For interferometers, in most cases, in ideal operation, the beamsplitter works from a single surface with an AR coating applied to the other surface. For both the Michelson and Fabry-Perot and others, they are most readily explained quantitatively by the Fresnel coefficients including the ## \pi ## phase change (essentially a minus sign) that results upon reflection off of the higher index surface. The Fabry-Perot interferometer works from two reflective surfaces, but similar interference results can occur from two beams incident on a single interface from opposite directions and can be readily computed using the Fresnel coefficients. One question that often arises with these interferometers: The Fresnel coefficients always apply for these systems, whereas energy reflection coefficients can work for a single beam but do not work in the case of two incident (coherent) beams because the system is completely linear w.r.t. the electromagnetic fields but not to the energy which is proportional to the second power of the electromagnetic field amplitude