Something that I need to clarify from you. When you aim to refine the structure I noticed that chi square value gets worse (above 1.5) even as density map looks better (let us say it changed from 1.2 to -1.2 to the range from 0.49 to -0.49). So what I want to say when I export the results of the grain size which result should I trust. The grain size values before structure refinement or after? Thank you.
One thing that comes to mind is that the calculation of Chi^2 depends on the number of refined parameters. If you release more parameters, Rexp gets a bit smaller, and hence chi^2 = (Rwp/Rexp)^2 increases. Whether or not this explains your observations I don't know. The crystallite size is derived from the peak width. So whatever fits your peak shape better should result in more accurate crystallite sizes. In a high-quality dataset, that would be the structure refinement. But in a low-quality dataset, a structure refinement might over-refine the data and result in bogus results. Without seeing your data I can't tell which of the two scenarios applies to you.
I keep getting Grain Size error, please advise, I've tried variations of isotropic, anisotropic refinement of k2 and B1, any help would be great thanks.
Is B1 refined to 0.0? That means infinitely large crystallites, which is reported as GrainSize=ERROR. The physical meaning is "too large to cause any detectable peak broadening".
While in theory it is possible to add the calculation of bond lengths to the STR files, it is extremely cumbersome. Profex version 5.3 will come with a dialog that lists bond lengths after the refinement.
Something that I need to clarify from you. When you aim to refine the structure I noticed that chi square value gets worse (above 1.5) even as density map looks better (let us say it changed from 1.2 to -1.2 to the range from 0.49 to -0.49). So what I want to say when I export the results of the grain size which result should I trust. The grain size values before structure refinement or after? Thank you.
One thing that comes to mind is that the calculation of Chi^2 depends on the number of refined parameters. If you release more parameters, Rexp gets a bit smaller, and hence chi^2 = (Rwp/Rexp)^2 increases. Whether or not this explains your observations I don't know. The crystallite size is derived from the peak width. So whatever fits your peak shape better should result in more accurate crystallite sizes. In a high-quality dataset, that would be the structure refinement. But in a low-quality dataset, a structure refinement might over-refine the data and result in bogus results. Without seeing your data I can't tell which of the two scenarios applies to you.
I keep getting Grain Size error, please advise, I've tried variations of isotropic, anisotropic refinement of k2 and B1, any help would be great thanks.
Is B1 refined to 0.0? That means infinitely large crystallites, which is reported as GrainSize=ERROR. The physical meaning is "too large to cause any detectable peak broadening".
Can we have an example were we can calculate bond lengths?
While in theory it is possible to add the calculation of bond lengths to the STR files, it is extremely cumbersome. Profex version 5.3 will come with a dialog that lists bond lengths after the refinement.