Created attachment 624 [details]
Energy deposited per photon (Geant4, and MCNP6)

Hello,

I have been using Geant4 to try and recreate a plot originally made in MCNP6 which is the attached plot with a nice linear distribution. The experimental setup is a 1 cm^3 cube of Cesium Iodide. The cube is hit with a monoenergetic, Circular Surface (radius .5 cm) source of photons, at various energies. The source is aligned perfectly with a side of the cube, so that the photons are incident at a 90 degree angle; the source is also placed just 5 mm away from the cube. The energy deposited in the cube per photon is graphed versus the photon energy. Geant4 seems to not follow this linear distribution at higher energies, as its curve seems to flatten out.

For the past 3 months, I have tried many of the pre-defined physics lists. I have also written my own, very basic physics list, using the EMStandardphysics constructor, and utilizing a step limiter, but to no avail. I have tried several different methods of scoring; via macros, or using the GetEnergyDeposit() function. I have had two other individuals, one from my lab and one from the g4forums replicate my results with their own implementations of the setup.

There are many papers which clearly illustrate the validity of the EM processes in Geant4. One important point that I noticed is that MCNP6 uses the NIST XCOM coefficients internally when carrying out some its calculations, while Geant does not. However, I have seen papers which show that various attenuation coefficients found in Geant4 should also match the NIST database despite not using them internally. Thus, it appears to me that Geant4 should be able to match MCNP for this simple simulation, and that this issue is merely a problem with my program. However, the fact that this behavior was replicated by other Geant4 users has led me to a dead end, and so I am hoping that someone with much more knowledge of the software will be able to lead me in the right direction. Any help would be greatly, greatly appreciated. Here is a link to the g4 forums where I have an ongoing discussion about this issue with a few other users.

https://geant4-forum.web.cern.ch/t/geant4-failing-to-match-with-mcnp6-and-nist-for-simple-test-of-energy-deposited-on-a-cesium-iodide-cube/2624/8