[Wien] I/O-Error in dstart

[Peter Blaha]

> > You really want to calculate Cu with RMT=0.2 bohr  ???
>
> Not really, this was a "last resort". I did not know how else to get around
> the errors when running nn:
>
>   ERROR !!!!!!!!!!!!!!!
>  RMT( 19)=1.20000 AND RMT( 19)=1.20000
>  SUMS TO 2.40000 GT NNN-DIST= 0.41517
>
> Even now I'm not quite sure what this means. If nn calculates the distances to
> the nearest neighbors, what's the "distance" between Atom 19 and Atom 19
> supposed to mean? Or can such errors be safely ignored?

No, you cannot ignore that.

> The only reason for this would be an error in the icsd-coordinates.

Maybe, maybe not.

a) Do you know that for those rhombohedral lattices you must give a "strange"
unconventional input: lattice parameters in hexagonal settings, positions
in rhombohedral setting ???
For sure this is NOT the way the data are stored in ICSD, but you must convert
them (see struct file description in UG)

b) Sometimes the crystallographers "see" a certain position, which, however,
is occupied only fractional.
Are those Cu positions fully occupied ?  Is your compound  (B_113 Cu_20) ?
At least this is the composition in your struct file.

Regards

                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------