GaussianSource bandwidth

[dalarev]

Hi, I am having trouble validating a known Gaussian propagation property: beam spread increases as wavelength decreases (provided you define a wavelength independent beam waist radius and focal distance).

My test setup includes a 2d vacuum model with a gaussian source (script is here). I define a fixed, frequency-independent waist radius and focal distance. I assign a fcen + fwidth that should cover the freqs assigned to the dft field monitor. However when I plot the field at each of the monitor freqs, it appears as though the beam diverges faster at longer wavelengths than shorter ones. See below.

I presume this is an issue related to the source bandwidth and monitor frequencies. I have attempted to widen the fwidth by large factors, but this doesn't seem to resolve the issue.

What am I missing? Any thoughts are appreciated.

[stevengj]

Hi, I am having trouble validating a known Gaussian propagation property: beam spread increases as wavelength decreases (provided you define a wavelength independent beam waist radius and focal distance).

That seems backwards to me. For a given beam waist, as the wavelength decreases, the beam spread (diffraction) decreases. The standard formula (in the paraxial approximation) is that the divergence angle is proportional to the wavelength divided by the beam diameter.

(In the limit of zero wavelength, you have ray optics, and there is no diffraction at all.)

This seems to match what you see.

[dalarev]

Oops - you're right. Probably easiest to see this in the Rayleigh range, which is inversely proportional to wavelength.

On the topic though, if fwidth is defined as fmax - fmin, is it necessary to apply some padding to ensure the fmin/fmax limits are valid for analyses? Thanks.