Discussion Forum

Did you try RF Module?

--
Pu, ZHANG
DTU Fotonik

We only have access to the wave optics module and the structural mechanics. I have heard that the RF module and wave optics module are similar.

Would you know how to get started with this at all?

Sorry, I'm not familiar with Wave Optics module, which is newly introduced only in 4.4. You can check what equations it is solving.

--
Pu, ZHANG
DTU Fotonik

How would you be able to do this in the RF module?

In principle it will be straightforward in RF module. Of course if it is doable depends on how large your system is.

--
Pu, ZHANG
DTU Fotonik

It turns out that I do have the RF module for COMSOL. What you be able to explain how to set up the parameters, variables, and boundary conditions that is needed to find transmission vs. wavelength?

Let's say for now the system is simple: There are 3 layers- air, glass, and air, all perfectly aligned on top of each other with a certain thickness.

You really need to read the documentation. Although I complained many times the documentation is quite bad, your problem definitely can be solved by reading documentation.

--
Pu, ZHANG
DTU Fotonik

Ryan,

There is example from the Comsol model library for analysis of reflectance and transmittance of light through the layered material:
www.comsol.com/model/fresnel-equations-14713

Note that both RF and Wave Optics can be used for this type of problem.

Regards,
Sergei

Hey Sergei, thanks for the reply.

I've gone through the tutorial and the documentation on what they used. I feel like I understand how they got their results.

I have a few questions that you hopefully might be able to answer.

1. If I now want to model transmission through multiple interfaces (more than one layer), do I still only need two ports, or would I need one at each interface?

2. How would this affect the number of wave propagation constants we need, as well as each transmitted and reflected variable for each interface? It seems we would need two wave propagation constants at each interface, as well as four coefficients for transmission and reflectance at each interface.

2. Is there any way I can change the transmission equations to depend on wavelength, and not the angles of incidence and refraction? Or if I could change the results to map transmission versus wavelength?

Ryan,

1. You still need two ports for transmission through multiple interfaces (you can even get away with one port for incident wave and replace exit port with PML).

2. Probably, you mean wave-vectors but not propagation constants (propagation constants are for guided waves). Wave-vector components in each layer are completely defined by incident wave-vector and refractive index of the layer since tangential component of the electric field remains the same for each interface. Transmittance and reflectance coefficients at each interface can be calculated using Poynting vector or by solving multilayer problem several times with one less layer for each simulation.

2.You can perform frequency sweep for the fixed incidence angle to get reflection and transmission as a function of wavelength.

Regards,
Sergei

Thanks Sergei, your answers are very helpful.

Another question I had is that currently the Fresnel model has the has the wave vectors coming in from the side of the material.

To study and model EM waves that come in from bottom to top instead, would you just have to change the Electric field vector from depending on the y-component to depending on the z-component?

Ryan,

The incident wave is coming in from the top surface. You can see this by plotting Poynting vector which gives the direction of the power flow and, hence, the direction of the wave propagation, as shown in the attached image.

The basic idea is that whatever direction incident wave is coming from, you decompose the wave-vector to the normal to the interface component and tangential components to the interface. Then, use normal component of the wave-vector to set-up port boundary condition and use corresponding tangential component for Floquet boundary condition.

Regards,
Sergei

Dear all,
The Fresnel example works well to understand the use of S parameters to calculate the transmitted and reflected power. A 2D model is not so difficult to derived from it.
In a 2D example the transmitted and reflected power can be calculated from the S11 and S21 parameters and the input (1) and output port (2). It works well as long as long as the ports are external boundaries. I didn't succeed to do the same thing when the ports are internal boundaries. Does anyone knows why ?

I am working on the flow of light within waveguides (thick waveguide) and along this waveguide some refraction index change may be present (like in a multilayer structure for example). The change in refraction can be ~500 nm long and as wide as the waveguide thickness, so light wave interferences occur and the transmitted power depends on the frequency. I need to calculate the transmitted and reflected power around these domains as a function of the frequency (visible light). Does anyone knows who how and can it be done without the use of S parameters
I have tried to use the operator side(x,y) but being a beginner I don't know how and there is no example in the documentation. In truth the Comsol support showed it me some month ago but I lost this info.
I thank you for your comments
PB

Hi Sergei
i really want to know how to set floquet boundary condition in weak form PDE manually ?
could you help me with that?
thank you
Yang