Dr. Stuart Wright, Senior Scientist, EDAX

A few weeks ago, my colleague at EDAX, Shawn Wallace, posed a question that has stayed with me, and so I thought my turn at the EDAX blog would be a good place to address it. Shawn was building an EBSD structure file for a new phase and encountered the following dialog for adding an atom to the unit cell.

Shawn asked how important it was to get the Ion Type correct for the structure he was working with. I realized I had implemented this capability several years ago for kinematical calculations of structure factors but had never really explored it’s impact on the calculations. I guessed that it would not have much of an impact, but I wasn’t entirely sure that was the case. The choice of ion type affects the atomic scattering factor used in structure factor calculations. I looked through our phase structure database for a binary compound containing Fe and decided to use a simple Al-Fe structure to check out the effect of the ion type selection on the structure factor calculation.

I calculated structure factors for the Fe and Fe+3 selections in the dialog box above. As shown, the difference in the structure calculation results is imperceptible in the kinematically simulated patterns. The maximum difference between the two patterns is only a 1.6% difference in the relative intensity of the {100} bands.

Here is a table showing that the structure factors are quite similar, confirming my initial guess. I haven’t tried any other structures, so it is not a complete study, but I suspect other structures will follow the trend shown by the simple Al-Fe structure. Thus, my conclusion is, Don’t Sweat the Small Stuff.

(hkl)	FFe	FFe+3
110	4.773	4.826
100	1.291	1.738
200	3.256	3.252
211	2.566	2.562
111	1.098	1.104
220	2.189	2.186
222	1.633	1.632
210	0.913	0.908
310	0.858	1.856
321	1.462	1.461

With that little study wrapped up, I turned my attention to choosing the Debye-Waller factor used in the dynamical simulation. In the dialog above, it says the default Debye-Waller factor for iron is “0.003106 for bcc, 0.533 for fcc”. Does the choice of Debye-Waller factor matter? Here are dynamically simulated patterns for these values.

The two patterns are quite different. To correctly use the new simulation tools, I need to expend some effort to learn more about Debye-Waller factors. Clearly, It’s the Little Things that Matter Most.