Discussion Forum

Hello,

I am working on a polariton waveguide to solve the lateral confinement dependence of dispersion curves. The core is made of Gallium Selenide, which is an anysotropic semi-conductor. My model has been tested with infinite width (using continuity boundary conditions) and showed results in agreement with the theory with perfect field maps. Hence, I am now adding width dependence to my model.

My model computes the out-of-plane wavenumber of a field propagating in the out-of-plane direction, and looking at the Ez field component. I am using the Mode analysis study, and searching for mode around 145000 rad/m for an energy "omega" of 230cm^-1 (approx. 6.9THz). These 2 parameters can be let unchanged for you to work on the model. Finally, I look for 20 modes, but only 1 is correct. Despite it succeeds to solve the good out-of-plane wavenumbers, it seems to struggle to solve a correct field map, especially in the waveguide (see the attached screenshot). As you can see, non physical shapes of the field appear inside the slab, whereas it seems correct outside the slab. The correct mode is the one depicted in the screenshot at 145240 rad/m (for a 5 um width), and should have the same shape but without the peculiar shapes inside the slab. This effect intensifies with decresing width, whereas it starts disappearing with larger width (100 um for example). Finally, it doesn't depend on the thickness.

I have tried to adress this problem by changing the mesh size, going from finer to coarser, without significative change. I also tried to add and remove Perfectly Match Layers, again without any change. Finally, I am using continuity lateral boundary conditions, but switching to scattering boundary condition doesn't change the result.

I was then wondering if you could have any idea of how to adress this problem, since I don't have anymore idea to troubleshoot this.

Thank you in advance for your help, you will find my model attached.

Best regards,

Clément Gradziel