For the original query - what about doing a Le Bail fit
to get the best profile fit without being biased by the
structure. Then lock the profile and refine on the structural
parameters? This might be a little tricky with GSAS as I
don't believe it recycles the HKL file like Fullprof does.
But with a good bit of damping - it should work OK.
> I am following with interest the discussion you started on the modeling of
> low angle XRD peaks. I, too, have had difficulty modeling data with our
> inhouse Philips diffractometer (even with function #3 as suggested to me by
> Bob von Dreele) and have begun to question Philips directly about the
> possibility of alignment problems. On similar samples, data refinement of
> Scintag data proceeded smoothly, so this again points to a problem with the
> diffractometer itself. May I ask what type of machine you are using?
>
> Thanks and good luck!
> Rick Jacubinas
While the Scintag I have used in the past gives wonderfully
nice peak shapes that are easy to fit, a well aligned Philips
XRD is usually a top workhorse machine and should be able
to do high quality Rietveld without too much of a problem -
especially with the new algorithms for handling peak assymmetry.
But the older 1050 style are definitely not fun to align unless
you are expert at it. After proper alignment (which was usually
done as part of a Philips Service Agreement), I am used to a
Philips lasting for at least a year before requiring another
alignment - providing it is not getting knocked around.
Philips used to sell a "gunsight" alignment kit that could
make aligning 1050 style goniometers much easier than using
rulers and the more primative jigs they supplied as a default.
For characterising the alignment of the XRD:
Have you run Silicon with peak profiling and unit-cell
refinement packages to check the alignment?
This can take a bit of time to do properly but there is
much software around to help - some being
HKLGEN or Dragon to Generate an HKL list
XFIT or Winfit for peak profiling
Xlat, Unitcell, Eracell or Refcell for Unit Cell
refinement.
For trying to validate various slit and geometry settings a
fundamental parameters program such as XFIT/Koalariet or BGMN
could be very useful. One of the more frustrating things
about Fundamental Parameters fitting being it can show
when a machine is out of kilter when other software
will happy cover this up. But this is very handy if you
are interested in alignment.
For checking the XRD is giving accurate intensities,
you could try running NIST LaB6 and doing a
Rietveld structure refinement. Then check the thermal
parameters to see that they have reasonable values.
Other suggestions for testing alignment appreciated.
Lachlan
------
Lachlan M. D. Cranswick
4th Jan - 16th March 1999 Visiting Scientist
CSIRO Division of Minerals, Melbourne, Australia
Tel: (613) 9545 8500 (switch)
Fax: (613) 9562 8919 [EMAIL PROTECTED]
Collaborative Computational Project No 14 (CCP14)
for Single Crystal and Powder Diffraction
Daresbury Laboratory, Warrington, WA4 4AD U.K
Tel: +44-1925-603703 Fax: +44-1925-603124
E-mail: [EMAIL PROTECTED] Ext: 3703 Room C14
NEW CCP14 Web Domain (Under heavy construction):
http://www.ccp14.ac.uk
