>One cure for the problem would be a short program that reads the [binary]
>reflection file and rewrites it with pseudorandom F(obs) values.
The important point is the initial intensity ratio given at, say, 2 exactly
overlapping reflections : this ratio will be kept at all the further
iterations of
the Rietveld decomposition formula even if the total intensity is rescaled. The
kind of dummy atom does not matter, hydrogen or plutonium will lead to the
same result, of.course, only the initial coordinates are important. Moreover,
those intensities starting at a zero intensity, will stay at zero intensity,
so that
if you put a dummy atom in a special position corresponding to special
extinctions, then a lot of reflections will not be accounted for. So, to cure
the problem, the best is to start with a set of |F| having all the same abitrary
value. OK, several treatments of the equipartitioned |Fobs| have been
proposed that could be applied further in order to improve the chances
to determine a structure from powder data, but this is another question.
>There are some advantages, though for the method used in GSAS -- it can be
>helpful for fitting an impurity phase with preferred orientation (as happens
>here when we use Al cans or don't mask properly) or when computing Fourier maps
>where one has a structure that is partly correct.
I think that all Rietveld programs should adopt this way to estimate |Fobs| at
the end of a refinement. Not only perform one time the Rietveld decomposition
formula but iterate it up to the minimum Rp. Then Fourier difference synthesis
may be more clear. Only one application of the decomposition formula is
generally unsufficient for a good fit. Of course the problem of overlapping
reflections remains : their intensity ratio will exactly correspond to the ratio
of the |Fcalc|, but the total intensity will be rescaled. If somebody is
interested
to submit a publication on this subject to J. Appl. Cryst. , I think it will be
accepted without difficulty :-).
Armel
Armel Le Bail, Laboratoire des Fluorures URA CNRS 449
Faculte des Sciences, Universite du Maine, F-72017 Le Mans, France
http://fluo.univ-lemans.fr:8001/ and/or http://pcb4122.univ-lemans.fr/