Correlating results from high resolution EBSD with TEM- and ECCI-based dislocation microscopy: Approaching single dislocation sensitivity via noise reduction

Identification of Star Defects in Gallium Nitride with HREBSD and ECCI

“…A previous study (Ruggles et al, 2020) showed similar good agreement for the Burgers vector and line direction at much smaller length scales between ECCI/TEM and HREBSD measurements. However, quantitative agreement in that study was poor, and it was speculated that this was due to a number of issues related to the small length scale of the EBSD scan, including but not limited to the presence of strain gradients and the high aspect ratio of the interaction volume for EBSD.…”

“…An EBSD scan may be used to generate maps of the dislocation density of each dislocation type at a continuum level. HREBSD dislocation analysis provides complimentary data to ECCI (Ram et al, 2016; Vilalta-Clemente et al, 2017; Ruggles et al, 2020). This method is broken down here into three steps: Use image correlation on EBSD patterns to measure the relative lattice distortion between points on a scan, Calculate Nye's tensor, a compact tensorial representation of dislocation content, with the measured lattice distortions, And finally, use Nye's tensor to determine dislocation densities for each combination of Burgers vector and line direction. …”

“…Cross-correlation measures relative lattice distortions () between diffraction patterns with an approximately 200 microstrain, or 1 part in 5,000, accuracy when the relative distortions are small. Here, a noise reduction strategy (enforced antisymmetry) is employed that takes advantage of the fact that strain part of these distortions is negligible compared to the rotations (Ruggles et al, 2020). Gaussian smoothing (using a kernel with a standard deviation of 0.8) was also performed on the distortion maps.…”

“…Because of the high number of unknowns in some crystals, equation (3) often fails to distinguish between dislocation types. The conventional Nye–Kröner method has the most success when characterizing FCC crystals, which have very few dislocation types, and when dislocation configurations are simple (Kysar et al, 2010; Ruggles et al, 2016 a , 2020; Dunlap et al, 2018). When only a single dislocation type is present, simulations have shown that equation (3) can recover the correct dislocation type in FCC crystals (Ruggles et al, 2016 a ).…”

“…In general, a shorter step size will result in more noise and blurring, but a larger step size will result in cancellation of some dislocations and of course lower spatial resolution. Despite these complications, HREBSD can successfully recover Burgers vector and line direction at very large (Ruggles et al, 2016 a ) and very small (Ruggles et al, 2020) step sizes, although the magnitude of the measured dislocation density is likely effected. For this study, two different step sizes are used, 1 m for a scan of the whole defect, and 300 nm for a smaller scan to do quantitative comparison between ECCI and HREBSD.…”