Caused by the high sample tilt required for EBSD measurements, the intensity profiles of bands become asymmetric. Horizontal bands are fully affected whereas vertical bands are practically not influenced (actually only the band which runs vertical through the pattern center). This should bias the band detection so that I would assume that the determined orientation cannot be exact. Does anybody know how this asymmetry affects the final orientation solution? It should be also dependent on the orientation as well. Certainly also the respective phase should be an additional parameter, since some phases show clear excess-deficiency effects, some other don't, cf. the attached patterns.
Hi Gert,
I am only really familiar with the algorithms used by Aztec and Flamenco software for orientation and phase determination. That said, the asymmetry you have demonstrated should not be much of a problem when using Aztec. Flamenco is more sensitive to problematic bands, but generally speaking, I believe it is not much of an issue for the band detection in Flamenco, either. Aztec will detect those bands fine and will conduct an analysis in which it isolates as many groups of 4 intersecting bands as exist in the Kikuchi pattern. From these sets it will compare the agreement of determinations and discard any bands which are systematically in disagreement. By this manner Aztec has improved orientation and phase determination over its predecessor, Flamenco. Side-by-side comparison of results from Aztec and Flamenco indicate that Aztec is more capable of deconvolving mixed patterns at grain boundaries, and also results in generally fewer mis-indexed and non-indexed points in your maps.
In short, with such nice patterns, I would not worry much about the asymmetry you observe for the orientation and phase identification. However, if you are storing patterns with plans to conduct a strain analysis (using cross-correlation of pattern images – probably with a software like CrossCourt), then these asymmetries might be more of an issue because the image analysis looks for changes in band width and angles between bands to estimate strain in the sample.
I hope this helps.
Cheers,
Zach
Hi Zach, I totally agree that for a common orientation determination excess-deficiency won't be that problematic. The accuracy of an orientation description is smaller than effect described. The final error is perhaps nothing else like a small additional samle tilt. Problems perhaps only occur for multiphase materials and diffferent excess-deficiency effect in combination with phase boundary characterization. Nevertheless, I do not have any idea about the final size of the systematic error.
The reason why Aztec has less misindexings or more indexings along boundaries (mixed patterns) is more related to the application of more bands. So the probability increses that the usually 6 or 7 required bands are part of the 12 detected bands. From my point it has nothing to do with the increase of accuracy. This is another new feature using an improved processing of band edge recognition. Both feature should require more ressorces, i.e. they take more time (what is not problematic if the accuracy will be inproved).
Volker, I guess that the described Kikuchi pattern are not acquired at high tilt angles what is the reason for the observed excess-deficiency in EBSD patterns.
This looks interesting. This isn't an EBSD band profile, is it? Channeling?
You could perhaps get some inspiration by the link. We did this years ago in order to check the possibility of dynaical theory.
Conference Paper Dynamical Simulation of EBSD Patterns of Imperfect Crystals
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Material Characterization