[sdpd] Interesting (?) Problem

["Rob Hart" <rohart...@indiana.edu>]

Dear Rietvelders-
I?m working on tellurite-niobate glasses for my thesis and as a test
crystal to use for comparison, I came across KNbTeO6.  The original
report of its diffraction pattern was a cubic pyrochlore in space group
Fd3m with a=10.245, circa 1970.  I?ve been able to get powder neutron
and x-ray data and done a refinement to find a=10.248, Te and Nb sharing
the 16c sites (49 and 51%), K in the 8b sites and O at 0.317, 1/8, 1/8
in 48f sites.  The average Rp for the three histograms in this case is
0.93.
But I also wondered if the metal ions couldn?t be layered, with all the
Telluriums at z=0 and ½ and all the Niobiums at z=1/4 and ¾.  That
refines in space group Imma with a=7.249, b=7.240 and c=10.246.  That
also refines nicely, but allows a little bit of variation in the Nb-O
bond lengths, like a SOJT distortion, which is seen in many other
niobates.  The final Rp for this refinement is 0.88.  
Being new to the Rietveld method, I don?t know what this really means.
Is the structure really layered, and the good fit in cubic is an
artifact of that?  Or is it that the two models are actually equivalent
for some reason that I don?t appreciate?  
It seems to me that this question could be answered by exafs at either
the Nb or Te edge, since the two models would predict different partial
g(r)?s.   I also thought to calculate the solid state NMR parameters of
the Nb using AM1 functional or so, then compare to experiment (since I
have access to an NMR spectrometer tomorrow, but not to a synchrotron
until July).  Any comments or ideas are welcome as well as advice on how
to proceed from here.
Thanks a lot-
Rob Hart



 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/