The principle of Mach-Zehnder interferometer can be described in reference with the following figure.
The half-silvered mirrors split the beam into two, L and R. Counting the total phase shift due to multiple reflections and refractions at the mirrors (M1 & M2) and splitters (HS1 & HS2), the beam L and R enters the photodiode D1 in phase, therefore, undergo a constructive interference. On the other hand, the beams entering photodiode D2 are out of phase by l/2, therefore, suffers a destructive interference and no light enters D2. Thus D1 "sees" all the incident beam. This situation occurs only for perfect mirrors.
If one of the mirrors is replaced by a reflecting surface (i.e., a sample), then the phase shift due to the specimen alters the phase relationships between the two beams, resulting into a partial interference (not completely destructive) at photodiode D2. The amount of light that enters D2 is thus proportional to the smoothness of the specimen surface. Therefore, by measuring the relative intensity of light entering D1 and D2, surface roughness of the specimen can be determined from know calibration. This technique can also be utilized in optical demultiplexing.
© 1999 Anis Rahman