Taper simulation in reverse direction (From large to small)



Hello, I am working on MMI with taper in Lumerical mode solution EME Solver. Since i am not getting a good output i just decided to do the simulation only on the taper part.
The following were the dimensions:
1.Input Waveguide
Width = 1 microns
Length = 5 microns
2.Taper Region
Width = 1 microns
Length = 7.5 microns
3.Output Waveguide
Width = 0.3 microns
Length = 5 microns
Thickness = 0.35 microns
Refractive indices
Silicon = 3.455 (Core)
Silicon Oxide =2.1 (Cladding)
Background index = 2.1

When i simulated the taper alone in forward direction (from small to large) i got 100% transmission. But when i simulated in reverse direction for Fundamental TE mode i am getting output more than 100% i.e. 120% which is not possible and negligible power in the Fundamental mode. I simulated under following conditions:
Number of cell groups = 3
Cells = 1,20,1
Subcell method = None
Eigensolver settings is not allowed
Modes = 60 in each
Energy conservation = None
Boundary Condition = Metal
Port 1 = Port 2 = Taper width = 1 microns
Fundamental TE mode is used.
I got the User s matrix as:
EME_user_s_matrix_slice (2, 2)
0.000554394 1.21074
1.21074 6.459e-05

Please see the simulation file attached.


Hi @minusunny

Can you please share your .lms file that has the material used in the script like SixNy (Glass) - Const(21) and Si (Silicon) - Const(3455)? Please make sure that your script runs with no error to avoid unnecessary delays.



Thank you for the response. Please see the lms file shared


Hi @minusunny

There are two things to pay attention to in your case:

  1. You need to increase the EME yz span so that the modes are not truncated by the boundaries:
    Reflection problem form straight and bend silicon waveguides

  2. The modes that you inject at two ports are different modes and may propagate differently. However, I expect that the 120% transmission to be due to small EME area.


Thank you for correcting the error. Above 100% transmission was due to the EME area as you mentioned.I have another doubt, is it possible to get 100% transmission for the above simulation (reverse taper) ?


Hi @minusunny

When I increased the y- and z-span of EME and port objects to 3\(\mu m\), I obtained identical results either choosing port 1 or port 2 as my source port:


As you can see, I am obtaining a 100% transmission when port 2 is selected.
Simple taper reverse_BK.lms (469.3 KB)


Thank you so much for your response