Unreasonable output while setting a large number timesteps.

[Manoranjan Minz]

I was simulating a grating based coupler having 4 ports (Endings of both waveguides). I am feeding a broadband pulse centered at 1550nm as input source (modal TE profile) to one port and observing output power spectrum at other 3 ports. As per the principle if we use PML boundaries after a long enough time there should not be any energy left in the observation area since it will be absorbed by PML layers. But in my case when I use timesteps of about 500000 and I see one of the electric fields(in time domain) at any observation points, its magnitude goes above 10^9 which is unreasonable. Electric field doesn’t die off with time, instead it increases exponentially to huge numbers. My simulation properties are as follows.

Simulation Properties:
Dimension of simulation domain = 20um x 2um
cell width = 0.005um (smaller than (lambda/10))
delta_t = 1e-17s (set less than the auto value)
timesteps = 500000 (larger than the auto value)
DFT for 200 points

What could be the reason for this unreasonable output. Image is attached. Can somebody help?

[Manoranjan Minz]

Waveguide dimensions are:

Waveguide1: Length=20um, Width=1um
Waveguide2: Length=20um, Width=140nm
Grating: height=300nm, period=360nm

[Scott Newman]

The results you are showing are the results of a divergent FDTD simulation. Typical reasons for this is mesh that is too large or a material defined such that it has gain.

1. I know you say you are using less than lambda/10 but that statement is based on your central wavelength. How broadband are we talking? What is your smallest wavelength?

2. What materials are you using in your structure? If you defined a dielectric with real and imaginary components what is the sign of your imaginary component? In OptiFDTD materials are defined as ~n = n – jk. Positive k is loss so the imaginary component must be negative for loss.

Scott

[Manoranjan Minz]

Sir,

1. Center frequency = 193.414 THz, FWHM = 147.09 THz. DFT spectrum was set to auto which was from some 1420nm-1700nm wavelength for 200 sample values.

2. I am using simple Si and SiO2 with constant real refractive index. n_Si = 3.473, n_SiO2 = 1.444. There is no imaginary part.

I can’t figure out where did I make the mistake.

[Scott Newman]

From what you have posted everything seems fine. What version of the software are you using?

Would it be possible for you to upload a copy of your design file so that I can take a look? I would suggest zipping the fdt file as the website sometimes has issues with that file extension.

Scott

[Manoranjan Minz]

I am using OptiFDTD 14.0. I am sorry I can’t upload it here, if you have any private mail ID I could send it over there.

Regards
Manoranjan

[Scott Newman]

Certainly, please send your file to scott.newman@optiwave.com

[Manoranjan Minz]

Sir, I have sent the file to your mail id. Please look into it and kindly revert back.

Thank you
Manoranjan

[Scott Newman]

Thank you Manoranjan for forwarding your design file. After running some tests it would appear that the number of layers selected for the PML is inadequate for this design.

While the default settings in FDTD are selected to maximize the range of effectiveness the PML along with your mesh size must be tested and verified using convergence testing.

Doubling the PML thickness to 20 layers removes the divergent behavior and returns expected results.

Scott

[Author]

Posts