Calibration of crystal orientation and pattern center of EBSD using integrated digital image correlation

Improved EBSD indexation accuracy by considering energy distribution of diffraction patterns

“…There could exist unknown errors in all these aspects, and one remedy is to use the calibrated PC coordinates to correct all these potential sources of errors. To calibrate the sample surface plane more accurately, a large EBSD scan with high-definition EBSPs is preferred, the roughness of sample surface should be as low as possible, and ideally the sample should be unstrained, as the strain state influences the calibrated PC coordinates [11,36,18]. These requirements converge to a large-area EBSD scan with full definition on an unstrained single crystal sample, for example the single crystal Si wafer.…”

“…Then a registration procedure similar to the IDIC algorithm described in Ref. [18] is proposed on the combined images with limited changes. To distinguish from the normal IDIC EBSD, the method described in the current paper is named as IDIC-E EBSD (Integrated Digital Image Correlation with Energy weighting).…”

“…Dynamical EBSP simulation proves to have a higher similarity with experimental pattern than kinematic simulations [12]. To achieve the optimization of the projection parameters, many algorithms have been suggested, such as a Nelder-Mead algorithm [13,14,15], cross-correlation with log-polar transform [16], evolution algorithm [17], SNOBFIT algorithm [15], integrated DIC algorithm [18] or non-disclosed ones used in commercial software [19]. IDIC is able to calibrate overlapped EBSPs at grain boundaries, to obtain multiple sets of Euler angles, one set of PC and the contribution ratio for each crystal orientation [20].…”

“…However, different groups studying FPM EBSD calibration on polycrystal samples have very recently reported grainwise discontinuity in PC fields [15,18]. The shift of PC components across grain boundaries could reach 1-2 pixels, that cannot be due to the sample topography nor to the elastic strain.…”

Kikuchi pattern simulations of backscattered and transmitted electrons

“…We demonstrate that systematic errors in the order of at least 1% in the relative position of the projection centre can be introduced by not including E/D effects in the simulations, which has corresponding effects for a quantitative determination of lattice parameters and strain states via simulated Kikuchi patterns. Several existing estimations of the absolute precision of pattern matching approaches for projection centre and strain determination have been based on analyses of simulated synthetic data without E/D effects, 19,30,73 which thus potentially introduces a corresponding bias. Finally, we will show that the refined simulation model is able to reproduce contrast formation under various different conditions which are relevant for both SEM as well as TEM Kikuchi diffraction patterns.…”