CLS XRD & Scattering Summer School 2024 Na0.5Bi0.5V0.4Ti0.

6O3
Rietveld Refinement with Constraints on Synchrotron Data

Rietveld Refinement of Na0.5Bi0.5V0.4Ti0.6O3 Structure using Constraints with

GSAS-2 using Synchrotron PXRD Data

Data Files: 00-072-0781.xy

Starting Instrument File: Inst-00-072-0781.instprm
Crystal Structure File: 00-072-0781-NBVTO.cif

In this exercise, we will perform Rietveld refinement of the Na0.5Bi0.5V0.4Ti0.6O3 structure

using constraints for atoms on identical sites (constrained coordinates, isotropic
displacement parameters and site occupancies) and constraints on the isotropic
displacement parameters of the oxygen atoms.

A. Creating and setting up the GSAS-2 GPX file for the refinement

The setup is similar to the steps in the Y2O3 and LaB6 Rietveld tutorials, with a few small
changes:

1. Use the following folder where the initial files for the Na0.5Bi0.5V0.4Ti0.6O3
refinement are located:

C:\RIETVELD-JR\NBVTO-Rietveld

Import the data file 00-072-0781.xy (Import → Powder Data, when opening the
data file, choose guess format from file) and GSAS-2 should ask if you want to
open an instrument parameter file (Inst-00-072-0781.instprm) after opening the
data file. You will have to select the instprm file type to see the file.

2. Import the Na0.5Bi0.5V0.4Ti0.6O3 structure (Import → Phase) from the CIF file
(00-072-0781-NBVTO.cif), name the phase NBVTO, and add the histogram to
the new phase.

3. In the data tree, choose Controls and change the Max cycles to 10.

4. In the data tree, choose Background and make sure the Background function is
log interpolate and the Number of coeff. is 16. The background of this pattern is a
little unusual, but the log interpolate function will model it quite flexibly.

5. In the data tree, choose Sample Parameters, and uncheck (turn off) the Histogram
scale factor. Change the Goniometer radius (mm) to 1000.

6. In the data tree, choose the NBVTO phase, go to the Data tab, and check the box
beside Phase fraction, indicating we wish to refine the phase fraction of NBVTO.
Then go to the Atoms tab. The atomic coordinates for the phase should look
like this:

1
CLS XRD & Scattering Summer School 2024 Na0.5Bi0.5V0.4Ti0.6O3
Rietveld Refinement with Constraints on Synchrotron Data

The atoms Bi1 and Na2 both reside on the same site, so we should constrain them
to be equivalent atoms. Similarly, the atoms V3 and Ti4 reside on one site, and
should be constrained to be equivalent. The atoms O5 and O6 are on different
sites, but should probably have a similar isotropic displacement parameter.

7. In the data tree, choose Constraints and go to the Phase tab. From the
pulldown menu, choose Edit constr. → Make atoms equivalent.

Select the first atom as Bi1 and the second atom as Na2. The program will note
that the parameters are not being varied, do you really want to keep the addition;
say Yes for all five constraints. The display should then look like this:

2
CLS XRD & Scattering Summer School 2024 Na0.5Bi0.5V0.4Ti0.6O3
Rietveld Refinement with Constraints on Synchrotron Data

We have constrained the changes in the three atomic coordinates (dAx, dAy, dAz)
and the isotropic displacement parameter (AUiso) to be equivalent for these two
atoms. The occupancies (Afrac) have been constrained to sum to 1 (full
occupancy).

8. Repeat the previous step to make the atoms V3 and Ti4 equivalent.

9. From the pulldown menu, choose Edit constr. → Add equivalence. For the
first variable, choose 0::AUiso:4 Atom O5 (the isotropic displacement
parameter for O5), and for the second variable, choose 0::AUiso:5 Atom O6
(the isotropic displacement parameter for O6). All the constraint should now
look like this:

3
CLS XRD & Scattering Summer School 2024 Na0.5Bi0.5V0.4Ti0.6O3
Rietveld Refinement with Constraints on Synchrotron Data
10. Choose File → Save Project As from the pulldown menu and save the project as
NBVTO.gpx in the working directory (C:\RIETVELD-JR\NBVTO-Rietveld).
We are now ready to begin the refinement.

B. Running the Rietveld Refinement

The following steps represent a reasonable order of refinement for the NBVTO data set.

Between steps, always (1) visually inspect the refinement and difference curve, (2)
choose Covariance in the data tree to check the goodness-of-fit (GOF, which should be
decreasing), plus to make sure the number of parameters being refined (Nvals) is correct,
(3) make sure your refined parameters look reasonable, and (4) check the GSAS-2
terminal to make sure the refinement converged.

11. From the pulldown menu, choose Calculate → Refine, to refine the Phase
Fraction (Scale) parameter and 16 Background parameters (17 parameters
total).

12. Refine the previous 17 parameters, plus the unit cell parameters (2), and the
Sample X displ. (20 parameters total).

When you Zoom in on the peaks, some are clearly broader than others,
suggesting anisotropic broadening.

4
CLS XRD & Scattering Summer School 2024 Na0.5Bi0.5V0.4Ti0.6O3
Rietveld Refinement with Constraints on Synchrotron Data
13. In the data tree, choose the NBVTO phase, and go to the Data tab. We are going to
try an anisotropic microstrain model. Change the Mustrain model to generalized,
and refine the S400, S004, S220 and S022 parameters, and press the Reset? button
for the Mustrain model to set the anisotropic coefficients at default starting values
(24 parameters total).

It is worth noting that for some lower symmetry crystal systems with more
anisotropic parameters, you should turn off the lattice parameters when you
begin refining the anisotropic microstrain parameters, to avoid parameter
correlation. We are okay to refine both lattice parameters and microstrain
parameters in this case.

14. Refine the previous 24 parameters, plus add the Atom parameters shown below
(11 parameters total).

Bi1 and Na2 are sufficiently different in mass (and electron count) that we can
refine their relative proportions by refining the occupancies with constraint. But
V3 and Ti4 are right next to one another on the periodic table, so it isn’t practical
to refine their occupancies with X-rays; they are too similar. With constraints, the
added atomic parameters make 31 parameters total.

The final refinement should have a GOF value of ~1.29 (under Covariance). The Bi1
and Na2 occupancies (frac) have changed marginally from 0.5, and all the Uiso
parameters have adjusted as well, but they are all still positive.

You could try refining the occupancies (frac) of the V3 and Ti4 atoms, to test what
happens, but they will not refine to sensible values. Vanadium and titanium are simply
too similar in scattering to refine with X-rays. There is a lot of information content in a
powder diffraction pattern, but there are also practical limits regarding what can be
refined.

5
CLS XRD & Scattering Summer School 2024 Na0.5Bi0.5V0.4Ti0.6O3
Rietveld Refinement with Constraints on Synchrotron Data

The final refinement for NBVTO should look fairly similar to the following plot:

Have questions?
Contact: Joel Reid
Industrial Scientist
Canadian Light Source Inc.
44 Innovation Boulevard
Saskatoon, SK, Canada
S7N 2V3

[email protected]

(306) 657-3854