More Complicated Self Calibration

Please note that these run-throughs do not in general go into "best practice" indexing methods but mainly demonstrate how to start interacting with the CRYSFIRE software. The tutorials make use of the self calibration option in CRYSFIRE purely to show its functionality (Self Calibration is to Powder Diffraction as is DIFABS to single crystal methods?). It is far better to make use of a well aligned and maintained Powder X-ray diffractometer and/or use of an appropriate internal position standard such as NBS/NIST Silicon 640(a-c) to correct for any systematic offsets that may have leaked into the experiment. Even minor systematic offset errors in the peak positions can cause all the indexing packages to fail in suggesting the correct result.

This tutorial will go through a more involved example which has many line pairs that could be used to help select a value for self calibration of the line positions. This is based on a data file from Armel Le Bail.

If running under Windows, go into a DOS Prompt and create/enter the appropriate subdirectory where you wish to perform the indexing. Run Crysfire, and load in the required CDT Crysfire file using the LO command (or import some raw data using the IM command)

Type SC for self calibration to be given the following menu options. The default is to do a "Zero" correction (Z), which we will go with by pressing [Enter]. Go with the defaults by pressing [Enter] until you obtain a histogram with 2 peaks (a major one with 7 pairs at -0.03 and a minor one with 4 pairs at +0.06). We'll use the major one and input it by hand in a moment. At this stage the mean and esd (33 pairs) are -0.00156 +- 0.05833 (i.e. this is not very meaningful yet)

So select S to manually provide a 2-theta offset and enter -0.03 to set an approx working Z2Theta of -0.03 deg. Go with the defaults but when prompted, don't save the changed dataset by typing N.

Run Self Calibrate again by typing SC. Then reselect Z for 2-theta offset correction. Enter a Max Q difference of 5. This tighter max-Q-diff limit now leaves only the major peak in the histogram, 7 pairs at an correction change dZ2Th of 0.00. The mean & esd (22 pairs) is now -0.02890 +- 0.02474. This is not bad, but we can improve it by tightening the limits to define the histogram peak better.

So keep pressing [Enter] until you read the Select H, L, D, M .. prompt. Now select D (Input max Q-difference for including line pairs in self-calibration) and enter 2 .. giving a symmetrical peak (7 pairs) at dZ2Th= 0.00 and a mean & esd (15 pairs) of -0.02960 +- 0.00752,

We will adopt this -0.02960 value by selecting M

Now when prompted, we will save the resulting pattern as suggested under the name BID8PSCZ (editing the title slightly to remove the truncated remains of "No Z2Th").

Now you can carry on with the indexing of this sample.