Discussion Forum

Dear Colleagues,

I'm working on 2D RF simulations using port and floquet periodic conditions with oblique incident angle in v4.1a. I have recently applied v4.1a update 3, and the results for TE (Ez, Hx, and Hy) case have been much improved and became reasonable. However, I still have strange results for TM (Hz, Ex, and Ey) case, as I explained below. I would appreciate if you could give me some suggestions.

Please find the attached mph. files. The structure is quite simple, a rectangular with 1(W) x 3(H) micro meters having the refractive index of 1.0 (vacuum). The bottom and top boundaries have ports with the phase delay exp(-i*K1x*x) along the x direction and the propagation constant beta K1y, to excite oblique incident wave propagating from the bottom to top ones. The left and right sides have floquet periodic conditions with K1x and K1y, where K1x and K1y are the wave vector related to the incident angle. I believe these settings are not special, as seen in somewhere at www.comsol.com/papers/5259/download/Yang_pres.pdf.

For TE (Ez, Hx, and Hy) case, I set Ez for the excitation and I got reasonable results, where the transmission and reflection are always 100 and 0% over any incident angle. I also confirmed that the values of Ez and Hx satisfied the equations on p.63-65 in RFModuleUsersGuide.pdf, where Ez = - ZTE*Hx, where ZTE is the impedance for TE wave, now is ZTE = 120*pi/cos(incident angle). These results are the exactly the same as what I got using v3.5a with the same settings.

For TM (Hz, Ex, and Ey) case, on the other hand, since we are not able to set Hz at port for the excitation in v4.1a, not like in v3.5a, I used Ex and Ey instead to excite TM wave. However, the results are strange and the transmission and reflection show unreasonable values over 0-90 degree incident. For some incident angles, the values of Hz and Ex satisfies the equations on p.63-65 in RFModuleUsersGuide.pdf, where Hz = 1/ZTM*Ex, where ZTM is the impedance for TM wave, now is ZTM = 120*pi*cos(incident angle), but this is not the case for certain angle (e.g. 45 deg.). Since I got reasonable results for TM case using v3.5a with similar settings where I used Hz for the excitation, I'm now wondering this could be bug in v4.1a.

Thank you for your help in advance.
Atsushi