When a sequential fit is performed, the results from that fit are placed into a new data tree entry, the Sequential Results, where the associated data window shows a table with the results from the fit. Each histogram has a separate row. An example of this is shown in Fig. 21.5. The first few columns in the table show values that track the refinement progress and sample parameters that change between histograms. The \(\Delta \chi ^2\) value is of particular value to know if the refinement has converged. This shows the change in the overall \(\chi ^2\) value before and after the last refinement as percentage. If this number is more than a percent or so, the refinement is still showing significant changes. If very large, you may want to repeat the refinement before varying additional parameters. For the final run, you want al; these values to be well under a percent. Larger values are highlighted in color to make them more obvious.
Later columns of the table show lattice parameters, when these are refined, and then later columns show refined parameters. Note that the “Hide columns...” command in the “Columns/Rows” menu is used to select the refined parameters that are included.
There are a lot of tools programmed for use with this table. Note that allowing the mouse to “rest” (allow the mouse to stay in one position for a few seconds) on a column heading will display some information about that column; allowing it to rest on a value in the table, will show that s.u. value, if one exists. Clicking on a column plots the values in that column. Selecting more than one column will plot those values together. The default x-axis values are the sequence numbers for each histogram, but pressing the “s” key in the plot (or using the “K” button on the plot toolbar and then selecting the “s” option) causes a list of columns to be shown. By selecting, for example, the temperature column, values can be plotted against temperature. Clicking on a row heading causes the histogram fit to be displayed and right-clicking on the heading shows the correlation matrix plot.
The “Pseudo Vars” menu commands allow you to compute values from the fitted parameters, which can be a distance between atoms or a formula that you specify, such as the ratio of lattice parameters. An example of this is in the “Parametric Fitting and Pseudo Variables for Sequential Fits” (https://advancedphotonsource.github.io/GSAS-II-tutorials/SeqParametric/ParametricFitting.htm) tutorial. These computed values are added to the table and are recomputed each time the table is shown. The s.u. for these values is computed. These values can also be used as the x-values for plotting. So, for example, should you wish to plot against temperatures in Celsius rather than Kelvin, you can create a pseudo-var that subtracts 273.15 from the temperature and then use that for plotting.
You can also use the “Parametric Fit” menu commands to define an equation that will be fitted to the values in the sequential table. This is also demonstrated in the “Parametric Fitting and Pseudo Variables for Sequential Fits” (https://advancedphotonsource.github.io/GSAS-II-tutorials/SeqParametric/ParametricFitting.htm) tutorial. Uncertainties in parameter values, as well as the parameters’ covariance are taken into account during fitting.
Note that if histograms are removed from the sequential dataset list, their results are not removed from the table. Thus, this table can be built up in sections. However, if one want to omit values from a histogram from being used in calculations and plots, unselect the “Use” flag in the second column. That will effectively remove that row from consideration. It is recommended that after all histograms have been successfully refined, a final fit be performed that includes all histograms, but this should not actually change the results in any meaningful way.
