Hello all
In reply to the question, "How far off the position of an atom may
be and the "refinement" still brings it in?", I would like to cite an
example of mine where a Rietveld analysis located the true positions
of atoms when our initial guess was clearly not correct. The case
was the low-temperature monoclinic phase of WO3. This has a
perovskite structure. Earlier X-ray diffraction had given rough
coordinates of the W atoms, but because of twinning the crystals
were poor and the oxygen atoms could not be located. In our
Rietveld refinement of our ISIS HRPD (nueton) data we put the W atoms
where the X-ray difraction had suggested, and we put the O atoms
where they would be if they had not displaced from their positions
in the cubic parent structure. It was clear that this was wrong,
not least because some of the low-hkl peaks were strong in the data
but vitually of zero intensity in this starting model. We first
worried about this large discrepancy between the calculation of
the diffraction pattern and the data, and initially tried to be
clever, but in the end we just tried letting all the coordinates
go all at once, and the Rietveld refinement quickly and easily
(and very dramatically I thought) moved the oxygen atoms into their
proper places and gave us all the missing intensity. In the end
we didn't need to be at all clever - the Rietveld code did it all
for us. For anyone interested, the final write-up is in
Crystal structure and paramagnetic behaviour of e-WO3�x.
EKH Salje, S Rehmann, F Pobell, D Morris, KS Knight,
T Herrmannsdorfer and M T Dove.
Journal of Physics: Condensed Matter 9, 6563�6577, 1997
I think the point is that Rietveld can be used as more than a
tool to simply tweak the final answer, since it can give you the
missing bits when you know only part of the answer. Possibly the
case where this happens most is when you have a phase transition
and need to determine the atomic displacements in the lower-symmetry
phase, when all you know is the structure of the higher-symmetry
phase.
Martin Dove
--
-----------------------------------------------------------------------
Martin Dove
Mineral Physics group, Department of Earth Sciences,
University of Cambridge, Downing Street, Cambridge CB2 3EQ
Tel. 01223-333482 or 01223-740075. Sec: 01223-333459. Fax: 01223-333450
email: [EMAIL PROTECTED] WWW: http://www.esc.cam.ac.uk/astaff/dove
-----------------------------------------------------------------------
