Far-field intensity distribution of a VCSEL

I want to calculate far-field of a VCSEL. But it seems that the results are not consistent. The results are changing with the simulation region. For diameters bigger than 2um results are quite good. However, for smaller diameters results are changing with the simulation region, actually, it’s the monitor size that affects. Since it crosses the simulation region the monitor size is equal to the simulation region.
I have tried 2 cases. One has 3.5um x 3.5um simulation region and the other one has 8um x 8um
Spectrums are very similar in both cases;

But the far-field distributions are different. The figure is normalized to 1 to make it clear to see the difference.

These are the sum of 500 frequency points from obtained from far-field monitor. I calculated them by using the following code.
div_angle.lsf (447 Bytes)

If we look at a single frequency, these are the distributions of peak values in the spectrums;

And these are the simulations files;
vcsel_3.5um.fsp (426.1 KB)
vcsel_8um.fsp (426.5 KB)

Thanks in advance.

Hello @e248299,

When calculating the far field, the monitor assumes that there is no light passing through the plane of the monitor beyond the edges of the monitor. If this is not the case, then increasing the span of the monitor will change the results.

Ideally the monitor would be infinite, but because this is not possible with FDTD it is best to keep expanding the monitor (and the FDTD region) until the far field results start to converge. It may also help to use spatial filtering to avoid truncating the field at the edges of the monitor, as described on this page:

I hope this helps. Let me know if you have any questions.