I have solved the lattice type and space group of high pressure phase of a thallium based ternary ABC2 type compound which crystallizes in orthorhombic structure. I tried to fit the high pressure [synchrotron ESRF, λ = 0.3738 Å) phase using profile matching method with Fullprof. Eventhough all the Bragg positions are matching, the intensity fluctuations are too high. The Chi square value is in the range of 10^7 (10 to the power 7). I am unable to reduce the value further. How to reduce the chi square value further. Please add your valuable suggesions.
Reply to Nita Dragoe: Dear sir, please find the attachment. It is a two phase fitting. The peaks after 9 degree and 10 degree (intense) are due to Re gasket.
I have used profile matching with constant scale factor and reflection list automatically generated from the space group.
your peak profile does't match that good (it seems to be asymmetric which is not unusual). But the major problem is the peak width which does not match for small angles. I am not sure whether you are using FWHM=f(U,V,W), but if you should check whether V is zero and U is not zero. Anyway, the asymmetry of the peak profile needs to be adjusted somehow.
Reply to Gert Nolze:
I thank for your suggestion. Since the initial value of chi square is very large (10^7), therefore even if I refine the FWHM, there won’t be appreciable change in the pattern. However, if I am able to bring the chi square value to less than 1000, then easily I can fit this data.
My question is that, did I wrongly edit the pcr file because the chi square is too large. Since it is in profile matching mode, I can’t adjust the scale. Please add your valuable suggestion.
YOu should start with only W and keep U and V zero. The reason: both parameter describe the 2theta variance in FWHM. In your case it is only 20° so that you can keep FWHM for sure in first approximation constant. Therefore you should manually adapt a W which already matches your FWHM. If then still remarkable problems occur they are related to other effects like anisotropic peak broadening. But I don't believe this at the moment. If you have the data as x,y table (angle intensity) you could send them. I could test it with PowderCell. However, since I do not have the phase files it might be too complicated. At least I need the phase files as CIF or cel-files...or somehow readable.
Reply to Gert Nolze: Dear sir, thanks a lot for your suggesion. Please find the attachment of dat , background and pcr files. It would be very nice if you look at my attached pcr and dat files.
I still need the phases (crystallographic data). Without them I cannot do anything. :-(
Reply to Gert Nolze: Thank you sir. Please find the attachment of cif files for the orthorhombic (S.G: P m c 21) phase and Re phase.
Lattice parameters for the orthorhombic phase is a = 3.7512 Å, b = 10.3738 Å and c = 4.0482 Å and the structure type is Al2CO (attached)
Second phase is due to the gasket peaks from Re.
For Re: a=b=2.696650 and c = 4.369567, α=β=90, γ = 120. S.G: P63/mmc, Z = 2
x=0.3333, y=0.6667, c=0.25, SOF=1
I guess you sent me the wrong structure file since in contrsat to your lattice parameters there are given
_cell_length_a 3.273
_cell_length_b 5.33
_cell_length_c 5.03
which does not fit at all.
Nita Dragoe: Thank you sir. The space group corresponding to the ambient pattern is Cmc21.
The first structural phase transition is initiated at around 5 GPa and it is completed after 20 GPa. The occurrence of extra peaks in the XRD pattern shows that the new phase could be a lower symmetry phase and hence either orthorhombic to another orthorhombic or orthorhombic to monoclinic phase is expected. In order to find the possible space group for the daughter phase, I looked at the possible subgroups of Cmc21 from Bilbao open data base. Using Pearson's data base, I tried to match the sub groups one by one with my XRD pattern. I found that the XRD pattern for ABC2 (here Al2CO) structure type (for λ = 0.3738 Å) with the space group Pmc21 is matching with my XRD pattern.
Then using powd software, I found that my lattice type is orthorhombic with lattice parameters a = 3.7512 Å, b = 10.3738 Å and c = 4.0482 Å. The above attached cif file belongs to Al2CO and it can be used to find the Wyckoff positions of the high pressure phase. That would be the next step.
Please find the cif file for my data. To run the pcr file, please select "Automatically generated from the Space Group symbol" for both of the phases.
could you post the original data and not the background-corrected? You shouldn't do any background correction since you don't know the background bevor finishing everything.
Answer to Nita Dragoe: Dear sir, Thank you very much for your valuable suggesion. I have used Lebail fitting since I am not sure about the Wyckoff positions. I didn't use any absorption model and I didn't use it for even parent phase also, but I could fit it. Please find my pcr file, dat and cif files.
Reply to Gert Nolze: Dear sir, thanks for your immediate reply. Please find the attachment of my dat file with the background uncorrected
This is an amazing background. Do you know where it is coming from?
If the background has this specific shape, didn't you think you cannot trust the intensity described by your diffraction peaks? I am not familiar with this geometry so that I have to ask some other colleague here who already did synchrotron measurements at very small angles. Sorry.
Reply to Gert Nolze: Dear sir, with the same background, I have fitted the ambient pattern. Please look at the refined data at 0.28 GPa after subtracting the background. Atleast, now I understood after conversation with you is that the issue is with background.
Fantastic :-). Backgroud handling is very sensitive. The problem is obviously solved and the reason a) the background and b) the sensitivity of the applied mathematics (with perhaps negative intensities). Sorry that I had to give up before since the software I used (PowderCell) got some problem and always ended up with an error message. I couldn't find the reason. Anyway, good end for you!
Congratulation! Ir should tell you that mathematics is not always correct, automatically.
Reply to Gert Nolze: Dear sir, thank you very much for spending your precious time for the discussion
Seems like fluorescence from Sb and Se.
Here I use MAUD rather than Fullprof because have to deal with anisotropic broadening from TlSbSe2 also preffered orientation.
Reply to Maykel Manawan: Dear sir, Really you did a great job. I am familiar with Fullprof. Can you please guide me to learn MAUD.
Dear Sorb here is the link: maud.radoigraphema.eu. U cud find anything u want there. There's also video tutorial uploaded in youtube. U cud email me also for that work ([email protected]).
Reply to Maykel Manawan: Thank you very much. Definitely I will contact you if some help is required. sir, you didn't attach the link.
Thank you very much to all those who engaged for this dicussion. Finally, I did the refinement using fullprof. The difficulty with the fitting was due to the improper background subtraction
- Badges
- Science method
- Similar topics
- Methodology
- Crystallography
- X-ray Diffraction