G4GNXS photonuclear cross section for ¹⁸⁶W

Physics Processes, Models and Cross Sections
1

Dear Geant4 developers and community,

While simulating laser-driven neutron production via LWFA electrons on a tungsten target, I noticed a large discrepancy in neutron yield depending on which photonuclear cross-section model is used. This was already flagged by Khue et al. (arXiv:2405.05996) for deformed nuclei including ¹⁸⁶W, but the origin of the discrepancy was left as an open question. I have traced the data chain and would like to share my findings and ask for clarification.

Using QGSP_BIC_HP with Geant4 11.2.2 on a W target (same electron input spectrum), switching between G4GNXS and G4PNXS changes the total neutron yield by a factor ~3. The G4PNXS result is consistent with an independent SMILEI+FLUKA simulation.

I compared three levels of the data pipeline at the GDR peak (~13–15 MeV):

Source σ peak (mb) vs Berman 1969
Berman et al. 1969 (experiment, ¹⁸⁶W) ~170 reference
w186_jendl.dat (IAEA/PD-1999, MT=5, raw) ~430 ×2.5
G4PARTICLEXS4.0 / inel74_186 (×10²⁵ mm²→mb) ~997 ×5.9
G4GNXS simulation output (W natural abundance) ~512 ×3.0

There is an additional ~×2.3 factor between the JENDL source data and the values stored in G4PARTICLEXS4.0/gamma/inel74_186.

My hypothesis is that this could originate from a double-counting of reaction channels during the ENDF-6 → Geant4 conversion (e.g. summing MT=5 and MT=201, or incorrect interpolation law handling). However, I am not familiar enough with the conversion pipeline to be certain.

Could the developers clarify: is this additional factor expected (e.g. a deliberate normalization), or does it suggest a bug in the data conversion for this isotope?

I am happy to share my comparison scripts and data files if useful. Thank you in advance for any clarification.

Best regards,
Eya Dammak (CEA-LIDYL / EPACE doctoral network, France)

2

Hello @EyaDammak ,

this is indeed a very clear report. Database handling and manipulation is always difficult. I think @civanch might be interested in having a look at this.

/Pico

3

Dear @EyaDammak ,

It seems that this issue was already reported in bugzilla

From Vladimir’s reply there, my understanding is that several targets were checked, that for several elements the tabulated cross sections were replaced by CHIPS, and that the problem may be related to the IAEA database (?)

This also seems consistent with the Geant4 11.3 release notes:

“G4GammaNuclearXS: Use CHIPS parameterisation for Zr and W at all energies, for Cr and Y above 25 MeV and for Sn and Gd above 16 MeV. … Addressing problem report #2609.”

4

Thank you very much for your response :folded_hands: