Because of the superimposition of multiple orders of interferences, but also because each band is affected by any other band since each lattice plane is intersected by all other planes, the interpretation of EBSD patterns is sometimes difficult. The effect is commonly described as pseudo-symmetry and can have several reasons, like small deviations from lattice symmetry, the same crystal lattice but lover Laue class, but also the very special effect that there are similar solutions available which are not symmetry-equivalent.
I am looking for phases which are specifically affected by misindexing. Could you help?
Definitely problems will occur if parts fo the Kikuchi bands show a similar geometry. You can find this in the case of polytypes.... Although Idid not do particular methodological research on this, I (or the people who did the EBSD for me) encountered that problem in the case of different Laves phase polytypes, but also in the case of cementite/Hägg carbide.
I am not very familiar with Laves phases, but is their metric so similar, or does this statement only considers polytypes where mainly the lattice parameters are mulitplied, and the lower symmetry density is only the result of small displacement of atoms? In my experience, Laves phases often generate very nice patterns, and they are commonly quite brittle and therefore more resistant against plastic deformation during sample preparation. Therefore, a nice example for misindexing.
Laves phases actually also follow the ABC stacking principle like closed-packed metals, ZnS etc. In our samples we had, besides polytypism, also layer faulting, which is known to broaden Bragg reflections, which also might blur the Kikuchi lines. I think such cases might be handled by use of more sophisticated indexing criteria in the indexing software.Another problem is extended sigma3 twinning for the 3C polytype (C15 in the case of Laves phases). The patterns of such twins can well appear together in one frame. Some optimisation in phase identification had been done in Z.Anorg.Allg. Chem. 635 (2009) 637. I have, however, not been part of that particular work.
I had trouble with mis-indexing glaucophane for barroisite due to their similar and was only able to correct it by looking at Na and Ca content for each indexed point with a combined EBSD/EDX data set. Have a look at my paper for more detail. I wonder if the software (I was using Channel at the time) has improved much in this regard?
Article Fabrics produced mimetically during static metamorphism in r...
I have had trouble with low trigonal (laue class -3) phases, specifically dolomite (CaMg(CO3)2)and ilmenite (FeTiO2). They often misindex about the m-direction (in the basal plane) because the bands that differentiate them from crystals of -3m symmetry are not well defined in the patterns. I have spent ages looking at the simulations produced by Bruker's Esprit software to identify where the critical bands are to differentiate the two patterns and it is not always clear to my eyes let alone an automated routine.
As a consequence I have taken to indexing dolomite as magnesian calcite (laue class -3m) and using a high trigonal ilmenite match unit. As with most misindexing (I spent my PhD working on plagioclase!) the severity of the misindexing is orientation dependent (i.e. some phases have the critical bands more visible in the pattern than others).
The misindexing phenomenon is also not software dependent. I have used both Bruker's Esprit and Oxford's Aztec and had misindexing in dolomite in both packages. Sometimes the dolomite patterns that I have been working with are of dubious quality. However, the ilmenite patterns are stunning and even using an eFlash HR camera with pattern resolution of 320 x 240 pixels it still misindexes.
Dear David and Mark, thanks for your detailed description of your problems. The differentiation between galucophane and barroisite I would also try to solve using EDS since the lattice parameters are quite similar and the point group is also the same. I havn't looked yet for the structure but if the intensity distribution is quite similar it might be hard to do it using band positions. This is comparable to chromite and magnetite which are only (fast) discriminatable using chemistry...with all disadvantages like the different information volume of EDS and EBSD. In so far i am not very optimistic that this problem can be solved without some other clever approaches. One would be the application of the backscattered coefficient (background signal). It might be that both phases generate a weak but sufficient difference which can be used. Until now no package is doing this, although EDAX launched their software PRIAS last year which enables at least a post processing. How much this can be used already during indexing, I don't know. But for several problems it works nicely and could help for critical questions, especially if a higherspatial resolution is required, see e.g the given link.
The problem of low-symmetric Laue groups as described by Mark is comparable to pyrite but . Unfortunately I do not have such a sample, but I am currently working exactly on the solution of such problems. Could you support me by a sample, Mark? It would be a pleasure for me to test the developed software on this sample. We could already map (!) GaP or InP with respect to their point group (not only Laue group). I guess that there is a good chance for -3 and -3m.
By the way... neither the application of the HR detector would help, nor the higher pattern resolution...since the pattern is binned down anyway (internally).
Conference Paper EBSD patterns: a mine of information
As already mentione above, chemical information should alwyas be used, particularly when one is dealing with compounds based on the same structure type (e.g., perovskites).
@ D.D. McNamara: How much time did you loose if you combine EDS and EBSD? Is it remarkable and you have to slow down the speed in order to get enough counts, or is there no big difference to measure EDS as well? How reliable is the chemical differentiation by EDS under 70 degree tilt? Or do you use element maps as mask?
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science