FDTD simulation of absorbed power


#1

dear all
I have learnt the FDTD simulation,now I want to reproduce the result of the paper below.But the result of my FDTD simulation is much different from the result of paper:
http://opticalengineering.spiedigitallibrary.org/article.aspx?articleid=1101226

Could you help me solve the question?

3x15meshfine1.31um.fsp (642.9 KB)
Some pictures of the simulation result of the paper:



#2

I wasn’t able to access the full text of the publication, but I checked the simulation file and images that you posted and a few things I would recommend checking are:
-Increasing the z span of the simulation region to include more distance between the top of the Ge material and the PML boundaries to prevent artificial coupling between the structure and the PML.
-Checking the thickness of the Ge layer since you mentioned 95% absorption in less than 10 um for 310 nm thickness whereas the structure that was set up in the simulation file had a 400 nm thickness.
-What is the value is in the figure showing the cross section of the structure - I think it may be the amplitude of Ex
-Convergence testing of the mesh - since mesh accuracy 1 is being used the mesh may be a bit too coarse for accurate results.

Hopefully this helps!


#3

dear nlui

I haved modified the programm as you said and now the Ge layer thickness is 310 nm.I also increased the Z span of the simulation.Now the length of the Ge layer is 10um.I also make the mesh accuracy from 1 to 3.
the picture below is got from my simulation,which is different from that from the paper.


Did I make any error or the result is just this?
waiting for you feedback.:relaxed:3x15meshfine1.31um.fsp (671.5 KB)
Thanks a lot.


#4

I checked the simulation file but did not find any obvious problems with the setup. If convergence testing of the mesh has already been done then it might be worth double-checking the description of the figure from the paper to see if there could be any possible difference in the structure or the value that is being plotted. For example, if I plot the real part of Ex I can see a pattern with more lobes that looks closer to the figure from the paper, so perhaps they are plotting the absolute value of the real part of Ex instead of the E amplitude or intensity and that could explain the difference.


#5

dear nlui
Thanks a lot for your feedback:relaxed:.