32Bit OptiFDTD modes

[Paulo Lourenco]

Hi there.

I have been using OptiFDTD 32Bit and the number of detected modes is not what I expect it to be…
I am using a 4 * 4 * 6 um (x, y, z) air wafer enclosing a 3 * 0.5 * 6 um (x, y, z) Si waveguide and an input field located at (0, 0, 0.5 um). The waveguide is coordinates centered, meaning it has a -y offset of 0.25 um and the input field can assume two wavelengths, either 1.5 um or 1.3 um, for test purposes. The input field has been set up with 1.55 or 1.3 um GMCW with default time offset and halfwidth, modal transverse and A = 1 V/m^2. On the former case, when using ADI method on 3D solver I am able to get as many as 3 detected TE modes which is what I expect (max number of modes = floor(2 * x dimension / lambda), floor being the closest lower integer value). On the latter case, unfortunately, I get the same 3 detected TE modes which does not verify the max number of modes expression, as far as I understand it…
Am I doing something wrong?
Thank you up front.

Best regards.
PL

P.S.: .fdt file attached.

[Damian Marek]

Paulo,

Do you have a reference for that number of modes equation? Usually those types of equations are only accurate for 2D modes (slab or fiber) and don’t work well with arbitrary profile shapes. I would also recommend the Finite Difference mode solver for this application, because I have personally tested it a lot more in various situations and I have found it gets the correct number of modes.

Regards

[Paulo Lourenco]

Damian,

First of all thank you for your answer.
As for the equation’s reference, yes I do: – Bahaa E. A. Saleh, Malvin Carl Teich – Fundamentals of Photonics, Chapter 7, page 244 – ISBN: 0471213748 (Electronic) 0471839655 (Print).
What I would like to find is the relationship between the number of modes and the waveguide’s width and thought that through OptiFDTD should be able to achieve this.

Best regards.

[Author]

Posts