|EXPGUI top||Previous page|
A common use for this is to adjust the lower 2theta/Q limit (maximum d-space/TOF) so that data where no reflections occur can be ignored. It is also common to start refinements with a somewhat limited range of data, raising the upper 2theta/Q limit (minimum d-space/TOF) later. (The ideal is to include all data where any peaks can be resolved in the final stages of refinement.) Note that there is a disadvantage to changing the data range, if one is using the Chebyschev polynomial (type 1 background function) to fit the background, as this function is scaled over the range of data. If the data range is changed, the background function will shift and thus must be refit each time the data range is changed.
A second purpose of EXCLEDT is to set and change excluded regions within the pattern. When a region is set as excluded, the data within that region are ignored with respect to the Rietveld refinement. Note that data should be removed only where there is a good explanation why the region cannot be accurately modeled. This is sometimes needed to deal with scattering from a sample container (although this can frequently be modeled as an additional phase, using Le Bail extraction for that phase when the container is badly textured). Excluded regions are also used occasionally when an impurity manifests itself with only a few minor peaks, although it is usually much better to model the impurity intensity as an additional phase (again using Le Bail extraction for that phase when the structure is unknown or the material has severe preferred orientation). Excluded regions are sometime needed when a section of pattern is affected by stacking faults. Another occasional appropriate use is to eliminate spurious intensity that is due instrumental malfunctions. It is important not to exclude peaks simply because they cannot be indexed. Such peaks may indicate a difference in symmetry or unit cell from the structural model, or some other error. It is also important that excluded regions be documented in the paper which describes the study.
The EXCLEDT program displays the data in a graph, (see the example to right) with the observed data as data points (black X's by default) and the computed values as a line (red by default). Any excluded data points are shown in a different color (orange, by default). Lines (again orange, by default) at the bottom of the plot show the allowed data range as well as any excluded regions. Note that if POWPREF has been run more recently then GENLES, the calculated values are zero. The EXCLEDT program can plot data in the native units (2-theta, TOF, energy), converted to d-space units, or in the most ideal unit for diffraction analysis, Q (2 pi/d).
In Zoom mode,
the cursor appears as
cross hairs, and the mouse is used to select a region to be magnified by
clicking on diagonal points in the region with the left mouse button
(see below) or "zoomed out" with the right mouse button.
In Add Region mode, the mouse is used to map out a region which that
will be excluded from the refinement. When this button is selected, the
mouse cursor changes to an arrow. After the left mouse is pressed once,
a yellow background is shown for the excluded region and when the left mouse
button is pressed a second time, the region is selected.
If the right mouse button is pressed instead, it resets the region selection.
Before the region is added to the excluded list, a message is displayed to confirm this. This confirmation message can optionally be eliminated (See Options/Prompt on add/del.)
In Delete Region mode, the mouse is used to select a region to be
deleted, causing the data in that region to be included in the refinement.
The cursor appears as a circle in the graph section of the window. After
a region has been selected to delete, a message is
displayed to confirm the region to be deleted from the exclusion list.
This confirmation message can optionally be eliminated (See
Options/Prompt on add/del.)
If the upper data limit is raised (or d-min ls lowered) this message is displayed. The reason is that EXCLEDT cannot display data above the upper data limit until the changed limit has been processed in POWPREF. If you also want to see the computed intensities in this region, you will need to run GENLES as well, so this is also offered as an option.
Note that it is possible to set the upper data limit (lower d-min limit)
well outside the range of data. If this is done, POWPREF will run slower
because many unneeded reflections are being included in the computation.
If this occurs and POWPREF is run inside EXCLEDT, a message such as
that to the right is shown. Unless there is some specific reason
(I do not know of one) it is probably best to accept this suggestion.
|EXPGUI top||Previous page|