This example uses example Cimetidine data from the EXPO software
Normally we would want to use Fundamental Parameters peak profiling. But as this is
synchrotron data (single wavelength and parallel beam), we should be able to get a
way with just using Pseudo-Voight profiling (plus XFIT cannot do Fundamental Parameters
profiling on Synchrotron data)
Open up the file and browse the pattern noting the range which gives fairly resolved peaks for profiling. At least 20 peaks are required. Using the File, Details option, set the background to refine with a bit more flexibility (3rd to 5th order). Selecting Ins/Del Peaks, then PV for Pseudo Voight, click left button on the peaks to profile. Right mouse button to remove a peak. If there are subtle shoulders or humps you are not sure about, ignore them for the moment. Come back to the shoulders if a single Pseudo-Voight peak cannot fit them. Also, if the width of a peak is enlarged compared to its neighbours, this could imply there are really two peaks. In the following screen dump, we have selected our range.
|
For convenience, now fit small sections of the powder pattern. Taking note
of any misfits which might indicate there is another peak present. At low
angle, these could be due to aberattions caused by sample packing and
geometry so be careful. It never hurts to run a standard to check what the
default profiles look like, and to check the alignment of the XRD (as well as
consider also running an extra sample with an internal standard
so the data can be validated against the standards. This type of
checking should be considered mandatory, not optional. This can help avoid
a lot of un-necessary pain with structure solution.
Note the low angle tail on some of the major peaks in the following screen dump. This could be legitimate peaks or an abberation of the XRD. For the moment, it will be treated as a machine abberation. If the indexing programs give extra "calculated" peaks at these points, it could add extra confidence to the indexing. This also shows why it is important to run standards of known characterstics (NBS/NIST Silicon, annealled Cubic Y2O3) so you can be more confident what is due to the XRD machine, and what is due to the sample.
|
When a small peak or shoulder is close to another larger peak (or when there
are a group of overlapped peaks), it can be effective to constrain the shapes and
widths to be equal. In XFIT, select Ins/Del Peaks, from the Peak
Edit Options select Code and you can now graphically link refinement
codewords together.
In the following screen image, the shape of two peaks are being linked together by giving them the same name (a letter must be the first character defining the codeword). Just click on the "@" to change it. The same goes with any value on the screen. (this can also be done via the "File Details" menu but this can be quicker for this type of local setting of constraints.
Check that the constraints are working and the widths, then the shapes are identical.
|
Once you are happy with the peak profiling, save the project. Then go into File, Details,
then Peaks, Two Theta All Pk Types; and create a text report saving the file as XFIT.TXT
ready for importing into the Crysfire powder diffraction indexing suite. To be prompted to save the file, close
the TXT Report window. Consider putting the file into a separate subdirectory called
crys1. It may be necessary to go back and re-analyse the diffraction data if indexing is
not satisfactory, so keeping things in separate subdirectories can limit the risk of "possibly"
good results getting overwritten.
|