Plane wave decay dramatically


#1

Dear users,
When I tried to simulate an air box with the plane wave, I found that there is a huge loss through the whole FDTD area. For the “T” monitor, the E-field is zero and transmission are zero. It’s very weird because there is no any lossy material in the FDTD area. Could you please help me solve this problem? I appreciate it. @fgomez
THz-polarizer.fsp (5.0 MB)


#2

Dear @xinbo.wang

Your simulation span is too large (200010001000 um^3) with min mesh step set to be 50 um. However, your source injects a pulse centred at 1.5 um. Thus, your source can not be injected properly. If you want to fix the problem, I recommend you to modify the span to be in the few micron scale, and also use finer mesh.

Please also note that in the current simulation file you are using plane wave with PML BCs. This will cause edge effects that is explained here.

I hope this answered your question.


#3

Thank you for replying. I tried to inject a source with 300um wavelength and also increase the simulation time. Now the decay along the propagating way became small. However, when I put two monitors close to the source and far away from the source respectively, the transmission(close to the source) is higher than 90%, but the transmission (far from the source) is only 70%. I don’t know why there is decay in this case. I think the transmission should be the same in the whole FDTD box, right?
I’m looking forward to your reply. Thank you.


#4

Dear @xinbo.wang

That is right, we should not see any decay in the transmission at the lossless material. Did you modify the boundaries to be periodic BCs? Can you please share your simulation file for a review?

Thanks


#5

Yes, when I use BC boundary along Y direction, I can get a perfect plane wave field. So I tried to increase the dimension of Z and Y of FDTD box. The result shows good. I think because my wavelength of the source is 300 um, and the FDTD box should also become very large. Even though there is a little decay, it’s not bad.


#6

Dear @xinbo.wang

Thanks for the update.
Just looking at the plot, it looks like you are experiencing some edge effects?
I expect that your results should confirm zero loss if everything is set properly.

Thanks