You are correct that FDTD/RCWA counts all interference effects since it is coherent by nature of the Maxwell Equations.
If the device is thick and peroid is relatively large, it may induce higher- diffraction orders. Thus the reflection and transmission can be oscillating. You may try to use more number of k vectors to do converging test and see if it is physical, provided that the number of layers are proper, eg the two "ports" are well resolved.
You may also try to use FDTD for verification with smaller autoshutoff min and much longer simulation time than 1e-5.
You are navigating away from the AIS Discovery experience