Kernel Average Misorientation Analysis of EBSD Patterns

This collection of MATLAB scripts was developed to perform kernel-average misorientation analysis of raw EBSD patterns. Misorientation mapping can help better understand subgrain structure and processes such as dynamic recrystallization and pseudo-strain measurements. This code is currently designed to handle cubic structure and pure samples.

If you are unfamiliar with the math behind crystallographic misorientation analysis, the background math section may be helpful.

The script main_kernel_calc.m controls the entirity of the analysis. The following provided functions must be in the same directory as the main script to run:

• ebsd_parse.m
• euler_convert.m
• euler2rotationmatrix.m
• KAM_calc.m
• KAM_clean.m
• plot_KAM.m

Varying EBSD detectors/softwares may vary in the structure of their raw EBSD data output. This script expects a .txt with a format resembling that given in the example file Ni.txt.

#Phase0:
#Name: Nickel (New)
#Spacegroup: F m#ovl3m
#A: 3.524
#B: 3.524
#C: 3.524
#Alpha: 9E1
#Beta: 9E1
#Gamma: 9E1
#Orientations:
#Index Phase x y phi1 PHI phi2 Bands BC GrainIndex
0 1 0 0 1.434201877E2 2.390389414E1 2.467704713E2 12 163 -1
1 1 1 0 1.43549729E2 2.422165518E1 2.469476464E2 12 177 -1
2 1 2 0 1.439342243E2 2.405383045E1 2.465798271E2 12 180 -1
3 1 3 0 1.445457072E2 2.380896362E1 2.465767945E2 12 180 -1
4 1 4 0 1.445333411E2 2.394108922E1 2.46444562E2 11 178 -1
5 1 5 0 1.442472109E2 2.368202045E1 2.463594578E2 11 185 -1
6 1 6 0 1.437432727E2 2.346567406E1 2.471365148E2 10 162 -1
7 1 7 0 1.443801434E2 2.353161058E1 2.468962834E2 11 135 -1

...etc

However the most import feature of the raw EBSD file is the phi, PHI, and phi2 values for each pixel with respect to its index number. If the raw data for each pixel can be extracted (or provided), the rest of the table can be user generated.

Note/ Given the number of nearest neighbors a rectangular kernel will be constructed. For example, if neighbors=3 the kernel considered during each misorientation calculation will be a 10x10 grid surrounding the point/pixel of interest. Kernel geometry can be modified by modifying KAM_calc.m.

To run the calculation the user must define the number of nearest neighbors to consider in main_kernel_calc.m. Depending on the power of your computer, it is recommended that you start with 1-3 neighbors because the calculation can be time consuming - though it will readily run in the background and time estimates to completion are provided.

The user is then prompted to select the .txt file that contains the EBSD data.

main_kernel_calc
Parsing the EBSD data...
Converting to orientation matricies...
Calculating the nearest neighbor misorientation...
3.22% complete - estimated 0h 26m remaining (0m elapsed)
6.45% complete - estimated 0h 26m remaining (1m elapsed)
9.67% complete - estimated 0h 26m remaining (2m elapsed)
12.89% complete - estimated 0h 25m remaining (3m elapsed)

(25 minutes later)

90.26% complete - estimated 0h 2m remaining (25m elapsed)
93.49% complete - estimated 0h 1m remaining (26m elapsed)
96.71% complete - estimated 0h 0m remaining (27m elapsed)
99.94% complete - estimated 0h 0m remaining (28m elapsed)
Cleaning data...
Plotting!