Please enable JavaScript to view this site.

Application Gallery

cover_picture_mmi_coupler_small_zoom15This example has been updated. Find the latest version at MMI Coupler .





In this example we calculate the optical loss through a 1x2 port multi-mode interference (MMI) coupler for different linear taper widths at the input and output ports of the device.





Associated files



Related publication

D.J. Thomson et al, “Low Loss MMI Couplers for High Performance MZI Modulators”, IEEE Photonics Technology Letters, Vol. 22, No. 20, October 2010.

See also

Simple MMI coupler

EME solver analysis



Optical loss through MMI coupler devices can occur due to mismatch between the input waveguide mode and the modes supported in the interference region. Based on the design in the publication listed above by Thomson et al., linear tapers are introduced at the input and output ports of the device to reduce the mode mismatch and reduce loss. We analyze the loss for a range of taper widths.




The MMI structure with 10 micron long linear tapers between the input and output waveguides and the interference region is set up automatically using a structure group.


Simulation region

Since we are interested in collecting the transmission of the fundamental mode, the fundamental mode is selected in the EME solver region's ports. The structure has symmetry across the y-direction; therefore, the y min boundary condition is set to “anti-symmetric” to reduce the simulation time and memory. Since only positive y half of the structure is simulated due to the symmetry condition, only one of the two output ports is included in the simulation region, so there are only two ports set up in the EME solver (the input port and one of the output ports). The total transmission is calculated after running the simulations by doubling the transmission through the output port that is simulated.


In the EME settings tab, five cell group regions are set up. In the cell group regions that contain the tapers, ten cells are used in the x-direction to resolve the tapers, and mesh override regions are used to increase the resolution of the transverse mesh over the tapers. The CVCS subcell method is also used in the taper regions to avoid artifacts due to discretization of the structure in the x-direction. The CVCS method is recommended in cell group regions where the structure cross section is continuously varying.


Because the interference region supports many modes compared to the input and output waveguides, we use an increased number of modes to represent the propagation in this region.


Wavelength sweep

The plot below shows the MMI transmission as a function of wavelength obtained using the wavelength sweep feature in the EME Analysis window. To obtain the plot, set the wavelength span and number of wavelength points in the wavelength sweep section of the EME Analysis window to 1.5 - 1.6 um and 100 points, respectively, and then press wavelength sweep. Once the calculation is finished, use "visualize wavelength sweep" button and then select Abs^2 in the visualizer as is explained above.




Parameter sweep

In the "Optimizations and Sweeps" window, a parameter sweep task is set up to sweep the taper width property of the structure group between 0.4 microns and 1.1 microns and collect the user s-matrix.



The script file is used to run the parameter sweep and collect the user s-matrix results. The transmission in the fundamental mode through port 2 from port 1 is then calculated from the S21 element from the user s-matrix, and the value is doubled to give the transmission through both output ports. The result is plotted below.



Copyright Lumerical Inc. | Privacy | Site Map