This video is taken from the FDTD Learning Track on Ansys Innovation Courses.
Transcript
The structure that has been set up here is an asymmetric directional waveguide coupler
where light will be injected from the curved waveguide at the top left.
Some power will be coupled from the curved waveguide into the straight bus waveguide
by evanescent coupling across the gap between the two waveguides.
A power monitor has been set up at the drop port of the straight waveguide on the bottom
right in order to measure the transmission spectrum through the drop port.
Add a mode source from the Sources drop down menu.
Edit the source and under the General tab, set the injection axis to the x-axis, and
the injection direction to Forward.
Keep the mode selection as fundamental mode, but before we can calculate the desired mode
to inject, the geometry of the source region needs to be set.
Select the multifrequency mode calculation option.
This option should be used whenever injecting a broadband mode source.
This will calculate the supported mode profile over the broadband range rather than just
at the center frequency of the range.
Since we want to inject the mode of the curved waveguide at the input, select the bent waveguide
option and specify the bending radius as 6 um to match the radius of curvature of the
waveguide.
Set the bend orientation to 0.
For more details about how to choose the bend orientation angle based on the propagation
direction of the source and the bending direction of the structure, see the Bent waveguide solver
page linked below.
Next, under the geometry tab, set the x position of the waveguide to -3.068 um, set the y position
to 1.536 um, and set the z position to 0.11 um.
It is important to make sure that the source is centered with the center of the waveguide
when using the bent waveguide option since the bending radius is measured from the center
of the source region.
Set the y span to 2.5 um and the z span to 2 um.
The values for the position of the waveguide have been pre-calculated based on the geometry
of the structure so that the position of the source is centered with the center of the
waveguide.
Now that the geometry is set, go back to the General tab and set the theta rotation angle
to -28.3576 degrees.
Again, this rotation angle has been calculated to match the angle of the waveguide at the
given location of the source.
Now plot the calculated fundamental mode using the “Visualize Data” button.
Here we can confirm that the size of the source region is large enough so that the fields
are not being truncated at the edges of the source.
We can also choose to plot the field profile of different components of the E and H fields
from the visualizer to confirm the polarization of the mode.
To get more information about the calculated effective index, loss and polarization fraction,
change the mode selection to “user select”, then click on the “select mode” button
and click ok to select a new mode.
Calculate the modes.
The fundamental mode is the top one in the mode list, and I can see the details about
the effective index, loss and polarization from here.
You can also hover the mouse cursor over the polarization fraction or waveguide TE/TM fraction
columns for a description of the meaning of these values.
Click Cancel and change the mode selection back to fundamental mode, then click OK to
accept the settings.
In the XY view, the white angled line indicates the rotation angle of the source.
The thicker white line along the y-direction indicates the source injection plane.
With the source selected, the effective index and modal fields of the selected mode can
be visualized from the result view window.
Click run, and after the simulation finishes, you can right-click on the monitor and visualize
the T result to view the transmission spectrum through the drop port of the device.
Here are some more tips for setting up a mode source.
Ensure that the source span is wide enough to completely contain the mode of the structure.
You can visually check this by plotting the field profile and making sure that the field
amplitude of the mode decays to 0 around the edges of the source, but another way you can
test this is by doing convergence testing by changing the span of the source.
Start by calculating the mode, then increasing the span of the mode source and re-calculating
the mode, then compare the calculated effective index of the mode between the two different
source spans.
If the mode source region is wide enough to contain the full mode, the calculated effective
index should stay the same when you further increase the source span.