Total energy deposited in sensitive detector dependant on step size

Particles, Track, Event, Run and Biasing
1

_Geant4 Version: 11.0.2
_Operating System: windows 11
_Compiler/Version: Visual studio

My understanding is that the step size of a simulation is just how frequently the various cross-sections are evaluation for the interaction. To me a smaller step size could result in inaccuracies occurring at material boundaries but I don’t understand why it would change the total energy deposition occurring in a large homogenous region.

For example, I have made a simulation using G4PVParameterised in which I have a large volume of lead 100010001000 pixels with each pixel being 0.1mm wide. I simulate 10 20keV photons from the centre of this cube and sum up the total amount of energy deposition that occurs using G4PSEnergyDeposit3D.
When I have the maximum stepsize in the volume limited to 0.001mm I observe the total energy deposition as being 0.2MeV as expected, but when I increase the maximum step size to 0.01mm then I observe less than 0.2MeV being deposited.

Assuming this is indeed intended behaviour can anyone explain why this changes the total energy deposition? I would expect the result to be the same total energy deposition but for the spatial resolution lower.

Thanks in advance for the help

2

There might be the effect of Geant4 dropping the track upon certain energy levels (too low), and the dropping may occur too soon. Your energies are very low in general; that is the reason for the guess.

3

Thank you Phirippu!

If it is a matter of the energy being too low then shouldn’t this be more influenced by cut values/ energy ranges? What is the best way to make sure that a track is not abandoned without depositing its energy?

In general I guess I am a bit surprised that the simulation would drop a track without depositing the remaining energy and am not entirely sure why this would be dependant on the step size rather than energy ranges. Any help would be really appreciated!

4

You can try to

  • make a record for each step to see when the track is stopped and how.
  • adjust cuts if applicable
    This must give a better picture of the issue. You can also try on a single volume or fewer volumes, as the issue may be connected to insufficient memory for the hit history or something similar.