Identification of Martensite in Steel by EBSD

Introduction

One of the most topical challenges within the field of microstructure analysis is the identification of martensite in steels. While the identification of martensite platelets in an austenitic matrix is quite straightforward by EBSD, the discrimination between ferritic and martensitic phases with this technique is more difficult. Latest advances by Oxford Instruments have resulted in much enhanced reliability. In this study, a 1.8715-steel was analysed using combined EBSD/EDS-analysis with the aim of identifying martensite and retained austenite regions in a ferrite matrix.

Identification of Martensite in Steel by EBSD

One of the most topical challenges within the field of microstructure analysis is the identification of martensite in steels. While the identification of martensite platelets in an austenitic matrix is quite straightforward by means of Electron Backscatter Diffraction (EBSD) due to the difference in crystallography (body centered tetragonal, as opposed to face centered cubic), the discrimination between ferritic and martensitic phases with this technique is more difficult. The crystal structure of martensite is very similar to that of ferrite; depending on the exact level of super-saturation with carbon (roughly between 10-100 times higher in austenite and martensite), the body centered cubic ferrite is being tetragonally distorted to about 5-8 % along the c-axis and contracted less than 1 % along the a-axes. This very slight difference therefore often gives rise to a high level of misindexing between the two phases, or makes discrimination impossible.

Latest Oxford Instruments advances in the sensitivity of the NordlysNano EBSD camera and the solver algorithm within the AZtec® suite of microanalysis software, have resulted in much enhanced reliability. Furthermore, the latest generations of X-Max® large area Silicon Drift Detectors (SDD) in combination with AZtecSynergy allow simultaneous X-ray chemical mapping. The differences in the carbon levels can be detected due to the high sensitivity and resolution in the low energy range of the X-Max. In addition, regions of higher concentration of substitutional austenite stabilising elements, such as nickel and manganese can be identified. This information can be used to confirm the EBSD indexing performance.

Sample Preparation and Experimental Conditions

In this study, a 1.8715-steel was analysed using combined EBSD/EDS analysis with the aim of identifying martensite and retained austenite regions in a ferrite matrix. The sample was mechanically polished using metallographic standard procedures with a colloidal silica finish. Subsequently, the sample was cleaned in an ultrasonic bath — first with methanol and then with distilled water, followed by blow drying. For this application, the quality of the sample preparation is key to the accuracy of the data collected. If the sample surface is not suitable, then the quality of the data collected will be impaired.

The sample was then analysed in a Zeiss Ultra55 FEG/SEM microscope at 15 kV accelerating voltage and 60 micron aperture size. For the coupled EBSD/EDS analysis, a NordlysNano EBSD detector and the X-Max 80 large area SDD were used in combination with the AZtec microanalysis suite. The EBSD camera was run in the 4×4 pixel binning mode at 10 ms exposure time and 2 frames averaging, while process time 4 was chosen for the EDS analysis. The step size was 30 nm.

Results and Discussion

Figure 1a shows the EBSD-phase map of the analysed material. Band contrast (BC) and band slope (BS) maps can also be used in order to discriminate between martensite and ferrite. Due to the higher lattice distortions in the martensite, the band contrast in the Electron Backscatter Patterns (EBSPs) is less steep and appears darker in the maps. Phase discrimination can then be performed by setting a filter with grey level threshold values. This is shown in Figure 1b, where the martensite phase was filtered out using a grey level threshold value of 60 (8-bit image with 256 grey levels). The discrimination between ferrite and austenite is usually not possible this way as they show similar BC and BS-levels.

The austenite phase is shown in Figure 1c, where it is overlayed on the EBSD-band contrast map. Figures 1d and 1e show the X-ray maps with the regions of highest carbon and manganese concentration respectively.

Fig. 1a) EBSD-Phase Map.

Fig. 1b) Martensite phase after BC-thresholding.

Fig.1c) EBSD-BC-map with austenite phase map overlayed. Red circles indicate regions indexed as austenite that coincide with regions of maximum Mn-contents.

Fig. 1b) Martensite phase after BC-thresholding.

Fig. 1e. Carbon X-ray map overlaying EBSD-BC Map.

Fig. 1f. EBSD-All Euler Map

From Figures 1a–e, it is evident that there is a good correlation between the X-ray elemental maps and the indexing of the EBSD-phase map. Regions indexed as martensite and austenite show higher carbon concentrations in general, which is in good agreement with the fact that the carbon (super)-saturation is much higher in both these phases than in the ferrite phase. Manganese shows maximum concentrations in the areas indexed as austenite, as one would expect from the theory. Also, the regions with low band contrast values show excellent correlation with the regions indexed as martensite/regions of highest carbon concentration.

Summary

This application note illustrates the potential to discriminate between phases in a steel sample. This is achieved using both the improved NordlysNano EBSD detector and the new solver algorithm in the AZtec microanalysis software, but also takes advantage of the low energy sensitivity and high count rate inherent to the X-Max 80 large area X-ray detector. The power of AZtecSynergy was shown on one of the most topical and challenging cases in the field of microstructural characterisation: a ferritic/martensitic steel with low fractions of retained austenite.