misfit of tick marks and peaks in Rietveld refinement

xiaojiao

Dear all,

I'm trying to do rietveld refinement on the data from Bruker D8. whenever I try to refine unit cell parameters, nearly all tick marks would move randomly and off the peak ranges, while the calculated and experimental patterns still match well. It doesn't make sense, and didn't show in my synchrotron XRPD data.

please, where possibly did I do wrong? How could I fix this?

Thanks very much!
Xiaojiao

johnsoevans

There are lots of reasons why observed peaks can shift relative to tick marks. For example if you have an asymmetric contribution to the peak shape then the peak maximum will no longer be at the 2-theta value you'd predict from Bragg's law.

In topas the tick marks are therefore placed where they would be for a given unit cell with no instrumental/experimental contributions. The calculated pattern is then offset according to corrections like zero point, sample height, peak asymmetry, etc, etc.

To prove this to yourself, try putting zero point to 0.0 and axial (or any other asymmetric contributions to peak shapes) to 0.0 and do a single cycle. You’ll see the red calculated peaks are exactly above the tick marks. When you let these parameters refine the red calculated peaks shift to fit the data but the tick marks don’t move.

Other programs choose a different approach to this and move the tick marks.

xiaojiao

Thank you very much John. I have tried to do refinement with zero point, axial and other asymmetric contributions fixed as well as refining. However, I still got the same outcome. please see the image below. The tick marks are completely off with regard to the peaks; yet the peak shape seems to fit ok. Do you still think this is about asymmetric contributions? Or do you have any other suggestions?

johnsoevans

Did you fix the parameters at zero rahther than just fixing them? It could also be sample height causing the offset if you have that in your model. If that still doesn't explain things I'd have to see the .inp file. If you want me to look at that send privately rather than to the forum.

alancoelho

Hi xiaojiao

Send me the INP/data file and will have a look
cheers
alan