Photonics

Photonics

Metamaterial

    • djihad-amina.djemmah
      Subscriber

      Hello,

      I am simulating a 2D matamaterial with FDTD in the range of 3-80GHz.

      I am trying to simulate a 2.5mm thick plate, with permittivity of 20 and permeability of 1.

      I used the grating S-parameters to calculate the S-parameters, and then I used Smith's script to extract the effective parameters.

      The problem I encountered is that when I plotted the effective index, impedance, permittivity and permeability curves, I have a weird result with a lot of fluctuation (see attachment), knowing that the initial values are correct and S_abs and S_phase seem correct.

      Could someone please help me to solve this problem.

      And I want to know if there is a way to plot the S phase directly in degrees rather than radians?


    • Guilin Sun
      Ansys Employee
      It is obviously that the result is distorted by the phase over 2pi as you can see the phase of S11. Since you are simulating a spectrum range more than 10X higher than the beginning frequency, it is normal to have phases over 2pi. Unfortunately the complex results automatically remove the extra phase, which you may need to add them back.
      Our current method based on Smith' method does not consider this extra phase change. There are some published methods which may handle this but need extra efforts. Please search some online.
    • djihad-amina.djemmah
      Subscriber
      Thank you for your answer.
      What intrigues me is that when I use a frequency range between [5-20GHz], I noticed that the material influences the effective parameters, i.e. for materials with low permittivity (e.g. eps=5) and with a non-zero imaginary part, I find correct results but if I put the imaginary part zero, in this case I find the fluctuations.
      and for materials with high permittivity (eps>10) I always find the fluctuations after the extraction of the effective parameters whatever the imaginary part (null or not).
      Do you think that the permittivity of the material influences the effective parameters?
      Do you have an idea to solve this problem?
      Looking forward to your answer.
      Thanks in advance
    • Guilin Sun
      Ansys Employee
      The physical material property does indeed affect this result since it effectively changes the wavelength inside the material.
      I am not sure how the material fitting looks like when you forced the imaginary part to be zero. In FDTD we use dispersive material models which means the imaginary part is essential, even it is small, if you do not use a constant dielectric index.

      I would suggest that you either use the dispersive material, or a constant refractive index for dielectric material disregarding minor material dispersion. However not to artificially force the imaginary to be zero other than from material model.

      As mentioned early, when the simulation spectrum is large to cover the range of phase change over 2pi, then you may need more careful for the post processing, and the extraction model. Smith' method applies to the cases where the phase change is less than 2pi.



Viewing 3 reply threads
  • You must be logged in to reply to this topic.