Problem 2466

Summary: degraded accuracy of the energy loss fluctuation (G4UniversalFluctuation model) in Geant4-11.0
Product: Geant4 Reporter: Mihaly Novak <mihaly.novak>
Component: processes/electromagnetic/standardAssignee: Vladimir.Ivantchenko
Status: RESOLVED FIXED    
Severity: minor    
Priority: P4    
Version: 11.0   
Hardware: All   
OS: All   
Attachments: demonstration of the problem (generated by using TestEm5): energy loss distribution of 100 MeV electrons in 5.6 [um] Silicon target
updated plot including the current master result

Description Mihaly Novak 2022-01-31 15:05:09 CET 
Created attachment 748 [details]
demonstration of the problem (generated by using TestEm5): energy loss distribution of 100 MeV electrons in 5.6 [um] Silicon target

Using Geant4-11.0 for modelling energy loss distributions in thin target the result found to be less accurate than the earlier versions. Anna Zaborowska drew my attention to this issue that pointed to the modifications 
done in the G4UniversalFluctuation model in version 11.0. 

The problem can be reproduced easily by using TestEm5 for simulating the energy loss distribution in 5.6 [um] Silicon target by 100 [MeV] electrons. As it can be seen in the attached figure, Geant4-10.7 gives a nice agreement with the available experimental data while Geant4-11.0 can reproduce this results only when rolling back the G4UniversalFluctuation model to that of version 10.7. 

There is an open MR that (https://gitlab.cern.ch/geant4/geant4-dev/-/merge_requests/2524)that already deals with this issue.
Comment 1 Vladimir.Ivantchenko 2022-02-22 09:51:04 CET 
Hi Mihaly,

can you, please, check the current master? If this test will not improve situation significantly, then we will change configuration for Opt4 EM physics in a way that the old fluctuation model is used.

Vladimir
Comment 2 Mihaly Novak 2022-02-25 11:50:46 CET 
Created attachment 751 [details]
updated plot including the current master result

Hi Vladimir, 

You can find the plot attached updated now with the results obtained by using the current master (as of 25th February 2022). 

Cheers,
Mihaly
Comment 3 Vladimir.Ivantchenko 2022-02-25 12:33:45 CET 
Hi, Mihaly,

Thanks! So, the improvement is significant but still the 10.7 model is better. So, let us include 10.7 variant of the model into Opt4 EM physics for e+-, mu+-, pi+-, p, pbar.

Vladimir