The Beam Propagation Method is usually not useful for simulation of surface roughness, owing to the spatial frequency of the roughness. Small variations in the surface are only significant to the light waves when the variations occur at a frequency that is on the same scale as the optical wavelength, or shorter. BPM, being a paraxial method, ignores those effects and concentrates on the effects of longer term variations. In BPM simulations, the structure is sampled at various points along the propagation. The typical distance between those points is often several wavelengths, and so sub-wavelength variations do not even enter the model.
Therefore we can’t recommend BPM for analysis of surface roughness effects. On the other hand, in the FDTD technique, the waveguide is sampled at a rate of a tenth of optical wavelength in all directions. Surface roughness will be entered and is properly simulated. Therefore we recommend OptiFDTD for this job.
.
