Optical Library problem with Geometry v02

[ggamezdiego]

The optical library generated with the new SBND geometry v02 presents unexpected values for the visibilities for some PMTs in the voxels intersecting with them. A sample of 30 PMTs in the TPC with X>0 and in the central APA-windows see a lot of more light than its symmetric situation for TPC with X<0:

To debug the issue, I have compared the result of the simulation for PMT-141 (a problematic one) and PMT 9 (a good one) (PMT layout below)

When I track the photons generated inside the PMT-141 (in the vacuum material), it typically results in the following:
ID Pre_PV Pre_material Pos_PV Pos_material OpBoundary Process
884 vol_PMT_in_PV matVacuum volOpDetSensitive_PV LAr NoGeomBoundary OpDetReadoutScoringProcess
884 volOpDetSensitive_PV LAr volPMT_PV LAr NoGeomBoundary OpDetReadoutScoringProcess
This means that those photons are being detected (and they should not! --> NO photon generated in a material different than LAr should survive/be tracked). It seems the photon does not see the boundary between the different volumes.

While the same simulation for the case of PMT-9 results typically in:
ID Pre_PV Pre_material Pos_PV Pos_material OpBoundary Process
884 vol_PMT_in_PV matVacuum volOpDetSensitive_PV LAr NoRINDEX CoupledTransportation
And in this case the photon is not detected/tracked, as should be!

I am using sbndcode v09_36_00 and geometry sbnd_v02_00.gdml (with wires). Without wires the issue is the same.
My impression is that it could be related with the geometry. To test this, I have done the same simulation for PMT-141 with another version of the geometry v02, and the problem is not present! This other geometry is a version that @gustavogx gave to me “in private” some months ago, before the new v02 geometry was published in sbndcode.

[pgreen135]

The optical library has now been updated in #245 and sbnd_data v01_11_00. The underlying problem with the geometry remains however and needs resolving to avoid issues with future library generation and/or full optical simulation.

[marcodeltutto]

Hi @gustavogx, @mrguzzo, just wondering if you have any update on this?

[bear-is-asleep]

Update - this issue persists still and will need to be fixed if the need arises to run the full optical simulation.