Dear Shahid,

In the most of the published research article there aren't a specific way to select the substrate thickness most of them done by trial and error till the best results were achieved. I think you should select the thickness as the antenna thickness.

Dear Hamzah M. Marhoon ,

Thank you for your Answer. In my case I have found that when I take substrate thickness as a multiple of 700 nm (e.g. 2700, 3400, .... 9700, 10,400 nm and so on), I get the best results. For other thicknesses, I get similar type of peak but with reduced transmission. I don't know much about the underlying physics but I think I should use the best results.

Thanks for your response :)

Shahid Hussain, Do you want to couple the metamaterial with an antenna?

Hamzah M. Marhoon , No, I am designing a metasurface for polarization change after transmission.

Dear Shahid Hussain , why did you choose multiplication of 700nm? Is that the resonant wavelength or the unit cell periodicity?

Dear The Linh Pham ,

After becoming familiar with CST software, I used semi-infinite substrate by using open boundary conditions instead of add space. As a result, the transmission did not change with substrate thickness. The substrate should not be lossy while using open boundary condition for metasurfaces in CST.

Dear Shahid Hussain , do you have any idea if the substrate is lossy, especially in mid-infrared region?

Hi The Linh Pham , usually we don't use lossy substrates when we are interested in transmission of structures. Even if it is slightly lossy, we ignore it. Aluminium oxide and calcium fluoride can be deemed lossless in the mid-infrared. If the substrate is lossy, where we cannot ignore the absorption, it is challenging to simulate in CST. Other photonic software such as Lumerical FDTD or COMSOL can be used for this purpose.

Hi Shahid Hussain , thanks a lot for your helpful advice.