Y2O3 Constant Wavelength Laboratory X-ray Data

This example will show you how to perform a simple Rietveld refinement of a high symmetry single phase oxide material using laboratory X-ray diffraction data. A high quality diffraction pattern (d5_00451.gsas) was recorded on a bulk sample of Y2O3 using a Siemens d5000 diffractometer equipped with a Cu tube, graphite diffracted beam monochromator and scintillation counter. The relevant instrument parameter file is d5.inst.

1. Copy d5_00451.gsas and d5.inst files to a new folder.

2. Launch expgui and navigate to the folder you want to work in.

3. Type the name you want for the .exp file (e.g. Y2O3_01.EXP) in the dialogue box and click "read".

4. Ask the program to create the .exp file and type in a title in the box that appears.

5. The main expgui window will launch. In the "LS Controls" tab set the number of cycles to be 20. Click on the yellow "extract fobs" box (this will allow the R(F**2) to be calculated).

6. Click on the "Phase" tab. Click on add phase.

7. Enter a phase title of "Y2O3", cell parameters of a=b=c=10.6 Å, space group "I a -3" (spaces are important!). Click on add. The program shows you the symmetry operations it has deduced. Click on continue.

8. Click on "Add new atoms". Click on "More atom boxes" until you have room to type in 3 atoms.

9. Type in the guessed coordinates below, leaving names and temperature factors at default values. Check that the site multiplicities correspond to sensible cell contents.
Y1 0.97 0.00 0.25
Y2 0.25 0.25 0.25
O1 0.39 0.15 0.38

10. Click on "Histogram" tab. Click on "Add new histogram". Select d5_00451.gsas as the data file and d5.inst as the instrument parameter file. Click on add.

11. Click on the "Edit Background" box and switch to function type 1 with 9 terms for the background. Click on "set".

12. Click on "powpref" so gsas prepares the necessary files for least squares refinement.

13. Click on "genles". A dos window should pop up as refinement proceeds. You should get Chi**2 of 59 and wRp of 23.98%. Press any key then click on "load new".

14. Click on "liveplot" to see the refinement progress.

15. Click on "powplot" to see the refinement in more detail. Choose option "C". Choose "N" to not save graphics. Choose options "h 1 m d t p" (read histogram 1, mark reflections, show difference curve, 2-theta scale, plot). Zoom to another region by typing "10 30 a".

16. Refine zero point (click yellow box contained on "Histogram" tab), then cell parameter (on "Phase" tab). Run genles after each. Chi**2 ~ 26.

17. In the "Phase" tab highlight the two Y atoms and click the yellow box to refine their xyz coordinates. Then do the same for the oxygen atom.

18. Highlight all the atoms and refine their U values. You should get Chi**2 ~ 14.

19. Go to the "Profile" tab. Try refining GU, GV, GW, LX, LY, asym.

20. Click on "lstview". Look for the value of the profile parameters. If any have refined to e.g. 0.0 with a large esd remove them from the refinement.

21. Under menu "Calc" run disagl and look at the bond distances to see if they are sensible.

22. Under menu "Export" write out a .cif file.

23. Use "Mercury" to view the structure from the .cif file (this is found in start/programs/academic software/chemistry/ccdc 2006).

24. [almost certainly not worth doing] If you really want to, you could try using"atoms" to view the structure. Try downloading atoms demo from shape software (this may not work!). Import structure and use the pull down box to select "gsas" file format. Click on input. Tick all the tick boxes. Say yes/ok to the next few questions. Go to input1/bonds and delete the bond between type 39 atoms (Y). Go to input2/unit cell and click to show unit cell. Go to input1/polyhedra to show polyhedra.

Refining 24 variables in total you should be able to achieve Chi**2 10.58, wRp=10.14%, Rp=6.12%, R(F**2)=3.2%.

Additional Work (optional/needs other software in i:\license\rietveld\school_software)
Many modern lab diffractometers use so-called "variable slits" rather than the traditional fixed divergence/anti-scatter slits assumed in many Rietveld codes. These slits give rise to a constant area of illumination on the sample giving higher diffracted intensity at high angle than fixed slits, without the beam being larger than the sample at low angle. The intensity correction can be approximated as 1/sin(theta). d5_00450.raw was recorded using variable slits. Use xch to convert the data to ascii (2theta, I) format. Using excel, convert the data to a pseudo-fixed angle setting (think about what to do with the esd's!) and compare refinements of the two data sets. Alternatively use v6xytogsas.


[Modified 21-Mar-2018 by John S.O. Evans. Pages checked for Google Chrome.]