The ABINIT Users Mailing List ( CLOSED )

[verstraete@pcpm.ucl.ac.be]

hello Gustavo,

The convergence for energy derivatives is always worse, but this seems 
ridiculous. You should however try with Trouiller-Martins psps. They are 
much softer than the HGH ones (which are notoriously hard). Do you have 
experimental values to compare with? Are you actually converging to 
something reasonable? You should also send in your output/log file 
extracts as attachments, and, more importantly, check them for convergence 
problems and/or warnings: the completely wrong values for ecut 40 seem to 
indicate a fluke somewhere.

Are the phonons near experimental values? If the lattice constants and the 
phonons are good for ecut=80, the elastic constants shouldn't need 400 Ha.

There may also be issues with the non-linear core correction, which is 
essential for Na. I think this has been coded and tested by Don Hamann, 
but check what is done in your pseudopotential. If it includes semicore 
2s2p electrons instead of the NLCC, you may be up against an authentic 
problem, with a pseudopotential which really does requires 200 Ha.


Matthieu

--
===================================================================
Matthieu Verstraete               mailto:verstraete@pcpm.ucl.ac.be 
PCPM, Boltzmann, pl. Croix du Sud, 1            tel: 010/ 47 33 59
B-1348 Louvain-la-Neuve  Belgium                fax: 010/ 47 34 52