| Summary: | Error in decay channel for neutrons from C-N-O ion primaries | ||
|---|---|---|---|
| Product: | Geant4 | Reporter: | David <david.francois.droz> |
| Component: | processes/hadronic/models/radioactive_decay | Assignee: | Gunter.Folger |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | Vladimir.Ivantchenko |
| Priority: | P4 | ||
| Version: | 10.5 | ||
| Hardware: | All | ||
| OS: | All | ||
This is not a radioactive decay problem, but rather the general particle decay class trying to decay a neutron which is off the mass shell. Such particles should not reach the tracking stage in Geant4, so my guess is that this is a problem with the Binary cascade that produced the neutron. I'm assigning your report to Gunter Folger, our Binary cascade expert. Could you please provide more information to re-produce this problem. Like, which physics list are you using, what is the target material; how do you define mat_Battery? Hi Gunter, we addressed this problem for 10.6beta. The problem is rare, is seen in the Binary cascade in ion-ion interaction, when a neutron is produced out of mass shell. The fix is inside hadronic interface, it is not simple but potentially we can back port it to 10.5 in the next patch if the patch will be provided. Vladimir Dear Gunter, Vladimir, Thank you for your comments. We tried Geant4.10.06beta and still observed the problem on two independent installations/teams. The problem happened on both QGSP_FTFP_BERT and FTFP_BERT physics list. Here is the target material and mat_Battery: <material name="mat_Battery" state="solid"> <T unit="K" value="293.15"/> <MEE unit="eV" value="144.881926987823"/> <D unit="g/cm3" value="2.165"/> <fraction n="0.19518445618745" ref="Carbon"/> <fraction n="0.398" ref="Aluminium"/> <fraction n="0.02" ref="Titanium"/> <fraction n="0.084" ref="Copper"/> <fraction n="0.0170098229419813" ref="Lithium"/> <fraction n="0.144570016541753" ref="Cobalt"/> <fraction n="0.134206611504321" ref="Oxygen"/> <fraction n="0.00702909282449493" ref="Hydrogen"/> </material> Hello, I have reproduce the problem and provide the fix, which will be available with the next public release of Geant4. For 10.6beta I may provide a private receipt how to remove this problem. This fix is done on level of hadronic interface allowing to keep running. Independently on that the model(s) may be improved. VI |
First, the error message: G4DecayTable::SelectADecayChannel :: no possible DecayChannel neutron -------- EEEE ------- G4Exception-START -------- EEEE ------- *** G4Exception : DECAY003 issued by : G4Decay::DoIt Can not determine decay channel for neutron mass of dynamic particle: 0.933405 (GEV) dacay table has 1 entries 0: BR 1, IsOK? 0, --> e- + anti_nu_e + proton *** Fatal Exception *** core dump *** G4Track (0x55d8264583c0) - track ID = 575, parent ID = 1 Particle type : neutron - creator process : ionInelastic, creator model : Binary Cascade Kinetic energy : 0 eV - Momentum direction : (1,0,0) Step length : 0 fm - total energy deposit : 0 eV Pre-step point : (-649.462,199.77,571.331) - Physical volume : BattaryA_physical (mat_Battery) - defined by : not available Post-step point : (-649.462,199.77,571.331) - Physical volume : BattaryA_physical (mat_Battery) - defined by : not available *** Note: Step information might not be properly updated. -------- EEEE -------- G4Exception-END --------- EEEE ------- We are using Geant4.10.05p01. This exception arises randomly on some events when the primary particle is Carbon-12, Carbon-13, Nitrogen-14, Nitrogen-15, Oxygen-16, and potentially others (untested). The physics list is FTFP_BERT, the same problem was observed with QGSP and with NeutronHP. We observed this crash on Geant4.10.05p01 and Geant4.10.06 beta The crash probability is very high in the energy range of 10-100GeV. (E.g. 40% failure rate for Nitrogen-14 at 10k events per file) but no such crashes at energies above 100 GeV.