• View
• Edit
• History
• Print

GlobalRefinement

Rietveld Refinement

Refinement of parameters with Maud

In the main Maud window, the parameters to refine are located in the main window, below the plots. You can force the refinement of a parameter by clicking on Fixed, it should bring up a menu to set it to Refined...

Incident intensity, detector distance, and main backgrounds

As a first step, we will refine:

• the incident intensity
• the sample to detector distance
• the main background paramaters, under dataFileSet_x, background polynomial coeff 0, background polynomial coeff 1 and

background polynomial coeff 2.

Once you have made those parameters as "Refinable", hit the hammer icon (on top of the window), and you'll see the refinement results.

Once the refinement is finished, in the bottom panel, you can see the new value of the parameters, and an error. If this error is equal to -1, there was an error: Maud is not able to refine this parameter.

You can plot the result in many way:

• by going back into the dataset, selecting a datafile, and hitting View.
• in the main window, by using the Plot 2D function: it will show you the original dataset (below) and the recalculated one (above).

At each step, you should always check that you fit is Ok and that Maud did not get out of hands. Save frequently, and use several kind of plots, in many orientations, to make sure everything is Ok...

Incident intensity and peak profiles

Next step is to get the peak profile to look a little bit more like the data:

• Fix all backgrounds
• Set the following parameters for refinement:
  • incident intensity
  • caglioti coeff 0
  • caglioti coeff 1
  • caglioti coeff 2
  • sample to detector distance

and run a refinement.

It should look a lot nicer. Depending on your data, you could also play with the caglioti and gaussian coefficients. It does not matter, as long as you get a nice match between your data and the fit.

Individual backgrounds

At this point, it is usefull to add background parameters for each spectra. I usually have 3 parameters for each one. To add those,

• go into the dataset window,
• select all datasets,
• click on Add bkg par three times.

Now we want to refine all those background... You could go and make them refineable one by one, but it would take for ever!! There is an easier way,

• in the Refinement menu (in the main menu bar, on top of the Maud window), select Parameter List.
• in the Parameter List window, select Fix all parameters
• select Free backgrounds
• in the table, you should see all backgrounds set to Refined
• launch a refinement (the hammer icon), it could take some time...

Your fit should get better and better!

Detector re-calibration

Now, we'll make sure that our detector calibration is Ok. In the parameter list window

• fix all parameters
• go into sample X > dataFileSet_X > Diffraction Instrument > Flat image transmission
• set the sample to detector distance, beam center, and tilt parameters to be refineable.
• start a refinement.

Your fit should get really good by now!

Material properties

At this point, you can start refining sample properties, fix everything else. In my experience I find that to get a perfect fit, you have to play with the crystallite size and microstrains (under isotropic).

Finally crystallographic parameters (talk with a Rietveld expert) should get you to the end...

Putting it all together

At this point, there is no reciepe:

• play around and refine various parameters
• avoid refining to many parameters at the same time
• avoid refining conflicting parameters together
• save frequently

until you get the perfect fit!!!

In my experience, for a calibration image:

• refine background and beam intensity together,
• refine peak profile and sample properties (crystallite size and microstrains) together.

If you can't get it right, it can be usefull to go back to the manual adjustment, from the Dataset window, View one of them, in the Change parameter in the Tool menu. You can use this function to see what is the effect of each parameter and figure out what to do next...