[Ifeffit] Errors from Webatoms?

[Michel Schlegel]

Hello Feng,

> I used WEBATOMS(version 1.8), to calculated the feff.inp for Si-K of 
> alpha-SiO2 (quartz phase), using the crystal structure (as listed below) 
> from Wyckoff's Crsytal Structures(P312-313), however, I could not get 
> correct atomic configuration, such as,  I did not get the tetrahedron (of 4 
> nearest neighboring oxygen atoms), I got two n.n. oxygen atoms only, please 
> refer to the results on the bottom.
> As suggested by Dr. Ravel, I tried the atoms in Artemis, but still I could 
> not get correct feff.inp.Would you please help me to check and suggest, 
> thank you so much.

The answer to your question lies in one of those cristallographic 
oddeness, namely that you can describe the crystallographic structure of 
quartz in two distinct sets of axis, the  obverse (a,b,c) and the 
reverse (a', b', c), with the latter being obtained by rotation of the 
former by pi/3 around c. The caveat is that the cristalloraphic data you 
used are relevant to the reverse axes (and usually is described so in 
the litterature), whereas the P31 2 1 is relevant to the obverse. Thus 
you have to make a slight coordinate change;

we had to deal with that before (Schlegel et al, Geochim. Cosmochim. 
Acta, vol. 66, p 3037, 2002), and we came up with the following 
cristallographic data

P 32 2 1
a = 4.9137
b = 5.4047

x(O) = 0.5867
y(O) = 0.7328
Z(O) = 0.7855

x(Si) = 0.5303
y(Si) = 0
z(Si) = 2/3

As a structural check, all Si-O distances must be within 1.60-1.62

angstroms (160-162 pm for purists)
Hope this will help; let me know if not

Best regards,

Michel