Discussion Forum

Posted: 1 decade ago 25.03.2013, 12:49 GMT-4

It's rather tricky to calculate transmission and reflection using "scattered field". It's much easier to do using full-field formulation and Port boundary conditions.

Posted: 1 decade ago 25.03.2013, 13:16 GMT-4

Thank you for your kind reply. I think so, but what I want to do in the end is to calculate the transmission of 2D diffraction grating. In case of 2D diffraction, it's highly complicated to think of all ports, especially the diagonal diffraction mode. Comsol support suggests me to use the scattered field and PMLs, instead of ports. However, when it comes down to it, I cannot find the way to get the transmission. This is the plot. Regards, Yutaro

Posted: 1 decade ago 25.03.2013, 18:30 GMT-4

[QUOTE] In case of 2D diffraction, it's highly complicated to think of all ports, especially the diagonal diffraction mode. [/QUOTE] For 2D gratings you have the same number of ports as for 1D: just two ports. You will need to setup more periodic BCs as well as excitation at the "in" port which is still easier than setting up the background field. And at the end, the transmission and reflection are automatically calculated for you. Another strong argument for periodic BC is that you will need to simulate only one period of the grating. In the case of PMLs, you will need to include many more unit cells in order to achieve the same accuracy.

Posted: 1 decade ago 26.03.2013, 05:44 GMT-4

Thank you for your reply. Could you elaborate a little bit more? "For 2D gratings you have the same number of ports as for 1D: just two ports. You will need to setup more periodic BCs as well as excitation at the "in" port" In the "plasmonic wire grating" tutorial, we have to prepare 6 ports (including 1 excitation port) because there are 0, ±1 order diffractions. If we want to simulate 2D diffraction grating, for example 2D nanowire square mesh, I think we have (0, 0), (0, ±1), (±1, 0) order diffraction, and what's more, depending on the wavelength and diffraction period, we have even (±1, ±1) order diffraction, which is the diagonal direction. Could you let me know why we need only two ports and the meaning of "more excitation ports"? "In the case of PMLs, you will need to include many more unit cells in order to achieve the same accuracy." In the "plasmonic wire grating" tutorial, we put ports on top and bottom and put periodic BCs on sides. I replace these ports with PMLs (so PMLs on top and bottom and periodic BCs on sides). I've tried only normal incident wave, but the 2D plot (electric field norm) from the PMLs case looks quite similar to the one I got from the "port" case. Is it difficult or inaccurate to get the transmission from this result? Sincerely, Yutaro

Posted: 1 decade ago 26.03.2013, 13:29 GMT-4

1. By "only 2 ports" I meant 2 boundaries that needs port BC. You are right saying that you will need additional BCs for non-zero orders. But you will need to put only 2 extra BC per boundary for each diffraction order, i.e. you can define ports for (0,±1) and (±1,0) in such a way that it automatically covers (±1,±1) case. At that definitions for ports (0,±1) and (±1,0) are almost identical due to symmetry. At the end, it all gets down to correctly writing one k-vector for each diffraction order (which is unfortunately not trivial). I suggest starting from simpler cases such as normal incidence, etc and testing transmission through a film as there is analytical solution to it. 2. I have never tested "scattered field" formulation together with periodic BC. I see several ways in which it can fail depending on the particular implementation. If it works for you, let us know! For "scattered field", I always use 6 PMLs.

Posted: 1 decade ago 27.03.2014, 07:01 GMT-4

Can anyone help me in this regard? I had made a 2D grating structures in 3D modeling Wizard of Comsol 4.4 version, While I am putting Diffraction orders( m=0, n=0 ; m=-1, m=-1, n=0 ; m=+1, n=0) with "port reference points" as diagonally opposite end of Unit square cell with periodic boundary conditions on all four sides( by doing copay face). Unfortunately it gives me zero electric filed while i am running., If I run the same model without diffraction order it gives the result but only from one order diffraction (normal reflection) . Since I need to calculate reflection from at least two orders (0 th order and 1 st order) I have to include those diffraction orders to my model Can you help me in this regard? can any one tell me how to put diffraction orders (with proper port reference points)in 2d grating on 3D model? Also How can i implement PML layer over the port, since if i add pml layer on the top of port, selecting port reference point is declined by default. I follow the pyramidal absorption model for making the infinite array of my 2d Grating nano structures( there a PML layer added over the port) the plasmonic wire grating model available in Comsol library is just a 2D model , so we can't able to follow the same thing in here in our model. I used this blog for making port reference points and putting periodic ports http://www.comsol.co.in/blogs/modeling-electromagnetic-waves-periodic-structures/ Please anybody help me in this regard.. Thank you Aneesh M Joseph

Posted: 1 decade ago 27.03.2014, 07:03 GMT-4

Can anyone help me in this regard? I had made a 2D grating structures in 3D modeling Wizard of Comsol 4.4 version, While I am putting Diffraction orders( m=0, n=0 ; m=-1, m=-1, n=0 ; m=+1, n=0) with "port reference points" as diagonally opposite end of Unit square cell with periodic boundary conditions on all four sides( by doing copay face). Unfortunately it gives me zero electric filed while i am running., If I run the same model without diffraction order it gives the result but only from one order diffraction (normal reflection) . Since I need to calculate reflection from at least two orders (0 th order and 1 st order) I have to include those diffraction orders to my model Can you help me in this regard? can any one tell me how to put diffraction orders (with proper port reference points)in 2d grating on 3D model? Also How can i implement PML layer over the port, since if i add pml layer on the top of port, selecting port reference point is declined by default. I follow the pyramidal absorption model for making the infinite array of my 2d Grating nano structures( there a PML layer added over the port) the plasmonic wire grating model available in Comsol library is just a 2D model , so we can't able to follow the same thing in here in our model. I used this blog for making port reference points and putting periodic ports www.comsol.co.in/blogs/modelin...tic-waves-periodic-structures/ Please anybody help me in this regard.. Thank you Aneesh M Joseph