I have done the Rietveld analysis of a oxide sample by using fullprof. The initial given occupancy values are much different than the obtained values after Rietveld. So, my doubt is, whether can we be sure the obtained values ( occupancies) after fitting is reliable.? if not, what is the best method to find them.?
Dear Sai Pranav,
Let's to assume that you have a very good diffractogram, i.e. small step size and a very good statistics. Let's to also assume that all intensity corrections have been done, i.e. LP correction. Let's to also asume that you have introduced the site occupancies in Fullprof correctly because they must be divided by the multiplicity of the general position (see the manual). In that case it is possible to refine the site occupancy of two atoms in the same site and get reliable values. However, the refinement of the site occupancies is influenced by the temperature factor parameter. So, you should take care of refining both because they are correlated. I usually start with a common isotropic temperature factor for all atoms, after that I refine one for each type of atoms and finnaly I refine one for each atom. Pay attention that the temperature factor should be always positive and within a expected values. After that you can start to refine the site occupancy only for the atoms that you expect some substitution remembering that the sum of the site occupancies in that site must be the unity. You should take into account that the substitution of atoms with similar Z is difficult to refine with X-rays, i.e Al and Si differ in only one electron so the effect in the cell electron density is almost negligible. In such cases you can try with neutron diffraction on you can derive the substitution from the cell parameters variation.
Best regards.
Dear Sai Pranav,
Let's to assume that you have a very good diffractogram, i.e. small step size and a very good statistics. Let's to also assume that all intensity corrections have been done, i.e. LP correction. Let's to also asume that you have introduced the site occupancies in Fullprof correctly because they must be divided by the multiplicity of the general position (see the manual). In that case it is possible to refine the site occupancy of two atoms in the same site and get reliable values. However, the refinement of the site occupancies is influenced by the temperature factor parameter. So, you should take care of refining both because they are correlated. I usually start with a common isotropic temperature factor for all atoms, after that I refine one for each type of atoms and finnaly I refine one for each atom. Pay attention that the temperature factor should be always positive and within a expected values. After that you can start to refine the site occupancy only for the atoms that you expect some substitution remembering that the sum of the site occupancies in that site must be the unity. You should take into account that the substitution of atoms with similar Z is difficult to refine with X-rays, i.e Al and Si differ in only one electron so the effect in the cell electron density is almost negligible. In such cases you can try with neutron diffraction on you can derive the substitution from the cell parameters variation.
Best regards.
I agree to Prof. Guirado. You should try this refinement in last one by one perhaps there may be results
I also do refinement using Fullprof and use the same procedure as described by Prof. Guirado. Here I have a query. After refining occupancy one by one for the expected atoms, should we refine all other parameters again like lattice parameters, background, isotropic temperature factor etc.?? if I do this then there is change in occupancy on refining again. I mean occupancy keeps on changing on refining it again and again if I do refinement of all other parameters once occupancy gets refined.
I know I am doing something wrong. So plz help me in this regard.
Thank you.
Dear Merry,
At the last step of the refinement all meaningful parameters should be refined simultaneously. If this results in divergence or meaningless value of some parameters, it means that the quality of your data does not allow to refine all that you want. In this case you have to make a smart choice: what should be refined, what should not be refined and what should be constrained.
You must post more information. Nature of your sample, sample preparation, instrument configuration, data acquisition and refinement strategy rules. If there's mix between light and heavy elements or between adjacent elements that x-ray unable to distinguish, than neutron is mandatory. Do combine refinement to get the advantages of both techniques.
Occupancy, atomic position and thermal displacement have a strong correlation, refine them at very end as pointed out by Prof Guirado . Nowday, the instrument configuration could give such issue as well as micro absorbtion due to inhomogenities in particles statistic, etc.
Post your diffractogram together with the instrument configuration if possible.
Or you could try your refinement strategy with this round robin data: http://www.iucr.org/resources/commissions/powder-diffraction/projects
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science