I get a werid gamma spectrum by 12C(p,gamma)12C’ reaction in the upper picture at 4MeV, it should be smooth simular as the next one. My model is quite simple: target is solid carbon and proton energy is 150MeV, gamma spectra is scored by USRBDX.
Is there any error in cross section in FLUKA?
Find attached below the gamma spectrum resulting from 150-MeV protons interacting with 12C in FLUKA.
What you should see with USRBDX is essentially the same, except that some of the photons may interact still in the target, so some mild spectral distortion is to be expected, depending on your geometry.
There is a priori no evidence for the artifact that you report. In order to figure out its origin, please upload your input file (and if applicable any user routine you might have used).
Thanks in advance,
Dear Francesc Salvat Pujol:
Here is the inp and also test different score card. shows same problem. My Fluka version is 4-0.0
P-cabon-spectrum.inp (1.6 KB)
Thanks for your feedback.
After further investigation the origin of the step is clear. In a
nutshell (details below), you are collecting photons not only from
12C(p,xg), but also from 12C(n,xg): the latter exhibits a a signature
of FLUKA’s multigroup treatment of low-energy neutron interactions.
Bombarding 12C with protons you will certainly produce neutrons, which
will be transported and may further interact in your target (depending on its size).
FLUKA’s multigroup treatment of low-energy neutrons (below 20 MeV) involves
42 groups for gamma emission, which you can consult in section 10.4.1 of the
manual. Gamma group number 14 happens to span the interval 4.0 to 4.5 MeV.
The 12C(n,xg) spectrum has a prominently sharp and intense peak at ~4.4
MeV (corresponding to the 1st excited level of 12C). In FLUKA’s
multigroup treatment this narrow peak is instead modelled as an intense
“bin” spanning gamma energies from 4.0 to 4.5 MeV. Thus, the contribution
from the 12C(n,xg) spectrum is “coarse-grained” such that the sharp peak
that should be at ~4.4 MeV is instead spread from 4.0 to 4.5 MeV, preventing
the user to adopt a finer resolution. This explains the step you witness at 4 MeV.
In practical terms, if neutron reinteraction is not an issue in your
problem, i.e. if your geometry is thin or if you are just interested in
12C(p,xg), you can either put a thin target (thin compared to the typical
nuclear inelastic scattering lengths of the order of 10s of cm) or discard
neutrons altogether (see the DISCARD card).
With kind regards,
@cesc I see. Here is the plot with neutron discard, seems good but less clinical reality in proton therapy.
Can FLUKA have a better handle here without a jump in code? or can the user do some postprocessing to smooth the sepectrum at 4 MeV?
Thanks a lot
If yours is a proton therapy problem, where neutron re-interaction plays a role, then you cannot discard neutrons. By properly coding the USDRAW entry of the mgdraw.f user routine, to be activated by the USERDUMP card (see the manual), you can identify and count the gammas contributing to the 4 MeV peak as produced by low energy neutrons (ICODE=300). This way, a posteriori you can correct the spectrum by removing their flat contribution over the 4.0 to 4.5 MeV interval and replacing it by a better resolution shape.
@ceruttif thanks for the suggestion, it’s a good idea.