As part of the job to do, analysing grains is important for both product engineering and quality-control. Mechanical properties e.g. tensile strength and formability, improve as grain size decreases. Therefore, material processing must be carefully controlled to obtain the desired grain size.
To accurately measure grain size, it is imperative that all of the grain boundaries are detected. Therefore the technique must produce the highest degree of grain boundary delineation.
Traditionally grain size is measured by analysing images collected by a light optical microscope. It requires a chemical etching of the surface in order to highlight the grain boundaries. In addition to the challenge of decreasing grain size there are other limitations to this technique.
The limitations of image analysis techniques are shown here.
The image at the top is an optical micrograph of a rolled galvanized steel. This material has a complex microstructure and boundaries, and it is not possible to accurately identify the individual grains in the LOM image.
The image at the bottom, from the same sample with a similar field of view, is a secondary electron image from a SEM. Here neither the grains nor grain boundaries are well defined.
EBSD is an ideal technique for these complex microstructures including microstructural characterisation and determining grain size.
Looking at the same sample and collecting EBSD data we can accurately detect the grain boundaries by measuring crystallographic orientations and this overcomes the limitations of image analysis techniques.
The upper image first shows the distribution of the grain boundaries as identified by EBSD. By identifying these boundaries the individual grains are located.
The grain map here shows each grain coloured differently. This map indicates the range of grain size in the material and from this a statistical distribution grain size measurements is derived.
It is clear that with EBSD we can easily and accurately identify the grains in the sample.
In addition to analysing a single field of view, the analysis can be automated to cover a large area, delivering reliable statistically valid data. This example shows the same steel, where a quarter of the rod has been analyzed. This area has a radius of 6mm. In this example covering this large area was significant, as trends in grain structure from the centre to edge of the rod were examined.
Back to Automotive & Aerospace