I also spent much time in trying to fit yttria-stabilized C+T zirconia
using a Rietveld program (SIROQUANT) and using a peak fit program
(PROFILE by SIEMENS/AXS). The main problem is, that you usually have
asymmetric and strongly overlapped peaks. Because of the asymmetric
peaks, you are inclined to use some kind of splitted function to get =
low
R's. But due to the strong overlap, the results tend to be unstable or
just nonsense.=20
For simple profile fitting on the (220)T/(400)C/(004)T peak cluster the
not splitted PearsonVII function turned out to be a good compromise. It
works well because at 2theta =3D 75=B0 my peaks are already rather
symmetric. 10% higher R's than with asymmetric PearsonVII, but far more
reasonable results.=20
For the Rietveld work, I have no workaround. As soon as I give free the
asymmetric parameters, the fit is getting unstable. Results of
quantitative phase analysis are far off reality. I gave up fitting
mixtures of phases with similar lattice constants using Rietveld. But
maybe, someone has made better experiences. I'm very much interested in
your suggestions.
Thomas
> ----------
> Von: beth fancio[SMTP:bethfan@hotmail.com]
> Gesendet: Montag, 29. September 1997 19:39
> An: rietveld_l@ill.fr
> Betreff: <none>
>=20
> Hi Rietveld users,
> I'm working in Rietveld refinements(dbws9411 version) for =
quantitative
>=20
> phase analysis of Zro2 at various concentrations of y2o3=20
> (2%mol-8.5%mol). The convergence for pseudo-Voight function in the =
T+C
>=20
> cases(4.5%-7%mol y203) is giving NA~2.
> I'm trying to apply the split-PearsonVII function, but the outputs =
are
>=20
> just errors messages for the m values.
> I need some suggestions for the initial values of NA,NB & NC, and
> other=20
> possible errors that could arise in the application of this function
>=20
> ______________________________________________________
> Get Your Private, Free Email at http://www.hotmail.com
>