Segmentation of 3D EBSD data for subgrain boundary identification and feature characterization.

Area of Science:

• Materials Science
• Crystallography
• Computational Materials Science

Background:

• Subgrain structures in plastically deformed metals are crucial for understanding material properties.
• Automatic identification of these structures using electron backscatter diffraction (EBSD) data is challenging due to subtle orientation variations.

Purpose of the Study:

• To adapt a 2D image segmentation technique, fast multiscale clustering (FMC), for automated analysis of 3D EBSD data.
• To develop a robust method for segmenting and grouping subgrain boundaries based on microstructural features.

Main Methods:

• Adaptation of the fast multiscale clustering (FMC) algorithm for 3D EBSD data, incorporating a novel variance function for quaternion data.
• Modification of FMC to group segmented boundaries into coherent surfaces using local surface normals.
• Integration of the enhanced FMC technique into the open-source MTEX software package.

Main Results:

• The adapted FMC technique successfully segments subtle and gradual orientation variations, as well as sharp boundaries in 3D EBSD data.
• The method effectively groups segmented boundaries into coherent surfaces, enabling detailed microstructural analysis.
• Demonstrated capabilities on diverse materials including aluminum, steel, and nickel, revealing microbands and varying boundary misorientations.

Conclusions:

• The adapted FMC technique provides a powerful and automated solution for identifying subgrain structures in 3D EBSD data.
• This advancement facilitates more accurate and efficient characterization of microstructural evolution during plastic deformation.
• The open-source implementation in MTEX promotes wider accessibility and application in materials research.