Effect of orientation noise on the determination of percolation thresholds from electron back-scatter pattern data

The effect of orientation noise in EBSP data on measurement of percolation threshold values has been investigated by use of computer simulated microstructures. A 2-D Monte Carlo Potts model run on a square lattice of size 200 x 200 was used to generate a microstructure containing approximately 150 grains. Orientations were then assigned to each of these grains to generate single texture component microstructures of differing texture tightness ("model" data). In order to simulate the effect of orientation noise on the experimental data, the orientation at each point in the 200 x 200 grid was adjusted in a manner consistent with experimental observations of the effect of orientation noise. The model data represent therefore an underlying real grain structure, and the noise-adjusted data represent the orientation map that would be measured using EBSP analysis at a given orientation noise level. The misorientation angle theta(70%) at which 70% of the grains were percolatively connected was then determined for both the ideal data and for the orientation noise adjusted data. A comparison of the two allows calibration of the extent to which percolation data may be incorrectly estimated by EBSP measurements.