Meaning
Electromagnetic wave modeling relies on mathematical methods to solve the Helmholtz equation in the spatial frequency domain. The method of angular spectrum propagation models the evolution of an optical field between parallel planes by decomposing the wave into a sum of plane waves, each travelling in a unique direction. Each component undergoes a phase shift determined by its propagation distance and spatial frequency before the field is reconstructed.
This modeling stops at the boundary of non-homogeneous media where the refractive index varies spatially.
Numerical Computation
Fourier transform algorithms facilitate the transitions between the spatial domain and the frequency domain. In this computation, angular spectrum propagation utilizes two fast Fourier transforms to execute the field transfer efficiently. The first transform decomposes the complex amplitude of the starting wavefront, and the second reconstructs the field at the target plane.
Numerical accuracy depends directly on the stability of the transfer function.
Grid Resolution
Discretization of the continuous wave field introduces constraints on the grid spacing of the computational window. If the grid spacing is too large, high frequency wave components suffer from aliasing, which introduces non-physical artifacts into the simulated wavefront. Filters must be applied to suppress these evanescent waves or large angle components that exceed the propagation limit.
This suppression ensures that the computed field remains physical. The maximum propagation distance is therefore bound by the sampling interval and the wavelength of the radiation, preventing the usage of this method over infinite distances without grid expansion.
Sampling Limit
Propagation distance dictates the required sampling density to avoid aliasing in the transfer phase factor. When the propagation distance is very small, the phase varies slowly, but at larger distances, the phase oscillates rapidly across the grid. This rapid oscillation demands a finer grid or a restricted propagation step to maintain execution accuracy.
Thus, the technique applies most effectively to short and medium distances.