Transfer characteristics of white light interferometers and confocal microscopes

Optical profilers present significant advantages in comparison with stylus profilometry and scanning probe microscopy due to their non-contacting technique and their ability to scan the entire measurement field simultaneously. However, optical profilers suffer from different systematical artifacts. The dissertation investigates the transfer characteristics of optical profilometers, namely white light interferometers and confocal microscopes. In white light interferometry dispersion in the optical system leads to measurement errors in the envelope evaluation and consequently severe ghost steps take place in the phase evaluation, especially at locations with high surface slope at the margins of the field of view. The dispersion errors can be corrected either physically by applying additional lenses or numerically by model-based calculation. Overshooting, also called batwing, takes place commonly at rectangular or step height structures. The batwings may be reduced or even eliminated if certain conditions of the step height and the wavelength of the illumination are met. Besides the systematical effects mentioned above, other effects in white light interferometry are studied as well, such as overestimation of the roughness, polarization-dependent batwings, shadow effect for rectangular structures with high aspect ratios, etc. The occurrence or disappearance of the batwing effect also occurs in measurements of step height structures in confocal microscopy. However, the systematical artifacts differs from those in white light interferometry. The motivation of this work is to explain diverse systematical discrepancies of measurement results obtained by white light interferometers and confocal microscopes compared to the "real" profile of a sample through mathematical modeling, numerical simulation and experimental measurements, so that a better understanding of the relevant effects is obtained and device-related improvements as well as new measurement and evaluation strategies can be derived. Besides physical phenomena, such as diffraction, temporal and spatial coherence, the physical properties of both, the measuring instruments and the measurement objects are considered in the modeling. Three mathematical modeling approaches are presented in this work: Fourier optics, Kirchhoff scattering theory and Richards-Wolf modeling. Each of these models considers different properties of the measuring instrument and measurement object. Diverse measuring instruments including Michelson, Mirau and Linnik interferometers, as well as confocal microscopes are used for measurement. Simulation and measurement results are compared and analyzed in detail.