I am trying to simulate a spectrum of X-rays produced by bremsstrahlung of primary protons injected into a thick-target. A very simple plane-parallel geometry is considered in the simulation, which consists of a cubic box centered at the origin of a Cartesian coordinate system (Ox,Oy,Oz) with faces perpendicular to the coordinate axes. The z-coordinate corresponds to the vertical depth in the target. A xy-plane at z = 0 divides the cubic box into two half-spaces. The half-space at z < 0 is just filled with vacuum. The half-space at z > 0 is filled with atomic Hydrogen. The primary protons are injected into the z >0 region from a point located in the z <0 at a distance very close to the xy-plane at z = 0, and are assumed to have a downward isotropic angular distribution and a power-law energy distribution of spectral index delta = 4 in the range from 1 MeV to 1 GeV. Two USRBDX detectors are used to score the downward and upward current of photons escaping the target.
I am using the PAIRBREM card to set a 10 keV photon energy threshold for explicit bremsstrahlung production and several cards to suppress the contributions from processes other than the bremsstrahlung of protons. However, no photons are produced in the simulation. I would be grateful if you could check the setup of the simulation and for any help to understand what is causing this issue.
Attached are the project and input files and the routine source.f used to generate the energy and angular distributions of primary protons.
This is an interesting case you are studying! I am afraid though that I actually do get some photons using your input - however by simulating many more than you attempted in the first place. I believe there is nothing strictly wrong with your input, it is simply the problem you are studying which is very difficult to obtain decent statistics from, due to the combination of the low-probability proton bremsstrahlung process itself with a very low density target (I could imagine that if you run the same input with a higher density, higher Z target material you will get more photons produced). Unfortunately in FLUKA I believe there is no way to bias the interaction producing bremsstrahlung photons to achieve better statistics, such as is e.g. the case for proton inelastic interactions with LAM-BIAS.
Have you tried taking more time to run orders of magnitude more particles? If you have the possibility to parallelize your runs I would highly recommend that. I also don’t fully see why you would need such a large simulation volume, I suppose there is no benefit from tracking other secondary particles which could still be generated in the large simulation volume.
Thank you for the help. I will run the simulation with more primary protons as you suggested (how many did you use?).
" I also don’t fully see why you would need such a large simulation volume…"
You are right. The CSDA range for 1 GeV protons in a hydrogen target is ~150 g/cm^2. The target density in the simulation is 3.19x10^-5 g/cm^3, so in order to have a thick-target I could set the size of the box to ~5x10^6 cm.
Thank you. I ran the simulation with 10^8 primary protons and indeed got just a few photons produced. So, in order to get a statistically reliable proton bremsstrahlung spectrum, I would have to run the simulation with a much larger number of primary protons. As you pointed out, it would be necessary to parallelize the run, but I have never done that.
As you already remarked in the linked post on Radiative processes included when using PAIRBREM, the photons you get in this problem are not originating from proton bremsstrahlung, but are rather secondaries from inelastic interactions, bremsstrahlung from delta-rays, … . A dedicated mgdraw.f routine (solliciting the usdraw subroutine) can help you clarify exactly how many pure proton bremsstrahlung photons you will obtain, if of course you are above the energy thresholds which are detailed in the post mentioned above. Let us know if you need more clarification!