17.6 Sequential Peak Fitting

If you have a number of similar datasets and want to perform a peak fit to all of them, this can be done as well. This might be desired if your only interest is to compare the relative intensities of a few peaks, or look at how the position of peak changes across a parametric experiment, where the data are not sufficient for a full Rietveld fit, or a full fit is not needed. To do this in a fairly simple way, GSAS-II provides sequential peak fitting.

To perform a sequential fit, you want to have a project where all the data have been imported. Also, you are best off performing a peak fit with one histogram. This is commonly the first histogram, but could be any histogram in the project. You then want to copy the following settings from that histogram to the rest of the histograms:

• Copy the limits. This is done by selecting the Limits data tree entry for the fitted histogram and then use the “Edit Limits”/Copy menu command to copy limits from the selected histogram to the others.
• Copy the background: This is done by selecting the Background data tree entry for the fitted histogram and then use the “Background”/Copy menu command to copy the background settings from the selected histogram to the others.
• Copy the peaks: For this select the “Peak List” data tree entry for the fitted histogram. Before you copy settings, you will likely want to turn off some of the refinement flags. For example, perhaps start with refining peak intensities only and refine more parameters later. Use the “Peak Fitting”/“Peak copy” menu command to copy the entire peak table from the selected histogram to all others.

At this point you can use the “Peak List”/“Seq Peakfit” menu command to start a fit to a series of histograms. You will be asked which histograms to include in the series. Note that there are two additional options available on the window where you select histograms to be fit:

• Copy to next: When this is selected, the refinement results from one histogram are copied to the next as the refinement starting point. This is very useful for parametric work where one seting of starting points would not be good for all histograms, but not very much is changing between histograms that are adjacent in the list.
• Reverse order: Normally histograms are process in a top-to-bottom order with respect to the histogram list, though only the histograms selected in the window are used for peak fitting. When this option is selected, the order is bottom-to-top.

After the histograms have been selected and OK is pressed, a peak fit is performed on all selected histograms and the results are displayed in a new data tree entry labeled “Sequential peak fit results.” This shows the fit metrics (\(\rm R_{wp}\) and GOF) as well as the refined terms. Clicking on a row in the table shows the fit in the graphics window.

To continue with more fitted parameters, select the “Peak List” data tree entry for any histogram that is being fit and select the refinement flags values to use. Then use the “Peak Fitting”/“Peak flags copy” menu command to copy the refinement flags in the current peak table to the other selected histograms. Then use the “Peak List”/“Seq Peakfit” menu command again to repeat the peak fit. Note that you do not want to use the “Copy to next” option, as this would overwrite your previous fit results. You likely want to repeat this step until you have included all peak parameters that you would want to fit.