Roughness of surfaces

Physics Processes, Models and Cross Sections
1

Dear all,

I’m simulating a plastic scintillator in order to tune the roughness of the surfaces (scintillator-air). I’m using the unified model with a rough surface. However when I set the sigma-alpha to 0.0 I have different results to the case when I set the surface to polished. If I understand correctly if I set the sigma_alpha to 0 it is equivalent to a polished surface? it’s correct? and the lower the sigma_alpha the closest to a polished surface?. In addition what are the default values when no G4MaterialPropertiesTable is set for the reflection probabilities (SPECULARLOBECONSTANT, SPECULARSPIKECONSTANT, etc), but of course the rindex of the two materials are defined for each material. This is my implementation:

surfaceAirTargetTopBottom = new G4OpticalSurface(“surfaceAirTargetTopBottom”,unified, ground, dielectric_dielectric);
//surfaceAirTargetTopBottom = new G4OpticalSurface(“surfaceAirTargetTopBottom”,unified, polished, dielectric_dielectric);
surfaceAirTargetTopBottom -> SetSigmaAlpha(fSigAlphaPENAirTopBottom);

new G4LogicalBorderSurface(“surfaceAirTargetTopBottomOut1”, GetPhysicalVolumeByName(“target_1”), physicWorldBox, surfaceAirTargetTopBottom);

Thanks in advance for any input!

Luis

2

If I understand correctly if I set the sigma_alpha to 0 it is equivalent to a polished surface?

No. Please see the figure in the documentation:

1226×874 61.1 KB

In addition what are the default values when no G4MaterialPropertiesTable is set for the reflection probabilities (SPECULARLOBECONSTANT, SPECULARSPIKECONSTANT

3

Thanks for the reply. Just be sure, no specifying the type of reflection implies that it is set to lambertian at 100%?

4

If SPECULARSPIKECONSTANT, SPECULARLOBECONSTANT, and BACKSCATTER aren’t specified, all reflections will be Lambertian.

5

Is there any physical justification to set all reflections to Lambertian? Reading the paper [Levin-1997] it seems that might be a better idea to set to 100 % SPECULARLOBECONSTANT when none is specified by the user when using the rough option in the unified model…

6

It’s not clear to me that default values for these parameters is a good idea. The user should be required to set them.