Problem 821

Summary:	Missing inelastic recoils from NeutronHP
Product:	Geant4	Reporter:	hans-herbert.fischer
Component:	processes/hadronic/models	Assignee:	tkoi
Status:	RESOLVED FIXED
Severity:	normal	CC:	joa
Priority:	P2
Version:	7.1
Hardware:	PC
OS:	Windows

Description hans-herbert.fischer 2005-12-12 11:06:07 CET

We consider neutron interaction with targetmaterial germanium, particularly the
non-ionising energy loss (NIEL) of the neutrons (which is dependent on the
number and energy of the generated recoils). We notice discrepancies in the
statistics of inelastic recoils (residual nuclei) when compared to expected
values and when switching from neutron_HP to binary cascade at 20 MeV. These
discrepancies persist after deleting the Ge-isotope data files in
"G4NDL3.7/Inelastic/CrossSection" (workaround for bug #526 to prevent the usage
of gallium-data). After using a modified "G4NeutronHPInelasticComp.cc", which
should take care of the conservation of linear momentum (bug #675) and has been
kindly provided by Joa Ljungvall, the recoil energy is closer to what is
expected, but with increasing neutron energy (< 20 MeV) our NIEL results get
worse. It turned out that there is a correlation with the contribution to final
states from Ge(n,2n)Ge reactions, which -according to the data in "/F04"
compared to "/F01"- increases with increasing neutron energy. It seems that no
residuals are generated for Ge(n,2n) reactions. This is probably not the
intended behaviour of neutron_hp.

PS: We initially suspected a mix-up or corruption of cross-sections due to a
faulty file reading or invalid memory pointers and coarsely checked the file
reading and the memory usage. Indeed we observe that in
G4NeutronHPPartial::InitData() the allocated data space is exceeded when
neg>nData.

Comment 1 tkoi 2006-07-24 11:44:59 CEST

Thank you for reporting problem.
This is under invetigation.
If you have any updated information, please let me know.

Tatsumi

Comment 2 tkoi 2008-07-01 19:58:14 CEST

This problem is identified and a temporal fix will be released soon.
However the fix is not guaranteed conservation laws at single event level.

Tatsumi

Comment 3 tkoi 2010-04-23 15:55:45 CEST

The recoil (residual) nucleus should exist in final state.
Energy and momentum conservation is not guaranteed at single event level.
This is a kind of model feature. However if you have very big difference in binary reaction (ex elastic, nA-pB and so on), please let me know. 

I changed status of this entry to “RESOLVED-FIXED”.

Tatsumi