234Th I will look to that! Best regards Laurent
(In reply to comment #2) Dear Laurent, thank you very much for your answer. 1) For Ac227 (and the same for 228Ra) our problem is not with the electron spectra, which you showed that are correctly simulated, but when detecting the whole energy released in the decay, that is, detecting also the emissions (gamma rays or internal conversion electrons) following the nuclear deexcitation of the daughter. The Ac227 spectrum I attached, and that we think it is not as it should, corresponds to the total energy absorbed in an ideal detector which contains also the Ac227 nuclei (so it must be the total energy released); for this reason, I would expect to register the beta spectra shifted the energy of the excited levels of the daughter, which is not always the case. We think that there is still some missing energy in the deexcitation process. 2) For K40 I have attached a pdf file and the corresponding txt files, with the energy spectra registered for 40K decays again inside the ideal detector, using versions 9.3.p02 y 9.4.p01. In the past days I have directly been in contact with Prof. Taiuti, who proposed the correction in the beta spectrum for 40K implemented in version 9.4 (bug report #1068) and I have understood that the correction is just the shape factor corresponding to a third forbidden transition. Our problem here is just that, as I reported, the agreement with the beta spectrum information from the Table of Isotopes was very good for simulation with version 9.3 and has worsened with version 9.4. So if you can compare the simulated beta spectra with experimental data would be perfect. I understand that for any other isotope, the simulated beta spectrum is always that of an allowed transition, is it OK? The three problems I presented in my bug report are all dealing with the Radioactive Decay Model in Geant4.9.4, but I consider that are quite independent. I hope they are clearer now. Thanks a lot again for your help! Susana
Created attachment 104 [details] Spectra obtained for 40K
Dear Laurent, have you had time to have a look at the reported problems when using the G4RadioactiveModel, specially the one dealing with the simulation of the decay of 234Th using version 9.4.p01? Thank you very much, best regards, Susana
Dear Susana, I will look at the 234Th issue when I am back from my holidays on 19 June. It will be my first task. Sorry for the delay. Laurent
Dear Suzanna, Considering your point c). The modeling of the 234Th decay is indeed wrong by default in G4.9.4. The problem is that after the beta minus decay of 234Th to 234Pa_73.92keV, the 234Pa_73.92keV is sent directly to the G4PhotoEvaporation process in order to process a gamma/IC_conversion decay without considering the Beta minus channels. To consider the beta minus channels correctly the metastable level 234Pa_73.92keV has to be produced has a track in G4. This can be done at the moment by setting a negative half life energy treshold in grdm. You can do that by using the macro command /grdm/hlThreshold -1.0000e+00 s This should solve your problem at the moment. But anyway we need some recoding in Geant4 to avoid this bug to occur by default. Please let me know if by using the macro commands specified above you get finally the expected decay results. With my best regards Laurent
Dear Laurent, I have checked the decay of Th234 using the proposed solution and I think it works fine; what I get in an ideal detector registering all the enegy is what I expect: - for Th234 decay, three beta spectra corresponding to transitions to the metastable level of 234Pa at 74 keV and to other levels with higher energies +92 keV and +114 keV. - for Pa234m decay, a beta spectrum and a peak with the correct area at 74 keV corresponding to transition to the ground level of Pa234. I attach these spectra. In summary, this problem is solved. In any case, I would like to further understand what the solution does. I have read that hlThreshold "sets hl threshold for isomer production"; being negative means that isomer is not produced? I guess that this correction can be dangerous for the proper simulation of other metastable states... is it right? From my first bug report I think that the only pending question is (1), the problems observed for Ra228 and Ac227 decays (which were not as dramatic as that of Th234...). I really thank you very much for your help. Best regards, Susana
Created attachment 123 [details] Spectrum of registered energy for 234Th decay after the proposed solution in comment #7
Created attachment 124 [details] Spectrum of registered energy for 234mPa decay after the proposed solution in comment #7
Dear Susana, The hlTreshold in Geant4 is by default set to 1e-6 s. This limits represents the minimum half life that and exited state should have to be produced as a track in Geant4. On the contrary the G4RDM passes an excited state directly to G4PhotoEvaporation for gamma/EC deexcitation. Setting hlTreshold to a negative value will make that all excited level will be produced as G4Track. The problem that occurs in your case is that Geant4 considers 234Pa_73.92keV as having T1/2=0s while it should be ~1.159m. In fact there are two 234Pa_73.92keV levels 3+ and 0-. The 3+ state has T1/2=0s but not the 0- that has T1/2=1.159 m It seems that G4PhotonEvaporation consider all 234Pa_73.92keV levels has having 0s hal-flife and that no differences is done between 3+ and 0-. So you see some corrections is needed in the code, to avoid that the same problems occurs with other isomers. But as you have seen to set T1/2<0 is sufficient to solve your problem at the moment. I still need to re-check your issue 1). I'll do that next week. Best regards Laurent
Dear Laurent, I realized by chance that the problem noticed for Ra228 decay had disappeared, and I checked that the origin of the solution was in fact the correction you gave us for the Th234 problem. Exactly the same happens for Ac227, so I think that now all the problems initially reported in the bug report have been solved. So, just to thank you again for your help. Best regards, Susana
