Hi I am doing TOPAS Rieltveld Refinement by JEdit on PbTiO3-based perovskite solid solution found that my peak fitting is not that good. I got Rwp 4.584%, Rexp 0.825%, Rp 2.780, and GOF 5.560. This data was collected by Bruker D8 Advance Diffractometer with 2mm filter, 40 kV, 40 mA, , step size 0.02degree/step and scan speed 4s/step. I used P4mm space group to refine the structure and I am trying to get a good refinement results for the cation displacement of A-site and B-site cations
I have tried to adjust all the parameters like thermal parameters Beq/ADPs, prefered orientation, peak type (I tried to put a different peak type in the coding in Jedit but it didn't work. The peak type JEdit used is TCHZ peak type), it didn't make the fitting any much better. When I observe the fitting pattern, I noticed interesting things on the pattern fitting. See in the diagram P4mm-1, P4MM-2, P4MM-3, it looks like the kalpha2 of my data is stronger than the calculated one which makes the fitting not as good. I first thought there is another phase inside then I tried to add one more tetragonal phase or orthorhombic phase to fit, the fitting difference is still the same. Then I thought it might due to the bigger fitter in the XRD machine so I used 0.2mm filter and scan the same powder with 0.01 degree/step and 4s/step and did the fitting again, which still shows the same problem with stronger kalpha2 intensity than the calculated one with the fitting parameters: Rwp 8.874%, Rexp 3.340%, Rp 6.429%, GOF 2.630 (see diagram P4MM-4 and P4MM-5). Does anyone know why it is like this ? Thank you for your help!
Dear Yi Yuan ,
If another phase's possibility is discharged, as tetragonal material, one possible problem is a stack fault. You can model it in TOPAS.
It is necessary to consider that the asymmetric peak profile may be responsible for your fitting result. Have you used a profile with divergent axial asymmetry? Once again, TOPAS can handle such a profile type.
Naturally, I am considering the material stoichiometric. If some vacancies or lose stoichiometry as a synthetic issue happens, those intensities can be explained.
The final fitting result is not so bad as you think; you are in the right way, need to refine the solution, try occupancy, divergent axial asymmetry, particle size distribution, stack fault, and, finally, consider the possibility to have a solid solution with different slight stoichiometry.
Best regards,
WNM
One thing which is not known among everybody is that you should use the Holzer profiles only in combination with a appropriately modelled instrument, but not with an empirical profile like the TCHZ function. If you use the latter (which I do in many cases), please use a simplified wavelength description with a single alpha1 and a single alpha2 with an equal "dummy" width of 1 (in that case only the relative values are used) and a 2:1 intensity ratio.
As you mentioned that problem is with stronger K alpha2 intensity compared to K alpha1. Generally ratio of Kalpha2/Kalpha1 ~0.5.
First try to put this ratio 0.51, 0.55, 0.6 etc. in steps and optimize this ratio to get best fit (to increase Kalpha2 share) using simple pV or TCHZ empirical profile fitting function as suggested by Prof Andreas Leineweber. If fitting is improved and good, then it may be an instrumental issue. Cross-check it with taking XRD data under same condition on NIST SRM powder sample and fix this ratio.
If there is another phase, then it will affect intensities of both Kalpha1,2 components simultaneously and it is difficult to guess such two-phase / multi-phase combination to selectively alter K alpha2 intensity only.
Dear Canh Nguyen ,
I didn't mean only refine asymmetric parameters; Please, use a divergent axial asymmetric peak function description.
I guess a better statistics, longer collecting time will improve you fitting.
Please, check the reference:A Correction for Powder Diffraction Peak Asymmetry Due to Axial Divergence, Journal of Applied Crystallography 27(6):892-900 DOI:10.1107/S0021889894004218
See the FullProf page attached.Best regards,WNM
I wish to emphasize part of what Ashok Bhakar said (and disagree with part).
"First try to put this ratio 0.51, 0.55, 0.6 etc. in steps and optimize this ratio to get best fit (to increase Kalpha2 share)"
No, do not do that, as you may improve fit in a way that is not meaningful. Your goal is not a good fit; your goal is to describe your sample.
"taking XRD data under same condition on NIST SRM powder sample"
Yes, you do need to characterize your instrument. Be sure to collect scan of SRM using same configuration as used for your sample. Fit to the SRM data. That fit describes your instrument. Then, in the fit to the sample data, fix the instrument parameters and vary only sample parameters.
The apparent ratio of kalpha1:Kalpha2 > 1:2 is meaningful, but does not indicate the ratio is really not 1:2. The asymmetry of your peaks with tails to the high angle side is not a typical instrumental or sample peak. Really looks like you have a second small peak that unfortunately falls right on top of Kalpha2; the high angle tail is caused by the Kalpha2 peak of the small second peak. Consider tetragonal or a second phase with just slightly smaller lattice parameters.
Or do you have an incident beam monochromator set completely wrong that is supposed to eliminate Kalpha2, but in set to select more Kalpha2?
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders