Problem in getting curves for the case of crystal structure in a defect mode

a12.fsp (1.0 MB)
I have drawn a crystal structure in a defect mode as shown in the file. The reflection curves which I am getting does not have perfect peaks as required. There are some disruptions. How can I remove those to get a perfect and smooth curve?

Dear @mohini.nagpal1

I think you are referring to the peaks marked red in the plot below?:

I think there are a few things that you can try:

  • Since you are using a plane wave and devices are periodic along the y-axis, you can reduce FDTD y-span. A few mesh cells along y-axis should be sufficient.
  • You need to leave at least half the maximum source wavelength (3um/2=1.5um) between the PML and edge of the stack.
  • If results didn’t improve, try increasing PML layers and make sure that it is absorbing incoming light properly.
  • Perform convergence test to make sure that your results do not change by changing mesh size, PML layer, and FDTD region.

You can also compare the results with stackrt command which analytically calculates transmisison and reflection results.

Please keep me updated with your results and I will be glad to be of a help.

I reduced the y span to few mesh cells along y axis.Also I tried working
out by making the PML layers to a maximum(64).But the results didn’t
improved much.What else can be done? What have you refered to as stack in
the second point?


Dear @mohini.nagpal1

Thank you for updating me with your findings.

I noticed that your simulation does not trigger auto shutoff level. Thus, I increased the simulation time to 10,000fs. Below is the new results (I have simulation file attached to this post):

The dips are more pronounced now.

I recommend you to perform convergence testing, and use override mesh to properly resolve the material (where they overlap with the edge of the geometry).

You can use stackrt command to calculate tranmission/reflection of a stack layer. Please check the link below for more information and let me know if you had any follow up question.

here is simulation file for you review:
a12_modified.fsp (1.1 MB)


Thankyou sir that problem is solved and results obtained are as desired.But I have uploaded one more file in which at the point of the peak, there are some values more than 1 which is obtained. How to correct that one and make it smooth ?a9.fsp (1.1 MB)

Dear @mohini.nagpal1

Ripples in transmission spectra or transmission grater than 1 are due to early shutoff, poor PML absorption, or other unwanted effects. In you case the problem seems to be due to early shutoff in which simulation finishes before reaching auto shutoff level.

To improve the results, decrease auto shutoff level to 1e-6 from advanced tab of FDTD, and increase time to 30,000. Here is the updated results with these modifications:

If you are planning on working with stack layers, I strongly recommend you to employ using stackrt command. It is fast and debugging is extremely easy.


Thank you Sir for resolving the issue.