Investigation on the traceability of three dimensional scanning electron microscope measurements based on the stereo-pair technique

An investigation was carried out concerning the traceability of dimensional measurements performed with the scanning electron microscope (SEM) using reconstruction of surface topography through stereo-photogrammetry. A theoretical description of the effects that the main instrumental variables and measurement parameters have on the reconstruction accuracy of any point on the surface of the object being imaged was undertaken. This description was based on the model function introduced by Piazzesi adapted for eucentrically tilted stereopairs. An analysis of the sensitivity coefficients for the vertical elevation model has identified the most important instrumental variables and measurement parameters to be the tilt angle and the magnification calibration. An equation describing the theoretically determined uncertainty of the vertical elevation was produced. An experimental investigation was carried out involving: (i) magnification calibration using a steel scale with 50 mu m pitch at low magnifications, and a 2D calibration grating with 2.120 mu m pitch at high magnifications; (ii) tilt angle calibration by laser interferometry; (iii) vertical elevation calibration using gauge-block steps; (iv) calibration in the vertical plane using two ISO 5436 type C roughness standards. Results from magnification calibration and tilt angle measurement were used for calculating the theoretical uncertainty on the vertical elevation. Experimental deviations measured on gauge-block steps showed slightly bigger values than those calculated from the theoretical model. Lateral measurement on the reconstructed three-dimensional topography of the type C roughness standards showed good agreement with the nominal profile wavelength values. (c) 2004 Elsevier Inc. All rights reserved.