Pawley Fitting

Files needed: d5_05005.raw

Learning outcomes: Pawley fitting is a process in which observed peaks in a powder pattern are fitted without a structural model but at 2-theta values constrained by the size and symmetry of the unit cell. It is a useful precursor to Rietveld fitting and gives an indication of the "best fit possible" from an eventual structural refinement. It is also important before structure solution to either extract intensities for e.g. direct methods or to determine e.g. cell parameters and peak shapes for simulated annealing structure solutions.

Topas advantages: very fast/stable compared to other packages; both Pawley and Le Bail algorithms are available.

1. Copy the file above to your area.

2. Work through the items on the "Simple Pawley Refinement" menu.

3. Click on "Select Data File" and navigate to select d5_05005.raw.

4. Click on "Instrument/Corrections" and select "Durham_d5000" for the diffractometer and "Refine zero point".

5. Click on "Pawley fit phase" and click on "Structural Information" then "Phase Name" and enter "TiO2".

6. Open "Lattice parameters" and select "Tetragonal". Enter the cell parameters of 4.595 and 2.959 for a/b and c (these are the values from the indexing tutorial which you may/may not have done yet). Type these over the top of the "#" symbols.

7. Click on space group and enter "P42nm" - the space group suggested by the earilier indexing tutorial.

8. Click on "Save/send to Topas". Launch topas if needed and run refinement.

9. You should get a Rwp of ~ 15.1%. Note this is similar to that obtained by both Rietveld fitting (~15.5%) and independent peak fitting (~15.2%) the same data.

10. Examine the input file in jedit. Reflections and their intensities have been written to the file. The cell parameters and zero point will have shifted slightly.

11. Further work: the true space group of rutile is P42/mnm. Try deleting the reflections from the input file (i.e. from the line "load hkl_m_d_th2 I" onwards). Then try the Pawley fit in this space group. You should get a similar Rwp (~14.95%). Try different peak shape models.


[Modified 31-Jul-2017 by John S.O. Evans. Pages checked for Google Chrome.]