Discussion Forum

Update
%%%%%%

Dear All,

I have simulated the following model 'Optical Scattering by Gold Nanospheres' using RF Module. As mentioned in the model, you can solve for 1/4 of the geometry due to electromagnetic and geometric symmetry. My goal is to solve for Maxwell Stress Tensor on the sphere.

I am expecting to get the exact answer if I solve for 1/2 of geometry or for the full geometry (providing that the boundary condition is right). However, the MST differs in orders of magnitudes (see table below). I checked the boundary conditions, but everything seem OK.

I set the far field boundary condition to satisfy the symmetry in both cases (1/4 and 1/2 of the geometry).
These are the values I got for MST:
emw.unTx emw.unTy emw.unTz
1/4 of the geometry 3.9207E-13 3.2820E-15 1.4930E-13
1/2 of the geometry 3.9200E-13 3.2287E-15 -2.4217E-17
1 of the geometry 3.9201E-13 -6.2472E-18 1.1781E-19

Note that the correct answer should be the one with almost zero force in the transverse directions of the propagation direction (x-axis). Note also that in all three cases, the x-axis force did not change, while the y and z component did. I realized that the boundary (PMC or PEC) enhances the force by orders of magnitude. For example, for only 1/2 of the geometry, I remove (PEC) and I have symmetry at y=0 plane, and the z-force goes down by 4 orders of magnitude. For full geometry, y and z forces go down by orders of magnitude.

My question is that why does this happen? Did I missed certain boundary condition? Or is it meshing quality problem?

The file is attached!

Many thanks.