Learning Forum

[matteo]

Hi,
I'm trying to simulate an EO phase modulator in LNOI. The design is quite simple, a rectangular waveguide with electrodes on the sides. To model it, I've been following the BTO example I found on your website. As far as I understood, there are three main steps to follow, for a given geometry:

1. Calculate the electric field distribution
2. Using a script, redefine the refractive index tensor in the region of interest. As far as I understood, the idea is that, since Lumerical does not support non-diagonal anisotropic tensors, one needs to define a "fake" diagonal refractive index tensor, and then apply a Matrix transformation to convert it to the desired non-diagonal refractive index tensor. Is that right?
3. Calculate the eigenmodes in the new structure.

I've tried to do so, but I have some questions regarding the actual implementation.

1. Is it possible to connect the designs in CHARGE and MODE, to make sure that the simulations are consistent (without external interfacing eg via Python)? Eg, waveguide dimensions in both simulations are the same. 
2. In the BTO example, the original, unperturbed n (used to calculate the electrooptic response) is hardcoded. In my .lms files I already have a wavelength-dependent, anisotropic model for the unperturbed material. Can I load that into the script - possibly with interpolating the wavelengths that are missing in the original database? 
3. In the BTO example, you define a matrix transformation for different (x,y) positions of the waveguide. How does the Matrix Transformation interact with the underlying meshgrid? For example, if my refractive index is defined on a 10nmx10nm grid while the matrix transformation is defined on a 500nmx500nm grid, how is the transformation mapped onto the material?
4. Is it possible to introduce a refractive index perturbation to a pre-existing model, or can I only override the refractive index values?

I have also two side questions, related to that.

1. My modulator design does not have semiconducors. What is the correct way to simulate it? Can I simulate it in CHARGE, or is it better in a different solver? At the moment I've been doing it in CHARGE, with the addition of a dummy semiconducting region with a high gap (to make it basically and insulator).
2. To check if I understood how to model things in CHARGE, I tried to calculate the electric field in a capacitor with parallel, infinite planes and 1V across the faces. The boundary conditions are 1V and 0V for the two electrodes. The resulting electric field is correct, but when looking at the voltage monitor, the faces have weird voltages, like -3V and -4V, i.e. they have the correct differences, but with an offset. What is the reason for such a behaviour?

[matteo]

In addition to the questions above, I have a follow up (I couldn't find how to edit the OP). In your example, you import a rectangle of "modified" refractive index. I need to modify more complex region (e.g, trapezoidal, or composed of multiple rectangles of different heigths). Moreover, I need to take into account possible claddings and so on. Therefore, is it possible to apply the imported sampled 3d material only on the region of interest (identified by a well-defined CAD element, like a rectangle)? Like, can I generate a region of modified refractive index which is bigger than the waveguide, and then appy it only where I have defined my original waveguide,with operations like an intersection or something like that?