Rapid Characterisation of Steel and Ni

Introduction

The groundbreaking Symmetry CMOS-based EBSD detector, together with the powerful AZtec® software, is capable of acquiring EBSD and EDS data at speeds in excess of 3000 indexed patterns per second (pps). These speeds, twice as fast as those achievable with conventional CCD-based detectors, are further enhanced by the fact that extreme pixel-binning of the diffraction patterns is not necessary. This means that the quality of the data collected using Symmetry is significantly better than with CCD-based detectors, both in terms of hit rate and also angular accuracy.

In this application note we give examples of how Symmetry can be used to characterise the microstructures in two example metals at high speeds, providing reliable statistics in affordable timescales.

Results

Two different samples were analysed using the Symmetry detector on a field emission gun SEM: a relatively simple sample – a partially recrystallised Ni-based superalloy and a large scale weld structure in a duplex steel. Both samples were mechanically polished down to a final stage using colloidal silica prior to analysis. All measurements were carried out using the AZtec EBSD-EDS software, with an EBSP pattern resolution of 156 x 128 pixels.

Ni-superalloy

This sample was a deformed and partially recrystallised Ni-based superalloy. An area of 835 by 592 μm was mapped at a step size of 0.25 μm, equating to almost 8 million analysis points in less than 45 minutes. The acquisition speed, with an indexing hit rate in excess of 99%, was just over 3000 patterns per second; this has allowed the analysis to cover a large area including numerous deformed grains, whilst characterising the recrystallised grains in superb detail.

Fig. 1a shows the mapped area using the grain relative orientation distribution (GROD) colouring scheme, highlighting the deformed grains in the matrix of recrystallised grains. This type of plot allows the separation of deformed and recrystallised areas for further examination. In this case the boundary misorientation distributions have been plotted (Fig. 1b), showing the prevalence of Σ3 twin boundaries (60° rotation about <111> axes) in the recrystallised area, and low angle boundaries in the deformed area.

Fig.1a. Grain relative orientation distribution (GROD) map, showing numerous deformed grains (brighter colours) in a recrystallised matrix. Σ3 twin boundaries are marked by red lines.

Fig.1b. Boundary misorientation frequency distribution, separated into the deformed grains and recrystallised grains.

Fig.1c. Forescatter detector orientation image of the analysis area

Fig.1d. Inverse pole figure orientation map showing grain boundaries in black and twin boundaries in red

Weld

In this example, a welded sample of duplex stainless steel was examined using Large Area Mapping. The challenge was to examine a wide transect (~5 mm across) yet to retain sufficient resolution to measure accurately the small grains in both the weld itself and the base metal. The area was mapped with a 0.25 μm step size, at ~2000 indexed patterns per second. In total 96 maps were collected, totalling more than 45 million points with a 98.5% hit rate. The resulting phase map (Fig. 2a) shows clearly the variation from the base metal on the right to the fusion zone on the left. This large area, high resolution scan allows a detailed study of the textures, grain sizes and phase distributions across the different zones of the weld.

Despite the obvious change in the microstructure, there is little change in the phase fraction between the base metal and the fusion zone (a ferrite to austenite ratio of 55:45 compared to 54:46 respectively). However, in the heat affected zone (HAZ), the phase fractions change significantly to 71% ferrite and 29% austenite.

Although the fusion zone appears to be significantly coarser grained, the full scale images shown in Figs. 2b and 2c show the presence of many very small austenite grains, <5 μm diameter. These reduce the mean grain size of the austenite in the fusion zone to only 6.3 μm, significantly less than the 9.5 μm mean in the base metal (both figures based on more than 20,000 grains).

The large area analysis also enables a statistically valid study of the variation in texture across this weld margin. The contoured pole figures in Fig. 2d and 2e show the change in texture in the ferrite between the base metal and the fusion zone. Within the base metal there is a fairly strong alignment of {100} poles normal to the rolling direction, but little clear texture is present in the fusion zone, with the higher densities resulting from a few very large grains.

Fig. 2a. Phase map across the weld margin in a duplex steel. Red – ferrite, blue – austenite. The fusion zone (weld) is to the left of the map, the base metal to the right. The yellow box marks the area highlighted below.

Fig. 2b. Phase map of the area marked in (a) in higher detail. Red – ferrite, blue – austenite.

Fig. 2c. Orientation map of the area marked in (a) in higher detail (inverse pole figure colour scheme, with grain boundaries in black).

Fig. 2d. Contoured pole figures ({100}, {110} and {111}) for ferrite within the fusion zone.

Fig. 2e. Contoured pole figures ({100}, {110} and {111}) for ferrite within the base metal.

Conclusion

These two contrasting examples demonstrate how the AZtec EBSD software in combination with the Symmetry EBSD detector is an ideal combination for the rapid and effective characterisation of a range of metal and alloy samples. In single phase metal samples, indexing speeds in excess of 3000 pps are achieved with no compromise on data quality or indexing hit rate, whereas in dual-phase materials speeds in excess of 2000 pps are commonplace. These speeds are achieved without extraordinary beam currents: typically ~10 nA is sufficient to achieve excellent pattern quality and indexing at 3000 pps.

These high speeds and associated, uncompromised data quality makes Symmetry an extremely powerful EBSD detector for grain size, texture and phase characterisation of all but the most complex samples. What took hours or even days with conventional CCD-based detectors can now be completed in a matter of minutes, expanding the appeal of EBSD as a routine characterisation tool.