Finally the amount of radiation escaping in the direction of observation is computed. To do so, the observation angle and the polarization must be specified. For each light source in the luminescent layer, the waves that end up in the finally observed direction are superimposed (with phase relations) taking into account Fresnels' equations and summing up all multiply reflected partial waves:

In the graph above, the multiply reflected partial waves within the top layer stack and the bottom layer stack are not drawn for simplicity. They are, however, included in the computation.

Squaring the electric field amplitudes the intensity for each point light source is computed. The contributions of all light sources are added up to the final value of the PL intensity. This is repeated for the whole spectral range (given by the settings of the internal efficiency). The result for the 500 nm porous silicon layer on silicon is this (400 nm excitation):

The comparison with the internal efficiency spectrum shows a shift of the emitted PL spectrum:

In contrast to an intuitive inspection of experimental spectra, in this case there is no doubt about the origin of the observed shift: Since the used internal efficiency is known and has a peak at 800 nm, any shift must be due to interference, reflection and re-absorption effects.