reflectivity and transmissivity of glass . after 80+ angle why results is incorrect

Answers

  • JaspreetSinghJaspreetSingh Posts: 11Member

    I am new on fdtd. i tried simulation for reflectivity and transmissivity of glass 1.5 index but after 80+ angle its incorrect. please help me

  • gsungsun Posts: 583Ansys Employee

    Do you use Standard PML 8 layers at an incident angle of 80+ degrees using Bloch BCs? it is mostly due to the degraded PML performance at large angle. You can use Steep Angle PML with more layers to suppress the reflection from PML:


    Bloch BC method: https://support.lumerical.com/hc/en-us/articles/360034382714-Bloch-boundary-conditions-in-FDTD-and-MODE

  • JaspreetSinghJaspreetSingh Posts: 11Member

    @gsun sir its not working. again its gave same answer

  • gsungsun Posts: 583Ansys Employee
    edited June 10

    Please provide more information: what is the wavelength? how many layers of the Steep Angle PML have you used? I can do a test.

  • JaspreetSinghJaspreetSingh Posts: 11Member

    @gsun sir I take glass(Sio2) with X,Y span 2 micrometer and Z start -1.5 with 4.5micrometer span, a plane wave source along z backward direction with 633nm wavelength above the glass , My FDTD region x, y span .6micrometer, Z span 2.6 micrometer with boundary condition x, y periodic Z bloch. in profile as you say i use steep angle with 34 layer. for Reflectivity put monitor above the source and for Transmissivity at z=0, then use sweep window for angle zero to 90.

  • gsungsun Posts: 583Ansys Employee

    At 90 deg, transmissivity should be zero, as you can find this principle from theory in a textbook.I guess what you used in Z is not Bloch BC (should be PML) other wise you would not get the result.

    I found a wiki page:


    when the incident angle theta approaches 90 deg, cosine is zero, so reflectivity is unit so the transmission is zero.


    To accurate simulate such case, I usually suggest to have two or three simulations with different PML layers for efficiency. The large the theta, the more number of PML layers (at large incident angle, the steep angle PML should be used).

  • JaspreetSinghJaspreetSingh Posts: 11Member

    @gsun sorry sir , last time i by mistake wrote wrong boundary conditions. i am used x, y bloch Bc and z is pml. i am really sorry for my mistake.

  • gsungsun Posts: 583Ansys Employee

    No problem. As mentioned previously, close to 1 reflection is theoretically correct when the incident angle is approaching 90 deg; for simulation, the PML will lead to high reflection in such cases. You might just simulate up to 85deg or 88 deg with different PML settings. If it is not actual device but you just want to verify the software, I would suggest just to simulate smaller angle not to very close to 90deg.

  • JaspreetSinghJaspreetSingh Posts: 11Member

    @gsun yes sir for 88 the of R is 1 but T is not zero.

  • gsungsun Posts: 583Ansys Employee

    Since at this large angle the PML degraded significantly, you will need to modify not only PML but also the Autoshutoff min and the simulation time:reduce autoshutoff min and increase simulation time. You may also increase the zmax of the simulation region. At such large angle it is challenge to get accurate result.

  • JaspreetSinghJaspreetSingh Posts: 11Member

    @gsun sir when i increase z max larger than z of my glass value the its show too much error. Even in that's case graph is not like our actually results.

    Also when i put my monitor for T value outside the galss in that sir case almost graph is too much distorted.

  • gsungsun Posts: 583Ansys Employee

    The PML should not be extended to air otherwise you are simulating a cavity. You should first increase zmin of SiO2 and then increase PML zmin. Similarly T monitor can only be inside SiO2 when you just want to simulate the air-SiO2 reflection and transmission.

    I guess you are just testing the FDTD reliability. If you want to get result for such device and compare with experiment, there is an analytical solver: stackrt: https://support.lumerical.com/hc/en-us/articles/360034406254-stackrt-Script-command Please try.

  • JaspreetSinghJaspreetSingh Posts: 11Member

    @gsun thankyou sir for yours help. sorry for my late respone.

Sign In or Register to comment.