Over the last few weeks I tried to derive the lattice parameters using electron backscatter diffraction (EBSD) in SEM. Comparable investigations using simulated patterns gave in promising results. Only experimental pattern showed systematically shorter lattice parameters caused by obviously wider bands. They can be explained by a change of the electron wavelength, i.e. by the electron energy. Often the so adapted wavelength shows a up to 10% smaller electron energy. Which technique enables a relatively simple derivation of the electron energy? How reliable is the EDS background signal? Does the sample tilt has any effect?
Simple question, how accurately have you determined your working distance ?
Have you calibrated for the effect of tilt, and any sample movement on that working distance ?
@John: I don't understand the meaning of your question. Is it related to the electron energy or to the geometrical conditions? The working distance is (from my point of view) totally irrelevant for the energy. The pattern center determines in a refinement process the (indirectly) working distance, but also for this application WD is totally irrelevant. Only the relative position of the diffraction source and the imaging EBSD screen is important. And...the electron energy E which practically scales with 1/sqrt(E) all band widths.
Dear Gert,
Isn't it possible that not the primary beam, but backscattered electrons diffract in material - with lowered energy by these 10%?
Is it material dependent? did you try other materials as well?
Yes, I tried different materials. Metals, semiconductors, minerals... The all show tendentially the same. For metals it is quite OK (there we are at the limit of the accuracy and one also has to consider the error in Bragg angle determination because of the intensity profile), but for different minerals it changes, and this is not explainable by the error in band width.
Regarding the backscattered electrons used for the pattern generation: until now the general accepted belief exists that only the zero-loss electrons form the visible diffraction patterns. Of course, all other also form diffraction patterns but a) their intensity is much lower because of the high efficiency of backscattered electrons directly on the surface, and b) electrons coming from higher depth commonly interact again and loose then energy so that they do not match anymore energetically. This is obviously also the reason why TKD patterns look identical like EBSD patterns. We only see the signal generated directly at the surface, anywhere where the surface is. Is the detector below the sample you see the signal there and if you would be able to move the detector looking from the other side you would see the same. Perhaps a good idea to tilt the TKD sample +70° and -70° in order to compare both signals. Do you have a system running at TESCAN which you an use to test it?
Yes i think its totally depend on the operating the SEM instrument on sample with a lower energy and also depends on the type of sample.
The obvious choice is to use the background signal in an EDS spectrum to measure the Duane-Hunt limit. I don't know the precision of this measurement as it will depend on the EDS signal intensity and how you fit the background, but it is the only direct method I can think of. See this reference:
http://dx.doi.org/10.1017/S143192760404070X
You would want to do this measurement under the same same conditions as your EBSD work as the sample tilt, beam current, etc. will all effect the result to some degree. We do this exercise as part of a lab course I teach and you can see several kV of charging induced deceleration on some samples.
My concern would be that the sampling volume is very different for EBSD and EDS and that could skew the results..
@ B. Myers: I think so as well. I don't assume it is correct enough. However, a remarkable difference would at least show that one has to expect some kind of effect. Moreover, not every times one is using EDS parallel to EBSD.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science