The ABINIT Users Mailing List ( CLOSED )

["Josef W. Zwanziger" <jzwanzig@dal.ca>]

Dear Colleagues,

I have been using ABINIT for some months now, with pleasure and
enthusiasm, and have also used the fhi98PP code to construct my own
pseudopotentials. I have made working and evidently accurate
pseudopotentials of Troullier-Martin type with PBE GGA, for Te-52, O-8,
and O-8 with a 1s hole (to simulate XPS spectra, which worked really
well by the way). But I am having a lot of trouble with Pb-82. The
valence configuration of lead is 6s2-6p2, and in fact the starter file
for fhi98PP has this configuration (actually, it's listed there as
6s2-6d2, which I assume is a typo) and the abinit LDA PS for lead has
this configuration too, but the fhi98PP checks show multiple ghost
states. The only configuration I can get to work is 5d10-6s2-6p2, with
lmax = 2 and lloc = 0 or 1. And, the lloc = 1 case seems to give subtler
problems. I notice that the Vanderbilt USPP configuration is also
5d10-6s2-6p2, and within CASTEP, the TM PS is of this configuration as
well. I assume I am re-inventing the wheel here, do any of you have
other experiences with lead, and also, can you give advice as to how to
choose the lloc parameter in generating a PS?

Thanks much for your help,

Joe Zwanziger

Josef W. Zwanziger             tel: (902) 494-1960
Professor of Chemistry and     fax: (902) 494-1867
  Canada Research Chair in     net: jzwanzig@dal.ca
  NMR Studies of Materials     web: http://jwz.chem.dal.ca
Dept. of Chemistry
Dalhousie University
Halifax, NS 
B3H 4J3 CANADA