On the Representativeness of Automated SEM/EDS Analyses for Inclusion Characterisation with Special Regard to the Measured Sample Area

Abstract

The cleanness of steel has emerged to an important quality criterion for a wide field of special applications. In order to ensure the reliable characterisation of non-metallic inclusions, the constant optimisation of current analysing methods is essential. The present study focuses on the automated SEM/EDS analysis, especially determining the influence of the analysed sample area on the obtained particle diameters and size distributions of non-metallic inclusions. Besides the experimental analysis of different area sizes, a model was formulated to estimate the error of area ratio as a function of inclusion content. For the assumed conditions, an area range combining a sufficient and significant area size with a feasible time effort is defined. The model also evaluates the significance of the results for the maximum and medium particle diameter gained out of the measurements. Moreover, the truncation of the data – resulting from an experimental analysing limit – and its consequences are discussed. The results underline the potential of this analysing method for inclusion characterisation but also demonstrate the limitations especially regarding the maximum particle diameter. This work contributes to a better understanding in terms of representativeness of automated SEM/EDS analysis and provides important information for further practical analyses.