Reflectivity using glisur vs unified

Physics Processes, Models and Cross Sections
1

I have a simulation of a liquid scintillator detector, consisting of a aluminum box filled with scintillating material. When I implement the glisur model for the surface on the inside of the detector between the aluminum and scintillator as below

  G4OpticalSurface* BaSO4_surfaceLT = new G4OpticalSurface("BaSO4_surfaceLT");
  BaSO4_surfaceLT->SetType(dielectric_metal);
  BaSO4_surfaceLT->SetFinish(ground);
  BaSO4_surfaceLT->SetModel(glisur);

  G4double photonEnergy9[2] = {1*eV, 5*eV};
  G4double other_coating[2] = {0, 0};
  G4double total_Al[2] = {0.5, 0.5};
  G4double specular_Al[2] = {0.5, 0.5};

  G4MaterialPropertiesTable* MTP_BaSOLT = new G4MaterialPropertiesTable();
  MTP_BaSOLT->AddProperty("SPECULARSPIKECONSTANT", photonEnergy9, other_coating, 2);
  MTP_BaSOLT->AddProperty("BACKSCATTERCONSTANT", photonEnergy9, other_coating, 2);
  MTP_BaSOLT->AddProperty("SPECULARLOBECONSTANT", photonEnergy9, specular_Al, 2);
  MTP_BaSOLT->AddProperty("REFLECTIVITY", photonEnergy9, total_Al, 2);

  BaSO4_surfaceLT->SetMaterialPropertiesTable(MTP_BaSOLT);
  G4LogicalSkinSurface* SurfaceLT = new G4LogicalSkinSurface("BaSO4_surfaceLT", ReflectBoxLT_log, BaSO4_surfaceLT);

the results come out the same regardless of the fraction of specular reflectivity. When I change the “REFLECTIVITY” parameter the results change but when the “SPECULARLOBECONSTANT” is modified using any number between 0-1 the results do not change. This is fixed by changing the model from glisur to unified without changing any other lines. Is there a problem with the way the glisur model is implemented here?

This is running on the most recent version of lxplus with the following
Geant4 Version: 11.2
Operating System: RHEL 9.4
Compiler/Version: gcc (GCC) 11.4.1 20231218 (Red Hat 11.4.1-3)
CMake Version: 3.18.3

2

Glisur dates from Geant3 days, and I haven’t seen any documentation for it.

Unified contains the parameters from Glisur, plus many others. Probably SPECULARLOBECONSTANT isn’t part of the Glisur model.

I recommend using the Unified model, for new applications at least.

Let me ask: why do you want to use Glisur, and why do you think there is a problem (that is, why do you think SPECULARLOBECONSTANT should be part of the model)?

3

Hi Daren,

Thanks very much for your reply. This is an old code passed down to me and originally based upon the OpNovice example, which uses the Glisur model. I could not find much documentation either, but since the Book for Application Developers states that it is very similar to the Unified I assumed that it could be implemented in the same way. Perhaps you are correct and the SPECULARLOBECONSTANT does not exist in this model. I will use Unified going forward.