The assessment of fractal dimensions with the direct use of scanning electron microscopy (SEM) images from blast-cleaned steel substrates

Abstract

Morphological parameters of substrate surfaces determine the adhesion of coatings and pressure-sensitive adhesives. This paper is concerned with the assessment of fractal parameters by means of the direct use of scanning microscope images (SEM) of blast-cleaned substrates. Mild steel samples were prepared with different abrasive materials. Images were taken from a total of 12 surface configurations with a scanning electron microscope. The fractal dimensions of the surfaces were estimated on four different image sizes with and without the use of a median filter by means of the box-counting method. All tests were performed twice to evaluate accuracy and repeatability of the assessment procedure. Design of Experiment (DoE) was applied in order to statistically analyze the relationships. The factors considered included abrasive type, surface preparation grade and surface roughness. Independently on filter and image size, no meaningful regression model could be established. Neither main effects nor interactive effects of the factors on the fractal dimensions could be identified. The deviations in the fractal dimension were larger for two measurements on a given factor combination than the deviations for two measurements at different factor combinations. The direct use of SEM images is, therefore, not practical for the estimation of fractal dimensions of blast-cleaned steel substrates.