Your simulation setup looks good. I made some changes in reflections_modified.lms (309.2 KB) according to the points discussed below.
To answer your questions:
The most important one is the position of the green cross because it determines where the solver will calculate the reference vertical slab mode used to generate the effective material and collapse the 3D simulation to a 2D simulation. I think you placed it correctly since the vertical line crosses the two materials that form the waveguide, InP and InGaAsP. You can always check how the vertical slab mode looks like in the plot shown in the “Effective index” tab:
The blue crosses are used for checking purposes and their position does not affect the simulation. You can place them at different positions in the simulation region and then use the Material Explorer to check the effective index used at those positions (the core material corresponds to the green cross and the test materials correspond to the blue crosses):
The simulation region should be big enough in the y and z directions to avoid any undesired effects from the boundaries. For example, if you use PML boundaries, you should make sure the distance from the structure to the PML is at least half the wavelength to avoid evanescent fields at the boundaries. This is a good rule of thumb but sometimes you need to increase the size even further. If changing the size of the simulation region changes the injected mode significantly, it might be necessary to increase the size even more until the change is small.
You have to make sure the monitor is not placed at the grey region around the source as explained here.
You can look at the transmission through the monitor you placed behind the source to determine the net reflection. It might also make sense to consider using an additional monitor at the output to check the transmission. Also, I think you might find useful to use a mode expansion monitor to study in more detail the mode composition of the reflected/transmitted light.
Some additional suggestions:
- I noticed that in your original simulation the source size was too small to get the correct field profile of the injected mode. After increasing the size the injected mode looks much better.
- The monitor was also too small so I increased its size to make sure it catches all the light passing through the position where it is placed.
- It is usually better to start with metal boundary conditions everywhere except for x min and x max. The reason is that the PML can introduce additional effects. If the fields decrease to very small values before they reach the boundary it should be safe to use metal boundaries. However, in some cases (for example for taking into account radiative losses) it is necessary to use PML.