Comparison of Michelson and Linnik interference microscopes with respect to measurement capabilities and adjustment efforts

Details zur Publikation
Kühnhold, P.; Lehmann, P.; Xie, W.
Peter H. Lehmann, Wolfgang Osten, Armando Albertazzi
Verlagsort / Veröffentlichungsort:
Titel der Buchreihe:
Optical Measurement Systems for Industrial Inspection

Zusammenfassung, Abstract
Scanning white-light interferometry (SWLI) provides the capability of fast and high-precision three-dimensional measurement of surface topography. The Mirau configuration exhibits a compact design and a high stability. Therefore, this configuration is actually mostly used in industrial applications. However, white-light interferometers can also be set up according to the Linnik or the Michelson configuration. This Paper is intended to analyze and compare Michelson and Linnik interferometers, especially in view of industrial applications of surface topography measurement. A Linnik interferometer consists of two conventional microscope objectives which are typically combined with a beam splitter cube and an additional tube lens in order to image the surface of the measuring object and to generate interference patterns. Because there are no optical elements in front of the objectives, no further correction of aberrations is necessary. However, the disadvantage of the Linnik configuration is the high effort which is needed for the mechanical precision and the alignment procedure. Especially the beam splitter cube and the two objectives must be of high optical quality and have to be aligned very precisely. A rather small misalignment of one of these components leads to dispersion effects and influences the measurement results. In contrast, Michelson interferometers use only one objective lens and are rather easy to align. Commonly this kind of interferometer needs only one degree of freedom of adjustment. Also the mechanics requires less accuracy compared to the Linnik configuration. Besides all these advantages in a Michelson interferometer the beam splitter cube has to be placed between the objective and the specimen. This reduces the overall working distance of the interferometer substantially, so that long working distance microscope objectives are generally necessary. As a consequence, these objective lenses show rather low magnification and low numerical aperture. The second negative effect of the beam splitter cube in a Michelson interferometer is that the light rays do not transmit the beam splitter cube in parallel. This leads to dispersive effects and the optical correction of the objective is no longer valid. As it is shown discrepancies arise between the position of the envelope of a SWLI signal compared to its phase. However, these effects can be avoided by combining the beam splitter cube with additional correction lenses. This is demonstrated by optics design computations based on objective lenses that are assumed to work perfect in absence of the beam splitter cube. Our contribution presents comparative measurement results obtained with a well aligned Linnik and a Michelson interferometer using the same artifacts, microscope objectives and tube lens. A measured sinusoidal standard of 100 µm cycle length and a peak-to-valley amplitude of 1 µm shows that it is possible to reduce dispersion influences in a Michelson interferometer significantly, so that results of the same accuracy as by using a Linnik interferometer are reached.

Autor(inn)en / Herausgeber(innen)

Zuletzt aktualisiert 2019-01-11 um 16:04