Absorption monitor issue - plasmonic nanograting

fbkfbk Member Posts: 1

Dear all,

I am trying to simulate a 1D metallic grating of Silver, over a double layers of silicon nitride and oxide, 11 nm and 5 nm, respectevely. The metallic nano grating is embedded in a PMMA passivation to avoid silver oxidation and the device is placed on 3um of Si. Schematic rapresentation of the strcuture in figure 1.

The purpose of the simulation is to evaluate the ART and the E and H fields in the device in the range (450-1100) nm.

Issue n1:

Evaluating the absorption in the metal, I noticed some intensification of the field near to the borders of the metal (1or2 orders of magnitude more than the field inside the metal), figure 2.

The field in figure 2 is evaluated at the wavelength at which a peak is present (550nm). In figure 2.2 and 2.3 the same abs metal monitor but with the max of the colorbar scale set respectevely at 1 and 2 orders of magnitude lower than figure 2.

Issue n2:

I have a issue also with the absorption in the silicon in the first tens of nm. It seems like the mesh is not able to accomodate a so strong field gradient, and this does not change with a more finer mesh (0.5nm), figure 3.

After every simulations, I performed some sanity checks:

- A+R+T=1, ok

- Doubling the mesh, ok

- Auto-shutoff ending, ok

I attached my simulation file,

I will appreaciate any helps.




  • kjohnsonkjohnson Posts: 203Ansys Employee

    Hello @fbk ,

    The fields will vary rapidly near the surfaces or corners of metals, so very high mesh density is often required in these areas. I would recommend that you use convergence testing to see if the results (A, R and T) are changing as you reduce the mesh size. While these are high field values, they don't cover a large area so it's possible they won't drastically affect the results. You should also use an integer number of mesh cells over the metal object to ensure the mesh cells overlap with the metals.

    Alternatively, you could use the DGTD solver for this simulation. It is well suited for plasmonic simulations like this. You might also consider replacing the metal rectangle with a rectangle with rounded corners. This should be more accurate, as the actual device won't have perfectly sharp corners.

    Please try that and let me know if it helps.

Sign In or Register to comment.