some question about the edge coupler


dear all
I have wanted to do a simulation about the edge coupler with my structure.I have learnt and follow the step as the official example taught.But the value of the power coupling in the wavelength(1550nm) is just about 20%,much lower than the value(50%) as we measure.Also the power coupling (6%) in the 1310nm wavelength is much lower than that (35%) we measure.I doubt that whether somewhere of my programm is set wrong or the software can’t simulate the edge coupler of my structure.
Hope to get your feedback.Thanks in advance.

1.lms (376.9 KB)


Hi, @jbwei

Please, provide the simulation file used.
It is much convenient to have one to answer you question.


Hi @jbwei

Your simulation file seems reasonable and I got the same value for 1310 nm even with finer mesh. I don’t have a clear idea why your simulation results does not match the experimental data.

Can you please check the dimensions of Si waveguide? I expect that fabrication of a waveguide with a cross section of 180 * 220 nm^2 will be very challenging and you will get at least 50nm error on the width. The thickness of 220nm is fine as its defined by wafer but I doubt that your fabrication result to match your layout.



@jbwei @bkhanaliloo
I didn’t catch the link to the simulation file


dear msaygin
the file is in the middle of the two can have a check again.
dear bkhanaliloo
the dimension of the waveguide is really 180nm*220nm.I have check it. In fact,we have simulated the coupling efficiency of waveguide and optical fiber with other software(fimmwave). Both of two simulated results are lower than the experiment value,but the result of the EME is lower and worse.I am not sure what is wrong with it.The waveguide is shown below and the thickness is 220nm with a taper structure.


Dear @jbwei

In your first post you mentioned that you have measured the spectra experimentally. As your simulation file seems reasonable, I was thinking maybe your fabricated devices is not exactly 180nm. Can you please provide references if you are comparing simulation results with reported results on scientific journals?