IR transmission through sample only correct with a vacuum layer (heat simulation)

[kathinkagerlinger]

When simulating a sample with 500nm Al as heatsink, I get a non zero transmission (1-4% transmission, depending on the sample architecture) of an IR excitation (1039nm) even though no IR should go through the Al. The multilayer absorption profile looks right though. If I add a very thin vacuum layer after the Al layer, then I always get a transmission of 0% which is expected.

Is it maybe possible to add an option, where an extra Vacuum layer is added to the stack?