The correct identification of phases is a fundamental information for a meaningful characterization of a microstructure by EBSD, especially when the phase is of minor fraction and cannot be discovered by other, more sensitive techniques like XRD.
How accurate lattice parameters need to be known if the phase is assumed to be identified properly? Is there any important application where it would be beneficial to know the used phase better than now?
Good point (which is however always a problem in electron, ion, or proton techniques). It certainly depends on the thermal flux in the sample (phase). On the other hand, I don't think that the accuracy in lattice parameters determination is sufficiently good that the local heating will make serious problems. But...it will influence the lattice parameter determination without any doubt.
A serious application for lattice parameter detection via EBSD would be the analysis of martensite in steel. Martensite is quite similar to ferrite (bcc structure), but due to rapid cooling the carbon in the austenite (fcc structure) had no time for duffusion and is stuck in interstictal sites in the bcc lattice. Therefore the lattice constant in the martensite is slightly higher than in ferrite. Due to the German wikipedia http://de.wikipedia.org/wiki/Martensit the martensite changes from cubic to tetragonal lattice depending on the carbon concentration.
So far we can not distinguish ferrite from martensite via EBSD. We can only see a higher band slope contrast in ferrite compared to martensite. If you could develop software for measuring the lattice parameters that would help a lot in understanding the steel better.
Dear Martin, in fact you are right, but you selected a very difficult example. It is not only pseudosymmetric, and does not only have practically the same lattice parameters, it is also highly defect-containing. I myself don't expect that EBSD will never ever have a better resolution than XRD or synchrotron. I was more looking into the direction of precipitations, intermetallic phases etc. where we still now only assume the selected phase is correct only because the software is indexing somehow. Or in mineralogy, where many phases can occur. But for materials like Ti-Al, ZrO2, martensite-ferrite-bainite (which are even not real phases but actually only a microstructures) I an convinced that the currently applied EBSD techniques are much too speculative (already, because of the "unknown" wave length of the used electrons). Nevertheless, thanks for your answer.
Apparently Gert, the answer is not! :). However, you should look perhaps on the direction of Cross Correlation techniques. Band width determination has only been approached by residual strain people.
Cross-correlation only makes sense when you have a good simulation which describes the band profiles. But this is not really the "problem". How accurate a lattice parameter has to be in order to be useful? Would be 5% OK? This is the nowadays accepted best accuracy, but if someone could do it with 2%... Wouldn't this help for certain applications? How accurate TEM people can do it with SAED? This should be in the same limit, or am I wrong?
Gert, your question is similar to mine. I have conducted two experiments in titanium alloys and titanium aluminide alloys. In some of titanium alloys after quenching from beta phase field, martensite prime occurs which owns a hexagonal crystal lattice, however, when I tried to analyze this sample using EBSD, it cannot tell the martensite and alpha phase apart (alpha phase has one crystal structure of HCP). Another one is the EBSD cannot distinguish the omega phase and alpha2 phase in TIAl alloys. I think maybe this phenomenon occurs because of similar crystal structure and lattice parameters in both phases we try to distinguish, and if we can get the very Kikuchi database of martensite or omega, which can help the EBSD system to distinguish them from alpha or alpha2, respectively. Also, maybe we should compare the value of MAD to make a decision.
"the martensite and alpha phase" & " the omega phase and alpha2 phase in TIAl alloys", do these similar crystallographic phases have distinct XRD reflections? If so, in that case, they may be distinguishable using Bragg XRD Microscopy. Presently, this method has a lot lower spatial resolution than the high tech EBSD tool. However, it has a dramatically higher resolution in reciprocal space for any given VOXEL of sample surface. Meaning, in terms of being able to distinguish minuscule changes in orientation (fractions of ArcSec).
Ravi, you are right. These pseudo-phases (like martensite, bainite, delta-ferrite, since these are more microstructures than phases you can find in a crystallographic database) or pseudo-symmetric phases like gamma and gamma' in superalloys won't generate patterns which form different band positions which are the major information used during phase discrimination. I don't want exclude that in future detectors and techniques will be availabel which are able to use reliably band intensities, but this is nowadays not supported at all. Nevertheless, there is a software (EBSDL) available which allows to determine lattice parameters using single EBSD patterns. It determines from unknown phases of any crystal symmetry the Bravais lattice type, and derives the lattice parameter. Presently, the precision in better than 3% which is surprizingly good for EBSD, wheras the accuracy in lattice parameter ratio is about ten times better. I think for many applications this should be a good first approach for a phase characterization.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials Chemistry
- Polymer Chemistry