Volume 77, Issue 11 -

A robust common-path interferometer, based on an interferometer according to Bath and Gates, is presented. In this setup, the distance of both the reference beam and the measuring beam can be varied by offsetting a beamsplitter cube laterally. Thus, an active positioning of the specimen is not required for measuring cross-sections. The measurable depth range and the minimal measurable size of microstructures can be adjusted by changing the focusing optics.

The increase of integration density in semiconductor industry demands the ability to measure structures with subnanometer resolution. Quantitative estimation of geometry parameters of edges by proximal probe microscopy forms a basis for CD measurements with various optical and electron-microscopic procedures. This article describes a procedure for sampling edges with single point probing. Main feature of this strategy is the approaching of nanosized structures from directions according to the surface slope in order to estimate slope, foot, and bulge of edges.

For the measurement of components geometries using fringe projection the sensor has to be calibrated. After the calibration procedure the sensor behaviour is described through a blackbox- or a physical model. This article describes first the model of the fringe projection system by means of the projective geometry in detail, with taking into account the object-sided telecentricity of the camera lens. Afterwards the calibration strategy for the identification of all system parameters is presented. The sensor model identified will be implemented as a numerical simulation into a computer program. Furthermore the simulation is combined with Monte-Carlo-methods to calculate measurement uncertainties. The simulation is planned to be used for the influence-analysis of the system parameters onto the overall measurement uncertainty of the system.

The analysis of tribological behaviour often requires the detection of worn areas on contact surfaces. Innovative optical methods allow to measure the surface topography in-plane with high resolution. Worn areas with low degree of wear, however, cannot be detected directly. This paper presents a new method for space-resolved 3D-roughness analysis that enables the detection of areas with very low degree of wear.

The article describes three procedures to estimate the task specific uncertainty of coordinate measurements according to the GUM, the internationally acknowledged “Guide to the Expression of Uncertainty in Measurement”: Numerical simulation, calibrated workpieces, and uncertainty budget. It compares three approaches to realize these procedures at workpieces with regular geometric features and discusses possible applications.

Computed tomography (CT) can be used for dimensional measurement. Reliable results require an effective artefact correction. In two earlier papers a Synchronous Artefact Correction (German: Einsatzsynchrone Artefakt-Reduktion EAR ) was introduced. In this paper some further developments of EAR are given though the key concept isn´t altered. It is still two-staged with a separation in fine and coarse registration, the latter is subdivided into translation and rotation. For the coarse rotation the idea of the Overall Cachment Area is applied. In-depth tests favoured “Normalized Mutual Information” that was only second-best in the prime research. Another topic is the analysis of the accuracy of the registration result concerning the applied similarity measure. In this test “Normalized Mutual Information” was figured out to be the measure with the highest accuracy. Zusammenfassung Computertomografie (CT) kann als Messmittel für das dimensionelle Messen genutzt werden. Um aussagekräftige Messergebnisse zu erhalten, ist es notwendig, im Rekonstruktionsvolumen entstehende Artefakte z. B. mit der IAR (Iterative Artefact Reduction) effektiv zu korrigieren. Die hier präsentierte Arbeit stellt mit der EAR eine Korrekturmethode vor, die auf der IAR aufbaut. Mit der EAR ist es möglich, die Korrekturkennlinie schon während der Datenaufnahme und damit deutlich schneller als noch bei der IAR zu erstellen. Die Korrekturmethode EAR benötigt ein CAD-Volumen des zu untersuchenden Bauteiles, welches während der Datenaufnahme registriert werden muss. Für diese Registrierung werden in dieser Arbeit Ähnlichkeitsmaße in Bezug auf ihre Registrierungseigenschaften verglichen.