Learning Forum

[Giovanni Piscopo]

Dear all,

as seen on another post on this forum, I'm interested in calculating the phase response of a 2D grating (where the actual computational domain corresponds to the xy cross section while the z extension is infinite).

In order to do that, I'm acquiring as results the "grating charachterization" however, while for a 3D domain it works, for a 2D domain, the software crashes (it automatically closes).

How can I solve this issue?

Is there another way for calculating the phase response?

Thank you in advance.

Regards

[Amrita Pati]

Hi Giovanni,

I am not sure why RCWA is crashing. We had a fee bugs in the initial releases that had been fixed. Could you install the latest release of Lumerical and see if you still see the crash? Secondly, would you also be able to share a screenshot of the geometry and the "Geometry" tab of the RCWA solver (which consists of information on the interfaces)?

Regards,
Amrita

[Giovanni Piscopo]

 

Dear Amrita, 

thank you very much fr your answer! I will proceede to install the latest version to see if this problem gets fixed.

In the meantime I attach two pictures: one regarding the geometry and the other regarding the interfaces.

Basically the geometry consists of a substrate on top of which there is a waveguide layer and a grating on top of it.

To ease the design, the top interface of the substrate is at y=0 which also coincides with the bottom interface of the waveguide layer.

Also the thickness of the waveguide layer is 1.3613 um while the thickness of the grating is 0.1 um (in this case I’m considering the substrate as infinite).

 

Thank you in advance.

 

Regards

 

Giovanni

 

[Amrita Pati]

Hi Giovanni,

Thanks for sharing the info! I do not see any interfaces in the RCWA setup. You need to setup interfaces to tell the RCWA solver about the different layers of your geometry along the propagation direction. Then RCWA solves Maxwell’s equations in each layer, following which the fields are matched at the interfaces to calculate the overall reflection/transmission of the geometry. You should have interfaces at the red lines shown below:

I will recommend using the “reference interface positions” as you can set the interfaces using the max/min cooridinates of the different objects instead of using absolute co-ordinate values. I will recommend going through this introductory article for more information on how the algorithm works: RCWA: Solver Introduction. I will also recommend you to check out this example: SRG for AR Systems to see how the interfaces have been setup. 

Please feel free to let me know if you have any questions.

Regards,
Amrita

 

[Giovanni Piscopo]

 

Dear Amrita,

thank you very much for your answer! I will definitely read the article you sent me and implement the example that you linked!

Just as a additional info, I set the interfaces with this command: 

 

“set(“interface absolute positions”, [0, h_wg, h_wg+t]);”

 

where h_wh = 1.3613 um and t = 0.1 um and automatically I got the interfaces set as in the picture I attached.

To confirm also the result, I calculated the refractive index distribution and I was able to see all the three indices of the three layers (actually about that I’ve seen that there is a bug in the visualization of the cross section because the thicknesses of the layers do not match the actual values but in FDTD there is no such a problem. However I’ve seen on another post on the forum that this is actually a visual bug that does not interfere with the simulation results).

Also, is there a difference in results between defining the interfaces between "reference positions" and absolute positions"? Because I find more intuitive defining them in terms of "absolute positions".

Thank you in advance

Regards

 

Giovanni

 

 

 

[Amrita Pati]

Hi Giovanni,

You are welcome!

I am not sure if the visualization artifact has been fixed or not. I can check and let you know. Are you seeing any discrepancy between the index result and the geometry you have set? I think the issue was that the indices in the top layer and bottom layer were extending into the substarte and superstrate. Even if you do see this, as you said, this shouldn't affect the simulation results. I remember the Development team had told me that the right way to visualize the index in RCWA is to look at the index of each transverse layer (instead of looking along the propagation direction), but I do think that this might not be insightful sometimes. 

You can use the absolute postions as well. The advantage of the "reference positions" is that if you change the height of any of your layers, either manually or using sweep, the RCWA layer positions will automatically update. But in absolute position, it won't get updated. So, if you are planning to change the height of the grating or any other layer, then the reference positions would be the better choice, otherwise the absolute positions should work as well.

Regards,
Amrita