problem about the simulation of grating coupler

ShelleyShelley Member Posts: 2

Hi,I have a problem on the simulation of grating coupler. When I optimized 2D grating coupler, I got a smooth transmission.



But, when I use the same parameter to run 3D grating coupler, the transmission have great fluctuations.

the transmission of port1(input)->T:

port1->mode expansion monitor->T_out:


the transmission of port2(output)->T:

port2->mode expansion monitor->T_in:


I increased the simulation time and PML layer but the transmission still have the fluctuations. So, I think the problem is the back reflection.

I do not know why the back reflection is less in 2D simulation but great in 3D simulation?

Where the back reflection is from?

the back reflection seems to introduce from port1(input)?

Answers

  • kjohnsonkjohnson Posts: 172Ansys Employee

    Hello @Shelley ,


    This difference between the 2D and 3D results does not necessarily indicate there is a problem. It's possible that there is some effect causing these transmission spectrum results in the 3D simulation that is not accounted for in the 2D simulation. However this difference is not seen in the original simulation results, so there might be something wrong with the 3D simulation. It might help to increase the number of frequency points the monitors are recording in the 3D simulation, it looks like the number of points is not high enough to fully capture the transmission spectrum features in these results.

    What have you changed from the original example files? And can you explain why you think the back reflections are causing this?

  • ZhangXueleiZhangXuelei Posts: 16Member
    edited February 23
  • ShelleyShelley Posts: 15Member
    edited February 23

    Hello @kjohnson ,

    The number of frequency points of the port which is from original example files is 50. According your advise, I increased the number of frequency points from 50 to 300. The curve got smooth but still had the fluctuation. It seem to have nothing to do with the points.



    points=300:


    points = 50:



    According my need, I changed a lot from the original example files like material,duty cycle ,etch depth... Which is based my 2D simulation .


    The reason why I think the back reflections are causing this is expansion for port monitor(port1:input).


    the T_out looks like great( from 0.01-0.08), So I think the back reflection cause the fluctuation. I increased the simulation time and PML layer but the transmission still have the fluctuations So I think the the fluctuation has nothing to do with the boundary condition.

    But It is just my guess?


  • kjohnsonkjohnson Posts: 172Ansys Employee

    Hi @Shelley ,

    Thank you for the information and screenshots, I see what you mean about the back reflections. I agree that this is larger than we should expect for this device. I'm not sure exactly what might be causing this. I would guess that this is an issue with the ports. This might be due to the large bandwidth. Does the problem persist if you use a smaller bandwidth, for example the original 1500-1600 nm bandwidth from the example? For the ports, are you using multifrequency injection and multiple field profiles?


  • ZhangXueleiZhangXuelei Posts: 16Member
    edited February 26
  • ShelleyShelley Posts: 15Member

    Hi @kjohnson ,

    I will use smaller bandwidth(1500nm-1600nm) to run the simulation. The setting of multifrequency injection and multiple field profiles is default:



    I will use mode source to run the simulation to verify whether the port lead to the problem.

    If I get a better result, I will tell you.

  • ShelleyShelley Posts: 15Member

    Hi @kjohnson ,


    I used  smaller bandwidth(1500nm-1600nm) to run the simulation. The result did not improve.

    I used  ModeSource to run the simulation. The result did not improve.

    I used  GaussianSource to run the simulation. The result did not improve.

    These simulation explain that the port have no influence on the transmission. I guess whether the 2D simulation is not good and the design of the grating is defective. I guess whether the taper influence the result.

    Can you give me some advise?

  • kjohnsonkjohnson Posts: 172Ansys Employee

    Hello @Shelley ,

    You might be correct that this is a physical effect caused by the taper. However I think the fact that the mode expansion monitor at the waveguide was showing back reflections indicates that there is an issue with the simulation, there shouldn't be any reflections from the PML boundary.

    The next thing I would recommend would be to try convergence testing to see if reducing the mesh, increasing the simulation time or increasing the number of PML layers changes the results. If changing the simulation settings like this does not affect the results this will help confirm that this is a physical effect.

  • ShelleyShelley Posts: 15Member

    Hi @kjohnson ,

    I have a improvement on my simulation, but I still have a problem.

    I change the angle of taper and get some smooth curves.

    I previously used around 70 degree on the taper. Now, I use around 30 degree on the taper in 3D simulation.


    But, I meet a new problem. When I use 30 degree one the taper, the transmission is better than 2D simulation in the same parameter.

    2D transmission(0.3-0.4) :


    3D transmission(0.4-0.5) :

    Normally the 3D simulation result should be worse than 2D simulation.

    Why my 3D simulation result is better than 2D simulation?

    Is it normal ?

  • kjohnsonkjohnson Posts: 172Ansys Employee

    Hello @Shelley ,

    I'm not sure why the 3D results would be better than the 2D results. Typically for a grating coupler I would expect the 2D transmission results to be higher than the 3D results because the 2D simulation doesn't include the taper losses. But I guess it's not impossible for the 3D results to be better. I would recommend that you double check that all of the other simulation settings are the same between the two simulations, and take a look at these posts to see if there are any errors in the simulation setup:



    Let me know if that helps.

  • mborhanmiamborhanmia Posts: 2Member

    Hi Shelley,

    I am having difficulty setting up the 3D-FDTD grating coupler simulation. Could you please share your 3D FDTD file? I am already done with 2D FDTD, and transmission is 0.2-0.45 from 1500-1600~nm. However, from 3D-FDTD I got the transmission less than 1%. It would be a great help if I can take a look at your file.

    Thanks

Sign In or Register to comment.